Trailing-Edge
-
PDP-10 Archives
-
decuslib20-05
-
decus/20-0141/fasp.doc
There are 2 other files named fasp.doc in the archive. Click here to see a list.
FASP User's Guide
(FORTRAN Alphameric Subroutine Package)
24 June 1978
FFFFFFFFFFFFFF AA SSSSSSSSSS PPPPPPPPPP
FFFF AAAA SSSSS PPPP PPPPP
FFFF AAAAAA SSSS PPPP PPPP
FFFF AAAA AA SSSSS PPPP PPPP
FFFF AAAA AA SSSSSS PPPP PPPPP
FFFFFFFFF AAAA AA SSSSS PPPPPPPPPP
FFFF AAAAAAAAAAAAAA SSSS PPPP
FFFF AAAA AA SSSSS PPPP
FFFF AAAA AA SSSSSSSSSS PPPP
UU UU SSSSSS EEEEEEEE RRRRRRR ''' SSSSSS
UU UU SS EE RR RR '' SS
UU UU SS EE RR RR '' SS
UU UU SSSS EEEEEE RRRRRRR SSSS
UU UU SS EE RR RR SS
UU UU SS EE RR RR SS
UUUUU SSSSSS EEEEEEEE RR RR SSSSSS
GGGGGGG UU UU IIIIII DDDDDDD EEEEEEEEE
GG UU UU II DD DD EE
GG UU UU II DD DD EE
GG GGGG UU UU II DD DD EEEEEE
GG GG UU UU II DD DD EE
GG GG UU UU II DD DD EE
GGGGGGG UUUUU IIIIII DDDDDDD EEEEEEEEE
Donald E. Barth
Division of Computer Services
Graduate School of Business Administration
Harvard University
An Introduction to the FORTRAN Alphameric Subroutine Package
-- ------------ -- --- ------- ---------- ---------- -------
FASP, a FORTRAN Alphameric Subroutine Package, is a package
of FORTRAN subroutines which provide alphameric input/output
features not provided directly by FORTRAN. Most of the
subroutines in this package either write characters with a
multiple of an A1 format, or else generate, manipulate or
interpret characters which the calling program has read in
with or will write out with a multiple of an A1 format. The
following types of routines are included in this package.
A. Routines which evaluate the words, numbers or
combinations of these which the user has typed.
B. Routines which generate the representations of numbers.
C. Routines which allow the user to specify by name and
subscripts, to examine and to manipulate the values of
multiply subscripted FORTRAN arrays which are
equivalenced with or otherwise loaded into singly
subscripted buffers.
D. Routines which perform tab character to space
conversions, lower case letter to upper case
conversions, and other manipulations of text.
E. Routines which generate bar charts, point and line
plots, vertical time axis plots and pin maps.
F. Routines which generate large multiple-line lettering.
This documentation and all programs and routines described
in this documentation were written at PDP-10 (DECsystem-10)
installations at Aiken Computation Laboratory of Harvard
University, and at the Harvard Business School by Donald E.
Barth, who can be contacted at the following address
Division of Computer Services, Baker Library 21
Graduate School of Business Administration
Harvard University, Soldiers Field
Boston, Massachusetts 02163
The author requests that the comment lines stating his name
be retained in the routines and programs which form this
package and which are described in this document.
The routines described in this documentation were written
for specific applications in which these routines appear to
produce the expected results. However, FASP includes over
6000 FORTRAN statements, and the routines in the package
have up to 37 arguments each in their calling sequences.
Any package of this size and complexity will contain errors,
and cannot produce reasonable results for all possible
illogical values of the input arguments. Neither the
author, nor Harvard University, assume any responsibility
for errors in the documentation of this package, for
malfunctions of the routines within this package, or for
difficulties which may arise in the use of these routines.
FASP, FORTRAN Alphameric Subroutine Package Page 2
An Introduction to the FORTRAN Alphameric Subroutine Package
This manual contains a collection of 1 line descriptions of
each of the routines in this package, followed by a
collection of 1 paragraph descriptions of each of the
routines, and finally a collection of detailed multiple-page
descriptions of each of the routines. Each of these
collections is arranged in the alphabetical order of the
routine names. Each detailed description contains a summary
of the purpose of the routine, a listing of the SUBROUTINE
statement as found in the routine and of the DIMENSION
statement if any, and a description of each of the variables
and arrays in the order in which they appear in the argument
list. A few of the routines in this package which have
argument lists which include all of the arguments of the
primitives which they call are described in the
documentation of these primitives, rather than separately so
as to prevent duplication of the descriptions of the
identical arguments. The 1 paragraph descriptions specify
the locations of the detailed descriptions of such wrapper
routines and similarly specify the locations of the detailed
descriptions of the reduced capability versions which are
supplied as alternatives for some of the routines. At the
end of the detailed description will usually be found a
listing of a program which calls the routine and a listing
of a typical program-user dialog. Such calling programs
range from simple interactive demonstrations to the
prototypes of the application programs for which the
routines were written.
In the argument lists of most of the routines described in
this documentation, input arguments which are returned
unchanged appear first, arguments which are used both for
input to this routine and for output from this routine
either to the calling program or to the subsequent call to
the same routine appear next, and arguments for which the
input values are ignored and which are used only for output
appear last. The argument lists of a few of the older
routines in this package are not arranged in this manner,
but these argument lists are usually short anyway.
Arguments having the same names in different routines often
have similar definitions. However, such definitions can
vary in obscure ways. For example, the argument named
LOWBFR is returned pointing to the character to the right of
an illegal character by DASPAN and DANEXT, but is returned
pointing to the illegal character itself by most of the
other routines.
The programs and routines in this package are written in
machine independent FORTRAN but do require that the compiler
support the 1H notation in DATA statements to define 1
character per array location, and that the runtime system be
able to use the A1 format to read a single character into a
single array location or to write out an array location
containing such a single character. The fill characters to
pad to the full word size of the particular computer being
FASP, FORTRAN Alphameric Subroutine Package Page 3
An Introduction to the FORTRAN Alphameric Subroutine Package
used are of no concern except that the fill characters
supplied during reading with an A1 format must match those
supplied when the 1H notation is used in DATA statements.
If the routines in this package are used on some IBM 370
systems in which the 1H notation in DATA statements results
in padding with zero character codes while characters read
with a multiple of an A1 format are padded with space
characters, then it will be necessary to change the
character definitions in the DATA statements in these
routines from 1H notation to 4H notation and to insert 3
extra spaces to the right of each character so defined.
Most of the arguments of these routines have names which
begin with the letters I through N and must contain either
integer values or else characters which have each been read
by the use of an A1 format or which have been defined using
a 1H notation. Those few arguments which have names
beginning with the letters A through H or else with the
letters O through Z are used for real values. None of the
machine-independent routines in this package store more than
a single character per computer location. A few of the
machine-dependent demonstration programs and routines which
are listed as examples in this document use single and
double precision variables to contain several characters
which have been packed into individually addressable
computer locations, but no special naming conventions have
been used for these variables.
Most of the routines in this package were developed using
the DECsystem-10 F40 compiler. The DIMENSION and DATA
statements have since been reordered to conform to the
somewhat stricter sequencing rules enforced by the newer
DECsystem-10 FORTRAN-10 compiler and this compiler has been
used for all of the more recent development of the package.
The use of the FORTRAN-10 compiler has allowed the DATA
statement in the DAFLAG routine which defines the characters
1H[ and 1H] to be changed from octal machine word notation
to Hollerith character notation. The octal notation DATA
statement has been commented out, but must be restored if
this routine is to be compiled using the F40 compiler. The
lower case letters which appear in Hollerith strings in a
few of the routines will be translated into upper case if
the F40 compiler is used, but this will cause no difficulty
other than that the routines will then be unable to process
lower case input.
Although the routines in the package use the 1H Hollerith
notation in DATA statements to define variables which are to
be matched against characters which the calling program has
read 1 to a computer storage location using a multiple of an
A1 format, the array arguments used for passing characters
to the routines are not used for any other numeric
information so that the routines in this package should be
easily converted to the character conventions of the
FASP, FORTRAN Alphameric Subroutine Package Page 4
An Introduction to the FORTRAN Alphameric Subroutine Package
FORTRAN77 standards. Similarly, to permit subscript range
checking, all of the argument lists include size information
for the array arguments even though this size information
may not be needed by the logic of the routines. Since the
FORTRAN77 standards require the explicit preservation of
values across calls to a particular routine, all
non-dimensioned variables and all fixed dimensioned arrays
which contain values which are needed by the subsequent call
to the same routine but which are not needed by the calling
program are placed into labeled COMMON, and all variably
dimensioned arrays appear in the argument lists.
The routines in this package make extensive use of
punctuation characters which are not part of the standard
FORTRAN character set. In particular, many of the routines
in FASP use the semicolon to separate statements which
appear together on a single line, the exclamation point to
indicate that the current statement is complete with the
remainder of the current line being taken as a comment, and
the ampersand to indicate that the current statement
continues on the next line with the remainder of the current
line being taken as a comment. If these characters are not
available, then it will be necessary to select other
characters which can be used for these purposes and to
change the DATA statements accordingly. Most of the
routines in this package accept a tabulation (tab) character
as equivalent to a space character. Each test for a tab
character is preceded by a test for a space character, so it
is merely necessary to change the tab characters in the DATA
statements to spaces if the computer system upon which these
routines are used does not include the tab character in its
character set.
The routines named DACASE (the lower case to upper case
converter) and DAVERB (the word interpreter) each contain a
list of the lower case letters and a list of the
corresponding upper case letters. The lower case letters
are defined by DATA statements to be in the order 1Ha
through 1Hz which on the DECsystem-10 computer results in
the associated integer values being sorted into an
increasing order. If these routines are used upon a
computer system in which the alphabetical order of the
letters 1Ha through 1Hz does not result in an increasing
order for the associated integer values, then the DATA
statements which define the arrays which contain the lower
case letters should be rewritten so that the values
associated with the letters will be in increasing numerical
order, and then the DATA statements which define the
corresponding upper case letters must be rewritten so that
the lower and upper case versions of each letter appear in
locations in the respective arrays having the same
subscripts. If these routines are used upon a computer
system which does not support lower case letters then the
arrays which would otherwise contain the list of lower case
FASP, FORTRAN Alphameric Subroutine Package Page 5
An Introduction to the FORTRAN Alphameric Subroutine Package
letters as well as the arrays which contain the list of
upper case letters can each contain the upper case letters
1HA through 1HZ in alphabetical order even if this order is
not the numerically sorted order.
Since the routines in this package are designed to be usable
in a wide variety of applications, many of their argument
lists are long. If any one routine is to be called several
times within a particular program, then the writing of
separate CALL statements is both error prone and consumptive
of machine space. If the argument lists in the various
calls would be similar then it is often worthwhile to write
a short wrapper routine having a minimal argument list
consisting only of those arguments which would differ in
value for the individual calls. An error which is commonly
made in writing calls to routines having such long argument
lists is the omission of an argument. Such omissions can be
detected by comparing the number of arguments in the CALL
and SUBROUTINE statements, but can usually be prevented by
writing the CALL statements with exactly the same number of
arguments per line as appear in the SUBROUTINE statements.
An error which is often made when using the routines which
permit the user to specify by name and to manipulate the
values of FORTRAN arrays is the failure to declare the array
arguments of these routines in a DIMENSION statement in the
calling program since these arrays usually are not
manipulated by the calling program itself.
FASP, FORTRAN Alphameric Subroutine Package Page 6
One-Line Descriptions of Each of the Routines in FASP
One-Line Descriptions of Each of the Routines in FASP
--- ---- ------------ -- ---- -- --- -------- -- ----
FASP includes the routines listed below. The names of most
of these routines are constructed from the 2 letter prefix
"DA", standing for the word DAta, together with a 4 letter
word indicating the purpose of the routine. Two versions
are provided of some routines. The shorter of these
versions, named with the DA prefix followed by the single
letter I and a 3 letter contraction of the normal 4 letter
name, provides only a selected subset of the capabilities of
the longer, more general version. For example, several of
the input text buffer interpretation routines, such as
DAHEST and DANEXT, are provided in general versions which
will handle both real numbers and integers, and in shorter
versions, such as DAIHST and DAINXT, which will accept only
integer numbers.
DABASE locates start of region in dictionary made by DALOAD
DABELT constructs a band of column identification numbers
DACASE converts lower case letters in buffer to upper case
DACOPY expands tabs to spaces while copying text
DAFILL expands tabs to spaces without using extra buffer
DAFLAG locates components of PDP-10 file specification
DAFONT constructs FORTRAN statements describing DATEXT font
DAGRID rationalizes scales of plots produced by DAPLAT
DAHEFT evaluates integer and real numbers
DAHELP determines if text begins with question mark
DAHEST identifies command and all associated items
DAIBLT version of DABELT for column numbers increasing by 1
DAIFLL version of DAFILL which always copies initial spaces
DAIHFT integer only version of DAHEFT
DAIHST integer only version of DAHEST
DAINXT integer only version of DANEXT
DAIPAR version of DAPAIR not accepting range specifications
DAIRNK version of DARANK not accepting range specifications
FASP, FORTRAN Alphameric Subroutine Package Page 7
One-Line Descriptions of Each of the Routines in FASP
DAISPN integer only version of DASPAN
DAJOIN evaluates whole numbers, fractions and mixed numbers
DALEAD identifies command and next associated item
DALINE bar chart plotter for printer
DALIST summarizes contents of dictionary made by DALOAD
DALOAD constructs dictionary describing FORTRAN arrays
DALONE summarizes single entry in dictionary made by DALOAD
DALOOP returns next loop indexes varied in any order
DALOSS entry for DAVERB allowing missing words
DAMENU constructs dictionaries for DAVERB or DAHEST
DAMISS entry for DAHEFT allowing missing values
DAMOVE justify or center group of characters in field
DANAME finds multiple subscripts corresponding to single
DANEXT returns next value in series of numbers or ranges
DANUMB represents any integer
DAPAIR returns next pair of leading and following values
DAPATH represent unsigned integers sequence in form 1.2.3
DAPICK interprets array notation using DALOAD dictionary
DAPLAT plots lines, points or point clusters on printer
DARANK returns evaluated integers sorted without duplicates
DARITE represents real number in floating or exponent form
DAROLL returns next loop indexes varied in fixed order
DAROME represents integer as Roman numeral
DASAVE makes DATA statements for integer or Hollerith array
DASHOW simple entry for DARITE
DASITE finds single subscript corresponding to multiple
DASPAN returns next number or description of range
FASP, FORTRAN Alphameric Subroutine Package Page 8
One-Line Descriptions of Each of the Routines in FASP
DASWAP interchanges adjacent groups of characters in buffer
DATALL constructs printer plot with extended vertical scale
DATEAM evaluates and returns several values at once
DATEXT represents text using large, multiple-line lettering
DATREE returns next line in tree structure representation
DATREK evaluates series of unsigned integers of form 1.2.3
DATURN version of DATEXT lettering top to bottom of page
DAVARY displays and changes value identified by DAPICK
DAVERB identifies word or abbreviation of word
TEXT1 description for DATEXT of letters 9 wide by 5 high
TEXT2 description for DATEXT of letters 5 wide by 5 high
TEXT3 description for DATEXT of letters 12 wide by 11 high
TEXT4 description for DATEXT of letters 8 wide by 7 high
TEXT5 description for DATEXT of letters 14 wide by 9 high
FASP, FORTRAN Alphameric Subroutine Package Page 9
One-Paragraph Descriptions of Each of the Routines in FASP
One-Paragraph Descriptions of Each of the Routines in FASP
--- --------- ------------ -- ---- -- --- -------- -- ----
The routines having asterisks following their names in the
list below are used for the specification by name,
examination and modification of the values of multiply
subscripted FORTRAN arrays equivalenced with or otherwise
loaded into singly subscripted buffers. These routines
share a common data base, and use the same naming
conventions for their arguments. In addition to the
routines devoted soley to the manipulation of named arrays,
the routines DASITE and either DAROLL or DALOOP must be
called by the user's program. The latter routines use the
same argument naming conventions as do the other routines
designed for the manipulation of arrays by name, but were
written independently, and so accept a somewhat more general
data base.
DABASE * finds the start of a particular logical group of
array names contained in the dictionary
constructed by the DALOAD routine.
DABELT outputs into a text buffer the representation of a
band of column numbers. The numbers can increase
or decrease from left to right. Each line of the
representation can contain digits either
corresponding to the same power of 10, or of the
same significance. Either the most or the least
significant digits can be generated first. The
following are typical bands of column numbers as
generated by this routine.
- ----------01234567891
1--------- 1 1987654321 0
098765432101234567890 0
DACASE converts all lower case letters in an input text
buffer to upper case so that these letters can be
recognized by the other routines in FASP.
DACOPY copies the contents of one input text buffer into
another, expanding tab characters to enough spaces
to fill to the corresponding tab stops.
DAFILL expands tab characters in an input text buffer to
enough spaces to fill to the corresponding tab
stops without the use of a separate intermediate
buffer.
DAFLAG locates in an input text buffer, but does not
evaluate, the components of a file specification
of the general form
WORD:WORD.WORD[WORD,WORD]/WORD:'TEXT'/WORD:WORD
in which there can be only one name field
consisting of words connected by periods, and only
FASP, FORTRAN Alphameric Subroutine Package Page 10
One-Paragraph Descriptions of Each of the Routines in FASP
one bracketed field consisting of words connected
by commas. There can be several switch fields
starting with slashes and consisting of words or
quoted text strings connected by colons. Any
field can be missing, or can consist of any number
of subfields, some of which can be missing.
DAGRID adjusts the limiting data unit coordinates and the
grid line separations and offsets to produce
neater scale numbers in the printer plots produced
by DAPLAT. DAGRID is described at the end of the
DAPLAT documentation.
DAHEFT evaluates the integers or the real numbers
represented in an input text buffer. These
numbers can be represented with decimal points
and/or in scientific notation. DAMISS should be
called instead of DAHEFT if comments and missing
numbers are to be allowed.
DAHELP determines whether the input text buffer contains
leading question marks.
DAHEST evaluates simple commands issued to interactive
programs. These commands consist of a leading key
word followed by arguments which can be numbers,
text strings and/or words. Only one type of
argument can appear more than once following a
particular command, and extra commas between the
arguments indicate missing arguments of this type.
Several statements can appear on a single line,
but statements cannot be continued onto a
following line. Each call to DAHEST returns
information about an entire command.
DAIBLT outputs into a text buffer the representation of a
band of column numbers. DAIBLT is described in
the documentation of DABELT. DABELT must be used
instead of DAIBLT if the numbers must decrease
from left to right or if each line of the
representation must contain digits of the same
significance.
DAIFLL expands tab characters in an input text buffer to
enough spaces to fill to the corresponding tab
stops without the use of a separate intermediate
buffer. DAIFLL is described in the documentation
of DAFILL. DAFILL must be used instead of DAIFLL
if initial tabs or spaces in the buffer must be
suppressed.
DAIHFT evaluates the integers represented in an input
text buffer. DAIHFT is described in the
documentation of DAHEFT. DAHEFT must be used
FASP, FORTRAN Alphameric Subroutine Package Page 11
One-Paragraph Descriptions of Each of the Routines in FASP
instead of DAIHFT if real numbers must be
evaluated.
DAIHST evaluates simple commands issued to interactive
programs. DAIHST is described in the
documentation of DAHEST. DAHEST must be used
instead of DAIHST if real numbers must be
evaluated in the argument lists following the
command words.
DAINXT returns the next integer value represented by the
contents of an input text buffer which can contain
representations of any combination of single
values, range specifications and/or values to be
repeated a specified number of times. DAINXT is
described in the documentation of DANEXT. DANEXT
must be used instead of DAINXT if real numbers
must be evaluated.
DAIPAR returns the next pair of values represented by the
contents of an input text buffer. If the
representation of a third value appears after the
representation of a pair of values, then the third
number can be interpreted as either an additional
number to be associated with the leading
(leftmost) number of the previous pair, or as the
leading number of a new pair. DAIPAR is described
in the documentation of DAPAIR. DAPAIR must be
used instead of DAIPAR if ranges of numbers must
be evaluated.
DAIRNK returns the decimal integer values represented by
the contents of an input text buffer, sorted and
excluding duplicate values. DAIRNK is described
in the documentation of DARANK. DARANK must be
used instead of DAIRNK if the values are to be
specified as ranges of values.
DAISPN interprets an input text buffer containing the
representation of a single integer value,
specification of an integer range, or an integer
value to be repeated a specified number of times.
DAISPN is described in the documentation of
DASPAN. DASPAN must be used instead of DAISPN if
real numbers must be evaluated.
DAJOIN evaluates the whole numbers, fractions and mixed
numbers represented in an input text buffer.
DAJOIN would return the value -3.5 if the input
text buffer contains -3 1/2.
DALEAD identifies a command word and a following
associated word or number or quoted text string
represented by the contents of an input text
FASP, FORTRAN Alphameric Subroutine Package Page 12
One-Paragraph Descriptions of Each of the Routines in FASP
buffer. If an associated word or number or string
is found without a preceding command word, then
the command word identified by the previous call
to this routine is again identified. An ampersand
at the right end of a line indicates that the
items found at the start of the following line are
to continue to be associated with the current
command word until a new command word is
encountered.
DALINE constructs printable plots containing horizontal
bars formed of segments the lengths of which
represent the magnitudes of the corresponding
values.
DALIST * types a summary of the array names and subscript
limits contained in the dictionary constructed by
the DALOAD routine.
DALOAD * reads FORTRAN statements which would define the
dimensions of arrays and, from these statements,
constructs a dictionary which can be used by the
rest of the routines in this list having an
asterisk to the right of their names.
DALONE * outputs into a text buffer the characters of an
array name contained in the dictionary constructed
by the DALOAD routine, together with either a
representation of the subscript limits of the
array or a representation of the current values of
these subscripts.
DALOOP simulates variably embedded DO loops to any
desired depth.
DALOSS identifies words and abbreviations appearing in an
input text buffer. An abbreviation consists of
sufficient initial characters to uniquely identify
the desired word from all other words in the
dictionary. Commas can separate the words and
extra commas can indicate missing items. Comments
in the input text buffer are ignored. DALOSS is
described at the end of the DAVERB documentation.
DAMISS evaluates the integers or the real numbers
represented in an input text buffer. Commas can
separate the number representations and extra
commas can indicate missing items. Comments in
the input text buffer are ignored. DAMISS is
described at the end of the DAHEFT documentation.
DAMOVE left justifies, centers or right justifies a group
of characters within a larger section of the text
buffer in which these characters are contained.
FASP, FORTRAN Alphameric Subroutine Package Page 13
One-Paragraph Descriptions of Each of the Routines in FASP
DANAME * finds the array name and its subscripts (as
defined by the dictionary constructed by the
DALOAD) routine) associated with a particular
location in the singly subscripted buffer with
which the array is equivalenced or into which the
array is otherwise loaded.
DANEXT returns the next numeric value represented by the
contents of an input text buffer which can contain
the representations of any combination of single
values, range specifications and/or values to be
repeated a specified number of times. The values
10, 15, 20, -120 and -120 would be returned by 5
calls to this routine if the input text buffer
contains 10/5/20,2*-120.
DANUMB outputs into a text buffer the representation of
an integer (not real) value.
DAPAIR returns the next pair of values represented by the
contents of an input text buffer. Either value
can be specified as a member of a range of values.
If the leading value is specified as a member of a
range of values, if the the associated value is
specified either as a single value or as a member
of a range of values containing fewer values than
the leading range, and if an additional
specification of a single value or of a range of
values appears to the right of the specification
of the associated value or range of values, then
the additional value or values will be returned
with the remaining values of the initial range.
The pairs of values 10 and 50, 15 and 60, 20 and
70 and 25 and 100 would be returned by 4 calls to
this routine if the input text buffer contains
10/5/25,50/10/70,100.
DAPATH outputs into a text buffer the representation of a
sequence of unsigned integers separated by
periods. This representation can be interpreted
by the DATREK routine.
DAPICK * identifies an array name and subscript ranges in
an input text buffer by matching the contents of
the buffer against the entries in a dictionary
constructed by the DALOAD routine.
DAPLAT generates printer or terminal plots of points,
point clusters and/or curves.
DARANK returns the decimal integer values represented by
the contents of an input text buffer, sorted and
excluding duplicate values. The values can be
specified as members of ranges of values.
FASP, FORTRAN Alphameric Subroutine Package Page 14
One-Paragraph Descriptions of Each of the Routines in FASP
DARITE outputs into a text buffer the representation of a
real (not integer) value in either floating point
notation or scientific (exponent) notation. If
necessary, DARITE can alter these formats within
ranges specified by the calling program.
DAROLL returns the next values of an integer (not text)
array being varied between a set of lower and
upper bounds. Either the lowest or the highest
subscript can be varied the most rapidly. DAROLL
is described in the documentation of DALOOP.
DALOOP must be used instead of DAROLL if the
values in the array must be advanced in a sequence
other than that specified by the subscripts of the
array.
DAROME outputs into a text buffer the representation in
Roman numeral notation of an integer value.
DASAVE generates compilable FORTRAN DATA statements
representing the contents of any single precision
array containing either integer values or text
characters.
DASHOW outputs into a text buffer the representation of a
real (not integer) value in either floating point
notation or scientific (exponent) notation. If
necessary, DASHOW can alter these formats within
ranges specified by the calling program.
Rightmost zeroes which are to the right of the
decimal point are suppressed. DARITE should be
used instead of DASHOW if centering or right
justification or rightmost fill with spaces is
needed.
DASITE returns the location within a singly dimensioned
buffer (containing any data type) of a particular
location within a multiply subscripted array
contained in the buffer.
DASPAN interprets an input text buffer containing the
representation of a single numeric value, range
specification or value to be repeated a specified
number of times. DANEXT would usually be called
instead of DASPAN if the values within the range,
rather than just a description of the range, is
required.
DASWAP swaps 2 adjacent groups of characters in an input
text buffer without the use of an additional
buffer.
DATALL generates a plot with a vertical axis extending
onto as many lines and pages as necessary to
FASP, FORTRAN Alphameric Subroutine Package Page 15
One-Paragraph Descriptions of Each of the Routines in FASP
represent the data.
DATEAM returns all of the values represented by the
contents of an input text buffer. Extra commas,
which indicate missing values to some routine, are
ignored.
DATEXT generates large multiple line lettering extending
from left to right across the width of the page.
Also included is DAFONT, a program which allows
user specification of character shapes.
DATREE identifies the items appearing in the next line of
a simple tree structure in which the root is
placed in the left column, those nodes which lie
immediately above the root are placed in the
second column, those which are above the nodes in
the second column are placed in the third column,
and so on.
DATREK evaluates a series of unsigned integers
represented by groups of digits separated by
periods in an input text buffer. A value of -1 is
returned for any integer which is indicated as
missing by an initial period, by a trailing
period, or by 2 adjacent periods. Signs and
exponents are not recognized. This representation
of a series of unsigned integers can be generated
by the DAPATH routine.
DATURN generates large multiple line lettering extending
from top to bottom of page and onto subsequent
pages. DATURN is described at the end of the
DATEXT documentation.
DAVARY * displays the values contained in the array
locations identified by the DAPICK routine, and
accepts the user specification of the values in
these array locations.
DAVERB identifies words and abbreviations appearing in an
input text buffer. An abbreviation consists of
sufficient initial characters to uniquely identify
the desired word from all other words in the
dictionary. DALOSS should be called instead of
DAVERB if comments and missing words are to be
allowed.
The FASP package also includes the following support
programs which are used to construct the DATA statements
necessary for the use of some of the routines described
above.
FASP, FORTRAN Alphameric Subroutine Package Page 16
One-Paragraph Descriptions of Each of the Routines in FASP
DAFONT reads a lettering pattern file which can be easily
created or modified by the user, and generates
from this file the FORTRAN statements which define
a bit coded array which contains the information
needed by DATEXT and by DATURN to generate this
style of lettering. DAFONT is described at the
end of the DATEXT documentation.
DAMENU reads a file containing the words which are to be
recognized by DAVERB, or by any routine which
calls DAVERB, and generates the FORTRAN statements
which define this dictionary. DAMENU can maintain
the array of argument types to be allowed by
DAHEST parallel to the words in the dictionary.
DAMENU is described at the end of the
documentation of DAHEST.
FASP, FORTRAN Alphameric Subroutine Package Page 17
Routines in FASP Used to Evaluate Typical Lines of Text
Routines in FASP Used to Evaluate Typical Lines of Text
-------- -- ---- ---- -- -------- ------- ----- -- ----
Several typical lines of text which could be interpreted by
the routines in this package are listed below together with
descriptions of the results which would be returned to the
calling program. The exclamation point is used in some of
the examples to start comments, but is not supported by the
more primitive routines such as DAHEFT, DAVERB and DATREK.
-7366.2
or
-7.3662K
or
-7.3662E3
evaluated by DAHEFT, or by most of the routines which
call DAHEFT (some don't allow octal), as an octal
integer having the decimal value -3830 (- 7x512 - 3x64 -
6x8 - 6) or as a decimal integer having the value -7366
or as a decimal real having -7366.2 as its value.
DARITE can represent the real value -7366.2 in any of
the forms shown in the above examples.
-1/2
evaluated by DAJOIN as a decimal real having -0.5 as its
value.
evaluated by DANEXT as the sequence of octal integers or
of decimal integers or of decimal reals having the
values -1, 0, 1 and 2. An increment of 1 is assumed.
-1 3/5
evaluated by DAJOIN as a decimal real having -1.6 as its
value. This interpretation does not allow a comma
between the -1 and the 3.
evaluated by DANEXT as the sequence of octal integers or
of decimal integers or of decimal reals having the
values -1, 3, 4 and 5. A comma could appear between the
-1 and the 3.
evaluated by DAPAIR as the pair of octal integers or of
decimal integers or of decimal reals having the values
-1 and 3, followed by the initial -1 associated with the
values 4 and 5 in turn. A comma could appear between
the -1 and the 3.
FASP, FORTRAN Alphameric Subroutine Package Page 18
Routines in FASP Used to Evaluate Typical Lines of Text
1.5377E5,,-1 3/5!ANY CHARACTERS FORMING COMMENT
evaluated by DAJOIN as a decimal real having the value
153770 followed by a missing number and then by a mixed
fraction having the value -1.6.
evaluated by DANEXT as a leading octal integer having
the value 55288 or as a leading decimal integer having
the value 153770 or as a leading decimal real having the
value 153770, the leading number being followed by a
missing item indicated by the extra comma and then
followed by a sequence having the values -1, 3, 4 and 5.
20/5/40!ANY CHARACTERS FORMING COMMENT
evaluated by DANEXT as the sequence of decimal integers
or of decimal reals having the values 20, 25, 30, 35 and
40 or as the sequence of octal integers having the
decimal values 16, 21, 26 and 31.
5*-100!ANY CHARACTERS FORMING COMMENT
evaluated by DANEXT as 5 occurrences of a decimal
integer having the value -100 or of a decimal real
having the value -100 or of an octal integer having the
decimal value -64.
44.120.0
evaluated by DATREK as the sequence of unsigned decimal
integers having the values 44, 120 and 0.
DAPATH can represent the sequence of unsigned decimal
integers having the values 44, 120 and 0 in this form.
-12 70 16 4!ANY CHARACTERS FORMING COMMENT
evaluated by DAPAIR as the pair of decimal integers or
of decimal reals having the values -12 and 70 followed
by the pair having the values 16 and 4, or as the pair
of octal integers having the decimal values -10 (- 1x8 -
2) and 56 (7x8) followed by the pair having the decimal
values 14 (1x8 + 6) and 4. A comma could appear between
any of the number representations.
evaluated by DAPAIR alternatively as a decimal integer
or as a decimal real having the value -12 which is
associated in turn with each of the numbers 70, 16 and 4
or as an octal integer having the decimal value -10
which is associated in turn with numbers having the
decimal values 56, 14 and 4. A comma could appear
between any of the number representations.
evaluated by DAMISS or by DATEAM or by DANEXT as the
FASP, FORTRAN Alphameric Subroutine Package Page 19
Routines in FASP Used to Evaluate Typical Lines of Text
sequence of decimal integers or of decimal reals having
the values -12, 70, 16 and 4, or as the sequence of
octal integers having the decimal values -10, 56, 14 and
4. A comma could appear between any of the number
representations.
10/5/20 100 200 300!ANY CHARACTERS FORMING COMMENT
evaluated by DAPAIR as the pair of decimal integers or
of decimal reals having the values 10 and 100, followed
by the pair having the values 15 (5 is the increment)
and 200 and finally by the pair having the values 20 and
300, or as the pair of octal integers having the decimal
values 8 and 64, followed by the pair having the decimal
values 13 (octal 10 plus 5) and 128. A comma could
appear between the 20 and the 100 or between any of the
following number representations. If the evaluation is
done in octal, then the 13 exhausts the first range of
numbers so that the octal 300 (decimal 192) can be
associated with an additional 13 or can be the first
number of the following pair of numbers.
evaluated by DANEXT as the sequence of decimal integers
or of decimal reals having the values 10, 15, 20, 100,
200 and 300 or as the sequence of octal integers having
the decimal values 8, 13, 64, 128 and 192. A comma
could appear between the 20 and the 100 or between any
of the following number representations.
20/5/10 -400/100/-200!ANY CHARACTERS FORMING COMMENT
evaluated by DAPAIR as the pair of decimal integers or
of decimal reals having the values 20 and -400, followed
by the pair having the values 15 and -300 and finally by
the pair having the values 10 and -200, or as the pair
of octal integers having the decimal values 16 and -256
(- 4x64) followed by the pair having the decimal values
11 (octal 20 minus 5) and -192 (octal -400 plus octal
100). A comma could appear between the 10 and the -400.
If the evaluation is done in octal, then the 11 exhausts
the first range of numbers so that the octal -200 can be
associated with an additional 11 or can be the first
number of the following pair of numbers.
evaluated by DANEXT as the sequence of decimal integers
or of decimal reals having the values 20, 15, 10, -400,
-300 and -200 or as the sequence of octal integers
having the decimal values 16, 11, -256, -192 and -128.
A comma could appear between the 10 and the -400.
FASP, FORTRAN Alphameric Subroutine Package Page 20
Routines in FASP Used to Evaluate Typical Lines of Text
WORD
identified by DAVERB or by DALOSS by matching the word
against all of the words in a dictionary supplied by the
calling program. DAVERB can also identify any
abbreviation which uniquely selects the word.
WORD,,WORD!ANY CHARACTERS FORMING COMMENT
identified by DALOSS as an appearance of the word named
WORD, followed by a missing item indicated by the extra
comma, and then by an appearance of the word named WORD.
COMMAND 5.3 WORD 'QUOTED TEXT STRING'!ANY COMMENT
evaluated by DALEAD or by DAHEST as the command word
named COMMAND followed by the associated decimal integer
having the value 5 or by the associated decimal real
having the value 5.3, followed by the word named WORD
associated with the same command and then by the text
between the delimiting apostrophes again associated with
the same command. If evaluated by DALEAD, commas could
appear between the command and the first following item
or between the following items, and the number could
also be evaluated as an octal integer. If evaluated by
DAHEST, a comma between the command and the first
following item would indicate a missing item, and 2 or
more commas between any pair of following items would
similarly indicate 1 or more missing items.
ARRAY(12,3/5)=-10.2,16E-3,3K!ANY CHARACTERS FORMING COMMENT
The combination of DAPICK and DAVARY used together with
several other routines would evaluate the above text as
the instructions to set ARRAY(12,3) to -10.2,
ARRAY(12,4) to 0.016 and ARRAY(12,5) to 3000. The
numbers appearing to the right of the equals sign could
also be evaluated as octal integers or as decimal
integers if indicated by the dictionary which describes
the named array.
DEVICE:NAME.NAME[NUMBER,NUMBER]/OPTION:NUMBER/OPTION:'TEXT'
DAFLAG locates but does not evaluate or verify the
component parts of a PDP-10 file specification such as
that shown above. The name field can consist of several
words separated by periods. The bracketed field can
contain several numbers or words separated by commas.
The several option fields can each contain several words
or numbers or quoted text strings separated by colons.
FASP, FORTRAN Alphameric Subroutine Package Page 21
Character Set Used by these Routines
The Character Set Used by these Routines
--- --------- --- ---- -- ----- --------
Although the logic of the routines in this package was
written in simple, machine independent FORTRAN, character
sets are not standardized. In addition to the space (or
blank), the upper and lower case letters of the alphabet,
and the characters customarily used for the representation
of numbers in scientific notation, the following characters
are also used
/ (slash) used in range specifications to separate
starting value from increment from final value. Used
in fractions to separate numerator from denominator.
Used in a file specification to start each switch
field.
* (asterisk) used in a range specification to separate
the number of times a particular value is to be
repeated from the value itself. In a file
specification interpreted by the DAFLAG routine, each
appearance of the asterisk becomes a separate component
of the corresponding field even if not separated within
the text from the other characters which form the
field.
' (apostrophe) used to mark the start and end of text
strings which appear either as the argument of a
command or as the argument of a switch in a file
specification.
! (exclamation) used to indicate that the remainder of
the buffer contents are to be ignored. Used by the
plotting routines for the ruling of vertical grid
lines.
& (ampersand) used to indicate that the remainder of the
buffer contents are to be ignored, but that the user
intends to continue input on the following line. Used
by the plotting routines to indicate curve
intersections and superimposed points.
; (semicolon) used to separate commands or file
specifications typed on the same line.
, (comma) used to separate adjacent words and/or numbers
in a list. Used to separate the arguments of a
command. Used in a file specification to separate the
components of a bracketed field. Used by the named
array manipulation routines to separate the subscripts
of an array.
= (equal) used by the named array manipulation routines
to separate the identification of an array location
FASP, FORTRAN Alphameric Subroutine Package Page 22
Character Set Used by these Routines
from the value to be inserted into that location. Used
in a file specification to separate destination file
from source file.
% (percent) used to indicate that a number is to be
divided by 100.
[ (left square bracket) used in a file specification to
start a bracketed field. A DATA statement in the
DAFLAG routine defines this character in 1H notation.
A second DATA statement which has been commented out
with a C in column 1 contains the octal value
corresponding to 5H[ on the PDP10, and must be restored
if this routine is compiled using the F40 compiler.
] (right square bracket) used in a file specification to
end a bracketed field. A DATA statement in the DAFLAG
routine defines this character in 1H notation. A
second DATA statement which has been commented out with
a C in column 1 contains the octal value corresponding
to 5H] on the PDP10, and must be restored if this
routine is compiled using the F40 compiler.
( (left parenthesis) used by the named array manipulation
routines to begin the specification of the subscripts
of an array. Used by DAHEST to mark the start of a
text string which will be terminated by a matching ),
or as the terminator of such a text string which has
been begun by an initial ).
) (right parenthesis) used by the named array
manipulation routines to terminate the specification of
the subscripts of an array. Used by DAHEST to mark the
start of a text string which will be terminated by a
matching (, or as the terminator of such a text string
which has been begun by an initial (.
: (colon) used in range specifications to separate
starting value from increment from final value. Used
in a file specification to terminate a device name or
to separate the components of a switch field.
@ (at) used in a file specification to indicate that the
file being specified itself contains a file
specification.
$ (dollar) used within text strings being converted to
large printable lettering to mark a following character
which is to be treated as a command rather than as a
character to be represented.
- (minus) in addition to its traditional use to indicate
that the number appearing immediately to its right is
negative, the minus sign is used by the plotting
FASP, FORTRAN Alphameric Subroutine Package Page 23
Character Set Used by these Routines
routines for the ruling of horizontal grid lines.
+ (plus) in addition to its traditional use to indicate
that the number appearing immediately to its right is
positive, the plus sign is used by the plotting
routines for the marking of grid line intersections.
? (question) typed by the user at the start of a line to
request assistance.
tab (nonprinting horizontal tabulation) equivalent to one
or more spaces such that the next character would
appear beyond the next whole number multiple of the tab
stop interval. Each test for a tab character is
preceded by a test for a space (blank). If the tab
character is not available, the variable named ITAB
should in each routine in which it appears be defined
as a Hollerith space character rather than a Hollerith
tab character.
In some cases, such as the inclusion of an @ sign or = sign
in a file specification, the FASP routines merely report to
the calling program that the character was found, and it is
the responsibility of the calling program to take the
appropriate action.
FASP, FORTRAN Alphameric Subroutine Package Page 24
Characters Recognized by Interpretative Routines in FASP
Characters Recognized by Interpretative Routines in FASP
---------- ---------- -- -------------- -------- -- ----
The most primitive routines in FASP recognize no printing
characters other than those which can form numbers or which
are contained in words in the dictionary supplied by the
calling program. Most of the routines which call these
primitives allow punctuation characters such as the comma
between adjacent items, the semicolon between adjacent
statements on a single line, the exclamation point before a
comment, and the ampersand before a comment when the current
statement is to be continued onto the next line. The
special printing characters recognized by each of the
interpretative routines in FASP are listed in the table
below. The routines devoted to the manipulation of named
arrays are not included in this table.
DAFLAG . * / : , ; ! & ' [ ] @ =
DAHEFT 0-9 EKM% + - .
DAIHFT 0-9 EKM% + - .
DAHELP ?
DAHEST A-Z 0-9 EKM% + - . / : , ; ! ' ( )
DAIHST A-Z 0-9 + - / : , ; ! ' ( )
DAJOIN 0-9 EKM% + - . /
DALEAD A-Z 0-9 EKM% + - . , ; ! & '
DALOSS A-Z , ; ! &
DAMISS 0-9 EKM% + - . , ; ! &
DANEXT 0-9 EKM% + - . * / : , ; ! &
DAINXT 0-9 + - * / : , ; ! &
DAPAIR 0-9 EKM% + - . * / : , ; ! &
DAIPAR 0-9 EKM% + - . , ; ! &
DARANK 0-9 EKM% + - . * / : , ; ! &
DAIRNK 0-9 EKM% + - . , ; ! &
DASPAN 0-9 EKM% + - . * / : , ; ! &
DAISPN 0-9 + - * / : , ; ! &
DATEAM 0-9 EKM% + - . , ; ! &
DATREK 0-9 .
DAVERB A-Z
FASP, FORTRAN Alphameric Subroutine Package Page 25
Calling Hierarchies and Sizes of the FASP Routines
Calling Hierarchies and Sizes of the FASP Routines
------- ----------- --- ----- -- --- ---- --------
For each of the routines listed in the left column in the
following table, the routines named to its right must also
be loaded. The externally called routines are listed
alphabetically except for those routines which have names in
which the third letter is I, these being listed just below
the corresponding routine with real number capabilities. In
parentheses to the right of each routine name is the number
of FORTRAN statements, both executable and nonexecutable but
excluding comments, in that particular routine. As an
example, the table entry
DATALL(119)-DAPLAT(455)-DARITE(312)
-PLTCUT( 61)
indicates that DATALL which contains 119 FORTRAN statements,
calls DAPLAT which contains 455 statements and which in turn
calls the 2 routines DARITE and PLTCUT.
DABASE( 36)
DABELT(133)
DAIBLT( 66)
DACASE( 29)
DACOPY( 44)
DAFILL( 82)
DAIFLL( 58)
DAFLAG(205)
DAGRID( 90)
DAHEFT(177)
DAIHFT(149)
DAHELP( 14)
DAHEST(328)-DAHEFT(177)
-DAVERB(112)
DAIHST(322)-DAVERB(112)
DAJOIN( 82)-DAHEFT(177)
DALEAD(104)-DAHEFT(177)
-DAVERB(112)
DALINE(263)-DARITE(312)
DALIST( 42)-DALONE( 75)-DANUMB( 62)
FASP, FORTRAN Alphameric Subroutine Package Page 26
Calling Hierarchies and Sizes of the FASP Routines
DALOAD(271)
DALONE( 75)-DANUMB( 62)
DALOOP( 76)
DALOSS( 49)-DAVERB(112)
DAMISS( 54)-DAHEFT(177)
DAMOVE( 33)
DANAME( 69)
DANEXT( 83)-DASPAN(120)-DAHEFT(177)
DAINXT( 42)-DAISPN(118)
DANUMB( 62)
DAPAIR(146)-DANEXT( 83)-DASPAN(120)-DAHEFT(177)
DAIPAR( 98)-DAHEFT(177)
DAPATH( 25)-DAMOVE( 33)
-DANUMB( 62)
DAPICK(199)
DAPLAT(455)-DARITE(312)
-PLTCUT( 61)
DARANK( 88)-DANEXT( 83)-DASPAN(120)-DAHEFT(177)
DAIRNK( 60)-DAMISS( 54)-DAHEFT(177)
DARITE(312)
DAROME( 71)
DASAVE(230)-DABOTH( 42)-DANUMB( 62)
-DANUMB( 62)
DASHOW( 23)-DANUMB( 62)
-DARITE(312)
DASITE( 76)
DASPAN(120)-DAHEFT(177)
DAISPN(118)
DASWAP( 25)
DATALL(119)-DAPLAT(455)-DARITE(312)
-PLTCUT( 61)
DATEAM( 38)-DAHEFT(177)
FASP, FORTRAN Alphameric Subroutine Package Page 27
Calling Hierarchies and Sizes of the FASP Routines
DATEXT(269)-one font of approximately 20 statements *
DATREE(105)
DATREK( 43)
DATURN(273)-one font of approximately 20 statements *
DAVARY( 58)-DANEXT( 83)-DASPAN(120)-DAHEFT(177)
-DANUMB( 62)
-DARITE(312)
DAVERB(112)
* The fonts contain descriptions of the character shapes,
and are named TEXT1, TEXT2, TEXT3 and so on.
FASP, FORTRAN Alphameric Subroutine Package Page 28
An Introduction to Character Manipulation in FORTRAN
An Introduction to Character Manipulation in FORTRAN
-- ------------ -- --------- ------------ -- -------
Interactive programs often allow the user to select one item
from a group of possible choices. The easiest way for the
programmer to provide this feature is to assign a different
number to each possible choice, then to have the program ask
for and accept the number. This places upon the user the
burden either of remembering or of finding the number
corresponding to the item being selected. Using such a
method, a typical interaction with the user could be
performed by the FORTRAN statements (in which the $ in the
format merely allows the user to answer on the same line)
1 TYPE 2
2 FORMAT(' METHOD OF SHIPMENT (TYPE 0 FOR LIST)? ',$)
ACCEPT 3,METHOD
3 FORMAT(I)
IF((METHOD.GT.0).AND.(METHOD.LE.3))GO TO 5
TYPE 4
4 FORMAT(' POSSIBLE CHOICES ARE:'/
1' 1 TRUCK'/
2' 2 TRAIN'/
3' 3 PLANE')
GO TO 1
5 CONTINUE
The sample question would be easier to answer correctly if
the words TRUCK, TRAIN and PLANE could be typed directly
instead of the numbers selecting these responses. If the
accepted words are selected to each begin with different
letters, then a single letter response can be read with an
A1 format. In the above example, the words TRUCK and TRAIN
both start with the letter T, but the TRAIN response can
instead be changed to RAIL. The FORTRAN statements to
perform the interaction could then be rewritten
DIMENSION LETTER(3)
DATA LETTER/1HT,1HR,1HP/
1 TYPE 2
2 FORMAT(' METHOD OF SHIPMENT? ',$)
ACCEPT 3,IRSPNS
3 FORMAT(1A1)
METHOD=1
4 IF(IRSPNS.EQ.LETTER(METHOD))GO TO 6
METHOD=METHOD+1
IF(METHOD.LE.3)GO TO 4
TYPE 5
5 FORMAT(' POSSIBLE CHOICES ARE:'/
1' TRUCK'/
2' RAIL'/
3' PLANE')
GO TO 1
6 CONTINUE
FASP, FORTRAN Alphameric Subroutine Package Page 29
An Introduction to Character Manipulation in FORTRAN
Free format responses are not accepted by the above FORTRAN
statements. The identifying character must be the first
character typed by the user. The response is not recognized
if preceded by a space. This difficulty could be
circumvented by reading the response with a larger A format,
perhaps A4, and testing against the possible sequences
4HT ,4H T ,4H T ,4H T
4HR ,4H R ,4H R ,4H R
4HP ,4H P ,4H P ,4H P
However, the user who typed more than a single character
abbreviation would find that such a response would be
rejected, so the range of character sequences tested against
for just the single word TRUCK would have to be expanded
still further to include
4HT ,4H T ,4H T ,4H T
4HTR ,4H TR ,4H TR
4HTRU ,4H TRU
4HTRUC
Such a scheme becomes impractical. An alternative method is
to read the user's response into an array, 1 character per
array location, with a multiple of an A1 format. If, as is
probable, the user types fewer characters than the array can
hold, then spaces are placed into the array locations to the
right of (having higher subscripts than) those containing
the characters actually typed. The FORTRAN statements shown
below can then search for the first printing character in
the array.
DIMENSION LETTER(3),IRSPNS(10)
DATA ISPACE/1H /,LETTER/1HT,1HR,1HP/
1 TYPE 2
2 FORMAT(' METHOD OF SHIPMENT? ',$)
ACCEPT 3,IRSPNS
3 FORMAT(10A1)
INITAL=1
4 IF(IRSPNS(INITAL).NE.ISPACE)GO TO 5
IF(INITAL.GE.10)GO TO 7
INITAL=INITAL+1
GO TO 4
5 METHOD=1
6 IF(IRSPNS(INITAL).EQ.LETTER(METHOD))GO TO 9
METHOD=METHOD+1
IF(METHOD.LE.3)GO TO 6
7 TYPE 8
8 FORMAT(' POSSIBLE CHOICES ARE:'/
1' TRUCK'/
2' RAIL'/
3' PLANE')
GO TO 1
9 CONTINUE
FASP, FORTRAN Alphameric Subroutine Package Page 30
An Introduction to Character Manipulation in FORTRAN
If the program must accept either an alphabetic or a numeric
response to the question, then the first printing character
must also be compared with each of the possible digits. If
a digit is found, then the currently stored numeric value,
if any, must be shifted, and the value of the new digit
inserted into the least significant end of the stored value,
following which the next character typed by the user must be
similarly tested. Either an alphabetic or a numeric
response is identified by the following FORTRAN statements
DIMENSION LETTER(3),IRSPNS(10),IDIGIT(10)
DATA IDIGIT/1H0,1H1,1H2,1H3,1H4,1H5,1H6,1H7,1H8,1H9/
DATA ISPACE/1H /,LETTER/1HT,1HR,1HP/
C
C ASK FOR AND ACCEPT RESPONSE
1 TYPE 2
2 FORMAT(' METHOD OF SHIPMENT OR MONTHS TO STORE? ',$)
ACCEPT 3,IRSPNS
3 FORMAT(10A1)
C
C LOOK FOR FIRST PRINTING CHARACTER
INITAL=1
4 IF(IRSPNS(INITAL).NE.ISPACE)GO TO 5
IF(INITAL.GE.10)GO TO 10
INITAL=INITAL+1
GO TO 4
C
C CHECK FOR WORDS TRUCK, RAIL AND PLANE
5 METHOD=1
6 IF(IRSPNS(INITAL).EQ.LETTER(METHOD))GO TO 12
METHOD=METHOD+1
IF(METHOD.LE.3)GO TO 6
MONTHS=0
C
C CHECK IF NUMBER ISSUED INSTEAD OF TRANSFER METHOD
7 LTRTST=IRSPNS(INITAL)
DO 8 JDIGIT=1,10
IF(LTRTST.NE.IDIGIT(JDIGIT))GO TO 8
MONTHS=(10*MONTHS)+JDIGIT-1
INITAL=INITAL+1
IF(INITAL.LE.10)GO TO 7
GO TO 9
8 CONTINUE
9 IF(MONTHS.GT.0)GO TO 12
C
C PROMPT USER IF WRONG RESPONSE GIVEN
10 TYPE 11
11 FORMAT(' POSSIBLE CHOICES ARE:'/
1' TRUCK'/
2' RAIL'/
3' PLANE'/
4' OR NUMBER OF MONTHS WHICH PRODUCT IS TO BE STORED')
GO TO 1
12 CONTINUE
FASP, FORTRAN Alphameric Subroutine Package Page 31
An Introduction to Character Manipulation in FORTRAN
The appearance of a similar group of FORTRAN statements
after each question for the identification of the user
response might be acceptable if only a few questions are
asked by the entire program. However, if the program is to
ask many questions, then it becomes reasonable to separate
the evaluation statements into one or more general
subroutines which are called whenever a response must be
evaluated. Each subroutine can then do much more evaluation
and validation since the statements appear only within the
subroutines rather than being duplicated at many sites.
FASP includes several such subroutines. Calling two of
these subroutines, DALOSS for the identification and
validation of words and DAHEFT for the evaluation of
numbers, the FORTRAN statements shown above can be rewritten
DIMENSION LETTER(14),KNTLTR(3),IRSPNS(10)
DATA LETTER/1HT,1HR,1HU,1HC,1HK,1HR,1HA,1HI,1HL,
11HP,1HL,1HA,1HN,1HE/,KNTLTR/5,4,5/
C
C ASK FOR AND ACCEPT RESPONSE
1 TYPE 2
2 FORMAT(' METHOD OF SHIPMENT OR MONTHS TO STORE? ',$)
ACCEPT 3,IRSPNS
3 FORMAT(10A1)
INITAL=1
MANY=0
C
C CHECK FOR WORDS TRUCK, RAIL AND PLANE
4 CALL DALOSS(1,14,LETTER,1,3,KNTLTR,IRSPNS,10,INITAL,
1KIND,METHOD,LCNWRD,LCNKNT,LCNBFR,MANY,LCNERR)
GO TO(1,5,8,8,6,6,6,6,4,1,4),KIND
C
C CHECK IF NUMBER ISSUED INSTEAD OF TRANSFER METHOD
5 METHOD=4
CALL DAHEFT(0,1,0,IRSPNS,10,INITAL,KIND,
1ISHIFT,JSHIFT,KSHIFT,LSHIFT,MONTHS,VALUE)
IF(KIND.NE.3)GO TO 6
IF(MONTHS.GT.0)GO TO 8
C
C PROMPT USER IF WRONG RESPONSE GIVEN
6 TYPE 7
7 FORMAT(' POSSIBLE CHOICES ARE:'/
1' TRUCK'/
2' RAIL'/
3' PLANE'/
4' OR NUMBER OF MONTHS WHICH PRODUCT IS TO BE STORED')
GO TO 1
8 CONTINUE
Through the use of the FASP routines, abbreviations are
rejected if ambiguous or if misspelled, numbers can include
sign and decimal point and exponent, and comments can be
included to the right of either exclamation points or
ampersands.
FASP, FORTRAN Alphameric Subroutine Package Page 32
DABASE, Routine to Locate Logical Group of Named Arrays
DDDDD AAA BBBBBB AAA SSSSSS EEEEEEEE
DD DD AAAA BB BB AAAA SS EE
DD DD AA AA BB BB AA AA SS EE
DD DD AA AA BBBBBB AA AA SSSS EEEEE
DD DD AAAAAAA BB BB AAAAAAA SS EE
DD DD AA AA BB BB AA AA SS EE
DDDDD AA AA BBBBBB AA AA SSSSSS EEEEEEEE
DABASE, Routine to Locate Logical Group of Named Arrays
------ ------- -- ------ ------- ----- -- ----- ------
DABASE locates a desired logical group of array names within
the dictionary which is created by DALOAD and which is used
by several other routines in the FASP package for the
manipulation of multiply subscripted arrays equivalenced
with or otherwise loaded into a singly subscripted buffer.
DABASE can identify the logical group by its name, by its
position in the dictionary relative to the other logical
groups, or by a particular value being stored as subscript
range information associated with the name.
After DABASE has located the logical group of names, the
routine DANAME can be called to identify the array name and
subscripts associated with a known location in the singly
subscripted buffer. The combination of DABASE and DANAME
enable a program to cycle through the data base in some
predetermined fashion. Typical applications of DABASE and
DANAME would be to allow a program to report the non-zero
values in its common blocks, or in the writing of a program
to compare the values in data files containing different
versions of the values to be loaded into such common blocks.
The DABASE Argument List
--- ------ -------- ----
The argument list of routine DABASE is
SUBROUTINE DABASE(LOCATE,LTRLFT,LTRRIT,LTRNAM,IVALUE,
1 LTRLOW,LTRUSD,LTRSTR,NUMLOW,NUMUSD,NUMSTR,LRGLTR,
2 LRGNUM,LRGKNT)
with the associated DIMENSION statement
DIMENSION LTRNAM(LTRRIT),LTRSTR(LTRUSD),NUMSTR(NUMUSD)
The following are input arguments left unchanged.
LOCATE = identifies how the logical group of array names is
to be selected.
= -1, look for the logical group having as its name
FASP, FORTRAN Alphameric Subroutine Package Page 33
DABASE, Routine to Locate Logical Group of Named Arrays
the characters stored in LTRNAM(LTRLFT) through and
including LTRNAM(LTRRIT). The name of the logical
group is the name appearing between slashes at the
start of the specification of the contents of the
group as read by the DALOAD routine. The
characters in LTRNAM are stored 1 character per
array location as read by a multiple of A1 format
or defined by several 1H fields. The name in
LTRNAM(LTRLFT) through and including LTRNAM(LTRRIT)
must match the name of the logical group exactly.
Abbreviations are not recognized, and excess
characters prevent a match. If a particular
logical group is not named, then that logical group
cannot be identified in this manner.
= 0, look for the logical group the position of which
relative to all logical groups in the dictionary
matches the value input in argument IVALUE. If
IVALUE has the value 4, then the fourth logical
group in the dictionary would be searched for.
= greater than zero, search for a logical group
having a value equal to the input value of the
argument IVALUE in the subscript range storage
location identified by the value of LOCATE.
LOCATE=1 would test against the lower bound of the
first subscript, =2 against the upper bound of the
first subscript, and =4 against the upper bound of
the second subscript. If the subscript range
information associated with a particular logical
group contains less than LOCATE values, then that
logical group cannot be identified in this manner.
LTRLFT = subscript of the LTRNAM array location containing
the first character of the name to be matched
against if LOCATE=-1.
LTRRIT = subscript of the LTRNAM array location containing
the final character of the name to be matched
against if LOCATE=-1.
LTRNAM = array containing in locations LTRNAM(LTRLFT)
through and including LTRNAM(LTRRIT) the characters
of the name to be matched against if LOCATE=-1.
The characters in LTRNAM are stored 1 character per
array location as read by a multiple of an A1
format or defined by several 1H fields.
IVALUE = the sequence number of the logical group to be
found (the count relative to all logical groups in
the dictionary) if LOCATE=0.
= the value to be matched against in the subscript
range storage location identified by the value of
LOCATE if LOCATE is greater than zero.
LTRLOW = lowest subscript of the locations in the LTRSTR
FASP, FORTRAN Alphameric Subroutine Package Page 34
DABASE, Routine to Locate Logical Group of Named Arrays
array containing the characters of the names in the
dictionary as originally read by DALOAD using a
multiple of an A1 format. LTRSTR(LTRLOW) contains
either the first letter of the name of the first
logical group of names in the dictionary or else
(if the first group itself isn't named) the first
letter of the first name within the first logical
group in the dictionary.
LTRUSD = highest subscript of the locations in the LTRSTR
array containing the characters of the names in the
dictionary as originally read by DALOAD using a
multiple of an A1 format. LTRSTR(LTRUSD) contains
the last character of the last name in the
dictionary.
LTRSTR = array containing the characters forming the names
in the dictionary, 1 character per array location
as originally read by DALOAD using a multiple of an
A1 format.
NUMLOW = lowest subscript of the locations in the NUMSTR
array containing the numeric information
corresponding to the names stored in the LTRSTR
array. NUMSTR(NUMLOW) must contain the start of
the description of a logical group of names, not
the start of the description of an individual name.
NUMUSD = highest subscript of the locations in the NUMSTR
array containing the numeric information
corresponding to the names stored in the LTRSTR
array.
NUMSTR = array containing the numeric information
corresponding to the names stored in the LTRSTR
array. The construction of the NUMSTR array is
described in detail in the DALOAD documentation.
For each name in the dictionary, the NUMSTR array
contains
a. the number of characters in the name
b. an indication of the associated data type
c. the number of subscript ranges
d. pairs of starting and ending values of these
ranges.
If the number of characters is instead zero or
negative, then its absolute value is the number of
characters in the name of a logical group of names,
and the next location, rather than indicating the
data type, contains the number of locations within
FASP, FORTRAN Alphameric Subroutine Package Page 35
DABASE, Routine to Locate Logical Group of Named Arrays
a singly subscripted buffer which would be needed
to store the values of the multiply subscripted
arrays which are within the logical group and
equivalenced with or otherwise loaded into such a
singly subscripted buffer.
The following arguments are used for output from this
routine. Their input values are ignored.
LRGLTR = returned containing the subscript of the LTRSTR
array location which contains the first letter of
the name associated with the logical group of names
in the dictionary which has been identified by this
routine. If the logical group does not have a name
(NUMSTR(LRGNUM) being zero), then LRGLTR is
returned pointing to the LTRSTR array location
containing the first letter of the first name
within the group.
LRGNUM = returned containing the subscript of the NUMSTR
array location which contains the first of the
numeric information associated with the logical
group of names in the dictionary which was
identified by this routine. NUMSTR(LRGNUM)
contains as its absolute value the number of
characters starting at LTRSTR(LRGLTR) which are
contained in the name, if any, of the logical group
of names. NUMSTR(LRGNUM+1) contains the number of
locations within a singly subscripted buffer which
would be needed to store the values of the multiply
subscripted arrays which are within the logical
group and equivalenced with or otherwise loaded
into such a singly subscripted buffer.
LRGKNT = returned containing the sequence number within the
dictionary of the logical group of names identified
by this routine. If the third logical group in the
dictionary is identified, then LRGKNT would be
returned containing the value 3. If more than one
logical group matches the given specifications,
then the first group which matches the
specifications is that identified to the calling
program. LRGKNT is returned containing zero if no
logical group in the dictionary matched the given
specifications.
FASP, FORTRAN Alphameric Subroutine Package Page 36
DABELT, Routine to Represent Column Numbers
DDDDD AAA BBBBBB EEEEEEEE LL TTTTTTTT
DD DD AAAA BB BB EE LL TT
DD DD AA AA BB BB EE LL TT
DD DD AA AA BBBBBB EEEEE LL TT
DD DD AAAAAAA BB BB EE LL TT
DD DD AA AA BB BB EE LL TT
DDDDD AA AA BBBBBB EEEEEEEE LLLLLLLL TT
DABELT, Routine to Represent Column Numbers
------ ------- -- --------- ------ -------
When a program needs to display to the user a line of
characters in which the characters are oriented by columns,
a band of numbers identifying the columns can be printed,
either above or below the main display, by calling the
DABELT routine to generate each line of the representation
of the column numbers in a buffer array which the calling
program can then print with a multiple of an A1 format.
DABELT is called as many times as there are lines in the
representation of the column numbers, the calling program
printing the returned characters before again asking DABELT
to generate the next line.
The numbers can be generated with each line containing
digits corresponding to the same power of 10 as in the
following example
5432109876543210123456789012345
111111--------- 111111
------
or with each line containing digits of the same significance
as in the following example.
---------------0123456789111111
111111987654321 012345
543210
The numbers can decrease from left to right as in the
following example
5432109876543210123456789012345
111111 ---------111111
------
and can have any desired spacing and increment between
adjacent numbers as in the following example.
- - 0 1 2 3 4 5 6 7 8 9 1 1 1 1 1 1 1 1 1 1 2
2 1 0 0 0 0 0 0 0 0 0 0 1 2 3 4 5 6 7 8 9 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0
FASP, FORTRAN Alphameric Subroutine Package Page 37
DABELT, Routine to Represent Column Numbers
Two versions of the routine are provided. DABELT can
produce all of the format variations described above.
DAIBLT can only produce numbers of the form shown in the
first illustration above. Each line of the number
representation produced by DAIBLT contains digits
corresponding to the same power of 10, with adjacent numbers
increasing by an increment of one from left to right with no
extra spacing. Both DABELT and DAIBLT can produce any
selected line of the representation of the numbers, so
columns can be printed to be read either from top to bottom
or from bottom to top.
The DABELT and DAIBLT Argument Lists
--- ------ --- ------ -------- -----
The argument lists of routines DABELT and DAIBLT are
SUBROUTINE DABELT(KOLUMN,INTRVL,JSTIFY,LINE ,ILEFT ,
1 IRIGHT,LFTCOL,MAXBFR,IBUFFR,MAXLIN,MAXPRT,MAXUSD)
and
SUBROUTINE DAIBLT( LINE ,ILEFT ,
1 IRIGHT,LFTCOL,MAXBFR,IBUFFR,MAXLIN,MAXPRT,MAXUSD)
both with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The argument lists of the 2 routines are identical except
that the first 3 arguments in the DABELT argument list do
not appear in the DAIBLT argument list. For DAIBLT these
missing arguments effectively have the values KOLUMN=1,
INTRVL=1 and JSTIFY=0. Also, for DAIBLT, if ILEFT is
greater than IRIGHT, then no representation is generated,
and MAXPRT and MAXUSD are both returned set equal to the
input value of LFTCOL.
The following are input arguments left unchanged.
KOLUMN = number of columns to be used for a single number.
If KOLUMN is greater than 1, then KOLUMN-1 columns
of spaces will be inserted between the numbers. No
extra spaces will appear to the right of the right
number even if KOLUMN is greater than 1. The
effective value of KOLUMN is 1 if KOLUMN is input
less than or equal to zero.
INTRVL = difference between adjacent numbers which are to be
represented in the IBUFFR array. If the sign of
INTRVL is wrong to proceed from ILEFT to IRIGHT,
then the sign of the effective value of INTRVL is
FASP, FORTRAN Alphameric Subroutine Package Page 38
DABELT, Routine to Represent Column Numbers
reversed, but the sign and value of the argument
supplied by the calling program are left unchanged.
The effective value of INTRVL is 1 if INTRVL is
input equal to zero.
JSTIFY = 0, each line is to contain digits corresponding to
the same power of 10 as in the first example at the
start of the documentation of this routine. LINE
equal to 1 selects the digits of the lowest
significance in the numbers. LINE equal to MAXLIN
will select either the minus sign or the digits of
highest significance if positive in the number or
numbers requiring the most characters to represent.
= 1, each line is to contain digits of the same
significance as in the second example at the start
of the documentation of this routine. LINE equal 1
selects either the minus signs or the digits of
highest significance if positive in the numbers.
LINE equal to MAXLIN will select the digits of
lowest significance in the number or numbers
requiring the most characters to represent.
LINE = selects which line of the representation of the
numbers is to be constructed. LINE equal 1 will
select the digits of lowest significance if JSTIFY
is zero, or the minus signs or digits of highest
significance of positive numbers if JUSTIFY equals
1. LINE equal to MAXLIN will select the minus
signs or the digits of highest significance if
positive of the number or numbers requiring the
most characters to represent if JSTIFY is zero, or
the digits of lowest significance of the number or
numbers requiring the most characters to represent
if JSTIFY equals 1. MAXPRT and MAXUSD are both
returned equal to the input value of LFTCOL if LINE
is input greater than MAXLIN.
The following examples illustrate the definition of
LINE for JSTIFY equal to both zero and one.
KOLUMN=2,INTRVL=75,JSTIFY=0,ILEFT=-1052,IRIGHT=2000
LINE=1 2 7 2 7 2 7 2 7 2 7 2 7 2 7 2 3 8 3 8 3 8 3
LINE=2 5 7 0 2 5 7 0 2 5 7 0 2 5 7 - 7 4 2 9 7 4 2
LINE=3 0 9 9 8 7 6 6 5 4 3 3 2 1 - 1 2 2 3 4 5
LINE=4 1 - - - - - - - - - - - -
LINE=5 -
LINE=6
FASP, FORTRAN Alphameric Subroutine Package Page 39
DABELT, Routine to Represent Column Numbers
KOLUMN=2,INTRVL=75,JSTIFY=1,ILEFT=-1052,IRIGHT=2000
LINE=1 - - - - - - - - - - - - - - - 7 1 2 2 3 4 5
LINE=2 1 9 9 8 7 6 6 5 4 3 3 2 1 7 2 3 4 2 9 7 4 2
LINE=3 0 7 0 2 5 7 0 2 5 7 0 2 5 7 8 3 8 3 8 3
LINE=4 5 7 2 7 2 7 2 7 2 7 2 7 2
LINE=5 2
LINE=6
Since the effective value of MAXLIN is not known
prior to the first call to this routine, LINE can
be set to zero to represent the same line as if
LINE was input equal to the returned value of
MAXLIN. LINE is returned unchanged, so the calling
program would in this case have to set LINE equal
to the returned value of MAXLIN-1 prior to the
second call to this routine.
ILEFT = the left or first number to be represented.
IRIGHT = the right or final limit of the numbers to be
represented. Unlike ILEFT which is always
represented, IRIGHT is represented only if the
buffer is large enough to include the numbers
through IRIGHT plus the extra spaces if KOLUMN is
greater than 1, and if IRIGHT-ILEFT is exactly a
whole number multiple of INTRVL. If IRIGHT-ILEFT
is not exactly a whole number multiple of INTRVL,
then the rightmost number which could be
represented if the buffer is large enough is the
number which is equal to ILEFT plus the next
smaller whole number multiple of INTRVL.
LFTCOL = the subscript of the IBUFFR array location to the
immediate left of the location into which is to be
placed the digit or sign forming the representation
upon the current line of the left number.
MAXBFR = subscript of the highest location in the IBUFFR
array into which can be placed the representations
of the numbers from ILEFT through IRIGHT. This
would normally be the dimension of the IBUFFR
array.
The following arguments are used for output. Their input
values are ignored.
IBUFFR = array in which the numbers are to be represented
and which can then be printed by the calling
program using a multiple of an A1 format.
MAXLIN = returned containing the number of lines needed to
represent the numbers ILEFT through IRIGHT. The
actual number of lines which would include printing
characters may be less since the right number
FASP, FORTRAN Alphameric Subroutine Package Page 40
DABELT, Routine to Represent Column Numbers
actually displayed can require fewer characters for
its representation than would IRIGHT.
MAXPRT = returned containing the subscript of the right
location in the IBUFFR array containing a printing
character generated by this routine. If the
current call to this routine has not added any
printing characters to IBUFFR, then MAXPRT will be
returned equal to LFTCOL.
MAXUSD = returned containing the subscript of the right
location in the IBUFFR array containing any
character generated by this routine. If LINE is
less than or equal to MAXLIN, then MAXUSD will be
returned equal to the value of MAXPRT which would
be returned if LINE was input as 1. Since the
rightmost printing character generated by the
current call to this routine can be to the left of
that generated if LINE is 1, the array locations
starting at IBUFFR(MAXPRT+1) through IBUFFR(MAXUSD)
will contain spaces. If LINE is greater than the
returned value of MAXLIN, then MAXUSD is returned
equal to LFTCOL.
FASP, FORTRAN Alphameric Subroutine Package Page 41
DABELT, Routine to Represent Column Numbers
DABELT output for Various Values of Arguments and JSTIFY=0
------ ------ --- ------- ------ -- --------- --- ------ -
KOLUMN= 1, INTRVL= 1, ILEFT= -25, IRIGHT= 25
LINE=3 ----------------
LINE=2 2222221111111111--------- 1111111111222222
LINE=1 543210987654321098765432101234567890123456789012345
KOLUMN= 1, INTRVL= 1, ILEFT= 25, IRIGHT= -25
LINE=3 ----------------
LINE=2 2222221111111111 ---------1111111111222222
LINE=1 543210987654321098765432101234567890123456789012345
KOLUMN= 1, INTRVL= 5, ILEFT= -125, IRIGHT= 125
LINE=4 ------
LINE=3 111111------------------ 111111
LINE=2 221100998877665544332211- 112233445566778899001122
LINE=1 505050505050505050505050505050505050505050505050505
KOLUMN= 2, INTRVL= 10, ILEFT= -125, IRIGHT= 125
LINE=4 - - -
LINE=3 1 1 1 - - - - - - - - - 1 1 1
LINE=2 2 1 0 9 8 7 6 5 4 3 2 1 - 1 2 3 4 5 6 7 8 9 0 1 2
LINE=1 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5 5
KOLUMN= 2, INTRVL= 10, ILEFT= -250, IRIGHT= 0
LINE=4 - - - - - - - - - - - - - - - -
LINE=3 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 - - - - - - - - -
LINE=2 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1
LINE=1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
KOLUMN= 2, INTRVL= 10, ILEFT= 0, IRIGHT= 250
LINE=3 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2
LINE=2 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5
LINE=1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
KOLUMN= 5, INTRVL= 6, ILEFT= -30, IRIGHT= 30
LINE=3 - - - -
LINE=2 3 2 1 1 - 1 1 2 3
LINE=1 0 4 8 2 6 0 6 2 8 4 0
KOLUMN= 5, INTRVL= 6, ILEFT= -29, IRIGHT= 31
LINE=3 - - - -
LINE=2 2 2 1 1 - 1 1 2 3
LINE=1 9 3 7 1 5 1 7 3 9 5 1
FASP, FORTRAN Alphameric Subroutine Package Page 42
DABELT, Routine to Represent Column Numbers
An Example of the Use of DABELT
-- ------- -- --- --- -- ------
The sample program listed on the following page demonstrates
the manner in which DABELT is used. The program asks the
user for the desired spacing, interval, and left and right
limits, then generates the highest numbered line through
line 1 and back again to and beyond the highest numbered
line for JSTIFY equal first to zero then to one.
A sample dialog between the program and the user is
presented following the listing of the program.
For a practical application of the DABELT routine, see the
first demonstration program in the description of the DALOAD
routine.
FASP, FORTRAN Alphameric Subroutine Package Page 43
DABELT, Routine to Represent Column Numbers
DATA MAXBFR,ITTY,ISPACE/51,5,1H /
DIMENSION IBUFFR(51),JBUFFR(51)
C
C GET VALUES OF ARGUMENTS FROM USER
1 WRITE(ITTY,2)
2 FORMAT(' LFTCOL, KOLUMN, INTRVL, ILEFT, IRIGHT ',$)
READ(ITTY,3)LFTCOL,KOLUMN,INTRVL,ILEFT,IRIGHT
3 FORMAT(10I)
C
C REPRESENT THE NUMBERS
DO 4 I=1,MAXBFR
4 IBUFFR(I)=ISPACE
DO 9 JSTIFY=0,1
WRITE(ITTY,5)JSTIFY
5 FORMAT(' JSTIFY=',1I1)
LINE=0
6 CALL DABELT(KOLUMN,INTRVL,JSTIFY,LINE,ILEFT,
1IRIGHT,LFTCOL,MAXBFR,IBUFFR,MAXLIN,MAXPRT,MAXUSD)
IF(LINE.EQ.0)LINE=MAXLIN
WRITE(ITTY,7)LINE,(IBUFFR(I),I=1,MAXPRT)
7 FORMAT(' LINE=',I1,3X,100A1)
LINE=LINE-1
IF(LINE.GT.0)GO TO 6
LINE=1
8 LINE=LINE+1
CALL DABELT(KOLUMN,INTRVL,JSTIFY,LINE,ILEFT,
1IRIGHT,LFTCOL,MAXBFR,IBUFFR,MAXLIN,MAXPRT,MAXUSD)
IF(MAXPRT.LE.LFTCOL)WRITE(ITTY,7)LINE
IF(MAXPRT.GT.LFTCOL)WRITE(ITTY,7)LINE,
1(IBUFFR(I),I=1,MAXPRT)
IF(LINE.LE.MAXLIN)GO TO 8
9 CONTINUE
C
C REPORT THE EXPECTED VALUES
KOUNT=0
IF(INTRVL.LT.0)INTRVL=-INTRVL
IF(INTRVL.EQ.0)INTRVL=1
IF(ILEFT.GT.IRIGHT)INTRVL=-INTRVL
IF(KOLUMN.LE.0)KOLUMN=1
10 LFTCOL=LFTCOL+1
IF(LFTCOL.GT.MAXBFR)GO TO 13
IF(INTRVL.GT.0)GO TO 11
IF(ILEFT.LT.IRIGHT)GO TO 13
GO TO 12
11 IF(ILEFT.GT.IRIGHT)GO TO 13
12 KOUNT=KOUNT+1
JBUFFR(KOUNT)=ILEFT
LFTCOL=LFTCOL+KOLUMN-1
ILEFT=ILEFT+INTRVL
GO TO 10
13 IF(KOUNT.GT.0)WRITE(ITTY,14)(JBUFFR(I),I=1,KOUNT)
14 FORMAT(6I10)
GO TO 1
END
FASP, FORTRAN Alphameric Subroutine Package Page 44
DABELT, Routine to Represent Column Numbers
Typical Dialog Between DABELT Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
LFTCOL, KOLUMN, INTRVL, ILEFT, IRIGHT 0 2 6 108 -108
JSTIFY=0
LINE=4
LINE=3 1 1 - - - - - -
LINE=2 0 0 9 9 8 7 7 6 6 5 4 4 3 3 2 1 1 - 1 1 2 3 3 4
LINE=1 8 2 6 0 4 8 2 6 0 4 8 2 6 0 4 8 2 6 0 6 2 8 4 0 6 2
LINE=2 0 0 9 9 8 7 7 6 6 5 4 4 3 3 2 1 1 - 1 1 2 3 3 4
LINE=3 1 1 - - - - - -
LINE=4
LINE=5
JSTIFY=1
LINE=4
LINE=3 8 2 2 8 4 0 6 2
LINE=2 0 0 6 0 4 8 2 6 0 4 8 2 6 0 4 8 2 6 1 1 2 3 3 4
LINE=1 1 1 9 9 8 7 7 6 6 5 4 4 3 3 2 1 1 6 0 - - - - - - -
LINE=2 0 0 6 0 4 8 2 6 0 4 8 2 6 0 4 8 2 6 1 1 2 3 3 4
LINE=3 8 2 2 8 4 0 6 2
LINE=4
LINE=5
108 102 96 90 84 78
72 66 60 54 48 42
36 30 24 18 12 6
0 -6 -12 -18 -24 -30
-36 -42
LFTCOL, KOLUMN, INTRVL, ILEFT, IRIGHT 0 2 6 -108 108
JSTIFY=0
LINE=4 - -
LINE=3 1 1 - - - - - - - - - - - - - - -
LINE=2 0 0 9 9 8 7 7 6 6 5 4 4 3 3 2 1 1 - 1 1 2 3 3 4
LINE=1 8 2 6 0 4 8 2 6 0 4 8 2 6 0 4 8 2 6 0 6 2 8 4 0 6 2
LINE=2 0 0 9 9 8 7 7 6 6 5 4 4 3 3 2 1 1 - 1 1 2 3 3 4
LINE=3 1 1 - - - - - - - - - - - - - - -
LINE=4 - -
LINE=5
JSTIFY=1
LINE=4 8 2
LINE=3 0 0 6 0 4 8 2 6 0 4 8 2 6 0 4 8 2
LINE=2 1 1 9 9 8 7 7 6 6 5 4 4 3 3 2 1 1 6 2 8 4 0 6 2
LINE=1 - - - - - - - - - - - - - - - - - - 0 6 1 1 2 3 3 4
LINE=2 1 1 9 9 8 7 7 6 6 5 4 4 3 3 2 1 1 6 2 8 4 0 6 2
LINE=3 0 0 6 0 4 8 2 6 0 4 8 2 6 0 4 8 2
LINE=4 8 2
LINE=5
-108 -102 -96 -90 -84 -78
-72 -66 -60 -54 -48 -42
-36 -30 -24 -18 -12 -6
0 6 12 18 24 30
36 42
FASP, FORTRAN Alphameric Subroutine Package Page 45
DACASE, Converts Lower Case Letters to Upper Case
DDDDD AAA CCCCC AAA SSSSSS EEEEEEEE
DD DD AAAA CC AAAA SS EE
DD DD AA AA CC AA AA SS EE
DD DD AA AA CC AA AA SSSS EEEEE
DD DD AAAAAAA CC AAAAAAA SS EE
DD DD AA AA CC AA AA SS EE
DDDDD AA AA CCCCC AA AA SSSSSS EEEEEEEE
DACASE, Converts Lower Case Letters to Upper Case
------ -------- ----- ---- ------- -- ----- ----
Although lower case alphabetic letters are treated as
equivalent to the upper case forms of these letters by the
FASP routines DAHEFT, DAIHFT and DAVERB (and by any other
FASP routines such as DAHEST and DANEXT which call DAHEFT,
DAIHFT or DAVERB either directly or indirectly), lower case
letters are not recognized by the FASP routines DALOAD,
DAPICK, DATEXT and DATURN. If these latter routines are
called upon a computer system which supports lower case, but
requiring that the user lock the terminal into upper case
shift is not acceptable, then the contents of the input text
buffer can instead be processed by the DACASE routine which
replaces lower case letters with the corresponding upper
case letters. All other characters are returned unchanged
by the DACASE routine.
DACASE contains a list of the lower case letters and a list
of the corresponding upper case letters. The lower case
letters are defined by a DATA statement to be in the order
1Ha through 1Hz which on the PDP10 computer results in the
associated integer values being sorted into an increasing
order. Each lower case letter in the input text buffer is
identified by a ternary search for a match within the sorted
list of lower case letters. If this routine is used upon a
computer system in which the alphabetical order of the
letters 1Ha through 1Hz does not result in an increasing
order for the associated integer values, then the DATA
statement which defines the array which contains the lower
case letters should be rewritten so that the values
associated with the letters will be in increasing numerical
order, and then the DATA statement which defines the
corresponding upper case letters must be rewritten so that
the lower and upper case versions of each letter appear in
locations in the respective arrays having the same
subscripts.
Operation of the DACASE routine requires that the compiler
not translate into upper case any lower case letters
included in Hollerith strings in the routine. On the PDP10
computer, this routine can be used with the F10 compiler,
but not with the older F40 compiler which does such a
translation of lower case letters in Hollerith strings into
upper case. Of course, this routine is not needed at all if
FASP, FORTRAN Alphameric Subroutine Package Page 46
DACASE, Converts Lower Case Letters to Upper Case
the run time system itself translates lower case letters to
upper case before these are sent to the user's program.
The DACASE Argument List
--- ------ -------- ----
The argument list of routine DACASE is
SUBROUTINE DACASE(MINBFR,MAXBFR,IBUFFR)
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The following arguments are used for input and are returned
unchanged.
MINBFR = subscript of the first location in the IBUFFR array
containing a character which is to be converted to
upper case if input in lower case.
MAXBFR = subscript of the final location in the IBUFFR array
containing a character which is to be converted to
upper case if input in lower case.
The following argument is used both for input to and for
output from this routine.
IBUFFR = array containing in locations having subscripts
MINBFR through MAXBFR the characters read by a
multiple of an 1A format which are to be converted
to upper case if input in lower case.
FASP, FORTRAN Alphameric Subroutine Package Page 47
DACOPY, Routine to Expand Tab Characters to Spaces
DDDDD AAA CCCCC OOOOO PPPPPP YY YY
DD DD AAAA CC OO OO PP PP YY YY
DD DD AA AA CC OO OO PP PP YYYY
DD DD AA AA CC OO OO PPPPPP YY
DD DD AAAAAAA CC OO OO PP YY
DD DD AA AA CC OO OO PP YY
DDDDD AA AA CCCCC OOOOO PP YY
DACOPY, Routine to Expand Tab Characters to Spaces
------ ------- -- ------ --- ---------- -- ------
Computer terminals often can generate 2 types of forward
spacing but nonprinting characters. The simpler of these,
the space, merely causes the character to its right to be
shifted over by 1 column and is exactly equivalent to an
empty column on a punched card. The tab character
(sometimes called HT or horizontal tab), however, shifts the
character to its right beyond the next column position which
is a whole multiple of 8 (or some other, generally fixed tab
stop interval), and is convenient for quickly spacing across
the width of the input line. The reading program can
distinguish a tab character from a space, so appropriate
treatment of the tab character poses little or no difficulty
to a program which is doing parsing or which is reading
numeric information, since the tab character can be treated
as equivalent to a space or not, whichever is appropriate to
the particular application.
If the program must insert text containing tab characters
into printed or typed output, the original alignment of the
text with the tab stop columns will probably not be
maintained, causing the appearance of the text upon output
to be quite different from that seen on the original
terminal from which the text was entered into the computer
system. Also, when this alignment changes, the effective
number of columns across the width of the text changes, so
that information inserted into fields to the right of the
text containing the tab characters will also be shifted into
different columns than intended. This difficulty can be
avoided by conversion of the tab characters into the correct
number of spaces prior to output.
DACOPY is one of two routines in the FASP package which
convert tab characters to spaces in buffers read by an A1
format. DACOPY performs the tab character to space
expansion while copying from an input buffer to a separate
output buffer which can be written by the calling program
with a multiple of an A1 format. The expanded text is not
copied back into the original buffer since the number of
characters can increase during the conversion. DAFILL, the
other tab to space conversion routine, copies the expanded
text back into the input buffer, overwriting any characters
which could not be represented due to the buffer being too
small.
FASP, FORTRAN Alphameric Subroutine Package Page 48
DACOPY, Routine to Expand Tab Characters to Spaces
the DACOPY Argument List
--- ------ -------- ----
The argument list of routine DACOPY is
SUBROUTINE DACOPY(INITAL,INTRVL,IBUFFR,IBEGIN,IFINAL,
1JFINAL,JUSED,JBUFFR,NXTINI,NXTBGN,MAXPRT)
with the associated DIMENSION statement
DIMENSION IBUFFR(IFINAL),JBUFFR(JFINAL)
The arguments INITAL, INTRVL, IBUFFR, IBEGIN, IFINAL, and
JFINAL are used only for input and their original values are
returned unchanged. The argument JUSED is used for both
input and output. The portion of the JBUFFR array above
JBUFFR(JUSED) as well as the arguments NXTINI, NXTBGN and
MAXPRT are used only for output and their original values
are ignored.
INITAL = less than zero and providing that INTRVL is greater
than zero, the absolute value of INITAL is the
number of extra spaces to be inserted at the start
of the output buffer before the contents of the
input buffer are copied into the output buffer. If
INITAL is less than zero, regardless of whether
INTRVL is positive or negative, the first tab stop
will be of the width specified by the absolute
value of INTRVL. If INTRVL is less than or equal
to zero, then no leading spaces will be inserted
into the output buffer whether requested by a
negative value of INITAL or by leading spaces or
tab characters in the input buffer.
= equal to or greater than zero, INITAL is the
distance to the first tab stop. (Distance is here
defined as a number of characters or, since the
buffers contain a single character in each location
of the arrays, as a subscript range.) If INITAL is
zero, then it is assumed that copying has already
passed beyond the first tab stop and the distance
to the next tab stop is then taken as the absolute
value of INTRVL. If the left tab stop is to be of
the same width as those to its right, then INITAL
can equal either zero or the absolute value of
INTRVL. If the first character in the input buffer
is a tab character, then it will be expanded to the
number of spaces specified by INITAL (or by a
positive value of INTRVL if INITAL is zero). If
the first character is not a tab character, but the
second character is a tab character, then the
second character will be expanded to INITAL-1
spaces, and so on through the first INITAL
characters providing in each case that none of the
characters to the left are tab characters. After
FASP, FORTRAN Alphameric Subroutine Package Page 49
DACOPY, Routine to Expand Tab Characters to Spaces
the sum of the number of characters which have been
inserted into the output buffer, whether these
characters are printing characters or spaces
originally in the input buffer or spaces resulting
from the expansion of tab characters, and, if
INTRVL is equal to or less than zero, of the number
of spaces which have been suppressed, equals the
value of INITAL, then the tab stop interval becomes
that given by the absolute value of the argument
INTRVL.
INTRVL = not equal to zero, the absolute value of INTRVL is
the tab stop interval. On the PDP-10 computer, the
appropriate absolute value of INTRVL is 8, and,
assuming that the first character in the input
buffer corresponds to column 1, the value of INITAL
at the start of the processing of the contents of
the input buffer can be either 0 or 8. After the
sum of the number of characters inserted into the
output buffer and of the number of spaces, if any,
suppressed by a negative or zero value of INTRVL,
has reached at least the absolute value of INITAL,
then a tab character encountered in the input
buffer causes sufficient spaces to be inserted into
the output buffer or else to be suppressed so that
the total number of characters inserted into the
output buffer or suppressed by the present call to
DACOPY equals the sum of the absolute value of
INITAL and the next higher whole multiple of the
absolute value of INTRVL.
= less than zero, no leading spaces are to be
inserted into the output buffer whether such
leading spaces are requested by a negative value of
INITAL or by leading spaces or tab characters in
the input buffer. Once a printing character has
been copied into the output buffer, however, then
all subsequent spaces will be copied and all
subsequent tab characters will be expanded to
spaces.
= zero, no spaces are to be inserted into the output
buffer. Tab characters and spaces in the input
buffer are ignored.
= greater than zero, all spaces whether requested by
a negative value of INITAL or by spaces or tab
characters in the input buffer are to be inserted
into the output buffer.
IBUFFR = the input buffer containing the characters which
are to be copied into the output buffer expanding
any tabs found. The characters in the IBUFFR array
must have been read by a multiple of an A1 format.
IBEGIN = the subscript of the IBUFFR array location at which
is to be found the first character to be copied.
FASP, FORTRAN Alphameric Subroutine Package Page 50
DACOPY, Routine to Expand Tab Characters to Spaces
IFINAL = the subscript of the IBUFFR array location at which
is to be found the final character to be copied.
JFINAL = the dimension of, or subscript of the highest
available location within, the JBUFFR array.
JUSED = the subscript of the lowest location in the JBUFFR
array which is currently in use and which contains
data which must be maintained. JUSED is returned
containing the subscript of the highest location
into which DACOPY has placed a character.
JBUFFR = the array into which the contents of IBUFFR are to
be copied expanding tab characters to spaces.
NXTINI = returned containing the value which should be next
given to INITAL if the current call could not
completely represent the contents of the input
buffer due to the room available in the output
buffer being to small. If a tab character was
encountered in the input buffer but could not be
completely represented in the output buffer, then
NXTINI will be returned negative. If the last
character encountered in the input buffer was not a
tab character, then NXTINI will be returned
containing the remaining distance to the next tab
stop.
NXTBGN = returned containing the subscript within the IBUFFR
array of the first character which could not be
represented in the output buffer. If all
characters could be represented, then NXTBGN will
be returned equal to IFINAL+1. Note that if a tab
character is represented even by single space, then
NXTBGN is passed beyond this tab character although
there might not be enough room in the output buffer
to fill completely to the next tab stop.
MAXPRT = returned containing the subscript of the highest
location in the JBUFFR array into which DACOPY has
placed a printing (nonspace) character. If no
printing characters are placed into the output
buffer, then MAXPRT will be returned equal to the
input value of JUSED.
FASP, FORTRAN Alphameric Subroutine Package Page 51
DACOPY, Routine to Expand Tab Characters to Spaces
Examples of Tab Character Expansions
-------- -- --- --------- ----------
In the following examples of the expansion of tabs to spaces
with a tab stop interval of 8 (INITAL=INTRVL=8), the symbol
<HT> represents a single tab character stored in a single
computer word, and the digits represent characters other
than the tab character stored one per computer word. If,
according to this notation, the input buffer contains
1<HT>90<HT>789<HT>5678<HT>34567<HT>123456<HT>9012345<HT>7
then after expansion, the output buffer would contain
1 90 789 5678 34567 123456 9012345 7
In similar manner, the input buffer contents
1234567<HT>901234<HT>78901<HT>5678<HT>345<HT>12<HT>9<HT>7
would be changed to the following in the output buffer
1234567 901234 78901 5678 345 12 9 7
An Example of the Use of DACOPY
-- ------- -- --- --- -- ------
The sample program listed on the following page demonstrates
the manner in which DACOPY is used. The program prompts the
user with an asterisk, then reads a line of text including
tabs to be expanded to the corresponding number of spaces.
The tab stop interval is 8. Since the asterisk typed by the
program appears in column 1, an initial tab would only
correspond to an additional 7 spaces.
The program allocates the output buffer in sections,
allowing as much as possible of the input buffer to be
represented in a particular section, then moving to the next
section of the output buffer and again calling DACOPY. This
sequence is repeated until either the output buffer fills or
the input buffer is exhausted. If the rightmost character
of the input buffer which has been represented in the output
section is a tab, then the location of the rightmost
character of the output section is marked with a digit
corresponding to the position of the tab relative to the
other tabs in the input buffer. The program displays the
expansion for output buffer section sizes 1, 5, 9, 13, 17
and 21.
A sample dialog between the program and the user is
presented following the listing of the program.
FASP, FORTRAN Alphameric Subroutine Package Page 52
DACOPY, Routine to Expand Tab Characters to Spaces
DIMENSION IBUFFR(70),JBUFFR(85),IDIGIT(10)
DATA INTRVL,IWIDTH,IFINAL,MAXOUT/8,59,70,85/
DATA IDIGIT/1H1,1H2,1H3,1H4,1H5,1H6,1H7,1H8,1H9,1H0/
DATA ITAB/1H /
C ASK FOR AND READ TEXT CONTAINING TABS TO BE CONVERTED
LOWER=IFINAL-IWIDTH+1
ISTART=MAXOUT-IWIDTH+1
1 TYPE 2
2 FORMAT(1X/' *',$)
ACCEPT 3,(IBUFFR(I),I=LOWER,IFINAL)
3 FORMAT(60A1)
C LOOP THRU DESTINATION FIELD SIZES 1 TO 21, INCREMENT 4
DO 9 ISIZE=1,21,4
IBEGIN=LOWER
JUSED=ISTART-1
INITAL=INTRVL-1
LIMIT=JUSED
4 JFINAL=JUSED+ISIZE
IF(JFINAL.GT.MAXOUT)JFINAL=MAXOUT
JSTART=JUSED+1
C CONVERT TAB CHARACTERS INTO SPACES
CALL DACOPY(INITAL,INTRVL,IBUFFR,IBEGIN,IFINAL,JFINAL,
1JUSED,JBUFFR,NXTINI,NXTBGN,MAXPRT)
IF(JUSED.LT.JSTART)GO TO 8
IF(MAXPRT.GE.JSTART)LIMIT=MAXPRT
C MARK RIGHT END OF AREA INTO WHICH TAB WAS EXPANDED
IF(IBUFFR(NXTBGN-1).NE.ITAB)GO TO 7
KOUNT=0
INDEX=LOWER-1
5 INDEX=INDEX+1
IF(INDEX.GE.NXTBGN)GO TO 6
IF(IBUFFR(INDEX).NE.ITAB)GO TO 5
KOUNT=KOUNT+1
IF(KOUNT.GT.10)KOUNT=1
GO TO 5
6 JBUFFR(JUSED)=IDIGIT(KOUNT)
C PREPARE FOR NEXT AREA INTO WHICH TAB CAN BE EXPANDED
7 INITAL=NXTINI
IBEGIN=NXTBGN
GO TO 4
C REPORT RESULTS TO USER
8 IF(LIMIT.LT.ISTART)GO TO 1
9 TYPE 10,(JBUFFR(I),I=ISTART,LIMIT)
10 FORMAT(2X,200A1)
C IDENTIFY COLUMNS
LINE=1
IRIGHT=LIMIT-ISTART+2
11 CALL DABELT(1,1,0,LINE,2,
1IRIGHT,0,MAXOUT,JBUFFR,MAXLIN,MAXPRT,MAXUSD)
TYPE 10,(JBUFFR(I),I=1,MAXPRT)
LINE=LINE+1
IF(LINE.LE.MAXLIN)GO TO 11
GO TO 1
END
FASP, FORTRAN Alphameric Subroutine Package Page 53
DACOPY, Routine to Expand Tab Characters to Spaces
Typical Dialog Between DACOPY Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
The first line typed by the user consisted merely of 7 tab
characters followed by the letter A. The second line
consists of alternating tabs and letters. The third line
consists of several repetitions of a sequence formed of
several spaces, a single letter and a single tab. The final
line consists of a random sequence of tabs, spaces and
letters.
* A
1111111222222223333333344444444555555556666666677777777A
1 2 2 3 4 4 5 6 6 7 7A
2 3 4 5 6 7 A
2 4 5 7 A
3 5 7 A
3 6 A
23456789012345678901234567890123456789012345678901234567
111111111122222222223333333333444444444455555555
* A B C D E F G
1111111A2222222B3333333C4444444D5555555E6666666F7777777G
1 A 2 2B 3 C4 4 D 5 E 6 F 7 7G
A2 B 3 C 4 D 5 E 6 F 7 G
A 2 B C 4 D 5E F 7 G
A B3 C D 5 E F 7 G
A B 3 C D E 6 F G
23456789012345678901234567890123456789012345678901234567
111111111122222222223333333333444444444455555555
* A B C D E F G
A111 B222 C333 D444 E555 F666 G
A1 B 2 C D 4 E F6 G
A B C D E F6 G
A B2 C D E 5 F G
A B C D E F G
A B C3 D E F G
2345678901234567890123456789012345678901234567890123
11111111112222222222333333333344444444445555
* A B C D EF GH I J K
1111111 222222A 3333B C 4444 D 5555 EF66666GH I7777J K
1 2 2A 3 B C 4 D EF 6 GH I 7J K
A B C D 5 EF 6 GH I 7 J K
2 A B C D 5 EF GH I7 J K
A B C D EF GH I J K
A 3 B C D EF GH I J K
2345678901234567890123456789012345678901234567890123456789
11111111112222222222333333333344444444445555555555
FASP, FORTRAN Alphameric Subroutine Package Page 54
DAFILL, Routine to Expand Tab Characters to Spaces
DDDDDDD AAA FFFFFFFFF IIIII LLL LLL
DDD DDD AAA AAA FFF III LLL LLL
DDD DDD AAA AAA FFFFFF III LLL LLL
DDD DDD AAAAAAAAA FFF III LLL LLL
DDDDDDD AAA AAA FFF IIIII LLLLLLLLL LLLLLLLLL
DAFILL, Routine to Expand Tab Characters to Spaces
------ ------- -- ------ --- ---------- -- ------
Computer terminals often can generate 2 types of forward
spacing but nonprinting characters. The simpler of these,
the space, merely causes the character to its right to be
shifted over by 1 column and is exactly equivalent to an
empty column on a punched card. The tab character
(sometimes called HT or horizontal tab), however, shifts the
character to its right beyond the next column position which
is a whole multiple of 8 (or some other, generally fixed tab
stop interval), and is convenient for quickly spacing across
the width of the input line. The reading program can
distinguish a tab character from a space, so appropriate
treatment of the tab character poses little or no difficulty
to a program which is doing parsing or which is reading
numeric information, since the tab character can be treated
as equivalent to a space or not, whichever is appropriate to
the particular application.
If the program must insert text containing tab characters
into printed or typed output, the original alignment of the
text with the tab stop columns will probably not be
maintained, causing the appearance of the text upon output
to be quite different from that seen on the original
terminal from which the text was entered into the computer
system. Also, when this alignment changes, the effective
number of columns across the width of the text changes, so
that information inserted into fields to the right of the
text containing the tab characters will also be shifted into
different columns than intended. This difficulty can be
avoided by conversion of the tab characters into the correct
number of spaces prior to output.
DAFILL is one of two routines in the FASP package which
convert tab characters to spaces in buffers read by an A1
format. DAFILL copies the expanded text back into the input
buffer, overwriting any characters which could not be
represented due to the buffer being too small. DACOPY, the
other tab to space conversion routine, copies the input
buffer into a separate output buffer and provides the
calling program with sufficient information to properly
continue the expansion of any text which might remain
unprocessed due to the output buffer being too small.
DAFILL requires more execution time since the processing is
iterative, and is a longer routine, but it does not require
the extra output buffer.
FASP, FORTRAN Alphameric Subroutine Package Page 55
DAFILL, Routine to Expand Tab Characters to Spaces
A shorter version of DAFILL, named DAIFLL, is also provided
which does not provide for the insertion of extra initial
spaces other than those resulting from the expansion of
initial tabs, and which does not provide for the suppression
of either tabs or spaces.
the DAFILL and DAIFLL Argument Lists
--- ------ --- ------ -------- -----
The argument lists of routines DAFILL and DAIFLL are
SUBROUTINE DAFILL(INITAL,INTRVL,IBEGIN,IFINAL,MAXBFR,
1IBUFFR,MAXPRT,MAXUSD)
and
SUBROUTINE DAIFLL(INITAL,INTRVL,IBEGIN,IFINAL,MAXBFR,
1IBUFFR,MAXPRT,MAXUSD)
both with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The argument lists of the 2 routines are identical except
that, for DAIFLL, the argument INITAL must be greater than
or equal to zero, and the argument INTRVL must be greater
than zero.
The arguments INITAL, INTRVL, IBEGIN, IFINAL, and MAXBFR are
used only for input and their original values are returned
unchanged. The array IBUFFR is used for both input and
output. The arguments MAXPRT and MAXUSD are used only for
output and their original values are ignored.
INITAL = less than zero and providing that INTRVL is greater
than zero, the absolute value of INITAL is the
number of extra spaces to be inserted at the start
of the processed text. If INITAL is less than
zero, regardless of whether INTRVL is positive or
negative, then the first tab stop will be of the
width specified by the absolute value of INTRVL.
If INTRVL is less than or equal to zero, then no
leading spaces will be inserted into the processed
text whether requested by a negative value of
INITAL or by leading spaces or tab characters in
the original text.
= equal to or greater than zero, INITAL is the
distance to the first tab stop. (Distance is here
defined as a number of characters or, since the
buffers contain a single character in each location
of the array, as a subscript range.) If INITAL is
zero, then the distance to the first tab stop is
FASP, FORTRAN Alphameric Subroutine Package Page 56
DAFILL, Routine to Expand Tab Characters to Spaces
taken to be the absolute value of INTRVL. If the
left tab stop is to be same width as those to its
right, then INITAL can equal either zero or the
absolute value of INTRVL. If the first character
in the original text is a tab character, then it
will be expanded to the number of spaces specified
by INITAL (or by a positive value of INTRVL if
INITAL is zero). If the first character is not a
tab character, but the second character is a tab
character, then the second character will be
expanded to INITAL-1 spaces, and so on through the
first INITAL characters providing in each case that
none of the characters to the left are tab
characters. After the sum of the number of
characters which have been inserted into the
processed text, whether these characters are
printing characters or spaces in the original text
or spaces resulting from the expansion of tab
characters, and, if INTRVL is equal to or less than
zero, of the number of spaces which have been
suppressed, equals the value of INITAL, then the
tab stop interval becomes that given by the
absolute value of the argument INTRVL.
INTRVL = not equal to zero, the absolute value of INTRVL is
the tab stop interval. On the PDP-10 computer, the
appropriate absolute value of INTRVL is 8, and,
assuming that the first character in the buffer
corresponds to column 1, the value of INITAL at the
start of the processing of the contents of the
original text can be either 0 or 8. After the sum
of the number of characters inserted into the
processed text and of the number of spaces, if any,
suppressed by a negative or zero value of INTRVL,
has reached at least the absolute value of INITAL,
then a tab character encountered in the original
text causes sufficient spaces to be inserted into
the processed text or else to be suppressed so that
the total number of characters inserted into the
processed text or suppressed by the present call to
DAFILL equals the sum of the absolute value of
INITAL and the next higher whole multiple of the
absolute value of INTRVL.
= less than zero, no leading spaces are to be
inserted into the processed text whether such
leading spaces are requested by a negative value of
INITAL or by leading spaces or tab characters in
the original text. Once a printing character has
been copied into the processed text, however, then
all subsequent spaces will be copied and all
subsequent tab characters will be expanded to
spaces.
= zero, no spaces are to be inserted into the
processed text. Tab characters and spaces in the
FASP, FORTRAN Alphameric Subroutine Package Page 57
DAFILL, Routine to Expand Tab Characters to Spaces
original text are ignored.
= greater than zero, all spaces whether requested by
a negative value of INITAL or by spaces or tab
characters in the original text are to be inserted
into the processed text.
IBEGIN = the subscript of the IBUFFR array location at which
is to be found the first character of the text to
be processed. Following the conversion of tab
characters in the text to the proper number of
spaces, the processed text is placed back into the
IBUFFR array starting at subscript IBEGIN.
IFINAL = the subscript of the IBUFFR array location at which
is to be found the final character of the text to
be processed.
MAXBFR = the subscript of the highest location in the IBUFFR
array at which a character of the processed text
can be placed. MAXBFR must equal or be greater
than IFINAL. Any characters which require shifting
to a higher subscript than that indicated by MAXBFR
are instead discarded.
IBUFFR = the array used for input of the text containing tab
characters which are to be expanded to the proper
number of spaces, and used for output of the text
after this expansion of tab characters to spaces
has been performed. The characters in the IBUFFR
array must have been read by a multiple of an A1
format.
MAXPRT = returned containing the subscript of the highest
location in the processed text into which DAFILL
has placed a printing (nonspace) character. If no
printing characters are placed into the processed
text, then MAXPRT will be returned containing the
value IBEGIN-1.
MAXUSD = returned containing the subscript of the highest
location in the IBUFFR array containing a character
of the text after processing. If INTRVL is less
than or equal to zero, and the original text
contained only spaces and/or tab characters, then
MAXUSD is returned containing the value IBEGIN-1.
If INTRVL is greater than zero, then MAXUSD will be
returned containing a value in the range IFINAL
through MAXBFR.
FASP, FORTRAN Alphameric Subroutine Package Page 58
DAFILL, Routine to Expand Tab Characters to Spaces
An Example of the Use of DAFILL
-- ------- -- --- --- -- ------
The sample program listed below and on the following page
demonstrates the manner in which DAFILL is used. The
program asks the user for the values of the arguments INITAL
and INTRVL, then prompts the user with an asterisk before
reading a line of text including tabs to be expanded to the
corresponding number of spaces. Since the asterisk is typed
in the first column, the first tab stop interval would be 1
less than the subsequent intervals. On the PDP-10,
appropriate values would be INITAL=7 and INTRVL=+/-8. The
use of other values for these arguments would give results
different than that shown directly on the terminal in
response to the user's typing.
The program calls DACOPY once for each character typed by
the user to mark with digits in a second buffer the
locations of the spaces resulting from the expansion of tab
characters. If merely copying with tab expansion was
desired, a single call to DACOPY similar to the following
single call to DAFILL would be sufficient to expand the tabs
while copying the text typed by the user.
A sample dialog between the program and the user is
presented following the listing of the program.
DATA ITTY,JTTY/5,5/
DATA IBEGIN,IFINAL,MAXBFR/6,55,72/
DATA IBLANK,ITAB/1H ,1H /
DIMENSION IBUFFR(72),JBUFFR(72),IDIGIT(10)
DATA IDIGIT/1H1,1H2,1H3,1H4,1H5,1H6,1H7,1H8,1H9,1H0/
C
C ASK USER FOR THE INITIAL AND SUBEQUENT TAB STOPS
1 WRITE(JTTY,2)
2 FORMAT(1X/15H INITAL,INTRVL=,$)
READ(ITTY,3)INITAL,INTRVL
3 FORMAT(2I)
C
C ASK FOR THE TEXT IN WHICH TABS ARE TO BE EXPANDED
WRITE(JTTY,4)
4 FORMAT(2H *,$)
READ(ITTY,5)(IBUFFR(I),I=IBEGIN,IFINAL)
5 FORMAT(60A1)
C
C USE DACOPY TO REPORT LOCATIONS OF TABS
JUSED=IBEGIN-1
JNITAL=INITAL
KOUNT=0
MOST=JUSED
JNTRVL=INTRVL
DO 8 JBEGIN=IBEGIN,IFINAL
FASP, FORTRAN Alphameric Subroutine Package Page 59
DAFILL, Routine to Expand Tab Characters to Spaces
KUSED=JUSED
CALL DACOPY(JNITAL,JNTRVL,IBUFFR,JBEGIN,JBEGIN,
1MAXBFR,JUSED,JBUFFR,NXTINI,NXTBGN,MAXPRT)
IF(JUSED.EQ.KUSED)GO TO 8
IF(JNTRVL.LT.0)JNTRVL=-JNTRVL
IF(IBUFFR(JBEGIN).EQ.IBLANK)GO TO 8
IF(IBUFFR(JBEGIN).NE.ITAB)GO TO 7
KOUNT=KOUNT+1
IF(KOUNT.GT.10)KOUNT=1
6 KUSED=KUSED+1
JBUFFR(KUSED)=IDIGIT(KOUNT)
IF(KUSED.LT.JUSED)GO TO 6
GO TO 8
7 MOST=MAXPRT
8 JNITAL=NXTINI
IF(MOST.GE.IBEGIN)WRITE(JTTY,11)
1(JBUFFR(I),I=IBEGIN,MOST)
C
C USE DAFILL TO EXPAND SAME TEXT
CALL DAFILL(INITAL,INTRVL,IBEGIN,IFINAL,MAXBFR,
1IBUFFR,MAXPRT,MAXUSD)
IF(MAXUSD.LT.IBEGIN)GO TO 1
DO 9 I=IBEGIN,MAXUSD
9 IF(IBUFFR(I).EQ.ITAB)WRITE(JTTY,10)
10 FORMAT(6H ERROR)
IF(MAXPRT.LT.IBEGIN)GO TO 1
WRITE(JTTY,11)(IBUFFR(I),I=IBEGIN,MAXPRT)
11 FORMAT(2H ,100A1)
C
C REPORT COLUMN NUMBERS
LINE=1
IRIGHT=MAXPRT-IBEGIN+2
12 CALL DABELT(1,1,0,LINE,2,
1IRIGHT,0,MAXBFR,IBUFFR,MAXLIN,MAXPRT,MAXUSD)
WRITE(JTTY,11)(IBUFFR(I),I=1,MAXPRT)
LINE=LINE+1
IF(LINE.LE.MAXLIN)GO TO 12
GO TO 1
END
Typical Dialog Between DAFILL Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
INITAL,INTRVL=7 8
* A B C D
1111111A 222222 B 333344444444 C 55 66666 77777 D
A B C D
23456789012345678901234567890123456789012345678901234567890
111111111122222222223333333333444444444455555555556
FASP, FORTRAN Alphameric Subroutine Package Page 60
DAFILL, Routine to Expand Tab Characters to Spaces
INITAL,INTRVL=7 8
* A B C D
1111111A2222222 3333333B4444444 555555C6666666 77777D
A B C D
23456789012345678901234567890123456789012345678901234567
111111111122222222223333333333444444444455555555
INITAL,INTRVL=7 8
*ABCDEF GHIJK LMNO PQR ST U
ABCDEF1GHIJK222LMNO3333PQR44444ST555555U
ABCDEF GHIJK LMNO PQR ST U
2345678901234567890123456789012345678901
11111111112222222222333333333344
INITAL,INTRVL=7 8
*A BC DEF GHIJ KLMNO PQRSTU
A111111BC222222DEF33333GHIJ4444KLMNO555PQRSTU
A BC DEF GHIJ KLMNO PQRSTU
234567890123456789012345678901234567890123456
1111111111222222222233333333334444444
INITAL,INTRVL=7 8
* ABC DEF GHI
1111122222222 ABC 344444444 DEF 566666666 GHI
ABC DEF GHI
2345678901234567890123456789012345678901234567890123
11111111112222222222333333333344444444445555
INITAL,INTRVL=7 -8
* ABC DEF GHI
ABC 122222222 DEF 344444444 GHI
ABC DEF GHI
23456789012345678901234567890123456
111111111122222222223333333
INITAL,INTRVL=-7 -8
* ABC DEF GHI
ABC 122222222 DEF 344444444 GHI
ABC DEF GHI
23456789012345678901234567890123456
111111111122222222223333333
FASP, FORTRAN Alphameric Subroutine Package Page 61
DAFLAG, Routine to Locate Components of File Specification
DDDDD AAA FFFFFFFF LL AAA GGGGG
DD DD AAAA FF LL AAAA GG
DD DD AA AA FF LL AA AA GG
DD DD AA AA FFFFF LL AA AA GG GGGG
DD DD AAAAAAA FF LL AAAAAAA GG GG
DD DD AA AA FF LL AA AA GG GG
DDDDD AA AA FF LLLLLLLL AA AA GGGGG
DAFLAG, Routine to Locate Components of File Specification
------ ------- -- ------ ---------- -- ---- -------------
DAFLAG locates the components of a file specification
contained in a buffer read by the calling program with a
multiple of an A1 format. Although the leading and
separating characters are those expected in file
specifications for Digital Equipment Corporation (DEC)
computers, no evaluation of the components is performed, so
this routine could be used to find the parts of
specifications having the same leading and separating
characters but which are meant for other purposes.
The basic form of a file specification is
DEVICE:NAME.EXT[NUMBER,NUMBER]/SWITCH:'TEXT'/SWITCH:NUMBER
or
[NUMBER,NUMBER]DEVICE:NAME.EXT/SWITCH:'TEXT'/SWITCH:NUMBER
in which NUMBER merely implies a location at which a number
might be present, but which could instead contain any
sequence of printing characters other than the separation
and delimiter characters. Only one device field, one name
field and one bracketed field can appear together in a
single file specification. The device field must appear
before the name field, but the bracketed field can be
before, between, or after the device and name fields. The
calling program is also informed if an at sign appears
anywhere within the file specification other than within a
switch field. Switch fields consisting of initial slashes
followed by words connected by colons are reported
separately from the rest of the file specification and
separately from each other by separate calls to this
routine. Complete file specifications can be separated by
commas, semicolons or equal signs. Comments which are
otherwise ignored can appear to the right of exclamation
points and ampersands.
Spaces and/or tab characters can appear to either side of
the items enclosed within the brackets, and can appear after
the leading slashes of switch fields and both before and
after the connecting colons in switch fields. The calling
program can specify whether other spaces are to terminate
FASP, FORTRAN Alphameric Subroutine Package Page 62
DAFLAG, Routine to Locate Components of File Specification
the file specification, or are to terminate a particular
field within the file specification or are to merely
terminate a particular component within the field. In a
bracketed field, commas are not necessary between items
which are separated by spaces and/or tab characters
providing the items being separated are present. For
example, the file specification DAFLAG[6001,56] could also
be written as DAFLAG[6001 56] or as DAFLAG[6001 , 56] but
the comma in DAFLAG[,56] or DAFLAG[ , 56] is necessary since
this comma indicates that the first number within the
bracketed field is missing.
The components of each field within the file specification
are identified to the calling program by length and by
starting location within the buffer. This information is
returned within the 2 arrays KNTLTR and INILTR respectively.
The descriptions of the components of each field are grouped
together in the arrays returned to the calling program.
Since each field can consist of any number of components
connected by the appropriate character (colon, comma or
period), and since for some applications a terminal
separator character has special meaning, an extra zero is
returned in the KNTLTR array if a connecting character
terminates the field but is not necessary. Since there is
never any question about whether an item in the device field
was followed by the connecting character, the description of
the device field does not include an extra zero. Whether an
extra zero appears within the description of the bracketed
field is independent of whether the right bracket is
present.
Text strings delimited by apostrophes are recognized only
within the switch fields. If a text string is found which
has not been preceded by a slash then this text string is
treated as though it were the only component of a switch
field which cannot be extended by a subsequent colon, so
that anything appearing to the right of the text string
would have to be evaluated by the next call to this routine.
The location of a text string as returned to the calling
program is that of the initial apostrophe, and the length of
the text string includes the initial, but not the final,
apostrophe. If the final apostrophe is missing, then the
text string is assumed to extend through the rightmost
printing character in the buffer. Within a text string, two
immediately adjacent apostrophes indicate a single
apostrophe which is to be included within the string. If
two immediately adjacent apostrophes are encountered within
the text string, then the remaining portion of the string is
moved 1 character to the left so that the returned contents
of the buffer and the returned length in the KNTLTR array do
not include the extra apostrophe.
FASP, FORTRAN Alphameric Subroutine Package Page 63
DAFLAG, Routine to Locate Components of File Specification
For example, if the contents of the buffer are
DSK:DAFLAG.F4[6001,56,FASP]/LINE:60:/TITLE:'(JAN 76)'
then the information returned by the first call to this
routine would be
field KNTLTR contents INILTR contents
device 3 1
name 6 5
2 12
bracketed 4 15
2 20
4 23
and the information returned by the second call to this
routine would be
switch 4 29
2 34
0 undefined
and the information returned by the third call to this
routine would be
switch 5 38
9 44
Some DECsystem-10 programs interpret an asterisk which is
attached directly to the right end of one of the components
of a file specification as indicating a wild-card match of
any and of all characters at or to the right of that
position within the component. A leftmost asterisk would
then be matched by any character sequence. Since the
asterisk logically terminates the specification of a
particular component, any asterisk which is followed
immediately by a printing character other than a punctuation
mark is treated by the DAFLAG routine as though separated
from this following character by a period if within a name
field, by a comma if within a bracketed field, or by a colon
if within a switch field. Therefore, the text strings
FASP*.F4[*,56] and FASP*F4[*56] would both be treated the
same by the DAFLAG routine. These text strings would
specify all files in which the first name begins with the
letters FASP, which have the second name F4 and which belong
to any user identified by the second number 56. This
interpretation of the asterisk is of course dependent upon
the existence of the proper logic within the program which
calls DAFLAG and is in no way supported by the DAFLAG
routine itself.
FASP, FORTRAN Alphameric Subroutine Package Page 64
DAFLAG, Routine to Locate Components of File Specification
The DAFLAG Argument List
--- ------ -------- ----
The argument list of routine DAFLAG is
SUBROUTINE DAFLAG(KONECT,LOWSTR,MAXSTR,MAXBFR,IBUFFR,
1 LOWBFR,MANY ,KIND ,INILTR,KNTLTR,MAXDSK,MAXNAM,
2 MAXNUM,MAXFLG,KONTNT,MINPRT,MAXPRT)
with the associated DIMENSION statement
DIMENSION KNTLTR(MAXSTR),INILTR(MAXSTR),IBUFFR(MAXBFR)
The following arguments are used only for input, and are
returned unchanged.
KONECT = -1, spaces and tab characters mark the end of the
current component of the current field, but do not
indicate either the end of the field or the end of
the file specification. Spaces and tab characters
can appear before and after any field, before and
after colons in the device field, between the
components of the name field, between the
components of the bracketed field, and following
the initial slash and before and after the
separating colons in the switch field. Periods are
not necessary between components of the name field
which are separated by spaces and/or tab
characters. Commas are not necessary between
components of the bracketed field which are
separated by spaces and/or tab characters. The
file specification will extend through the end of
the buffer, or up to the following ampersand,
equals sign, semicolon or exclamation point, or up
to the following comma which is not in a bracketed
field. All of the special characters mentioned
here including the space and tab characters are of
course treated no differently than any other
characters when these appear within a text string
delimited by apostrophes. If KONECT=-1, then the
text strings
DSK : DAFLAG F4 [ 6007 56 ] / HEADER : 0 / DISPOSE
and
DSK:DAFLAG.F4[6007,56]/HEADER:0/DISPOSE
are equivalent.
= 0, spaces and tab characters which are not within
either bracketed fields or switch fields mark the
end of the file specification. Spaces and tab
characters can appear between the components of the
bracketed field, and following the initial slash
and before and after the separating colons in the
switch field. Commas are not necessary between
components of the bracketed field which are
separated by spaces and/or tab characters. The
FASP, FORTRAN Alphameric Subroutine Package Page 65
DAFLAG, Routine to Locate Components of File Specification
file specification will extend through the end of
the buffer, or up to the following ampersand,
equals sign, semicolon or exclamation point, or up
to the following comma which is not in a bracketed
field, or up the the next space or tab character
which is within neither a bracketed field nor
within a switch field. All of the special
characters mentioned here including the space and
tab characters are of course treated no differently
than any other characters when these appear within
a text string delimited by apostrophes. If
KONECT=0, then the text strings
DSK:DAFLAG.F4[6007 56]/ HEADER : 0 / DISPOSE
and
DSK:DAFLAG.F4[6007,56]/HEADER:0/DISPOSE
are equivalent.
= 1, spaces and tab characters which are not within
either bracketed fields or switch fields mark the
end of the device or name field, but do not
indicate the end of the file specification. Spaces
and tab characters can appear before and after any
field, between the components of the bracketed
field, and following the initial slash and before
and after the separating colons in the switch
field. Commas are not necessary between components
of the bracketed field which are separated by
spaces and/or tab characters. The file
specification will extend through the end of the
buffer, or up to the following ampersand, equals
sign, semicolon or exclamation point, or up to the
following comma which is not in a bracketed field.
All of the special characters mentioned here
including the space and tab characters are of
course treated no differently than any other
characters when these appear within a text string
delimited by apostrophes. If KONECT=1, then the
text strings
DSK: DAFLAG.F4 [6007 56] / HEADER : 0 / DISPOSE
and
DSK:DAFLAG.F4[6007,56]/HEADER:0/DISPOSE
are equivalent.
LOWSTR = subscript of the first location within the INILTR
and KNTLTR arrays which can be used to hold a
description of the components of the file
specification.
MAXSTR = subscript of the final location within the INILTR
and KNTLTR arrays which can be used to hold a
description of the components of the file
specification.
MAXBFR = subscript of the final (rightmost) location within
the IBUFFR array which contains a character which
FASP, FORTRAN Alphameric Subroutine Package Page 66
DAFLAG, Routine to Locate Components of File Specification
can be part of the file specification.
The following arguments are used for both input to, and
output from this routine.
IBUFFR = array containing in locations IBUFFR(LOWBFR)
through IBUFFR(MAXBFR) the characters read by the
calling program with a multiple of an A1 format and
which can form the file specification. The
contents of the IBUFFR array are returned
unchanged, with the exception that the portion of a
text string to the right of adjacent apostrophes in
a text string in a switch field is moved 1
character to the left.
LOWBFR = subscript of the first (leftmost) location within
the IBUFFR array which contains a character which
can be part of the file specification. LOWBFR is
returned pointing to the first character which
should be evaluated by the subsequent call to this
routine, or else is returned pointing beyond the
end of the buffer if the buffer is empty or if the
buffer contains merely a comment indicated by a
leading exclamation point or by a leading
ampersand.
MANY = should be input containing zero each time this
routine is called to begin processing of a new
logical section of text, as for example when
beginning processing of a line of text not tied to
the previous line by an ampersand at the end of the
previous line, or when processing the text to the
right of a semicolon or to the right of an equals
sign. The initial zeroing of this argument must be
done by the calling program, but thereafter the
value returned by the previous call to this routine
can usually be used.
= returned containing the value which MANY should
have when this routine or any other in the FASP
package having MANY as an argument is next called.
The returned value of MANY should not be changed by
the calling program unless the interpretation of
the contents of the buffer is being abandoned
prematurely, in which case MANY should be reset to
have a zero value.
= -1, returned if a missing item is to be indicated
if the next routine encounters a leading comma or
finds that the buffer contains nothing other than a
possible comment indicated by a leading exclamation
point. MANY is returned containing -1 if a comma
precedes either an ampersand or a switch field.
= 0, returned if a missing item is to be indicated if
the next routine encounters a leading comma, but a
missing item is not indicated if the buffer
FASP, FORTRAN Alphameric Subroutine Package Page 67
DAFLAG, Routine to Locate Components of File Specification
contains nothing other than a possible comment
indicated by a leading exclamation point. MANY is
returned containing zero if the buffer is found to
be empty, or if the first printing character at or
to right of IBUFFR(LOWBFR) is found to be an
exclamation point, a semicolon or an equals sign.
These are all conditions under which the next call
to this routine would evaluate the start of a new
group of file specifications. MANY is returned
unchanged if a switch field is found at the start
of the contents of the buffer.
= 1, returned if a missing item is not to be
indicated if the next routine encounters a leading
comma or finds that the buffer contains nothing
other than a possible comment indicated by a
leading exclamation point. MANY is returned
containing one if a file specification consisting
of more than just a switch field is found, or if a
missing item is being indicated.
The following arguments are used only for output. Their
input values are ignored.
KIND = returned describing the type of item encountered.
= 1, nothing, except possibly a comment indicated by
a leading exclamation point, was found at or to the
right of IBUFFR(LOWBFR). LOWBFR is returned
pointing beyond the end of the buffer.
= 2, the first printing character at or to the right
of IBUFFR(LOWBFR) is a semicolon. LOWBFR is
returned pointing to the character to the right of
the semicolon. It is suggested that the calling
program treat this as an indication by the user
that the preceding command has been completed and
that a subsequent command will follow on the same
line.
= 3, the first printing character at or to the right
of IBUFFR(LOWBFR) is an equals sign. LOWBFR is
returned pointing to the character to the right of
the equals sign. On the PDP-10 computer, an equals
sign is used in file transfer commands to separate
destination and source file specifications.
= 4, the first printing character at or to the right
of IBUFFR(LOWBFR) is an ampersand. The characters
to the right of the ampersand are taken to be a
comment. LOWBFR is returned pointing beyond the
end of the buffer. It is suggested that the
calling program treat this as a request by the user
that the command be continued on the following
line. The effect is not quite the same as if the
user had typed all of the file specifications on a
single line since a file specification cannot be
split across a line boundary.
= 5, a missing file specification was indicated by an
FASP, FORTRAN Alphameric Subroutine Package Page 68
DAFLAG, Routine to Locate Components of File Specification
extra comma.
= 6, a partial file specification was found which
will be continued by the subsequent call to this
routine. A single call to this routine can only
return the description of a single switch field, or
of the portion of a file specification exclusive of
a trailing switch field if any.
For example, if the text buffer contains:
NAME.EXT/FIRST:SECOND/THIRD:FOURTH
then the first call to this routine would return
the description of the name field NAME.EXT and
would return KIND=6, the second call would return
the description of the switch field /FIRST:SECOND
and would again return KIND=6, and the third call
would return the description of the switch field
/THIRD:FOURTH and would finally return KIND=7.
= 7, a file specification (possibly consisting merely
of a switch field) was found. If the previous call
to this routine returned KIND=6 indicating a
partial specification, then the information
returned by the current call to this routine
completes the file specification. If the entire
file specification consisted only of switch fields,
and if these switch fields were preceded by a comma
and/or are followed by a comma, then the next call
to this routine will return KIND=5 indicating a
missing item unless the calling program first sets
MANY to have the value 1.
= 8, a file specification was found, but this was
followed by an unexpected character pointed to by
the returned value of LOWBFR. For example, the
text DEVICE:NAME.EXTENSION: would be evaluated as
though the buffer terminated prior to the second
colon, and LOWBFR would be returned pointing to
this second colon. The text
DEVICE:NAME.EXTENSION[6001,56][22,56] would be
evaluated as though the buffer terminated prior to
the second bracketed field, and LOWBFR would be
returned pointing to the second left bracket.
INILTR = array returned containing the subscripts within the
the IBUFFR array of the initial characters of the
words forming the file specification.
Locations INILTR(LOWSTR) through INILTR(MAXDSK) are
returned containing the locations in the buffer of
the initial characters of the device names.
Locations INILTR(MAXDSK+1) through INILTR(MAXNAM)
are returned containing the locations in the buffer
of the initial characters of the words in the name
field.
FASP, FORTRAN Alphameric Subroutine Package Page 69
DAFLAG, Routine to Locate Components of File Specification
Locations INILTR(MAXNAM+1) through INILTR(MAXNUM)
are returned containing the locations in the buffer
of the initial characters of the words in the
bracketed field.
If MAXFLG is returned greater than or equal to
LOWSTR, then MAXDSK and MAXNAM and MAXNUM are all
returned set to LOWSTR-1, KONTNT is returned set to
zero, and locations INILTR(LOWSTR) through
INILTR(MAXFLG) are returned containing the
locations in the buffer of the initial characters
of the words and text strings appearing in the
switch field. If a text string appears in a switch
field, the location in the IBUFFR array indicated
by the INILTR array will contain an apostrophe.
If a switch is found which can take a file
specification as its argument, and if an additional
switch field component was found to the right of
this switch name, then LOWBFR should be reset to
the value returned for this next component in the
INILTR array and this routine should then be called
again to evaluate the file specification which
appears as an argument.
KNTLTR = array returned containing the numbers of characters
in each of the words for which the first characters
are in the buffer locations indicated by the values
in the INILTR array. The subscripts of the INILTR
array and KNTLTR array locations describing a
particular word are identical. If a location
within the KNTLTR array is returned set to zero,
then the corresponding location within the INILTR
array is returned undefined.
MAXDSK = returned containing the subscript of the INILTR and
KNTLTR array locations describing the rightmost
component of the device field of the file
specification. If a device field is not found,
then MAXDSK will be returned containing LOWSTR-1.
MAXNAM = returned containing the subscript of the INILTR and
KNTLTR array locations describing the rightmost
component of the name field of the file
specification. If a name field is not found, then
MAXNAM will be returned equal to MAXDSK.
MAXNUM = returned containing the subscript of the INILTR and
KNTLTR array locations describing the rightmost
component of the bracketed field of the file
specification. If a bracketed field is not found,
then MAXNUM will be returned equal to MAXNAM.
MAXFLG = returned containing the subscript of the INILTR and
FASP, FORTRAN Alphameric Subroutine Package Page 70
DAFLAG, Routine to Locate Components of File Specification
KNTLTR array locations describing the rightmost
component of the switch field. If a switch field
is not found, then MAXFLG is returned equal to
LOWSTR-1.
KONTNT = bit coded number returned describing the location
of the bracketed field relative to the device and
name fields. The right bit is one if and only if a
name field is found. The second bit from the right
is one if and only if a device field is found. The
fourth and third bits from the right are 00 if no
bracketed field is found, 01 if a bracketed field
appears first, 10 if a bracketed field follows a
device field, and 11 if a bracketed field follows a
name field.
The following table presents the values of KONTNT
returned for all possible combinations of device,
name and bracketed fields. The minus signs
represent values of KONTNT which cannot be
returned. The value zero indicates that neither
device, nor name nor bracketed fields were found,
but does not indicate if a switch field was found.
decimal binary decimal binary
0 0 nothing 8 1000 ------
1 1 NAME 9 1001 ------
2 10 DEVICE: 10 1010 DEVICE:[]
3 11 DEVICE:NAME 11 1011 DEVICE:[]NAME
4 100 [] 12 1100 ------
5 101 []NAME 13 1101 NAME[]
6 110 []DEVICE: 14 1110 ------
7 111 []DEVICE:NAME 15 1111 DEVICE:NAME[]
16 is added to KONTNT if an at sign (@) is found
before or after device, name or bracketed field.
The at sign, like the brackets and period, are
treated no differently than any alphabetic
character within a switch field.
MINPRT = the subscript of the IBUFFR array location which
contains the leftmost printing character in the
portion of the file specification evaluated by the
current call to this routine if KIND is returned
containing the value 6 or greater. MINPRT and
MAXPRT can be used as the limits of the IBUFFR
array subscripts if the text evaluated by the
current call to this routine must be displayed to
the user. MINPRT and MAXPRT are returned undefined
if KIND is returned set to a value less than 6.
MAXPRT = the subscript of the IBUFFR array location which
contains the rightmost printing character of the
portion of the file specification evaluated by the
FASP, FORTRAN Alphameric Subroutine Package Page 71
DAFLAG, Routine to Locate Components of File Specification
current call to this routine if KIND is returned
containing the value 6 or greater. MAXPRT can be
returned set to less than the returned value of
LOWBFR if the text to the right of the file
specification had to be searched for a continuation
of the file specification.
Demonstration Program to Interactively Test DAFLAG
------------- ------- -- ------------- ---- ------
The program listed on the following pages accepts a line of
text from the user, then summarizes the components of the
file specifications located within this text by DAFLAG. The
argument named KONECT initially has the value zero, but this
value can be changed to -1, back to 0 or to 1 if a minus
sign, a zero or a plus sign respectively is is found as the
first character of a new line typed by the user. After each
call to DAFLAG, the program reports the returned value of
the argument named KONTNT, together with a description of
the type of item found as indicated by the returned value of
the argument named KIND. This is followed by a listing of
the length of and the characters forming each of the
components of each of the fields of the file specification.
A sample dialog between the program and the user is
presented following the listing of the program.
C PROGRAM TO DEMONSTRATE DAFLAG ROUTINE
C
DIMENSION KNTLTR(25),INILTR(25),IBUFFR(60)
DATA ITTY,JTTY/5,5/
DATA MINUS,IZERO,IPLUS,IONE,IGREAT/
11H-,1H0,1H+,1H1,1H>/
DATA LOWSTR,MAXSTR,MAXBFR/4,20,60/
KONECT=0
MANY=0
WRITE(JTTY,1)
1 FORMAT(1X,37HPROGRAM TO DEMONSTRATE DAFLAG ROUTINE/
146H INDICATE KONECT BY -, 0 OR + AS 1ST CHARACTER/
244H KONECT IS UNCHANGED IF -, 0 OR + IS NOT 1ST)
2 WRITE(JTTY,3)
3 FORMAT(1X,1H*,$)
READ(ITTY,4)IBUFFR
4 FORMAT(60A1)
LOWBFR=1
C
C OBTAIN VALUE OF KONECT FROM START OF TEXT
I=-2
IF(IBUFFR(1).EQ.MINUS)I=-1
IF(IBUFFR(1).EQ.IZERO)I=0
IF(IBUFFR(1).EQ.IPLUS)I=1
IF(IBUFFR(1).EQ.IONE)I=1
FASP, FORTRAN Alphameric Subroutine Package Page 72
DAFLAG, Routine to Locate Components of File Specification
IF(I.EQ.-2)GO TO 5
KONECT=I
LOWBFR=2
C
C LOCATE COMPONENTS OF FILE DESCRIPTION
5 CALL DAFLAG(KONECT,LOWSTR,MAXSTR,MAXBFR,IBUFFR,
1LOWBFR,MANY,KIND,INILTR,KNTLTR,MAXDSK,MAXNAM,
2MAXNUM,MAXFLG,KONTNT,MINPRT,MAXPRT)
C
C REPORT TYPE OF ITEM LOCATED
IF(KIND.GE.6)WRITE(JTTY,6)
1(IBUFFR(I),I=MINPRT,MAXPRT),IGREAT
6 FORMAT(1X,1H<,100A1)
IF(KIND.EQ.1)WRITE(JTTY,7)
7 FORMAT(6H EMPTY)
IF(KIND.EQ.2)WRITE(JTTY,8)
8 FORMAT(10H SEMICOLON)
IF(KIND.EQ.3)WRITE(JTTY,9)
9 FORMAT(7H EQUALS)
IF(KIND.EQ.4)WRITE(JTTY,10)
10 FORMAT(10H AMPERSAND)
IF(KIND.EQ.5)WRITE(JTTY,11)
11 FORMAT(8H MISSING)
IF(KIND.EQ.6)WRITE(JTTY,12)KONTNT
12 FORMAT(8H PARTIAL,I4)
IF(KIND.EQ.7)WRITE(JTTY,13)KONTNT
13 FORMAT(9H TERMINAL,I3)
IF(KIND.EQ.8)WRITE(JTTY,14)KONTNT
14 FORMAT(8H ERROR ,I4)
C
C SUMMARIZE COMPONENTS OF FILE SPECIFICATION
LIMIT=LOWSTR-1
DO 20 IPART=1,4
ISTART=LIMIT+1
IF(IPART.EQ.1)LIMIT=MAXDSK
IF(IPART.EQ.2)LIMIT=MAXNAM
IF(IPART.EQ.3)LIMIT=MAXNUM
IF(IPART.EQ.4)LIMIT=MAXFLG
15 IF(ISTART.GT.LIMIT)GO TO 20
J=INILTR(ISTART)
K=KNTLTR(ISTART)
K=J+K-1
IF(K.LT.J)K=J
IF(IPART.EQ.1)WRITE(JTTY,16)KNTLTR(ISTART),
1(IBUFFR(I),I=J,K)
16 FORMAT(9H DEVICE,I3,1X,100A1)
IF(IPART.EQ.2)WRITE(JTTY,17)KNTLTR(ISTART),
1(IBUFFR(I),I=J,K)
17 FORMAT(9H NAME ,I3,1X,100A1)
IF(IPART.EQ.3)WRITE(JTTY,18)KNTLTR(ISTART),
1(IBUFFR(I),I=J,K)
18 FORMAT(9H BRCKTD,I3,1X,100A1)
IF(IPART.EQ.4)WRITE(JTTY,19)KNTLTR(ISTART),
1(IBUFFR(I),I=J,K)
FASP, FORTRAN Alphameric Subroutine Package Page 73
DAFLAG, Routine to Locate Components of File Specification
19 FORMAT(9H SWITCH,I3,1X,100A1)
ISTART=ISTART+1
GO TO 15
20 CONTINUE
IF(KIND.EQ.1)GO TO 2
IF(KIND.EQ.4)GO TO 2
GO TO 5
END
Typical Dialog Between DAFLAG Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
PROGRAM TO DEMONSTRATE DAFLAG ROUTINE
INDICATE KONECT BY -, 0 OR + AS 1ST CHARACTER
KONECT IS UNCHANGED IF -, 0 OR + IS NOT 1ST
*/ONE:TWO [THREE,FOUR,FIVE]SIX:SEVEN.EIGHT.NINE/'TEN':'ZERO'
</ONE:TWO>
PARTIAL 0
SWITCH 3 ONE
SWITCH 3 TWO
<[THREE,FOUR,FIVE]SIX:SEVEN.EIGHT.NINE>
PARTIAL 7
DEVICE 3 SIX
NAME 5 SEVEN
NAME 5 EIGHT
NAME 4 NINE
BRCKTD 5 THREE
BRCKTD 4 FOUR
BRCKTD 4 FIVE
</'TEN':'ZERO>
TERMINAL 0
SWITCH 4 'TEN
SWITCH 5 'ZERO
EMPTY
*-/ ONE : TWO [ THREE FOUR , FIVE ] SIX : SEVEN EIGHT . NINE
</ ONE : TWO>
PARTIAL 0
SWITCH 3 ONE
SWITCH 3 TWO
<[ THREE FOUR , FIVE ] SIX : SEVEN EIGHT . NINE>
TERMINAL 7
DEVICE 3 SIX
NAME 5 SEVEN
NAME 5 EIGHT
NAME 4 NINE
BRCKTD 5 THREE
BRCKTD 4 FOUR
BRCKTD 4 FIVE
EMPTY
FASP, FORTRAN Alphameric Subroutine Package Page 74
DAFLAG, Routine to Locate Components of File Specification
*0/ ONE : TWO [ THREE FOUR , FIVE ] SIX : SEVEN EIGHT . NINE
</ ONE : TWO>
PARTIAL 0
SWITCH 3 ONE
SWITCH 3 TWO
<[ THREE FOUR , FIVE ]>
TERMINAL 4
BRCKTD 5 THREE
BRCKTD 4 FOUR
BRCKTD 4 FIVE
<SIX>
TERMINAL 1
NAME 3 SIX
<:>
TERMINAL 2
DEVICE 0
<SEVEN>
TERMINAL 1
NAME 5 SEVEN
<EIGHT>
TERMINAL 1
NAME 5 EIGHT
<.>
TERMINAL 1
NAME 0
NAME 0
<NINE>
TERMINAL 1
NAME 4 NINE
EMPTY
*+/ ONE : TWO [ THREE FOUR , FIVE ] SIX : SEVEN EIGHT . NINE
</ ONE : TWO>
PARTIAL 0
SWITCH 3 ONE
SWITCH 3 TWO
<[ THREE FOUR , FIVE ] SIX>
ERROR 5
NAME 3 SIX
BRCKTD 5 THREE
BRCKTD 4 FOUR
BRCKTD 4 FIVE
<: SEVEN>
ERROR 3
DEVICE 0 :
NAME 5 SEVEN
<EIGHT>
ERROR 1
NAME 5 EIGHT
<.>
ERROR 1
NAME 0 .
NAME 0
FASP, FORTRAN Alphameric Subroutine Package Page 75
DAFLAG, Routine to Locate Components of File Specification
<NINE>
TERMINAL 1
NAME 4 NINE
EMPTY
*,'ONE'TWO'THREE',=,FOUR,;@FIVE,SIX@&SPECIAL CHARACTERS
MISSING
<'ONE>
PARTIAL 0
SWITCH 4 'ONE
<TWO>
PARTIAL 1
NAME 3 TWO
<'THREE>
TERMINAL 0
SWITCH 6 'THREE
MISSING
EQUALS
MISSING
<FOUR>
TERMINAL 1
NAME 4 FOUR
MISSING
SEMICOLON
<@FIVE>
TERMINAL 17
NAME 4 FIVE
<SIX@>
TERMINAL 17
NAME 3 SIX
AMPERSAND
*,SEVEN,&NO MISSING ITEM AT START OF THIS LINE
<SEVEN>
TERMINAL 1
NAME 5 SEVEN
AMPERSAND
*,EIGHT!COMMAS IN BOTH LINES INDICATE SINGLE MISSING ITEM
MISSING
<EIGHT>
TERMINAL 1
NAME 5 EIGHT
EMPTY
*!DEMONSTRATE POSSIBLE VALUES OF KONTNT IN RANGE 1 THRU 15
EMPTY
*A,B:,C:D,[E],[F]G,[H]I:,[J]K:L,M:[N],O:[P]Q,R[S],T:U[V]
<A>
TERMINAL 1
NAME 1 A
<B:>
TERMINAL 2
DEVICE 1 B
<C:D>
TERMINAL 3
DEVICE 1 C
NAME 1 D
FASP, FORTRAN Alphameric Subroutine Package Page 76
DAFLAG, Routine to Locate Components of File Specification
<[E]>
TERMINAL 4
BRCKTD 1 E
<[F]G>
TERMINAL 5
NAME 1 G
BRCKTD 1 F
<[H]I:>
TERMINAL 6
DEVICE 1 I
BRCKTD 1 H
<[J]K:L>
TERMINAL 7
DEVICE 1 K
NAME 1 L
BRCKTD 1 J
<M:[N]>
TERMINAL 10
DEVICE 1 M
BRCKTD 1 N
<O:[P]Q>
TERMINAL 11
DEVICE 1 O
NAME 1 Q
BRCKTD 1 P
<R[S]>
TERMINAL 13
NAME 1 R
BRCKTD 1 S
<T:U[V]>
TERMINAL 15
DEVICE 1 T
NAME 1 U
BRCKTD 1 V
EMPTY
*!DEMONSTRATE EMPTY STATEMENTS AND MISSING COMPONENTS
EMPTY
*;:ONE::TWO:.THREE..FOUR.[,FIVE,,SIX,]/:SEVEN::EIGHT:;
SEMICOLON
FASP, FORTRAN Alphameric Subroutine Package Page 77
DAFLAG, Routine to Locate Components of File Specification
<:ONE::TWO:.THREE..FOUR.[,FIVE,,SIX,]>
PARTIAL 15
DEVICE 0 :
DEVICE 3 ONE
DEVICE 0 :
DEVICE 3 TWO
NAME 0 .
NAME 5 THREE
NAME 0 .
NAME 4 FOUR
NAME 0 [
BRCKTD 0 ,
BRCKTD 4 FIVE
BRCKTD 0 ,
BRCKTD 3 SIX
BRCKTD 0 ]
</:SEVEN::EIGHT:>
TERMINAL 0
SWITCH 0 :
SWITCH 5 SEVEN
SWITCH 0 :
SWITCH 5 EIGHT
SWITCH 0 ;
SEMICOLON
EMPTY
*!DEMONSTRATE ASTERISKS IN ALL BUT DEVICE FIELD
EMPTY
*DSK:ONE*[THREE*]/FIVE*/'SEVEN'*
<DSK:ONE*[THREE*]>
PARTIAL 15
DEVICE 3 DSK
NAME 4 ONE*
BRCKTD 6 THREE*
</FIVE*>
PARTIAL 0
SWITCH 5 FIVE*
</'SEVEN>
PARTIAL 0
SWITCH 6 'SEVEN
<*>
TERMINAL 1
NAME 1 *
EMPTY
*DSK:*TWO[*FOUR]/*SIX/*'EIGHT'
<DSK:*TWO[*FOUR]>
PARTIAL 15
DEVICE 3 DSK
NAME 1 *
NAME 3 TWO
BRCKTD 1 *
BRCKTD 4 FOUR
FASP, FORTRAN Alphameric Subroutine Package Page 78
DAFLAG, Routine to Locate Components of File Specification
</*SIX>
PARTIAL 0
SWITCH 1 *
SWITCH 3 SIX
</*'EIGHT>
TERMINAL 0
SWITCH 1 *
SWITCH 6 'EIGHT
EMPTY
*DSK:**[**]/** !**:** WOULD BE TAKEN AS *.* FOLLOWED BY :*.*
<DSK:**[**]>
PARTIAL 15
DEVICE 3 DSK
NAME 1 *
NAME 1 *
BRCKTD 1 *
BRCKTD 1 *
</**>
TERMINAL 0
SWITCH 1 *
SWITCH 1 *
EMPTY
Typical but Machine Dependent Wrapper for the DAFLAG Routine
------- --- ------- --------- ------- --- --- ------ -------
DAFLAG locates the components of a single file
specification, but does not pack the characters which form
these components into the single or double precision
computer locations which can be used as arguments in the
system subroutine calls or FORTRAN OPEN statements which are
necessary to prepare for reading or writing of the desired
file. Listed on the following pages is a wrapper for the
DAFLAG routine which packs the components of a file
specification into the form required for DECsystem-10
FORTRAN OPEN statements. The wrapper routine, named GETFIL,
supports a multiple file specification of the form
LIST OF OUTPUT FILES=LIST OF INPUT FILES
or
LIST OF INPUT FILES
and so must scan the command from left to right until one
more than the number of files which can appear in the list
of output files has been found. If the user desires to
input several lines, all but the last line can be terminated
by an ampersand, or, if the command consists of a list of
file specifications, all but the last line can be terminated
by a rightmost comma. Although GETFIL initially interacts
with the user, the user can at any point specify that the
remainder of the command is to be read from a file by giving
its name along with an at (@) sign. If the command is being
read from a file, then the continuation indications are not
FASP, FORTRAN Alphameric Subroutine Package Page 79
DAFLAG, Routine to Locate Components of File Specification
necessary since the entire command file will be read until
the end of file is encountered.
Switches can appear either before or after the file
specifications with which they are associated. The
interpretation of switches is left to the calling program.
Switches are reported individually when they are found,
since many switches appearing on more than just a single
command line could precede the file specification to which
the switches apply. The file specification associated with
the switch is identified only by its serial position in the
arrays in which the file specification will be reported by a
subsequent call of this routine, both since the file
specification might not yet have been read, and because it
is not immediately known whether a leading file
specification appears to the left of an equal sign or if no
equal sign at all is present.
The argument list of routine GETFIL is
SUBROUTINE GETFIL(MAXFIL, ITTY, JTTY,KMDNUM,KMDDVC,
1 KMDNAM,KMDEXT,MAXSTR,MAXBFR, KIND,NEWNUL,NEWDSK,
2 NEWNAM,NEWPTH,LCNRIT,IBUFFR,MAXFLG,INILTR,KNTLTR,
3 LCNOWN)
with the associated DIMENSION and DOUBLE PRECISION
statements
DIMENSION NEWNUL(MAXFIL),NEWDSK(MAXFIL),
1IBUFFR(MAXBFR),INILTR(MAXSTR),KNTLTR(MAXSTR)
DOUBLE PRECISION KMDNAM,NEWNAM(MAXFIL),
1NEWPTH(3,MAXFIL)
The following arguments are used only for input and are
returned unchanged.
MAXFIL = 1 more than the number of files which can appear to
the left of the equal sign. The single precision
arrays NEWNUL and NEWDSK, and the double precision
array NEWNAM must be dimensioned to at least the
value of MAXFIL. The double precision and doubly
dimensioned array NEWPTH must have 3 as its first
dimension and MAXFIL as its second dimension.
Switches are always returned individually by
separate calls to this routine prior to the call or
calls to this routine which return a description of
the file specifications with which these switches
are associated. All switches left of the equal
sign and the switches associated with the first
file specification right of the equal sign will
have been returned to the calling program before
any file specifications are returned. The first
call to this routine which returns file
specifications can return up to MAXFIL file
FASP, FORTRAN Alphameric Subroutine Package Page 80
DAFLAG, Routine to Locate Components of File Specification
specifications, of which only the file
specification which is returned in the location in
each array having the value of LCNRIT as its
subscript is to the right of the equal sign. The
subsequent calls to this routine will return either
a single switch specification or a single file
specification. An equal sign found after an equal
sign has already been found or after more than
MAXFIL file specifications have been found is taken
to be equivalent to a comma.
ITTY = number of the unit from which commands are
initially to be read.
JTTY = number of the unit to which error messages and
prompts for more user input are to be written.
KMDNUM = number of the unit from which the command file
indicated by an at (@) sign appearing with a file
specification is to be read.
KMDDVC = name in 5H form of the device from which the
command file is to be read if the user fails to
supply a device name followed by a colon.
KMDNAM = name in 6H form which is to be used as the first
component of the name of the command file if none
is supplied by the user. On the PDP10, KMDNAM must
be a double precision variable if the name of this
variable appears in the argument list.
KMDEXT = name in 3H form which is to be used as the second
component (the file name extension) of the name of
the command file if none is supplied by the user.
If the command file name is not to have such a
second component, then the user must type a period
following the first component of the name.
MAXSTR = dimension of the INILTR and KNTLTR arrays in which
the descriptions of the components of the switches
are returned and which are used internally within
this routine for the storage of the descriptions of
the components of each file specification. MAXSTR
should have a value of at least 6 which would be
sufficient to allow either a 6 component switch or
else the combination of a 1 part device name, a 2
part file name, and a 3 part path (directory)
description.
MAXBFR = dimension of the IBUFFR array into which each line
of the commands typed by the user or read from the
command file are stored in a multiple of an A1
format. MAXBFR is the maximum number of characters
which can appear in a single command line. MAXBFR
FASP, FORTRAN Alphameric Subroutine Package Page 81
DAFLAG, Routine to Locate Components of File Specification
must not exceed 132.
The following argument must be zeroed by the calling program
before this routine is first called, but then the value
returned by this routine should be sent to the following
call of this routine unchanged.
KIND = should be set to zero before this routine is first
called, or whenever the interpretation of the
previous set of commands is to be abandoned. KIND
is returned describing the reason why control has
been transferred back to the calling program, and
should not be changed by the calling program if
this routine is to be called again to continue the
interpretation of the same sequence of commands.
= 1, returned if no more file specifications remain
to be evaluated. The description of the last of
the previous set of file specifications was
returned to the calling program by the previous
call to this routine and that file specification
was not followed by either an ampersand or a
semicolon.
= 2, returned if a semicolon was found. If this
routine is called again without KIND having first
been zeroed, then the evaluation of a new set of
file specifications will be begun in the text
appearing to the right of the semicolon. The
appearance of a semicolon when the file
specification is known by this routine to be
incomplete will not be reported to the calling
program since the text to the right of the
semicolon is treated as if it continued the file
specifications on a subsequent input line. For
example, the file specifications
A=B,;C
and
A=B,
C
are both equivalent to
A=B,C
= 3, returned if this routine is currently reporting
all of the file specifications appearing to the
left of the equal sign together with the first file
specification to the right of the equal sign, or if
this routine is currently reporting the first file
specification in a series of file specifications
which does not include an equal sign.
= 4, returned if this routine is currently reporting
the second or a subsequent file specification to
the right of the equal sign, or if this routine is
currently reporting the second or a subsequent file
specification in a series of file specifications
which does not include an equal sign.
= 5, returned if this routine is currently returning
FASP, FORTRAN Alphameric Subroutine Package Page 82
DAFLAG, Routine to Locate Components of File Specification
the description of a switch in the INILTR and
KNTLTR array locations having the subscripts 1
through the returned value of MAXFLG. LCNOWN is
returned containing the value of the subscript of
the locations in the NEWNUL, NEWDSK, NEWNAM and
NEWPTH arrays which will describe the file
specification when KIND is next returned set to
either 3 or 4.
The following arguments are used both for returning
information to the calling program and for transferring
information from the current call of this routine to the
subsequent call of this routine. The values of these
arguments returned by the current call of this routine
should be sent to the following call of this routine
unchanged. The original contents of these arguments are
ignored.
NEWNUL = if KIND is returned set to either 3 or 4, then the
locations in the NEWNUL array having the subscripts
1 through the returned value of LCNRIT are returned
describing whether a file specification was found,
and if so, whether the file name consisted of 1 or
2 components.
= 0, an extra comma indicated that no file was being
specified. The locations in the NEWDSK, NEWNAM and
NEWPTH arrays having the same subscripts as the
location in the NEWNUL array containing the zero
are returned undefined.
= 1, either a device name or a path (a directory) was
specified, but a file name was not specified. The
unspecified items are returned set either to zeroes
or to spaces as is appropriate.
= 2, a file name was specified, but this did not
include either a period or a second component. A
null second component consisting of 3 spaces is
being returned following a period in the NEWNAM
array. The calling program can supply a default
second component if such is appropriate.
= 3, a file name was specified which included a
period. A second component consisting of 3 spaces
is being returned if no second component was
specified.
= 4, a file name was specified which included a
period and a second component, but no first
component was included. A null first component
consisting of 6 spaces is being returned preceding
a period in the NEWNAM array. It is expected that
the calling program will supply a default first
component.
NEWDSK = if KIND is returned set to either 3 or 4, then the
locations in the NEWDSK array having the subscripts
1 through the returned value of LCNRIT are returned
FASP, FORTRAN Alphameric Subroutine Package Page 83
DAFLAG, Routine to Locate Components of File Specification
containing in 5H form the device names associated
with each of the file specifications. If no device
names are specified in the file specifications to
the left of the equal sign, then the locations
corresponding to these file specifications in the
NEWDSK array will contain spaces. If no device
names are specified in the file specifications to
the right of the equal sign, then these locations
either contain the previously specified device name
if any has been specified to the right of the equal
sign, or else contain spaces if no devices name has
yet been specified to the right of the equal sign.
NEWNAM = if KIND is returned set to either 3 or 4, then the
locations in the NEWNAM double precision array
having the subscripts 1 through the returned value
of LCNRIT are returned containing both components
of the file name (the 6 character file name and its
3 character extension) in an A10 form (format
1A6,1H.,1A3).
NEWPTH = if KIND is returned set to either 3 or 4, then the
locations in the NEWPTH double precision and doubly
dimensioned array having 1 through the returned
value of LCNRIT as their second subscripts are
returned containing the path (the directory) upon
which the file is located. If no paths are
specified in the file specifications to the left of
the equal sign, then the locations corresponding to
these file specifications in the NEWPTH array will
contain zeroes. If no paths are specified in the
file specifications to the right of the equal sign,
then these locations either contain the previously
specified path if any has been specified to the
right of the equal sign, or else contain zeroes if
no path has yet been specified to the right of the
equal sign. NEWPTH(1,...) contains in its left
half the octal project number and in its right half
the octal programmer number (the first two numbers
appearing within the square brackets), or contains
zero if no path has been specified. NEWPTH(2,...)
contains the sub file directory (SFD) name (the
third component within the square brackets) in 6H
form if any has been specified, or contains zero
otherwise. NEWPTH(3,...) always contains zero.
LCNRIT = if KIND is returned set to either 3 or 4, then
LCNRIT is returned containing the value of the
subscript of the locations in the NEWNUL, NEWDSK,
NEWNAM and NEWPTH arrays which describe the file
specification which appears to the right of the
equal sign. If KIND is returned set to 3 and if
LCNRIT is greater than one, then the lower
locations in these arrays describe the file
FASP, FORTRAN Alphameric Subroutine Package Page 84
DAFLAG, Routine to Locate Components of File Specification
specifications appearing to the left of the equal
sign. If KIND is returned set to 4, then the
values of the locations within these arrays having
subscripts less than the returned value of LCNRIT
should be ignored.
IBUFFR = array in which this routine can return characters
typed by the user or read from the command file.
These characters represent only the most recent
command line and have been read by a multiple of an
A1 format. It is the responsibility of the calling
program to evaluate the switch appearing within
this array if KIND is returned set to 5.
The following arguments are used only for returning
information to the calling program. Their input values are
ignored.
MAXFLG = if KIND is returned set to 5, then MAXFLG is
returned containing the subscript of the locations
in the INILTR and KNTLTR arrays which describe the
rightmost component of the switch.
INILTR = if KIND is returned set to 5, then the locations in
the INILTR array having the subscripts 1 through
the returned value of MAXFLG are returned
containing the values of the subscripts of the
locations within the IBUFFR array in which are to
be found the initial characters of each of the
components of the switch.
KNTLTR = if KIND is returned set to 5, then the locations in
the INILTR array having the subscripts 1 through
the returned value of MAXFLG are returned
containing the number of characters within each of
the components of the switch. A missing component
of the switch is indicated by a zero value in the
KNTLTR array.
LCNOWN = if KIND is returned set to 5, then LCNOWN is
returned containing the value of the subscript of
the locations in the NEWNUL, NEWDSK, NEWNAM and
NEWPTH arrays which will describe the file
specification when KIND is next returned set to
either 3 or 4.
The GETFIL routine is listed below. The first half of the
routine calls DAFLAG and then decides what is to be done
with the information returned by DAFLAG. The GETFIL routine
manipulates the value of the DAFLAG argument named MANY to
simulate a rightmost ampersand when a statement ends with a
rightmost equal sign or rightmost comma. The second half of
the routine uses DECsystem-10 FORTRAN ENCODE statements to
pack the characters read 1 to a computer storage location
FASP, FORTRAN Alphameric Subroutine Package Page 85
DAFLAG, Routine to Locate Components of File Specification
into a form which might have been read with an A5 or double
precision A10 format. The octal numbers are evaluated by
matching the individual digits against an array containing
1H0 through 1H7 rather than by use of ENCODE and DECODE
statements since these statements do not have any provision
for continuing after a non-numeric character is found in a
numeric field.
SUBROUTINE GETFIL(MAXFIL, ITTY, JTTY,KMDNUM,KMDDVC,
1 KMDNAM,KMDEXT,MAXSTR,MAXBFR, KIND,NEWNUL,NEWDSK,
2 NEWNAM,NEWPTH,LCNRIT,IBUFFR,MAXFLG,INILTR,KNTLTR,
3 LCNOWN)
C RENBR(/EVALUATE FORM FILE,FILE=FILE,FILE)
COMMON/FASPZ/KNTFIL,MANY,IEOF,IAFTER,LSTPTH,LSTDSK,
1LOWBFR
DIMENSION INILTR(MAXSTR),KNTLTR(MAXSTR),
1IBUFFR(MAXBFR),NEWNUL(MAXFIL),NEWDSK(MAXFIL),
2LETTER(8),KOLECT(10),NUMTWO(2)
DOUBLE PRECISION NEWNAM(MAXFIL),NEWPTH(3,MAXFIL),
1KOMAND,ONEPTH(3),LSTPTH(3),TWONUM,KMDNAM
EQUIVALENCE (TWONUM,NUMTWO),(NEWPRJ,NUMTWO(1)),
1(NEWUSR,NUMTWO(2))
DATA LETTER/1H0,1H1,1H2,1H3,1H4,1H5,1H6,1H7/
DATA IBLANK,JBLANK,IDOT/1H ,5H ,1H./
C
C DECIDE WHETHER ARE STARTING OR CONTINUING EVALUATION
IF(KIND.EQ.5)GO TO 9
IF(KIND.GE.3)GO TO 30
LCNRIT=0
KNTFIL=0
IAFTER=0
MANY=0
IF(KIND.EQ.2)GO TO 9
C
C READ CONTENTS OF NEXT LINE
WRITE(JTTY,1)
1 FORMAT(2H *,$)
GO TO 4
2 WRITE(JTTY,3)
3 FORMAT(2H &,$)
4 READ(ITTY,5,END=10)IBUFFR
5 FORMAT(132A1)
IEOF=0
GO TO 7
6 READ(KMDNUM,5,END=10)IBUFFR
IEOF=1
7 LOWBFR=1
GO TO 9
8 LCNRIT=1
KNTFIL=0
C
C LOCATE NEXT FILE SPECIFICATION
9 CALL DAFLAG(1,1,MAXSTR,MAXBFR,IBUFFR,
1LOWBFR,MANY,KIND,INILTR,KNTLTR,MAXDSK,MAXNAM,
FASP, FORTRAN Alphameric Subroutine Package Page 86
DAFLAG, Routine to Locate Components of File Specification
2MAXNUM,MAXFLG,KONTNT,MINPRT,MAXPRT)
GO TO(11,11,13,15,16,17,17,17),KIND
C
C END OF LINE, END OF FILE OR SEMICOLON FOUND
10 KIND=1
IEOF=0
11 IF(IAFTER.LT.0)GO TO 14
IF(IAFTER.EQ.1)GO TO 14
IF(KNTFIL.GT.0)GO TO 12
IF(LCNRIT.EQ.0)GO TO 15
IF(KIND.EQ.2)GO TO 37
IF(IEOF.EQ.0)GO TO 37
GO TO 6
12 IF(KIND.EQ.2)LOWBFR=LOWBFR-1
GO TO 29
C
C EQUAL SIGN FOUND
13 IF(IAFTER.GT.0)GO TO 27
MANY=-1
IAFTER=0
LCNRIT=-1
GO TO 9
C
C AMPERSAND FOUND OR MORE FILES NEEDED
14 KNTFIL=KNTFIL-1
MANY=-1
IAFTER=0
15 IF(KIND.EQ.2)GO TO 9
IF(IEOF.NE.0)GO TO 6
GO TO 2
C
C EXTRA COMMA FOUND
16 IF(IAFTER.GT.0)GO TO 27
KNTFIL=KNTFIL+1
NEWNUL(KNTFIL)=0
IAFTER=1
GO TO 19
C
C EVALUATE LOCATION AND REPACK DEVICE AND FILE NAME
17 IF(KONTNT.EQ.0)GO TO 20
IF(IAFTER.GT.0)GO TO 28
NXTFIL=KNTFIL+1
GO TO 38
18 IF(KONTNT.GE.16)GO TO 21
KNTFIL=NXTFIL
IAFTER=2
19 IF(LCNRIT.NE.0)GO TO 9
IF(KNTFIL.LT.MAXFIL)IAFTER=IAFTER-2
GO TO 9
C
C ALLOW CALLING PROGRAM TO EVALUATE SWITCH
20 LCNOWN=KNTFIL
KIND=5
IF(MANY.GT.0)GO TO 37
FASP, FORTRAN Alphameric Subroutine Package Page 87
DAFLAG, Routine to Locate Components of File Specification
IF(IAFTER.GT.0)GO TO 28
MANY=-1
LCNOWN=LCNOWN+1
GO TO 37
C
C OPEN COMMAND FILE SPECIFIED BY USER
21 KOMAND=NEWNAM(NXTFIL)
IF(NEWNUL(NXTFIL).LE.1)ENCODE(10,22,KOMAND)
1KMDNAM,KMDEXT
IF(NEWNUL(NXTFIL).EQ.2)ENCODE(10,22,KOMAND)
1NEWNAM(NXTFIL),KMDEXT
22 FORMAT(1A6,1H.,1A3)
IF(NEWNUL(NXTFIL).EQ.4)ENCODE(10,23,KOMAND)
1KMDNAM,(KOLECT(I),I=8,10)
23 FORMAT(1A6,1H.,3A1)
INDISK=NEWDSK(NXTFIL)
IF(INDISK.EQ.JBLANK)INDISK=KMDDVC
DO 24 I=1,3
24 ONEPTH(I)=NEWPTH(I,NXTFIL)
OPEN(UNIT=KMDNUM,DEVICE=INDISK,FILE=KOMAND,
1DIRECTORY=ONEPTH,ACCESS=5HSEQIN,ERR=25)
GO TO 6
25 WRITE(JTTY,26)(IBUFFR(I),I=MINPRT,MAXPRT)
26 FORMAT(26H CANNOT READ COMMAND FROM ,100A1)
GO TO 2
C
C PREPARE TO RETURN RESULTS TO CALLING PROGRAM
27 MANY=-1
GO TO 29
28 LOWBFR=MINPRT
29 IAFTER=0
IF(LCNRIT.GT.0)GO TO 31
KIND=3
IF(LCNRIT.LT.0)GO TO 33
LCNRIT=1
GO TO 34
30 LCNRIT=LCNRIT+1
31 IF(LCNRIT.GT.KNTFIL)GO TO 8
KIND=4
IF(NEWNUL(LCNRIT).EQ.0)GO TO 37
IF(NEWDSK(LCNRIT).EQ.JBLANK)NEWDSK(LCNRIT)=LSTDSK
IF(NEWPTH(1,LCNRIT).NE.0)GO TO 34
DO 32 I=1,3
32 NEWPTH(I,LCNRIT)=LSTPTH(I)
GO TO 36
33 LCNRIT=KNTFIL
34 DO 35 I=1,3
35 LSTPTH(I)=NEWPTH(I,LCNRIT)
36 LSTDSK=NEWDSK(LCNRIT)
C
C RETURN TO CALLING PROGRAM
37 RETURN
C
C *****************************************************
FASP, FORTRAN Alphameric Subroutine Package Page 88
DAFLAG, Routine to Locate Components of File Specification
C * *
C * A10 PACK NAME, A5 PACK DEVICE, EVALUATE NUMBERS *
C * *
C *****************************************************
C
C SET SWITCHES WHICH STATE IF ANYTHING WAS FOUND
38 NEWPRJ=0
NEWUSR=0
DO 39 I=1,3
39 NEWPTH(I,NXTFIL)=TWONUM
NEWNUL(NXTFIL)=0
NEWDSK(NXTFIL)=JBLANK
C
C PACK DEVICE NAME INTO A5 FORM
IF(MAXDSK.LT.1)GO TO 42
KOUNT=KNTLTR(1)
IF(KOUNT.LE.0)GO TO 42
IBGN=INILTR(1)
DO 40 I=1,5
KOLECT(I)=IBLANK
IF(KOUNT.GT.0)KOLECT(I)=IBUFFR(IBGN)
IBGN=IBGN+1
40 KOUNT=KOUNT-1
ENCODE(5,41,NEWDSK(NXTFIL))(KOLECT(I),I=1,5)
41 FORMAT(5A1)
NEWNUL(NXTFIL)=1
C
C EVALUATE OCTAL PROJECT, PROGRAMMER NUMBERS
42 INDEX=MAXNAM+1
IF(INDEX.GE.MAXNUM)GO TO 48
43 KOUNT=KNTLTR(INDEX)
IF(KOUNT.LE.0)GO TO 48
IBGN=INILTR(INDEX)
IEND=IBGN+KOUNT-1
NEWPRJ=NEWUSR
NEWUSR=0
DO 45 I=IBGN,IEND
LTRNOW=IBUFFR(I)
NEWUSR=8*NEWUSR
DO 44 J=1,8
IF(LETTER(J).NE.LTRNOW)GO TO 44
NEWUSR=NEWUSR+J-1
GO TO 45
44 CONTINUE
45 CONTINUE
IF(NEWUSR.LE.0)GO TO 48
INDEX=INDEX+1
IF(INDEX.LE.(MAXNAM+2))GO TO 43
NEWPTH(1,NXTFIL)=TWONUM
NEWNUL(NXTFIL)=1
C
C PACK SUB FILE DIRECTORY NAME
IF(MAXNUM.LE.(MAXNAM+2))GO TO 48
KOUNT=KNTLTR(MAXNAM+3)
FASP, FORTRAN Alphameric Subroutine Package Page 89
DAFLAG, Routine to Locate Components of File Specification
IF(KOUNT.LE.0)GO TO 48
IBGN=INILTR(MAXNAM+3)
DO 46 I=1,6
KOLECT(I)=IBLANK
IF(KOUNT.GT.0)KOLECT(I)=IBUFFR(IBGN)
IBGN=IBGN+1
46 KOUNT=KOUNT-1
ENCODE(10,47,NEWPTH(2,NXTFIL))(KOLECT(I),I=1,6)
47 FORMAT(6A1,4X)
C
C PACK FILE NAME AND ITS EXTENSION INTO A10 FORM
48 DO 49 I=1,10
49 KOLECT(I)=IBLANK
IF(MAXNAM.LE.MAXDSK)GO TO 55
KOUNT=KNTLTR(MAXDSK+1)
IF(KOUNT.LE.0)GO TO 51
NEWNUL(NXTFIL)=2
IBGN=INILTR(MAXDSK+1)
IF(KOUNT.GT.6)KOUNT=6
DO 50 I=1,KOUNT
KOLECT(I)=IBUFFR(IBGN)
50 IBGN=IBGN+1
IF(MAXNAM.LE.(MAXDSK+1))GO TO 54
NEWNUL(NXTFIL)=3
KOUNT=KNTLTR(MAXDSK+2)
GO TO 52
51 IF(MAXNAM.LE.(MAXDSK+1))GO TO 55
KOUNT=KNTLTR(MAXDSK+2)
IF(KOUNT.LE.0)GO TO 55
NEWNUL(NXTFIL)=4
52 IBGN=INILTR(MAXDSK+2)
IF(KOUNT.GT.3)KOUNT=3
DO 53 I=8,10
IF(KOUNT.GT.0)KOLECT(I)=IBUFFR(IBGN)
IBGN=IBGN+1
53 KOUNT=KOUNT-1
54 KOLECT(7)=IDOT
55 ENCODE(10,56,NEWNAM(NXTFIL))KOLECT
56 FORMAT(10A1)
GO TO 18
C610045095007$&
END
The program listed below calls GETFIL, reports the results,
then repeats the process.
C DEMONSTRATE GETFIL WHICH IN TURN DEMONSTRATES DAFLAG
DIMENSION IBUFFR(72),INILTR(20),KNTLTR(20),
1NEWNUL(10),NEWDSK(10),INTPTH(4)
DOUBLE PRECISION NEWNAM(10),NEWPTH(3,10),DBLPTH(2),
1KMDNAM
EQUIVALENCE(INTPTH,DBLPTH)
DATA KMDNAM,KMDEXT,KMDDVC/6HGETFIL,3HCCL,3HDSK/
DATA ITTY,JTTY,KMDNUM/5,5,1/
FASP, FORTRAN Alphameric Subroutine Package Page 90
DAFLAG, Routine to Locate Components of File Specification
WRITE(JTTY,1)
1 FORMAT(8H MAXFIL ,$)
READ(ITTY,2)MAXFIL
2 FORMAT(I)
IF(MAXFIL.LE.0)MAXFIL=1
IF(MAXFIL.GT.10)MAXFIL=10
KIND=0
C
C GET NEXT PORTION OF FILE SPECIFICATIONS
3 CALL GETFIL(MAXFIL,ITTY,JTTY,KMDNUM,KMDDVC,
1KMDNAM,KMDEXT,20,72,KIND,NEWNUL,NEWDSK,
2NEWNAM,NEWPTH,LCNRIT,IBUFFR,MAXFLG,INILTR,KNTLTR,
3LCNOWN)
GO TO(4,6,9,8,21),KIND
C
C END OF FILE SPECIFICATIONS
4 WRITE(JTTY,5)
5 FORMAT(5H DONE)
GO TO 3
6 WRITE(JTTY,7)
7 FORMAT(10H SEMICOLON)
GO TO 3
C
C LIST FILES
8 INDEX=LCNRIT
GO TO 11
9 IF(LCNRIT.LE.0)GO TO 3
INDEX=0
10 INDEX=INDEX+1
11 IF(NEWNUL(INDEX).NE.0)GO TO 13
IF(INDEX.GE.LCNRIT)GO TO 16
WRITE(JTTY,12)INDEX
12 FORMAT(5H LEFT,I4,8H MISSING)
GO TO 10
13 IF(NEWPTH(1,INDEX).NE.0)GO TO 15
IF(INDEX.GE.LCNRIT)GO TO 18
WRITE(JTTY,14)INDEX,NEWNUL(INDEX),NEWDSK(INDEX),
1NEWNAM(INDEX)
14 FORMAT(5H LEFT,I4,I2,1X,A5,1H:,1A10,1X,O6,1X,O6,
11X,A5,A1)
GO TO 10
15 DBLPTH(1)=NEWPTH(1,INDEX)
DBLPTH(2)=NEWPTH(2,INDEX)
IF(INDEX.GE.LCNRIT)GO TO 20
WRITE(JTTY,14)INDEX,NEWNUL(INDEX),NEWDSK(INDEX),
1NEWNAM(INDEX),INTPTH
GO TO 10
16 WRITE(JTTY,17)INDEX
17 FORMAT(6H RIGHT,I3,8H MISSING)
GO TO 3
18 WRITE(JTTY,19)INDEX,NEWNUL(INDEX),NEWDSK(INDEX),
1NEWNAM(INDEX)
19 FORMAT(6H RIGHT,I3,I2,1X,A5,1H:,1A10,1X,O6,1X,O6,
11X,A5,A1)
FASP, FORTRAN Alphameric Subroutine Package Page 91
DAFLAG, Routine to Locate Components of File Specification
GO TO 3
20 WRITE(JTTY,19)INDEX,NEWNUL(INDEX),NEWDSK(INDEX),
1NEWNAM(INDEX),INTPTH
GO TO 3
C
C LIST SWITCHES
21 M=0
DO 24 I=1,MAXFLG
J=INILTR(I)
K=J+KNTLTR(I)-1
IF(K.LT.J)GO TO 24
IF(M.EQ.0)WRITE(JTTY,22)LCNOWN,(IBUFFR(L),L=J,K)
22 FORMAT(7H SWITCH,I2,3X,72A1)
IF(M.NE.0)WRITE(JTTY,23)LCNOWN,(IBUFFR(L),L=J,K)
23 FORMAT(7H " " ,I2,3X,72A1)
M=1
24 CONTINUE
GO TO 3
END
The following is a typical dialog between the user and the
program listed above.
MAXFIL 3
*FIRST.1ST,SECOND.2ND=THIRD.3RD,FOURTH.4TH
LEFT 1 3 :FIRST .1ST
LEFT 2 3 :SECOND.2ND
RIGHT 3 3 :THIRD .3RD
RIGHT 1 3 :FOURTH.4TH
DONE
*FIRST.1ST,SECOND.2ND,THIRD.3RD,FOURTH.4TH
RIGHT 1 3 :FIRST .1ST
RIGHT 2 3 :SECOND.2ND
RIGHT 3 3 :THIRD .3RD
RIGHT 1 3 :FOURTH.4TH
DONE
*FIRST,
&SECOND=
&THIRD,
LEFT 1 2 :FIRST .
LEFT 2 2 :SECOND.
RIGHT 3 2 :THIRD .
&FOURTH
RIGHT 1 2 :FOURTH.
DONE
*FIRST,
&SECOND,
&THIRD,
RIGHT 1 2 :FIRST .
RIGHT 2 2 :SECOND.
RIGHT 3 2 :THIRD .
&FOURTH
RIGHT 1 2 :FOURTH.
DONE
FASP, FORTRAN Alphameric Subroutine Package Page 92
DAFLAG, Routine to Locate Components of File Specification
*/A 1ST/B,/C 2ND/D=/E 3RD/F,/G 4TH/H
SWITCH 1 A
SWITCH 1 B
SWITCH 2 C
SWITCH 2 D
SWITCH 3 E
SWITCH 3 F
LEFT 1 2 :1ST .
LEFT 2 2 :2ND .
RIGHT 3 2 :3RD .
SWITCH 1 G
SWITCH 1 H
RIGHT 1 2 :4TH .
DONE
*/A 1ST/B,/C 2ND/D,/E 3RD/F,/G 4TH/H
SWITCH 1 A
SWITCH 1 B
SWITCH 2 C
SWITCH 2 D
SWITCH 3 E
SWITCH 3 F
RIGHT 1 2 :1ST .
RIGHT 2 2 :2ND .
RIGHT 3 2 :3RD .
SWITCH 1 G
SWITCH 1 H
RIGHT 1 2 :4TH .
DONE
*LPT:FIRST.[6000,56],SECOND=
&THIRD,DSK:FOURTH,FIFTH[6002,56,RENBR],SIXTH
LEFT 1 3 LPT :FIRST . 006000 000056
LEFT 2 2 :SECOND.
RIGHT 3 2 :THIRD .
RIGHT 1 2 DSK :FOURTH.
RIGHT 1 2 DSK :FIFTH . 006002 000056 RENBR
RIGHT 1 2 DSK :SIXTH . 006002 000056 RENBR
DONE
*=FIRST
RIGHT 1 2 :FIRST .
DONE
*/A::C=SECOND
SWITCH 1 A
" " 1 C
LEFT 1 MISSING
RIGHT 2 2 :SECOND.
DONE
FASP, FORTRAN Alphameric Subroutine Package Page 93
DAFLAG, Routine to Locate Components of File Specification
*,SECOND=,FOURTH;FIRST,=THIRD,!TERMINAL COMMA IMPLIES &
LEFT 1 MISSING
LEFT 2 2 :SECOND.
RIGHT 3 MISSING
RIGHT 1 2 :FOURTH.
SEMICOLON
LEFT 1 2 :FIRST .
LEFT 2 MISSING
RIGHT 3 2 :THIRD .
&FOURTH
RIGHT 1 2 :FOURTH.
DONE
*,=,!THIS AND NEXT 3 LINES ARE REPEATED WITH ; BELOW
LEFT 1 MISSING
LEFT 2 MISSING
RIGHT 3 MISSING
&FOURTH
RIGHT 1 2 :FOURTH.
DONE
*/A,/B=/C,/D
SWITCH 1 A
SWITCH 2 B
SWITCH 3 C
LEFT 1 MISSING
LEFT 2 MISSING
RIGHT 3 MISSING
SWITCH 1 D
&FOURTH
RIGHT 1 2 :FOURTH.
DONE
*,=,;FOURTH;/A,/B=/C,/D;FOURTH!SAME AS ABOVE 4 LINES
LEFT 1 MISSING
LEFT 2 MISSING
RIGHT 3 MISSING
RIGHT 1 2 :FOURTH.
SEMICOLON
SWITCH 1 A
SWITCH 2 B
SWITCH 3 C
LEFT 1 MISSING
LEFT 2 MISSING
RIGHT 3 MISSING
SWITCH 1 D
RIGHT 1 2 :FOURTH.
DONE
The GETFIL routine demonstrated above repacks each file
specification into a form which can be used in a PDP-10
FORTRAN OPEN statement, but does not actually open the files
and does not evaluate the associated switches. The routine
named OPNFIL which is listed below is a typical example of
the manner in which GETFIL could be called for an
application in which the user is allowed to specify up to 2
FASP, FORTRAN Alphameric Subroutine Package Page 94
DAFLAG, Routine to Locate Components of File Specification
output files and many input files. OPNFIL identifies each
output file as being of either of 2 types, designated as
types one and two, depending upon the letters which form the
switches which appear with the output file specification.
Although the calling program specifies default extensions
for the output file names of each type, the first component
of the name of the first input file is used as the first
component of the name of the output file if the first
component of the name of the output file is not specified by
the user. If no more than 1 output files is specified, and
if a switch does not identify the type of this output file,
then the single output file is of type one.
For example, if the default extensions for the type one
output file, for the type two output file, and for the input
files are .ONE, .TWO and .SRC respectively, and if the
switch /O identifies a type one output file while the switch
/T identifies a type two output file, then the short forms
of the commands shown at the left in the table below are
equivalent to the longer forms shown at the right.
THIRD or =THIRD to THIRD.ONE=THIRD.SRC
THIRD/O or /O=THIRD to THIRD.ONE/O=THIRD.SRC
THIRD/T or /T=THIRD to THIRD.TWO/T=THIRD.SRC
THIRD/O/T to THIRD.ONE/O,THIRD.TWO/T=THIRD.SRC
/O=THIRD/T to " " "
/T=THIRD/O to " " "
/O/T=THIRD to " " "
/O,SECOND=THIRD to THIRD.ONE/O,SECOND.TWO=THIRD.SRC
/T,SECOND=THIRD to THIRD.TWO/T,SECOND.ONE=THIRD.SRC
,SECOND/O=THIRD to THIRD.TWO,SECOND.ONE/O=THIRD.SRC
,SECOND/T=THIRD to THIRD.ONE,SECOND.TWO/T=THIRD.SRC
FIRST/O,SECOND=THIRD to FIRST.ONE/O,SECOND.TWO=THIRD.SRC
FIRST/T,SECOND=THIRD to FIRST.TWO/T,SECOND.ONE=THIRD.SRC
FIRST,SECOND/O=THIRD to FIRST.TWO,SECOND.ONE/O=THIRD.SRC
FIRST,SECOND/T=THIRD to FIRST.ONE,SECOND.TWO/T=THIRD.SRC
FIRST/O,SECOND/T=THIRD to FIRST.ONE/O,SECOND.TWO/T=THIRD.SRC
FIRST/O,SECOND=THIRD/T to " " "
FIRST,SECOND/T=THIRD/O to " " "
FIRST/T,SECOND/O=THIRD to FIRST.TWO/T,SECOND.ONE/O=THIRD.SRC
FIRST/T,SECOND=THIRD/O to " " "
FIRST,SECOND/O=THIRD/T to " " "
As can be seen in the above examples, switches implying both
types of output files can be present if no output file is
specified since then the default characteristics are used
for both output files. If a single output file is
specified, but switches implying both types are present,
then the OPNFIL routine does not attempt to determine which
output file has been specified and which has not been
specified, so the command is considered to be ambiguous.
None of the following commands would be allowed.
FASP, FORTRAN Alphameric Subroutine Package Page 95
DAFLAG, Routine to Locate Components of File Specification
FIRST=THIRD/O/T FIRST/O=THIRD/T FIRST,SECOND=THIRD/O/T
FIRST/O/T=THIRD FIRST/T=THIRD/O FIRST/O/T,SECOND=THIRD
FIRST,SECOND/O/T=THIRD
The argument list of routine OPNFIL is
SUBROUTINE OPNFIL( ITTY, JTTY,KMDNUM,KMDDVC,KMDNAM,
1 KMDEXT,ID1NUM,ID1DVC,ID1EXT,ID2NUM,ID2DVC,ID2EXT,
2 ID3NUM,ID3DVC,ID3EXT,MAXTTL,MAXBFR, KIND,KNDFLG,
3 NUMFLG,LTRTTL,ID1OPN,ID2OPN,IBUFFR)
with the associated DIMENSION and DOUBLE PRECISION
statements
DIMENSION KNDFLG(27),NUMFLG(27),LTRTTL(MAXTTL),
1IBUFFR(MAXBFR)
DOUBLE PRECISION KMDNAM
The OPNFIL arguments named ITTY, JTTY, KMDNUM, KMDDVC,
KMDNAM, KMDEXT, MAXBFR and IBUFFR are identical to the
GETFIL arguments having the same names and have already been
described. The following arguments, together with ITTY,
JTTY, KMDNUM, KMDDVC, KMDNAM, KMDEXT and MAXBFR, are used
only for input and are returned unchanged.
ID1NUM = number of the unit upon which this routine is to
open an output file which is to be of type one. An
output file will be of type one if its file
specification is associated with a switch for which
the corresponding value in the KNDFLG array is 1.
ID1DVC = name in 5H form of the device upon which the type
one output file is to be written if the user does
not supply a device name followed by a colon at the
start of the specification of the type one output
file.
ID1EXT = name in 3H form which is to be used as the second
component (the file name extension) of the name of
the type one output file if no second component of
the name of the type one output file is supplied by
the user. If the name of the type one output file
is not to have such a second component, then the
user must type a period following the first
component of its name. If the user does not
specify the first component of the name of the type
one output file, then the first component of the
name of the first input file is used as the first
component of the name of the type one output file.
ID2NUM = number of the unit upon which this routine is to
open an output file which is to be of type two. An
output file will be of type two if its file
specification is associated with a switch for which
FASP, FORTRAN Alphameric Subroutine Package Page 96
DAFLAG, Routine to Locate Components of File Specification
the corresponding value in the KNDFLG array is 2.
ID2DVC = name in 5H form of the device upon which the type
two output file is to be written if the user does
not supply a device name followed by a colon at the
start of the specification of the type two output
file.
ID2EXT = name in 3H form which is to be used as the second
component (the file name extension) of the name of
the type two output file if no second component of
the name of the type two output file is supplied by
the user. If the name of the type two output file
is not to have such a second component, then the
user must type a period following the first
component of its name. If the user does not
specify the first component of the name of the type
two output file, then the first component of the
name of the first input file is used as the first
component of the name of the type two output file.
ID3NUM = number of the unit upon which this routine is to
open the next input file specified by the user.
ID3DVC = name in 5H form of the device upon which the next
input file is to be opened if no device name has
been specified by the user for any previous input
file and if no device name is specified by the user
for this new input file. Once the user has
specified a device name for an input file, however,
then the device name specified by the user
continues to be applied to any subsequent files for
which the device name is not specified until the
processing of the current command has been
completed.
ID3EXT = name in 3H form which is to be used as the second
component (the file name extension) of the name of
each input file for which no second component of
the file name is supplied by the user. If the file
name is not to have such a second component, then
the user must type a period following the first
component of the name.
MAXTTL = maximum number of locations in the LTRTTL array
which can be used to return the text which was
found to the right of a leading apostrophe. The
text extends through the next apostrophe or through
the rightmost printing character on the line if the
text is not terminated by a second apostrophe. The
output file with which the text appears will be
identified as either a type one or a type two
output file if the initial absolute value of
KNDFLG(27) is either 1 or 2 respectively. If such
FASP, FORTRAN Alphameric Subroutine Package Page 97
DAFLAG, Routine to Locate Components of File Specification
text is found, then KNDFLG(27) is returned
containing the negative of its initial absolute
value and NUMFLG(27) is returned containing the
number of characters returned in the LTRTTL array.
The following argument must be zeroed by the calling program
before this routine is first called, but then the value
returned by this routine should be sent to the following
call of this routine unchanged.
KIND = defined similarly to the argument of the same name
in the argument list of the GETFIL routine, except
that the OPNFIL routine does not return KIND=5
since the OPNFIL routine itself evaluates the
switches which the GETFIL routine has indicated by
setting KIND=5. KIND should be set to zero (or to
1) before this routine is first called, or whenever
the interpretation of the current set of file
specifications is to be abandoned. KIND is
returned describing the reason why control has been
transferred back to the calling program, and should
not be changed by the calling program if this
routine is to be called again to continue the
interpretation of the same sequence of file
specifications.
= 1, returned if the previous call to this routine
opened the final input file specified by the user.
The next call to this routine will ask the user to
type a new set of file specifications.
= 2, returned if the previous call to this routine
opened the final input file specified to the left
of a semicolon. If this routine is called again
without KIND having first been zeroed, then the
evaluation of a new set of file specifications will
be begun in the text appearing to the right of the
semicolon.
= 3, the current call to this routine has opened the
output file or files and has opened the first input
file.
= 4, the current call to this routine has opened the
second or a subsequent input file.
The following argument is used both for sending information
to this routine when KIND is input set to a value less than
3, and for returning information to the calling program when
KIND is returned set to 3. The contents of this array are
returned unchanged if KIND is returned set to a value other
than 3. The calling program must define the contents of
this array before this routine is first called.
KNDFLG = an array dimensioned at 27 which must initially
indicate the output file types to be associated
with each of the single letter switches /A through
/Z and /' (or lone ') respectively. If this
FASP, FORTRAN Alphameric Subroutine Package Page 98
DAFLAG, Routine to Locate Components of File Specification
routine is called with KIND set to 2 or less, then
this routine redefines the KNDFLG array to contain
the absolute values of its original contents. If
the current call to this routine begins the
processing of a new command such that KIND is
returned containing the value 3, then those
locations within the KNDFLG array having as their
subscripts the serial locations within the alphabet
of the letters which are found as switches (taking
the apostrophe to be the 27th letter in the
alphabet) are then returned containing the
negatives of their absolute values. The contents
of the KNDFLG array are returned unchanged if KIND
is returned set to 4 indicating that the current
call to this routine has continued the processing
of a command begun by a previous call to this
routine.
If either of the values -1 or 1 appear in the
KNDFLG array location having as its subscript the
serial location within the alphabet of the letter
which is used as a switch appearing with an output
file specification, then this type one output file
will be opened upon the unit number identified by
ID1NUM, upon the device specified by ID1DVC if no
device is specified by the user, and with the file
name extension specified by ID1EXT if no extension
is specified by the user, and ID1OPN will be
returned set to 1. If either of the values -2 or 2
appear in the KNDFLG array location having as its
subscript the serial location within the alphabet
of the letter which is used as a switch appearing
with an output file specification, then this type
two output file will be opened upon the unit number
identified by ID2NUM, upon the device specified by
ID2DVC if no device is specified by the user, and
with the file name extension specified by ID2EXT if
no extension is specified by the user, and ID2OPN
will be returned set to 1. For example, if either
/C or /c appear with the name of an output file and
if KNDFLG(3)=2 or -2, then KNDFLG(3) will be
returned containing -2 and this output file will be
opened upon the unit identified by ID2NUM.
Switches for which the corresponding locations in
the KNDFLG array contain the value zero cause the
routine named HLPFIL to be called to display a help
message to the user, then force the user to supply
a new set of file specifications. For example,
KNDFLG(8) might contain the value zero so that /H
would generate the help message. HLPFIL has as its
single argument the terminal unit number. The
HLPFIL routine must be supplied by the calling
program.
FASP, FORTRAN Alphameric Subroutine Package Page 99
DAFLAG, Routine to Locate Components of File Specification
Switches for which the corresponding locations in
the KNDFLG array do not contain one of the values
-2 through 2 can appear with either output file
specification or with the leftmost input file
specification but do not identify the type of the
output files.
The following arguments are used only for returning
information to the calling program when the processing of a
new set of commands is being begun by the current call to
this routine. Their input values are returned unchanged if
KIND is returned set to a value other than 3.
NUMFLG = an array dimensioned at 27 which is used for
returning the values which appeared with separating
colons after the switches in the file
specifications. The NUMFLG array locations in
which these values are returned have as their
subscripts the serial locations within the alphabet
of the letters which are used as the switches.
Such numbers must be specified as decimal integers,
but cannot contain exponents. If the switch /C:123
appears in the file specification, then KNDFLG(3)
will be returned containing the negative of its
original absolute value, and NUMFLG(3) will be
returned containing the integer value 123. If an
apostrophe is found with the set of file
specifications, then NUMFLG(27) is returned
containing the number of characters which appeared
to the right of the apostrophe and which are
returned in the LTRTTL array.
LTRTTL = array in which the characters which were found to
the right of a leading apostrophe are returned with
1 character per array location as read by a
multiple of an A1 format. The LTRTTL array must be
dimensioned to at least the value of MAXTTL. If
such text is found to the right of a leading
apostrophe, then KNDFLG(27) is returned containing
the negative of its initial absolute value and
NUMFLG(27) is returned containing the number of
characters returned in the LTRTTL array. The
terminal apostrophes are never returned in the
LTRTTL array and are not included in the character
count returned in NUMFLG(27).
ID1OPN = 0, returned if a type one output file was not
opened.
= 1, returned if a type one output file was opened.
ID2OPN = 0, returned if a type two output file was not
opened.
= 1, returned if a type two output file was opened.
FASP, FORTRAN Alphameric Subroutine Package Page 100
DAFLAG, Routine to Locate Components of File Specification
The OPNFIL routine is listed below. It is expected that
this routine will have to be modified to fit the particular
application for which this routine is used.
SUBROUTINE OPNFIL( ITTY, JTTY,KMDNUM,KMDDVC,KMDNAM,
1 KMDEXT,ID1NUM,ID1DVC,ID1EXT,ID2NUM,ID2DVC,ID2EXT,
2 ID3NUM,ID3DVC,ID3EXT,MAXTTL,MAXBFR, KIND,KNDFLG,
3 NUMFLG,LTRTTL,ID1OPN,ID2OPN,IBUFFR)
C RENBR(/OPEN FILES FOR FILE,FILE=FILE,FILE COMMAND)
C
COMMON/FASPY/NEWNUL(3),NEWDSK(3),NEWNAM(3),
1NEWPTH(3,3),LCNRIT
DIMENSION KNDFLG(27),NUMFLG(27),LTRTTL(MAXTTL),
1IBUFFR(MAXBFR),LTRABC(27),LWRABC(27),LTRDGT(10),
2INILTR(6),KNTLTR(6)
DOUBLE PRECISION KMDNAM,NEWNAM,NEWPTH,PTHONE(3),
1PTHTWO(3),PTHTHR(3),FILONE,FILTWO,FILTHR,FILNAM
DATA LTRABC/1HA,1HB,1HC,1HD,1HE,1HF,1HG,1HH,1HI,1HJ,
1 1HK,1HL,1HM,1HN,1HO,1HP,1HQ,1HR,1HS,1HT,
2 1HU,1HV,1HW,1HX,1HY,1HZ,1H'/
DATA LWRABC/1Ha,1Hb,1Hc,1Hd,1He,1Hf,1Hg,1Hh,1Hi,1Hj,
1 1Hk,1Hl,1Hm,1Hn,1Ho,1Hp,1Hq,1Hr,1Hs,1Ht,
2 1Hu,1Hv,1Hw,1Hx,1Hy,1Hz,1H'/
DATA LTRDGT/1H0,1H1,1H2,1H3,1H4,1H5,1H6,1H7,1H8,1H9/
DATA LTRPLS,LTRMNS,IBLANK/1H+,1H-,5H /
C
C INITIALIZE STORAGE OF COMMAND
INITAL=KIND-2
IF(INITAL.GT.0)GO TO 4
GO TO 2
1 KIND=0
2 ID1OPN=0
ID2OPN=0
DO 3 I=1,27
IF(KNDFLG(I).LT.0)KNDFLG(I)=-KNDFLG(I)
3 NUMFLG(I)=0
KNDONE=0
KNDTWO=0
KNDTHR=0
C
C GET NEXT COMPONENT OF COMMAND TYPED BY USER
4 CALL GETFIL(3,ITTY,JTTY,KMDNUM,KMDDVC,
1KMDNAM,KMDEXT,6,MAXBFR,KIND,NEWNUL,NEWDSK,
2NEWNAM,NEWPTH,LCNRIT,IBUFFR,MAXFLG,INILTR,KNTLTR,
3LCNOWN)
GO TO(45,45,23,31,5),KIND
C
C DECIDE WHICH SWITCH WAS GIVEN
5 IF(MAXFLG.LE.0)GO TO 4
IF(KNTLTR(1).LE.0)GO TO 4
IF(INITAL.GT.0)GO TO 4
LOCLTR=INILTR(1)
LTRNOW=IBUFFR(LOCLTR)
NEWFLG=0
FASP, FORTRAN Alphameric Subroutine Package Page 101
DAFLAG, Routine to Locate Components of File Specification
6 NEWFLG=NEWFLG+1
IF(NEWFLG.GT.27)GO TO 4
IF(LTRABC(NEWFLG).EQ.LTRNOW)GO TO 7
IF(LWRABC(NEWFLG).NE.LTRNOW)GO TO 6
7 IF(KNDFLG(NEWFLG).EQ.0)GO TO 35
IF(KNDFLG(NEWFLG).GT.0)KNDFLG(NEWFLG)=-KNDFLG(NEWFLG)
IF(KNDFLG(NEWFLG).EQ.-2)GO TO 11
IF(KNDFLG(NEWFLG).NE.-1)GO TO 15
C
C MARK THAT FILE BEARS TYPE ONE SWITCH
ID1OPN=1
GO TO(8,9,10),LCNOWN
8 IF(KNDONE.EQ.1)GO TO 15
IF(KNDONE.NE.3)KNDONE=KNDONE+1
GO TO 15
9 IF(KNDTWO.EQ.1)GO TO 15
IF(KNDTWO.NE.3)KNDTWO=KNDTWO+1
GO TO 15
10 IF(KNDTHR.EQ.1)GO TO 15
IF(KNDTHR.NE.3)KNDTHR=KNDTHR+1
GO TO 15
C
C MARK THAT FILE BEARS TYPE TWO SWITCH
11 ID2OPN=1
GO TO(12,13,14),LCNOWN
12 IF(KNDONE.LE.1)KNDONE=KNDONE+2
GO TO 15
13 IF(KNDTWO.LE.1)KNDTWO=KNDTWO+2
GO TO 15
14 IF(KNDTHR.LE.1)KNDTHR=KNDTHR+2
C
C STORE QUOTED TEXT STRING
15 IVALUE=0
IF(NEWFLG.LT.27)GO TO 17
LMTLTR=LOCLTR+KNTLTR(1)
16 IF(IVALUE.GE.MAXTTL)GO TO 22
LOCLTR=LOCLTR+1
IF(LOCLTR.GE.LMTLTR)GO TO 22
IVALUE=IVALUE+1
LTRTTL(IVALUE)=IBUFFR(LOCLTR)
GO TO 16
C
C EVALUATE NUMBERS IN RANGE OF SWITCH
17 IF(MAXFLG.LE.1)GO TO 22
IF(KNTLTR(2).LE.0)GO TO 22
LOCLTR=INILTR(2)
LMTLTR=LOCLTR+KNTLTR(2)
I=0
IF(IBUFFR(LOCLTR).EQ.LTRPLS)GO TO 18
IF(IBUFFR(LOCLTR).NE.LTRMNS)GO TO 19
I=1
18 LOCLTR=LOCLTR+1
19 IF(LOCLTR.GE.LMTLTR)GO TO 21
LTRNOW=IBUFFR(LOCLTR)
FASP, FORTRAN Alphameric Subroutine Package Page 102
DAFLAG, Routine to Locate Components of File Specification
DO 20 L=1,10
IF(LTRDGT(L).NE.LTRNOW)GO TO 20
IVALUE=(10*IVALUE)+L-1
GO TO 18
20 CONTINUE
21 IF(I.NE.0)IVALUE=-IVALUE
22 NUMFLG(NEWFLG)=IVALUE
GO TO 4
C
C SET DEFAULT OUPUT DEVICE NAMES AND PATHS
23 IF(NEWNUL(LCNRIT).LE.1)GO TO 38
IF(NEWNUL(LCNRIT).EQ.4)GO TO 38
IF(ID1OPN.EQ.ID2OPN)ID1OPN=1
FILONE=NEWNAM(LCNRIT)
FILTWO=FILONE
MORONE=ID1EXT
MORTWO=ID2EXT
PTHONE(1)=0
PTHTWO(1)=0
LOCONE=ID1DVC
LOCTWO=ID2DVC
GO TO(31,24,25),LCNRIT
C
C SINGLE FILE LEFT OF EQUAL SIGN
24 IF(NEWNUL(1).EQ.0)GO TO 31
IF(ID1OPN.EQ.ID2OPN)GO TO 36
KNDONE=1
KNDTWO=1
GO TO 26
C
C TWO FILES LEFT OF EQUAL SIGN
25 IF(KNDONE.GE.3)GO TO 36
IF(KNDTWO.GE.3)GO TO 36
IF(KNDONE.EQ.KNDTWO)GO TO 36
ID1OPN=1
ID2OPN=1
IF(KNDONE.EQ.0)KNDONE=3-KNDTWO
IF(KNDTWO.EQ.0)KNDTWO=3-KNDONE
26 IF(NEWNUL(KNDONE).EQ.0)GO TO 29
IF(NEWDSK(KNDONE).NE.IBLANK)LOCONE=NEWDSK(KNDONE)
DO 27 I=1,3
27 PTHONE(I)=NEWPTH(I,KNDONE)
IF(NEWNUL(KNDONE).LE.1)GO TO 29
FILNAM=NEWNAM(KNDONE)
IF(NEWNUL(KNDONE).LE.3)FILONE=FILNAM
IF(NEWNUL(KNDONE).GE.3)DECODE(10,28,FILNAM)MORONE
28 FORMAT(7X,1A3)
29 IF(NEWNUL(KNDTWO).EQ.0)GO TO 31
IF(NEWDSK(KNDTWO).NE.IBLANK)LOCTWO=NEWDSK(KNDTWO)
DO 30 I=1,3
30 PTHTWO(I)=NEWPTH(I,KNDTWO)
IF(NEWNUL(KNDTWO).LE.1)GO TO 31
FILNAM=NEWNAM(KNDTWO)
IF(NEWNUL(KNDTWO).LE.3)FILTWO=FILNAM
FASP, FORTRAN Alphameric Subroutine Package Page 103
DAFLAG, Routine to Locate Components of File Specification
IF(NEWNUL(KNDTWO).GE.3)DECODE(10,28,FILNAM)MORTWO
C
C OPEN INPUT FILE
31 IF(NEWNUL(LCNRIT).EQ.0)GO TO 4
FILTHR=NEWNAM(LCNRIT)
MORTHR=ID3EXT
IF(NEWNUL(LCNRIT).GE.3)DECODE(10,28,FILTHR)MORTHR
LOCTHR=NEWDSK(LCNRIT)
IF(LOCTHR.EQ.IBLANK)LOCTHR=ID3DVC
DO 32 I=1,3
32 PTHTHR(I)=NEWPTH(I,LCNRIT)
ENCODE(10,33,FILNAM)FILTHR,MORTHR
33 FORMAT(1A6,1H.,1A3)
OPEN(UNIT=ID3NUM,FILE=FILNAM,DIRECTORY=PTHTHR,
1DEVICE=LOCTHR,ACCESS='SEQIN',ERR=43)
IF(INITAL.GT.0)GO TO 47
C
C OPEN OUTPUT FILES
IF(ID1OPN.EQ.0)GO TO 34
ENCODE(10,33,FILNAM)FILONE,MORONE
OPEN(UNIT=ID1NUM,FILE=FILNAM,DIRECTORY=PTHONE,
1DEVICE=LOCONE,ACCESS='SEQOUT',ERR=40)
IF(ID2OPN.EQ.0)GO TO 46
34 ENCODE(10,33,FILNAM)FILTWO,MORTWO
OPEN(UNIT=ID2NUM,FILE=FILNAM,DIRECTORY=PTHTWO,
1DEVICE=LOCTWO,ACCESS='SEQOUT',ERR=42)
GO TO 46
C
C ISSUE HELP MESSAGE AND THEN CLEAR COMMAND IF ANY
35 CALL HLPFIL(JTTY)
GO TO 1
C
C ERROR IN COMMAND TYPED BY USER
36 WRITE(JTTY,37)
37 FORMAT(31H AMBIGUOUS OUTPUT SPECIFICATION)
GO TO 1
38 WRITE(JTTY,39)
39 FORMAT(34H 1ST SOURCE FILE MUST BE SPECIFIED)
GO TO 1
40 WRITE(JTTY,41)LOCONE,FILNAM
41 FORMAT(26H CANNOT WRITE OUTPUT FILE ,1A5,1H:,1A10)
GO TO 1
42 WRITE(JTTY,41),LOCTWO,FILNAM
GO TO 1
43 WRITE(JTTY,44)LOCTHR,FILNAM
44 FORMAT(25H CANNOT READ SOURCE FILE ,1A5,1H:,1A10)
GO TO 4
C
C RETURN TO CALLING PROGRAM
45 IF(INITAL.LE.0)GO TO 2
GO TO 47
46 KIND=3
47 RETURN
C024147266970'abcdefghijklmnopqrstuvwxyz:
FASP, FORTRAN Alphameric Subroutine Package Page 104
DAFLAG, Routine to Locate Components of File Specification
END
The following program demonstrates the manner in which the
OPNFIL routine might be called.
C RENBR(TSTOPN/DEMONSTRATE OPNFIL ROUTINE)
DOUBLE PRECISION KMDNAM
DIMENSION IBUFFR(72),JBUFFR(72),KNDFLG(27),
1NUMFLG(27),LTRTTL(30),LTRABC(26)
DATA LTRABC/1HA,1HB,1HC,1HD,1HE,1HF,1HG,1HH,1HI,1HJ,
1 1HK,1HL,1HM,1HN,1HO,1HP,1HQ,1HR,1HS,1HT,
2 1HU,1HV,1HW,1HX,1HY,1HZ/
DATA KMDDVC,KMDNAM,KMDEXT/3HDSK,6HOPNFIL,3HCCL/
DATA ID1DVC,ID1EXT/3HDSK,3HONE/
DATA ID2DVC,ID2EXT/3HDSK,3HTWO/
DATA ID3DVC,ID3EXT/3HDSK,3HSRC/
DATA MAXTTL,MAXBFR/30,72/
DATA KMDNUM,ITTY,JTTY,ID1NUM,ID2NUM,ID3NUM/
124,5,5,1,20,21/
C
C DEFINE FILE TYPE ASSOCIATED WITH EACH FLAG
C /H, /O, /T GIVE HELP, TYPE ONE, TYPE TWO RESPECTIVELY
C NO OTHER LETTERS IMPLY ASSOCIATED FILE TYPES
DO 1 I=1,27
1 KNDFLG(I)=3
KNDFLG(8)=0
KNDFLG(15)=1
KNDFLG(20)=2
KIND=0
C
C OPEN NEXT INPUT FILE
2 CALL OPNFIL(ITTY,JTTY,KMDNUM,KMDDVC,KMDNAM,
1KMDEXT,ID1NUM,ID1DVC,ID1EXT,ID2NUM,ID2DVC,ID2EXT,
2ID3NUM,ID3DVC,ID3EXT,MAXTTL,MAXBFR,KIND,KNDFLG,
3NUMFLG,LTRTTL,ID1OPN,ID2OPN,IBUFFR)
GO TO(12,14,3,10),KIND
C
C REPORT VALUES OF SWITCHES
3 NEEDED=0
4 NEEDED=NEEDED-1
DO 6 INDEX=1,26
IF(KNDFLG(INDEX).EQ.NEEDED)WRITE(JTTY,5)
1LTRABC(INDEX),NUMFLG(INDEX)
5 FORMAT(1X,1A1,I5)
6 CONTINUE
IF(NEEDED.GE.-2)GO TO 4
J=NUMFLG(27)
IF(J.GT.0)WRITE(JTTY,7)(LTRTTL(I),I=1,J)
7 FORMAT(8H TITLE: ,30A1)
IF(ID1OPN.NE.0)WRITE(JTTY,8)
8 FORMAT(21H TYPE ONE OUTPUT FILE)
IF(ID2OPN.NE.0)WRITE(JTTY,9)
9 FORMAT(21H TYPE TWO OUTPUT FILE)
C
FASP, FORTRAN Alphameric Subroutine Package Page 105
DAFLAG, Routine to Locate Components of File Specification
C COPY INPUT FILE INTO OUTPUT FILES
10 READ(ID3NUM,11,END=2)JBUFFR
11 FORMAT(72A1)
IF(ID1OPN.NE.0)WRITE(ID1NUM,11)JBUFFR
IF(ID2OPN.NE.0)WRITE(ID2NUM,11)JBUFFR
GO TO 10
C
C END OF USER COMMAND FOUND
12 WRITE(JTTY,13)
13 FORMAT(' DONE')
GO TO 16
14 WRITE(JTTY,15)
15 FORMAT(' SEMICOLON')
16 IF(ID1OPN.NE.0)CLOSE(UNIT=ID1NUM)
IF(ID2OPN.NE.0)CLOSE(UNIT=ID2NUM)
GO TO 2
END
SUBROUTINE HLPFIL(JTTY)
WRITE(JTTY,1)
1 FORMAT(13H HELP MESSAGE)
RETURN
END
FASP, FORTRAN Alphameric Subroutine Package Page 106
DAHEFT, Free Format Numeric Input Routine
DDDDD AAA HH HH EEEEEEEE FFFFFFFF TTTTTTTT
DD DD AAAA HH HH EE FF TT
DD DD AA AA HH HH EE FF TT
DD DD AA AA HHHHHHHH EEEEE FFFFF TT
DD DD AAAAAAA HH HH EE FF TT
DD DD AA AA HH HH EE FF TT
DDDDD AA AA HH HH EEEEEEEE FF TT
DAHEFT, Free Format Numeric Input Routine
------ ---- ------ ------- ----- -------
DAHEFT interprets an array read by the calling program with
a multiple of an A1 format and returns the values contained
in this array. Numbers interpreted by DAHEFT can contain
leading sign, embedded decimal point and/or trailing E with
possibly signed exponent. A percent sign following the
number implies E-2, trailing letter K implies E3 and
trailing letter M implies E6. To be recognized as part of a
number, the percent sign or letters K or M or E must follow
either a sign or a digit or a decimal point.
If a number contains no value digits but does contain a sign
or decimal point, then the value is taken as zero. If the
letter E appears in a number but no digits appear to the
right of the letter E, then the value of the exponent is
taken as zero. Trailing E, E0, E- and E+ are therefore
equivalent to multiplication by 10**0 or 1.
DAHEFT can return the value either as a real number or as an
integer. If only integer values are required, then the
shorter routine named DAIHFT can be called instead of DAHEFT
to evaluate the numbers. DAIHFT can evaluate integer
numbers in the full range of E, K and M notations, and will
accept percent signs and negative exponents. As an example,
DAIHFT would return the integer value 123 after evaluation
of any of the following alphameric representations.
123. 12.3E+1 0.0123E4 0.123K .000123M 12300% +12300E-2
The DAHEFT and DAIHFT Argument Lists
--- ------ --- ------ -------- -----
The argument lists of routines DAHEFT and DAIHFT are
SUBROUTINE DAHEFT(KONTRL,ITRAIL,IEXTRA,IBUFFR,MAXBFR,
1 LOWBFR,KIND ,ISHIFT,JSHIFT,KSHIFT,LSHIFT,IVALUE,
2 VALUE )
and
SUBROUTINE DAIHFT(KONTRL,ITRAIL,IEXTRA,IBUFFR,MAXBFR,
1 LOWBFR,KIND ,ISHIFT,JSHIFT,KSHIFT,LSHIFT,IVALUE)
FASP, FORTRAN Alphameric Subroutine Package Page 107
DAHEFT, Free Format Numeric Input Routine
both with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The argument lists of the 2 routines are identical except
that the DAHEFT argument VALUE does not appear in the DAIHFT
argument list. Since DAIHFT cannot evaluate real numbers,
DAIHFT evaluates decimal integers if KONTRL is input greater
than or equal to zero.
The following arguments are used for input and are returned
unchanged.
KONTRL = 1 or greater, the item in the IBUFFR array
represents a real number. If possible, the real
number will be accumulated as an integer, then be
converted to a real number and shifted if
necessary. KONTRL is then the maximum number of
digits in the integer. The value is output as the
argument VALUE. If the item has more than KONTRL
digits (not counting leftmost zeroes), then the
entire evaluation is done as a real number. The
advantage of calculating the real values in integer
as long as the precision of the computer is not
overflowed is that the calculation of the portion
of the number right of the decimal point is more
exact. As an example, if KONTRL is greater than or
equal to 4, then the number 33.33 can be stored as
the integer 3333, then be converted to the real
value 3333.0 and divided by 100.0 to obtain the
final answer. If it makes no difference whether
the number typed as 33.33 has value 33.33 or
33.32999... then KONTRL can be given the value 1.
= 0, the item in the IBUFFR array is the
representation of a decimal integer. The number
can be typed with a decimal point (for example
1.23K or 1.23E3 equals 1230), but is stored as an
integer in DAHEFT, and is output as argument
IVALUE. Any decimal integer which the computer can
represent can be evaluated. This includes, on twos
complement computers, the largest negative number
the absolute value of which cannot be stored. On
the PDP-10, a 36 bit computer with twos complement
notation, the range of decimal integers is
-34359738368 through 34359738367 (octal notation of
bit patterns being 400000000000 through
377777777777).
= -1, the item in the IBUFFR array is the
representation of an octal integer. The number can
be typed with a decimal point and/or with an
exponent. However, the number following the letter
E of the exponent is evaluated in decimal. The
value of the octal number is returned as the
argument IVALUE. It must be noted that numbers
FASP, FORTRAN Alphameric Subroutine Package Page 108
DAHEFT, Free Format Numeric Input Routine
evaluated as negative octal integers have the
negative octal integer as their value, not as their
bit representation in computer storage. For
example, on a 36 bit twos complement computer, the
octal number -400000000000 (which could also be
typed as -4E11 or -4E+11 where the 11 after the E
is in decimal) is represented as the bit pattern
having the octal notation 400000000000 and the
octal number -377777777777 is represented by the
bit pattern 400000000001.
= -2, do not evaluate numbers. LOWBFR will be
returned pointing to the leftmost printing
character in the buffer, or will be returned
pointing beyond the right end of the buffer if the
buffer contains no printing characters.
ITRAIL = selects whether exponents are to be recognized. If
not, then each number will terminate prior to the
exponent, and LOWBFR will be returned pointing to
the letter starting the exponent. The subsequent
call to this routine will return KIND=2 indicating
that an illegal character has been found if the
calling program does not first increment the value
of LOWBFR.
= -1, exponents expressed in E notation are to be
recognized if and only if preceded either by a
digit, a decimal point or by a plus or minus sign.
The returned value will already have been
multiplied by the radix raised to the indicated
power. The percent sign and the letters K and M
are to be treated the same as any other alphabetic
character.
= 0, no exponents are to be recognized. Numbers will
be terminated prior to percent signs or to the
letters E or K or M.
= 1, percent signs, the letters K and M, and
exponents expressed in E notation, are all to be
recognized if and only if preceded either by a
digit, a decimal point or by a plus or minus sign.
The returned value will already have been
multiplied by the radix raised to the indicated
power.
IEXTRA = an extra shift, stated as the power of the radix
(the radix being 8 if KONTRL is -1, and 10
otherwise), which is always to be applied to the
evaluated number (regardless of the value of
ITRAIL) before the evaluated number is returned to
the calling program. For example, IEXTRA=2 would
return the number of cents corresponding to a
dollar price specified by the contents of the input
text buffer. The shift specified by IEXTRA does
not change the returned values of ISHIFT, JSHIFT,
FASP, FORTRAN Alphameric Subroutine Package Page 109
DAHEFT, Free Format Numeric Input Routine
KSHIFT or LHSIFT which reflect the shifts specified
by the contents of the input text buffer.
IBUFFR = input buffer array containing characters typed by
user, read by a multiple of an A1 format, which is
to be searched for numbers. IBUFFR then contains 1
character per computer storage location.
MAXBFR = subscript of the final (rightmost) location in the
IBUFFR array which can be searched for a number.
The following argument must be set by the calling program
before this routine is first called, and then is returned by
this routine describing the location of the first character
in the input buffer not yet processed.
LOWBFR = subscript within the IBUFFR array of the first
(leftmost) character which can be scanned for the
representation of numbers. If a number is found,
then LOWBFR will be returned pointing to the first
character beyond the number which could not be part
of the number. If a number is not found, then
LOWBFR will be returned pointing to the first
printing character (which would have to be a
character other than a plus sign or a minus sign or
a decimal point or a digit through 7 if octal or
through 9 if decimal), or beyond the end of the
buffer if the buffer does not contain any printing
characters at or to the right of IBUFFR(LOWBFR).
LOWBFR must be set by the calling program before
this routine is first called to process the
contents of the buffer, and should not be changed
by the calling program if this routine locates a
number. If a printing character is found which
cannot start a number, then the contents of the
buffer should be processed by some other FASP
routine, or at least the calling program should
increment the value of LOWBFR by 1 before again
calling this routine.
The following arguments are used for returning information
to the calling program. Their input values are ignored.
KIND = returned describing the kind of item located in the
IBUFFR array.
= 1, no printing character was found at or to the
right of IBUFFR(LOWBFR). LOWBFR is returned
pointing beyond the end of the IBUFFR array. The
calling program should read a new line into the
IBUFFR array and reset LOWBFR to point to the first
character in this array.
= 2, a number was not found, but a printing character
which cannot start a number was found at or to the
right of IBUFFR(LOWBFR). LOWBFR will be returned
FASP, FORTRAN Alphameric Subroutine Package Page 110
DAHEFT, Free Format Numeric Input Routine
pointing to the printing character. The calling
program must increment the value of LOWBFR before
again calling this routine since supplying the same
initial character would produce identical results.
VALUE or IVALUE, whichever is appropriate, is
returned with the value zero (even if ITRAIL is in
the range -13 to -7 as described later in this
documentation), so KIND=2 can be considered
equivalent to KIND=3 if such is appropriate to the
application for which DAHEFT is being used.
= 3, a number was found. LOWBFR is returned pointing
to the character to the right of the representation
of the number. If ITRAIL=0 and the number is
followed by a percent sign or the letters K, M or
E, or if ITRAIL is equal to -1 and the number is
followed by a percent sign or the letters K or M,
then LOWBFR will be returned pointing to this
printing character and no exponent will be
evaluated.
ISHIFT = if KIND is returned containing 3, then ISHIFT is
returned describing which of the characters E, %, K
or M, if any, appeared in the number specification.
If KIND is returned containing a value other than
3, then ISHIFT, JSHIFT, KSHIFT and LSHIFT are all
returned undefined but probably changed.
= -4, the number was followed by the letter E, but
neither sign nor digits appeared in the exponent.
JSHIFT is returned zeroed. No exponent other than
that specified by IEXTRA has been applied to the
returned value (providing that ITRAIL is not in the
range -13 to -7 as described later in this
documentation).
= -3, the number was followed by the letter E and
then by a minus sign, but no digits appeared in the
exponent. JSHIFT is returned zeroed. No exponent
other than that specified by IEXTRA has been
applied to the returned value (providing that
ITRAIL is not in the range -13 to -7 as described
later in this documentation).
= -2, the number was followed by the letter E and
then by a plus sign, but no digits appeared in the
exponent. JSHIFT is returned zeroed. No exponent
other than that specified by IEXTRA has been
applied to the returned value (providing that
ITRAIL is not in the range -13 to -7 as described
later in this documentation).
= -1, the number was followed by the letter E and
then by the number which is returned as the value
of JSHIFT.
= 0, returned if none of the characters E, %, K or M
follow the number. JSHIFT is returned containing
0.
= 1, the number was followed by a percent sign.
FASP, FORTRAN Alphameric Subroutine Package Page 111
DAHEFT, Free Format Numeric Input Routine
JSHIFT is returned containing -2.
= 2, the number was followed by the letter K. JSHIFT
is returned containing 3.
= 3, the number was followed by the letter M. JSHIFT
is returned containing 6.
JSHIFT = if KIND is returned containing 3, then JSHIFT is
returned containing the value of the exponent
evaluated at the right end of the number
representation. JSHIFT is the shift stated as the
power of the radix. If the letter E follows the
number, then JSHIFT is the number appearing to the
right of the letter E, or is returned containing
zero if no digits were found to the right of the
letter E. If one of the characters %, K or M
follow the number, then JSHIFT is returned
containing -2, 3 and 6 respectively. The returned
value of JSHIFT is independent of the input value
of IEXTRA. JSHIFT is returned undefined but
probably changed if KIND is returned containing a
value other than 3.
KSHIFT = if KIND is returned containing 3, then KSHIFT is
returned describing the precision of the least
significant digit specified by the contents of the
input text buffer. KSHIFT is the power of the
radix which if used as the shift would result in
the same value if the same digits where found but
without a decimal point. KSHIFT would be returned
containing -2 if the input text buffer contains
12.34 since this could have been written as 1234E-2
instead. A KSHIFT value of -2 would indicate that
the tenths and hundredths digits were specified.
If the input text buffer contains 12.34K then
KSHIFT would be returned containing 1 since this
could be written as 1234E1 instead. A KSHIFT value
of 1 would indicate that the ones (units) digit was
not specified. The returned value of KSHIFT is
independent of the input value of IEXTRA. KSHIFT
is returned undefined but probably changed if KIND
is returned containing a value other than 3.
LSHIFT = returned specifying the number of digits which were
found counting the leftmost non-zero digit and all
digits which were specified to its right. LSHIFT
is returned undefined but probably changed if KIND
is returned containing a value other than 3.
= -4, the number representation contained neither
leading sign nor decimal point nor value digits.
In order for this value of LSHIFT to be returned,
ITRAIL would have to have a value other than -11,
-10, -9, -1, 0, 1, 9, 10 or 11 as described later
in this documentation, and the first printing
characters in the text buffer would have to form
FASP, FORTRAN Alphameric Subroutine Package Page 112
DAHEFT, Free Format Numeric Input Routine
the representation of an exponent. VALUE or
IVALUE, whichever is appropriate, is returned
containing zero (providing that ITRAIL is not in
the range -13 to -7).
= -3, the number representation began with a minus
sign but contained no value digits. VALUE or
IVALUE, whichever is appropriate, is returned
containing zero (providing that ITRAIL is not in
the range -13 to -7).
= -2, the number representation began with a plus
sign but contained no value digits. VALUE or
IVALUE, whichever is appropriate, is returned
containing zero (providing that ITRAIL is not in
the range -13 to -7).
= -1, the number representation began with a decimal
point but contained no value digits. VALUE or
IVALUE, whichever is appropriate, is returned
containing zero (providing that ITRAIL is not in
the range -13 to -7).
= 0, one or more zero value digits were found, but
the number representation contained no value digits
other than zero. The number representation may or
may not have contained either sign or decimal
point.
= greater than zero, LSHIFT is the number of digits
which were specified in the input text buffer,
counting the leftmost non-zero value digit and all
value digits which were specified to its right.
LSHIFT is the number of significant digits which
were evaluated and is independent of the location
of the decimal point or the value of the exponent
or of IEXTRA. IF KONTRL is less than or equal to
zero so that the value is returned as the integer
IVALUE, then those digits which are discarded due
to their being to the right of adjusted decimal
point (after application of IEXTRA and of the
exponent indicated in the number representation)
are still included within the value of LSHIFT.
IVALUE = returned with the value of the evaluated number if
KONTRL is less than or equal to zero. Note that if
KONTRL is less than or equal to zero, then the
original content of IVALUE is always destroyed. In
particular, if KONTRL is less than or equal to zero
and if KIND is returned not equal to 3 then IVALUE
will be zeroed.
VALUE = returned with the value of the evaluated number if
KONTRL is greater than zero. Note that if KONTRL
is greater than zero, then the original content of
VALUE is always destroyed. In particular, if
KONTRL is greater than zero and if KIND is returned
not equal to 3 then VALUE will be zeroed.
FASP, FORTRAN Alphameric Subroutine Package Page 113
DAHEFT, Free Format Numeric Input Routine
An Example of the Use of DAHEFT
-- ------- -- --- --- -- ------
The following sample subroutine reads a requested number of
real values. Additional lines are read until the requested
number of values has been obtained. Input lines which do
not contain numbers are ignored. The routine does not
return to the calling program until the requested number of
items has been read.
The arguments of the sample routine are as follow
IUNIT = the number of the input unit from which the values
are to be read.
IWANT = the number of values to be read.
VALUE = the array into which the floating point values are
to be stored.
SUBROUTINE READIN(IUNIT,IWANT,VALUE)
DIMENSION IBUFFR(100),VALUE(IWANT)
IF(IWANT.LE.0)GO TO 4
IHAVE=1
1 READ(IUNIT,2)IBUFFR
2 FORMAT(100A1)
LOWBFR=1
3 CALL DAHEFT(7,1,0,IBUFFR,100,LOWBFR,KIND,ISHIFT,
1JSHIFT,KSHIFT,LSHIFT,IVALUE,VALUE(IHAVE))
IF(KIND.EQ.1)GO TO 1
IF(KIND.EQ.2)LOWBFR=LOWBFR+1
IF(KIND.EQ.3)IHAVE=IHAVE+1
IF(IHAVE.LE.IWANT)GO TO 3
4 RETURN
END
Recognition of Leading Exponents and Changing Default Values
----------- -- ------- --------- --- -------- ------- ------
In addition to its function of selecting whether trailing
percent signs and the letters K, M and E are to be
recognized, ITRAIL can enable the recognition of exponents
which are not preceded by value specifications, and can
change the default value used when no value is specified to
be one rather than zero. ITRAIL values of -1, 0 and 1 have
been described previously, and enable the recognition of
trailing E's, of no exponents whatever, and of trailing
exponents of all sorts, respectively. ITRAIL values of -3,
-2, 2 and 3 as described below allow exponents to be
recognized even if not preceded by a sign, decimal point or
digit. The value zero is always returned by this routine if
no value digits are encountered and ITRAIL is in the range
FASP, FORTRAN Alphameric Subroutine Package Page 114
DAHEFT, Free Format Numeric Input Routine
-3 through 3 (or in the range 7 through 13). ITRAIL values
of -2 and 2 will allow the evaluation of either a number
possibly containing a trailing exponent, or of an exponent
appearing by itself without a leading value representation.
ITRAIL values -3 and 3 accept only an exponent without
leading value representation. If ITRAIL has either of the
values -3 or 3 and a leading sign or digit or decimal point
is encountered, then KIND is returned containing 2 to
indicate an unknown character.
ITRAIL = -3, only exponents specified as the letter E
followed by a possibly signed number are to be
recognized. Leading sign, digits or decimal point
are not allowed. The value zero is always
returned.
= -2, same as ITRAIL=-1, except that in addition,
leading exponents specified as the letter E
followed by possible signed numbers are allowed.
The value zero is returned if the letter E is found
not preceded by any value digits.
= 2, same as ITRAIL=1, except that in addition
leading exponents specified by percent signs or the
letters K or M, and leading exponents specified as
the letter E followed by a possibly signed number
are allowed. The value zero is returned if the
percent sign or the letters K, M or E are not
preceded by any value digits.
= 3, only exponents specified as the percent sign or
the letters K or M, or specified as the letter E
followed by a possible signed number are to be
recognized. Leading sign, digits or decimal point
are not allowed. The value zero is always
returned.
ITRAIL values of -13 through -7 evaluate the same sequences
of characters as ITRAIL values of -3 through 3 respectively,
except that if no value digits are found for ITRAIL values
of -13 through -7, then the value is assumed to be one
rather than zero before this value is negated and/or shifted
by the amount indicated by the exponent if necessary, and
except that if the letter E is found but not followed by any
digits, then the single exponent digit 1 is similarly
assumed. For ITRAIL values of -9, -8 or -7 (1-10, 2-10 and
3-10 respectively), the number representation -K would be
equivalent to -1K having the value -1000, and -E would be
the same as -1E1 having the value -10.
ITRAIL values of 7 through 13 evaluate the same sequences of
characters as ITRAIL values of -3 through 3 respectively,
except that the ITRAIL values of 7 through 13 return the
evaluated number as though neither an exponent nor the
decimal point had been included, although the location of
the decimal point and the value of the exponent are
identified to the calling program. The returned value will
FASP, FORTRAN Alphameric Subroutine Package Page 115
DAHEFT, Free Format Numeric Input Routine
have been multiplied by the radix raised to the power
indicated by IEXTRA. ITRAIL being set to 9 (-1+10) would,
for example, allow an E exponent in the number, but would
return the value as though neither exponent nor decimal
point had been found. Since neither ITRAIL values of -3 nor
3 allow a leading number anyway, ITRAIL=-3 is equivalent to
ITRAIL=7, and ITRAIL=3 is equivalent to ITRAIL=13. If any
of the values of 7 through 13 are used for ITRAIL, then the
value indicated by the full combination of value digits,
decimal point and/or exponent can be obtained as either
VALUE*radix**KSHIFT or IVALUE*radix**KSHIFT (where the radix
is 8 if KONTRL is -1 and 10 otherwise, and where **
indicates exponentiation), whichever is appropriate. The
value indicated by the digits and/or decimal point but
ignoring the exponent can similarly be obtained as either
VALUE*radix**(KSHIFT-JSHIFT) or
IVALUE*radix**(KSHIFT-JSHIFT).
The table below presents the values which would be returned
when the number representations shown along the top are
evaluated using the ITRAIL values appearing along the left
edge. Blank entries in the table indicate that some portion
of the representation could not be evaluated using the
corresponding value of ITRAIL. For example, ITRAIL=0 does
not allow the recognition of any exponents so, although the
number representation 1.23K would be evaluated for ITRAIL=0
as the number 1.23 followed by the unknown character K, the
entry for this is left out of the table.
1.23 12.3 1.23K 1.23E3 1.23E 1.23E- E2 E E- -E -E-
-------------------------------------------------------
-13 100 10 .1
-12 1.23 12.3 1230 12.3 .123 100 10 .1 -10 -.1
-11 1.23 12.3 1230 12.3 .123 -10 -.1
-10 1.23 12.3
-9 1.23 12.3 1230 1230 12.3 .123 -10 -.1
-8 1.23 12.3 1230 1230 12.3 .123 100 10 .1 -10 -.1
-7 100 10 .1
-------------------------------------------------------
-3 0 0 0
-2 1.23 12.3 1230 1.23 1.23 0 0 0 0 0
-1 1.23 12.3 1230 1.23 1.23 0 0
0 1.23 12.3
1 1.23 12.3 1230 1230 1.23 1.23 0 0
2 1.23 12.3 1230 1230 1.23 1.23 0 0 0 0 0
3 0 0 0
-------------------------------------------------------
7 0 0 0
8 123 123 123 123 123 0 0 0 0 0
9 123 123 123 123 123 0 0
10 123 123
11 123 123 123 123 123 123 0 0
12 123 123 123 123 123 123 0 0 0 0 0
13 0 0 0
FASP, FORTRAN Alphameric Subroutine Package Page 116
DAHEFT, Free Format Numeric Input Routine
DAMISS, Extends DAHEFT to Allow Comments and Missing Numbers
------ ------- ------ -- ----- -------- --- ------- -------
If the first printing character found by DAHEFT in the line
of input text is not a digit, minus sign, plus sign or
period (and cannot begin an exponent specification if
ITRAIL= has a value other than -11, -10, -9, -1, 0, 1, 9, 10
or 11), then the character is considered to be unknown
causing DAHEFT to return control to the calling program. In
particular, DAHEFT does not recognize the punctuation
characters allowed by many of the other routines in the
FORTRAN Alphameric Subroutine Package. DAMISS, a short
subroutine which calls upon DAHEFT for numeric evaluation,
allows commas between numbers, identifies missing items
indicated by extra commas, skips over any text which is to
the right of either an exclamation point or an ampersand,
and reports any semicolons found in the text being
evaluated. Although DAMISS checks the character to the
right of the number to insure that this character is one of
the allowed punctuation marks or else is either a space or a
tab character, the finding of some other character to the
right of the number might not actually be an error since the
number might be evaluated again by a routine such as DASPAN
or DATREK which allow additional characters to continue the
representation.
The argument list of routine DAMISS is
SUBROUTINE DAMISS(KONTRL,ITRAIL,IEXTRA,IBUFFR,MAXBFR,
1 LOWBFR,KIND ,ISHIFT,JSHIFT,KSHIFT,LSHIFT,IVALUE,
2 VALUE ,MANY ,LCNBFR,LCNERR)
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The argument lists of DAMISS and DAHEFT are identical except
for the DAMISS arguments MANY, LCNBFR and LCNERR which do
not appear in the DAHEFT argument list, and except that
DAMISS can return the argument named KIND containing the
additional values 4, 5, 6 and 7. The argument named MANY
must be set to zero by the calling program before calling
either this routine or any of the other routines in the FASP
package (such as DANEXT, DASPAN and DALOSS) which define
this argument in a similar manner. The arguments named
KIND, LCNBFR and LCNERR are used only for output to the
calling program and their input values are ignored. These
arguments are described below. The documentation of DAHEFT
should be consulted for descriptions of the remaining
arguments.
KIND = 1, nothing, except perhaps a comment indicated by a
leading exclamation point, was found at or to the
right of IBUFFR(LOWBFR). The calling program
FASP, FORTRAN Alphameric Subroutine Package Page 117
DAHEFT, Free Format Numeric Input Routine
should read a new line into the IBUFFR array before
again calling this routine if additional values are
required. LOWBFR is returned pointing beyond the
end of the buffer. MANY is returned set to zero.
IVALUE and VALUE are returned undefined.
= 2, the first printing character (other than a
possible comma if MANY was input greater than zero)
in or to right of IBUFFR(LOWBFR) was not a
character which could begin the representation of a
number and was not a comma, semicolon, ampersand or
exclamation point. LOWBFR is returned pointing to
this printing character. It is expected that the
calling program will otherwise process this
printing character since DAMISS would return the
same results if called again with the same value of
LOWBFR and with the same buffer contents. MANY is
returned containing one plus its input absolute
value. IVALUE and VALUE are returned undefined.
= 3, a number was found which extended through the
end of the buffer or which was followed by a space,
a tab character, a semicolon, an exclamation point
or an ampersand. The value of the number is
returned in either VALUE or IVALUE, whichever is
appropriate, and ISHIFT, JSHIFT, KSHIFT and LSHIFT
are returned describing the representation of the
number. MANY is returned containing one plus its
input absolute value. LOWBFR is returned pointing
to the character to the right of the number
representation.
= 4, a number was followed by a printing character
other than a comma, a semicolon, an exclamation
point or an ampersand. The value of the number is
returned in either VALUE or IVALUE, whichever is
appropriate, and ISHIFT, JSHIFT, KSHIFT and LSHIFT
are returned describing the representation of the
number. LCNBFR is returned pointing in the buffer
to the first character of the number. LOWBFR is
returned pointing in the buffer to the character to
the right of the number. LCNERR is returned
pointing in the buffer to the next space, tab
character, comma, semicolon, exclamation point or
ampersand to the right of the number, or is
returned pointing beyond the end of the buffer if
no space, tab character, comma, semicolon,
exclamation point or ampersand is found to the
right of the number. MANY is returned containing
one plus its input absolute value.
KIND will be returned set to 4 if one number
immediately follows another as in 18-36 (18
followed by -36) or in 12.34.56 (which might be
interpreted as 12.34 followed by .56). If
immediately adjacent numbers are to be allowed,
then KIND=4 should be considered to be equivalent
FASP, FORTRAN Alphameric Subroutine Package Page 118
DAHEFT, Free Format Numeric Input Routine
to KIND=3. If the character in IBUFFR(LOWBFR),
where LOWBFR is the returned not input value, does
not itself start a number, then the subsequent call
to DAMISS will return KIND=2.
= 5, a semicolon was found as the first printing
character at or to the right of IBUFFR(LOWBFR).
LOWBFR is returned pointing to the next character
beyond the semicolon. It is assumed that the
calling program will treat the appearance of the
semicolon as marking the end of a statement. MANY
is returned set to zero. IVALUE and VALUE are
returned undefined.
= 6, an ampersand was found as the first printing
character at or to the right of LOWBFR. The text
to the right of the ampersand is taken as a comment
so LOWBFR is returned pointing beyond the right end
of the buffer. It is assumed that the calling
program will read in the contents of a new buffer,
then again request a new number evaluation from
this routine. The value of MANY must not be
changed by the calling program prior to this
following call. The effect is not quite the same
as if the user had typed all of the text on a
single line since a single number cannot be split
across a line boundary. IVALUE and VALUE are
returned undefined.
= 7, a number was not found, but an extra comma was
found indicating a missing number. MANY is
returned containing one plus its input absolute
value. IVALUE or VALUE, whichever is appropriate,
is returned set to zero. ISHIFT, JSHIFT, KSHIFT
and LSHIFT are returned undefined.
MANY = should be input containing zero each time this
routine is called to begin processing of a new
logical section of text, as for example when
beginning processing of a line of text not tied to
the previous line by an ampersand at the end of the
previous line, or when processing the text to the
right of a semicolon. The initial zeroing of this
argument must be done by the calling program, but
thereafter the value returned by the previous call
to this routine can usually be used. MANY is
returned set to zero each time a semicolon (KIND=5)
is found, and each time an end of line not tied to
the following line by an ampersand (KIND=1) is
found. MANY is returned containing one plus its
input absolute value each time a number is found,
each time an unknown character is found, or each
time an indication of a missing number is found.
KIND is returned containing the value 6 and MANY is
returned containing the negative of the number of
items found if the next printing character
following a comma is an ampersand. MANY should not
FASP, FORTRAN Alphameric Subroutine Package Page 119
DAHEFT, Free Format Numeric Input Routine
be changed by the calling program if an ampersand
(KIND being returned=6) is found indicating that
the subsequent call to this routine is to process
text which is to be treated as though it appeared
in place of the ampersand and the characters to its
right. The effect is not quite the same as if the
user had typed all of the text on a single line
since a single number cannot be split across the
line boundary.
If MANY is input containing zero, then an initial
comma in the input text buffer is taken to indicate
an initial missing item, and MANY is then returned
containing 1. If MANY is input greater than zero,
then an initial comma is ignored if followed by a
number. If MANY is input greater than zero, then
an initial comma followed by no other printing
characters, by a semicolon, or by an exclamation
point indicates a missing item. If MANY is input
greater than zero, then an initial comma followed
by an ampersand will cause the remaining characters
in the buffer to be ignored, and MANY will be
returned containing the negative of its input
value. If MANY is input negative, then it is
assumed that the contents of the current buffer
continue a previous line which terminated with a
comma followed by an ampersand, and MANY is
returned greater than zero.
LCNBFR = if a number representation is found, KIND being
returned containing either 3 or 4, then LCNBFR is
returned containing the subscript of the IBUFFR
array location which contains the first (leftmost)
character of the number representation.
If the first printing character at or to the right
of IBUFFR(LOWBFR) cannot begin a number and is not
one of the allowed punctuation marks so that KIND
is returned containing 2, then LCNBFR and LOWBFR
are both returned pointing to the unexpected
character and LCNERR is returned pointing to the
following space, tab character, comma, semicolon,
exclamation point or ampersand.
LCNERR = if a number representation is found, KIND being
returned containing either 3 or 4, then LCNERR is
returned containing the subscript of the IBUFFR
array location which contains the next space, tab
character, comma, semicolon, exclamation point or
ampersand to the right of the number representation
or else LCNERR is returned containing MAXBFR+1 if
none of these characters appear anywhere to the
right of the number representation. If KIND is
returned containing 3, then LCNERR and LOWBFR are
FASP, FORTRAN Alphameric Subroutine Package Page 120
DAHEFT, Free Format Numeric Input Routine
both returned pointing to the character to the
immediate right of the number representation. If
KIND is returned containing 4, then LOWBFR is
returned pointing to the unexpected printing
character to the immediate right of the number
representation.
If KIND is returned containing 2, indicating that a
printing character which could not begin a number
and which was not one of the allowed punctuation
marks was found, then LCNERR is returned pointing
to the next space, tab character, comma, semicolon,
exclamation point or ampersand, or else LCNERR is
returned containing MAXBFR+1 if none of these
characters appear anywhere to the right of the
initial unexpected character.
Demonstration Program to Interactively Test DAHEFT
------------- ------- -- ------------- ---- ------
The program listed on the following pages accepts a line of
text from the user, then reports each value encountered in
the text for all requested values of ITRAIL, together with
the portion of the text which has just been evaluated. The
values of ITRAIL which give identical results are listed
together in parentheses to the right of the returned value
or to the right of the EMPTY (KIND=1) or UNKNOWN (KIND=2)
condition descriptions. If a number has been evaluated
(KIND=3), then the values of ISHIFT, JSHIFT, KSHIFT and
LSHIFT are also reported to the right of the ITRAIL values.
The typing of a completely empty line allows the the
respecification of whether DAHEFT or DAIHFT is being tested,
of the values of the arguments KONTRL and IEXTRA, and of
another group of values of ITRAIL to be tested.
C PROGRAM TO DEMONSTRATE DAHEFT AND DAIHFT ROUTINES
DIMENSION IBUFFR(58),JBUFFR(72),LOCATE(20),NEEDED(20),
1 KNDSAV(20),ISHSAV(20),JSHSAV(20),KSHSAV(20),
2 LSHSAV(20),VALSAV(20),IVASAV(20), IDENT(20),
3 ITEST(20),IEMPTY( 5),IUNKNO( 8)
DATA ILEFT,IRIGHT,ILESS,IGREAT,IYES,IBLANK/
11H(,1H),1H<,1H>,1HY,1H /
DATA IEMPTY/1HE,1HM,1HP,1HT,1HY/
DATA IUNKNO/1HU,1HN,1HK,1HN,1HO,1HW,1HN,1H /
DATA ITTY,JTTY/5,5/
C
C MAXBFR = NUMBER OF LOCATIONS IN IBUFFR
C MAXSHO = NUMBER OF LOCATIONS IN JBUFFR
C LIMIT = MAXIMUM NUMBER OF ITRAIL VALUES TO TEST
DATA MAXBFR,MAXSHO,LIMIT/58,72,20/
C
C GET VALUES OF KONTRL AND IEXTRA
1 WRITE(JTTY,2)
FASP, FORTRAN Alphameric Subroutine Package Page 121
DAHEFT, Free Format Numeric Input Routine
2 FORMAT(22H TEST DAIHFT (Y OR N) ,$)
READ(ITTY,3)IANS
3 FORMAT(1A1)
WRITE(JTTY,4)
4 FORMAT(15H KONTRL,IEXTRA ,$)
READ(ITTY,5)KONTRL,IEXTRA
5 FORMAT(3I)
IWHICH=KONTRL
IF(IANS.EQ.IYES)IWHICH=0
WRITE(JTTY,6)
6 FORMAT(26H TEST WHICH ITRAIL VALUES ,$)
READ(ITTY,7)ITEST
7 FORMAT(20I)
MAXTRL=LIMIT
8 IF(ITEST(MAXTRL).NE.0)GO TO 10
MAXTRL=MAXTRL-1
IF(MAXTRL.GT.0)GO TO 8
GO TO 1
C
C GET NEXT TEXT BUFFER TO BE EVALUATED
9 IF(INITAL.EQ.1)GO TO 1
10 WRITE(JTTY,11)
11 FORMAT(2H *,$)
READ(ITTY,12)IBUFFR
12 FORMAT(72A1)
DO 13 I=1,MAXTRL
13 LOCATE(I)=1
INITAL=1
GO TO 19
C
C GET NEXT ITEM IN TEXT BUFFER AND STORE DESCRIPTION
14 IF(LOWBFR.GT.MAXBFR)GO TO 9
INITAL=LOWBFR
15 ITRAIL=ITEST(JTRAIL)
IF(IANS.EQ.IYES)GO TO 16
CALL DAHEFT(KONTRL,ITRAIL,IEXTRA,IBUFFR,MAXBFR,
1LOWBFR,KIND,ISHIFT,JSHIFT,KSHIFT,LSHIFT,IVALUE,
2VALUE)
GO TO 17
16 CALL DAIHFT(KONTRL,ITRAIL,IEXTRA,IBUFFR,MAXBFR,
1LOWBFR,KIND,ISHIFT,JSHIFT,KSHIFT,LSHIFT,IVALUE)
17 IF(KIND.EQ.2)LOWBFR=LOWBFR+1
LOCATE(JTRAIL)=LOWBFR
NEEDED(JTRAIL)=1
KNDSAV(JTRAIL)=KIND
IF(KIND.NE.3)GO TO 19
ISHSAV(JTRAIL)=ISHIFT
JSHSAV(JTRAIL)=JSHIFT
KSHSAV(JTRAIL)=KSHIFT
LSHSAV(JTRAIL)=LSHIFT
IF(IWHICH.LE.0)GO TO 18
VALSAV(JTRAIL)=VALUE
GO TO 19
18 IVASAV(JTRAIL)=IVALUE
FASP, FORTRAN Alphameric Subroutine Package Page 122
DAHEFT, Free Format Numeric Input Routine
C
C SEARCH FOR NEXT VALUE OF ITRAIL TO BE USED
19 LOWBFR=MAXBFR+1
DO 20 I=1,MAXTRL
IF(LOWBFR.LE.LOCATE(I))GO TO 20
LOWBFR=LOCATE(I)
JTRAIL=I
20 CONTINUE
IF(INITAL.EQ.LOWBFR)GO TO 15
C
C REPORT RESULTS IF ALL VALUES OF ITRAIL DONE
ISEEN=INITAL-1
21 LEAST=MAXBFR+1
DO 22 I=1,MAXTRL
IF(NEEDED(I).EQ.0)GO TO 22
IF(LEAST.LT.LOCATE(I))GO TO 22
LEAST=LOCATE(I)-1
INDEX=I
22 CONTINUE
IF(LEAST.GT.MAXBFR)GO TO 14
NEEDED(INDEX)=0
ISAME=1
IDENT(1)=ITEST(INDEX)
DO 25 I=INDEX,MAXTRL
IF(NEEDED(I).EQ.0)GO TO 25
IF(KNDSAV(INDEX).NE.KNDSAV(I))GO TO 25
IF(KNDSAV(INDEX).NE.3)GO TO 24
IF(ISHSAV(INDEX).NE.ISHSAV(I))GO TO 25
IF(JSHSAV(INDEX).NE.JSHSAV(I))GO TO 25
IF(KSHSAV(INDEX).NE.KSHSAV(I))GO TO 25
IF(LSHSAV(INDEX).NE.LSHSAV(I))GO TO 25
IF(IWHICH.LE.0)GO TO 23
IF(VALSAV(INDEX).NE.VALSAV(I))GO TO 25
GO TO 24
23 IF(IVASAV(INDEX).NE.IVASAV(I))GO TO 25
24 ISAME=ISAME+1
IDENT(ISAME)=ITEST(I)
NEEDED(I)=0
25 CONTINUE
C
C DISPLAY EVALUATED TEXT
IF(ISEEN.EQ.LEAST)GO TO 30
NOWSHO=0
26 NOWSHO=NOWSHO+1
IF(NOWSHO.GE.INITAL)GO TO 27
JBUFFR(NOWSHO)=IBLANK
GO TO 26
27 JBUFFR(NOWSHO)=ILESS
DO 28 I=INITAL,LEAST
NOWSHO=NOWSHO+1
28 JBUFFR(NOWSHO)=IBUFFR(I)
NOWSHO=NOWSHO+1
JBUFFR(NOWSHO)=IGREAT
WRITE(JTTY,29)(JBUFFR(I),I=1,NOWSHO)
FASP, FORTRAN Alphameric Subroutine Package Page 123
DAHEFT, Free Format Numeric Input Routine
29 FORMAT(1X,100A1)
ISEEN=LEAST
C
C REPORT ALL VALUES OF ITRAIL WHICH GIVE SAME RESULTS
30 NOWSHO=0
KIND=KNDSAV(INDEX)
GO TO(31,33,35),KIND
31 DO 32 I=1,5
NOWSHO=NOWSHO+1
32 JBUFFR(NOWSHO)=IEMPTY(I)
GO TO 36
33 DO 34 I=1,8
NOWSHO=NOWSHO+1
34 JBUFFR(NOWSHO)=IUNKNO(I)
NOWSHO=NOWSHO+1
I=LOCATE(INDEX)-1
JBUFFR(NOWSHO)=IBUFFR(I)
GO TO 36
35 CALL DASHOW(IWHICH,0,6,6,6,5,
1IVASAV(INDEX),VALSAV(INDEX),MAXSHO,NOWSHO,JBUFFR,IERR)
36 NOWSHO=NOWSHO+1
IF(NOWSHO.GE.12)GO TO 37
JBUFFR(NOWSHO)=IBLANK
GO TO 36
37 JBUFFR(NOWSHO)=ILEFT
J=NOWSHO
DO 38 I=1,ISAME
CALL DANUMB(0,IDENT(I),10,JBUFFR,NOWSHO,J,MAXSHO)
38 CONTINUE
NOWSHO=NOWSHO+1
JBUFFR(NOWSHO)=IRIGHT
IF(KIND.NE.3)GO TO 39
CALL DANUMB(0,ISHSAV(INDEX),10,JBUFFR,NOWSHO,0,MAXSHO)
CALL DANUMB(0,JSHSAV(INDEX),10,JBUFFR,NOWSHO,0,MAXSHO)
CALL DANUMB(0,KSHSAV(INDEX),10,JBUFFR,NOWSHO,0,MAXSHO)
CALL DANUMB(0,LSHSAV(INDEX),10,JBUFFR,NOWSHO,0,MAXSHO)
39 WRITE(JTTY,40)(JBUFFR(I),I=1,NOWSHO)
40 FORMAT(2X,72A1)
GO TO 21
END
Typical Dialog Between DAHEFT Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
TEST DAIHFT (Y OR N) N
KONTRL,IEXTRA 1
TEST WHICH ITRAIL VALUES -11 -10 -9 -2 -1 0 1 2 9 10 11
*12.34E-4 .1234 12.34 1234 12.34E4
<12.34>
12.34 (-10 0) 0 0 -2 4
1234 (10) 0 0 -2 4
<12.34E-4>
1.234E-3 (-11 -9 -2 -1 1 2) -1 -4 -6 4
FASP, FORTRAN Alphameric Subroutine Package Page 124
DAHEFT, Free Format Numeric Input Routine
1234 (9 11) -1 -4 -6 4
<E>
UNKNOWN E (-10 0 10)
<-4>
-4 (-10 0 10) 0 0 0 1
< .1234>
.1234 (-11 -10 -9 -2 -1 0 1 2) 0 0 -4 4
1234 (9 10 11) 0 0 -4 4
< 12.34>
12.34 (-11 -10 -9 -2 -1 0 1 2) 0 0 -2 4
1234 (9 10 11) 0 0 -2 4
< 1234>
1234 (-11 -10 -9 -2 -1 0 1 2 9 10 11) 0 0 0 4
< 12.34>
12.34 (-10 0) 0 0 -2 4
1234 (10) 0 0 -2 4
< 12.34E4>
123400 (-11 -9 -2 -1 1 2) -1 4 2 4
1234 (9 11) -1 4 2 4
<E>
UNKNOWN E (-10 0 10)
<4>
4 (-10 0 10) 0 0 0 1
< >
EMPTY (-11 -10 -9 -2 -1 0 1 2 9 10 11)
*-18.3K -18.3M -18.3%
<-18.3>
-18.3 (-11 -10 -2 -1 0) 0 0 -1 3
-183 (9 10) 0 0 -1 3
<-18.3K>
-18300 (-9 1 2) 2 3 2 3
-183 (11) 2 3 2 3
<K>
UNKNOWN K (-11 -10 -2 -1 0 9 10)
< -18.3>
-18.3 (-11 -10 -2 -1 0) 0 0 -1 3
-183 (9 10) 0 0 -1 3
< -18.3M>
-1.83E7 (-9 1 2) 3 6 5 3
-183 (11) 3 6 5 3
<M>
UNKNOWN M (-11 -10 -2 -1 0 9 10)
< -18.3>
-18.3 (-11 -10 -2 -1 0) 0 0 -1 3
-183 (9 10) 0 0 -1 3
< -18.3%>
-.183 (-9 1 2) 1 -2 -3 3
-183 (11) 1 -2 -3 3
<%>
UNKNOWN % (-11 -10 -2 -1 0 9 10)
< >
EMPTY (-11 -10 -9 -2 -1 0 1 2 9 10 11)
*
< >
FASP, FORTRAN Alphameric Subroutine Package Page 125
DAHEFT, Free Format Numeric Input Routine
EMPTY (-11 -10 -9 -2 -1 0 1 2 9 10 11)
TEST DAIHFT (Y OR N) N
KONTRL,IEXTRA 6
TEST WHICH ITRAIL VALUES -11 -10 -9 -2 -1 0 1 2 9 10 11
*12.34E-4 .1234 12.34 1234 12.34E4
<12.34>
12.34 (-10 0) 0 0 -2 4
1234 (10) 0 0 -2 4
<12.34E-4>
1.234E-3 (-11 -9 -2 -1 1 2) -1 -4 -6 4
1234 (9 11) -1 -4 -6 4
<E>
UNKNOWN E (-10 0 10)
<-4>
-4 (-10 0 10) 0 0 0 1
< .1234>
.1234 (-11 -10 -9 -2 -1 0 1 2) 0 0 -4 4
1234 (9 10 11) 0 0 -4 4
< 12.34>
12.34 (-11 -10 -9 -2 -1 0 1 2) 0 0 -2 4
1234 (9 10 11) 0 0 -2 4
< 1234>
1234 (-11 -10 -9 -2 -1 0 1 2 9 10 11) 0 0 0 4
< 12.34>
12.34 (-10 0) 0 0 -2 4
1234 (10) 0 0 -2 4
< 12.34E4>
123400 (-11 -9 -2 -1 1 2) -1 4 2 4
1234 (9 11) -1 4 2 4
<E>
UNKNOWN E (-10 0 10)
<4>
4 (-10 0 10) 0 0 0 1
< >
EMPTY (-11 -10 -9 -2 -1 0 1 2 9 10 11)
*-18.3K -18.3M -18.3%
<-18.3>
-18.3 (-11 -10 -2 -1 0) 0 0 -1 3
-183 (9 10) 0 0 -1 3
<-18.3K>
-18300 (-9 1 2) 2 3 2 3
-183 (11) 2 3 2 3
<K>
UNKNOWN K (-11 -10 -2 -1 0 9 10)
< -18.3>
-18.3 (-11 -10 -2 -1 0) 0 0 -1 3
-183 (9 10) 0 0 -1 3
< -18.3M>
-1.83E7 (-9 1 2) 3 6 5 3
-183 (11) 3 6 5 3
<M>
UNKNOWN M (-11 -10 -2 -1 0 9 10)
< -18.3>
-18.3 (-11 -10 -2 -1 0) 0 0 -1 3
FASP, FORTRAN Alphameric Subroutine Package Page 126
DAHEFT, Free Format Numeric Input Routine
-183 (9 10) 0 0 -1 3
< -18.3%>
-.183 (-9 1 2) 1 -2 -3 3
-183 (11) 1 -2 -3 3
<%>
UNKNOWN % (-11 -10 -2 -1 0 9 10)
< >
EMPTY (-11 -10 -9 -2 -1 0 1 2 9 10 11)
*
< >
EMPTY (-11 -10 -9 -2 -1 0 1 2 9 10 11)
TEST DAIHFT (Y OR N) N
KONTRL,IEXTRA 0
TEST WHICH ITRAIL VALUES -11 -10 -9 -2 -1 0 1 2 9 10 11
*12.34E-4 .1234 12.34 1234 12.34E4
<12.34>
12 (-10 0) 0 0 -2 4
1234 (10) 0 0 -2 4
<12.34E-4>
0 (-11 -9 -2 -1 1 2) -1 -4 -6 4
1234 (9 11) -1 -4 -6 4
<E>
UNKNOWN E (-10 0 10)
<-4>
-4 (-10 0 10) 0 0 0 1
< .1234>
0 (-11 -10 -9 -2 -1 0 1 2) 0 0 -4 4
1234 (9 10 11) 0 0 -4 4
< 12.34>
12 (-11 -10 -9 -2 -1 0 1 2) 0 0 -2 4
1234 (9 10 11) 0 0 -2 4
< 1234>
1234 (-11 -10 -9 -2 -1 0 1 2 9 10 11) 0 0 0 4
< 12.34>
12 (-10 0) 0 0 -2 4
1234 (10) 0 0 -2 4
< 12.34E4>
123400 (-11 -9 -2 -1 1 2) -1 4 2 4
1234 (9 11) -1 4 2 4
<E>
UNKNOWN E (-10 0 10)
<4>
4 (-10 0 10) 0 0 0 1
< >
EMPTY (-11 -10 -9 -2 -1 0 1 2 9 10 11)
*-18.3K -18.3M -18.3%
<-18.3>
-18 (-11 -10 -2 -1 0) 0 0 -1 3
-183 (9 10) 0 0 -1 3
<-18.3K>
-18300 (-9 1 2) 2 3 2 3
-183 (11) 2 3 2 3
<K>
UNKNOWN K (-11 -10 -2 -1 0 9 10)
FASP, FORTRAN Alphameric Subroutine Package Page 127
DAHEFT, Free Format Numeric Input Routine
< -18.3>
-18 (-11 -10 -2 -1 0) 0 0 -1 3
-183 (9 10) 0 0 -1 3
< -18.3M>
-18300000 (-9 1 2) 3 6 5 3
-183 (11) 3 6 5 3
<M>
UNKNOWN M (-11 -10 -2 -1 0 9 10)
< -18.3>
-18 (-11 -10 -2 -1 0) 0 0 -1 3
-183 (9 10) 0 0 -1 3
< -18.3%>
0 (-9 1 2) 1 -2 -3 3
-183 (11) 1 -2 -3 3
<%>
UNKNOWN % (-11 -10 -2 -1 0 9 10)
< >
EMPTY (-11 -10 -9 -2 -1 0 1 2 9 10 11)
*
< >
EMPTY (-11 -10 -9 -2 -1 0 1 2 9 10 11)
TEST DAIHFT (Y OR N) N
KONTRL,IEXTRA 1
TEST WHICH ITRAIL VALUES -8 2 12
*3.2 3.2E 3.2E+ 3.2E-
<3.2>
3.2 (-8 2) 0 0 -1 2
32 (12) 0 0 -1 2
< 3.2E>
32 (-8) -4 1 0 2
3.2 (2) -4 0 -1 2
32 (12) -4 0 -1 2
< 3.2E+>
32 (-8) -2 1 0 2
3.2 (2) -2 0 -1 2
32 (12) -2 0 -1 2
< 3.2E->
.32 (-8) -3 -1 -2 2
3.2 (2) -3 0 -1 2
32 (12) -3 0 -1 2
< >
EMPTY (-8 2 12)
*E3 -E3 +E3 .E3 -.E3 +.E3 K -K +K .K -.K +.K
<E3>
1000 (-8) -1 3 3 -4
0 (2 12) -1 3 3 -4
< -E3>
-1000 (-8) -1 3 3 -3
0 (2 12) -1 3 3 -3
< +E3>
1000 (-8) -1 3 3 -2
0 (2 12) -1 3 3 -2
< .E3>
1000 (-8) -1 3 3 -1
FASP, FORTRAN Alphameric Subroutine Package Page 128
DAHEFT, Free Format Numeric Input Routine
0 (2 12) -1 3 3 -1
< -.E3>
-1000 (-8) -1 3 3 -3
0 (2 12) -1 3 3 -3
< +.E3>
1000 (-8) -1 3 3 -2
0 (2 12) -1 3 3 -2
< K>
1000 (-8) 2 3 3 -4
0 (2 12) 2 3 3 -4
< -K>
-1000 (-8) 2 3 3 -3
0 (2 12) 2 3 3 -3
< +K>
1000 (-8) 2 3 3 -2
0 (2 12) 2 3 3 -2
< .K>
1000 (-8) 2 3 3 -1
0 (2 12) 2 3 3 -1
< -.K>
-1000 (-8) 2 3 3 -3
0 (2 12) 2 3 3 -3
< +.K>
1000 (-8) 2 3 3 -2
0 (2 12) 2 3 3 -2
< >
EMPTY (-8 2 12)
*- + . -. +. 0 -0 +0 0. .0 0.00 0.00E-4 0.00E4
<->
-1 (-8) 0 0 0 -3
0 (2 12) 0 0 0 -3
< +>
1 (-8) 0 0 0 -2
0 (2 12) 0 0 0 -2
< .>
1 (-8) 0 0 0 -1
0 (2 12) 0 0 0 -1
< -.>
-1 (-8) 0 0 0 -3
0 (2 12) 0 0 0 -3
< +.>
1 (-8) 0 0 0 -2
0 (2 12) 0 0 0 -2
< 0>
0 (-8 2 12) 0 0 0 0
< -0>
0 (-8 2 12) 0 0 0 0
< +0>
0 (-8 2 12) 0 0 0 0
< 0.>
0 (-8 2 12) 0 0 0 0
< .0>
0 (-8 2 12) 0 0 -1 0
< 0.00>
FASP, FORTRAN Alphameric Subroutine Package Page 129
DAHEFT, Free Format Numeric Input Routine
0 (-8 2 12) 0 0 -2 0
< 0.00E-4>
0 (-8 2 12) -1 -4 -6 0
< 0.00E4>
0 (-8 2 12) -1 4 2 0
< >
EMPTY (-8 2 12)
* E -E +E E- E+ -E- -E+ +E- +E+
< E>
10 (-8) -4 1 1 -4
0 (2 12) -4 0 0 -4
< -E>
-10 (-8) -4 1 1 -3
0 (2 12) -4 0 0 -3
< +E>
10 (-8) -4 1 1 -2
0 (2 12) -4 0 0 -2
< E->
.1 (-8) -3 -1 -1 -4
0 (2 12) -3 0 0 -4
< E+>
10 (-8) -2 1 1 -4
0 (2 12) -2 0 0 -4
< -E->
-.1 (-8) -3 -1 -1 -3
0 (2 12) -3 0 0 -3
< -E+>
-10 (-8) -2 1 1 -3
0 (2 12) -2 0 0 -3
< +E->
.1 (-8) -3 -1 -1 -2
0 (2 12) -3 0 0 -2
< +E+>
10 (-8) -2 1 1 -2
0 (2 12) -2 0 0 -2
< >
EMPTY (-8 2 12)
FASP, FORTRAN Alphameric Subroutine Package Page 130
DAHELP, Determines Number of Leading Question Marks
DDDDD AAA HH HH EEEEEEEE LL PPPPPP
DD DD AAAA HH HH EE LL PP PP
DD DD AA AA HH HH EE LL PP PP
DD DD AA AA HHHHHHHH EEEEE LL PPPPPP
DD DD AAAAAAA HH HH EE LL PP
DD DD AA AA HH HH EE LL PP
DDDDD AA AA HH HH EEEEEEEE LLLLLLLL PP
DAHELP, Determines Number of Leading Question Marks
------ ---------- ------ -- ------- -------- -----
Many interactive programs interpret the appearance of a
question mark at the start of the text typed by the user to
be a request by the user for information. However, a
question mark appearing within or to the right of an
intelligible user response should probably be treated the
same as any other unknown character. The routines within
the FASP package do not have available to them any
information regarding the location of the text currently
being processed relative to the total text within the
buffer. These routines do not treat the question mark
differently than any other unknown or alphabetic character
since the interpretation of the question mark depends upon
its location relative to the total text within the buffer
rather than upon the location of the question mark relative
to the text forming a particular statement which might be
delimited by semicolons.
Initial question marks can, however, be identified by the
DAHELP routine immediately after a new line of text has been
read by the calling program. DAHELP reports how many
question marks were found at the start of the buffer, and
specifies the buffer location which contains the first
printing character which is not itself a question mark. If
a question mark in not found, then the location of the first
printing character can be supplied to the other routines in
the FASP package as the location in the buffer at which
these routines are to begin interpretation. The calling
program should issue the appropriate informative message if
the first printing character is found to be a question mark,
and then, ignoring the remainder of the current contents of
the buffer, should ask for and accept a new response from
the user, again assuring that this new message does not
itself begin with a question mark before resuming normal
processing. Either the number of question marks found on a
single line, or else the number of consecutive lines found
to start with question marks might be used to select the
length of the informative message.
FASP, FORTRAN Alphameric Subroutine Package Page 131
DAHELP, Determines Number of Leading Question Marks
The DAHELP Argument List
--- ------ -------- ----
The argument list of routine DAHELP is
SUBROUTINE DAHELP(IBUFFR,MAXBFR,LOWBFR,IQUERY)
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The following arguments are used for input and are returned
unchanged.
IBUFFR = input buffer array containing characters typed by
the user, read by a multiple of an A1 format, which
is to be searched for initial question marks.
IBUFFR then contains 1 character per computer
storage location.
MAXBFR = subscript of the final (rightmost) location in the
IBUFFR array which can be searched for initial
question marks.
The following argument must be set by the calling program
before this routine is called, and then is returned by this
routine describing the location of the first printing
character which is not itself a question mark.
LOWBFR = input containing the subscript of the IBUFFR array
location which contains the first (leftmost)
character which must be tested to determine whether
it is a question mark.
= returned containing the subscript of the IBUFFR
array location which contains the leftmost printing
character which is not itself a question mark, or
returned pointing beyond the end of the buffer (set
to MAXBFR+1) if no printing characters other than
question marks were found in the buffer.
The following argument is used for returning the number of
question marks found to the calling program. Its input
value is ignored.
IQUERY = 0, returned if the input buffer contained no
printing characters, or if no question marks were
found before the first printing character which is
not itself a question mark.
= greater than zero, IQUERY is returned containing
the number of question marks which were found
before any other printing characters in the buffer.
These question marks can be preceded by, be
separated by, or be followed by spaces and/or by
tab characters.
FASP, FORTRAN Alphameric Subroutine Package Page 132
DAHEST, Parser for Simple Commands to Interactive Programs
DDDDD AAA HH HH EEEEEEEE SSSSSS TTTTTTTT
DD DD AAAA HH HH EE SS TT
DD DD AA AA HH HH EE SS TT
DD DD AA AA HHHHHHHH EEEEE SSSS TT
DD DD AAAAAAA HH HH EE SS TT
DD DD AA AA HH HH EE SS TT
DDDDD AA AA HH HH EEEEEEEE SSSSSS TT
DAHEST, Parser for Simple Commands to Interactive Programs
------ ------ --- ------ -------- -- ----------- --------
The main routine of an interactive program often merely
interrogates the user concerning the sequence in which
subroutines are to be executed. Providing that any
subroutine will request relatively few items of data from
the user, and that the requests for this data are
predictable by the user, then the commands typed by the user
can be statements containing command words which identify
the subroutines and containing numeric or textual arguments.
DAHEST is a FORTRAN subroutine which can be called by the
main program to identify the command words and to evaluate
the arguments in such statements.
The following statements are typical of those which can be
interpreted by the DAHEST routine. The comments shown to
the right of the statements are optional and would not
normally be used unless DAHEST is processing commands which
have been read from a file, rather than typed by the user.
OPAQUE '=',,'*' !USE = FOR OUTLINE, * FOR ORIGIN
INVISIBLE ,'$' !REST OF OBJECT WILL SHOW
WINDOW 10/30,5/25 !WINDOW COLUMN 10 TO 30, LINE 5 TO 25
EXAMINE !LOOK AT CONTENTS OF WINDOW
GROUP 25,12 !GROUP CONTIGUOUS TO COLUMN 25, LINE 12
EXAMINE !LOOK AT WINDOW OUTLINING GROUP
MOVE +5,-3 !MOVE GROUP +5 COLUMNS, -3 LINES
EXAMINE !LOOK AT GROUP IN NEW LOCATION
If the possible command words have been chosen to start with
different letters of the alphabet, then the above example
could be reduced to the following single line.
O'=',,'*';I,'$';W 10/30,5/25;E;G 25,12;E;M+5,-3;E
or, if unnecessary commas are removed or replaced by spaces
O'=',,'*';I,'$';W 10/30 5/25;E;G 25 12;E;M+5-3;E
Of course, the user probably would not combine statements to
the extent shown above, since the results of one statement
would not be available before the next is issued, causing
the interactive nature of the program to be lost.
FASP, FORTRAN Alphameric Subroutine Package Page 133
DAHEST, Parser for Simple Commands to Interactive Programs
DAHEST interprets an array read by the calling program with
a multiple of an A1 format. The calling program indicates
to DAHEST the position in the buffer array of the leftmost
character which has not yet been evaluated. DAHEST moves
this pointer through the buffer and returns it to the
calling program specifying the leftmost character to be
evaluated by the next call to DAHEST. After DAHEST has
finally indicated that nothing more remains to be processed
in the line of text which is contained in the buffer, then
the calling program must read another line of text, and must
reset the pointer to indicate the start of the new line of
text before DAHEST is called again.
The line of text which is being evaluated can contain
several statements if these statements are separated by
semicolons. Extra semicolons can be used to indicate empty
statements. The text can also contain a comment indicated
by an exclamation point to the left of the comment. When
DAHEST encounters an exclamation point which is not within a
quoted text string (one of the argument types), then the
contents of the buffer to the right of the exclamation point
are not evaluated and the pointer is returned indicating the
character beyond the right end of the buffer. A statement
which is processed by DAHEST cannot be continued across the
end of a line. Although many other routines in FASP take a
rightmost ampersand to indicate that the current statement
is being continued on the following line, DAHEST provides
not special treatment of the ampersand.
A statement starts with a command word which can be
abbreviated providing that the abbreviation is not
ambiguous. If the command word typed by the user is an
exact match of a word in the dictionary supplied by the
calling program, but is also an abbreviation of a longer
word, then the shorter word is assumed to be the word which
was desired. For example, if the dictionary contains both
of the words NO and NONE, then the single letter N is an
ambiguous abbreviation, the letter sequence NO selects the
word NO, and the letter sequence NON is an abbreviation of
the word NONE.
If the word in the dictionary does not contain a space, then
a space or tab character typed by the user following a
command word will be assumed to mark the end of the command
word. If the word in the dictionary does contain a space,
then the portions of the command word to either side of the
space can be abbreviated by the user whether or not the user
has typed one or more spaces and/or tab characters at this
location. For example, if the dictionary contains the word
specification
1HN,1HO,1H ,1HL,1HI,1HM,1HI,1HT
then this word could be selected by any of the following
FASP, FORTRAN Alphameric Subroutine Package Page 134
DAHEST, Parser for Simple Commands to Interactive Programs
letter sequences
N, NL, N L, NLI, N LI, NO, NOL, NO L or NO LI
in the input buffer, providing in each case that the
sequence is not ambiguous.
The arguments which appear to the right of the command word
can be possibly signed numbers, or alphabetic words to be
evaluated by the calling program, or (quoted) text strings
begun and terminated either by the apostrophe character or
else by matched parentheses.
The signed value of a numeric argument and an indication of
its sign, if any, are both returned to the calling program
so that the appearance of a plus sign can, if necessary, be
treated differently than the absence of any sign. For
example, an unsigned number might indicate a location upon a
positive coordinate scale, but a number preceded by a sign
might indicate a relative shift or vector. Numbers can be
separated by slashes or by colons (the 2 characters are
equivalent) if they are to be associated in some manner.
Two such possibly signed numbers separated only by a single
slash or by a single colon might indicate a range. Three
possibly signed numbers separated only by slashes or by
colons might be used to indicate the start of a range, the
increment, and the end of the range.
A word appearing as an argument of a command must begin with
a character which cannot start a number and which is not one
of the delimiter characters such as the space, tab
character, slash, colon, semicolon, exclamation point,
comma, apostrophe or parentheses. Digits can appear
anywhere to the right of the leading character of the word,
but the other prohibited characters will, if encountered,
terminate the word. DAHEST indicates to the calling program
the character locations at which the word begins and ends.
Text strings delimited by the apostrophe character can
contain any legal FORTRAN readable characters. If the text
string is to contain the apostrophe itself, then this is
indicated by 2 successive appearances of the apostrophe
neither of which is taken to be the string terminator.
DAHEST indicates to the calling program the character
locations at which the text inside the delimiting
apostrophes begins and ends. If the text string contains
apostrophes indicated by 2 appearances of the apostrophe,
then DAHEST removes the extra apostrophe from the string and
moves the position of the rightmost pointer 1 character to
the left. If the apostrophe which indicates the right end
of the string is missing, then the text string is assumed to
extend through the rightmost printing character on the line.
The start of a text string can also be indicated by a
leading left or right parenthesis. Starting with the
leading parenthesis, a level count is obtained by adding 1
FASP, FORTRAN Alphameric Subroutine Package Page 135
DAHEST, Parser for Simple Commands to Interactive Programs
for each left parenthesis which is not preceded by an
apostrophe, and by subtracting 1 for each right parenthesis
which is not preceded by an apostrophe. The text string
then continues to the next parenthesis which returns the
parenthetical level count to zero, or through the rightmost
printing character on the line if a closing parenthesis is
not found. The text string includes embedded parentheses if
these do not return the parenthetical level count to zero.
If an apostrophe is encountered within the text string, then
the apostrophe is retained and both the apostrophe and the
character to its right are included in the text string.
Although DAHEST does not reveal to the calling program
whether the text string started to the right of a leading
apostrophe or of a left or a right parenthesis, the calling
program can easily determine which of these three characters
appears to the immediate left of the contents of the text
string. In the application for which DAHEST was written,
text strings begun by apostrophes contain characters to be
handled in the order encountered, strings begun by left
parentheses contain characters to be included in a class,
and strings begun by right parentheses contain characters to
be excluded from a class.
One, but only one, of the three types of arguments, words or
numbers or text strings, can be used more than once as an
argument. If the type of argument which can be repeated is
specified before the first argument is found, either being
the same for all commands, or else being specified
separately in the dictionary for each command, then
arguments of the other two types can appear at most once in
the argument list. If the repeatable type is word or text
string, then a set of numbers indicating a range can still
be supplied. If the repeatable type is numeric, then more
than one set of numbers indicating ranges will be accepted.
Alternatively, the type of argument which can be repeated
can be the type of the first argument encountered, in which
case arguments of the other two types are not allowed in the
argument list.
Any number of spaces and/or tab characters can appear before
the command word and between the command word and its first
argument. Successive arguments can be separated by a single
comma and/or by any number of spaces and/or tab characters.
No separating characters are necessary if the leading
character of an argument indicates that it cannot continue
the preceding command word or preceding argument. A single
comma appearing between 2 arguments of either the same or of
different types merely indicates the separation between the
arguments, and is entirely equivalent to one or more spaces
and/or tab characters. Two commas, possibly separated by
spaces and/or by tab characters, indicate a missing argument
of the repeatable type. A comma appearing immediately
(except for optional spaces and/or tabs) after the command
FASP, FORTRAN Alphameric Subroutine Package Page 136
DAHEST, Parser for Simple Commands to Interactive Programs
word is taken to indicate that the first argument of the
repeatable type is missing. For example, in the statements
OPAQUE'=',,'*';INVISIBLE,'$'
the quoted text character * is the third argument of the
command word OPAQUE, the second argument being missing, and
the quoted text character $ is the second argument of the
command word INVISIBLE, the first argument being missing.
The arguments of the repeatable type are returned to the
calling program in the order in which they appear in the
statement, and, in particular, the calling program is able
to determine whether any are missing. No information
regarding ordering between arguments of different types is
returned to the calling program. If the command word FETCH
takes both an object name and a pair of coordinates as
arguments, then the following statements would all be
equivalent.
FETCH HEXAGON,20,44
FETCH HEXAGON,20 44
FETCH HEXAGON 20,44
FETCH HEXAGON 20 44
FETCH 20,HEXAGON,44
FETCH 20,HEXAGON 44
FETCH 20 HEXAGON,44
FETCH 20 HEXAGON 44
FETCH 20,44,HEXAGON
FETCH 20,44 HEXAGON
FETCH 20 44,HEXAGON
FETCH 20 44 HEXAGON
DAHEST can evaluate numeric arguments either as integers or
as real numbers. The type of numeric evaluation which is to
be performed is associated with the command word at the
start of the statement, and is specified by the dictionary
containing the descriptions of the possible command words.
DAHEST calls the routine DAHEFT to perform the evaluation,
and all numeric variations which are permitted by DAHEFT are
recognized by DAHEST. Any number, whether being evaluated
as an integer or as a real number, can be specified in
floating point form or in exponent form. The characters %,
K and M are accepted in place of the exponents E-2, E3 and
E6 respectively.
If the program which calls DAHEST does not require the
evaluation of real numbers and does not otherwise call
DAHEFT, and if the specification of integers in exponent
form is not necessary, then the routine DAIHST should be
called instead of DAHEST. The 2 routines are of
approximately the same length, but DAIHST does not call
DAHEFT for numeric evaluation. Numbers evaluated by DAIHST
must consist only of digits following the optional sign.
FASP, FORTRAN Alphameric Subroutine Package Page 137
DAHEST, Parser for Simple Commands to Interactive Programs
DAIHST treats the characters ., %, E, K and M the same as
any alphabetic letter.
DAIHST was developed as the command scanner for the MIRAGE
printable image sketchpad program (not yet completed when
this documentation was written). Most of the command
statements shown above are written in the command language
designed for MIRAGE. DAHEST uses the same logic as DAIHST
and was developed from DAIHST since it was felt that other
users would reject the command scanner if it could not
evaluate real numbers.
the DAHEST and DAIHST Argument Lists
--- ------ --- ------ -------- -----
The argument lists of routines DAHEST and DAIHST are
SUBROUTINE DAHEST(KMDTYP,LSTTYP,NAMLOW,NAMMAX,MRKLOW,
1 MRKMAX,NUMLOW,NUMMAX,INTRVL,LOWWRD,MAXWRD,IWORD ,
2 LOWKNT,MAXKNT,KNTLTR,LEGAL ,MAXBFR,IBUFFR,LOWBFR,
3 KIND ,KOMAND,LCNWRD,LCNKNT,INIPRT,MIDPRT,LMTPRT,
4 NAMKNT,NAMLFT,NAMRIT,MRKKNT,MRKLFT,MRKRIT,NUMKNT,
5 NUMSIN,NUMVAL,VALNUM,IFLOAT)
with the associated DIMENSION statement
DIMENSION IWORD (MAXWRD),KNTLTR(MAXKNT),
1LEGAL (MAXKNT),IBUFFR(MAXBFR),NAMLFT(NAMMAX),
2NAMRIT(NAMMAX),MRKLFT(MRKMAX),MRKRIT(MRKMAX),
3NUMSIN(NUMMAX),NUMVAL(NUMMAX),VALNUM(NUMMAX)
and
SUBROUTINE DAIHST(KMDTYP,LSTTYP,NAMLOW,NAMMAX,MRKLOW,
1 MRKMAX,NUMLOW,NUMMAX,INTRVL,LOWWRD,MAXWRD,IWORD ,
2 LOWKNT,MAXKNT,KNTLTR,LEGAL ,MAXBFR,IBUFFR,LOWBFR,
3 KIND ,KOMAND,LCNWRD,LCNKNT,INIPRT,MIDPRT,LMTPRT,
4 NAMKNT,NAMLFT,NAMRIT,MRKKNT,MRKLFT,MRKRIT,NUMKNT,
5 NUMSIN,NUMVAL)
with the associated DIMENSION statement
DIMENSION IWORD (MAXWRD),KNTLTR(MAXKNT),
1LEGAL (MAXKNT),IBUFFR(MAXBFR),NAMLFT(NAMMAX),
2NAMRIT(NAMMAX),MRKLFT(MRKMAX),MRKRIT(MRKMAX),
3NUMSIN(NUMMAX),NUMVAL(NUMMAX)
The ordering of the arguments is the same for both routines,
but the last 2 arguments of DAHEST are not included in the
DAIHST argument list. The argument definitions are also
identical, with the exception that those argument values
which request evaluation of numbers as real numbers for
DAHEST instead produce evaluation as integers for DAIHST.
FASP, FORTRAN Alphameric Subroutine Package Page 138
DAHEST, Parser for Simple Commands to Interactive Programs
The following arguments are used for input only, and are
returned unchanged.
KMDTYP = specifies whether a command word is to be
recognized at the start of the statement. KMDTYP
equal to 1 or 2 specifies that a leading number
encountered in place of a command word is to be
treated specially. KMDTYP equal to 2 indicates
that such a leading number is to be evaluated as a
real number, and for routine DAIHST this value of
KMDTYP is exactly equivalent to the value 1.
= -1, a command word is not to be recognized at the
start of the statement. The contents of the
statement are to be evaluated as an argument list
of the type indicated by the value of LSTTYP. If
LSTTYP also has the value -1 and if an argument is
found, then KIND will be returned containing one of
the values 7, 8, 9 or 10, depending upon the type
of argument.
= 0, the statement must start with a known command
word. If a known command word does begin the
statement, then KIND will be returned with the
value 3 if there is no error in the argument list,
or with one of the values 7, 8, 9 or 10 if an error
is found in the argument list.
If the line of text contains no printing characters
or contains merely an exclamation point followed by
a comment, or if the line of text contains an extra
semicolon, then KIND will be returned with the
value 2 indicating an empty statement.
If the statement starts with a number, a text
string, a comma, a slash or a colon, then KIND will
be returned containing the value 5 to indicate a
missing command word.
If the statement does not start with a known
command word, and if the first printing character
in the statement is not one of the characters
exclamation point, semicolon, apostrophe, comma,
slash, colon or either parenthesis, and if the
first printing character in the statement is not
one of the characters which can start a number,
then KIND will be returned with the value 6
indicating an unknown command word, and the
arguments INIPRT and MIDPRT will be returned
pointing to the leftmost and rightmost characters
in this unknown command word.
= 1, same as KMDTYP=0, except that if an initial
number is found in the statement, then the value of
the number is evaluated as an integer and is
returned in NUMVAL(NUMLOW), an indication of the
sign if any is returned in NUMSIN(NUMLOW), KIND is
FASP, FORTRAN Alphameric Subroutine Package Page 139
DAHEST, Parser for Simple Commands to Interactive Programs
returned with the value 4, and MIDPRT and LMTPRT
are returned pointing to the character to the right
of the number, and to the rightmost printing
character in the IBUFFR array respectively. If the
first printing character on the line is a slash or
a colon, then KIND is returned with the value 5 to
indicate a missing command word, since the slash or
colon is not considered to be part of a number.
= 2, same as KMDTYP=1, except that if an initial
number is found in the statement, then the value of
the number is evaluated as a real number and is
returned in VALNUM(NUMLOW).
LSTTYP = specifies the type of numeric arguments, real or
integer, which can be evaluated, and the type of
argument, word or set of numbers or text string,
which can be present more than once in the argument
list. The values 5, 6, 7 and 8 specify that
numeric arguments are to be evaluated as real
numbers, and for the routine DAIHST these values of
LSTTYP are exactly equivalent to the values 1, 2, 3
and 4 respectively.
= -1, the type of argument list is specified for each
possible command word by the value in the LEGAL
array parallel to the character count in the KNTLTR
array.
= 0, no descriptions of arguments are to be returned
to the calling program. The arguments are,
however, interpreted to find the right end of the
statement. If an argument is found, then KIND will
be returned with one of the value 7, 8, 9 or 10,
depending upon the type of the argument, instead of
being returned with the value 3.
= 1, the type of argument which can be repeated is
the type of the first argument encountered, whether
or not this is preceded by commas. The commas do,
however, indicate missing arguments of the same
type as that eventually found. Additional
arguments of types other than that of the first
argument encountered are not allowed. If an
additional argument of another type is found, then
evaluation of the argument list will be terminated
except insofar as is necessary to detect the end of
the statement, and KIND will be returned containing
one of the values 8, 9 or 10 indicating the type of
the illegal argument. If a set of numbers is
found, it is evaluated as a set of integers.
= 2, allow a series of words. A single text string
and/or a single set of numbers can also appear in
the argument list. If a set of numbers is found,
it is evaluated as a set of integers.
= 3, allow a series of sets of integers. A single
word and/or a single text string can also appear in
FASP, FORTRAN Alphameric Subroutine Package Page 140
DAHEST, Parser for Simple Commands to Interactive Programs
the argument list.
= 4, allow a series of text strings. A single word
and/or a single set of numbers can also appear in
the argument list. If a set of numbers is found,
it is evaluated as a set of integers.
= 5, 6, 7 and 8, same as the LSTTYP values of 1, 2, 3
and 4 respectively, except that if a number or set
of numbers or series of sets of numbers are
encountered, these numbers are evaluated as real
numbers and returned in the VALNUM array rather
than in the NUMVAL array. For the routine DAIHST,
the LSTTYP values of 5, 6, 7 and 8 are exactly
equivalent to the LSTTYP values of 1, 2, 3 and 4
respectively.
= 9, all arguments will be text strings delimited by
parentheses, not by apostrophes. Each text string
begins with the first printing character and
extends to the next punctuation character or space,
or if a parenthesis is encountered within the text
string, to the parenthesis which returns the
parenthetical level count to zero. An apostrophe
can be included anywhere within the text string to
indicate that the following character is to be
included within the text string and is not to
change the parenthetical level count. Since a
leading parenthesis is not necessary, it is
included within the text string if found. The
closing parenthesis is never included within the
contents of the text string. For example
ABC DEF,GHI(JKL)MNO)PQR((STU))VWX(' '(')
would include the following text strings
ABC
DEF
GHI(JKL
MNO)PQR
(STU
)VWX
' '(')
It must be noted that in order for this particular
type of parenthetical expression to be recognized
either LSTTYP=9 or else LSTTYP=-1 while the value
in the LEGAL array corresponding to the command is
9. Under all other conditions, quoted text strings
begin either with an apostrophe or with a
parenthesis and this initial delimiting character
is not then included within the text string.
NAMLOW = subscript of the NAMLFT and NAMRIT array locations
into which can be placed the description of the
first word encountered in the argument list.
FASP, FORTRAN Alphameric Subroutine Package Page 141
DAHEST, Parser for Simple Commands to Interactive Programs
NAMLOW would normally have the value 1 to allow the
description of the first word argument to be placed
into the first locations in the NAMLFT and NAMRIT
arrays.
NAMMAX = highest subscript of the NAMLFT and NAMRIT array
locations into which can be placed the description
of a word in the argument list. NAMMAX would
normally be the dimension of the NAMLFT and NAMRIT
arrays. If NAMMAX equals NAMLOW, then at most a
single word can be evaluated in the argument list
even if a series of words is enabled by either of
the arguments LSTTYP or LEGAL. If NAMMAX is less
than NAMLOW, then no words can be evaluated in the
argument list.
MRKLOW = subscript of the MRKLFT and MRKRIT array locations
into which can be placed the description of the
first text string encountered in the argument list.
MRKLOW would normally have the value 1 to allow the
description of the first text string argument to be
placed into the first locations in the MRKLFT and
MRKRIT arrays.
MRKMAX = highest subscript of the MRKLFT and MRKRIT array
locations into which can be placed the description
of a text string in the argument list. MRKMAX
would normally be the dimension of the MRKLFT and
MRKRIT arrays. If MRKMAX equals MRKLOW, then at
most a single text string can be evaluated in the
argument list even if a series of text strings is
enabled by either of the arguments LSTTYP or LEGAL.
If MRKMAX is less than MRKLOW, then no text strings
can be evaluated in the argument list.
NUMLOW = subscript of the NUMSIN, NUMVAL and VALNUM array
locations into which can be placed the description
of the first number encountered in the argument
list. NUMLOW would normally have the value 1 to
allow the description of the sign of the first
numeric argument to be placed into the first
location in the NUMSIN array and the value of the
argument to be placed, depending upon whether the
number is evaluated as integer or real, into the
first location in either the NUMVAL or VALNUM
arrays.
NUMMAX = highest subscript of the NUMSIN, NUMVAL and VALNUM
array locations into which can be placed the
description of a number in the argument list.
NUMMAX would normally be the dimension of the
NUMSIN, NUMVAL and VALNUM arrays. If NUMMAX equals
NUMLOW, then at most a single number can be
evaluated in the argument list even if a series of
FASP, FORTRAN Alphameric Subroutine Package Page 142
DAHEST, Parser for Simple Commands to Interactive Programs
sets of numbers is enabled by either of the
arguments LSTTYP or LEGAL. If NUMMAX is less than
NUMLOW, then no numbers can be evaluated in the
argument list.
INTRVL = the maximum number of numeric arguments which can
be separated by slashes and/or by colons to form a
set of numeric arguments. Each slash or colon
causes the description of the immediately following
numeric argument to be placed into the next higher
location in the NUMSIN, NUMVAL and VALNUM arrays.
If 2 numeric arguments are separated by something
other than a slash or colon, then these arguments
are taken to be part of a series of sets of
numbers, and the description of the second number
is placed into the locations in the NUMSIN, NUMVAL
and VALNUM arrays having subscripts greater by the
value of INTRVL than the subscripts of the
locations into which was placed the description of
the first number of the previous set. For example,
if
NUMLOW=1, NUMMAX=36 (or greater) and INTRVL=5,
then the argument list
1/2//4 //+8/,11+16WORD+21'TEXT STRING'26,31 36
would consist of the word WORD, the text string
TEXT STRING, and the numbers 1, 2, 4, 8, 11, 16,
21, 26, 31 and 36 which were chosen for the example
so as to be stored at subscripts equal to their
values. The NUMSIN array locations having
subscripts 3, 6, 7 and 9 would be returned set to
-1 to show that the corresponding numbers were
indicated by slashes or colons to be missing.
LOWWRD = subscript of the location in the IWORD array which
contains the first character of the first word in
the dictionary. Note that if KNTLTR(LOWKNT) is
less than or equal to zero, then the first
character of the first word is instead contained in
IWORD(LOWWRD-KNTLTR(LOWKNT)).
MAXWRD = maximum dimension of the IWORD array.
IWORD = dictionary array containing the characters of the
command words which are to be recognized, 1
character per array location as read by an A1
format or else defined by 1H field. All alphabetic
letters within the IWORD array must be supplied in
upper case. The command word typed by the user and
read into the input buffer array IBUFFR can be
split into 2 or more portions any of which can be
FASP, FORTRAN Alphameric Subroutine Package Page 143
DAHEST, Parser for Simple Commands to Interactive Programs
abbreviated and/or separated by spaces or tabs if
the word in the IWORD array contains a single space
at the location at which the split is allowed, and
if the length stored in the KNTLTR array is 100
more than the number of characters including the
space or spaces which form the word in the IWORD
array. For example, if the IWORD array contains
1HN,1HO,1H ,1HL,1HI,1HM,1HI,1HT
and if the KNTLTR array contains the corresponding
length of 108, then this word could be selected by
any of the letter sequences
N, NO, N L, NL, NO L, NOL, N LI or NLI
providing in each case that the sequence is not
ambiguous.
LOWKNT = subscript of the KNTLTR array location containing
the length of the first word which can be matched
in the IWORD array. This first word will start at
IWORD(LOWWRD). If LSTTYP is less than zero
indicating that the argument list type is specified
in the LEGAL array for each possible command word,
then LOWKNT is also the subscript of the LEGAL
array location containing the argument list type
associated with the command word having its
character count in KNTLTR(LOWKNT).
MAXKNT = subscript of the KNTLTR array location containing
the length of the final word which can be matched
in the IWORD array. If LSTTYP is less than zero
indicating that the argument list type is specified
in the LEGAL array for each possible command word,
then MAXKNT is also the subscript of the LEGAL
array location containing the argument list type
associated with the command word having its
character count in KNTLTR(MAXKNT).
KNTLTR = array containing the numbers of characters in the
words in the IWORD array. A zero or negative value
in the KNTLTR array offsets the next possible word
which can be matched in the IWORD array by the
number of letters given by the absolute value of
the negative number in the KNTLTR array. The
dimension of KNTLTR must be at least MAXKNT. For
example to recognize the words
YES, NO, MAYBE
the contents of the IWORD array would be
1HY,1HE,1HS,1HN,1HO,1HM,1HA,1HY,1HB,1HE
FASP, FORTRAN Alphameric Subroutine Package Page 144
DAHEST, Parser for Simple Commands to Interactive Programs
and the contents of the KNTLTR array would be
3,2,5
LEGAL = if LSTTYP has the value -1, then LEGAL is an array
which specifies for each possible command word the
type of numeric arguments, real or integer, which
can be evaluated, and the type of argument, word or
set of numbers or text string, which can be present
more than once in the argument list. The
specification of the allowable argument list
construction for a particular command word is found
at the same subscript in the LEGAL array as the
character count in the KNTLTR array. If LSTTYP is
greater than or equal to zero, then the contents of
the LEGAL array are ignored.
The values 0 through 8 in the LEGAL array specify
the following allowable argument list
constructions. The values 5, 6, 7 and 8 specify
that numeric arguments are to be evaluated as real
numbers, and for the routine DAIHST these values in
the LEGAL array are exactly equivalent to the
values 1, 2, 3 and 4 respectively.
= 0 (or -1), no descriptions of arguments are to be
returned to the calling program. The arguments
are, however, interpreted to find the right end of
the statement. If an argument is found, then KIND
will be returned with one of the value 7, 8, 9 or
10, depending upon the type of the argument,
instead of being returned with the value 3.
= 1, the type of argument which can be repeated is
the type of the first argument encountered, whether
or not this is preceded by commas. The commas do,
however, indicate missing arguments of the same
type as that eventually found. Additional
arguments of types other than that of the first
argument encountered are not allowed. If an
additional argument of another type is found, then
evaluation of the argument list will be terminated
except insofar as is necessary to detect the end of
the statement, and KIND will be returned containing
one of the values 8, 9 or 10 indicating the type of
the illegal argument. If a set of numbers is
found, it is evaluated as a set of integers.
= 2, allow a series of words. A single text string
and/or a single set of numbers can also appear in
the argument list. If a set of numbers is found,
it is evaluated as a set of integers.
= 3, allow a series of sets of integers. A single
word and/or a single text string can also appear in
the argument list.
= 4, allow a series of text strings. A single word
and/or a single set of numbers can also appear in
FASP, FORTRAN Alphameric Subroutine Package Page 145
DAHEST, Parser for Simple Commands to Interactive Programs
the argument list. If a set of numbers is found,
it is evaluated as a set of integers.
= 5, 6, 7 and 8, same as the LEGAL array values of 1,
2, 3 and 4 respectively, except that if a number or
set of numbers or series of sets of numbers are
encountered, these numbers are evaluated as real
numbers and returned in the VALNUM array rather
than in the NUMVAL array. For the routine DAIHST,
the LEGAL array values of 5, 6, 7 and 8 are exactly
equivalent to the LEGAL array values of 1, 2, 3 and
4 respectively.
= 9, all arguments will be text strings delimited by
parentheses, not by apostrophes. Each text string
begins with the first printing character and
extends to the next punctuation character or space,
or if a parenthesis is encountered within the text
string, to the parenthesis which returns the
parenthetical level count to zero. An apostrophe
can be included anywhere within the text string to
indicate that the following character is to be
included within the text string and is not to
change the parenthetical level count. Since a
leading parenthesis is not necessary, it is
included within the text string if found. The
closing parenthesis is never included within the
contents of the text string.
MAXBFR = subscript of the IBUFFR array location containing
the rightmost (highest subscript) character in the
line being interpreted. MAXBFR would normally be
the dimension of the IBUFFR array.
The following arguments are used for both input and output.
IBUFFR = the input buffer array containing the characters of
the line being interpreted, one character per array
location, as read by a multiple of an A1 format.
Such a line can contain one or more statements.
The alphabetic letters forming the command words
and the numeric exponents which appear in the
IBUFFR array can be either upper or lower case.
If a quoted text string itself containing
apostrophes is found in the contents of the IBUFFR
array, then the extra apostrophes needed to mark
the apostrophes which are to remain in the text
string are removed from the text string.
LOWBFR = subscript within the IBUFFR array of the location
which contains the first (leftmost) character of
the line of text to be interpreted. LOWBFR will be
returned pointing to the first character beyond (to
the right of) the interpreted statement. If a
semicolon appears at the end of the statement, then
FASP, FORTRAN Alphameric Subroutine Package Page 146
DAHEST, Parser for Simple Commands to Interactive Programs
LOWBFR is returned pointing to the semicolon and
will be advanced beyond the semicolon by the
subsequent call to this routine. If an exclamation
point appears at the end of the statement, or if
there are no more printing characters to the right
of the statement, then LOWBFR will be returned
containing MAXBFR+1, but the calling program should
not read in a new line of text and should not reset
LOWBFR to the start of the new contents of the
buffer until after DAHEST has returned KIND=1
indicating that the evaluation of the line of text
has been completed.
KIND = must be set to zero by the calling program before
this routine is first called to evaluate a line of
text. KIND is then returned describing the type of
statement which was evaluated. The calling program
should reset KIND to have the value zero if the
evaluation of the contents of the line of text is
being abandoned by the calling program before this
routine has indicated by returning KIND=1 that it
has completed the evaluation of the line of text.
Except for this instance in which the
interpretation is being abandoned by the calling
program, the value of KIND is otherwise passed
unchanged to the subsequent call to this routine.
= 1, (processing completed) returned if the previous
calls to this routine have completed the evaluation
of the contents of the line of text. If the
original line of text contained no printing
characters or contained a leading exclamation point
indicating that the characters to its right formed
a comment, then the previous call to this routine
returned KIND=2 to indicate an empty statement.
The calling program should read a new line of text
and reset LOWBFR to point to the first character in
the new text.
= 2, (empty statement) returned if the original line
of text contained no printing characters or
contained a leading exclamation point indicating
that the characters to its right formed a comment.
KIND is also returned set to 2 if an extra
semicolon indicates a missing statement. A leading
semicolon would indicate an initial missing
statement. A final semicolon, possibly followed by
an exclamation point and comment, would indicate a
final missing statement. Two adjacent semicolons
would indicate a missing statement between them. A
line of text in which the only printing characters
are 2 semicolons, and possibly a following
exclamation point and comment, would specify 3
missing statements.
= 3, (correct statement) returned if the statement
was not empty and was evaluated without errors. If
FASP, FORTRAN Alphameric Subroutine Package Page 147
DAHEST, Parser for Simple Commands to Interactive Programs
KMDTYP is greater than or equal to zero, then a
known command word, or else a nonambiguous
abbreviation thereof, was found and the sequence
number of this command word within the dictionary
is returned as the value of KOMAND. If KMDTYP is
less than zero, then KOMAND is returned with the
value zero, and the statement contained at least a
comma, a slash, a colon or an argument.
= 4, (initial number) returned if KMDTYP is greater
than zero, and if a number was found at the start
of the statement. MIDPRT is returned containing
the subscript within the IBUFFR array of the
character to the immediate right of the number.
LMTPRT is returned containing the subscript within
the IBUFFR array of the rightmost printing
character within the IBUFFR array. If no printing
characters appear to the right of the number, then
LMTPRT will be returned pointing to the rightmost
character of the number, and will be less than
MIDPRT. Other than to determine the rightmost
printing character, the characters to the right of
the number are not interpreted, and in particular
are not tested to find apostrophes, semicolons and
exclamation points.
= 5, (missing command) returned if a command word or
else a leading number was required but not found,
but the statement is not empty. No argument
descriptions are returned to the calling program.
This value of KIND is never returned if KMDTYP is
less than zero. If KMDTYP is zero, then the
statement starts with a number, a text string, a
comma, a slash or a colon since any other initial
printing characters which do not match a word in
the dictionary would be assumed to form an unknown
or misspelled command word. If KMDTYP is greater
than zero, then the statement starts with a text
string, a comma, a slash or a colon.
= 6, (unknown command) returned if an initial command
word was required, but the statement starts with a
sequence of printing characters which could form a
command word, but which do not match a word in the
dictionary, or which form an ambiguous abbreviation
of 2 or more words in the dictionary, or which do
match a single word in the dictionary but are
followed immediately by additional alphabetic
characters or digits. INIPRT and MIDPRT are
returned pointing to the leftmost and rightmost
characters in this unknown command word. No
argument descriptions are returned to the calling
program. This value of KIND is never returned if
KMDTYP is less than zero. If KMDTYP is greater
than or equal to zero, then the first printing
character in the statement is not one of the
characters exclamation point, semicolon,
FASP, FORTRAN Alphameric Subroutine Package Page 148
DAHEST, Parser for Simple Commands to Interactive Programs
apostrophe, comma, slash or colon, and is not one
of the characters which can start a number.
= 7, 8, 9 or 10, (too many arguments) same as when
KIND is returned containing 3, except that the
maximum number of arguments of a single type was
exceeded during the evaluation of the argument
list. The description of the argument which
exceeded the limit, as well as the descriptions of
any arguments to its right, are not returned to the
calling program, although the scanning of the
statement continues to determine the right end of
the statement.
= 7, returned if too many slashes and/or too many
colons were encountered in a set of numbers.
= 8, returned if too many words were found.
= 9, returned if too many sets of numbers were found.
= 10, returned if too many text strings were found.
The following arguments are used only for output. Their
input values are ignored.
KOMAND = if a command word is recognized, then KOMAND is
returned containing the sequence number of a
command word matched in the IWORD array. For
example, if the second command word is matched,
then KOMAND would be returned containing 2. The
sequence number of the command word in the IWORD
array does not include the letters skipped over by
the value of LOWWRD being greater than 1, and does
not include the letters skipped over by negative
values encountered in the KNTLTR array. If a
command word in the IWORD array is matched, KMDTYP
being greater than or equal to zero and KIND being
returned containing 3 or containing 7 or greater,
then KOMAND is the number of values in the KNTLTR
array which are greater than zero starting at
KNTLTR(LOWKNT) up to and including the KNTLTR
location which contains the number of letters in
the command word which is successfully matched.
KOMAND is returned containing zero if a command
word could not be matched, regardless of whether
this was due to KMDTYP being less than zero, or to
a command word being missing or misspelled, or to
the statement starting with a number.
LCNWRD = if a command word was recognized, KIND being
returned containing 3 or containing 7 or greater,
and KOMAND being returned greater than zero, then
LCNWRD is returned containing the subscript of the
IWORD array location which contains the first
character of the command word matched in the
dictionary.
FASP, FORTRAN Alphameric Subroutine Package Page 149
DAHEST, Parser for Simple Commands to Interactive Programs
LCNWRD is returned undefined if KOMAND is returned
set to zero indicating that no command word was
recognized.
LCNKNT = if a command word was recognized, KIND being
returned containing 3 or containing 7 or greater,
and KOMAND being returned greater than zero, then
LCNKNT is returned containing the subscript of the
KNTLTR array location which contains the number of
characters in the command word matched in the
dictionary. If there are no zero or negative
entries in the KNTLTR array, then LCNKNT is
returned equal to LOWKNT+KOMAND-1.
LCNKNT is returned undefined if KOMAND is returned
set to zero indicating that no command word was
recognized.
INIPRT = returned containing the subscript of the IBUFFR
array location containing the leftmost printing
character in the statement if KIND is returned
containing 3 or greater. If an error is found in
the text being interpreted, then IBUFFR(INIPRT)
through and including IBUFFR(LMTPRT) can be
displayed by the calling program as the statement
containing the error, although this text may not be
exactly what was input by the user since when 2
adjacent apostrophes appear within text strings
which are delimited by apostrophes, one of the 2
adjacent apostrophes will already have been
removed. INIPRT is returned undefined if KIND is
returned set to either 1 or 2.
= if KIND is returned containing 6 indicating that an
unknown command word was found at the start of the
statement, then INIPRT is returned pointing to the
left character of the unknown command word. MIDPRT
will then be returned pointing to the right
character of the unknown command word.
MIDPRT = returned undefined if KIND is returned containing
any value other than 4 or 6.
= if KMDTYP was input greater than zero allowing an
initial number in place of an initial command word,
and if such an initial number was found so that
KIND is returned containing 4, then MIDPRT is
returned pointing to the character to the immediate
right of the number. LMTPRT will then be returned
pointing to the rightmost printing character on the
line. If no printing characters appear to the
right of the number, then LMTPRT will be returned
pointing to the rightmost character of the number
and will be less than the returned value of MIDPRT.
= if KIND is returned containing 6 indicating that an
unknown command word was found at the start of the
FASP, FORTRAN Alphameric Subroutine Package Page 150
DAHEST, Parser for Simple Commands to Interactive Programs
statement, then MIDPRT is returned pointing to the
right character of the unknown command word.
LMTPRT = returned containing the subscript of the IBUFFR
array location containing the rightmost printing
character in the statement if KIND is returned
containing 3 or greater. If the statement is
followed by either a semicolon or an exclamation
point, then LMTPRT is returned pointing to the
rightmost printing character to the left of the
semicolon or exclamation point. LMTPRT is returned
undefined if KIND is returned set to either 1 or 2.
= if KMDTYP was input greater than zero allowing an
initial number in place of an initial command word,
and if such an initial number was found so that
KIND is returned containing 4, then LMTPRT is
returned pointing to the rightmost printing
character in the buffer. If no printing characters
appear to the right of the number, then LMTPRT will
be returned pointing to the rightmost character of
the number and will be less than the returned value
of MIDPRT.
NAMKNT = returned containing the highest subscript used in
the NAMLFT and NAMRIT arrays to store the locations
within the buffer of the ends of words found in the
argument list. If no words were found in the
argument list, then NAMKNT will be returned
containing NAMLOW-1. NAMKNT cannot be returned
greater than NAMLOW unless the evaluation of a
series of words is enabled.
NAMLFT = array returned containing the subscripts within the
IBUFFR array of the locations which contain the
left characters of the words in the argument list.
If the evaluation of a series of words is enabled,
but some intermediate words are indicated as
missing by the appearance of extra commas in the
argument list, then for these missing words the
NAMRIT array will contain values which are less
than those in the NAMLFT array. Missing words at
the right end of the statement indicated by extra
commas at the right end of the statement are not
included in the value of NAMKNT and are not
indicated as being missing by the values in the
NAMLFT and NAMRIT arrays.
NAMRIT = array returned containing the subscripts within the
IBUFFR array of the locations which contain the
right characters of the words in the argument list.
If the evaluation of a series of words is enabled,
but some intermediate words are indicated as
missing by the appearance of extra commas in the
argument list, then for these missing words the
FASP, FORTRAN Alphameric Subroutine Package Page 151
DAHEST, Parser for Simple Commands to Interactive Programs
NAMRIT array will contain values which are less
than those in the NAMLFT array. Missing words at
the right end of the statement indicated by extra
commas at the right end of the statement are not
included in the value of NAMKNT and are not
indicated as being missing by the values in the
NAMLFT and NAMRIT arrays.
MRKKNT = returned containing the highest subscript used in
the MRKLFT and MRKRIT arrays to store the locations
within the IBUFFR array of the ends of quoted text
strings found in the argument list. If no quoted
text strings were found in the argument list, then
MRKKNT will be returned containing MRKLOW-1.
MRKKNT cannot be returned greater than MRKLOW
unless the evaluation of a series of quoted text
strings is enabled.
MRKLFT = array returned containing the subscripts within the
IBUFFR array of the locations which contain the
left characters of the quoted text strings in the
argument list. The character pointed to by a value
in the MRKLFT array is the character to the
immediate right of the initial apostrophe or of the
initial parenthesis at the left end of the quoted
text string. If LSTTYP=9, or if LSTTYP=-1 and the
value in the LEGAL array corresponding to the
command is 9, then the character pointed to by a
value in the MRKLFT array is either a leading
parenthesis, or is the leftmost printing character
which is neither a punctuation character nor a
parenthesis.
The character to the left of the delimiting
character at the right end of the quoted text
string is pointed to by the location in the MRKRIT
array having the same subscript as the MRKLFT array
location pointing to the character at the left end
of the quoted text string. If a leading apostrophe
is followed immediately by another apostrophe which
is not itself followed by an apostrophe, or if a
leading left parenthesis is followed immediately by
a closing right parenthesis, or if a leading right
parenthesis is followed immediately by a closing
left parenthesis, then the quoted text string is
considered to be empty and the location in the
MRKLFT array will point to the closing delimiting
character while the location in the MRKRIT array
points to the leading delimiting character so that
the value in the MRKRIT array is 1 less than that
of the corresponding location in the MRKLFT array.
If the closing delimiting character is not found,
then the location in the MRKRIT array will be
returned pointing to the rightmost printing
FASP, FORTRAN Alphameric Subroutine Package Page 152
DAHEST, Parser for Simple Commands to Interactive Programs
character in the IBUFFR array, or to the leading
delimiting character if no printing characters
appear to the right of the leading delimiting
character.
If the evaluation of a series of quoted text
strings is enabled, but some intermediate quoted
text strings are indicated as missing by the
appearance of extra commas in the argument list,
then for these missing quoted text strings the
MRKRIT array will contain values which are 2 less
than those in the MRKLFT array. Missing quoted
text strings at the right end of the statement
indicated by extra commas at the right end of the
statement are not included in the value of MRKKNT
and are not indicated as being missing by the
values in the MRKLFT and MRKRIT arrays.
MRKRIT = array returned containing the subscripts within the
IBUFFR array of the locations which contain the
right characters of the quoted text strings in the
argument list. The character pointed to by the
value in the MRKRIT array is the character to the
immediate left of the final apostrophe at the right
end of a quoted text string which is begun by an
apostrophe, or is the character to the immediate
left of the final parenthesis at the right end of a
quoted text string which is begun by a parenthesis,
or is the rightmost printing character on the line
if no final delimiting character is found. If a
quoted text string is begun by an apostrophe and
contains 2 adjacent apostrophes marking the single
appearance of an apostrophe as part of the quoted
text string, then the extra apostrophe is removed
from the quoted text string and the portion of the
quoted text string to its right is moved left an
extra character position, and the pointer in the
MRKRIT array then marks the right end of the quoted
text string after the shift is completed.
If the evaluation of a series of quoted text
strings is enabled, but some intermediate quoted
text strings are indicated as missing by the
appearance of extra commas in the argument list,
then for these missing quoted text strings the
MRKRIT array will contain values which are 2 less
than those in the MRKLFT array. Missing quoted
text strings at the right end of the statement
indicated by extra commas at the right end of the
statement are not included in the value of MRKKNT
and are not indicated as being missing by the
values in the MRKLFT and MRKRIT arrays.
NUMKNT = returned containing the highest subscript used in
FASP, FORTRAN Alphameric Subroutine Package Page 153
DAHEST, Parser for Simple Commands to Interactive Programs
the NUMSIN, NUMVAL and VALNUM arrays to store the
sign and value of numeric arguments found in the
argument list. If no numeric arguments were found
in the argument list, then NUMKNT will be returned
containing NUMLOW-1. NUMKNT cannot be returned
greater than NUMLOW+INTRVL-1 unless the evaluation
of a series of sets of numbers is enabled. If
KMDTYP is input greater than zero and if an initial
number is found in the statement, then NUMKNT will
be returned equal to NUMLOW and the description of
the initial number will be returned in
NUMSIN(NUMLOW) and in either NUMVAL(NUMLOW) or
VALNUM(NUMLOW).
NUMSIN = array returned indicating the sign, if any, which
preceded each numeric argument. The value of the
argument is returned in the NUMVAL or VALNUM array
location having the same subscript as the NUMSIN
array location describing the sign. Each location
from NUMSIN(NUMLOW) through NUMSIN(NUMKNT) is
returned containing one of the following values.
= -1, returned if the corresponding numeric argument
was indicated as missing either by the absence of a
number before a slash or a colon, by the absence of
a number after a slash or a colon, or by the
absence of a number between two adjacent slashes or
colons. The corresponding location in either the
NUMVAL or VALNUM arrays, whichever is appropriate,
is also set to zero.
= 0, returned if subsequent numeric arguments (which
are returned in NUMSIN, NUMVAL or VALNUM array
locations having higher subscripts) appear in the
argument list, but the current numeric argument was
indicated as missing by less than INTRVL numbers
being included in a set of numbers or, by 2
adjacent commas in the argument list. The
corresponding location in either the NUMVAL or
VALNUM arrays, whichever is appropriate, is also
set to zero. Missing numeric arguments indicated
by extra commas at the right end of the statement
are not included in the value of NUMKNT and are not
represented in the NUMSIN, NUMVAL and VALNUM
arrays.
= 1, returned if a numeric argument was evaluated,
but no sign appeared to its left.
= 2, returned if a minus sign appeared to the left of
the numeric argument.
= 3, returned if a plus sign appeared to the left of
the numeric argument.
NUMVAL = array returned containing the values of numeric
arguments evaluated as integers. If the numeric
argument was preceded by a minus sign, then the
value will be returned negative and the
FASP, FORTRAN Alphameric Subroutine Package Page 154
DAHEST, Parser for Simple Commands to Interactive Programs
corresponding location in the NUMSIN array will
contain 2.
The following arguments are used only for output, and appear
only in the argument list of routine DAHEST.
VALNUM = array returned containing the values of numeric
arguments evaluated as real numbers. If the
numeric argument was preceded by a minus sign, then
the value will be returned negative and the
corresponding location in the NUMSIN array will
contain 2.
IFLOAT = returned describing whether numeric arguments were
evaluated as integers or as real numbers.
= 0, returned if numeric arguments were evaluated as
integers. The values of these numeric arguments
are then returned in the NUMVAL array.
= 1, returned if numeric arguments were evaluated as
real numbers. The values of these numeric
arguments are then returned in the VALNUM array.
an Example of the Use of DAHEST
-- ------- -- --- --- -- ------
The following sample program demonstrates how most of the
DAHEST arguments are used. The program accepts a line of
text from the user, evaluates the contents of the line and
reports the results. This reading, evaluation and reporting
sequence is repeated until the user terminates execution of
the program. Typical output is shown after the program
listing. Argument types associated with the 11 command
words are described in comments at the start of the program.
C RENBR(TSTHST/TEST DAHEST ROUTINE FROM FASP PACKAGE)
DIMENSION IBUFFR(72),IWORD(148),KNTLTR(21),ISIGN(3),
1NAMLFT(100),NAMRIT(100),MRKLFT(100),MRKRIT(100),
2NUMSIN(100),NUMVAL(100),VALNUM(100),LEGAL(21)
C FOLLOWING WORDS ARE IN DICTIONARY
C REAL NUMBERS TAKES SERIES OF REAL SETS
C INTEGER NUMBERS TAKES SERIES OF INTEGER SETS
C NUMBERS TAKES SERIES OF INTEGER SETS
C STRINGS TAKES INTEGERS AND SERIES OF STRINGS
C WORDS TAKES INTEGERS AND SERIES OF WORDS
C TEXT STRINGS TAKES INTEGERS AND SERIES OF STRINGS
C NO ARGUMENTS NO ARGUMENTS ALLOWED
C NUMERIC SERIES TAKES SERIES OF INTEGER SETS
C NUMERAL SETS TAKES SERIES OF REAL SETS
C GENERAL REAL SERIES OF 1ST FOUND TYPE, REAL
C GENERAL INTEGER SERIES OF 1ST FOUND TYPE, INTEGER
C PARENTHESES LEADING CHARACTERS AND PARENTHESES
DATA IWORD/ 1HI,1HG,1HN,1HO,1HR,1HE,1H*,1HR,1HE,1HA,
11HL,1H ,1HN,1HU,1HM,1HB,1HE,1HR,1HS,1HI,1HN,1HT,1HE,
FASP, FORTRAN Alphameric Subroutine Package Page 155
DAHEST, Parser for Simple Commands to Interactive Programs
21HG,1HE,1HR,1H ,1HN,1HU,1HM,1HB,1HE,1HR,1HS,1H*,1H*,
31H*,1H*,1HN,1HU,1HM,1HB,1HE,1HR,1HS,1HS,1HT,1HR,1HI,
41HN,1HG,1HS,1H*,1HW,1HO,1HR,1HD,1HS,1H*,1HT,1HE,1HX,
51HT,1H ,1HS,1HT,1HR,1HI,1HN,1HG,1HS,1H*,1HN,1HO,1H ,
61HA,1HR,1HG,1HU,1HM,1HE,1HN,1HT,1HS,1HN,1HU,1HM,1HE,
71HR,1HI,1HC,1H ,1HS,1HE,1HR,1HI,1HE,1HS,1HN,1HU,1HM,
81HE,1HR,1HA,1HL,1H ,1HS,1HE,1HT,1HS,1HG,1HE,1HN,1HE,
91HR,1HA,1HL,1H ,1HR,1HE,1HA,1HL,1HG,1HE,1HN,1HE,1HR,
11HA,1HL,1H ,1HI,1HN,1HT,1HE,1HG,1HE,1HR,1HP,1HA,1HR,
21HE,1HN,1HT,1HH,1HE,1HS,1HE,1HS/
DATA KNTLTR/ 99, 99, 99, -1,112,115, -4, 7, 7, -1,
1 5, -1,112, -1,112,114, 0,112,112,115,
2 11/
DATA LEGAL / 0, 0, 0, 9, 7, 3, 9, 3, 4, 9,
1 2, 9, 4, 9, 0, 3, 9, 7, 5, 1,
2 9/
DATA KMDTYP,LSTTYP,NAMLOW,NAMMAX,MRKLOW,MRKMAX,
1NUMLOW,NUMMAX,INTRVL,LOWWRD,MAXWRD,LOWKNT,MAXKNT,
2MAXBFR/1,-1,21,25,76,80,15,39,5,7,148,4,21,72/
DATA ISIGN/1H ,1H-,1H+/
DATA IGREAT,IYES/1H>,1HY/
DATA ITTY,JTTY/5,5/
WRITE(JTTY,1)
1 FORMAT(13H TEST DAIHST ,$)
READ(ITTY,7)IANS
KIND=0
2 WRITE(JTTY,3)KMDTYP,LSTTYP
3 FORMAT(8H KMDTYP(,I2,10H), LSTTYP(,I2,2H) ,$)
READ(ITTY,4)KMDTYP,LSTTYP
4 FORMAT(2I)
5 WRITE(JTTY,6)
6 FORMAT(1X,1H*,$)
READ(ITTY,7)IBUFFR
7 FORMAT(100A1)
LOWBFR=1
C
C FIND NEXT STATEMENT IN BUFFER
IFANY=0
8 IF(IANS.EQ.IYES)GO TO 9
CALL DAHEST(KMDTYP,LSTTYP,NAMLOW,NAMMAX,MRKLOW,
1MRKMAX,NUMLOW,NUMMAX,INTRVL,LOWWRD,MAXWRD,IWORD,
2LOWKNT,MAXKNT,KNTLTR,LEGAL,MAXBFR,IBUFFR,LOWBFR,
3KIND,KOMAND,LCNWRD,LCNKNT,INIPRT,MIDPRT,LMTPRT,
4NAMKNT,NAMLFT,NAMRIT,MRKKNT,MRKLFT,MRKRIT,NUMKNT,
5NUMSIN,NUMVAL,VALNUM,IFLOAT)
GO TO 10
9 CALL DAIHST(KMDTYP,LSTTYP,NAMLOW,NAMMAX,MRKLOW,
1MRKMAX,NUMLOW,NUMMAX,INTRVL,LOWWRD,MAXWRD,IWORD,
2LOWKNT,MAXKNT,KNTLTR,LEGAL,MAXBFR,IBUFFR,LOWBFR,
3KIND,KOMAND,LCNWRD,LCNKNT,INIPRT,MIDPRT,LMTPRT,
4NAMKNT,NAMLFT,NAMRIT,MRKKNT,MRKLFT,MRKRIT,NUMKNT,
5NUMSIN,NUMVAL)
IFLOAT=0
10 IF(KIND.LE.2)GO TO 12
FASP, FORTRAN Alphameric Subroutine Package Page 156
DAHEST, Parser for Simple Commands to Interactive Programs
IFANY=1
WRITE(JTTY,11)(IBUFFR(I),I=INIPRT,LMTPRT),IGREAT
11 FORMAT(10X,1H<,100A1)
12 IF(KOMAND.LE.0)GO TO 14
IEND=KNTLTR(LCNKNT)
IF(IEND.GE.100)IEND=IEND-100
IEND=LCNWRD+IEND-1
WRITE(JTTY,13)(IWORD(I),I=LCNWRD,IEND),IGREAT
13 FORMAT(11H COMMAND =<,100A1)
14 GO TO(15,16,25,18,21,23,25,25,25,25),KIND
C
C CONDITIONS FOR WHICH NO ARGUMENTS ARE REPORTED
15 IF(IFANY.LE.0)GO TO 2
GO TO 5
16 WRITE(JTTY,17)
17 FORMAT(16H EMPTY STATEMENT)
GO TO 8
18 J=NUMSIN(NUMLOW)
WRITE(JTTY,19)NUMVAL(NUMLOW),ISIGN(J)
19 FORMAT(10H NUMBER =,1I12,1A1)
IF(MIDPRT.LE.LMTPRT)WRITE(JTTY,20)
1(IBUFFR(I),I=MIDPRT,LMTPRT),IGREAT
20 FORMAT(11H TEXT =<,100A1)
GO TO 8
21 WRITE(JTTY,22)
22 FORMAT(16H MISSING COMMAND)
GO TO 8
23 WRITE(JTTY,24)(IBUFFR(I),I=INIPRT,MIDPRT),IGREAT
24 FORMAT(11H UNKNOWN =<,100A1)
GO TO 8
C
C REPORT WORD ARGUMENTS
25 I=NAMLOW-1
N=0
26 IF(I.GE.NAMKNT)GO TO 28
I=I+1
N=N+1
J=NAMLFT(I)
K=NAMRIT(I)
IF(K.GE.J)WRITE(JTTY,27)N,(IBUFFR(L),L=J,K),IGREAT
27 FORMAT(6H WORD(,I2,3H)=<,100A1)
GO TO 26
28 IF(KIND.EQ.8)WRITE(JTTY,29)
29 FORMAT(15H TOO MANY WORDS)
C
C REPORT QUOTED TEXT STRING ARGUMENTS
I=MRKLOW-1
N=0
30 IF(I.GE.MRKKNT)GO TO 32
I=I+1
N=N+1
J=MRKLFT(I)
K=MRKRIT(I)
IF(K.EQ.(J-1))WRITE(JTTY,31)N,IGREAT
FASP, FORTRAN Alphameric Subroutine Package Page 157
DAHEST, Parser for Simple Commands to Interactive Programs
IF(K.GE.J)WRITE(JTTY,31)N,(IBUFFR(L),L=J,K),IGREAT
31 FORMAT(6H TEXT(,I2,3H)=<,100A1)
GO TO 30
32 IF(KIND.EQ.10)WRITE(JTTY,33)
33 FORMAT(22H TOO MANY TEXT STRINGS)
C
C REPORT NUMERIC ARGUMENTS
I=NUMLOW-1
N=0
34 IF(I.GE.NUMKNT)GO TO 39
I=I+1
N=N+1
J=NUMSIN(I)
IF(J.EQ.0)GO TO 34
IF(J.LT.0)GO TO 37
IF(IFLOAT.EQ.0)WRITE(JTTY,35)N,NUMVAL(I),ISIGN(J)
35 FORMAT(6H NMBR(,I2,2H)=,1I12,1A1)
IF(IFLOAT.NE.0)WRITE(JTTY,36)N,VALNUM(I),ISIGN(J)
36 FORMAT(6H NMBR(,I2,2H)=,1E12.4,1A1)
GO TO 34
37 WRITE(JTTY,38)N
38 FORMAT(6H NMBR(,I2,9H) MISSING)
GO TO 34
39 IF(KIND.EQ.7)WRITE(JTTY,40)
40 FORMAT(27H TOO MANY SLASHES OR COLONS)
IF(KIND.EQ.9)WRITE(JTTY,41)
41 FORMAT(17H TOO MANY NUMBERS)
GO TO 8
END
Sample Dialog Between a User and the Demonstration Program
------ ------ ------- - ---- --- --- ------------- -------
In the following dialog between a user and the demonstration
program listed on the previous pages, the lines of text
typed by the user appear right of asterisks in the left
column. The program types the characters found in the each
statement and a summary of the contents of the statement.
The demonstration program defines the maximum number of
numbers which can appear in each numeric set to be 5 so that
the first number of each numeric set is at a subscript 5
greater than the first number of the previous set if a
series of numeric sets is both enabled and found.
FASP, FORTRAN Alphameric Subroutine Package Page 158
DAHEST, Parser for Simple Commands to Interactive Programs
TEST DAIHST N
KMDTYP( 1), LSTTYP(-1) 1 -1
*R ONE;R N TWO;R NUM THREE;REA FOUR;REALN FIVE;REA NUM
<R ONE>
COMMAND =<REAL NUMBERS>
WORD( 1)=<ONE>
<R N TWO>
COMMAND =<REAL NUMBERS>
WORD( 1)=<TWO>
<R NUM THREE>
COMMAND =<REAL NUMBERS>
WORD( 1)=<THREE>
<REA FOUR>
COMMAND =<REAL NUMBERS>
WORD( 1)=<FOUR>
<REALN FIVE>
COMMAND =<REAL NUMBERS>
WORD( 1)=<FIVE>
<REA NUM>
COMMAND =<REAL NUMBERS>
*S'A STRING';STRINGSGARBAGE'A STRING';STRINGS'A STRING'
<S'A STRING'>
COMMAND =<STRINGS>
TEXT( 1)=<A STRING>
<STRINGSGARBAGE'A STRING'>
UNKNOWN =<STRINGSGARBAGE>
<STRINGS'A STRING'>
COMMAND =<STRINGS>
TEXT( 1)=<A STRING>
*WORD ONE,,'FIRST',1/2/,,FOUR;WORD,,THREE/2/3'FIRST'FOUR
<WORD ONE,,'FIRST',1/2/,,FOUR>
COMMAND =<WORDS>
WORD( 1)=<ONE>
WORD( 4)=<FOUR>
TEXT( 1)=<FIRST>
NMBR( 1)= 1
NMBR( 2)= 2
NMBR( 3) MISSING
<WORD,,THREE/2/3'FIRST'FOUR>
COMMAND =<WORDS>
WORD( 3)=<THREE>
WORD( 4)=<FOUR>
TEXT( 1)=<FIRST>
NMBR( 1) MISSING
NMBR( 2)= 2
NMBR( 3)= 3
FASP, FORTRAN Alphameric Subroutine Package Page 159
DAHEST, Parser for Simple Commands to Interactive Programs
*TEXT,1//3'TWO' 'THREE'FIRST'FOUR';TEXT/2,,'TWO'FIRST,,'FOUR
<TEXT,1//3'TWO' 'THREE'FIRST'FOUR'>
COMMAND =<TEXT STRINGS>
WORD( 1)=<FIRST>
TEXT( 2)=<TWO>
TEXT( 3)=<THREE>
TEXT( 4)=<FOUR>
NMBR( 1)= 1
NMBR( 2) MISSING
NMBR( 3)= 3
<TEXT/2,,'TWO'FIRST,,'FOUR>
COMMAND =<TEXT STRINGS>
WORD( 1)=<FIRST>
TEXT( 2)=<TWO>
TEXT( 4)=<FOUR>
NMBR( 1) MISSING
NMBR( 2)= 2
*I 1//3 6FIRST//13'ONE',,21;I+1//+3+6,FIRST,//+13,'ONE',,+21
<I 1//3 6FIRST//13'ONE',,21>
COMMAND =<INTEGER NUMBERS>
WORD( 1)=<FIRST>
TEXT( 1)=<ONE>
NMBR( 1)= 1
NMBR( 2) MISSING
NMBR( 3)= 3
NMBR( 6)= 6
NMBR(11) MISSING
NMBR(12) MISSING
NMBR(13)= 13
NMBR(21)= 21
<I+1//+3+6,FIRST,//+13,'ONE',,+21>
COMMAND =<INTEGER NUMBERS>
WORD( 1)=<FIRST>
TEXT( 1)=<ONE>
NMBR( 1)= 1+
NMBR( 2) MISSING
NMBR( 3)= 3+
NMBR( 6)= 6+
NMBR(11) MISSING
NMBR(12) MISSING
NMBR(13)= 13+
NMBR(21)= 21+
*TEXT'APOSTROPHES''IN''TEXT'(APOSTROPHES''IN''TEXT'
<TEXT'APOSTROPHES'IN'TEXT'(APOSTROPHES''IN''TEXT' >
COMMAND =<TEXT STRINGS>
TEXT( 1)=<APOSTROPHES'IN'TEXT>
TEXT( 2)=<APOSTROPHES''IN''TEXT' >
FASP, FORTRAN Alphameric Subroutine Package Page 160
DAHEST, Parser for Simple Commands to Interactive Programs
*TEXT '' 'TWO' '''' 'FOUR' '''''';TEXT'ONE',,'
<TEXT '' 'TWO' ''' 'FOUR' ''''>
COMMAND =<TEXT STRINGS>
TEXT( 1)=<>
TEXT( 2)=<TWO>
TEXT( 3)=<'>
TEXT( 4)=<FOUR>
TEXT( 5)=<''>
<TEXT'ONE',,'>
COMMAND =<TEXT STRINGS>
TEXT( 1)=<ONE>
TEXT( 3)=<>
*N A;NOA;NO A;N A-12;N A'STRING';N A WORD!ALLOWS NO ARGUMENT
<N A>
COMMAND =<NO ARGUMENTS>
<NOA>
COMMAND =<NO ARGUMENTS>
<NO A>
COMMAND =<NO ARGUMENTS>
<N A-12>
COMMAND =<NO ARGUMENTS>
TOO MANY NUMBERS
<N A'STRING'>
COMMAND =<NO ARGUMENTS>
TOO MANY TEXT STRINGS
<N A WORD>
COMMAND =<NO ARGUMENTS>
TOO MANY WORDS
*'MISSING COMMANDS'; ; /55 ;; ,SAME; -123LEADING NUMBER
<'MISSING COMMANDS'>
MISSING COMMAND
EMPTY STATEMENT
</55>
MISSING COMMAND
EMPTY STATEMENT
<,SAME>
MISSING COMMAND
<-123LEADING NUMBER>
NUMBER = -123-
TEXT =<LEADING NUMBER>
*PARENTH ONE TWO(')TWO),THREE)THREE''((FOUR))FIVE(
<PARENTH ONE TWO(')TWO),THREE)THREE''((FOUR))FIVE(>
COMMAND =<PARENTHESES>
TEXT( 1)=<ONE>
TEXT( 2)=<TWO(')TWO>
TEXT( 3)=<THREE)THREE''>
TEXT( 4)=<(FOUR>
TEXT( 5)=<)FIVE>
FASP, FORTRAN Alphameric Subroutine Package Page 161
DAHEST, Parser for Simple Commands to Interactive Programs
*;N;NUM;NUMER;NUMERA;NUMERI;N SE;N SET;N SER;
EMPTY STATEMENT
<N>
UNKNOWN =<N>
<NUM>
UNKNOWN =<NUM>
<NUMER>
UNKNOWN =<NUMER>
<NUMERA>
COMMAND =<NUMERAL SETS>
<NUMERI>
COMMAND =<NUMERIC SERIES>
<N SE>
UNKNOWN =<N SE>
<N SET>
COMMAND =<NUMERAL SETS>
<N SER>
COMMAND =<NUMERIC SERIES>
EMPTY STATEMENT
*INT-12//6.3K/+1.33E+2-123.45E-1/88E2WORD/36
<INT-12//6.3K/+1.33E+2-123.45E-1/88E2WORD/36>
COMMAND =<INTEGER NUMBERS>
WORD( 1)=<WORD>
NMBR( 1)= -12-
NMBR( 2) MISSING
NMBR( 3)= 6300
NMBR( 4)= 133+
NMBR( 6)= -12-
NMBR( 7)= 8800
NMBR(11) MISSING
NMBR(12)= 36
*REAL 12.34E+6 -945.12;REAL +44E2,16K;REAL,-22M,,0.002
<REAL 12.34E+6 -945.12>
COMMAND =<REAL NUMBERS>
NMBR( 1)= 0.1234E+08
NMBR( 6)= -0.9451E+03-
<REAL +44E2,16K>
COMMAND =<REAL NUMBERS>
NMBR( 1)= 0.4400E+04+
NMBR( 6)= 0.1600E+05
<REAL,-22M,,0.002>
COMMAND =<REAL NUMBERS>
NMBR( 6)= -0.2200E+08-
NMBR(16)= 0.2000E-02
FASP, FORTRAN Alphameric Subroutine Package Page 162
DAHEST, Parser for Simple Commands to Interactive Programs
A Program to Maintain the DAHEST or DAVERB Dictionaries
- ------- -- -------- --- ------ -- ------ ------------
DAMENU is a program which constructs the DATA statements
defining the dictionary required by FASP routines such as
DAVERB and DAHEST for word identification. The first 80
characters are read from each line of the input file. The
first line read from the input file is copied into the
output file as a FORTRAN comment line beginning with the
letter C in column 1, but is otherwise ignored. The second
line of the input file must contain separated by spaces (or
commas) the names by which the following 6 items are to be
represented in the DATA statements generated by this program
1 the variable which contains the total number of
characters in all words in the dictionary. This would
correspond to the variable named MAXWRD in the DAHEST
argument list.
2 the variable which contains the total number of words
in the dictionary. This would correspond to the
variable named MAXKNT in the DAHEST argument list.
3 the array which contains identical values for words
which are synonyms. This array does not appear in the
DAHEST argument list. Once all 6 names have been
specified, then a number at the left end of each of the
subsequent lines specifies the value to be placed into
this array for each of the words appearing to its
right. DATA statements defining this array are
generated only if at least one value within this array
is specified, but the name of the array must still be
supplied.
4 the array which contains the characters forming the
words in the dictionary. This would correspond to the
array named IWORD in the DAHEST argument list.
5 the array which contains the length of each word in the
dictionary. This would correspond to the array named
KNTLTR in the DAHEST argument list.
6 the array which contains the argument type associated
with each word in the dictionary. This would
correspond to the array named LEGAL in the DAHEST
argument list. Once all 6 names have been specified,
then a single digit at the right end of each of the
subsequent lines specifies the value to be placed into
this array for each of the words appearing to its left.
DATA statements defining this array are generated only
if at least one value within this array is specified,
but the name of the array must still be supplied.
If the second line does not contain at least 6 groups of
FASP, FORTRAN Alphameric Subroutine Package Page 163
DAHEST, Parser for Simple Commands to Interactive Programs
characters, then additional names will be read from
subsequent lines until 6 have been specified.
Once all 6 names have been established, then each subsequent
line contains a word identification number (which can be
expressed in E, K or M notation), followed by one or more
spaces and then by the spelling of the word. The FASP
routines return to the calling program the sequence number
of the word within the dictionary counting as separate words
all character sequences even though some of these might
logically be synonyms or allowed ambiguous abbreviations.
The word identification numbers are not needed by the FASP
routines, but are instead meant to be used by the calling
program to identify the operation to be performed when the
word or any of its synonyms is encountered. The word
identification number is assumed to be zero if no number is
found at the start of the line. The line can begin with a
comma if a word identification number is not required, but
the leftmost word on the line itself begins with a digit or
a sign or a decimal point. The DATA statements representing
the word identification numbers are generated only if at
least one word identification number is found.
If abbreviations of a word are to be recognized even though
they are not unique across the entire dictionary, then these
abbreviations should follow the complete spelling separated
from it and each other by commas, with the longest
abbreviations coming first. Spaces are ignored at the start
or at the end of the spelling of a word. If two or more
words appear on a single line but are not separated by
commas, then the resulting dictionary will include these
words as a single entry in which these words are separated
by single spaces. For example, if the first 3 letters of
the word duplicate those of some other word, but the word
being described on the current line is to be selected by its
1, 2 or 3 letter abbreviations, then the full spelling would
be followed by a comma and the 3 letter abbreviation, then
by a comma and the 2 letter abbreviation and finally by a
comma and the single letter abbreviation. Alternatively,
the abbreviations can be entered in subsequent lines, but
again the full spelling should appear first and be followed
by its longest abbreviation.
A single digit can appear at the right end of each line to
indicate the type of argument list which the words in the
line are to accept if recognized by DAHEST. The argument
description digit is assumed to be zero if no digit is found
at the right end of the line. The line can end with a comma
if the argument description digit is not required, but the
rightmost character of the rightmost word on the line is a
digit. The DATA statements describing the argument types
are generated only if at least one argument description
digit is found. The values 0 through 9 specified by the
digits 0 through 9 appearing at the right end of a line in
FASP, FORTRAN Alphameric Subroutine Package Page 164
DAHEST, Parser for Simple Commands to Interactive Programs
the input file, if used in the LEGAL array for either DAHEST
or DAIHST, would select the following variations in the
argument lists accepted by the corresponding words
0 (or absent number) will not allow the word to accept
any arguments.
1 allows the word to accept multiple arguments of the
same type as its first argument.
2 allows the word to accept multiple word arguments.
3 allows the word to accept multiple numeric arguments.
4 allows the word to accept multiple text string
arguments.
5, 6, 7 or 8, same as 1, 2, 3 or 4 respectively, except
that numbers if found are returned as reals.
9 allows the word to accept multiple text strings which
begin with any characters which cannot serve as
punctuation, but not to accept any other types of
arguments.
The input file is terminated by a line which either is empty
or contains only a leading number.
For example, if the dictionary is to contain the word
EXAMINE which is to
be identified by the number 135
be abbreviated as either EX or E
take multiple arguments of the same type as its first
argument
and the word EXCHANGE which is to
be identified by the number -8
not be abbreviated as either EX or E
take multiple numeric arguments
and if the names of the variables and arrays in the
generated DATA statements are to be MAXWRD, MAXKNT, IDNTFY,
IWORD, KNTLTR and LEGAL
then the input file would contain
THIS IS A COMMENT LINE
MAXWRD MAXKNT,IDNTFY IWORD KNTLTR LEGAL
135 EXAMINE,EX,E 1
-8,EXCHANGE,3
0
FASP, FORTRAN Alphameric Subroutine Package Page 165
DAHEST, Parser for Simple Commands to Interactive Programs
or
THIS IS A COMMENT LINE
MAXWRD MAXKNT IDNTFY IWORD KNTLTR LEGAL
135 EXAMINE 1
135 EX 1
135 E 1
-8 EXCHANGE 3
0
The following FORTRAN comments and statements would be
generated when either of the above forms of the input file
are read. The initial comment lines stating the argument
types are generated only if at least one line contains a
rightmost non-zero digit. Although not necessary in the
example, the extra arrays and the EQUIVALENCE statements are
generated to allow large arrays to be represented by several
DATA statements each of which defines a portion of the large
array.
CTHIS IS A COMMENT LINE
C
C MULTIPLE NUMERIC LENGTH
C
C 135 DYNAMIC INTEGER 7 EXAMINE
C 2 EX(AMINE)
C 1 E(XAMINE)
C -8 NUMBER INTEGER 8 EXCHANGE
C
DIMENSION IDNTFY(4),IDNTF1(4)
DIMENSION IWORD (18),IWORD1(18)
DIMENSION KNTLTR(4),KNTLT1(4)
DIMENSION LEGAL (4),LEGAL1(4)
C
EQUIVALENCE (IDNTF1(1),IDNTFY(1))
EQUIVALENCE (IWORD1(1),IWORD(1))
EQUIVALENCE (KNTLT1(1),KNTLTR(1))
EQUIVALENCE (LEGAL1(1),LEGAL(1))
C
C NUMBER OF CHARACTERS AND WORDS
DATA MAXWRD,MAXKNT/ 18, 4/
C
C NUMBER IDENTIFYING EACH COMMAND
DATA IDNTF1/135,135,135, -8/
C
C LETTERS FORMING EACH COMMAND
DATA IWORD1/1HE,1HX,1HA,1HM,1HI,1HN,1HE,1HE,1HX,1HE,
1 1HE,1HX,1HC,1HH,1HA,1HN,1HG,1HE/
C
C LENGTH OF EACH COMMAND
DATA KNTLT1/ 7, 2, 1, 8/
C
C TYPE OF ARGUMENT LIST
DATA LEGAL1/ 1, 1, 1, 3/
FASP, FORTRAN Alphameric Subroutine Package Page 166
DAJOIN, Free Format Fraction Evaluation Routine
DDDDD AAA JJ OOOOO IIIIII NN NN
DD DD AAAA JJ OO OO II NNN NN
DD DD AA AA JJ OO OO II NNNN NN
DD DD AA AA JJ OO OO II NN NN NN
DD DD AAAAAAA JJ JJ OO OO II NN NNNN
DD DD AA AA JJ JJ OO OO II NN NNN
DDDDD AA AA JJJJ OOOOO IIIIII NN NN
DAJOIN, Free Format Fraction Evaluation Routine
------ ---- ------ -------- ---------- -------
DAJOIN interprets an array read by the calling program with
a multiple of an A1 format and returns the values of the
possibly signed whole numbers, of the possibly signed
fractions and of the mixed numbers represented by the
characters in the array. A mixed number consists of a
possibly signed whole number followed by one or more spaces
and/or tab characters and then by an unsigned fraction. The
numbers forming the numerators and denominators of fractions
can contain decimal points. The value 12.475 could be
represented as 12.1 1.5/4 since DAJOIN sums the values of
the leading number and the following unsigned fraction.
However, the representation 12.1+1.5/4 would be interpreted
as the value 12.1 followed by the value 0.375 which would be
returned by the next call to this routine.
A number can be followed immediately with no intervening
spaces or tabs by the letter E and a possibly signed
exponent. A percent sign following the number implies E-2,
trailing letter K implies E3 and trailing letter M implies
E6. If an exponent is used with either a fraction or a
mixed number, then the exponent must appear to the immediate
right of the denominator of the fraction without any
intervening spaces or tab characters. The value -1500 could
be represented in any of the following forms
-1500 -1.5K -1 1/2K -3/2K -1.5E3 -1 1/2E3 -3/2E3
Evaluation of a number is terminated whenever an exponent is
found. The buffer contents -1K 1/2 would be interpreted as
the number -1000 followed by the fraction +1/2 which would
be returned by the next call to this routine. The buffer
contents -1 1K/2 would be treated as containing the number
-1 followed by the numbers +1000 and 0 (evaluated as +0/2)
which would be returned by the next 2 calls to this routine.
Fractions can have more than a single denominator. If an
exponent is included with a fraction containing more than a
single slash, then the exponent must appear to the immediate
right of the rightmost denominator. For example, 1/2/3K
would have the value 166.666.... A denominator having a
zero value is assumed instead to have the value 1 to prevent
division by zero. Therefore, 2///3, 2/ and -3/ would be
evaluated as 0.66666...., 2 and -3 respectively.
FASP, FORTRAN Alphameric Subroutine Package Page 167
DAJOIN, Free Format Fraction Evaluation Routine
The DAJOIN Argument List
--- ------ -------- ----
The argument list of routine DAJOIN is
SUBROUTINE DAJOIN(ITRAIL,IBUFFR,MAXBFR,LOWBFR,KIND ,
1 VALUE ,DIVISR,ISHIFT,JSHIFT)
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The following arguments are used for input and are returned
unchanged.
ITRAIL = selects whether exponents are to be recognized. If
not, then each number will terminate prior to the
exponent, and LOWBFR will be returned pointing to
the letter starting the exponent. The subsequent
call to this routine will return KIND=2 indicating
that an illegal character has been found if the
calling program does not first increment the value
of LOWBFR.
= -1, only exponents expressed in E notation are to
be recognized. The returned value will already
have been multiplied by 10.0 raised to the
indicated power. The percent sign and the letters
K and M are to be treated the same as any other
alphabetic character.
= 0, no exponents are to be recognized. Numbers will
be terminated prior to percent signs or to the
letters E or K or M.
= 1, percent signs, the letters K and M, and
exponents expressed in E notation, are all to be
recognized. The returned value will already have
been multiplied by 10.0 raised to the indicated
power.
IBUFFR = input buffer array containing characters typed by
user, read by a multiple of an A1 format, which is
to be searched for numbers. IBUFFR then contains 1
character per computer storage location.
MAXBFR = subscript of the final (rightmost) location in the
IBUFFR array which can be searched for a number.
The following argument must be set by the calling program
before this routine is first called, and then is returned by
this routine describing the location of the first character
in the input buffer not yet processed.
LOWBFR = subscript within the IBUFFR array of the first
(leftmost) character which can be scanned for the
FASP, FORTRAN Alphameric Subroutine Package Page 168
DAJOIN, Free Format Fraction Evaluation Routine
representation of numbers. If a number is found,
then LOWBFR will be returned pointing to the first
character beyond the number which could not be part
of the number. If a number is not found, then
LOWBFR will be returned pointing to the first
printing character (which would have to be a
character other than a plus sign or a minus sign or
a decimal point or a slash or a digit), or beyond
the end of the buffer if the buffer does not
contain any printing characters at or to the right
of IBUFFR(LOWBFR). LOWBFR must be set by the
calling program before this routine is first called
to process the contents of the buffer, and should
not be changed by the calling program if this
routine locates a number. If a printing character
is found which cannot start a number, then the
contents of the buffer should be processed by some
other FASP routine, or at least the calling program
should increment the value of LOWBFR by 1 before
again calling this routine.
The following arguments are used for returning information
to the calling program. Their input values are ignored.
KIND = returned describing the kind of item located in the
IBUFFR array.
= 1, no printing character was found at or to the
right of IBUFFR(LOWBFR). LOWBFR is returned
pointing beyond the end of the IBUFFR array. The
calling program should read a new line into the
IBUFFR array and reset LOWBFR to point to the first
character in this array before again calling this
routine.
= 2, a number was not found, but a printing character
which cannot start a number was found at or to the
right of IBUFFR(LOWBFR). LOWBFR will be returned
pointing to the printing character. The calling
program must increment the value of LOWBFR before
again calling this routine since supplying the same
initial character would produce identical results.
DIVISR is returned containing 1 and VALUE, ISHIFT
and JSHIFT are set to zero if a number is not found
so that KIND equal 1 or 2 can be considered
equivalent to KIND=3 if such is appropriate to the
application for which DAJOIN is being used. It
should be noted that an exponent must be preceded
by either a digit or a sign or a decimal point if
it is to be recognized. KIND will be returned
containing 2 if the first printing character at or
to the right of IBUFFR(LOWBFR) is a percent sign or
one of the letters E, K or M.
= 3 or 4 or 5, a number was found. If this number
included a fraction, then the rightmost denominator
was present. LOWBFR is returned pointing to the
FASP, FORTRAN Alphameric Subroutine Package Page 169
DAJOIN, Free Format Fraction Evaluation Routine
character to the right of the representation of the
number. If ITRAIL=0 and the number is followed by
a percent sign or the letters K, M or E, or if
ITRAIL is equal to -1 and the number is followed by
a percent sign or the letters K or M, then LOWBFR
will be returned pointing to this printing
character and no exponent will be evaluated.
= 3, a possibly signed whole number was found which
was not followed by a fraction. DIVISR is returned
containing one. LOWBFR is returned pointing to the
character to the immediate right of the number.
= 4, a mixed number consisting of a possibly signed
whole number followed by an unsigned fraction was
found. LOWBFR is returned pointing to the
character to the immediate right of the fraction.
= 5, a possibly signed fraction was found which was
not preceded by a whole number. LOWBFR is returned
pointing to the character to the immediate right of
the fraction.
= 6, a mixed number or fraction was found in which
the rightmost denominator was missing. LOWBFR is
returned pointing to the character to the immediate
right of the rightmost slash in the fraction. The
fraction representation 1/2/3/ would return KIND=6,
VALUE=0.16666.... and DIVISR=3.
VALUE = returned containing the value of the number if KIND
is returned containing 3 or greater. This value
has been multiplied by 10.0 raised to the power
indicated by the exponent if any. VALUE is
returned set to zero if a number is not found.
DIVISR = returned containing the value of the rightmost
denominator which is specified in a fraction or a
mixed number. The returned value of DIVISR has not
been multiplied by the power of 10.0 indicated by
the exponent if any. DIVISR is returned set to one
if a number is not found, or if a number is found
which does not contain a fraction, or if a fraction
is found in which a denominator is not specified.
The value 0.666666.... represented as 2/3/ would
return DIVISR containing 3, while the value 2
represented as 2/ would return DIVISR containing 1.
ISHIFT = if a number is found, then ISHIFT is returned
describing which of the characters E, %, K or M, if
any, appeared in the number specification. ISHIFT
is returned set to zero if a number is not found.
= -4, the number was followed by the letter E, but
neither sign nor digits appeared in the exponent.
JSHIFT is returned zeroed. No exponent has been
applied to the returned value.
= -3, the number was followed by the letter E and
then by a minus sign, but no digits appeared in the
FASP, FORTRAN Alphameric Subroutine Package Page 170
DAJOIN, Free Format Fraction Evaluation Routine
exponent. JSHIFT is returned zeroed. No exponent
has been applied to the returned value.
= -2, the number was followed by the letter E and
then by a plus sign, but no digits appeared in the
exponent. JSHIFT is returned zeroed. No exponent
has been applied to the returned value.
= -1, the number was followed by the letter E and
then by the number which is returned as the value
of JSHIFT.
= 0, returned if none of the characters E, %, K or M
follow the number. JSHIFT is returned containing
0.
= 1, the number was followed by a percent sign.
JSHIFT is returned containing -2.
= 2, the number was followed by the letter K. JSHIFT
is returned containing 3.
= 3, the number was followed by the letter M. JSHIFT
is returned containing 6.
JSHIFT = if a number is found, then JSHIFT is returned
containing the value of the exponent evaluated at
the right end of the number representation. The
returned value will have been multiplied by 10.0
raised to the power returned in JSHIFT. If the
letter E follows the number, then JSHIFT is the
number appearing to the right of the letter E, or
is returned containing zero if no digits were found
to the right of the letter E. If one of the
characters %, K or M follow the number, then JSHIFT
is returned containing -2, 3 and 6 respectively.
JSHIFT is returned set to zero if a number is not
found.
An Example of the Use of DAJOIN
-- ------- -- --- --- -- ------
The sample program listed below displays the values
represented by a line of text typed by the user. Each value
is displayed to the right of the original fraction as
reconstructed from the denominator and the exponent if any.
DIMENSION IBUFFR(72),JBUFFR(50)
DATA ISLASH,IEXPNT,IGREAT,IEQUAL,ISPACE/
11H/,1HE,1H>,1H=,1H /
DATA ITTY,JTTY/5,5/
WRITE(JTTY,1)
1 FORMAT(8H ITRAIL ,$)
READ(ITTY,2)ITRAIL
2 FORMAT(I)
3 WRITE(JTTY,4)
4 FORMAT(2H *,$)
READ(ITTY,5)IBUFFR
5 FORMAT(72A1)
FASP, FORTRAN Alphameric Subroutine Package Page 171
DAJOIN, Free Format Fraction Evaluation Routine
LOWBFR=1
C
C EVALUATE NEXT NUMBER IN TEXT
6 INIBFR=LOWBFR
CALL DAJOIN(ITRAIL,IBUFFR,72,LOWBFR,KIND,
1VALUE,DIVISR,ISHIFT,JSHIFT)
IF(KIND.EQ.1)GO TO 3
IF(KIND.EQ.2)GO TO 9
J=LOWBFR-1
IF(J.GE.INIBFR)WRITE(JTTY,7)(IBUFFR(I),I=INIBFR,J),
1IGREAT
7 FORMAT(2H <,132A1)
C
C CONSTRUCT DESCRIPTION OF DENOMINATOR AND OF EXPONENT
KOUNT=0
FRCTN=(VALUE*DIVISR)/(10.0**JSHIFT)
CALL DASHOW(1,0,6,1,6,
15,IVALUE,FRCTN,50,KOUNT,JBUFFR,IERR)
KOUNT=KOUNT+1
JBUFFR(KOUNT)=ISLASH
CALL DASHOW(1,0,6,1,6,
15,IVALUE,DIVISR,50,KOUNT,JBUFFR,IERR)
IF(JSHIFT.EQ.0)GO TO 8
KOUNT=KOUNT+1
JBUFFR(KOUNT)=IEXPNT
CALL DASHOW(0,0,6,0,6,
15,JSHIFT,VALUE,50,KOUNT,JBUFFR,IERR)
8 KOUNT=KOUNT+3
JBUFFR(KOUNT-2)=ISPACE
JBUFFR(KOUNT-1)=IEQUAL
JBUFFR(KOUNT)=ISPACE
CALL DASHOW(1,0,6,1,6,
15,IVALUE,VALUE,50,KOUNT,JBUFFR,IERR)
C
C DESCRIBE THE RESULT, DENOMINATOR AND EXPONENT
GO TO(3,9,11,13,15,17),KIND
9 WRITE(JTTY,10)IBUFFR(LOWBFR)
10 FORMAT(10H UNKNOWN ,1A1)
LOWBFR=LOWBFR+1
GO TO 6
11 WRITE(JTTY,12)(JBUFFR(I),I=1,KOUNT)
12 FORMAT(8H WHOLE ,50A1)
GO TO 6
13 WRITE(JTTY,14)(JBUFFR(I),I=1,KOUNT)
14 FORMAT(8H MIXED ,50A1)
GO TO 6
15 WRITE(JTTY,16)(JBUFFR(I),I=1,KOUNT)
16 FORMAT(8H FRCTN ,50A1)
GO TO 6
17 WRITE(JTTY,18)(JBUFFR(I),I=1,KOUNT)
18 FORMAT(8H ABSNT ,50A1)
GO TO 6
END
FASP, FORTRAN Alphameric Subroutine Package Page 172
DAJOIN, Free Format Fraction Evaluation Routine
Typical Dialog Between DAJOIN Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
ITRAIL 1
*1 2 3 1 2/3 1/2/3
<1>
WHOLE 1/1 = 1
< 2>
WHOLE 2/1 = 2
< 3>
WHOLE 3/1 = 3
< 1 2/3>
MIXED 5/3 = 1.66667
< 1/2/3>
FRCTN .5/3 = .166667
*1K 2/3 1 2K/3 1 2/3K
<1K>
WHOLE 1/1E3 = 1000
< 2/3>
FRCTN 2/3 = .666667
< 1>
WHOLE 1/1 = 1
< 2K>
WHOLE 2/1E3 = 2000
</3>
FRCTN 0/3 = 0
< 1 2/3K>
MIXED 5/3E3 = 1666.67
*-12 3/4 -12-3/4 -12 3/-4
<-12 3/4>
MIXED -51/4 = -12.75
< -12>
WHOLE -12/1 = -12
<-3/4>
FRCTN -3/4 = -.75
< -12 3/>
ABSNT -15/1 = -15
<-4>
WHOLE -4/1 = -4
*12.4, 12.4// 43.4//3.5//
<12.4>
WHOLE 12.4/1 = 12.4
UNKNOWN ,
< 12.4//>
ABSNT 12.4/1 = 12.4
< 43.4//3.5//>
ABSNT 43.4/3.5 = 12.4
FASP, FORTRAN Alphameric Subroutine Package Page 173
DALEAD, Identify Command and Trailing Word, Number or String
DDDDD AAA LL EEEEEEEE AAA DDDDD
DD DD AAAA LL EE AAAA DD DD
DD DD AA AA LL EE AA AA DD DD
DD DD AA AA LL EEEEE AA AA DD DD
DD DD AAAAAAA LL EE AAAAAAA DD DD
DD DD AA AA LL EE AA AA DD DD
DDDDD AA AA LLLLLLLL EEEEEEEE AA AA DDDDD
DALEAD, Identify Command and Trailing Word, Number or String
------ -------- ------- --- -------- ---- ------ -- ------
DALEAD identifies to the calling program a command word and
an associated word or number or quoted text string
represented by the characters in an input buffer read by the
calling program with a multiple of an A1 format. The
calling program supplies to DALEAD two separate dictionaries
which identify all possible command words and all possible
associated words. Any abbreviation which is unique across
both dictionaries is allowed. A word or word abbreviation
must be followed by a known delimiter character such as the
blank, tab, comma, apostrophe, semicolon, exclamation point
or ampersand, or by an end of line. The appearance of any
other character following a word or word abbreviation is
considered to be an error and causes the word or word
abbreviation and the following characters to be identified
as an unknown sequence of characters. A quoted text string
includes the characters to the right of an initial
apostrophe and extends to the second unpaired apostrophe or
through the rightmost printing character in the line if a
second unpaired apostrophe is not found. A pair of
apostrophes within the string indicates the inclusion of a
single apostrophe and the extra apostrophe is removed. Any
character (except of course another apostrophe) can follow
the second unpaired apostrophe which marks the right end of
a quoted text string. A number is terminated to the left of
any character which could not logically extend the number.
If the leftmost printing characters in the buffer (ignoring
leading commas) form an associated word or its abbreviation
or form a number or quoted text string, then the command
word identified by the previous call to this routine and the
new associated word or number or quoted text string are
identified to the calling program. If the printing
characters in the buffer form a command word or its
abbreviation followed to its right by spaces and/or tabs
and/or commas and then by an associated word or its
abbreviation or by a number or by a quoted text string, then
the new command word and the new associated word or number
or quoted text string are identified to the calling program.
If a command word or its abbreviation is followed by spaces
and/or tabs and/or commas and then by a second command word
or its abbreviation, then the first command is identified by
FASP, FORTRAN Alphameric Subroutine Package Page 174
DALEAD, Identify Command and Trailing Word, Number or String
the current call to this routine, and the second command is
identified by the subsequent call.
The range of a command is the sequence of associated words
and/or numbers and/or quoted text strings for which the
command continues to be identified by this routine if a new
command is not encountered. The range of a command extends
to the next appearance of another command, to the next
appearance of a semicolon or exclamation point, or to the
end of a line not preceded by an ampersand. An exclamation
point and any characters to its right are taken to be a
comment and are ignored. An ampersand and any characters to
its right are similarly ignored, but the associated words
and/or numbers and/or quoted text strings at the start of
the next text read into the buffer and interpreted by
subsequent calls to this routine are within the range of the
command in effect when the ampersand was encountered. Words
and numbers cannot extend across commas, but commas are
otherwise ignored. Semicolons are reported to the calling
program.
The DALEAD Argument List
--- ------ -------- ----
The argument list of routine DALEAD is
SUBROUTINE DALEAD(INITYP,NXTTYP,ITRAIL,KWRDLO,KWRDHI,
1 KWORD ,KCNTLO,KCNTHI,KCOUNT,NUMTYP,LWRDLO,LWRDHI,
2 LWORD ,LCNTLO,LCNTHI,LCOUNT,IBUFFR,MAXBFR,LOWBFR,
3 KIND ,KOMAND,KWRDID,KCNTID,LOCAL ,LWRDID,LCNTID,
4 INITAL,IVALUE,VALUE ,IFLOAT)
with the associated DIMENSION statement
DIMENSION KWORD(KWRDHI),KCOUNT(KCNTHI),
1NUMTYP(KCNTHI),LWORD(LWRDHI),LCOUNT(LCNTHI),
2IBUFFR(MAXBFR)
The following arguments are used for input only, and are
returned unchanged. ITRAIL is identical to the DAHEFT
argument having the same name. The dictionary of command
words is defined by KWRDLO, KWRDHI, KWORD, KCNTLO, KCNTHI
and KCOUNT which are identical to the DAVERB arguments
LOWWRD, MAXWRD, IWORD, LOWKNT, MAXKNT and KNTLTR
respectively. The dictionary of associated words is
similarly defined by LWRDLO, LWRDHI, LWORD, LCNTLO, LCNTHI
and LCOUNT.
INITYP = if the representation of a number is found outside
the range of a command, then INITYP specifies
whether the value is to be returned as the integer
argument IVALUE or as the real argument VALUE. The
FASP, FORTRAN Alphameric Subroutine Package Page 175
DALEAD, Identify Command and Trailing Word, Number or String
number can be typed with a decimal point and/or an
exponent regardless of the value of INITYP.
= -2, if the representation of a number is found
outside the range of a command, then it is treated
as an unknown sequence of characters, and KIND is
returned containing the value 7.
= -1, the value is calculated as an octal integer and
returned as the argument IVALUE. However, the
number following the letter E of an exponent is
evaluated in decimal.
= 0, the value is calculated as a decimal integer and
is returned as the argument IVALUE.
= 1 or greater, the value is returned as the real
argument VALUE. If possible, the real number will
be accumulated as an integer, then be converted to
real and shifted as necessary. INITYP is the
maximum number of digits in the integer.
NXTTYP = if the representation of a number is found within
the range of a command, then NXTTYP specifies
whether the value is to be returned as the integer
argument IVALUE or as the real argument VALUE. The
number can be typed with a decimal point and/or an
exponent regardless of the value of NXTTYP.
= -3, whether the number is interpreted as an octal
integer or as a decimal integer or as a decimal
real, and whether the value is returned in the
argument IVALUE or VALUE is specified by the NUMTYP
array entry having the same subscript as the KCOUNT
array entry containing the number of characters in
the command. The calling program must not change
the value of KCNTID returned by the previous call
to this routine if NXTTYP is set to -3.
= -2, if the representation of a number is found
within the range of a command, then it is treated
as an unknown sequence of characters, and KIND is
returned containing the value 7 or 14 depending
upon whether a command word was found by the
current call to this routine.
= -1, the value is calculated as an octal integer and
returned as the argument IVALUE. However, the
number following the letter E of an exponent is
evaluated in decimal.
= 0, the value is calculated as a decimal integer and
is returned as the argument IVALUE.
= 1 or greater, the value is returned as the real
argument VALUE. If possible, the real number will
be accumulated as an integer, then be converted to
real and shifted as necessary. NXTTYP is the
maximum number of digits in the integer.
ITRAIL = selects whether exponents are to be recognized. If
not, then each number will terminate prior to the
exponent, and LOWBFR will be returned pointing to
FASP, FORTRAN Alphameric Subroutine Package Page 176
DALEAD, Identify Command and Trailing Word, Number or String
the letter starting the exponent.
= -1, exponents expressed in E notation are to be
recognized, but the percent sign and the letters K
and M are to be treated the same as any other
alphabetic characters.
= 0, no exponents are to be recognized. Numbers will
be terminated prior to percent signs or to the
letters E or K or M.
= 1, percent signs, the letters K and M, and
exponents expressed in E notation are all to be
recognized.
KWRDLO = subscript of the location in the KWORD array
containing the first character of the first command
word which can be recognized.
KWRDHI = subscript of the location in the KWORD array
containing the final character of the final command
word which can be recognized.
KWORD = array containing the characters of the command
words which can be recognized, 1 character per
array location as though read by a multiple of an
A1 format or defined by several 1H fields. All
alphabetic letters within the KWORD array must be
supplied in upper case.
KCNTLO = subscript of the location in the KCOUNT array
containing the number of characters in the first
command word which can be recognized. If KCNTLO is
greater than KCNTHI, then no command words can be
recognized, causing KIND to always be returned
containing 7 or less.
KCNTHI = subscript of the location in the KCOUNT array
containing the number of characters in the final
command word which can be recognized.
KCOUNT = array containing in locations KCOUNT(KCNTLO)
through and including KCOUNT(KCNTHI) the numbers of
characters in each of the several command words
which can be recognized. A negative entry in the
KCOUNT array causes the number of characters
indicated by the absolute value of the negative
number to be skipped over in the KWORD array
without forming a recognizable word. If the
command word is formed of two or more sections
which can be matched when separated by any number
of spaces or tab characters, then the word stored
in the KWORD array must contain a single space at
each of the locations in the word at which a split
is allowed, and the length stored in the KCOUNT
array must be 100 more than the actual number of
characters including the spaces which are stored in
FASP, FORTRAN Alphameric Subroutine Package Page 177
DALEAD, Identify Command and Trailing Word, Number or String
the KWORD array.
NUMTYP = specifies the manner in which a number is evaluated
if the number is within the range of a command and
if NXTTYP=-3. Whether the number is interpreted as
an octal integer or as a decimal integer or as a
decimal real, and whether the value is returned in
the argument IVALUE or VALUE is specified by the
NUMTYP array entry having the same subscript as the
KCOUNT array entry containing the number of
characters in the command. If NXTTYP is greater
than -3, then the NUMTYP array is ignored. The
values of the individual entries within the NUMTYP
array are defined similarly to those of the
nondimensioned argument NXTTYP (other than
NXTTYP=-3) and are as follow
= -2, if the representation of a number is found
within the range of a command, then it is treated
as an unknown sequence of characters, and KIND is
returned containing the value 7 or 14 depending
upon whether a command word was found by the
current call to this routine.
= -1, the value is calculated as an octal integer and
returned as the argument IVALUE. However, the
number following the letter E of an exponent is
evaluated in decimal.
= 0, the value is calculated as a decimal integer and
is returned as the argument IVALUE.
= 1 or greater, the value is returned as the real
argument VALUE. If possible, the real number will
be accumulated as an integer, then be converted to
real and shifted as necessary. NUMTYP(KCNTID) is
the maximum number of digits in the integer.
LWRDLO = subscript of the location in the LWORD array
containing the first character of the first
associated word which can be recognized.
LWRDHI = subscript of the location in the LWORD array
containing the final character of the final
associated word which can be recognized.
LWORD = array containing the characters of the associated
words which can be recognized, 1 character per
array location as though read by a multiple of an
A1 format or defined by several 1H fields. All
alphabetic letters within the LWORD array must be
supplied in upper case.
LCNTLO = subscript of the location in the LCOUNT array
containing the number of characters in the first
associated word which can be recognized. If LCNTLO
is greater than LCNTHI, then no associated words
can be recognized so that KIND cannot be returned
FASP, FORTRAN Alphameric Subroutine Package Page 178
DALEAD, Identify Command and Trailing Word, Number or String
with either of the values 4 or 11.
LCNTHI = subscript of the location in the LCOUNT array
containing the number of characters in the final
associated word which can be recognized.
LCOUNT = array containing in locations LCOUNT(LCNTLO)
through and including LCOUNT(LCNTHI) the numbers of
characters in each of the several associated words
which can be recognized. A negative entry in the
LCOUNT array causes the number of characters
indicated by the absolute value of the negative
number to be skipped over in the LWORD array
without forming a recognizable word. If the
associated word is formed of two or more sections
which can be matched when separated by any number
of spaces or tab characters, then the word stored
in the LWORD array must contain a single space at
each of the locations in the word at which a split
is allowed, and the length stored in the LCOUNT
array must be 100 more than the actual number of
characters including the spaces which are stored in
the LWORD array.
IBUFFR = input buffer array, containing characters typed by
the user and read by a multiple of an A1 format,
which is to be searched for known command words,
known associated words, numbers and quoted text
strings. IBUFFR then contains 1 character per
computer storage location. The alphabetic letters
forming the command words, associated words and
numeric exponents which appear within the IBUFFR
array can be either upper or lower case. IBUFFR is
returned modified if a quoted text string is found
which contains a pair of apostrophes indicating a
single apostrophe within the string or if a quoted
text string is found which is terminated by a
second unpaired apostrophe.
MAXBFR = subscript of the final (rightmost) location in the
IBUFFR array which can be searched for known words,
numbers and quoted text strings.
The following arguments must be set by the calling program
before this routine is first called, then are returned by
this routine containing information to be used by the
calling program and, usually, to be passed unchanged to the
subsequent call to this routine.
LOWBFR = should be input containing the subscript within the
IBUFFR array of the first (leftmost) character
which can be scanned for known words and numbers
and quoted text strings. LOWBFR is returned
containing the subscript of the IBUFFR array
FASP, FORTRAN Alphameric Subroutine Package Page 179
DALEAD, Identify Command and Trailing Word, Number or String
location to the right of the rightmost character of
the rightmost identified item. If a quoted text
string is found, then LOWBFR is returned pointing
to the character to the right of the rightmost
character within the string, rather than to the
character to the right of the second unpaired
apostrophe. If the quoted text string was not
terminated by a second unpaired apostrophe, then
LOWBFR is returned pointing to the character to the
right of the rightmost printing character in the
buffer. If the quoted text string was terminated
with a second unpaired apostrophe, then this
apostrophe is changed to a space which is pointed
to by the returned value of LOWBFR. If an unknown
sequence of printing characters, or a known word or
abbreviation followed by an unknown sequence of
printing characters, is found, then LOWBFR is
returned pointing to the character to the right of
the unknown sequence. If a new command word is
followed by spaces and/or tabs and/or commas and
then by a second new command word, then LOWBFR is
returned containing the subscript of the IBUFFR
array location which contains the first character
of the second command word which will then be
identified by the subsequent call to this routine.
If an exclamation point or ampersand is found, then
LOWBFR is returned containing MAXBFR+1. LOWBFR
must be reset by the calling program to point to
the leftmost character in the IBUFFR array each
time new text is read into this array.
KIND = input containing 0, this routine has not yet been
called during the execution of the current logical
section of the calling program. This routine is to
begin evaluation of the contents of the input text
buffer, not continue the range of a command word
identified by a previous call to this routine. The
following values of KIND are returned to the
calling program describing the type of item or
items located in the IBUFFR array. Except where
mentioned below, these values are usually passed
unchanged to the subsequent call to this routine.
= 1, either the input buffer was empty, or else the
next printing character at or to the right of
IBUFFR(LOWBFR), ignoring commas, was an exclamation
point. LOWBFR is returned pointing beyond the end
of the buffer. It is expected that the calling
program will read new text into the input buffer
and reset LOWBFR to point to the first character in
the buffer before again calling this routine.
Unless KIND is changed to 3 by the calling program,
the subsequent call to this routine will return
KOMAND zeroed if a new command word is not
identified.
FASP, FORTRAN Alphameric Subroutine Package Page 180
DALEAD, Identify Command and Trailing Word, Number or String
= 2, the next printing character at or to the right
of IBUFFR(LOWBFR), ignoring commas, was a
semicolon. LOWBFR is returned pointing to the
character to the right of the semicolon. The
subsequent call to this routine will return KOMAND
zeroed if a new command word is not identified.
= 3, the next printing character at or to the right
of IBUFFR(LOWBFR), ignoring commas, was an
ampersand. LOWBFR is returned containing MAXBFR+1.
It is expected that the calling program will read
new text into the input buffer and reset LOWBFR to
point to the first character in the buffer before
again calling this routine. Unless KIND is changed
to 1 by the calling program, the subsequent call to
this routine will return the values of KOMAND,
KWRDID and KCNTID unchanged if a new command word
is not identified.
= 4, the next printing characters at or to the right
of IBUFFR(LOWBFR), ignoring commas except for their
use as word separators, uniquely identified an
associated word in the dictionary contained in the
LWORD and LCOUNT arrays. The sequence number of
the identified word among all possible associated
words is returned as the argument LOCAL. LWRDID is
returned containing the subscript of the LWORD
array location containing the first character of
the word, and LCNTID is returned containing the
subscript of the LCOUNT array location containing
the number of characters within the word. INITAL
is returned containing the subscript of the IBUFFR
array location which contains the first character
of the associated word or abbreviation of an
associated word. LOWBFR is returned pointing to
the character to the right of the word or word
abbreviation. The subsequent call to this routine
will return the values of KOMAND, KWRDID and KCNTID
unchanged if a new command word is not identified.
= 5, the next printing characters at or to the right
of IBUFFR(LOWBFR), ignoring commas except for their
use as number separators, formed a number. The
value of the number is returned in the argument
IVALUE if evaluated as an octal integer or as a
decimal integer, or in the argument VALUE if
evaluated as a decimal real. INITAL is returned
containing the subscript of the IBUFFR array
location which contains the first character of the
number. LOWBFR is returned pointing to the
character to the right of the number. The
subsequent call to this routine will return the
values of KOMAND, KWRDID and KCNTID unchanged if a
new command word is not identified.
= 6, the next printing character at or to the right
of IBUFFR(LOWBFR), ignoring leading commas, was an
apostrophe which is pointed to by the returned
FASP, FORTRAN Alphameric Subroutine Package Page 181
DALEAD, Identify Command and Trailing Word, Number or String
value of INITAL. LOWBFR is returned pointing to
the character to the right of the rightmost
character within the quoted text string. If the
quoted text string is not terminated by an unpaired
apostrophe, then LOWBFR is returned pointing to the
character to the right of the rightmost printing
character in the buffer. If the quoted text string
is terminated by an unpaired apostrophe, then this
apostrophe is changed to a space pointed to by the
returned value of LOWBFR. If a pair of adjacent
apostrophes is found within the quoted text string,
then the portion of the quoted text string to the
right of the pair of apostrophes is moved to the
left overwriting the second apostrophe of the pair,
and LOWBFR is returned pointing to the character to
the right of the rightmost character of the quoted
text string in its new location. The location
vacated by the rightmost character is filled with a
space after the shift is complete.
= 7, the next printing characters at or to the right
of IBUFFR(LOWBFR), ignoring commas except for their
use as character sequence separators, formed an
unknown sequence. This value of KIND is returned
also if a word or word abbreviation is not unique
across both dictionaries, or if a command word or
its abbreviation or an associated word or its
abbreviation is followed by any character other
than a space, tab, comma, apostrophe, semicolon,
exclamation point or ampersand. INITAL is returned
containing the subscript of the IBUFFR array
location which contains the first character of the
unknown sequence or of the word or abbreviation
which is followed by an unknown sequence. LOWBFR
is returned pointing to the following space, tab,
comma, apostrophe, semicolon, exclamation point or
ampersand or beyond the end of the line if the
unknown printing characters extend through the end
of the line. The subsequent call to this routine
will return the values of KOMAND, KWRDID and KCNTID
unchanged if a new command word is not identified.
= 8 through 14, same as KIND values 1 through 7
respectively except that a command word or
abbreviation was recognized prior to the end of
line (8), semicolon (9), ampersand (10), associated
word (11), number (12), quoted text string (13) or
unknown sequence (14). The sequence number of the
identified command among all possible commands is
returned as the argument KOMAND. KWRDID is
returned containing the subscript of the KWORD
array location containing the first character of
the command, and KCNTID is returned containing the
subscript of the KCOUNT array location containing
the number of characters within the command. The
following additional information should be noted.
FASP, FORTRAN Alphameric Subroutine Package Page 182
DALEAD, Identify Command and Trailing Word, Number or String
= 8, a command word extended through the rightmost
printing characters in the line, or else the next
printing character to the right of a command word,
ignoring commas except for their use as word
separators, was an exclamation point. Unless KIND
is changed to 10 by the calling program, the
subsequent call to this routine will return KOMAND
zeroed unless a new command is found.
= 10, the next printing character to the right of a
command word, ignoring commas except for their use
as word separators, was an ampersand. The calling
program should read new text into the input buffer
and the subsequent call to this routine is then to
continue the interpretation of the new text as
though this command word appeared at the start of
the new contents of the buffer. Unless KIND is
changed to 8 by the calling program, the subsequent
call to this routine will return KOMAND, KWRDID and
KCNTID unchanged and will return KIND greater than
7.
= 15, ignoring commas except for their use as word
terminators, a command word or its abbreviation was
followed by spaces and/or tabs and/or commas and
then by a second command word or its abbreviation.
The first command is identified by the returned
values of KOMAND, KWRDID and KCNTID. LOWBFR is
returned pointing to the first character of the
second command word which will in turn be
identified by the subsequent call to this routine.
If KIND was input containing the value 10, then the
second command word is of course actually the first
to appear in the current contents of the buffer.
The following arguments are returned to the calling program
identifying the command word found by this routine or which
is still in effect from the previous call to this routine.
The previous command is still in effect and the input values
of these arguments are returned unchanged if KIND is input
with a value other than 0, 1, 2, 8 or 9, and if KIND is
returned with a value less than or equal to 7.
KOMAND = returned containing the sequence number of the
current command word among all possible command
words. The sequence number does not include the
letters skipped over by the value of KWRDLO being
greater than 1, and does not include the letters
skipped over by negative values encountered in the
KCOUNT array. KOMAND is the number of values in
the KCOUNT array which are greater than zero
starting at KCOUNT(KCNTLO) up to and including the
KCOUNT array location which contains the number of
letters in the command word which is successfully
matched. KOMAND is returned containing zero if
KIND is input containing 0, 1, 2, 8 or 9 indicating
FASP, FORTRAN Alphameric Subroutine Package Page 183
DALEAD, Identify Command and Trailing Word, Number or String
that the range of the previous command is being
terminated, and if KIND is returned less than or
equal to 7 indicating that no new command was
found. KOMAND is returned unchanged if KIND is
input containing a value other than 0, 1, 2, 8 or
9, and if KIND is returned less than or equal to 7.
KWRDID = returned containing the subscript of the KWORD
array location which contains the first character
of the matched command word. KWRDID is returned
unchanged if the command identified by a previous
call to this routine is still in effect.
KCNTID = returned containing the subscript of the KCOUNT
array location which contains the number of
characters in the matched command word. KCNTID is
returned unchanged if the command identified by a
previous call to this routine is still in effect.
The following arguments are returned to the calling program
identifying the associated word found if KIND is returned
containing either 4 or 11. These arguments are returned
unchanged if KIND is returned with any value other than 4 or
11.
LOCAL = returned containing the sequence number of the
associated word among all possible associated
words. The sequence number does not include the
letters skipped over by the value of LWRDLO being
greater than 1, and does not include the letters
skipped over by negative values encountered in the
LCOUNT array. LOCAL is the number of values in the
LCOUNT array which are greater than zero starting
at LCOUNT(LCNTLO) up to and including the LCOUNT
array location which contains the number of letters
in the associated word which is successfully
matched.
LWRDID = returned containing the subscript of the LWORD
array location which contains the first character
of the matched associated word.
LCNTID = returned containing the subscript of the LCOUNT
array location which contains the number of
characters in the matched associated word.
The following arguments are used only for output. Their
input values are ignored.
INITAL = returned containing the subscript of the IBUFFR
array location which contains the first character
of an associated word (KIND returned as 4 or 11),
of a number (KIND returned as 5 or 12), or of an
unknown sequence of characters (KIND returned as 7
FASP, FORTRAN Alphameric Subroutine Package Page 184
DALEAD, Identify Command and Trailing Word, Number or String
or 14). If a quoted text string is found (KIND
returned containing 6 or 13), then INITAL is
returned pointing to the apostrophe at the left end
of the string.
IVALUE = returned containing the value if a number is found
(KIND being returned containing 5 or 12) and is
evaluated as an octal integer or as a decimal
integer.
VALUE = returned containing the value if a number is found
(KIND being returned containing 5 or 12) and is
evaluated as a decimal real.
IFLOAT = specifies how a number was evaluated. IFLOAT is
returned set equal to INITYP if the number is
outside the range of a command, or set equal to
NXTTYP if the number is within the range of a
command and NXTTYP has a value greater than -3, or
set equal to the entry in the NUMTYP array having
the same subscript as the KCOUNT array entry which
specifies the number of characters in the command
if NXTTYP=-3. IFLOAT is returned undefined, but
probably changed, if KIND is returned with a value
other than 5 or 12.
= -1, the number was evaluated as an octal integer
and its value is returned in IVALUE.
= 0, the number was evaluated as a decimal integer
and its value is returned in IVALUE.
= 1 or greater, the number was evaluated as a decimal
real and its value is returned in the argument
named VALUE.
FASP, FORTRAN Alphameric Subroutine Package Page 185
DALEAD, Identify Command and Trailing Word, Number or String
An Example of the Use of DALEAD
-- ------- -- --- --- -- ------
The program listed on the following pages reports the
results of the interpretation by DALEAD of a line of text
typed by the user. The user can type the commands LEADING,
TRAILING or BOTH followed by the associated words NONE,
OCTAL, DECIMAL or REAL to modify the interpretation of
numbers in the indicated positions. For example, if the
command TRAILING has been issued with the associated word
DECIMAL, then numbers associated with any command will be
evaluated as decimal integers. The command TRAILING can
also be followed by the associated word DEPENDENT to cause
the subsequent commands TEST NONE, TEST OCTAL, TEST DECIMAL,
and TEST REAL to take associated numbers of the types
indicated by the command names. For example, if the command
TRAILING has been issued with the associated word DEPENDENT,
then the command TEST NONE will not allow any associated
numbers, TEST OCTAL will accept only octal associated
numbers, and so forth. The demonstration program does not
take special action for the other words which it can
recognize. A typical dialog between the program and the
user is presented following the listing of the program.
C PROGRAM TO DEMONSTRATE DALEAD ROUTINE
C
DIMENSION KCOUNT(18),NUMTYP(18),KWORD(109)
DIMENSION LCOUNT(21),LWORD(85)
DIMENSION JAND(5),LGND1(11),LGND7(8),LGND15(7)
DIMENSION IBUFFR(60),JBUFFR(60)
C
C ARRAYS DEFINING DICTIONARY OF COMMANDS
C 1 LEADING 6 EXAMINE 11 EDIT
C 2 BOTH 7 LIST 12 TEST NONE
C 3 TRAILING 8 REMOVE 13 TEST OCTAL
C 4 CHANGE 9 RESTORE 14 TEST DECIMAL
C 5 MODIFY 10 NOTE 15 TEST REAL
C
DATA KCNTLO,KCNTHI,KWRDLO,KWRDHI/4,18,7,109/
DATA KCOUNT/ 99, 99, 99, 7, 4, 8, 6, 6, 7, 4,
1 6, 7, 4, 4,109,110,112,109/
DATA NUMTYP/ 0, 0, 0,-2,-2,-2, 0, 0, 0, 0,
1 0, 0, 0, 0,-2,-1, 0, 1/
DATA KWORD /1HI,1HG,1HN,1HO,1HR,1HE,1HL,1HE,1HA,1HD,
11HI,1HN,1HG,1HB,1HO,1HT,1HH,1HT,1HR,1HA,1HI,1HL,1HI,
21HN,1HG,1HC,1HH,1HA,1HN,1HG,1HE,1HM,1HO,1HD,1HI,1HF,
31HY,1HE,1HX,1HA,1HM,1HI,1HN,1HE,1HL,1HI,1HS,1HT,1HR,
41HE,1HM,1HO,1HV,1HE,1HR,1HE,1HS,1HT,1HO,1HR,1HE,1HN,
51HO,1HT,1HE,1HE,1HD,1HI,1HT,1HT,1HE,1HS,1HT,1H ,1HN,
61HO,1HN,1HE,1HT,1HE,1HS,1HT,1H ,1HO,1HC,1HT,1HA,1HL,
71HT,1HE,1HS,1HT,1H ,1HD,1HE,1HC,1HI,1HM,1HA,1HL,1HT,
81HE,1HS,1HT,1H ,1HR,1HE,1HA,1HL/
C
C ARRAYS DEFINING DICTIONARY OF ASSOCIATED WORDS
FASP, FORTRAN Alphameric Subroutine Package Page 186
DALEAD, Identify Command and Trailing Word, Number or String
C 1 DEPENDENT 6 SOME 11 TELL NOTHING
C 2 NONE 7 ALL 12 TRY NUMBER
C 3 OCTAL 8 FEW 13 NOT ALL
C 4 DECIMAL 9 MANY 14 TEST
C 5 REAL 10 MORE
C
DATA LCNTLO,LCNTHI,LWRDLO,LWRDHI/8,21,6,85/
DATA LCOUNT/ 99, 99, 99, 99, 99, 99, 99, 9, 4, 5,
1 7, 4, 4, 3, 3, 4, 4,112,110,107,
2 4/
DATA LWORD /1HD,1HU,1HM,1HM,1HY,1HD,1HE,1HP,1HE,1HN,
11HD,1HE,1HN,1HT,1HN,1HO,1HN,1HE,1HO,1HC,1HT,1HA,1HL,
21HD,1HE,1HC,1HI,1HM,1HA,1HL,1HR,1HE,1HA,1HL,1HS,1HO,
31HM,1HE,1HA,1HL,1HL,1HF,1HE,1HW,1HM,1HA,1HN,1HY,1HM,
41HO,1HR,1HE,1HT,1HE,1HL,1HL,1H ,1HN,1HO,1HT,1HH,1HI,
51HN,1HG,1HT,1HR,1HY,1H ,1HN,1HU,1HM,1HB,1HE,1HR,1HN,
61HO,1HT,1H ,1HA,1HL,1HL,1HT,1HE,1HS,1HT/
C
C ARRAYS USED TO GENERATE TEXT DESCRIPTIONS
DATA JAND/1H ,1HA,1HN,1HD,1H /
DATA LGND1/1HE,1HN,1HD,1H ,1HO,1HF,1H ,1HL,1HI,1HN,
11HE/
DATA LGND7/1HU,1HN,1HK,1HN,1HO,1HW,1HN,1H /
DATA LGND15/1HC,1HO,1HM,1HM,1HA,1HN,1HD/
DATA ILEFT,IRIGHT,IAND,IEND,IQUOTE/
11H(,1H),1H&,1H;,1H'/
C
C INITIAL VALUES
DATA ITTY,JTTY,MAXBFR,IFRCTN,INITYP,NXTTYP,ITRAIL,
1KIND/5,5,60,0,0,0,0,0/
C
C ASK USER FOR TEXT TO BE PROCESSED
1 WRITE(JTTY,2)
2 FORMAT(2H *,$)
READ(ITTY,3)IBUFFR
3 FORMAT(72A1)
LOWBFR=1
C
C FIND NEXT COMMAND AND ASSOCIATED WORD OR NUMBER
4 CALL DALEAD(INITYP,NXTTYP,ITRAIL,KWRDLO,KWRDHI,
1KWORD ,KCNTLO,KCNTHI,KCOUNT,NUMTYP,LWRDLO,LWRDHI,
2LWORD ,LCNTLO,LCNTHI,LCOUNT,IBUFFR,MAXBFR,LOWBFR,
3KIND ,KOMAND,KWRDID,KCNTID,LOCAL ,LWRDID,LCNTID,
4INITAL,IVALUE,VALUE ,IFLOAT)
C
C INSERT COMMAND WORD INTO TEXT TO BE PRINTED
INDEX=0
IF(KOMAND.EQ.0)GO TO 7
JBUFFR(INDEX+1)=ILEFT
IF(KIND.LE.7)INDEX=INDEX+1
J=KCOUNT(KCNTID)
IF(J.GE.100)J=J-100
J=KWRDID+J-1
DO 5 I=KWRDID,J
FASP, FORTRAN Alphameric Subroutine Package Page 187
DALEAD, Identify Command and Trailing Word, Number or String
INDEX=INDEX+1
5 JBUFFR(INDEX)=KWORD(I)
JBUFFR(INDEX+1)=IRIGHT
IF(KIND.LE.7)INDEX=INDEX+1
DO 6 I=1,5
INDEX=INDEX+1
6 JBUFFR(INDEX)=JAND(I)
C
C BRANCH TO CODE TO REPRESENT ASSOCIATED WORD OR NUMBER
7 NEW=KIND
IF(NEW.GT.7)NEW=NEW-7
GO TO(8,10,11,12,14,17,15,19),NEW
C
C END OF LINE
8 DO 9 I=1,11
INDEX=INDEX+1
9 JBUFFR(INDEX)=LGND1(I)
GO TO 21
C
C SEMICOLON
10 INDEX=INDEX+1
JBUFFR(INDEX)=IEND
GO TO 21
C
C AMPERSAND
11 INDEX=INDEX+1
JBUFFR(INDEX)=IAND
GO TO 21
C
C ASSOCIATED WORD
12 J=LCOUNT(LCNTID)
IF(J.GE.100)J=J-100
J=LWRDID+J-1
DO 13 I=LWRDID,J
INDEX=INDEX+1
13 JBUFFR(INDEX)=LWORD(I)
C
C CHECK IF DALEAD ARGUMENTS ARE TO BE CHANGED
IF(LOCAL.GT.5)GO TO 21
IF((KOMAND.EQ.1).OR.(KOMAND.EQ.2))INITYP=LOCAL-4
IF((KOMAND.EQ.2).OR.(KOMAND.EQ.3))NXTTYP=LOCAL-4
GO TO 21
C
C NUMBER
14 CALL DASHOW(IFLOAT,0,6,4,6,
13,IVALUE,VALUE,MAXBFR,INDEX,JBUFFR,IERR)
GO TO 21
C
C QUOTED TEXT STRING OR UNKNOWN CHARACTER SEQUENCE
15 DO 16 I=1,8
INDEX=INDEX+1
16 JBUFFR(INDEX)=LGND7(I)
17 J=LOWBFR-1
DO 18 I=INITAL,J
FASP, FORTRAN Alphameric Subroutine Package Page 188
DALEAD, Identify Command and Trailing Word, Number or String
INDEX=INDEX+1
18 JBUFFR(INDEX)=IBUFFR(I)
JBUFFR(INDEX+1)=IQUOTE
IF(NEW.EQ.6)INDEX=INDEX+1
GO TO 21
C
C FOLLOWING COMMAND
19 DO 20 I=1,7
INDEX=INDEX+1
20 JBUFFR(INDEX)=LGND15(I)
C
C REPORT RESULTS TO USER
21 WRITE(JTTY,22)(JBUFFR(I),I=1,INDEX)
22 FORMAT(2X,100A1)
IF(NEW.EQ.1)GO TO 1
IF(NEW.EQ.3)GO TO 1
GO TO 4
END
Typical Dialog Between DALEAD Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
*EXAM ALL!SIMPLE COMMAND AND ARGUMENT
EXAMINE AND ALL
(EXAMINE) AND END OF LINE
*EXAM ALL;!SEMICOLON
EXAMINE AND ALL
(EXAMINE) AND ;
END OF LINE
*EXAM ALL&ERSAND
EXAMINE AND ALL
(EXAMINE) AND &
*NONE LISTING SOMETHING UNKNOWN;;&UNKNOWN WORDS
(EXAMINE) AND NONE
(EXAMINE) AND UNKNOWN LISTING
(EXAMINE) AND UNKNOWN SOMETHING
(EXAMINE) AND UNKNOWN UNKNOWN
(EXAMINE) AND ;
;
&
*T N TE N T NO TEST TEL N TE NOT!AMBIGUOUS AND UNIQUE
UNKNOWN T N
UNKNOWN TE N
UNKNOWN T NO
TEST
TELL NOTHING
TELL NOTHING
END OF LINE
FASP, FORTRAN Alphameric Subroutine Package Page 189
DALEAD, Identify Command and Trailing Word, Number or String
*NOTE NOTA NON 12.34 ;ALL REAL 12.34 CHANGE SOM&CONTINUE
NOTE AND NOT ALL
(NOTE) AND NONE
(NOTE) AND 12
(NOTE) AND ;
ALL
REAL
12
CHANGE AND SOME
(CHANGE) AND &
*MORE 12.3 MANNY EXAMINE&CONTINUE WHEN NO ARGUMENT YET FOUND
(CHANGE) AND MORE
(CHANGE) AND 12
(CHANGE) AND UNKNOWN MANNY
EXAMINE AND &
*LIST,,MANY EDIT MODIFY!COMMAND AFTER COMMAND
EXAMINE AND COMMAND
LIST AND MANY
EDIT AND COMMAND
MODIFY AND END OF LINE
*0169.20 NOTE 0169.20 LEAD REAL;0169.20 NOTE 0169.20
169
NOTE AND 169
LEADING AND REAL
(LEADING) AND ;
169.2
NOTE AND 169
(NOTE) AND END OF LINE
*LEAD DEC TRAI REAL;0169.20 NOTE 0169.20 TRAIL DEPENDENT
LEADING AND DECIMAL
TRAILING AND REAL
(TRAILING) AND ;
169
NOTE AND 169.2
TRAILING AND DEPENDENT
(TRAILING) AND END OF LINE
*TEST NONE 0169.20 TEST OCTAL 0169.20
TEST NONE AND UNKNOWN 0169.20
TEST OCTAL AND 16
(TEST OCTAL) AND UNKNOWN 9.20
(TEST OCTAL) AND END OF LINE
*TEST DECIMAL 0169.20 TEST REAL 0169.20
TEST DECIMAL AND 169
TEST REAL AND 169.2
(TEST REAL) AND END OF LINE
FASP, FORTRAN Alphameric Subroutine Package Page 190
DALINE, Bar Chart Plotter for Printer
DDDDD AAA LL IIIIII NN NN EEEEEEEE
DD DD AAAA LL II NNN NN EE
DD DD AA AA LL II NNNN NN EE
DD DD AA AA LL II NN NN NN EEEEE
DD DD AAAAAAA LL II NN NNNN EE
DD DD AA AA LL II NN NNN EE
DDDDD AA AA LLLLLLLL IIIIII NN NN EEEEEEEE
DALINE, Bar Chart Plotter for Printer
------ --- ----- ------- --- -------
DALINE is a FORTRAN subroutine which constructs printable
plots containing horizontal bars formed of segments the
lengths of which represent the magnitudes of the
corresponding values. The segments can each be represented
with a different character. Either the entire lengths of
the segments, or merely their tips, can be represented. The
bars can be superimposed upon a background grid and can
extend to either side of a central column corresponding to
zero. Negative segments (debits) are accumulated to one
side of the zero column, and positive segments (credits) to
the other, although the user has the option of suppressing
the display of either portion. DALINE is called as many
times as there are lines in the plot, each subsequent call
generating the next lower line of the plot. Either a number
or an alphabetic label can be printed to the left of each
line of the plot. Tick marks and scale numbers can be
generated below the plot when the final bar is plotted, or
can be added afterwards.
The DALINE Argument List
--- ------ -------- ----
The argument list of routine DALINE is
SUBROUTINE DALINE(IGRID ,KNDBAR,LSTLIN,MARGIN,MSHLFT,
1 MSHRIT,LNGLFT,LNGRIT,XLEFT ,XRIGHT,SEGMNT,MINSEG,
2 MAXSEG,LETTER,MINLTR,MAXLTR,YVALUE,LABEL ,MINLBL,
3 MAXLBL,LINPRT,IDISK )
with the associated DIMENSION statement
DIMENSION SEGMNT(MAXSEG),LETTER(MAXLTR),LABEL(MAXLBL)
All of the arguments of this routine are used only for input
and are returned unchanged.
IGRID = selects the background characters to be shown where
not hidden by the characters used to represent the
bar segments, selects the character, if any, to
appear in the central zero column, and selects the
type of caption, either numeric or alphabetic, to
be shown to the left of the currently generated
FASP, FORTRAN Alphameric Subroutine Package Page 191
DALINE, Bar Chart Plotter for Printer
panel of the bar chart. These options are selected
by the digits in the ones, tens, and hundreds
positions respectively in the decimal integer value
of IGRID.
If the bar chart is to be superimposed upon a
rectangular grid formed of horizontal lines
represented by minus signs and of vertical lines
represented by exclamation points, then IGRID would
be assigned 2 different values, the first value
selecting exclamation points in the columns which
are to ruled with vertical lines and spaces
elsewhere in each panel which is not to be ruled
with a horizontal grid line, and a second value
selecting plus signs at the grid line intersections
and minus signs elsewhere in each panel which is to
be ruled with a horizontal grid line. If a
rectangular grid is not to be included in the plot
but border lines are desired, then IGRID would have
a value selecting horizontal grid lines in the top
and bottom panels, and another value selecting only
the left and right grid lines in the intervening
panels.
The meanings assigned to each position in the value
of IGRID are as follow
ONES DIGIT
selects the characters to be included across the
width of the plot in columns not occupied by the
characters used to represent the bar segments, but
does not dictate the characters appearing in either
the left margin or in the column representing zero.
0 spaces will be shown where not hidden by the
characters used to represent the bar segments.
1 minus signs will be shown where not hidden by
the characters used to represent the bar
segments.
2 if not hidden by the characters used to
represent the bar segments, then exclamation
points will be shown in the leftmost column
which could bear a lower scale number to the
left of the zero column, and in the rightmost
column which could bear a lower scale number
to the right of the zero column. Spaces will
be shown in the remaining columns where not
hidden by the characters used to represent the
bar segments.
3 same as IGRID=2, except that minus signs are
included instead of spaces in the columns
which contain neither exclamation points nor
the characters which represent the bar
FASP, FORTRAN Alphameric Subroutine Package Page 192
DALINE, Bar Chart Plotter for Printer
segments.
4 same as IGRID=2, except that if not hidden by
the characters used to represent the bar
segments, then plus signs will be shown in the
leftmost column which could bear a lower scale
number to the left of the zero column, and in
the rightmost column which could bear a lower
scale number to the right of the zero column.
5 same as IGRID=4, except that minus signs are
included instead of spaces in the columns
which contain neither plus signs nor the
characters which represent the bar segments.
6 if not hidden by the characters used to
represent the bar segments, then exclamation
points will be shown in each column which
could bear a lower scale number. Spaces will
be shown in the remaining columns where not
hidden by the characters used to represent the
bar segments.
7 same as IGRID=6, except that minus signs are
included instead of spaces in the columns
which contain neither exclamation points nor
the characters which represent the bar
segments.
8 same as IGRID=6, except that if not hidden by
the characters used to represent the bar
segments, then plus signs will be shown in
each column which could bear a lower scale
number.
9 same as IGRID=8, except that minus signs are
included instead of spaces in the columns
which contain neither plus signs nor the
characters which represent the bar segments.
The following examples demonstrate the results
produced by the various values of the ones digit.
ones digit = 0
= 1 --------------- ---------------
= 2 ! !
= 3 !-------------- --------------!
= 4 + +
= 5 +-------------- --------------+
= 6 ! ! ! ! ! !
= 7 !----!----!---- ----!----!----!
= 8 + + + + + +
= 9 +----+----+---- ----+----+----+
tick marks ! ! ! ! ! ! !
scale numbers -15 -10 -5 0 5 10 15
TENS DIGIT
selects the character, if any, to be shown in the
zero column, and whether the lower scale numbers,
FASP, FORTRAN Alphameric Subroutine Package Page 193
DALINE, Bar Chart Plotter for Printer
if generated by the current call to this routine,
are to include the identification of the zero
column.
0 the zero column is not to be included within
the representation of the current panel. If
the tens digit has the value zero for any
panel of the plot, then it must have the value
zero for all panels of the plot or else the
bar segments to the right of the zero column
will be improperly aligned.
1 the zero column is to be represented by the
space character, but is not to be identified
in the lower scale numbers which might be
generated by the current call to this routine.
2 the zero column is to be represented by the
minus sign, but is not to be identified in the
lower scale numbers which might be generated
by the current call to this routine.
3 the zero column is to be represented by the
exclamation point, but is not to be identified
in the lower scale numbers which might be
generated by the current call to this routine.
4 the zero column is to be represented by the
plus sign, but is not to be identified in the
lower scale numbers which might be generated
by the current call to this routine.
5 the zero column is to be represented by the
space character. If the current call to this
routine also generates the lower scale
numbers, then the zero column will be
identified by a scale number having the value
zero.
6 the zero column is to be represented by the
minus sign. If the current call to this
routine also generates the lower scale
numbers, then the zero column will be
identified by a scale number having the value
zero.
7 the zero column is to be represented by the
exclamation point. If the current call to
this routine also generates the lower scale
numbers, then the zero column will be
identified by a scale number having the value
zero.
8 the zero column is to be represented by the
plus sign. If the current call to this
routine also generates the lower scale
numbers, then the zero column will be
identified by a scale number having the value
zero.
The following examples demonstrate the results
produced by the various values of the tens digit.
FASP, FORTRAN Alphameric Subroutine Package Page 194
DALINE, Bar Chart Plotter for Printer
tens digit = 0
+-+--+-+
! ! ! !
-4-2 2 4
tens digit = 1 2 3 4
+-+- -+-+ +-+---+-+ +-+-!-+-+ +-+-+-+-+
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4
tens digit = 5 6 7 8
+-+- -+-+ +-+---+-+ +-+-!-+-+ +-+-+-+-+
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-4-2 0 2 4 -4-2 0 2 4 -4-2 0 2 4 -4-2 0 2 4
HUNDREDS DIGIT
selects the type of caption, either numeric or
alphabetic, to be shown to the left of the
currently generated panel of the bar chart.
0 the left margin is to contain spaces.
1 the representation of the value input as the
argument named YVALUE is to be right justified
in the margin to the left of the left border
of the bar chart.
2 the characters input in LABEL(MINLBL) through
and including LABEL(MAXLBL) are to be right
justified in the margin to the left of the
left border of the bar chart.
3 a space character is to be placed into the
column to the immediate left of the left
border of the bar chart, and the
representation of the value input as the
argument named YVALUE is to be right justified
in the remaining portion of the left margin.
4 a space character is to be placed into the
column to the immediate left of the left
border of the bar chart, and the characters
input in LABEL(MINLBL) through and including
LABEL(MAXLBL) are to be right justified in the
remaining portion of the left margin.
5 a minus sign is to be placed into the column
to the immediate left of the left border of
the bar chart, and the representation of the
value input as the argument named YVALUE is to
be right justified in the remaining portion of
the left margin.
6 a minus sign is to be placed into the column
to the immediate left of the left border of
the bar chart, and the characters input in
FASP, FORTRAN Alphameric Subroutine Package Page 195
DALINE, Bar Chart Plotter for Printer
LABEL(MINLBL) through and including
LABEL(MAXLBL) are to be right justified in the
remaining portion of the left margin.
The following examples demonstrate the results
produced by the various values of the hundreds
digit.
hundreds digit = 0 +----+---- ----+----+
= 1 12.34+----+---- ----+----+
= 2 LABEL+----+---- ----+----+
= 3 12.34 +----+---- ----+----+
= 4 LABEL +----+---- ----+----+
= 5 12.34-+----+---- ----+----+
= 6 LABEL-+----+---- ----+----+
tick marks ! ! ! !
scale numbers -10 -5 5 10
KNDBAR = specifies which bar segment is to be represented if
more than one bar segment appears in a single
column of the plot, and which columns are to
represent any bar segment which extends across more
than a single column.
= -1, no bar segments are to be plotted in the
current panel. The resulting panel is identical to
that which would be produced if MINSEG is set
greater than MAXSEG.
= 0, each bar segment is to be represented only in
the final column (that furthest from the zero
column) in which it is the largest component.
Spaces are inserted into any other columns in which
the bar segment is dominant.
= 1, each bar segment is to be represented only in
the final column (that furthest from the zero
column) in which it is the largest component. The
grid characters selected by the ones digit of the
decimal integer value of IGRID are inserted into
any other columns in which the bar segment is
dominant.
= 2, each bar segment is to be represented in all
columns in which it is the largest component.
= 3, 4 or 5, same as KIND=0, 1 or 2 respectively,
except that the final columns (those furthest to
the left and to the right of the zero column) which
contain the final negative and the final positive
values specified by the SEGMNT array instead
contain the grid characters selected by the ones
digit of the decimal integer value of IGRID if the
largest components within these columns are smaller
than the space remaining in these columns.
= 6, each bar segment is to be represented only in
the final column (that furthest from the zero
column) in which the bar segment appears. If two
or more bar segments terminate in the same column,
FASP, FORTRAN Alphameric Subroutine Package Page 196
DALINE, Bar Chart Plotter for Printer
then the bar segment specified by the SEGMNT array
location having the higher subscript is
represented, and the bar segment specified by the
the SEGMENT array location having the lower
subscript is instead represented in the adjacent
column next closer to the zero column if this lower
bar segment is the only bar segment which appears
in this adjacent column. Spaces are inserted into
any other columns across which a bar segment
extends.
= 7, each bar segment is to be represented only in
the final column (that furthest from the zero
column) in which the bar segment appears. If two
or more bar segments terminate in the same column,
then the bar segment specified by the SEGMNT array
location having the higher subscript is
represented, and the bar segment specified by the
the SEGMENT array location having the lower
subscript is instead represented in the adjacent
column next closer to the zero column if this lower
bar segment is the only bar segment which appears
in this adjacent column. The grid characters
selected by the ones digit of the decimal integer
value of IGRID are inserted into any other columns
which do not contain the termination of a bar
segment.
= 8, each bar segment is represented in the final
column (that furthest from the zero column) in
which the bar segment appears, and in each column
in which the bar segment is the only component.
As a demonstration of the plotting modifications
caused by the various values of KNDBAR, the numbers
.3 .4 .2 3.5 .1 .3 3.5 .2 .1 3.5 .3 .2
and
-.3 -.4 -.2 -3.5 -.1 -.3 -3.5 -.2 -.1 -3.5 -.3 -.2
are plotted below represented by the letters A
through X respectively with LNGLFT=LNGRIT=15,
XLEFT=-15 and XRIGHT=15 for all possible values of
KNDBAR
FASP, FORTRAN Alphameric Subroutine Package Page 197
DALINE, Bar Chart Plotter for Printer
KNDBAR =-1 +----+----+----+----+----+----+
= 0 +-WV S P N+B D G JK-+
= 1 +-WV-+-S--P---N+B---D--G-+-JK-+
= 2 +-WVVVVSSSPPPPN+BDDDDGGGJJJJK-+
= 3 +--V S P N+B D G J--+
= 4 +--V-+-S--P---N+B---D--G-+-J--+
= 5 +--VVVVSSSPPPPN+BDDDDGGGJJJJ--+
= 6 +-XV US RP O+C DF GI JL-+
= 7 +-XV-+US--RP--O+C--DF--GI+-JL-+
= 8 +-XVVVUSSSRPPPO+CDDDFGGGIJJJL-+
tick marks ! ! ! ! ! ! !
scale numbers -15 -10 -5 0 5 10 15
The letters which appear in columns 1, 5, 9 and 13
when KNDBAR is less than or equal to 5 are
different than whose which appear when KNDBAR is
greater than or equal to 6. The correspondence of
the letters to the columns is shown in the chart
below.
columns columns
column 1 ! 2 - 4 ! column 5 ! 6 - 8 !
! ! ! !
.3A .4B .2C .1D ! 3.D ! .4D .1E .3F .2G ! 3.G !
columns
column 9 !10 - 12! column 13
! !
.3G .2H .1I .4J ! 3.J ! .1J .3K .2L and .4 space
For example, when KNDBAR is less than or equal to
5, column 5 contains the letter D which represents
the largest component in column 5. When KNDBAR is
greater than or equal to 6, column 5 instead
contains the letter F which represents the highest
subscripted component which terminates in this
column, and, since the component represented by the
letter F does not appear in column 4, the letter D
is forced to appear in column 4.
LSTLIN = -1, only tick marks and lower scale numbers are to
be generated, but a panel of the bar chart will not
be generated. This option might be used to add
tick marks and scale numbers to a bar chart after
it has been completed if it is not possible to
determine whether a particular panel will terminate
the bar chart before the panel is to be generated.
= 0, the current panel finishes the bar chart. Scale
numbers ranging in value from that of XLEFT (if
LNGLFT is greater than zero) through that of XRIGHT
(if LNGRIT is greater than zero) are to be written
below the current panel of the bar chart.
= greater than zero, the current panel does not
finish the bar chart. Scale numbers are not to be
FASP, FORTRAN Alphameric Subroutine Package Page 198
DALINE, Bar Chart Plotter for Printer
written below the current panel of the bar chart.
Subsequent calls to DALINE will add additional
panels to the bar chart. The value of LSTLIN is
ignored other than to determine whether it is
greater than zero. If the main program knows the
total number of time periods, it can count LSTLIN
down to zero. If the main program does not know
the total number of time periods, it is sufficient
to set LSTLIN to 1 until the final panel.
MARGIN = number of characters which must appear in the
output between the carriage control character in
column 1 and the left edge of the current panel of
the bar chart. The width of this margin must be
the same for all panels forming the bar chart. The
margin must include sufficient room for the left
scale number or for the left alphameric label
requested by a nonzero value of the hundreds digit
of the decimal integer value of IGRID, and for the
extra space or for the extra minus sign if the
hundreds digit of IGRID is greater than or equal to
3. If scale numbers are to be written below the
bar chart, and if LNGLFT is greater than zero, then
MARGIN should have a value of at least 5 (or of at
least half of MSHLFT if MSHLFT is less than 10) to
allow a number to be centered below the left border
line. If LNGLFT is not greater than zero, then
room does not need to be allowed in the left margin
for the lower scale numbers.
MSHLFT = width of the grid divisions indicated either by
exclamation points or plus signs in the portion of
the bar chart to the left of the zero column. If
MSHLFT is equal to zero, then a grid division width
of 10 is assumed. MSHLFT is equal to one more than
the number of columns of characters appearing
between the exclamation points or plus signs. If
LSTLIN is less than or equal to zero, then scale
numbers are written below the bar chart every
MSHLFT columns to the left of the zero column.
MSHRIT = width of the grid divisions indicated either by
exclamation points or plus signs in the portion of
the bar chart to the right of the zero column. If
MSHRIT is equal to zero, then a grid division width
of 10 is assumed. MSHRIT is equal to one more than
the number of columns of characters appearing
between the exclamation points or plus signs. If
LSTLIN is less than or equal to zero, then scale
numbers are written below the bar chart every
MSHRIT columns to the right of the zero column.
LNGLFT = 0, the bar chart is not to include any columns to
the left of the zero column. This does not change
FASP, FORTRAN Alphameric Subroutine Package Page 199
DALINE, Bar Chart Plotter for Printer
the width of the left margin specified by the value
of MARGIN. The value of XLEFT is ignored.
= greater than zero, LNGLFT is the number of columns
of characters to be included in the bar chart to
the left of the zero column. The portion of the
bar chart to the left of the zero column will
represent negative segments if XLEFT is less than
zero, or positive segments if XLEFT is greater than
zero. The left border of the bar chart will
represent an accumulated bar length equal to the
value of XLEFT. If requested by the ones digit of
the decimal integer value of IGRID, a vertical grid
line will be indicated every MSHLFT columns to the
left of the zero column.
LNGRIT = 0, the bar chart is not to include any columns to
the right of the zero column. The value of XRIGHT
is ignored.
= greater than zero, LNGRIT is the number of columns
of characters to be included in the bar chart to
the right of the zero column. The portion of the
bar chart to the right of the zero column will
represent negative segments if XRIGHT is less than
zero, or positive segments if XRIGHT is greater
than zero. The right border of the bar chart will
represent an accumulated bar length equal to the
value of XRIGHT. If requested by the ones digit of
the decimal integer value of IGRID, a vertical grid
line will be indicated every MSHRIT columns to the
right of the zero column.
XLEFT = the accumulated total of SEGMNT values having the
same sign as XLEFT which is to be represented by
the left border of the portion of the bar chart to
the left of the zero column if LNGLFT is greater
than zero. If XLEFT is negative, then only
negative values are represented to the left of the
zero column. IF XLEFT is positive, then only
positive values are represented to the left of the
zero column. The magnitude of XLEFT should depend
upon the magnitudes of the values in the SEGMNT
array. If XLEFT has a value equal to -100 times
the number of columns specified by LNGLFT, then the
first column to the left of the zero column would
represent an accumulated total in the range -100 to
0, the second column to the left of the zero column
would represent an accumulated total in the range
-200 to -100 and so on.
XRIGHT = the accumulated total of SEGMNT values having the
same sign as XRIGHT which is to be represented by
the right border of the portion of the bar chart to
the right of the zero column if LNGRIT is greater
than zero. If XRIGHT is negative, then only
FASP, FORTRAN Alphameric Subroutine Package Page 200
DALINE, Bar Chart Plotter for Printer
negative values are represented to the right of the
zero column. IF XRIGHT is positive, then only
positive values are represented to the right of the
zero column. The magnitude of XRIGHT should depend
upon the magnitudes of the values in the SEGMNT
array. If XRIGHT has a value equal to 100 times
the number of columns specified by LNGRIT, then the
first column to the right of the zero column would
represent an accumulated total greater than zero
but not greater than 100, the second column to the
right of the zero column would represent an
accumulated total greater than 100 but not greater
than 200 and so on.
SEGMNT = array containing in the array locations
SEGMNT(MINSEG) through and including SEGMNT(MAXSEG)
the lengths of the bar segments specified in the
same coordinate systems as used for the definitions
of XLEFT and XRIGHT. Zero values in the SEGMNT
array are ignored.
MINSEG = subscript of the first location in the SEGMNT array
which can specify the length of a bar segment. If
MINSEG is greater than MAXSEG, then no bar segments
will be represented.
MAXSEG = subscript of the final location in the SEGMNT array
which can specify the length of a bar segment.
LETTER = array containing in the array locations
LETTER(MINLTR) through and including LETTER(MAXLTR)
the characters by which the bar segments are to be
represented, one character per LETTER array
location as though read by a multiple of an A1
format or defined by several 1H fields in a DATA
statement. The segment having its length specified
by SEGMNT(MINSEG) will be represented by the
character found in LETTER(MINLTR), that having its
length in SEGMNT(MINSEG+1) by the character in
LETTER(MINLTR+1), and so on. If MAXLTR-MINLTR is
less than MAXSEG-MINSEG, then the assignment of
characters recycles through the LETTER array as
many times as are necessary to represent all of the
bar segments so that the segment having its length
specified by SEGMNT(MINSEG+MAXLTR-MINLTR+1) is
represented by the character in LETTER(MINLTR),
that having its length in
SEGMNT(MINSEG+MAXLTR-MINLTR+2) by the character in
LETTER(MINLTR+1), and so on.
MINLTR = subscript of the LETTER array location which
specifies the character to be used for the
representation of the bar segment having its length
in SEGMNT(MINSEG).
FASP, FORTRAN Alphameric Subroutine Package Page 201
DALINE, Bar Chart Plotter for Printer
MAXLTR = subscript of the LETTER array location which
specifies the character to be used for the
representation of the bar segment having its length
in SEGMNT(MINSEG+MAXLTR-MINLTR).
YVALUE = number to be placed to the left of the bar chart if
the hundreds digit of the decimal integer value of
IGRID has the value 1, 3 or 5.
LABEL = array containing in the array locations
LABEL(MINLBL) through and including LABEL(MAXLBL)
the characters to be shown to the left of the bar
chart if the hundreds digit of the decimal integer
value of IGRID has the value 2, 4 or 6. The
characters in the LABEL array are stored one
character per array location as though read by a
multiple of an A1 format or defined by several 1H
fields in a DATA statement. If more characters are
specified in the LABEL array than can be shown in
the left margin, then the rightmost excess
characters are not shown.
MINLBL = subscript of the LABEL array location containing
the leftmost character to be shown to the left of
the bar chart if the hundreds digit of the decimal
integer value of IGRID has the value 2, 4 or 6.
MAXLBL = subscript of the LABEL array location containing
the rightmost character to be shown to the left of
the bar chart if the hundreds digit of the decimal
integer value of IGRID has the value 2, 4 or 6.
LINPRT = -1, do not include a carriage control character to
the left of each line of the plot. Since the minus
sign of a negative scale number or the left
character of a label can then appear in column 1,
the resulting output must not be treated as though
the left column contains carriage control
characters.
= 0, the plot will be viewed by the user on a
terminal, either typed directly with IDISK being
given the terminal unit number, or typed by the
user after this routine has written the plot into a
file on the unit the number of which is contained
in IDISK. A blank or space will be used as
carriage control character to give single spacing.
= 1, the plot will be printed on the line printer by
the user after the program has written the plot
into a file. An asterisk will be used as carriage
control character to give single spacing with
suppression of skipping extra lines at the page
boundaries. On the PDP-10, an asterisk as the
carriage control character gives overprinting on
the terminal as opposed to single spacing.
FASP, FORTRAN Alphameric Subroutine Package Page 202
DALINE, Bar Chart Plotter for Printer
IDISK = the unit number of the device onto which the plots
are to be written. This routine will only generate
the plot. It is the responsibility of the calling
program to open the output file and to write the
captions, the form feeds and/or the separating
lines.
FASP, FORTRAN Alphameric Subroutine Package Page 203
DALINE, Bar Chart Plotter for Printer
Examples of Grid Rulings Selected by Various IGRID Values
-------- -- ---- ------- -------- -- ------- ----- ------
In each illustration, a positive segment is represented by
the character P and a negative segment by the character N.
IGRID=500 501 502 503 504
1- NP 1----NP--- 1-! NP ! 1-!--NP--! 1-+ NP +
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4
IGRID=505 506 507 508 509
1-+--NP--+ 1-! !NP! ! 1-!-!NP!-! 1-+ +NP+ + 1-+-+NP+-+
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4
IGRID=510 511 512 513 514
1- N P 1----N P--- 1-! N P ! 1-!--N P--! 1-+ N P +
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4
IGRID=515 516 517 518 519
1-+--N P--+ 1-! !N P! ! 1-!-!N P!-! 1-+ +N P+ + 1-+-+N P+-+
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4
IGRID=540 541 542 543 544
1- N+P 1----N+P--- 1-! N+P ! 1-!--N+P--! 1-+ N+P +
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4
IGRID=545 546 547 548 549
1-+--N+P--+ 1-! !N+P! ! 1-!-!N+P!-! 1-+ +N+P+ + 1-+-+N+P+-+
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4 -4-2 2 4
IGRID=580 581 582 583 584
1- N+P 1----N+P--- 1-! N+P ! 1-!--N+P--! 1-+ N+P +
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-4-2 0 2 4 -4-2 0 2 4 -4-2 0 2 4 -4-2 0 2 4 -4-2 0 2 4
IGRID=585 586 587 588 589
1-+--N+P--+ 1-! !N+P! ! 1-!-!N+P!-! 1-+ +N+P+ + 1-+-+N+P+-+
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-4-2 0 2 4 -4-2 0 2 4 -4-2 0 2 4 -4-2 0 2 4 -4-2 0 2 4
FASP, FORTRAN Alphameric Subroutine Package Page 204
DALINE, Bar Chart Plotter for Printer
An Example of the Use of DALINE
-- ------- -- --- --- -- ------
The following program was used to generate the 3 plots shown
on the next page. The major portion of the program defines
the values of IGRID and KNDBAR needed for the production in
the first plot of the hollow bars separated by blank lines.
DIMENSION SEGMNT(7),LETTER(7),VALUES(7,6),LABEL(1)
DATA ((VALUES(I,J),I=1,7),J=1,6)/
1-2.,-4.,-2., 3., 8., 2., 1.,
2-2.,-3.,-3., 4., 7., 2., 3.,
3-1.,-2.,-4., 9., 6., 2., 7.,
4 0.,-2.,-1., 6., 5., 3., 5.,
5 0.,-2.,-6., 5., 5., 4., 3.,
6 0.,-2.,-8., 4., 9.,10., 5./
DATA LETTER/1HA,1HB,1HC,1HD,1HE,1HF,1HG/
DATA IDISK/1/
LSTLIN=1
DO 9 LINE=1,6
YVALUE=LINE
DO 1 I=1,7
1 SEGMNT(I)=VALUES(I,LINE)
DO 9 INNER=1,7
GO TO(2,3,4,5,4,6,7),INNER
2 IF(LINE.NE.1)GO TO 9
KNDBAR=-1
IGRID=49
GO TO 8
3 IF(LINE.NE.1)GO TO 9
IGRID=36
GO TO 8
4 IGRID=36
KNDBAR=2
GO TO 8
5 IGRID=549
KNDBAR=0
GO TO 8
6 KNDBAR=-1
IGRID=36
GO TO 8
7 IF(LINE.NE.6)GO TO 9
LSTLIN=0
IGRID=49
8 CALL DALINE(IGRID,KNDBAR,LSTLIN,10,10,
110,10,30,-10.,30.,SEGMNT,1,
27,LETTER,1,7,YVALUE,LABEL,2,
31,0,IDISK)
9 CONTINUE
DO 12 KNDBAR=1,2
WRITE(IDISK,10)
10 FORMAT(1X)
DO 12 LINE=1,6
DO 11 I=1,7
FASP, FORTRAN Alphameric Subroutine Package Page 205
DALINE, Bar Chart Plotter for Printer
11 SEGMNT(I)=VALUES(I,LINE)
12 CALL DALINE(176,KNDBAR,6-LINE,10,10,
110,10,30,-10.,30.,SEGMNT,1,
27,LETTER,1,7,FLOAT(LINE),LABEL,2,
31,0,IDISK)
STOP
END
+---------+---------+---------+---------+
! ! ! ! !
! CCBBBBAA!DDDEEEEEEEEFFG ! !
1-+-C B A + D E FG-----+---------+
! CCBBBBAA!DDDEEEEEEEEFFG ! !
! ! ! ! !
! CCCBBBAA!DDDDEEEEEEEFFGGG ! !
2-+-C B A + D E F G---+---------+
! CCCBBBAA!DDDDEEEEEEEFFGGG ! !
! ! ! ! !
! CCCCBBA!DDDDDDDDDEEEEEEFFGGGGGGG !
3-+--C B A+ D E F G-----+
! CCCCBBA!DDDDDDDDDEEEEEEFFGGGGGGG !
! ! ! ! !
! CBB!DDDDDDEEEEEFFFGGGGG! !
4-+------CB + D E F G+---------+
! CBB!DDDDDDEEEEEFFFGGGGG! !
! ! ! ! !
! CCCCCCBB!DDDDDEEEEEFFFFGGG ! !
5-+-C B + D E F G--+---------+
! CCCCCCBB!DDDDDEEEEEFFFFGGG ! !
! ! ! ! !
CCCCCCCCBB!DDDDEEEEEEEEEFFFFFFFFFFGGGGG !
6-C B + D E F G-+
CCCCCCCCBB!DDDDEEEEEEEEEFFFFFFFFFFGGGGG !
! ! ! ! !
+---------+---------+---------+---------+
! ! ! !
-10 10 20 30
1! C B A ! D !E FG ! !
2! C B A ! D !E F G ! !
3! C B A! D! E F ! G !
4! CB ! D !E F G! !
5! C B ! D E F G ! !
6C B ! D ! E ! F G !
! ! ! ! !
-10 0 10 20 30
1! CCBBBBAA!DDDEEEEEEEEFFG ! !
2! CCCBBBAA!DDDDEEEEEEEFFGGG ! !
3! CCCCBBA!DDDDDDDDDEEEEEEFFGGGGGGG !
4! CBB!DDDDDDEEEEEFFFGGGGG! !
5! CCCCCCBB!DDDDDEEEEEFFFFGGG ! !
6CCCCCCCCBB!DDDDEEEEEEEEEFFFFFFFFFFGGGGG !
! ! ! ! !
-10 0 10 20 30
FASP, FORTRAN Alphameric Subroutine Package Page 206
DALIST, Routine to List Dictionary Constructed by DALOAD
DDDDD AAA LL IIIIII SSSSSS TTTTTTTT
DD DD AAAA LL II SS TT
DD DD AA AA LL II SS TT
DD DD AA AA LL II SSSS TT
DD DD AAAAAAA LL II SS TT
DD DD AA AA LL II SS TT
DDDDD AA AA LLLLLLLL IIIIII SSSSSS TT
DALIST, Routine to List Dictionary Constructed by DALOAD
------ ------- -- ---- ---------- ----------- -- ------
DALIST summarizes the array names and subscript ranges
specified in the dictionary constructed by the DALOAD
routine. DALIST is used along with several other routines
in the FASP package for the purpose of specifying by name,
examining and modifying the values of multiply subscripted
arrays equivalenced with or otherwise loaded into a singly
subscripted buffer. The interaction between these routines
is described at the start of the DALOAD documentation. The
sample program at the end of the DALOAD documentation
illustrates the use of most of these routines.
The DALIST Argument List
--- ------ -------- ----
The argument list of routine DALIST is
SUBROUTINE DALIST(JTTY ,LTRLOW,LTRUSD,LTRSTR,NUMLOW,
1NUMUSD,NUMSTR,NAMMAX,NAME)
with the associated DIMENSION statement
DIMENSION LTRSTR(LTRUSD),NUMSTR(NUMUSD),NAME(NAMMAX)
The following arguments are used for input only and are
returned unchanged.
JTTY = number of the unit onto which the description of
the contents of the dictionary is to be written.
This could be the number associated with a terminal
or be the number of the unit to which a file is to
be written. If the number is that of the unit to
which a file is to be written, then the calling
program must open the file before calling this
routine, and close it, if necessary, afterwards.
LTRLOW = lowest subscript of the locations in the LTRSTR
array containing the characters of the names in the
dictionary as originally read by DALOAD using a
multiple of an A1 format. LTRSTR(LTRLOW) contains
FASP, FORTRAN Alphameric Subroutine Package Page 207
DALIST, Routine to List Dictionary Constructed by DALOAD
either the first letter of the name of the first
logical group of names in the dictionary or else
(if the first group itself isn't named) the first
letter of the first name within the first logical
group in the dictionary.
LTRUSD = highest subscript of the locations in the LTRSTR
array containing the characters of the names in the
dictionary as originally read by DALOAD using a
multiple of an A1 format. LTRSTR(LTRUSD) contains
the last character of the last name in the
dictionary.
LTRSTR = array containing the characters forming the names
in the dictionary, 1 character per array location
as originally read by DALOAD using a multiple of an
A1 format.
NUMLOW = lowest subscript of the locations in the NUMSTR
array containing the numeric information
corresponding to the names stored in the LTRSTR
array. NUMSTR(NUMLOW) must contain the start of
the description of a logical group of names, not
the start of the description of an individual name.
NUMUSD = highest subscript of the locations in the NUMSTR
array containing the numeric information
corresponding to the names stored in the LTRSTR
array.
NUMSTR = array containing the numeric information
corresponding to the names stored in the LTRSTR
array. The construction of the NUMSTR array is
described in detail in the DALOAD documentation.
For each name in the dictionary, the NUMSTR array
contains
a. the number of characters in the name
b. an indication of the associated data type
c. the number of subscript ranges
d. pairs of starting and ending values of these
ranges.
If the number of characters is instead zero or
negative, then its absolute value is the number of
characters in the name of a logical group of names,
and the next location, rather than indicating the
data type, contains the number of locations within
a singly subscripted buffer which would be needed
to store the values of the multiply subscripted
arrays which are within the logical group and
FASP, FORTRAN Alphameric Subroutine Package Page 208
DALIST, Routine to List Dictionary Constructed by DALOAD
equivalenced with or otherwise loaded into such a
singly subscripted buffer.
NAMMAX = highest subscript of any location in the NAME array
which can be used by this routine for the temporary
storage of characters which are to be included in
the summary of the dictionary contents. NAMMAX is
the maximum width of any line which can be included
in the summary. The contents of NAME(1) through
NAME(NAMMAX) will be destroyed by this routine.
The following argument is used internally by this routine.
Its original contents are destroyed.
NAME = array used internally by this routine for the
construction of the lines of text forming the
summary of the contents of the dictionary.
FASP, FORTRAN Alphameric Subroutine Package Page 209
DALOAD, Routine to Make Dictionary Describing Named Arrays
DDDDD AAA LL OOOOO AAA DDDDD
DD DD AAAA LL OO OO AAAA DD DD
DD DD AA AA LL OO OO AA AA DD DD
DD DD AA AA LL OO OO AA AA DD DD
DD DD AAAAAAA LL OO OO AAAAAAA DD DD
DD DD AA AA LL OO OO AA AA DD DD
DDDDD AA AA LLLLLLLL OOOOO AA AA DDDDD
DALOAD, Routine to Make Dictionary Describing Named Arrays
------ ------- -- ---- ---------- ---------- ----- ------
DALOAD reads the DIMENSION, COMMON, INTEGER, REAL and
IMPLICIT statements which could be used to declare
subscripted arrays in a FORTRAN program, and from the
information in these statements constructs a dictionary
containing the names, data types, and subscript limits of
the arrays. The calling program must open the input file
containing the FORTRAN statements. DALOAD is then called to
read the statements from this input file, returning control
only after an END statement or an end-of-file is read or if
the space allocated for the dictionary fills before the
dictionary is complete.
The dictionary constructed by DALOAD can be used by several
other routines in the FASP package for the purpose of
specifying by name, examining and modifying the values of
multiply subscripted arrays equivalenced with or otherwise
loaded into a singly subscripted buffer. The sample program
at the end of the DALOAD documentation illustrates the use
of most of these routines. A summary of the dictionary can
be typed by the DALIST routine. Array names typed by the
user and read by the calling program into an input buffer
with a multiple of an A1 format can be matched against
entries in the dictionary by the DAPICK routine. The DASITE
routine uses the dictionary and the subscripts typed by the
user as evaluated by DAPICK to locate the desired value of
the multiply subscripted array which is being simulated
within the singly subscripted buffer. The DALONE routine
constructs in a third buffer (the first two being the buffer
containing the text typed by the user and the buffer
containing the array values being manipulated) the name and
alphameric representation of the current subscripts even if
the user has typed a range or ranges through which the
subscripts are to be varied. The DAVARY routine displays
the name and current subscripts as represented by DALONE,
displays the value contained in the indicated location, and
accepts a new value for the array location if the user has
typed the character = to the right of the subscript range.
DAVARY will ask the user to supply a new value if he did not
type such a value to the right of the = character. Either
of the routines DAROLL or DALOOP are then used to advance
the subscripts through the ranges typed by the user, the
main program calling DALONE, DASITE and DAVARY for each new
FASP, FORTRAN Alphameric Subroutine Package Page 210
DALOAD, Routine to Make Dictionary Describing Named Arrays
set of subscripts until the ranges have been completed.
The FORTRAN statements read as data by DALOAD are written
either in conventional card format, or with a leading tab
character. In card format, the statement begins in or
continues with the text starting in column 7. If the card
format is chosen, a space or zero in column 6 indicates the
start of a new statement, while any other printing character
indicates the continuation of the current statement. In tab
format, any printing character other than the digits 0
through 9 to the immediate right of the initial tab
character starts the text of a new statement. If the
character right of the tab character is a zero, then the
statement starts with the character to the right of the
zero. A digit other than zero to the immediate right of the
tab character indicates that the text to the right of the
digit continues the previously begun statement. The left 72
characters, counting the tab character as a single
character, are usually read from each line of the input
file, but the calling program can specify that some other
number of characters is to be read.
Unrecognizable statements and lines beginning with the
letter C in column 1 are ignored. Statements recognized by
DALOAD cannot begin with statement numbers. Command words
such as COMMON, DIMENSION, IMPLICIT, REAL, INTEGER and END
must not be split across a continuation. Array names and
the parenthetical expressions used to define subscript
limits and the letters to be associated with various data
types can, however, be split across continuations. Such
array names and numbers will be handled correctly even if
comment lines indicated by the letter C in column 1 appear
before the name or number is completed by a subsequent
continuation line. Spaces and/or tabs can appear anywhere
within statements but are ignored.
Each array name in the dictionary constructed by DALOAD has
stored with it a specification of the type of data which the
array is expected to contain. Array names appearing in an
INTEGER or REAL statement are added to the dictionary with
the specification of their data type matching that indicated
by the statement in which they are declared. An additional
statement OCTAL can be used to declare arrays containing
octal information. The array names appearing on the
INTEGER, OCTAL and REAL statements are added to the
dictionary in the order in which they are found. The array
names are not searched for within the dictionary before
being added, so it is not possible to declare an array in a
DIMENSION or COMMON statement, then later change its type.
Within a COMMON or DIMENSION statement, the data type to be
associated with the name is indicated by the first character
of the array name. The default association of the letters I
through N with integer, and A through H and O through Z with
FASP, FORTRAN Alphameric Subroutine Package Page 211
DALOAD, Routine to Make Dictionary Describing Named Arrays
real is restored each time DALOAD is called. These
associations of characters with data types can, however, be
changed for the duration of a particular execution of DALOAD
by the use of the IMPLICIT statement. Although there is no
default association of the octal data type with any initial
letters, the words INTEGER, OCTAL and REAL are all accepted
in IMPLICIT statements as data type names.
The IMPLICIT statement consists of the word IMPLICIT
followed by a data type name and enclosed in parentheses the
letters to be associated with that data type. Within the
parentheses, single letters separated by commas indicate
that these letters are to begin array names associated with
the data type. Two (or more) letters not separated by
commas but the first being lower in the alphabet than the
second indicate that all letters within the range from the
first through the last are to begin array names associated
with the data type. Minus signs can appear between the
letters indicating ranges, but are not required. If a
second (or additional) parenthetical expression follows the
first without the declaration of another data type and with
no printing character other than an optional comma between
the pair of expressions, then the second parenthetical
expression will be assumed to continue the declaration of
the letters to be associated with the same data type.
Another data type name and its associated parenthetical
expressions can follow directly or with intervening spaces,
tabs and/or commas. The association of data types with the
initial letters of the array names applies only to arrays
declared by the following COMMON or DIMENSION statements and
only until the particular letters involved are associated
with another data type by a subsequent IMPLICIT statement.
The statement
IMPLICIT REAL(A-G,R-Z),OCTAL(W-Z,C),INTEGER(P,H,O)
would associate the initial letters A, B, D, E, F, G, R, S,
T, U and V with the real data type. The letters C and W
through Z are temporarily associated with the real data
type, but then are associated with the octal data type. The
sample statement also associates the letters H, O and P with
the integer data type. The letters not assigned by the
current statement will retain their previous data type
associations.
The formats of COMMON, DIMENSION, INTEGER, OCTAL and REAL
statements are identical. The initial command word is
followed by the array names each of which can be followed in
turn by its subscript limits enclosed in parentheses. Array
names can be of any length, but must start with an
alphabetic letter and can contain only alphabetic letters
and digits. Commas are required between array names only if
subscripts are not specified. Following an array name, the
subscripts enclosed in parentheses are separated by commas
FASP, FORTRAN Alphameric Subroutine Package Page 212
DALOAD, Routine to Make Dictionary Describing Named Arrays
and are specified as ranges consisting of the starting
value, a slash character or a colon (the 2 characters are
equivalent), and the ending value. The numbers indicating
the subscript ranges can be signed, but cannot contain an
exponent, neither the E of scientific notation nor K nor M
being accepted. If a slash or a colon is present but either
number is missing, then the missing number is assumed to be
1. If a single number appears without a slash and without a
colon, then the range is assumed to start at 1 and extend
through the indicated number. If nothing or just a slash or
just a colon appears between the parentheses and/or commas,
then the range is assumed to be 1/1. For example, the
statement
DIMENSION ABC,DEF(-123),GHI(20/4,,4/20)
would declare a nonsubscripted variable named ABC, an array
named DEF with dimensions starting at 1 and extending
through -123, and a triply subscripted array named GHI with
first dimension starting at 20 and extending through 4, with
second dimension having the single value 1, and with the
third dimension starting at 4 and extending through 20.
The appearance outside of a parenthetical expression of two
slashes in an array declaration statement (colons are not
recognized in this context) indicates that the arrays which
follow it in the current statement or which are declared by
the following statements are to be considered to be a
logical group (perhaps the contents of a single record of a
file). The appearance of the two slashes causes a specially
marked entry to be added to the dictionary. If nothing
appears between the slashes, then this entry has no
associated name and no associated subscript limits. The
entry does, however, specify the total number of buffer
locations which would be necessary to store the values
associated with the names which follow it in the dictionary.
If no slashes appear at the start of the first array
declaration read by the current call to DALOAD, then such a
zero character zero subscript entry is constructed anyway.
If the sample DIMENSION statement shown above was the first
array declaration read by DALOAD, then an entry would be
created prior to the ABC entry stating that a total of 415
buffer locations (1+125+17*1*17) would be necessary to store
the values associated with the names. If a name and/or
subscript information appears between the slashes, then this
information is stored along with the number of required
buffer locations. The name and/or subscript information, is
meant only to identify to the user, or to the calling
program, the particular logical grouping involved. Although
the names enclosed within slashes are stored in the
dictionary, the DAPICK routine cannot match such names with
the text typed by the user. DAPICK will, however, identify
to the calling program the location within the dictionary of
the entry which specifies the length of the grouping
FASP, FORTRAN Alphameric Subroutine Package Page 213
DALOAD, Routine to Make Dictionary Describing Named Arrays
containing the user selected name.
The dictionary is stored in two arrays, LTRSTR containing
the characters of the names, and NUMSTR containing the
numeric information. The numeric information consists of
the following
a. the number of characters in the array name. This must
be 1 or greater since zero length array names are not
allowed.
b. the associated data type. The coding is as follows
-1 = octal. This is an integer type. The DAVARY
routine will use a radix value of 8 to display the
values to the user and to accept new values from
the user. It must be noted that it is the value
that is displayed by DAVARY, not an octal
representation of the bit pattern. If the number
would be displayed as the radix 10 integer -10,
then it is displayed as the radix 8 integer -12
regardless of whether these octal numerals can be
directly mapped to the binary bits used to store
the number in the computer.
0 = standard FORTRAN integer. The DAVARY routine uses
radix 10 to display such values to the user and to
accept new values from the user.
1 = standard FORTRAN real. The DAVARY routine will
display such values to the precision specified by
the calling program.
c. the number of subscripts. If this is zero, then the
name is nonsubscripted, but DAPICK will accept either
the nonsubscripted name or the name and the single
subscript 1.
d. If the number of subscripts is greater than zero, then
the following pairs of numbers specify the beginning
and ending values of each subscript range.
The buffer size needed for a logical grouping is stored
instead of an associated data type in an entry the first
number of which is the negative of the number of characters
in the name found between the two slashes. Since a name is
not required between the two slashes, the first number of
such an entry can be zero.
The characters of each name are merely appended to the end
of the LTRSTR array 1 character per array location as read
by a multiple of an A1 format.
The DIMENSION statement used earlier as an example would
FASP, FORTRAN Alphameric Subroutine Package Page 214
DALOAD, Routine to Make Dictionary Describing Named Arrays
generate the following dictionary information if read as the
first array declaration.
contents of the LTRSTR array
1HA,1HB,1HC,1HD,1HE,1HF,1HG,1HH,1HI
contents of the NUMSTR array (shown with each entry on
a separate line)
0, 415, 0,
3, 1, 0,
3, 1, 1, 1,-123,
3, 1, 3, 20, 4, 1, 1, 4, 20
The DALOAD Argument List
--- ------ -------- ----
The argument list of routine DALOAD is
SUBROUTINE DALOAD(IDSK ,LTRMAX,NUMMAX,MAXBFR,LTRUSD,
1NUMUSD,LTRSTR,NUMSTR,IBUFFR,IFULL )
with the associated DIMENSION statement
DIMENSION LTRSTR(LTRMAX),NUMSTR(NUMMAX),IBUFFR(MAXBFR)
The following are input arguments left unchanged.
IDSK = number of the input unit from which the FORTRAN
statements are to be read as data. The file being
read from this unit must previously have been
opened by the calling program. If several sets of
array specifications appear in the file separated
by END statements, then subsequent calls to DALOAD
might read from the same file.
LTRMAX = maximum subscript of the locations in the LTRSTR
array which can be used to store the characters of
names in the dictionary.
NUMMAX = maximum subscript of the locations in the NUMSTR
array which can be used to store the numeric
information associated with the names in the LTRSTR
array.
MAXBFR = maximum subscript of the locations in the IBUFFR
array which can be used for temporary storage of
each line of the input file, the characters of
which are read into IBUFFR(1) through
IBUFFR(MAXBFR). MAXBFR is the number of characters
to be read from each line of the input file. It is
suggested that IBUFFR be dimensioned to at least
FASP, FORTRAN Alphameric Subroutine Package Page 215
DALOAD, Routine to Make Dictionary Describing Named Arrays
72, and that MAXBFR be input with the value 72.
The format used in DALOAD to read the input file is
100A1 so MAXBFR should not exceed 100 unless this
format is increased.
The following arguments are used both for input to the
DALOAD routine, and for output to the calling program.
LTRUSD = input containing the number of locations in the
LTRSTR array already in use and which must be left
intact. If the LTRSTR array is empty, then LTRUSD
would be input containing the value 0.
= returned specifying the number of locations in the
LTRSTR array in use after the current call to
DALOAD has appended new names to the dictionary.
If LTRSTR is returned equal to its input value,
then no names were added, but it is possible that a
slash or slashes in an array declaration statement
has caused some numeric information to be added to
the NUMSTR array.
NUMUSD = input containing the number of locations in the
NUMSTR array already in use and which must be left
intact. If the NUMSTR array is empty, then NUMUSD
would be input containing the value 0.
= returned specifying the number of locations in the
NUMSTR array in use after the current call to
DALOAD has appended new array specifications to the
dictionary. If NUMUSD is returned unchanged, then
nothing was appended to the dictionary. If the
computer system being used does not support
end-of-file tests in READ statements, and if more
than one set of data is contained in the input
file, then the end of the file might be marked by
an additional END statement which can be detected
by checking the returned values of both NUMUSD and
IFULL.
The following arguments are used to accumulate the
dictionary. Depending upon the application, a previously
constructed dictionary at the start of these arrays can be
kept intact.
LTRSTR = array used to store the characters of the names, 1
character per storage location as read by a
multiple of an A1 format. LTRSTR must be
dimensioned to at least the value of LTRMAX.
NUMSTR = array used to store the numeric information
associated with the names in the dictionary.
NUMSTR must be dimensioned to at least the value of
NUMMAX.
The following argument is used internally by the DALOAD
FASP, FORTRAN Alphameric Subroutine Package Page 216
DALOAD, Routine to Make Dictionary Describing Named Arrays
routine. The input contents of this buffer are destroyed.
No attempt is made to supply any meaningful information to
the calling program through this argument.
IBUFFR = array used by the DALOAD routine to store each line
of the input file as it is being processed. IBUFFR
must be dimensioned at least to the value of
MAXBFR.
The following argument is used only for output. Its input
value is ignored.
IFULL = returned describing the reason why DALOAD has
relinquished control back to the calling program.
= 1, returned if an END statement was read in the
input file.
= 2, returned if an end-of-file was sensed.
= 3, returned if adding a requested array declaration
to the dictionary would have caused the NUMSTR
array to overflow. NUMUSD is returned pointing to
the end of the numeric information concerning the
last successfully appended array specification.
Since reading an END statement, or sensing an
end-of-file, can cause the used portion of the
NUMSTR array (but not of the LTRSTR array) to
increase slightly, IFULL returned containing 3 does
not assure that an END statement or an end-of-file
was not found.
= 4, returned if adding a requested array declaration
to the dictionary would have caused the LTRSTR
array to overflow. LTRUSD is returned pointing to
the end of the name of the last successfully
appended array specification.
An Example of the Use of DALOAD
-- ------- -- --- --- -- ------
The sample program listed on the following page calls DALOAD
to read array specifications from the user's terminal. The
contents of the dictionary are summarized after DALOAD has
returned control back to the calling program. This report
consists first of the characters in the LTRSTR array, then
of the numbers in the NUMSTR array shown together with the
first 6 letters of each of the associated names in the
LTRSTR array.
A sample dialog between the program and user is presented
following the listing of the program.
DABELT is called to prepare a buffer which can be typed to
identify the locations of the characters of the names within
the LTRSTR array.
FASP, FORTRAN Alphameric Subroutine Package Page 217
DALOAD, Routine to Make Dictionary Describing Named Arrays
C PROGRAM TO DEMONSTRATE DALOAD ROUTINE
C
DIMENSION IBUFFR(72),LTRSTR(1000),NUMSTR(1000)
DATA ITTY,MAXBFR,LTRMAX,NUMMAX/5,72,1000,1000/
DATA ISPACE/1H /
LTRUSD=0
NUMUSD=0
1 LTRBGN=LTRUSD
NUMBGN=NUMUSD
WRITE(ITTY,2)
2 FORMAT(' TYPE ARRAY SPECIFICATIONS'/)
CALL DALOAD(ITTY,LTRMAX,NUMMAX,MAXBFR,LTRUSD,
1NUMUSD,LTRSTR,NUMSTR,IBUFFR,IFULL)
C
C REPORT CHARACTER INFORMATION IN LTRSTR ARRAY
LTRRIT=LTRBGN
3 IF(LTRRIT.GE.LTRUSD)GO TO 6
LTRLFT=LTRRIT+1
LTRRIT=LTRRIT+50
IF(LTRRIT.GT.LTRUSD)LTRRIT=LTRUSD
WRITE(ITTY,4)(LTRSTR(I),I=LTRLFT,LTRRIT)
4 FORMAT(1X,50A1)
LINE=0
5 LINE=LINE+1
CALL DABELT(1,1,1,LINE,LTRLFT,
1LTRRIT,0,MAXBFR,IBUFFR,MAXLIN,MAXPRT,MAXUSD)
WRITE(ITTY,4)(IBUFFR(I),I=1,MAXPRT)
IF(LINE.LT.MAXLIN)GO TO 5
WRITE(ITTY,4)
GO TO 3
C
C REPORT NUMERIC INFORMATION IN NUMSTR ARRAY
6 IF(NUMBGN.GE.NUMUSD)GO TO 11
NUMLFT=NUMBGN+1
LTRLFT=LTRBGN+1
NUMBGN=NUMBGN+3+(2*NUMSTR(NUMLFT+2))
LTRBGN=LTRBGN+IABS(NUMSTR(NUMLFT))
DO 7 MAXPRT=1,6
7 IBUFFR(MAXPRT)=ISPACE
MAXPRT=0
8 IF(LTRLFT.GT.LTRBGN)GO TO 9
MAXPRT=MAXPRT+1
IBUFFR(MAXPRT)=LTRSTR(LTRLFT)
LTRLFT=LTRLFT+1
GO TO 8
9 WRITE(ITTY,10)NUMLFT,NUMBGN,(IBUFFR(I),I=1,6),
1(NUMSTR(I),I=NUMLFT,NUMBGN)
10 FORMAT(' NUMSTR(',I4,'/',1I4,') ',6A1,100I4)
GO TO 6
11 WRITE(ITTY,12)IFULL
12 FORMAT(/' IFULL=',1I2/)
IF(IFULL.LE.2)GO TO 1
STOP
END
FASP, FORTRAN Alphameric Subroutine Package Page 218
DALOAD, Routine to Make Dictionary Describing Named Arrays
Typical Dialog Between DALOAD Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
TYPE ARRAY SPECIFICATIONS
COMMON FIRST(10),SECOND,THIRD(-100/-91)
1IFOURTH(10,10/,11/20)
IMPLICIT INTEGER(O-V)OCTAL(A-H,W-Z)
COMMON/SLASHES/FIFTH,SIXTH(-5)
REAL SEVENTH(1),EIGHTH(3/3)
DIMENSION/(4)/NINTH,TENTH//ELEVENTH
END
FIRSTSECONDTHIRDIFOURTHSLASHESFIFTHSIXTHSEVENTHEIG
12345678911111111112222222222333333333344444444445
01234567890123456789012345678901234567890
HTHNINTHTENTHELEVENTH
555555555666666666677
123456789012345678901
NUMSTR( 1/ 3) 01021 0
NUMSTR( 4/ 8) FIRST 5 1 1 1 10
NUMSTR( 9/ 11) SECOND 6 1 0
NUMSTR( 12/ 16) THIRD 5 1 1-100 -91
NUMSTR( 17/ 25) IFOURT 7 0 3 1 10 10 1 11 20
NUMSTR( 26/ 28) SLASHE -7 10 0
NUMSTR( 29/ 31) FIFTH 5 -1 0
NUMSTR( 32/ 36) SIXTH 5 0 1 1 -5
NUMSTR( 37/ 41) SEVENT 7 1 1 1 1
NUMSTR( 42/ 46) EIGHTH 6 1 1 3 3
NUMSTR( 47/ 51) 0 2 1 1 4
NUMSTR( 52/ 54) NINTH 5 0 0
NUMSTR( 55/ 57) TENTH 5 0 0
NUMSTR( 58/ 60) 0 1 0
NUMSTR( 61/ 63) ELEVEN 8 -1 0
IFULL= 1
TYPE ARRAY SPECIFICATIONS
C TEST DEFAULT ASSOCIATIONS OF LETTERS WITH TYPES
DIMENSION/PRESENT(22,55)/ABC,IJK,OPQ,WXY
END
PRESENTABCIJKOPQWXY
7777777788888888889
2345678901234567890
NUMSTR( 64/ 70) PRESEN -7 4 2 1 22 1 55
NUMSTR( 71/ 73) ABC 3 1 0
NUMSTR( 74/ 76) IJK 3 0 0
NUMSTR( 77/ 79) OPQ 3 1 0
NUMSTR( 80/ 82) WXY 3 1 0
IFULL= 1
FASP, FORTRAN Alphameric Subroutine Package Page 219
DALOAD, Routine to Make Dictionary Describing Named Arrays
Example of Use of Several Named Array Manipulation Routines
------- -- --- -- ------- ----- ----- ------------ --------
An interactive program which was used to prove the
sufficiency of the named array manipulation routines for a
particular application is listed on the following pages
together with a dialog between the user and the program.
Also listed is a shorter noninteractive support program
which constructs a data base for the interactive program.
The interactive program allows the user to examine and
change the values in 6 different groups of simulated arrays.
Each data file read and written by the program stores the
values of a single group of simulated arrays. In those
cases where different versions of the values of a single
group are maintained, these versions are selected by GLOBAL
and LOCAL identification numbers typed by the user. The
GLOBAL number selects the final character of the file name.
The LOCAL number selects a record within the file. If the
user types the name of an array which does not vary
according to either GLOBAL or LOCAL numbers, then these
numbers are ignored. To select a new LOCAL number, the user
types an empty line in response to the request for another
array name. To select a new GLOBAL number, the user types
an empty line (or zero) in response to the request for a new
LOCAL number. When the contents of a new record must be
read, the contents of previous record are first written back
into the file if these values have been changed by the user.
The program reports to the user when files are opened and
when records are read and written.
The values initially stored in the data files are merely the
sequence numbers of the locations containing the values
relative to all locations storing values for that particular
group of simulated arrays. In the assignment of these
values, the LOCAL number is varied more rapidly than the
GLOBAL numbers, i.e. where different versions are stored
for both LOCAL and GLOBAL numbers, the counts continue in
the record selected by the next higher LOCAL number, and
when the end of the file has been reached, are continued in
the file selected by the next higher GLOBAL number. To aid
the user in checking the operation of the program, the name
of each simulated array is constructed of a single
alphabetic letter followed by the digits of the number which
is the subscript of the buffer location which contains the
start of the simulated array, and each group of simulated
arrays which can vary by either GLOBAL or LOCAL numbers
contains a multiple of 1000 locations.
FASP, FORTRAN Alphameric Subroutine Package Page 220
DALOAD, Routine to Make Dictionary Describing Named Arrays
Program Which Constructs Data Base for Interactive Program
------- ----- ---------- ---- ---- --- ----------- -------
C RENBR(MAKDAT/MAKE DATA FILES FOR NAMED ARRAY DEMO)
C
DIMENSION LTRSTR(1000),NUMSTR(1000),IARRAY(2000),
1AARRAY(2000),IBUFFR(72),ITWO(6),IFOUR(6)
EQUIVALENCE(IARRAY(1),AARRAY(1))
C
C NAMES OF INPUT FILE CONTAINING VALUES TO MANIPULATE
DATA ITWO/
15HTWO1 ,5HTWO2 ,5HTWO3 ,5HTWO4 ,5HTWO5 ,5HTWO6 /
DATA IFOUR/
15HFOUR1,5HFOUR2,5HFOUR3,5HFOUR4,5HFOUR5,5HFOUR6/
C
C SIZES OF VARIOUS ARRAYS
C MAXBFR = DIMENSION OF IBUFFR ARRAY
C LTRMAX = DIMENSION OF LTRSTR ARRAY
C NUMMAX = DIMENSION OF NUMSTR ARRAY
DATA MAXBFR,LTRMAX,NUMMAX/72,1000,1000/
C
C UNIT NUMBERS
C ITTY = UNIT NUMBER OF USER TERMINAL FOR INPUT
C JTTY = UNIT NUMBER OF USER TERMINAL FOR OUTPUT
C IDSK = UNIT NUMBER FOR READING ALL FILES
DATA ITTY,JTTY,IDSK/5,5,1/
C
C ESTABLISH ARRAY MANIPULATION DICTIONARY
LTRUSD=0
NUMUSD=0
CALL IFILE(IDSK,5HARRAY)
CALL DALOAD(IDSK,LTRMAX,NUMMAX,MAXBFR,LTRUSD,
1NUMUSD,LTRSTR,NUMSTR,IBUFFR,IFULL)
CALL RELEAS(IDSK)
WRITE(JTTY,1)NUMUSD,NUMMAX,LTRUSD,LTRMAX
1 FORMAT(' USED STORAGE',1I5,1H/,1I5,' NUMERIC'/
113X,1I5,1H/,1I5,' CHARACTER')
C
C OBTAIN DESCRIPTION OF NEXT FILE
KNDFIL=0
2 KNDFIL=KNDFIL+1
CALL DABASE(2,2,1,IBUFFR,KNDFIL,
11,LTRUSD,LTRSTR,1,NUMUSD,NUMSTR,LRGLTR,
2LRGNUM,LRGKNT)
IF(LRGKNT.LE.0)GO TO 19
NEWSIZ=NUMSTR(LRGNUM+1)
KNTLOC=0
NUMINI=LRGNUM+3+(2*NUMSTR(LRGNUM+2))
KONTRL=NUMSTR(NUMINI+1)
MAXFIL=1
MAXRCD=1
NEWFIL=1
3 GO TO(4,5,6,7,8,9),KNDFIL
4 NEWNAM=3HONE
FASP, FORTRAN Alphameric Subroutine Package Page 221
DALOAD, Routine to Make Dictionary Describing Named Arrays
GO TO 10
5 MAXFIL=6
NEWNAM=ITWO(NEWFIL)
GO TO 10
6 NEWNAM=5HTHREE
GO TO 10
7 MAXFIL=6
MAXRCD=3
NEWNAM=IFOUR(NEWFIL)
GO TO 10
8 MAXRCD=3
NEWNAM=4HFIVE
GO TO 10
9 NEWNAM=3HSIX
C
C WRITE THE FILE
10 CALL OFILE(IDSK,NEWNAM)
DO 18 NEWRCD=1,MAXRCD
IF(KONTRL.GT.0)GO TO 12
DO 11 I=1,NEWSIZ
KNTLOC=KNTLOC+1
11 IARRAY(I)=KNTLOC
WRITE(IDSK)(IARRAY(I),I=1,NEWSIZ)
GO TO 14
12 DO 13 I=1,NEWSIZ
KNTLOC=KNTLOC+1
13 AARRAY(I)=KNTLOC
WRITE(IDSK)(AARRAY(I),I=1,NEWSIZ)
14 IF((NEWRCD.EQ.1).AND.(NEWFIL.EQ.1))
1WRITE(JTTY,15)KNDFIL,NEWNAM,NEWRCD,KNTLOC
15 FORMAT(' TYPE =',1I2,', NAME = ',1A5,
1', RCRD =',1I2,', THRU =',1I6)
IF((NEWRCD.EQ.1).AND.(NEWFIL.GT.1))
1WRITE(JTTY,16)NEWNAM,NEWRCD,KNTLOC
16 FORMAT(11X,'NAME = ',1A5,
1', RCRD =',1I2,', THRU =',1I6)
IF(NEWRCD.GT.1)WRITE(JTTY,17)NEWRCD,KNTLOC
17 FORMAT(25X,'RCRD =',1I2,', THRU =',1I6)
18 CONTINUE
END FILE IDSK
NEWFIL=NEWFIL+1
IF(NEWFIL.LE.MAXFIL)GO TO 3
GO TO 2
19 STOP
END
FASP, FORTRAN Alphameric Subroutine Package Page 222
DALOAD, Routine to Make Dictionary Describing Named Arrays
Data File Read by DALOAD to Define Array Names in Dictionary
---- ---- ---- -- ------ -- ------ ----- ----- -- ----------
C FILE TO BE READ BY NAMED ARRAY DEMONSTRATION PROGRAM
C
C SECOND FILE TYPE, VARIES BY GLOBAL NUMBER ONLY
COMMON/SECOND(2)/A1(24,4),A97(16,4),A161(16/1,4),A225(
116,4/1),A289(16/1,4/1),A353(-1/-100),A453(6,64),A837(4
2,4),A853(4,8),A885(4,4),A901(100/)
C
C FOURTH FILE TYPE, VARIES BY GLOBAL AND LOCAL NUMBERS
COMMON/FOURTH(4)/B1,B2(6,6),B38(6,4),B62(8,6),B110(2,1
12,4),B206(3),B209(12,6),B281(3,2,6),B317(6,2),B329(6),
2B335(4,4),B351(8,4),B383(2,3,4),B407(2,4),B415(2,4,3,4
3),B511(2,3),B517(2,2,4)
COMMON B533(2,12,4),B629(2,15,4),B749(40,2),B829(2,12,
12),B877(12,4,2),B973(7,4,2),B1029(13,20),B1289(8,4),B1
2321(8,20),B1481(6),B1487(10),B1497(4,4),B1513(2,6),B15
325(4,4)
COMMONB1541(5,4,4),B1621(20),B1641(4,4),B1657(10,4),
1 B 1 6 9 7 ( 2 , 1 2 , 3 , 4 ) , B 1 9 8 5 ( 4 , 4 )
C
C FIRST FILE TYPE, 1 COPY ONLY
COMMON/FIRST(1)/C1(10),C11(10),C21(30),C51(8),C59(2,4)
1,C67(3,3,2),C85(2,4),C93(3,3,3,2),C147(4,2,4),
2C179(3,2,4,2),C227(2,3,4)
C
C FIFTH FILE TYPE, VARIES BY LOCAL NUMBER ONLY
COMMON/FIFTH(5)/D1(3,3),D10(14,4),D66(4,3),D78(4,22),
1D166(2,3,6),D202(2,4),D210(3,4),D222(4),D226(4,3),
2D238(5),D243(31,18),D801(2,2,4,8),D929(4,3,2,16),
3D1313(3,4,3,16),D1889(2,4,14)
C
C THIRD FILE TYPE, 1 COPY ONLY
OCTAL/THIRD(3)/I1(4,2),I9(4),I13(25)
COMMON I38(2,4),I46,I47(5/5),I48(1),I49(2)
C
C SIXTH FILE TYPE, 1 COPY ONLY
COMMON/SIXTH(6)/J1(4,4,2),J33(18)
END
FASP, FORTRAN Alphameric Subroutine Package Page 223
DALOAD, Routine to Make Dictionary Describing Named Arrays
File Description Produced During Construction of Data Base
---- ----------- -------- ------ ------------ -- ---- ----
USED STORAGE 632/ 1000 NUMERIC
348/ 1000 CHARACTER
TYPE = 1, NAME = ONE , RCRD = 1, THRU = 250
TYPE = 2, NAME = TWO1 , RCRD = 1, THRU = 1000
NAME = TWO2 , RCRD = 1, THRU = 2000
NAME = TWO3 , RCRD = 1, THRU = 3000
NAME = TWO4 , RCRD = 1, THRU = 4000
NAME = TWO5 , RCRD = 1, THRU = 5000
NAME = TWO6 , RCRD = 1, THRU = 6000
TYPE = 3, NAME = THREE, RCRD = 1, THRU = 50
TYPE = 4, NAME = FOUR1, RCRD = 1, THRU = 2000
RCRD = 2, THRU = 4000
RCRD = 3, THRU = 6000
NAME = FOUR2, RCRD = 1, THRU = 8000
RCRD = 2, THRU = 10000
RCRD = 3, THRU = 12000
NAME = FOUR3, RCRD = 1, THRU = 14000
RCRD = 2, THRU = 16000
RCRD = 3, THRU = 18000
NAME = FOUR4, RCRD = 1, THRU = 20000
RCRD = 2, THRU = 22000
RCRD = 3, THRU = 24000
NAME = FOUR5, RCRD = 1, THRU = 26000
RCRD = 2, THRU = 28000
RCRD = 3, THRU = 30000
NAME = FOUR6, RCRD = 1, THRU = 32000
RCRD = 2, THRU = 34000
RCRD = 3, THRU = 36000
TYPE = 5, NAME = FIVE , RCRD = 1, THRU = 2000
RCRD = 2, THRU = 4000
RCRD = 3, THRU = 6000
TYPE = 6, NAME = SIX , RCRD = 1, THRU = 50
FASP, FORTRAN Alphameric Subroutine Package Page 224
DALOAD, Routine to Make Dictionary Describing Named Arrays
Interactive Program Using Named Array Manipulation Routines
----------- ------- ----- ----- ----- ------------ --------
C RENBR(TSTNAM/TEST NAMED ARRAY MANIPULATION ROUTINES)
C
DIMENSION LTRSTR(1000),NUMSTR(1000),IARRAY(2000),
1AARRAY(2000),IBUFFR(72),INISUB(7),LMTSUB(7),
2NOWSUB(7),INCSUB(7),NAME(60),ITWO(6),IFOUR(6),
3IDOUBL(2)
EQUIVALENCE(IARRAY(1),AARRAY(1))
C
C INITIAL VALUES FOR INCREMENT ARRAY
DATA INCSUB/7*0/
C
C NAMES OF INPUT FILE CONTAINING VALUES TO MANIPULATE
DATA ITWO/
15HTWO1 ,5HTWO2 ,5HTWO3 ,5HTWO4 ,5HTWO5 ,5HTWO6 /
DATA IFOUR/
15HFOUR1,5HFOUR2,5HFOUR3,5HFOUR4,5HFOUR5,5HFOUR6/
C
C SIZES OF VARIOUS ARRAYS
C NAMMAX = DIMENSION OF NAME ARRAY
C MAXBFR = DIMENSION OF IBUFFR ARRAY
C LTRMAX = DIMENSION OF LTRSTR ARRAY
C NUMMAX = DIMENSION OF NUMSTR ARRAY
C MAXSUB = DIMENSIONS OF INISUB,LMTSUB,NOWSUB,INCSUB
DATA NAMMAX,MAXBFR,LTRMAX,NUMMAX,MAXSUB/
160,72,1000,1000,7/
C
C UNIT NUMBERS
C ITTY = UNIT NUMBER OF USER TERMINAL FOR INPUT
C JTTY = UNIT NUMBER OF USER TERMINAL FOR OUTPUT
C IDSK = UNIT NUMBER FOR READING ALL FILES
DATA ITTY,JTTY,IDSK/5,5,1/
C
C ESTABLISH ARRAY MANIPULATION DICTIONARY
LTRUSD=0
NUMUSD=0
CALL IFILE(IDSK,5HARRAY)
CALL DALOAD(IDSK,LTRMAX,NUMMAX,MAXBFR,LTRUSD,
1NUMUSD,LTRSTR,NUMSTR,IBUFFR,IFULL)
CALL RELEAS(IDSK)
IF(IFULL.EQ.1)WRITE(JTTY,1)
1 FORMAT(' DALOAD - END STATEMENT READ')
IF(IFULL.EQ.2)WRITE(JTTY,2)
2 FORMAT(' DALOAD - END-OF-FILE READ')
IF(IFULL.EQ.3)WRITE(JTTY,3)
3 FORMAT(' DALOAD - NUMSTR ARRAY OVERFLOW')
IF(IFULL.EQ.4)WRITE(JTTY,4)
4 FORMAT(' DALOAD - LTRSTR ARRAY OVERFLOW')
WRITE(JTTY,5)NUMUSD,NUMMAX,LTRUSD,LTRMAX
5 FORMAT(' USED STORAGE',1I5,1H/,1I5,' NUMERIC'/
113X,1I5,1H/,1I5,' CHARACTER')
C
FASP, FORTRAN Alphameric Subroutine Package Page 225
DALOAD, Routine to Make Dictionary Describing Named Arrays
C TYPE DESCRIPTION OF DICTIONARY
WRITE(JTTY,6)
6 FORMAT(' TYPE ARRAY DESCRIPTION (Y OR N) ',$)
READ(ITTY,7)IFYES
7 FORMAT(1A1)
IF(IFYES.EQ.1HY)CALL DALIST(JTTY,1,LTRUSD,LTRSTR,1,
1NUMUSD,NUMSTR,NAMMAX,NAME)
C
C ASK USER FOR WHAT RECORD HE WANTS
LSTRCD=0
8 WRITE(JTTY,9)
9 FORMAT(' GLOBAL NUMBER ',$)
READ(ITTY,10)MAJOR
10 FORMAT(1I)
IF(MAJOR.LE.0)GO TO 11
IF(MAJOR.LE.6)GO TO 13
11 WRITE(JTTY,12)
12 FORMAT(' EXIT (Y OR N) ',$)
READ(ITTY,7)IFYES
IF(IFYES.NE.1HY)GO TO 8
IF(LSTRCD.EQ.0)GO TO 43
MAJOR=0
GO TO 33
13 WRITE(JTTY,14)
14 FORMAT(' LOCAL NUMBER ',$)
READ(ITTY,10)MINOR
IF(MINOR.LE.0)GO TO 8
IF(MINOR.GT.3)GO TO 8
C
C GET FIRST USER ARRAY SPECIFICATION
15 WRITE(JTTY,16)
16 FORMAT(' ARRAY NAME ',$)
READ(ITTY,17)IBUFFR
17 FORMAT(72A1)
LOWBFR=1
KIND=-1
18 CALL DAPICK(MAXBFR,IBUFFR,1,LTRUSD,LTRSTR,
11,NUMUSD,NUMSTR,MAXSUB,LOWBFR,KIND,LRGLTR,
2LRGNUM,LRGKNT,INITAL,KOUNT,LTRINI,NUMINI,KNTSUB,
3INISUB,LMTSUB)
GO TO(25,25,13,15,23,23,21,19),KIND
C
C SEMICOLON REQUIRED IF TEXT PREVIOUSLY SPECIFIED
19 WRITE(JTTY,20)
20 FORMAT(' SEMICOLON REQUIRED')
GO TO 15
C
C UNKNOWN NAME
21 WRITE(JTTY,22)
22 FORMAT(' UNKNOWN')
GO TO 15
C
C REPORT CORRECT SUBSCRIPT LIMITS IF IN ERROR
23 NAMUSD=0
FASP, FORTRAN Alphameric Subroutine Package Page 226
DALOAD, Routine to Make Dictionary Describing Named Arrays
CALL DALONE(-1,LTRINI,LTRUSD,LTRSTR,NUMINI,
1NUMUSD,NUMSTR,7,NOWSUB,NAMMAX,NAME,NAMUSD)
IF(NAMUSD.GT.0)WRITE(JTTY,24)(NAME(I),I=1,NAMUSD)
24 FORMAT(' CORRECT LIMITS ARE ',72A1)
GO TO 15
C
C DETERMINE THE NEWLY SELECTED FILE AND RECORD
25 NEWSIZ=NUMSTR(LRGNUM+1)
NEWRCD=1
IDENT=NUMSTR(LRGNUM+4)
GO TO(26,27,28,29,30,31),IDENT
26 NEWNAM=3HONE
GO TO 32
27 NEWNAM=ITWO(MAJOR)
GO TO 32
28 NEWNAM=5HTHREE
GO TO 32
29 NEWNAM=IFOUR(MAJOR)
NEWRCD=MINOR
GO TO 32
30 NEWNAM=4HFIVE
NEWRCD=MINOR
GO TO 32
31 NEWNAM=3HSIX
C
C TEST IF NEED TO WRITE OUT FORMER DATA
32 IF(LSTRCD.EQ.0)GO TO 36
IF(LSTRCD.NE.NEWRCD)GO TO 33
IF(LSTNAM.EQ.NEWNAM)GO TO 41
33 IF(MODIFY.LE.0)GO TO 35
WRITE(JTTY,34),LSTRCD
34 FORMAT(1X,'WRITING RECORD',1I3)
WRITE(IDSK#LSTRCD)(AARRAY(I),I=1,LSTSIZ)
35 IF(MAJOR.EQ.0)GO TO 43
IF(LSTNAM.EQ.NEWNAM)GO TO 39
CALL RELEAS(IDSK)
C
C OPEN NEXT FILE
36 ENCODE(10,37,IDOUBL)NEWNAM
37 FORMAT(1A5,5H.DAT )
WRITE(JTTY,38)IDOUBL,NEWSIZ
38 FORMAT(1X,'OPENING ',2A5,'SIZE',1I5)
CALL DEFINE FILE(IDSK,NEWSIZ,IDUMMY,IDOUBL,0,0)
C
C READ NEW INFORMATION
39 WRITE(JTTY,40)NEWRCD
40 FORMAT(1X,'READING RECORD',1I3)
READ(IDSK#NEWRCD)(AARRAY(I),I=1,NEWSIZ)
MODIFY=0
LSTNAM=NEWNAM
LSTRCD=NEWRCD
LSTSIZ=NEWSIZ
C
C ADVANCE THE SUBSCRIPTS THROUGH THE RANGE
FASP, FORTRAN Alphameric Subroutine Package Page 227
DALOAD, Routine to Make Dictionary Describing Named Arrays
41 INLOOP=0
KONTRL=NUMSTR(NUMINI+1)
LSTKNT=KOUNT
42 CALL DAROLL(0,1,KNTSUB,INISUB,LMTSUB,
1INCSUB,INLOOP,NOWSUB)
IF(INLOOP.EQ.0)GO TO 18
C
C FIND THE NEWLY SELECTED ITEM IN THE BUFFER
CALL DASITE(0,KOUNT,1,KNTSUB,NOWSUB,
1-2,LRGNUM,NUMUSD,NUMSTR,LSTKNT,NUMINI,INITAL,
2LOCATN)
C
C CONSTRUCT ALPHAMERIC REPRESENTATION OF ARRAY NAME
NAMUSD=0
CALL DALONE(0,LTRINI,LTRUSD,LTRSTR,NUMINI,
1NUMUSD,NUMSTR,KNTSUB,NOWSUB,NAMMAX,NAME,NAMUSD)
C
C TYPE CURRENT VALUE AND MODIFY IF REQUESTED.
CALL DAVARY(KONTRL,ITTY,JTTY,LOCATN,NAMUSD,
1NAMMAX,MAXBFR,0,-1,8,6,6,
25,AARRAY,IARRAY,NAME,IBUFFR,LOWBFR,KIND,
3MODIFY)
IF(KIND.NE.0)GO TO 42
GO TO 18
C
C USER HAS REQUESTED EXIT
43 STOP
END
FASP, FORTRAN Alphameric Subroutine Package Page 228
DALOAD, Routine to Make Dictionary Describing Named Arrays
Portion of Dialog Text Shown Below Which Was Typed by User
------- -- ------ ---- ----- ----- ----- --- ----- -- ----
The user typed the following text which is included in the
dialog.
Y
1
1
B1985(4,4)!END OF 1ST 2K RECORD OF 1ST FILE
3
B1985(4,4)!END OF 3RD 2K RECORD OF 1ST FILE
4
2
B1985(4,4)!END OF 2ND 2K RECORD OF 4TH 6K FILE
D1889(2,4,14!END OF 2ND 2K RECORD OF SINGLE FILE
&NEXT LINE TESTS ARRAYS WITH WEIRD DIMENSIONS
A901(100);A161(16,1);A161(1,4);A353(/-4)
&NEXT LINES TEST ABILITY TO CHANGE VALUES
I9()=100/100/400;I46(1)=
&NOTE THAT I9 IS OCTAL, I46 DECIMAL
100;I48()=200;I48(1)=
&BLANK RESPONSE KEPT THE ABOVE VALUE FOR I48
B407(,)=2*,2*.023,1K/1/2K;I9(4/1);I46;I48;B407(,)
Dialog With Program Using Named Array Manipulation Routines
------ ---- ------- ----- ----- ----- ------------ --------
DALOAD - END STATEMENT READ
USED STORAGE 632/ 1000 NUMERIC
348/ 1000 CHARACTER
TYPE ARRAY DESCRIPTION (Y OR N) Y
1000 SECOND(2)
REAL A1(24,4) A97(16,4) A161(16/1,4) A225(16,4/1)
REAL A289(16/1,4/1) A353(-1/-100) A453(6,64) A837(4,4)
REAL A853(4,8) A885(4,4) A901(100/1)
2000 FOURTH(4)
REAL B1 B2(6,6) B38(6,4) B62(8,6) B110(2,12,4) B206(3)
REAL B209(12,6) B281(3,2,6) B317(6,2) B329(6) B335(4,4)
REAL B351(8,4) B383(2,3,4) B407(2,4) B415(2,4,3,4)
REAL B511(2,3) B517(2,2,4) B533(2,12,4) B629(2,15,4)
REAL B749(40,2) B829(2,12,2) B877(12,4,2) B973(7,4,2)
REAL B1029(13,20) B1289(8,4) B1321(8,20) B1481(6) B1487(10)
REAL B1497(4,4) B1513(2,6) B1525(4,4) B1541(5,4,4)
REAL B1621(20) B1641(4,4) B1657(10,4) B1697(2,12,3,4)
REAL B1985(4,4)
250 FIRST(1)
REAL C1(10) C11(10) C21(30) C51(8) C59(2,4) C67(3,3,2)
FASP, FORTRAN Alphameric Subroutine Package Page 229
DALOAD, Routine to Make Dictionary Describing Named Arrays
REAL C85(2,4) C93(3,3,3,2) C147(4,2,4) C179(3,2,4,2)
REAL C227(2,3,4)
2000 FIFTH(5)
REAL D1(3,3) D10(14,4) D66(4,3) D78(4,22) D166(2,3,6)
REAL D202(2,4) D210(3,4) D222(4) D226(4,3) D238(5)
REAL D243(31,18) D801(2,2,4,8) D929(4,3,2,16)
REAL D1313(3,4,3,16) D1889(2,4,14)
50 THIRD(3)
OCTAL I1(4,2) I9(4) I13(25)
INTEGER I38(2,4) I46 I47(5/5) I48(1) I49(2)
50 SIXTH(6)
INTEGER J1(4,4,2) J33(18)
GLOBAL NUMBER 1
LOCAL NUMBER 1
ARRAY NAME B1985(4,4)!END OF 1ST 2K RECORD OF 1ST FILE
OPENING FOUR1.DAT SIZE 2000
READING RECORD 1
B1985(4,4) = 2000
ARRAY NAME
LOCAL NUMBER 3
ARRAY NAME B1985(4,4)!END OF 3RD 2K RECORD OF 1ST FILE
READING RECORD 3
B1985(4,4) = 6000
ARRAY NAME
LOCAL NUMBER
GLOBAL NUMBER 4
LOCAL NUMBER 2
ARRAY NAME B1985(4,4)!END OF 2ND 2K RECORD OF 4TH 6K FILE
OPENING FOUR4.DAT SIZE 2000
READING RECORD 2
B1985(4,4) = 22000
ARRAY NAME D1889(2,4,14!END OF 2ND 2K RECORD OF SINGLE FILE
OPENING FIVE .DAT SIZE 2000
READING RECORD 2
D1889(2,4,14) = 4000
ARRAY NAME &NEXT LINE TESTS ARRAYS WITH WEIRD DIMENSIONS
ARRAY NAME A901(100);A161(16,1);A161(1,4);A353(/-4)
OPENING TWO4 .DAT SIZE 1000
READING RECORD 1
A901(100) = 3901
A161(16,1) = 3161
A161(1,4) = 3224
A353(-1) = 3353
A353(-2) = 3354
A353(-3) = 3355
A353(-4) = 3356
ARRAY NAME &NEXT LINES TEST ABILITY TO CHANGE VALUES
ARRAY NAME I9()=100/100/400;I46(1)=
OPENING THREE.DAT SIZE 50
READING RECORD 1
I9(1) = 11 = 100
I9(2) = 12 = 200
I9(3) = 13 = 300
I9(4) = 14 = 400
FASP, FORTRAN Alphameric Subroutine Package Page 230
DALOAD, Routine to Make Dictionary Describing Named Arrays
I46 = 46 = &NOTE THAT I9 IS OCTAL, I46 DECIMAL
I46 = 46 = 100;I48()=200;I48(1)=
I48(1) = 48 = 200
I48(1) = 200 =
ARRAY NAME &BLANK RESPONSE KEPT THE ABOVE VALUE FOR I48
ARRAY NAME B407(,)=2*,2*.023,1K/1/2K;I9(4/1);I46;I48;B407(,)
WRITING RECORD 1
OPENING FOUR4.DAT SIZE 2000
READING RECORD 2
B407(1,2) = 20409 = .023
B407(2,2) = 20410 = .023
B407(1,3) = 20411 = 1000
B407(2,3) = 20412 = 1001
B407(1,4) = 20413 = 1002
B407(2,4) = 20414 = 1003
WRITING RECORD 2
OPENING THREE.DAT SIZE 50
READING RECORD 1
I9(4) = 400
I9(3) = 300
I9(2) = 200
I9(1) = 100
I46 = 100
I48(1) = 200
OPENING FOUR4.DAT SIZE 2000
READING RECORD 2
B407(1,1) = 20407
B407(2,1) = 20408
B407(1,2) = .023
B407(2,2) = .023
B407(1,3) = 1000
B407(2,3) = 1001
B407(1,4) = 1002
B407(2,4) = 1003
FASP, FORTRAN Alphameric Subroutine Package Page 231
DALONE, Routine to Represent Array Name and Subscript Limits
DDDDD AAA LL OOOOO NN NN EEEEEEEE
DD DD AAAA LL OO OO NNN NN EE
DD DD AA AA LL OO OO NNNN NN EE
DD DD AA AA LL OO OO NN NN NN EEEEE
DD DD AAAAAAA LL OO OO NN NNNN EE
DD DD AA AA LL OO OO NN NNN EE
DDDDD AA AA LLLLLLLL OOOOO NN NN EEEEEEEE
DALONE, Routine to Represent Array Name and Subscript Limits
------ ------- -- --------- ----- ---- --- --------- ------
DALOAN represents the name of an array contained in the
dictionary constructed by the DALOAD routine and also
represents the subscripts of this array so that the name and
subscripts can be written by the calling program with a
FORTRAN format statement containing a multiple of an A1
alphameric specification.
DALONE is used along with several other routines in the FASP
package for the purpose of specifying by name, examining and
modifying the values of multiply subscripted arrays
equivalenced with or otherwise loaded into a singly
subscripted buffer. The interaction between these routines
is described at the start of the DALOAD documentation. The
sample program at the end of the DALOAD documentation
illustrates the use of most of these routines.
The DALONE Argument List
--- ------ -------- ----
The argument list of routine DALONE is
SUBROUTINE DALONE(LMTTYP,LTRINI,LTRUSD,LTRSTR,NUMINI,
1 NUMUSD,NUMSTR,KNTSUB,NOWSUB,NAMMAX,NAME ,NAMUSD)
with the associated DIMENSION statement
DIMENSION LTRSTR(LTRUSD),NUMSTR(NUMUSD),
1NAME(NAMMAX),NOWSUB(KNTSUB)
The following arguments are used only for input and are
returned unchanged.
LMTTYP = -2, the characters placed into the NAME array are
to include
a. an identification of the data type associated
with the name of the array. This will be one
of the words OCTAL, INTEGER or REAL. If the
item being represented is instead the name
FASP, FORTRAN Alphameric Subroutine Package Page 232
DALONE, Routine to Represent Array Name and Subscript Limits
and/or subscript limits associated with an
entire logical group of array names, then the
number of locations required in a singly
subscripted buffer to contain the entire group
of arrays is represented rather than the data
type.
b. the name of the array or of the logical group
of arrays.
c. within parentheses, the subscript limits
stored in the dictionary for this particular
array or for this logical group of arrays.
When such information is included with the
name of a logical group of arrays, it is meant
to be searched for by the DABASE routine,
rather than to indicate subscript limits.
= -1, same as LMTTYP=-2, except that the associated
data type or buffer size is not identified.
= 0, the representation is to include the name of the
array and, enclosed in parentheses, the values of
the current subscripts contained in the NOWSUB
array. If the dictionary indicates that the array
is not dimensioned, then no subscripts will follow
the name even if one or more subscripts are
contained in the NOWSUB array.
= greater than zero, the value of LMTTYP is to be
enclosed in parentheses following the array name.
The subscript ranges indicated by the dictionary,
and the values of the current subscripts contained
in the NOWSUB array, are ignored. The value of
LMTTYP is represented even if the dictionary
indicates that the name is not dimensioned.
LTRINI = subscript of the LTRSTR array location containing
the first character to be included in the name of
the array or in the name of the logical group of
arrays.
LTRUSD = highest subscript of any location in the LTRSTR
array used for the storage of a character of the
name of any array.
LTRSTR = array containing the characters of the array names
in the dictionary, 1 character per array location
as read by a multiple of an A1 format or as defined
by several 1H fields.
NUMINI = subscript of the NUMSTR array location containing
the first of the numeric information associated
with the name of the array or with the name of the
logical group of arrays.
FASP, FORTRAN Alphameric Subroutine Package Page 233
DALONE, Routine to Represent Array Name and Subscript Limits
NUMUSD = highest subscript of any location used in the
NUMSTR array to hold numeric information associated
with any name in the dictionary.
NUMSTR = array containing the numeric information
corresponding to the names stored in the LTRSTR
array. The construction of the NUMSTR array is
described in detail in the DALOAD documentation.
For each name in the dictionary, the NUMSTR array
contains
a. the number of characters in the name
b. an indication of the associated data type
c. the number of subscript ranges
d. pairs of starting and ending values of these
ranges.
If the number of characters is instead zero or
negative, then its absolute value is the number of
characters in the name of a logical group of names,
and the next location, rather than indicating the
data type, contains the number of locations within
a singly subscripted buffer which would be needed
to store the values of the multiply subscripted
arrays which are within the logical group and
equivalenced with or otherwise loaded into such a
singly subscripted buffer.
KNTSUB = the number of current subscripts contained in the
NOWSUB array.
NOWSUB = array containing in locations NOWSUB(1) through and
including NOWSUB(KNTSUB) the current values of the
subscripts of the array being represented. This is
used only if LMTTYP=0.
NAMMAX = highest subscript of any location in the NAME array
which can be used for the storage of characters by
this routine.
The following arguments are used for both input to this
routine and for output to the calling program.
NAME = the array into which the representation of the
array name and its subscripts is to be placed so
that it can be written by the calling program using
a multiple of an A1 format. The locations which
can be used for this purpose include NAME(NAMUSD+1)
through NAME(NAMMAX).
NAMUSD = input containing the highest subscript of any
FASP, FORTRAN Alphameric Subroutine Package Page 234
DALONE, Routine to Represent Array Name and Subscript Limits
location in the NAME array which must be preserved.
The representation of the array name and its
subscripts is placed in the NAME array locations
having subscripts above the input value of NAMUSD.
= returned containing the highest subscript of a
location in the NAME array into which this routine
has placed a character of the representation of the
array name and its subscripts. If the number of
locations remaining in the NAME array was
insufficient to hold the representation of the
array name and its subscripts, or if for any other
reason the representation could not be generated,
then NAMUSD is returned unchanged.
FASP, FORTRAN Alphameric Subroutine Package Page 235
DALOOP, Variably Embedded DO Loop Simulator
DDDDDD AAAA LL OOOOO OOOOO PPPPPP
DD DD AA AA LL OO OO OO OO PP PP
DD DD AA AA LL OO OO OO OO PP PP
DD DD AA AA LL OO OO OO OO PP PP
DD DD AAAAAAAA LL OO OO OO OO PPPPPP
DD DD AA AA LL OO OO OO OO PP
DDDDDD AA AA LLLLLLLL OOOOO OOOOO PP
DALOOP, Variably Embedded DO Loop Simulator
------ -------- -------- -- ---- ---------
DALOOP simulates any number of variably embedded DO loops
which have the same statements within their ranges. The
order in which the loop indexes are varied is specified by
the calling program, but can not be changed while the
statements within the range of the loops are being executed.
The index of each loop has its own initial and final values,
and its own increment by which it varies between these
limits. The final value can be less than, equal to, or
greater than the initial value.
DALOOP is called at the start of each execution of the
statements within the loop structure to define the array
named NOWSUB which contains the current values of the loop
indexes. The argument named INLOOP must be set to zero by
the calling program before DALOOP is first called. DALOOP
returns INLOOP nonzero until the outermost loop has been
completed. As an example of the use of DALOOP, if the order
specification array named NXTSUB contains the values 1, 2, 3
and 4 in locations having the subscripts 5 (LOWSUB) through
8 (KNTSUB), and if the loop with the index varying the most
rapidly (the innermost loop) is selected by the largest
value within the NXTSUB array, then the statements
INLOOP=0
1 CALL DALOOP(1,5,8,INISUB,LMTSUB,
1INCSUB,NXTSUB,INLOOP,NOWSUB)
IF(INLOOP.EQ.0)GO TO 2
*
*
text to be executed within the loop structure
*
*
GO TO 1
2 CONTINUE
would simulate the corresponding DO loop structure
FASP, FORTRAN Alphameric Subroutine Package Page 236
DALOOP, Variably Embedded DO Loop Simulator
DO 2 NOWSUB(5)=INISUB(5),LMTSUB(5),INCSUB(5)
DO 2 NOWSUB(6)=INISUB(6),LMTSUB(6),INCSUB(6)
DO 2 NOWSUB(7)=INISUB(7),LMTSUB(7),INCSUB(7)
DO 2 NOWSUB(8)=INISUB(8),LMTSUB(8),INCSUB(8)
*
*
text to be executed within the loop structure
*
*
2 CONTINUE
However, merely by changing the contents of the NXTSUB array
to 3, 1, 4 and 2, the order of the simulated DO loops would
become
DO 2 NOWSUB(6)=INISUB(6),LMTSUB(6),INCSUB(6)
DO 2 NOWSUB(8)=INISUB(8),LMTSUB(8),INCSUB(8)
DO 2 NOWSUB(5)=INISUB(5),LMTSUB(5),INCSUB(5)
DO 2 NOWSUB(7)=INISUB(7),LMTSUB(7),INCSUB(7)
*
*
text to be executed within the loop structure
*
*
2 CONTINUE
A simpler version of DALOOP is also provided which bases the
order of the embedded loops strictly upon the relative
values of the subscripts of the array locations containing
the loop descriptions. This simpler version, named DAROLL,
is meant for use when the order of the embedded loops is not
subject to change during execution of the program.
FASP, FORTRAN Alphameric Subroutine Package Page 237
DALOOP, Variably Embedded DO Loop Simulator
The DALOOP and DAROLL Argument Lists
--- ------ --- ------ -------- -----
The argument lists of routines DALOOP and DAROLL are
SUBROUTINE DALOOP(IRAPID,LOWSUB,KNTSUB,INISUB,LMTSUB,
1 INCSUB,NXTSUB,INLOOP,NOWSUB)
with the associated DIMENSION statement
DIMENSION INISUB(KNTSUB),LMTSUB(KNTSUB),
1INCSUB(KNTSUB),NXTSUB(KNTSUB),NOWSUB(KNTSUB)
and
SUBROUTINE DAROLL(IRAPID,LOWSUB,KNTSUB,INISUB,LMTSUB,
1 INCSUB,INLOOP,NOWSUB)
with the associated DIMENSION statement
DIMENSION INISUB(KNTSUB),LMTSUB(KNTSUB),
1INCSUB(KNTSUB),NOWSUB(KNTSUB)
The argument lists of the 2 routines are identical except
that the DALOOP argument NXTSUB does not appear in the
DAROLL argument list. Since the NXTSUB array dictates the
order of the embedded loops for routine DALOOP, the
definition of the argument IRAPID is somewhat different for
the 2 routines. The definitions of the remaining arguments
are identical for the 2 routines.
The following arguments are used as input and are returned
unchanged. (Under circumstances described below, the
contents of the INCSUB array can be returned changed
slightly.) The contents of these arguments must not be
changed by the calling program until the outermost loop has
been completed.
IRAPID = 0 for routine DALOOP, the innermost loop (the loop
having the index which is varied the most rapidly)
is that with the smallest value in the NXTSUB
array.
= 1 for routine DALOOP, the innermost loop is that
with the largest value in the NXTSUB array.
= 0 for routine DAROLL, the innermost loop is that
represented by the contents of INISUB(LOWSUB),
LMTSUB(LOWSUB), INCSUB(LOWSUB) and NOWSUB(LOWSUB).
The loops represented by the values in the array
locations having higher subscripts are performed
less rapidly than those having lower subscripts.
= 1 for routine DAROLL, the innermost loop is that
represented by the contents of INISUB(KNTSUB),
LMTSUB(KNTSUB), INCSUB(KNTSUB) and NOWSUB(KNTSUB).
FASP, FORTRAN Alphameric Subroutine Package Page 238
DALOOP, Variably Embedded DO Loop Simulator
The loops represented by the values in the array
locations having lower subscripts are performed
less rapidly than those having higher subscripts.
LOWSUB = the lowest subscript of the arrays which are to be
used as the loop descriptors. The portions, if
any, of the arrays below subscript LOWSUB and above
subscript KNTSUB are ignored by this routine and
are returned unchanged. A total of KNTSUB-LOWSUB+1
loops will be simulated by this routine. If
I=LOWSUB+N-1, then the Nth loop is described by the
contents of INISUB(I), LMTSUB(I) and INCSUB(I), its
order relative to the rest of the loops is
specified by NXTSUB(I), and its current subscript
is returned by this routine in NOWSUB(I).
KNTSUB = the highest subscript of the arrays which are to be
used as the loop descriptors. KNTSUB must be
greater than or equal to LOWSUB.
INISUB = array containing the starting values of the
individual loop indexes.
LMTSUB = array containing ending values of the individual
loop indexes. An individual item in the LMTSUB
array can be either less than, equal to, or greater
than the corresponding item in the INISUB array.
INCSUB = array containing the increments by which the
indexes returned in the NOWSUB array are varied
between the starting values in the INISUB array,
and the ending values in the LMTSUB array. If an
increment is zero, its value in the array is
changed to 1. If an increment has the wrong sign,
its sign in the array is changed.
The following argument is used as an input to the DALOOP
routine, then is returned changed for use by subsequent
DALOOP calls. The contents of this array from subscripts
LOWSUB through KNTSUB must not be changed by the calling
program until the outermost loop has been completed. This
argument does not appear in the DAROLL argument list.
NXTSUB = array set by the calling program before the initial
call to DALOOP to control the order of the embedded
loops. DALOOP will return NXTSUB as an array
containing pointers for use by subsequent calls to
DALOOP. The original contents of the NXTSUB array
are sorted to determine the order of the pointers,
but the original contents are destroyed after being
used. The returned values must not be altered by
the calling program.
The original contents could, but need not, equal
FASP, FORTRAN Alphameric Subroutine Package Page 239
DALOOP, Variably Embedded DO Loop Simulator
the subscripts of the loops. Only the relative
sizes of the original contents of the NXTSUB array
are used and all of the numbers from largest to
smallest need not be represented (i.e., the order
of the embedded loops is not changed if the same
offset is added to each of the original values in
the NXTSUB array and the difference between the
largest and smallest numbers in the NXTSUB array
can be greater than the number of embedded loops).
If IRAPID is zero, then the loops corresponding to
the smaller original values in the NXTSUB array
will be done the most rapidly. If IRAPID is 1,
then the loops corresponding to the larger original
values in the NXTSUB array will be done the most
rapidly. Loops which have identical original
values within the NXTSUB array are performed as a
single unit, a single call to DALOOP advancing the
NOWSUB array values for all members of the group
until the NOWSUB array value of any member of the
group exceeds its own corresponding LMTSUB array
value.
When DALOOP finally signals that all loops have
been completed, by returning INLOOP=0, it will also
attempt to restore the original values of the
NXTSUB array, making the assumption that the
smallest value in the range NXTSUB(LOWSUB) through
NXTSUB(KNTSUB) has the value LOWSUB, and that the
values are then incremented by 1. The relative
order within the used portion of the NXTSUB array
will be returned correct, but if the assumptions
are not correct, then the order will be incorrect
relative to that of the unused portion, if any, of
the NXTSUB array.
The following argument is used as input, then is returned
changed for use by both the calling program and by
subsequent calls to this routine. The contents of this
argument must not be changed by the calling program until
the outermost loop has been completed.
INLOOP = must be set to zero before the initial call to this
routine. This initial call will set INLOOP
nonzero, and will copy the INISUB array into the
NOWSUB array for use as the initial loop indexes.
Each subsequent call to this routine will either
load the indexes for the next set of loops into the
NOWSUB array or will set INLOOP to zero if the
outermost loop has been completed.
The following argument is returned changed for use by both
the calling program and by subsequent calls to this routine.
The original contents are ignored. The contents of this
FASP, FORTRAN Alphameric Subroutine Package Page 240
DALOOP, Variably Embedded DO Loop Simulator
argument must not be changed by the calling program until
the outermost loop has been completed.
NOWSUB = array returned containing the indexes for the
current loop structure. The contents of
NOWSUB(LOWSUB) through NOWSUB(KNTSUB) are used in
the same manner as the indexes of FORTRAN DO loops.
The values upon input to the initial call to this
routine are ignored.
Contents of NXTSUB Array After First DALOOP EXECUTION
-------- -- ------ ----- ----- ----- ------ ---------
To prevent unnecessary searching by subsequent calls to
DALOOP of the NXTSUB array which originally contains values
indicating the loop order, the first call to DALOOP converts
the values in this array into pointers. The first location
in the used portion of the NXTSUB array after its conversion
to a pointer array contains the subscript of the index which
is being varied the most rapidly. If additional indexes are
being varied at the same rate, then the second location in
the pointer array will contain the sum of the subscript of
the index being varied at the same rate and the total number
of all indexes so that the sum is greater than the maximum
possible subscript. Scanning from left to right, the next
location in the pointer array which contains a value less
than or equal to the maximum subscript of an index will
point to the index being varied the next most rapidly.
For example, if the initial contents of the NXTSUB array are
1, 2, 3, 2, 3, 4, 3, 4, 5
meaning (if smallest number indicates innermost loop) that
the first index is varied most rapidly, the second and
fourth are varied next most rapidly, and so on through the
ninth which is varied the least rapidly, and if LOWSUB is 1,
then the contents of the NXTSUB array after conversion to
pointers are
1, 2, (4+9)=13, 3, (5+9)=14, (7+9)=16, 6, (8+9)=17, 9
FASP, FORTRAN Alphameric Subroutine Package Page 241
DAMOVE, Justify Group of Characters within Field of Spaces
DDDDD AAA MM MM OOOOO VV VV EEEEEEEE
DD DD AAAA MMM MMM OO OO VV VV EE
DD DD AA AA MMMM MMMM OO OO VV VV EE
DD DD AA AA MM MMMM MM OO OO VV VV EEEEE
DD DD AAAAAAA MM MM MM OO OO VVVV EE
DD DD AA AA MM MM OO OO VVV EE
DDDDD AA AA MM MM OOOOO VV EEEEEEEE
DAMOVE, Justify Group of Characters within Field of Spaces
------ ------- ----- -- ---------- ------ ----- -- ------
DAMOVE left justifies, centers or right justifies a group of
characters within a larger section of the array in which
these characters are contained. The remainder of the
section is filled with spaces. The characters are each
contained in separate array locations as though they had
been read using a multiple of an A1 format or defined with
several 1H fields.
The DAMOVE Argument List
--- ------ -------- ----
The argument list of routine DAMOVE is
SUBROUTINE DAMOVE(JSTIFY,IFILL ,LFTCOL,LTREND,IERR ,
1 IBUFFR,KOUNT )
with the associated DIMENSION statement
DIMENSION IBUFFR(LTREND)
The following are input arguments returned unchanged
JSTIFY = -1, do not move the group of characters input in
IBUFFR(LFTCOL+1) through IBUFFR(KOUNT).
= 0, the group of characters input in
IBUFFR(LFTCOL+1) through IBUFFR(KOUNT) is to be
centered in the field starting at IBUFFR(LFTCOL+1)
and extending through IBUFFR(LTREND). Spaces are
then placed into the IBUFFR array locations
starting with IBUFFR(LFTCOL+1) and extending up to
but not including the IBUFFR array location
containing the leftmost character of the group.
= 1, the group of characters input in
IBUFFR(LFTCOL+1) through IBUFFR(KOUNT) is to be
moved to the right so that the character input in
IBUFFR(KOUNT) is placed into IBUFFR(LTREND).
Spaces are then placed into the IBUFFR array
locations starting with IBUFFR(LFTCOL+1) and
extending up to but not including the IBUFFR array
location containing the leftmost character of the
group.
IFILL = 0, do not place spaces into any IBUFFR array
FASP, FORTRAN Alphameric Subroutine Package Page 242
DAMOVE, Justify Group of Characters within Field of Spaces
locations to the right of those returned containing
the characters of the group. The value of IFILL
has no effect on the placement of the characters of
the group. KOUNT will be returned pointing to the
rightmost character of the group.
= 1, place spaces into the IBUFFR array locations to
the right of those returned containing the group
and extending through IBUFFR(LTREND). KOUNT will
be returned returned pointing to IBUFFR(LTREND).
LFTCOL = subscript of the IBUFFR array location to the
immediate left of the first character in the group.
LTREND = subscript of the IBUFFR array location which forms
the right end of the field which is to be returned
containing the group. If JSTIFY=1, then the
rightmost character of the group will be placed
into IBUFFR(LTREND).
IERR = -1, fill the IBUFFR array locations starting with
IBUFFR(LFTCOL+1) and extending through
IBUFFR(LTREND) with asterisks and return KOUNT
pointing to IBUFFR(LTREND). The input values of
JSTIFY, IFILL and KOUNT are ignored.
= zero or greater, move the group of characters input
in IBUFFR(LFTCOL+1) through IBUFFR(KOUNT) into the
position indicated by JSTIFY.
The following arguments are used both for input to this
routine, and for returning information to the calling
program. The contents of the IBUFFR array locations with
subscripts which are less than or equal to LFTCOL or which
are greater than LTREND are not changed.
IBUFFR = input containing the group of characters to be
moved in locations IBUFFR(LFTCOL+1) through
IBUFFR(KOUNT). Each IBUFFR array location contains
a single character as though read by an A1 format
or defined by a 1H field.
= returned containing this group of characters left
justified, centered or right justified (as directed
by the value of JSTIFY) in the locations
IBUFFR(LFTCOL+1) through IBUFFR(LTREND). If IERR
is input containing -1, then the locations
IBUFFR(LFTCOL+1) THROUGH IBUFFR(LTREND) are instead
returned containing asterisks.
KOUNT = input containing the subscript of the IBUFFR array
location containing the rightmost character of the
group to be positioned as directed by JSTIFY.
= returned containing the subscript of the IBUFFR
array location containing the rightmost character
defined after the group has been moved, and, if
requested by IFILL, after spaces have been added.
FASP, FORTRAN Alphameric Subroutine Package Page 243
DANAME, Convert Buffer Location to Array Name and Subscripts
DDDDD AAA NN NN AAA MM MM EEEEEEEE
DD DD AAAA NNN NN AAAA MMM MMM EE
DD DD AA AA NNNN NN AA AA MMMM MMMM EE
DD DD AA AA NN NN NN AA AA MM MMMM MM EEEEE
DD DD AAAAAAA NN NNNN AAAAAAA MM MM MM EE
DD DD AA AA NN NNN AA AA MM MM EE
DDDDD AA AA NN NN AA AA MM MM EEEEEEEE
DANAME, Convert Buffer Location to Array Name and Subscripts
------ ------- ------ -------- -- ----- ---- --- ----------
DANAME converts a subscript of a singly subscripted buffer
which is considered to hold one or more possibly multiply
subscripted arrays into the array name and its subscripts as
defined by the dictionary constructed by the DALOAD routine.
This conversion is the opposite of that performed by the
DASITE routine.
For example, if the first statement read during the
construction of the dictionary by DALOAD was
DIMENSION FIRST(10),SECOND(10,10),THIRD(10,10,10)
and if it is assumed that the first (left) subscript of each
simulated array varies the most rapidly (the usual FORTRAN
convention for array names appearing without subscripts in
READ and WRITE statements) then the singly subscripted
buffer location having 123 as its subscript would be
equivalent to the simulated array location THIRD(3,2,1).
The buffer location subscript is converted by DANAME to the
array name and its subscripts by determining that the
locations below the desired location include
a. all 10 locations of the FIRST array
b. all 100 locations of the SECOND array
c. the 10 locations THIRD(1,1,1) through THIRD(10,1,1)
d. the 2 locations THIRD(1,2,1) and THIRD(2,2,1)
Since the dictionary constructed by DALOAD is divided into
logical groups of array names, the proper logical group must
be located by calling the DABASE routine prior to the first
call to DANAME, unless it is known that the desired logical
group is the first (or only) group in the dictionary.
FASP, FORTRAN Alphameric Subroutine Package Page 244
DANAME, Convert Buffer Location to Array Name and Subscripts
The DANAME Argument List
--- ------ -------- ----
The argument list of routine DANAME is
SUBROUTINE DANAME(IRAPID,LOCATN,LRGNUM,NUMUSD,NUMSTR,
1 LRGLTR,MAXSUB,INITAL,KOUNT ,LTRINI,NUMINI,KNTSUB,
2 NOWSUB)
with the associated DIMENSION statement
DIMENSION NUMSTR(NUMUSD),NOWSUB(MAXSUB)
The following arguments are used for input only and are
returned unchanged.
IRAPID = 0, multiply subscripted arrays have their left
subscripts vary the most rapidly
= 1, multiply subscripted arrays have their right
subscripts vary the most rapidly.
LOCATN = the value of the subscript of the singly
subscripted buffer which is to be converted to the
corresponding array name and its subscripts.
LRGNUM = subscript of the NUMSTR array location containing
the start of the numeric portion of the array
descriptions.
If NUMSTR(LRGNUM) is less than or equal to zero,
then NUMSTR(LRGNUM) is the first location in the
description of the logical group of arrays the
values of which are equivalenced with or otherwise
contained in the singly subscripted buffer, and the
subsequent description is of the array which if it
contained the location corresponding to the input
value of LOCATN would return KOUNT=1. If
NUMSTR(LRGNUM) is less than or equal to zero, then
its absolute value is the number of characters
within the name of the group of arrays, not the
number of characters in the name of the first
array, and the dictionary is effectively terminated
prior to the next array description which does not
start with a value greater than zero.
If NUMSTR(LRGNUM) is greater than zero, then the
dictionary starts with the description of the array
which if it contained the location corresponding to
the input value of LOCATN would return KOUNT=1. If
NUMSTR(LRGNUM) is greater than zero, then
NUMSTR(LRGNUM) contains the number of characters in
the name of the first array within the logical
group, and the dictionary is effectively terminated
prior to the first array description which does not
FASP, FORTRAN Alphameric Subroutine Package Page 245
DANAME, Convert Buffer Location to Array Name and Subscripts
start with a value greater than zero.
NUMUSD = highest subscript of the locations in the NUMSTR
array containing the numeric information
corresponding to the possibly multiply subscripted
arrays the values of which are equivalenced with or
otherwise contained in the singly subscripted
buffer. NUMUSD is the highest subscript used in
the NUMSTR array for the storage of information
about any array in any logical group, and is not
necessarily the highest subscript used in the
NUMSTR array for the storage of information about
an array in the current logical group.
NUMSTR = the array containing the numeric information
corresponding to the possibly multiply subscripted
arrays the values of which are equivalenced with or
otherwise contained in the singly subscripted
buffer. The construction of the NUMSTR array is
described in detail in the DALOAD documentation.
For each name in the dictionary, the NUMSTR array
contains
a. the number of characters in the name
b. an indication of the associated data type
c. the number of subscript ranges
d. pairs of starting and ending values of these
ranges.
If the number of characters is instead zero or
negative, then its absolute value is the number of
characters in the name of a logical group of names,
and the next location, rather than indicating the
data type, contains the number of locations within
a singly subscripted buffer which would be needed
to store the values of the multiply subscripted
arrays which are within the logical group and
equivalenced with or otherwise loaded into such a
singly subscripted buffer.
LRGLTR = the subscript of the LTRSTR array (not an argument
of this routine but constructed by DALOAD in
parallel with NUMSTR) location which contains the
first letter of the name associated with the
logical group of names in the dictionary if
NUMSTR(LRGNUM) is negative, or which contains the
first letter of the first array name in the logical
group of names if NUMSTR(LRGNUM) is positive or
zero.
MAXSUB = highest subscript of the locations in the NOWSUB
FASP, FORTRAN Alphameric Subroutine Package Page 246
DANAME, Convert Buffer Location to Array Name and Subscripts
array which can be used by this routine to store
the values of the subscripts of the array name
corresponding to the subscript (input as the value
of LOCATN) of the singly subscripted buffer.
The following arguments are used for both input and output.
If this routine is asked to convert a value of LOCATN which
is the subscript of a location in or further beyond the
array identified by the previous call to this routine but
which is still within the same logical group of arrays as
dealt with by the previous call, then the values of these
arguments as output by the previous call are used as the new
offsets in the dictionary and buffer, rather than repeating
the calculations for the lower portion of the logical group.
INITAL = should be set to zero by the calling program before
this routine is first called, and again set to zero
whenever this routine is called to reference a
different logical group of array names in the
dictionary.
= returned greater than zero if the subscript of the
singly subscripted buffer input as the value of the
argument LOCATN could be converted into the name
and subscripts of a possibly multiply subscripted
array equivalenced with or otherwise contained in
part or all of the singly subscripted buffer.
INITAL is returned containing the subscript of the
singly subscripted buffer location which contains
the start of (the first location within) the
possibly multiply subscripted array some location
within which corresponds to the singly subscripted
buffer location having as its subscript the input
value of argument LOCATN. If the possibly multiply
subscripted array identified by this routine
consists of a single location, then INITAL is
returned containing the input value of LOCATN.
= 0, returned if the subscript of the singly
subscripted buffer input as the value of the
argument LOCATN could not be converted into a name
and subscripts of an array equivalenced with or
otherwise contained in part or all of the singly
subscripted buffer. In other words, LOCATN was
input containing a value not indicated by the
dictionary as being within the singly subscripted
buffer containing the arrays forming the logical
group.
= -1, MAXSUB is less than the number of subscripts of
the possibly multiply subscripted array identified
by this routine so that not all of the subscripts
could be represented in the NOWSUB array.
KOUNT = input value is ignored if INITAL is input
containing zero or if INITAL is input greater than
the input value of LOCATN. KOUNT is set by each
FASP, FORTRAN Alphameric Subroutine Package Page 247
DANAME, Convert Buffer Location to Array Name and Subscripts
call to this routine, and should never be set by
the calling program.
= if INITAL is input greater than zero but less than
or equal to LOCATN, then KOUNT, as returned by the
previous call to this routine, is the sequence
number of the possibly multiply subscripted array
corresponding to the input value of INITAL.
= returned containing the sequence number of the
identified array relative to all arrays in the
logical group of arrays. If the third array in the
logical group contains the location corresponding
to the input value of LOCATN, then KOUNT is
returned containing the value 3.
If NUMSTR(LRGNUM) is less than or equal to zero,
then NUMSTR(LRGNUM) is the first location in the
description of the logical group of arrays the
values of which are equivalenced with or otherwise
contained in the singly subscripted buffer, and the
subsequent description is of the array which if it
contained the location corresponding to the input
value of LOCATN would return KOUNT=1. If
NUMSTR(LRGNUM) is less than or equal to zero, then
its absolute value is the number of characters
within the name of the group of arrays, not the
number of characters in the name of the first
array, and the dictionary is effectively terminated
prior to the next array description which does not
start with a value greater than zero.
If NUMSTR(LRGNUM) is greater than zero, then the
dictionary starts with the description of the array
which if it contained the location corresponding to
the input value of LOCATN would return KOUNT=1. If
NUMSTR(LRGNUM) is greater than zero, then
NUMSTR(LRGNUM) contains the number of characters in
the name of the first array within the logical
group, and the dictionary is effectively terminated
prior to the first array description which does not
start with a value greater than zero.
LTRINI = input value is ignored if INITAL is input
containing zero or if INITAL is input greater than
the input value of LOCATN. LTRINI is set by each
call to this routine, and should never be set by
the calling program.
= if INITAL is input greater than zero but less than
or equal to LOCATN, then LTRINI, as returned by the
previous call to this routine, is the subscript of
the LTRSTR location containing the first character
of the name of the possibly multiply subscripted
array corresponding to the input value of INITAL.
= returned containing the subscript of the LTRSTR
array location containing the first character of
FASP, FORTRAN Alphameric Subroutine Package Page 248
DANAME, Convert Buffer Location to Array Name and Subscripts
the name of the identified array.
NUMINI = input value is ignored if INITAL is input
containing zero or if INITAL is input greater than
the input value of LOCATN. NUMINI is set by each
call to this routine, and should never be set by
the calling program.
= if INITAL is input greater than zero but less than
or equal to LOCATN, then NUMINI, as returned by the
previous call to this routine, is the subscript of
the NUMSTR location containing the start of the
numeric description of the possibly multiply
subscripted array corresponding to the input value
of INITAL.
= returned containing the subscript of the NUMSTR
array location containing the start of the numeric
description of the identified array.
The following arguments are used only for output. Their
input values are ignored. These arguments are returned
undefined if INITAL is returned less than or equal to zero.
KNTSUB = returned containing the number of subscripts of the
identified array. If the NUMSTR array indicates
that the identified array is nondimensioned, then
KNTSUB is returned containing 1, and NOWSUB(1) is
returned also containing 1. This precaution is
performed so that any subsequent FASP routine call
which includes KNTSUB in its argument list for use
in a DIMENSION statement within the routine, will
not have that dimension input as the value zero
which could be considered illegal during execution
of the program (depending upon whether the compiler
generates code to check for this condition).
Without this precaution, a subsequent call to
DALONE, for example, could cause the DIMENSION
statement in that routine to effectively become
DIMENSION NOWSUB(0)
Although such a statement has meaning to DALOAD
when read as data (in particular, to DALOAD such a
statement means that NOWSUB includes the 2
locations NOWSUB(1) and NOWSUB(0) with the higher
subscript value of 1 appearing first in storage),
it probably does not have a valid meaning in the
code generated by the compiler. Of course, under
these circumstances, the logic in DALONE is such
that no location whatever is used in the NOWSUB
array so that probably no damage would be done if
the zero dimension is not detected as being illegal
by the entry code generated by the compiler.
NOWSUB = array returned containing in locations NOWSUB(1)
FASP, FORTRAN Alphameric Subroutine Package Page 249
DANAME, Convert Buffer Location to Array Name and Subscripts
through and including NOWSUB(KNTSUB) the values of
the subscripts of the possibly multiply subscripted
array location corresponding to the subscript of
the singly subscripted buffer input as the argument
LOCATN.
FASP, FORTRAN Alphameric Subroutine Package Page 250
DANEXT, Returns Next Value Typed as Single Number or Series
DDDDD AAA NN NN EEEEEEEE XX XX TTTTTTTT
DD DD AAAA NNN NN EE XX XX TT
DD DD AA AA NNNN NN EE XXXX TT
DD DD AA AA NN NN NN EEEEE XX TT
DD DD AAAAAAA NN NNNN EE XXXX TT
DD DD AA AA NN NNN EE XX XX TT
DDDDD AA AA NN NN EEEEEEEE XX XX TT
DANEXT, Returns Next Value Typed as Single Number or Series
------ ------- ---- ----- ----- -- ------ ------ -- ------
DANEXT returns to the calling program the next number
represented by the characters in an input buffer read by the
calling program with a multiple of an A1 format. If a
series specification is encountered in the input buffer,
then the individual numbers of the series are returned by
the calls to this routine. When all of the numbers forming
the series have been returned, then the subsequent number or
series of numbers specified by the contents of the buffer
are evaluated.
A series evaluated by this routine can be written as a lower
bound, increment and upper bound separated by slashes or by
colons (the 2 characters are equivalent). If the increment
is to be one, then the lower and upper bounds need be
separated only by a single slash or by a single colon. If
either bound is missing, it is assumed to be zero. The
lower bound is the first number of the series reported to
the calling program. The series can then either increase or
decrease depending upon whether the upper bound is less
than, equal to, or greater than the lower bound. The sign
of the increment is changed if it does not conform to the
relative values of the bounds.
If the series is to consist of several repetitions of the
same value, then the series is instead written as the number
of times the value is to be used followed immediately by an
asterisk and the value itself. The value to be repeated is
assumed to be null if it is missing.
The only printing delimiter character allowed between series
specifications is the comma. Two commas with no other
printing characters between them are taken to indicate a
missing series specification. The calling program is
informed if a semicolon is encountered in the text being
evaluated by this routine. The semicolon might be typed by
the user to signal that the calling program is to terminate
its requests for data from the user. An exclamation point
and any characters to its right are taken to be a comment
and are otherwise ignored. An ampersand and the characters
to its right are similarly considered to form a comment, but
the calling program is informed that an ampersand was found.
An ampersand might be typed by the user to indicate that the
FASP, FORTRAN Alphameric Subroutine Package Page 251
DANEXT, Returns Next Value Typed as Single Number or Series
input text is to be continued on the following line.
Two versions of the routine are supplied. DANEXT can
evaluate real numbers as well as either decimal or octal
integers including E, K and M notations for specifying
either of these. If the calling program does not otherwise
reference the free format input routine DAHEFT, if the
evaluation of either real or octal numbers is not needed,
and if integers can be specified without resorting to the E,
K and M notations, then the routine DAINXT should be used
instead of DANEXT. Numbers evaluated by DAINXT must consist
only of digits following the optional sign. DAINXT treats
the characters ., %, E, K and M as illegal characters.
The DANEXT and DAINXT Argument Lists
--- ------ --- ------ -------- -----
The argument lists of routines DANEXT and DAINXT are
SUBROUTINE DANEXT(KONTRL,SLACK ,IBUFFR,MAXBFR,LOWBFR,
1 MANY ,KIND ,NEWVAL,INCVAL,LMTVAL,VALNEW,VALINC,
2 VALLMT)
and
SUBROUTINE DAINXT (IBUFFR,MAXBFR,LOWBFR,
1 MANY ,KIND ,NEWVAL,INCVAL,LMTVAL)
both with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The ordering of the arguments is the same for both routines,
but the first 2 arguments of DANEXT, and the last 3
arguments of DANEXT are not included in the DAINXT argument
list. The argument definitions for DAINXT are identical to
those of DANEXT with the exception that the new value is
always returned in integer form for DAINXT.
The following arguments are used for input only and are
returned unchanged.
KONTRL = -1, the number (or series of numbers) in the IBUFFR
array is an octal integer (or a series of octal
integers). The number can be typed with a decimal
point and/or with an exponent. However, the number
following the letter E of the exponent is evaluated
in decimal. The value of the octal number is
returned as the argument NEWVAL. It must be noted
that numbers evaluated as negative octal integers
have the negative octal integer as their value, not
as their bit representation in computer storage.
For example, on a 36 bit twos complement computer,
FASP, FORTRAN Alphameric Subroutine Package Page 252
DANEXT, Returns Next Value Typed as Single Number or Series
the octal number -400000000000 (which could also be
typed as -4E11 or -4E+11 where the 11 after the E
is in decimal) is represented as the bit pattern
having octal notation 400000000000 and the octal
number -377777777777 is represented by the bit
pattern 400000000001.
= 0, the number (or series of numbers) in the IBUFFR
array is a decimal integer (or a series of decimal
integers). The number can be typed with a decimal
point (for example 1.23K or 1.23E3 equals 1230),
but is stored as an integer, and is output as
argument NEWVAL. Any decimal integer which the
computer can represent can be evaluated. This
includes, on twos complement computers, the largest
negative number the absolute value of which cannot
be stored. On the PDP10, a 36 bit computer with
twos complement notation, the range of decimal
integers is -34359738368 through 34359738367 (the
octal notation of the bit patterns being
400000000000 through 377777777777).
= 1 or greater, the number (or series of numbers) in
the IBUFFR array is a real number (or a series of
real numbers). If possible, the real number will
be accumulated as an integer, then be converted to
real and shifted if necessary. KONTRL is then the
maximum number of digits in the integer. The value
is output as the argument VALNEW. If the item has
more than KONTRL digits, then the entire evaluation
is done as a real number. The advantage of
calculating the real values in integer as long as
the precision of the computer is not overflowed is
that the calculation of the portion of the number
right of the decimal point is more exact. As an
example, if KONTRL is greater than or equal to 4,
then the number 33.33 can be stored as the integer
3333, then be converted to the real value 3333.0
and divided by 100.0 to obtain the final answer.
If it makes no difference whether the number typed
as 33.33 has value 33.33 or 33.32999... then
KONTRL can be given the value 1.
SLACK = the fraction of the increment of a series expressed
in slash or colon notation, if KONTRL is greater
than zero, by which the returned value of the
argument named VALNEW can exceed the upper bound
and still be considered to be within the series.
SLACK is necessary because of the imprecision of
calculations involving real numbers. For example,
if the user wished to have the increment be
0.666... between a lower bound of 3.0 and an upper
bound of 5.0, he might type 3/.667/5 and expect
that the numbers 3.000, 3.667, 4.333, and 5.000
would be included within the series, while,
actually, the number near 5.00 would be calculated
FASP, FORTRAN Alphameric Subroutine Package Page 253
DANEXT, Returns Next Value Typed as Single Number or Series
to be 5.001 and would therefore be outside the
series. Although SLACK can be set to zero, a value
of about 0.01 is recommended. In the above
example, SLACK=0.01 would cause the actual upper
bound to be 5.00667 so that the number calculated
near the upper bound would be included within the
series. SLACK should almost certainly never exceed
0.5 since larger values would often lead to the
inclusion of values not meant by the user. SLACK
is ignored if KONTRL is less than or equal to zero.
IBUFFR = input buffer array containing characters typed by
user, read by a multiple of an A1 format, which is
to be searched for series specifications. IBUFFR
then contains one character per computer storage
location.
MAXBFR = maximum subscript of IBUFFR array to be searched
The following arguments are used both for input to these
routines, and for output to the calling program.
LOWBFR = input containing the subscript within the IBUFFR
array of the first (leftmost) character which can
be scanned for a series specification. LOWBFR is
returned unchanged if this call to this routine
merely generates the next member of a series begun
by a previous call. If a new series specification
is evaluated by this call to this routine, then
LOWBFR will be returned pointing to the next
character beyond the series specification. If a
second number immediately follows a first without a
separating slash or colon or asterisk, then LOWBFR
will be returned pointing to the first character of
the second number. If either a space or a tab
character follows a series specification, then
LOWBFR will be returned pointing to the space or
tab character. If the first printing character at
or to right of IBUFFR(LOWBFR) is an exclamation
point, or if there are no printing characters at or
to the right of IBUFFR(LOWBFR), then LOWBFR will be
returned containing MAXBFR+1 and KIND will be
returned containing one. LOWBFR, MANY and KIND
must be set by the calling program before anything
is processed in the current contents of the IBUFFR
array, but then should not be modified by the
calling program until the entire contents of the
IBUFFR array have been processed (except that KIND
can be returned to zero to abandon the generation
of a particular series).
MANY = should be input containing zero each time this
routine is called to begin processing of a new
logical section of text, as for example when
FASP, FORTRAN Alphameric Subroutine Package Page 254
DANEXT, Returns Next Value Typed as Single Number or Series
beginning processing of a line of text not tied to
the previous line by an ampersand at the end of the
previous line, or when processing the text to the
right of a semicolon. The initial zeroing of this
argument must be done by the calling program, but
thereafter the value returned by the previous call
to this routine can usually be used. MANY is
returned set to zero each time a semicolon (KIND=2)
is found, and each time an end of line not tied to
the following line by an ampersand (KIND=1) is
found. MANY is returned containing one plus its
input absolute value each time a new specification
of either a single number or of a series is found,
each time an erroneous series specification is
found, each time an unknown character is found, or
each time an indication of a missing series
specification is found. MANY is returned unchanged
if this call to this routine merely generates the
next member of a series already begun by a previous
call to this routine. KIND is returned containing
the value 3 and MANY is returned containing the
negative of the number of items found if the next
printing character following a comma is an
ampersand. MANY should not be changed by the
calling program if an ampersand (KIND being
returned=3) is found indicating that the subsequent
call to this routine is to process text which is to
be treated as though it appeared in place of the
ampersand and the characters to its right. The
effect is not quite the same as if the user had
typed all of the text on a single line since a
single series specification cannot be split across
the line boundary.
If MANY is input containing zero, then an initial
comma in the input text buffer is taken to indicate
an initial missing item, and MANY is then returned
containing 1. If MANY is input greater than zero,
then an initial comma is ignored if followed by a
series specification. If MANY is input greater
than zero, then an initial comma followed by no
other printing characters, by a semicolon, or by an
exclamation point indicates a missing item. If
MANY is input greater than zero, then an initial
comma followed by an ampersand will cause the
remaining characters in the buffer to be ignored,
and MANY will be returned containing the negative
of its input value. If MANY is input negative,
then it is assumed that the contents of the current
buffer continue a previous line which terminated
with a comma followed by an ampersand, and MANY is
returned greater than zero.
KIND = should be input containing zero the first time this
FASP, FORTRAN Alphameric Subroutine Package Page 255
DANEXT, Returns Next Value Typed as Single Number or Series
routine is called, or to abandon generation of
values within a particular series. KIND is
returned describing the kind of item located in the
IBUFFR array. Except in the cases of when the
buffer is found to be empty, or when the calling
program resets KIND to zero to terminate the
expansion of the current series, the returned value
of KIND should be used for the next call to this
routine to inform this routine whether a series is
being expanded. KIND is returned containing one of
the following values.
= 1, nothing, except perhaps a comment indicated by a
leading exclamation point, was found at or to the
right of IBUFFR(LOWBFR). The calling program
should read a new line into the IBUFFR array before
again calling this routine if additional values are
required. MANY is returned set to zero.
= 2, a semicolon was found as the first printing
character at or to the right of IBUFFR(LOWBFR).
LOWBFR is returned pointing to the next character
beyond the location of the semicolon. It is
assumed the calling program will treat the
appearance of the semicolon as marking the end of a
statement. MANY is returned set to zero.
= 3, an ampersand was found as the first printing
character at or to the right of LOWBFR. The text
to the right of the ampersand is taken as a comment
so LOWBFR is returned pointing beyond the right end
of the buffer. It is assumed that the calling
program will read in the contents of a new buffer,
then again request a new series evaluation from
this routine. The value of MANY must not be
changed by the calling program prior to this
following call. The effect is not quite the same
as if the user had typed all of the text on a
single line since a series specification cannot be
split across a line boundary.
= 4, a number or series specification was not found,
but a comma was found indicating a missing series
specification. Either NEWVAL or VALNEW, whichever
is appropriate, is returned with the value zero so
KIND=4 can be considered equivalent to KIND=5 if
such is appropriate to the application for which
this routine is being used.
= 5, the next number specified by the contents of the
input buffer is being returned as the value of the
argument NEWVAL or VALNEW, whichever is
appropriate.
= 6, a series was specified in asterisk notation, but
no number appeared to right of the asterisk.
NEWVAL or VALNEW, whichever is appropriate, is
returned with value zero. Subsequent calls to this
routine will continue to return KIND=6 until the
specified number of null values has been returned
FASP, FORTRAN Alphameric Subroutine Package Page 256
DANEXT, Returns Next Value Typed as Single Number or Series
to the calling program. Note that if a number does
not precede the asterisk, then KIND is returned
with the value 7 whether or not a number follows
the asterisk.
= 7, a series was specified in asterisk notation, but
no number appeared to the left of the asterisk or
the number to left of the asterisk was less than 1
(or less than 0.5 if evaluated as a real number as
this number is rounded when used). The value of
the number to the right of the asterisk will be
returned in either NEWVAL or VALNEW, whichever is
appropriate. If no number appears to the right of
the asterisk, then either NEWVAL or VALNEW,
whichever is appropriate, will be returned
containing zero.
= 8, a series specification was found which contained
too many numbers, too many asterisks, too many
slashes or too many colons.
= 9, the first printing character (other than a
possible comma if MANY was input greater than zero)
in or to right of IBUFFR(LOWBFR) was not a
character which could appear in a number or number
range, and was not a comma, semicolon, ampersand or
exclamation point. LOWBFR is returned pointing to
the next character beyond this character. The
calling program must decrement LOWBFR by 1 if the
unknown character is to be identified by other
routines in this package.
The following argument is used to return the next value
represented by the contents of the text buffer if this is
being evaluated as an integer.
NEWVAL = returned containing the next integer value
evaluated by this routine (the value of KONTRL
being less than or equal to zero). The contents of
this argument should not be changed by the calling
program.
The following 2 arguments are used, together with KIND and
NEWVAL, by the current call to this routine to specify the
characteristics of the currently generated integer series to
the subsequent call to this routine. Neither of these
arguments should be changed by the calling program, and
their returned values can probably otherwise be ignored by
the calling program.
INCVAL = returned containing the increment if an integer
series was specified in slash or colon notation.
INCVAL is used by the subsequent call to this
routine even if the series is composed of real
numbers. INCVAL is returned nonzero for use as a
simple switch by the subsequent call to this
routine if a real number series was specified in
FASP, FORTRAN Alphameric Subroutine Package Page 257
DANEXT, Returns Next Value Typed as Single Number or Series
slash or colon notation. INCVAL is set to zero if
a series specified in asterisk notation was found,
whether or not the series was evaluated in integer
or real mode.
LMTVAL = returned containing the ending value if an integer
series was specified in slash or colon notation.
If the integer series was specified in asterisk
notation, then LMTVAL is returned containing the
number of times that this routine must return the
same value of argument NEWVAL, including the
current return. In other words, the first time a
new integer value expressed in asterisk notation is
returned, LMTVAL is equal to the number left of the
asterisk, not to one less than the number left of
the asterisk.
The following argument is used to return the next value
represented by the contents of the text buffer if this is
being evaluated as a real number.
VALNEW = returned containing the next real value evaluated
by this routine (the value of KONTRL being greater
than zero). The contents of this argument should
not be changed by the calling program.
The following 2 arguments are used, together with KIND and
VALNEW (and INCVAL, see above), by the current call to this
routine to specify the characteristics of the currently
generated real number series to the subsequent call to this
routine. Neither of these arguments should be changed by
the calling program, and their returned values can probably
otherwise be ignored by the calling program.
VALINC = returned containing the increment if a real number
series was specified in slash or colon notation.
VALLMT = returned containing the ending value if a real
number series was specified in slash or colon
notation. If the real number series was specified
in asterisk notation, then VALLMT is returned
containing the number of times that this routine
must return the same value of argument VALNEW,
including the current return. In other words, the
first time a new real number value expressed in
asterisk notation is returned, VALLMT is equal to
the number found left of the asterisk, not to one
less than the number left of the asterisk.
FASP, FORTRAN Alphameric Subroutine Package Page 258
DANEXT, Returns Next Value Typed as Single Number or Series
Demonstration Program to Interactively Test DANEXT
------------- ------- -- ------------- ---- ------
The program listed on the following pages accepts a line of
text from the user, then reports each punctuation mark, each
single value, and each of the values in the ranges included
in the text. A sample dialog between the program and the
user is presented following the listing of the program.
C PROGRAM TO DEMONSTRATE DANEXT ROUTINE
C
DATA ITTY,JTTY,MAXBFR/5,5,58/
DIMENSION IBUFFR(58),JBUFFR(20)
DATA IGREAT/1H>/
WRITE(JTTY,1)
1 FORMAT(1X,37HPROGRAM TO DEMONSTRATE DANEXT ROUTINE/
124H TEST DAINXT (Y OR N) = ,$)
READ(ITTY,2)IANS
2 FORMAT(1A1)
KONTRL=0
IF(IANS.EQ.1HY)GO TO 7
WRITE(JTTY,3)
3 FORMAT(10H KONTRL = ,$)
READ(ITTY,4)KONTRL
4 FORMAT(I)
WRITE(JTTY,5)
5 FORMAT(10H SLACK = ,$)
READ(ITTY,6)SLACK
6 FORMAT(F)
7 IRADIX=10
IF(KONTRL.LT.0)IRADIX=8
MANY=0
C
C READ TEXT TYPED BY USER
8 WRITE(JTTY,9)
9 FORMAT(1X,1H*,$)
READ(ITTY,10)IBUFFR
10 FORMAT(100A1)
LOWBFR=1
KIND=0
C
C GET NEXT NUMBER IN SERIES
11 INIBFR=LOWBFR
IF(IANS.NE.1HY)CALL DANEXT(KONTRL,SLACK,IBUFFR,MAXBFR,
1LOWBFR,MANY,KIND,NEWVAL,INCVAL,LMTVAL,VALNEW,
2VALINC,VALLMT)
IF(IANS.EQ.1HY)CALL DAINXT(IBUFFR,MAXBFR,
1LOWBFR,MANY,KIND,NEWVAL,INCVAL,LMTVAL)
IF(LOWBFR.LE.INIBFR)GO TO 13
J=LOWBFR-1
WRITE(JTTY,12)(IBUFFR(I),I=INIBFR,J),IGREAT
12 FORMAT(1X,1H<,100A1)
13 GO TO(14,16,18,20,22,27,22,29,31),KIND
FASP, FORTRAN Alphameric Subroutine Package Page 259
DANEXT, Returns Next Value Typed as Single Number or Series
C
C REPORT RESULT OF CURRENT CALL TO DANEXT
14 WRITE(JTTY,15)
15 FORMAT(6H EMPTY)
GO TO 8
16 WRITE(JTTY,17)
17 FORMAT(17H END OF STATEMENT)
GO TO 11
18 WRITE(JTTY,19)
19 FORMAT(10H AMPERSAND)
GO TO 8
20 WRITE(JTTY,21)
21 FORMAT(8H MISSING)
GO TO 11
22 KOUNT=0
IF(KONTRL.GT.0)GO TO 23
CALL DANUMB(0,NEWVAL,IRADIX,JBUFFR,KOUNT,0,20)
GO TO 24
23 CALL DARITE(VALNEW,-1,0,0,0,
1-3,0,20,-1,8,6,6,
2-1,0,5,0,0,0,0,
320,JBUFFR,KOUNT,IERR)
24 IF(KIND.EQ.5)WRITE(JTTY,25)(JBUFFR(I),I=1,KOUNT)
25 FORMAT(7H VALUE ,20A1)
IF(KIND.EQ.7)WRITE(JTTY,26)(JBUFFR(I),I=1,KOUNT)
26 FORMAT(21H ZERO DUPLICATION OF ,20A1)
GO TO 11
27 WRITE(JTTY,28)
28 FORMAT(32H ASTERISK NOT FOLLOWED BY NUMBER)
GO TO 11
29 WRITE(JTTY,30)
30 FORMAT(15H TOO MANY ITEMS)
GO TO 11
31 WRITE(JTTY,32)
32 FORMAT(27H SEPARATOR OTHER THAN COMMA)
GO TO 11
END
FASP, FORTRAN Alphameric Subroutine Package Page 260
DANEXT, Returns Next Value Typed as Single Number or Series
Typical Dialog Between DANEXT Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
PROGRAM TO DEMONSTRATE DANEXT ROUTINE
TEST DAINXT (Y OR N) = N
KONTRL = 4
SLACK = 0.01
*1.23K 0.012M 1.067E-8!TEST SIMPLE NUMBERS
<1.23K>
VALUE 1230
< 0.012M>
VALUE 12000
< 1.067E-8>
VALUE 1.067E-8
<!TEST SIMPLE NUMBERS >
EMPTY
*1E6/2E5/2E6,3*,,2*0.04618,,;0.0245/0.002/0.03
<1E6/2E5/2E6>
VALUE 1E6
VALUE 1.2E6
VALUE 1.4E6
VALUE 1.6E6
VALUE 1.8E6
VALUE 2E6
<,3*>
ASTERISK NOT FOLLOWED BY NUMBER
ASTERISK NOT FOLLOWED BY NUMBER
ASTERISK NOT FOLLOWED BY NUMBER
<,>
MISSING
<,2*0.04618>
VALUE .04618
VALUE .04618
<,>
MISSING
<,>
MISSING
<;>
END OF STATEMENT
<0.0245/0.002/0.03>
VALUE .0245
VALUE .0265
VALUE .0285
< >
EMPTY
* 4*1352.01 , , 2/5 ; , -17/2/-25 &WITH SPACES
< 4*1352.01>
VALUE 1352.01
VALUE 1352.01
VALUE 1352.01
VALUE 1352.01
< , >
MISSING
<, 2/5>
FASP, FORTRAN Alphameric Subroutine Package Page 261
DANEXT, Returns Next Value Typed as Single Number or Series
VALUE 2
VALUE 3
VALUE 4
VALUE 5
< ;>
END OF STATEMENT
< >
MISSING
<, -17/2/-25>
VALUE -17
VALUE -19
VALUE -21
VALUE -23
VALUE -25
< &WITH SPACES>
AMPERSAND
*0/0.667/2 &TESTING SLACK
<0/0.667/2>
VALUE 0
VALUE .667
VALUE 1.334
VALUE 2.001
< &TESTING SLACK >
AMPERSAND
FASP, FORTRAN Alphameric Subroutine Package Page 262
DANUMB, Integer Number FORTRAN Output Routine
DDDDDDD AAA NNN NNN UUU UUU MMM MMM BBBBBBBBB
DDD DDD AAA AAA NNNN NNN UUU UUU MMMM MMMM BBB BBB
DDD DDD AAA AAA NNN N NNN UUU UUU MMM MM MMM BBBBBBB
DDD DDD AAAAAAAAA NNN NNNN UUU UUU MMM MMM BBB BBB
DDDDDDD AAA AAA NNN NNN UUUUUUUUU MMM MMM BBBBBBBBB
DANUMB, Integer Number FORTRAN Output Routine
------ ------- ------ ------- ------ -------
DANUMB represents an integer value so that it can be written
with a FORTRAN format statement containing a multiple of an
A1 alphameric specification. DANUMB is much shorter than
the free format floating point output routine DARITE, both
because fewer options are provided by DANUMB and because
integer numbers are easier to represent.
DANUMB can represent any integer which can be calculated by
a computer which uses two's complement notation, including
the largest negative number which consists of the sign bit
on and all other bits off and for which the absolute value
cannot be stored. If DANUMB is used to generate FORTRAN
source code, as for example when used by the integer data
statement generator, it is possible that the FORTRAN
compiler will not accept this largest negative number since
the compiler may evaluate a negative number by negating its
absolute value. Similarly, on computers which take the sign
bit on and all other bits off to mean negative zero, DANUMB
will produce the representation of positive zero for the
value negative zero.
DANUMB has no output argument used to signal an error
condition. Instead, the character count KOUNT is returned
unchanged if the representation of the number will not fit
into the buffer provided.
FASP, FORTRAN Alphameric Subroutine Package Page 263
DANUMB, Integer Number FORTRAN Output Routine
The DANUMB Argument List
--- ------ -------- ----
The argument list of routine DANUMB is
SUBROUTINE DANUMB(KONTRL,NUMBER,IRADIX,LETTER,
1KOUNT,LFTCOL,MAX)
with the associated DIMENSION statement
DIMENSION LETTER(MAX)
The following are input arguments left unchanged
KONTRL = 0, left justify the representation of the number.
If KOUNT is less than LFTCOL, then the output
buffer starting at LETTER(KOUNT+1) and extending
through LETTER(LFTCOL) will be filled with spaces.
If KOUNT is less than or equal to LFTCOL, then the
leftmost character, which will be either a minus
sign or the leftmost digit of the representation of
the number, will be placed in LETTER(LFTCOL+1).
If KOUNT is greater than LFTCOL meaning that
characters are already in use to the right of
LFTCOL, then a space will be placed in
LETTER(KOUNT+1), and the leftmost digit of the
representation of the number will be placed in
LETTER(KOUNT+2).
If the representation of the number will fit into
the output buffer, then the character count KOUNT
will be left pointing to the location of the
rightmost digit of the representation of the
number. If the representation of the number would
extend to the right of LETTER(MAX), then the
character count KOUNT is returned unchanged.
= 1, right justify the representation of the number
so that the rightmost digit is placed in
LETTER(LFTCOL).
If the representation of the number requires less
than LFTCOL-KOUNT characters, then fill the output
buffer with spaces starting at LETTER(KOUNT+1) and
extending up to the leftmost character of the
representation of the number.
If the representation of the number will fit into
the output buffer, then the character count KOUNT
will be returned equal to LFTCOL. If LFTCOL is
greater than the dimension of the buffer, or if the
FASP, FORTRAN Alphameric Subroutine Package Page 264
DANUMB, Integer Number FORTRAN Output Routine
number is too large to be represented starting at
LETTER(KOUNT+1) and ending at LETTER(LFTCOL), then
the character count KOUNT is returned unchanged.
NUMBER = the integer value to be represented.
IRADIX = the radix with which the value is to be
represented. IRADIX can vary from 2 through 10,
but IRADIX values 2 for binary, 8 for octal, and 10
for decimal representations are the most
reasonable. It must be remembered that for
negative values of the argument NUMBER, it is the
value of the argument NUMBER which is represented,
not the particular sequence of bits by which the
computer stores this argument. Thus the value
decimal -8 would be represented as -10 for IRADIX=8
and as -1000 for IRADIX=2 even if the computer uses
two's complement notation to store negative values.
The following argument is an array to which the
representation of the value is appended.
LETTER = the output buffer into which DANUMB will place the
representation of the number and which can be
written by a multiple of an A1 format by the
calling program.
The following argument is used for both input to, and output
from this routine.
KOUNT = is input containing the number of characters
already in use in the output buffer LETTER array.
If the representation of the value will fit into
the letter array, then KOUNT will be returned
containing the location within the LETTER array of
the rightmost character of the representation of
the value. If the representation of the value will
not fit into the LETTER array, then KOUNT is
returned unchanged.
The following are input arguments which describe the size of
the output text buffer. These arguments are returned
unchanged.
LFTCOL = number of characters in the buffer which are to be
to the left of the representation of the number if
KONTRL equals 0.
= position in the buffer at which is to be placed the
right digit of the representation of the number if
KONTRL equals 1.
MAX = the dimension of the output buffer LETTER array.
The character count KOUNT will be returned
FASP, FORTRAN Alphameric Subroutine Package Page 265
DANUMB, Integer Number FORTRAN Output Routine
unchanged if the representation of the value would
extend beyond LETTER(MAX).
An Example of the use of DANUMB
-- ------- -- --- --- -- ------
The following sample program demonstrates the manner in
which DANUMB is called.
DIMENSION NUMBER(5),LETTER(50)
DATA NUMBER/-123456,-123,0,123,123456/
C
C PRODUCE MARGIN
NMB=0
KOUNT=0
1 NMB=NMB+1
IF(NMB.GT.9)NMB=0
J=KOUNT
CALL DANUMB(0,NMB,10,LETTER,KOUNT,KOUNT,50)
IF(J.LT.KOUNT)GO TO 1
TYPE 2,LETTER
2 FORMAT(1X,50A1)
C
C PRODUCE LEFT JUSTIFIED OUTPUT
KOUNT=0
DO 3 I=1,5
J=KOUNT
CALL DANUMB(0,NUMBER(I),10,LETTER,KOUNT,0,50)
IF(J.GE.KOUNT)GO TO 4
3 CONTINUE
4 IF(KOUNT.GT.0)TYPE 2,(LETTER(I),I=1,KOUNT)
C
C PRODUCE OUTPUT RIGHT JUSTIFIED IN WIDTH 10 FIELDS
KOUNT=0
DO 5 I=1,5
LFTCOL=KOUNT+10
CALL DANUMB(1,NUMBER(I),10,LETTER,KOUNT,LFTCOL,50)
5 CONTINUE
TYPE 2,(LETTER(I),I=1,KOUNT)
STOP
END
The output generated by the above program is shown below.
12345678901234567890123456789012345678901234567890
-123456 -123 0 123 123456
-123456 -123 0 123 123456
FASP, FORTRAN Alphameric Subroutine Package Page 266
DAPAIR, Returns an Index and an Associated Value
DDDDD AAA PPPPPP AAA IIIIII RRRRRR
DD DD AAAA PP PP AAAA II RR RR
DD DD AA AA PP PP AA AA II RR RR
DD DD AA AA PPPPPP AA AA II RRRRRR
DD DD AAAAAAA PP AAAAAAA II RR RR
DD DD AA AA PP AA AA II RR RR
DDDDD AA AA PP AA AA IIIIII RR RR
DAPAIR, Returns an Index and an Associated Value
------ ------- -- ----- --- -- ---------- -----
DAPAIR returns to the calling program a pair of values
represented by the characters in an input buffer read by the
calling program with a multiple of an A1 format. The pair
of values is assumed to consist of an index (or a
subscript), which is represented first in the input buffer,
and of a value to be associated with this index. If
subsequent calls to this routine are to assign a sequence of
values with a constant increment to either the index or the
associated value, then the representation of a single index
or of a single range of indexes is followed by the
representations of the several associated values and/or of
the several ranges of associated values. If both numbers
are specified as members of ranges, then the first call to
this routine returns the first value of the first range as
the index and the first value of the second range as the
associated value, the second call to this routine returns
the second value of the first range as the index and the
second value of the second range as the associated value,
and so on. If the range of indexes is exhausted prior to
the generation of the final member of the range of
associated values, then the final value of the index is
returned unchanged with the subsequent associated values,
but the calling program is informed of this condition and
can adjust the index as necessary. Once both the range of
indexes and the current range of associated values have been
exhausted, then the following representation of a single
number or of a range of numbers can specify the indexes of a
subsequent pair, or can specify additional values to be
associated with the same final index.
Spaces and/or tabs are allowed between the representations
of numbers and/or of number ranges. The calling program can
prohibit the appearance of printing characters between the
representations of the index and its associated value, or
between the representations of subsequent associated values,
in which case a comma indicates that the following number is
to be taken as a new index, or else the calling program can
allow commas between the index and the associated value or
between subsequent associated values, in which case 2
adjacent commas indicate a missing associated value.
The ranges of values to be assigned to the indexes or to the
FASP, FORTRAN Alphameric Subroutine Package Page 267
DAPAIR, Returns an Index and an Associated Value
associated values can be specified in either slash or colon
or asterisk notations. In asterisk notation, a value to be
repeated a given number of times is preceded by the number
of times and an asterisk. In slash or colon notation (the 2
characters are equivalent), a value which is to start at a
specified initial value, then be varied until it reaches a
final bound is expressed as the initial value, the increment
by which the value is to be varied, and the final bound
separated by slashes or by colons. If either the initial
value or the final bound is missing in slash or colon
notation, then the missing value is assumed to be zero. If
the increment is not included in slash or colon notation,
then the increment is assumed to be 1 and only a single
slash or a single colon is needed to separate the initial
value and the final bound. For example, the text
1/3 4 5/7 8/10 4*11
would specify the 4 pairs 1 and 4, 2 and 5, 3 and 6, and an
unchanged 3 and 7. Depending upon how this routine is
called, the next pairs returned could be 3 and 8, 3 and 9, 3
and 10, and then 4 appearances of 3 and 11. The routine
could instead inform the calling program that a new index
was found but not evaluated, then subsequent calls would
return the pairs 8 and 11, 9 and 11, 10 and 11, and a final
unchanged 10 and 11. A semicolon (or a comma if not allowed
between adjacent associated values) appearing between the 7
and 8 would force the latter interpretation, since the
appearance of a semicolon (or of a comma if not allowed
between adjacent associated values) causes the following
number to be interpreted as the index of a subsequent pair.
An exclamation point and anything to its right are taken to
be a comment and are otherwise ignored. An ampersand and
anything to its right are also ignored, but the text being
interpreted is assumed to continue with that supplied to the
subsequent call of this routine. Additional ranges of
associated values can appear in the text next read into the
input buffer after an ampersand has been found, but the
specification of a single range cannot be continued in this
manner across a line boundary. For example, the pair
specification shown above could instead have been given as
1/3 4 5/7&THIS IS THE FIRST LINE
8/10 4*11!THIS IS THE SECOND LINE
Two versions of the routine are supplied. DAPAIR requires
the routines DANEXT and DASPAN for the evaluation of range
specifications. DAIPAR cannot evaluate ranges, but might be
sufficient if the index varies in a predetermined manner,
and if ranges are not needed for the specification of the
associated values.
FASP, FORTRAN Alphameric Subroutine Package Page 268
DAPAIR, Returns an Index and an Associated Value
The DAPAIR and DAIPAR Argument Lists
--- ------ --- ------ -------- -----
The argument lists of routines DAPAIR and DAIPAR are
SUBROUTINE DAPAIR(SLACK ,KNDBGN,KNDEND,KONECT,IBUFFR,
1 MAXBFR,LOWBFR,KIND ,NEWBGN,NEWEND,BGNNEW,ENDNEW,
2 IWHERE,INCBGN,LMTBGN,BGNINC,BGNLMT,INCEND,LMTEND,
3 ENDINC,ENDLMT)
and
SUBROUTINE DAIPAR( KNDBGN,KNDEND,KONECT,IBUFFR,
1 MAXBFR,LOWBFR,KIND ,NEWBGN,NEWEND,BGNNEW,ENDNEW)
both with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The argument lists of these routines are identical with the
exception that the first argument and the last 9 arguments
of DAPAIR do not appear in the DAIPAR argument list. In
addition, the argument KIND will not be assigned its 3
highest possible values by DAIPAR. These missing arguments
and the missing values of KIND are all used in the
evaluation of value ranges and so are not needed by DAIPAR.
The following arguments are used for input only, and are
returned unchanged.
SLACK = the fraction of the increment of a real number (not
integer) range expressed in slash or colon
notation, by which the returned value can exceed
the final bound and still be considered to be
within the range. SLACK is necessary because of
the imprecision of calculations involving
fractional real numbers. For example, if the user
wished to have the increment be 0.666... with an
initial value of 3.0 and a final bound of 5.0, then
the user might type 3/.667/5 and expect that the
numbers 3.000, 3.667, 4.333, and 5.000 would be
included within the range, while, actually, the
number near 5.00 would be calculated to be 5.001
and would therefore be outside the range. Although
SLACK can be set to zero, a value of about 0.01 is
recommended. In the above example, SLACK=0.01
would cause the actual final bound to be 5.00667 so
that the number calculated near the final bound
would be within the range. SLACK is ignored if
numbers in the range are returned as integers.
KNDBGN = specifies whether the index is returned as the
integer argument NEWBGN or as the real argument
BGNNEW. The number can be typed with a decimal
FASP, FORTRAN Alphameric Subroutine Package Page 269
DAPAIR, Returns an Index and an Associated Value
point and/or an exponent regardless of the value of
KNDBGN.
= -1, the index is calculated as an octal integer and
is returned as the argument NEWBGN. However, the
number following the letter E of an exponent is
evaluated in decimal.
= 0, the index is calculated as a decimal integer and
is returned as the argument NEWBGN.
= 1 or greater, the index is returned as the real
argument BGNNEW. If possible, the real number will
be accumulated as an integer, then be converted to
real and shifted as necessary. KNDBGN is the
maximum number of digits in the integer.
KNDEND = specifies whether the associated value is returned
as the integer argument NEWEND or as the real
argument ENDNEW. The number can be typed with a
decimal point and/or an exponent regardless of the
value of KNDEND.
= -1, the associated value is calculated as an octal
integer and is returned as the argument NEWEND.
However, the number following the letter E of an
exponent is evaluated in decimal.
= 0, the associated value is calculated as a decimal
integer and is returned as the argument NEWEND.
= 1 or greater, the associated value is returned as
the real argument ENDNEW. If possible, the real
number will be accumulated as an integer, then be
converted to real and shifted as necessary. KNDEND
is the maximum number of digits in the integer.
KONECT = specifies the manner of interpretation of a
following number when the range of indexes and the
range of associated values have both been
exhausted. KONECT also specifies whether commas
are to indicate that the following number, if any,
is to be the index of a subsequent pair, or whether
commas are allowed between subsequent values
associated with the same index or range of indexes.
Regardless of the value of KONECT, all commas
appearing to the left of an index are ignored.
= -1, 0 or 1, commas cannot appear between the
specification of an index and of its associated
value and cannot appear between the representations
of subsequent associated values. If a comma is
found, then the following number, if any, is taken
to be the index of a subsequent pair of numbers.
= -1, when the range of indexes and the range of
associated values have both been exhausted, a
following representation of a number or of a range
of numbers on the same line is taken to specify
additional associated values.
If an end of line is found before the range of
FASP, FORTRAN Alphameric Subroutine Package Page 270
DAPAIR, Returns an Index and an Associated Value
indexes has been exhausted, then the calling
program can request (by setting KIND=4 to simulate
the finding of an ampersand) that the evaluation of
associated values is to be continued by the next
call to this routine in the new text read into the
input buffer. If an end of line is found before
the range of indexes has been exhausted, but the
calling program leaves KIND unchanged before again
calling this routine, then the remaining index
values will be returned without associated values,
following which the first number in the new
contents of the text buffer will be interpreted as
the index of a new pair.
If an ampersand is found to the right of an index
specification, regardless of whether the range of
indexes has been exhausted, then the evaluation of
associated values will be continued by the next
call to this routine in the new text read into the
input buffer. If an ampersand is found prior to
the exhaustion of the range of indexes, but the
next line read into the input buffer is to start a
new pair, then KIND should be set to 6 by the
calling program causing the remaining indexes to be
generated first but marked as having missing
associated values. If the ampersand is found after
the exhaustion of the range of indexes, but the
next line read into the input buffer is to start a
new pair, then KIND should be set to one.
= 0, when the range of indexes and the range of
associated values have both been exhausted, a
following representation of a number or of a range
of numbers on the same line causes this routine to
report a missing comma, and the subsequent call to
this routine then begins the evaluation of this
following number as an index or range of indexes of
a subsequent pair.
If an end of line is found before the range of
indexes has been exhausted, then the calling
program can request (by setting KIND=4 to simulate
the finding of an ampersand) that the evaluation of
associated values is to be continued by the next
call to this routine in the new text read into the
input buffer. If an end of line is found before
the range of indexes has been exhausted, but the
calling program leaves KIND unchanged before again
calling this routine, then the remaining index
values will be returned without associated values,
following which the first number in the new
contents of the text buffer will be interpreted as
the index of a subsequent pair.
FASP, FORTRAN Alphameric Subroutine Package Page 271
DAPAIR, Returns an Index and an Associated Value
If an ampersand is found prior to the exhaustion of
the range of indexes, then the evaluation of
associated values will be continued by the next
call to this routine in the new text read into the
input buffer unless the calling program sets KIND
to 6 prior to the subsequent call to this routine,
causing the remaining indexes to be generated but
marked as having missing associated values. If an
ampersand is found after the range of indexes and
the range of associated values have both been
exhausted, then the first number evaluated by the
subsequent call to this routine in the new text
read into the input buffer is assumed to be the
index of a new pair of values.
= 1, when the range of indexes and the range of
associated values have both been exhausted, a
following representation of a number or of a range
of numbers on the same line is taken to specify
additional associated values.
If either an ampersand or an end of line is found
before the range of indexes has been exhausted,
then the remaining index values will be returned
without associated values, following which the
calling program will be informed that the end of
line or ampersand has been reached.
= 2, 3 or 4, same as when KONECT=-1, 0 or 1
respectively, except that commas can appear (but
are not required) between the representations of
the index and the first associated value and
between the representations of subsequent
associated values. An extra comma indicates a
missing associated value. If KONECT=3, and if the
range of indexes and the range of associated values
have both been exhausted, then KIND will be
returned set to 10 to indicate a missing semicolon
if a following number is found whether or not this
following number is preceded by a comma, and the
subsequent call to this routine will begin the
evaluation of the following number as an index or
range of indexes of a subsequent pair. Unlike
commas, semicolons are always reported to the
calling program regardless of the value of KONECT.
IBUFFR = input buffer array containing characters typed by
the user, read by a multiple of an A1 format, which
is to be searched for number specifications.
IBUFFR then contains one character per computer
storage location.
MAXBFR = maximum subscript of the IBUFFR array locations to
be searched.
FASP, FORTRAN Alphameric Subroutine Package Page 272
DAPAIR, Returns an Index and an Associated Value
The following arguments are used both for input to this
routine, and for output to the calling program.
LOWBFR = input containing the subscript within the IBUFFR
array of the first (leftmost) character which can
be scanned for a number specification. LOWBFR is
returned pointing to the next character which would
be interpreted by the subsequent call to this
routine if this subsequent call needs to evaluate
further number specifications. LOWBFR is returned
unchanged if this routine merely generates the next
associated value within a previously determined
range. If there are no printing characters at or
to the right of IBUFFR(LOWBFR), but the
interpretation of additional characters is
attempted, then LOWBFR will be returned containing
MAXBFR+1. LOWBFR must be reset by the calling
program to point to the first character in the
IBUFFR array each time a new line of text is read.
KIND = input containing -1, an error was detected by the
calling program (rather than by this routine) in
the results returned by the previous call to this
routine and any additional numbers through the next
appearance of a comma (unless KONECT is greater
than 1) or a semicolon or an ampersand or an
unknown character or the end of line are to be
ignored. The next number to the right of the comma
or on the line following the ampersand or end of
line is to be interpreted as the index of the next
pair. An unknown character or a semicolon is
reported to the calling program if it is found
instead.
= input containing 0, this routine has not yet been
called during the execution of the current logical
section of the calling program. This routine is to
begin evaluation of the contents of the input text
buffer, not continue the generation of ranges
evaluated by previous calls to this routine.
LOWBFR should point to the first character in the
IBUFFR array to be interpreted.
= input greater than zero, the value of KIND
indicates the condition returned by the previous
call to this routine and in most cases has not been
modified by the calling program. KIND being
returned with one of the values 7, 8, 9, 10 (if
KONECT is 2 or 4), 13 and 14 indicates that a pair
of values is being returned, although either or
both members of the pair could be missing. The
other values of KIND indicate special conditions
for which a pair of numbers is not returned.
= 1, nothing, except perhaps leading commas or else a
comment indicated by a leading exclamation point,
was found at or to the right of IBUFFR(LOWBFR).
FASP, FORTRAN Alphameric Subroutine Package Page 273
DAPAIR, Returns an Index and an Associated Value
The calling program should read a new line of text
into IBUFFR.
= 2, a semicolon was found as the first printing
character at or to the right of IBUFFR(LOWBFR). If
the range of indexes has not been exhausted when
the semicolon is found, then the remaining members
of the range are reported to the calling program
before the semicolon is reported. It is assumed
that the calling program will treat the semicolon
as marking the end of a logical group of pair
specifications if KONECT is not greater than 1, or
the end of a single pair specification if KONECT is
greater than 1.
= 3, an ampersand was found while searching for the
next range of indexes. LOWBFR is returned pointing
beyond the end of the buffer. The calling program
should read a new line of text into the input
buffer before again calling this routine. This
value of KIND is not returned if an ampersand is
found while searching for a range of associated
values.
= 4, the range of indexes was not exhausted, but an
ampersand was found instead of a range of
associated values. The calling program should read
a new line of text into the buffer before again
calling this routine. The subsequent call to this
routine will continue the evaluation of associated
ranges in the new contents of the buffer. If a
semicolon (or a comma if KONECT is not greater than
1) is found at the start of the new contents of the
buffer, then the remaining index values will be
returned to the calling program but the calling
program will be informed that the associated values
are missing. If the subsequent call to this
routine is instead to begin the evaluation of a new
set of leading indexes and associated values, then
the calling program should set KIND to 6 to
simulate the lack of an ampersand, causing the
remaining indexes to be generated, but marked as
having missing associated values. If the remaining
indexes are not wanted, then KIND should be set to
one.
= 5, the range of indexes was exhausted, but an
ampersand was found instead of a range of
associated values. The calling program should read
a new line of text into the buffer before again
calling this routine. The subsequent call to this
routine will continue the evaluation of associated
ranges in the new contents of the buffer. Since
the range of indexes has been exhausted, no
additional missing values will be reported if a
semicolon (or a comma if KONECT is not greater than
1) is found at the start of the new contents of the
buffer. If the subsequent call to this routine is
FASP, FORTRAN Alphameric Subroutine Package Page 274
DAPAIR, Returns an Index and an Associated Value
always to begin the evaluation of a new set of
leading indexes and associated values, then the
calling program should set KIND to one.
= 6, an end of line was found but the range of
indexes was not exhausted. If KIND is sent
unchanged to the subsequent call of this routine,
then the remaining indexes will be returned to the
calling program, but the calling program will be
informed that the associated values are missing.
If the subsequent call to this routine is instead
to continue the evaluation of associated ranges in
the new contents of the buffer read by the calling
program before this routine is again called, then
KIND should be set to 4 to simulate the appearance
of an ampersand at the end of the preceding line.
If the remaining indexes are not wanted, then KIND
should be set to one.
= 7, both an index and an associated value are being
returned to the calling program.
= 8, an index is being returned to the calling
program, but the associated value is missing.
NEWEND and ENDNEW are returned undefined.
= 9, an associated value is being returned to the
calling program, but the range of indexes has been
exhausted. NEWBGN or BGNNEW, whichever is
appropriate, is returned unchanged (but can be
modified as desired by the calling program).
= 10, the ranges of both the indexes and the
associated values have been exhausted. This value
is never returned if KONECT=-1 or 1.
If KONECT has either of the values 2 or 4, then the
ranges of both the indexes and the associated
values have been exhausted, but an extra comma was
found indicating a missing associated value.
NEWBGN or BGNNEW, whichever is appropriate, is
returned unchanged (but can be modified as desired
by the calling program).
If KONECT=0, then the ranges of both the indexes
and associated values have been exhausted, but a
following number was found which was not preceded
by a comma. KIND should be set to -1 by the
calling program before this routine is called next
if the lack of a comma indicates a serious enough
error that all numbers should be ignored until the
next comma is found. If KONECT=3, then the ranges
of both the indexes and associated values have been
exhausted, but a following number was found which
was not preceded by a semicolon, but which may or
may not have been preceded by one or more commas.
KIND should be set to -1 by the calling program
before this routine is called next if the lack of a
semicolon indicates a serious enough error that all
FASP, FORTRAN Alphameric Subroutine Package Page 275
DAPAIR, Returns an Index and an Associated Value
numbers should be ignored until the next semicolon
is found. If KONECT is 0 or 3, then LOWBFR is
returned pointing to the first character of the
number specification, and the number will be
evaluated as an index if this routine is called
again without KIND having been changed.
= 11, an unknown character was found. If the range
of indexes has not been exhausted when the unknown
character is found, then the remaining members of
the range are reported to the calling program
before the unknown character is reported. LOWBFR
is returned pointing to the next character to the
right of the unknown character. If this routine is
called without changing the value of KIND, then the
number to the right of the unknown character will
be interpreted as the index of the next pair.
= 12, an error was found within a range
specification. The finding of such an error
terminates the generation of the range of indexes.
IWHERE is returned pointing to the character at the
start of the specification containing the error.
LOWBFR is returned pointing to the character to the
right of the specification containing the error.
If this routine is called without changing the
value of KIND, then the number to the right of the
unknown character will be interpreted as the index
of the next pair.
= 13, an index and the associated value zero are
being returned to the calling program. The zero
was indicated by the lack of a number to the right
of an asterisk. A missing number to the right of
an asterisk in the specification of a range of
indexes is not allowed and would return KIND
containing the value 12.
= 14, same as KIND=13 except that the range of
indexes was exhausted. NEWBGN or BGNNEW, whichever
is appropriate, is returned unchanged (but can be
modified as desired by the calling program).
NEWBGN = returned containing the next value of an integer
index if KIND is returned containing one of the
values 7, 8 or 13 and if KNDBGN is less than or
equal to zero. NEWBGN should not be changed by the
calling program unless KIND is returned containing
one of the values 9, 10 or 14 indicating that the
range of indexes has been exhausted.
NEWEND = returned containing the next associated integer
value if KIND is returned with one of the values 7,
9, 13 or 14 and if KNDEND is less than or equal to
zero. NEWEND can be returned changed by this
routine even if an associated value is not found
but then must not be changed by the calling
program. If the range of associated values has
FASP, FORTRAN Alphameric Subroutine Package Page 276
DAPAIR, Returns an Index and an Associated Value
been exhausted, or if the associated value is
indicated by an extra comma to be missing, then
NEWEND does not contain an associated integer value
but instead indicates to the subsequent call to
this routine the reason why a value could not be
returned.
BGNNEW = returned containing the next value of a real index
if KIND is returned containing one of the values 7,
8 or 13 and if KNDBGN is greater than zero. BGNNEW
should not be changed by the calling program unless
KIND is returned containing one of the values 9, 10
or 14 indicating that the range of indexes has been
exhausted.
ENDNEW = returned containing the next associated real value
if KIND is returned with one of the values 7, 9, 13
or 14 and if KNDEND is greater than zero. ENDNEW
should not be changed by the calling program.
The following argument is used for output only, and is not
included in the DAIPAR argument list.
IWHERE = returned pointing to the left character of a range
notation which was found to contain an error such
as too many asterisks, too many slashes, too many
colons, or the absence of a number to the right of
an asterisk in the specification of a range of
leading indexes.
The remaining DAPAIR arguments are used only during the
generation of values within ranges and should be sent to the
subsequent DAPAIR call unchanged. These arguments should be
stated by name in the call statements and should be ignored
by the calling program except insofar as is necessary to
transmit their values to the subsequent call to this
routine.
Values Returned for Typical Range Specification Combinations
------ -------- --- ------- ----- ------------- ------------
Shown on the following pages are results returned by DAPAIR
during the evaluation of several typical lines of input
text. The line of text is shown first, followed by a
description of the results obtained with KONECT values of -1
(left column), 0 (center column) and 1 (right column). The
returned values of the argument named KIND are indicated by
the following descriptions in which the digits 1 and 2
represent the returned index and returned associated value
respectively, and in which # represents any number. The
index is enclosed in parentheses if the index range has been
exhausted but associated values are still being found.
FASP, FORTRAN Alphameric Subroutine Package Page 277
DAPAIR, Returns an Index and an Associated Value
KIND=1, empty KIND= 7, 1 and 2
=2, ; = 8, 1 and missing #
=3, & = 9, (1) and 2
=4, & but # needed =10, unexpected # (KONECT=0,3)
=5, & but # expected =10, (1) and missing #(KONECT=2,4)
=6, empty but # needed
In the examples shown on the following pages, it is assumed
that KIND will be changed by the calling program to have the
value 4 whenever it is returned set to 6 by this routine.
The initial values returned during the evaluation of a
subsequent line when KIND is input set to 4 are indicated by
"# next line" rather than by merely a #.
KONECT=-1 KONECT=0 KONECT=1
!
empty empty empty
&
& & &
,
empty empty empty
,&
& & &
1
empty but # needed empty but # needed 1 and missing #
1 and # next line 1 and # next line empty
1!FOLLOWED BY A LINE BEGINNING WITH A COMMA
empty but # needed empty but # needed
1 and missing # 1 and missing # (same as above)
values after comma values after comma
1&
& but # needed & but # needed 1 and missing #
1 and # next line 1 and # next line &
1&FOLLOWED BY A LINE BEGINNING WITH A COMMA
& but # needed & but # needed
1 and missing # 1 and missing # (same as above)
values after comma values after comma
1,
1 and missing # 1 and missing # 1 and missing #
empty empty empty
1,&
1 and missing # 1 and missing # 1 and missing #
& & &
FASP, FORTRAN Alphameric Subroutine Package Page 278
DAPAIR, Returns an Index and an Associated Value
1 2
1 and 2 1 and 2 1 and 2
empty empty empty
1 2&
1 and 2 1 and 2 1 and 2
& but # expected & &
(1) and # next line
1 2&FOLLOWED BY A LINE BEGINNING WITH A COMMA
1 and 2
& but # expected (same as above) (same as above)
values after comma
1 2,
1 and 2 1 and 2 1 and 2
empty empty empty
1 2,&
1 and 2 1 and 2 1 and 2
& & &
1 2 3
1 and 2 1 and 2 1 and 2
(1) and 3 unexpected # (1) and 3
empty empty but # needed empty
3 and # next line
1 2 3!FOLLOWED BY A LINE BEGINNING WITH A COMMA
1 and 2
(same as above) unexpected # (same as above)
empty but # needed
3 and missing #
values after comma
1 2 3&
1 and 2 1 and 2 1 and 2
(1) and 3 unexpected # (1) and 3
& but # expected & but # needed &
(1) and # next line 3 and # next line
1 2 3&FOLLOWED BY A LINE BEGINNING WITH A COMMA
1 and 2 1 and 2
(1) and 3 unexpected # (same as above)
& but # expected & but # needed
values after comma 3 and missing #
values after comma
1 4/6
1 and 4 1 and 4 1 and 4
(1) and 5 (1) and 5 (1) and 5
(1) and 6 (1) and 6 (1) and 6
empty empty empty
FASP, FORTRAN Alphameric Subroutine Package Page 279
DAPAIR, Returns an Index and an Associated Value
1 4/6&
1 and 4 1 and 4 1 and 4
(1) and 5 (1) and 5 (1) and 5
(1) and 6 (1) and 6 (1) and 6
& but # expected & &
(1) and # next line
1 4/6&FOLLOWED BY A LINE BEGINNING WITH A COMMA
1 and 4
(1)and 5 (same as above) (same as above)
(1) and 6
& but # expected
values after comma
1 4/6 7/9
1 and 4 1 and 4 1 and 4
(1) and 5 (1) and 5 (1) and 5
(1) and 6 (1) and 6 (1) and 6
(1) and 7 unexpected # (1) and 7
(1) and 8 empty but # needed (1) and 8
(1) and 9 7 and # next line (1) and 9
empty 8 and # next line empty
9 and # next line
1 4/6 7/9!FOLLOWED BY A LINE BEGINNING WITH A COMMA
1 and 4
(same as above) (1) and 5 (same as above)
(1) and 6
unexpected #
empty but # needed
7 and missing #
8 and missing #
9 and missing #
values after comma
1 4/6 7/9&
1 and 4 1 and 4 1 and 4
(1) and 5 (1) and 5 (1) and 5
(1) and 6 (1) and 6 (1) and 6
(1) and 7 unexpected # (1) and 7
(1) and 8 & but # needed (1) and 8
(1) and 9 7 and # next line (1) and 9
& but # expected 8 and # next line &
(1) and # next line 9 and # next line
1 4/6 7/9&FOLLOWED BY A LINE BEGINNING WITH A COMMA
1 and 4 1 and 4
(1) and 5 (1) and 5 (same as above)
(1) and 6 (1) and 6
(1) and 7 unexpected #
(1) and 8 & but # needed
(1) and 9 7 and missing #
& but # expected 8 and missing #
values after comma 9 and missing #
values after comma
FASP, FORTRAN Alphameric Subroutine Package Page 280
DAPAIR, Returns an Index and an Associated Value
1/3 4
1 and 4 1 and 4 1 and 4
empty but # needed empty but # needed 2 and missing #
2 and # next line 2 and # next line 3 and missing #
3 and # next line 3 and # next line empty
1/3 4!FOLLOWED BY A LINE BEGINNING WITH A COMMA
1 and 4 1 and 4
empty but # needed empty but # needed (same as above)
2 and missing # 2 and missing #
3 and missing # 3 and missing #
values after comma values after comma
1/3 4&
1 and 4 1 and 4 1 and 4
& but # needed & but # needed 2 and missing #
2 and # next line 2 and # next line 3 and missing #
3 and # next line 3 and # next line &
1/3 4&FOLLOWED BY A LINE BEGINNING WITH A COMMA
1 and 4 1 and 4
& but # needed & but # needed (same as above)
2 and missing # 2 and missing #
3 and missing # 3 and missing #
values after comma values after comma
1/3 4/6 ! or 1/3 4 5 6 or 1/3 4/5 6 or 1/3 4 5/6
1 and 4 1 and 4 1 and 4
2 and 5 2 and 5 2 and 5
3 and 6 3 and 6 3 and 6
empty empty empty
1/3 4 7/9 10 13
1 and 4 1 and 4 1 and 4
2 and 7 2 and 7 2 and 7
3 and 8 3 and 8 3 and 8
(3) and 9 (3) and 9 (3) and 9
(3) and 10 unexpected # (3) and 10
(3) and 13 10 and 13 (3) and 13
empty empty empty
Each of the examples shown above which does not include
commas would be interpreted the same for KONECT values 2, 3
and 4 as for -1, 0 and 1 respectively. The rest of the
examples would also be interpreted the same for KONECT
values of 2, 3 and 4 as for -1, 0 and 1 respectively if the
commas are replaced by semicolons. The examples shown below
illustrate the results obtained using the higher KONECT
values when commas appear to the right of the index. Commas
appearing to the left of the index are ignored for all
KONECT values.
FASP, FORTRAN Alphameric Subroutine Package Page 281
DAPAIR, Returns an Index and an Associated Value
KONECT=2 KONECT=3 KONECT=4
1/3 4 5 6! OR 1/3,4,5,6
1 and 4 1 and 4 1 and 4
2 and 5 2 and 5 2 and 5
3 and 6 3 and 6 3 and 6
empty empty empty
1/3,,5 6! OR 1/3,,5,6
1 and missing # 1 and missing # 1 and missing #
2 and 5 2 and 5 2 and 5
3 and 6 3 and 6 3 and 6
empty empty empty
1/3 4,,6! OR 1/3,4,,6
1 and 4 1 and 4 1 and 4
2 and missing # 2 and missing # 2 and missing #
3 and 6 3 and 6 3 and 6
empty empty empty
1/3 4;! OR 1/3,4; OR 1/3 4,; OR 1/3,4,; OR 1/3 4,,;
1 and 4 1 and 4 1 and 4
2 and missing # 2 and missing # 2 and missing #
3 and missing # 3 and missing # 3 and missing #
; ; ;
empty empty empty
1;1,4;1,4,5;,,1,4,,;,,1,4,,6,,8!WITH NUMBER ON NEXT LINE
1 and missing # 1 and missing # 1 and missing #
; ; ;
1 and 4 1 and 4 1 and 4
; ; ;
1 and 4 1 and 4 1 and 4
(1) and 5 unexpected # (1) and 5
; 5 and missing # ;
1 and 4 ; 1 and 4
(1) and missing # 1 and 4 (1) and missing #
(1) and missing # ; (1) and missing #
; 1 and 4 ;
1 and 4 unexpected # 1 and 4
(1) and missing # 6 and missing # (1) and missing #
(1) and 6 unexpected # (1) and 6
(1) and missing # empty but # needed (1) and missing #
(1) and 8 8 and # next line (1) and 8
empty empty
1/3 4!FOLLOWED BY LINE WITH OPTIONAL COMMA AND 2 NUMBERS
1 and 4 1 and 4 1 and 4
empty but # needed empty but # needed 2 and missing #
2 and # next line 2 and # next line 3 and missing #
3 and # next line 3 and # next line empty
FASP, FORTRAN Alphameric Subroutine Package Page 282
DAPAIR, Returns an Index and an Associated Value
1/3 4,!FOLLOWED BY LINE BEGINNING WITH 2 NUMBERS
(same as above) (same as above) (same as above)
1/3 4,!FOLLOWED BY LINE BEGINNING WITH COMMA THEN NUMBER
1 and 4 1 and 4 1 and 4
empty but # needed empty but # needed 2 and missing #
2 and missing # 2 and missing # 3 and missing #
3 and # next line 3 and # next line empty
An Example of the Use of DAPAIR
-- ------- -- --- --- -- ------
The program listed on the following pages reports the
results returned during the processing of a line of text for
3 values of KONECT. Processing of the line of text is
discontinued for the particular value of KONECT when the
routine returns either an ampersand or an end of line
indication. If an end of line is found when an associated
value is needed, then an ampersand is simulated. The user
is asked to supply a new line of text when processing of the
previous line of text has been completed for all 3 values of
KONECT. An asterisk is used to prompt the user to type the
line of text if an empty indication has been returned for
all 3 values of KONECT. An ampersand is used to prompt the
user if the routine has indicated the need for a
continuation line, in which case processing continues only
for the values of KONECT which requested the continuation.
The table of returned values on the preceding pages is a
slightly edited dialog between the user and this program.
DIMENSION IBUFFR(72),JBUFFR(72),IFDONE(3)
C ARRAYS USED TO GENERATE TEXT DESCRIPTIONS
DIMENSION LGND1 ( 5),LGND4 (14),LGND5 (16),
1LGND6 (18),LGND7 ( 5),LGND8 ( 9),LGND10(12),
2LGND11( 7),LGND12( 5),LGND13(6)
DATA LGND1 /1HE,1HM,1HP,1HT,1HY/
DATA LGND4 /1H&,1H ,1HB,1HU,1HT,1H ,1H#,1H ,1HN,1HE,
11HE,1HD,1HE,1HD/
DATA LGND5 /1H&,1H ,1HB,1HU,1HT,1H ,1H#,1H ,1HE,1HX,
11HP,1HE,1HC,1HT,1HE,1HD/
DATA LGND6 /1HE,1HM,1HP,1HT,1HY,1H ,1HB,1HU,1HT,1H ,
11H#,1H ,1HN,1HE,1HE,1HD,1HE,1HD/
DATA LGND7 /1H ,1HA,1HN,1HD,1H /
DATA LGND8 /1HM,1HI,1HS,1HS,1HI,1HN,1HG,1H ,1H#/
DATA LGND10/1HU,1HN,1HE,1HX,1HP,1HE,1HC,1HT,1HE,1HD,
11H ,1H#/
DATA LGND11/1HU,1HN,1HK,1HN,1HO,1HW,1HN/
DATA LGND12/1HE,1HR,1HR,1HO,1HR/
DATA LGND13/1HN,1HU,1HL,1HL,1H ,1H#/
C ARRAYS USED TO KEEP 3 COLUMNS SEPARATED
DIMENSION LOWBFR(3),KIND (3),NEWBGN(3),
1NEWEND(3),BGNNEW(3),ENDNEW(3),IWHERE(3),
2INCBGN(3),LMTBGN(3),BGNINC(3),BGNLMT(3),
FASP, FORTRAN Alphameric Subroutine Package Page 283
DAPAIR, Returns an Index and an Associated Value
3INCEND(3),LMTEND(3),ENDINC(3),ENDLMT(3)
DATA ITTY,JTTY,SLACK,MAXBFR/5,5,0.01,72/
DATA ILEFT,IRIGHT,IAND,IEND,IBLANK/
11H(,1H),1H&,1H;,1H /
C
WRITE(JTTY,1)
1 FORMAT(22H TEST DAIPAR (Y OR N) ,$)
READ(ITTY,2)IANSWR
2 FORMAT(1A1)
IF(IANSWR.NE.1HY)WRITE(JTTY,3)
3 FORMAT(22H DAPAIR WILL BE CALLED)
WRITE(JTTY,4)
4 FORMAT(41H COMMAS BETWEEN INDEX AND VALUE (Y OR N) ,$)
READ(ITTY,2)JANSWR
IBEGIN=-2
IF(JANSWR.EQ.1HY)IBEGIN=1
IF(JANSWR.NE.1HY)WRITE(JTTY,5)
5 FORMAT(28H COMMAS WILL TERMINATE PAIRS)
WRITE(JTTY,6)
6 FORMAT(15H KNDBGN,KNDEND ,$)
READ(ITTY,7)KNDBGN,KNDEND
7 FORMAT(3I)
C
C REQUEST NEXT LINE OF TEXT TO BE EVALUATED
8 DO 9 KNCT=1,3
KIND(KNCT)=0
9 IFDONE(KNCT)=0
WRITE(JTTY,10)
10 FORMAT(2H *,$)
GO TO 13
11 WRITE(JTTY,12)
12 FORMAT(2H &,$)
13 DO 14 KNCT=1,3
LOWBFR(KNCT)=1
14 IF(IFDONE(KNCT).GT.0)IFDONE(KNCT)=0
READ(ITTY,15)IBUFFR
15 FORMAT(72A1)
C
C PROCESS CONTENTS OF BUFFER FOR ALL 3 VALUES OF KONECT
16 INDEX=0
LIMIT=0
DO 47 KNCT=1,3
KONECT=KNCT+IBEGIN
LIMIT=LIMIT+20
IF(IFDONE(KNCT).NE.0)GO TO 46
IF(IANSWR.NE.1HY)
1CALL DAPAIR (SLACK ,KNDBGN,KNDEND,KONECT,IBUFFR,
2MAXBFR ,LOWBFR(KNCT),KIND (KNCT),NEWBGN(KNCT),
3NEWEND(KNCT),BGNNEW(KNCT),ENDNEW(KNCT),IWHERE(KNCT),
4INCBGN(KNCT),LMTBGN(KNCT),BGNINC(KNCT),BGNLMT(KNCT),
5INCEND(KNCT),LMTEND(KNCT),ENDINC(KNCT),ENDLMT(KNCT))
IF(IANSWR.EQ.1HY)
1CALL DAIPAR ( KNDBGN,KNDEND,KONECT,IBUFFR,
2MAXBFR ,LOWBFR(KNCT),KIND (KNCT),NEWBGN(KNCT),
FASP, FORTRAN Alphameric Subroutine Package Page 284
DAPAIR, Returns an Index and an Associated Value
3NEWEND(KNCT),BGNNEW(KNCT),ENDNEW(KNCT))
GO TO(18,20,21,22,24,26,29,29,28,17,39,41,29,28),
1KIND(KNCT)
17 IF(KONECT.EQ.2)GO TO 28
IF(KONECT.EQ.4)GO TO 28
GO TO 37
C
C BUFFER IS EMPTY
18 DO 19 I=1,5
INDEX=INDEX+1
19 JBUFFR(INDEX)=LGND1(I)
GO TO 43
C
C SEMICOLON
20 INDEX=INDEX+1
JBUFFR(INDEX)=IEND
GO TO 45
C
C AMPERSAND INSTEAD OF LEADING NUMBER
21 INDEX=INDEX+1
JBUFFR(INDEX)=IAND
GO TO 44
C
C AMPERSAND WHEN ASSOCIATED NUMBER NEEDED
22 DO 23 I=1,14
INDEX=INDEX+1
23 JBUFFR(INDEX)=LGND4(I)
GO TO 44
C
C AMPERSAND WHEN ASSOCIATED NUMBER EXPECTED
24 DO 25 I=1,16
INDEX=INDEX+1
25 JBUFFR(INDEX)=LGND5(I)
GO TO 44
C
C END OF LINE WHEN ASSOCIATED NUMBER NEEDED
26 DO 27 I=1,18
INDEX=INDEX+1
27 JBUFFR(INDEX)=LGND6(I)
KIND(KNCT)=4
GO TO 44
C
C REPRESENT LEADING NUMBER
28 INDEX=INDEX+1
JBUFFR(INDEX)=ILEFT
29 CALL DASHOW(KNDBGN,0,4,4,4,
13,NEWBGN(KNCT),BGNNEW(KNCT),MAXBFR,INDEX,JBUFFR,IERR)
IF(KIND(KNCT).EQ.9)GO TO 30
IF(KIND(KNCT).EQ.10)GO TO 30
IF(KIND(KNCT).NE.15)GO TO 31
30 INDEX=INDEX+1
JBUFFR(INDEX)=IRIGHT
C
C INSERT ' AND ' BETWEEN LEADING AND TRAILING NUMBERS
FASP, FORTRAN Alphameric Subroutine Package Page 285
DAPAIR, Returns an Index and an Associated Value
31 DO 32 I=1,5
INDEX=INDEX+1
32 JBUFFR(INDEX)=LGND7(I)
IF(KIND(KNCT).EQ.8)GO TO 33
IF(KIND(KNCT).EQ.10)GO TO 33
IF(KIND(KNCT).GE.13)GO TO 35
C
C REPRESENT TRAILING NUMBER
CALL DASHOW(KNDEND,0,4,4,4,
13,NEWEND(KNCT),ENDNEW(KNCT),MAXBFR,INDEX,JBUFFR,IERR)
GO TO 45
C
C REPRESENT MISSING NUMBER
33 DO 34 I=1,9
INDEX=INDEX+1
34 JBUFFR(INDEX)=LGND8(I)
GO TO 45
C
C REPRESENT NULL NUMBER
35 DO 36 I=1,6
INDEX=INDEX+1
36 JBUFFR(INDEX)=LGND13(I)
GO TO 45
C
C NUMBER WHEN COMMA EXPECTED
37 DO 38 I=1,12
INDEX=INDEX+1
38 JBUFFR(INDEX)=LGND10(I)
GO TO 45
C
C UNKNOWN CHARACTER
39 DO 40 I=1,7
INDEX=INDEX+1
40 JBUFFR(INDEX)=LGND11(I)
KIND(KNCT)=-1
GO TO 45
C
C RANGE NOTATION ERROR
41 DO 42 I=1,5
INDEX=INDEX+1
42 JBUFFR(INDEX)=LGND12(I)
KIND(KNCT)=-1
GO TO 45
C
C MARK THAT PROCESSING DONE FOR THIS VALUE OF KONECT
43 IFDONE(KNCT)=-1
GO TO 45
44 IFDONE(KNCT)=1
C
C FILL REST OF COLUMN WITH SPACES
45 MAXPRT=INDEX
46 IF(INDEX.GE.LIMIT)GO TO 47
INDEX=INDEX+1
JBUFFR(INDEX)=IBLANK
FASP, FORTRAN Alphameric Subroutine Package Page 286
DAPAIR, Returns an Index and an Associated Value
GO TO 46
47 CONTINUE
C
C REPORT RESULTS FOR ALL 3 VALUES OF KONECT
WRITE(JTTY,48)(JBUFFR(I),I=1,MAXPRT)
48 FORMAT(2X,100A1)
C
C DETERMINE IF PROCESSING DONE FOR ALL VALUES OF KONECT
DO 49 KNCT=1,3
IF(IFDONE(KNCT).EQ.0)GO TO 16
49 CONTINUE
C
C DETERMINE IF SERIES IS TO CONTINUE ON FOLLOWING LINE
DO 50 KNCT=1,3
IF(IFDONE(KNCT).GT.0)GO TO 11
50 CONTINUE
GO TO 8
END
Typical Dialog Between DAPAIR Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
TEST DAIPAR (Y OR N) N
DAPAIR WILL BE CALLED
COMMAS BETWEEN INDEX AND VALUE (Y OR N) N
COMMAS WILL TERMINATE PAIRS
KNDBGN,KNDEND 0 0
*1/3 4;5/7 8!END OF LINE WHEN VALUE NEEDED
1 AND 4 1 AND 4 1 AND 4
2 AND MISSING # 2 AND MISSING # 2 AND MISSING #
3 AND MISSING # 3 AND MISSING # 3 AND MISSING #
; ; ;
5 AND 8 5 AND 8 5 AND 8
EMPTY BUT # NEEDED EMPTY BUT # NEEDED 6 AND MISSING #
7 AND MISSING #
EMPTY
&9/11 12 13
6 AND 9 6 AND 9
7 AND 10 7 AND 10
(7) AND 11 (7) AND 11
(7) AND 12 UNEXPECTED #
(7) AND 13 12 AND 13
EMPTY EMPTY
*1/3 4!TEST OF FOLLOWING COMMA
1 AND 4 1 AND 4 1 AND 4
EMPTY BUT # NEEDED EMPTY BUT # NEEDED 2 AND MISSING #
3 AND MISSING #
EMPTY
&,5 6
2 AND MISSING # 2 AND MISSING #
3 AND MISSING # 3 AND MISSING #
5 AND 6 5 AND 6
EMPTY EMPTY
FASP, FORTRAN Alphameric Subroutine Package Page 287
DAPATH, Routine to Represent Integer Sequence of Form 1.2.3
DDDDD AAA PPPPPP AAA TTTTTTTT HH HH
DD DD AAAA PP PP AAAA TT HH HH
DD DD AA AA PP PP AA AA TT HH HH
DD DD AA AA PPPPPP AA AA TT HHHHHHHH
DD DD AAAAAAA PP AAAAAAA TT HH HH
DD DD AA AA PP AA AA TT HH HH
DDDDD AA AA PP AA AA TT HH HH
DAPATH, Routine to Represent Integer Sequence of Form 1.2.3
------ ------- -- --------- ------- -------- -- ---- -----
DAPATH represents a sequence of unsigned integers so that
they can be written with a FORTRAN format statement
containing a multiple A1 alphameric specification. Values
within the sequence which are equal to or greater than zero
are represented directly and are separated by periods.
Values less than zero are not represented, but the same
separating periods are inserted into the output text buffer
as if the values were represented. No printing characters
are inserted into the output text buffer if the sequence
consists of a single negative value. For example, the
sequence
100, 200, -1, 400
would insert the following into the output text buffer
1H1,1H0,1H0,1H.,1H2,1H0,1H0,1H.,1H.,1H4,1H0,1H0
which when written with a 12A1 format (where the number to
the left of the A1 can be greater than 12) would produce
100.200..400
The DAPATH Argument List
--- ------ -------- ----
The argument list of routine DAPATH is
SUBROUTINE DAPATH(LOWVLU,KNTVLU,IVALUE,JSTIFY,IFILL ,
1 IWIDTH,LFTCOL,MAXBFR,IBUFFR,KOUNT ,IERR )
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR),IVALUE(KNTVLU)
The following arguments are used for input only, and are
returned unchanged.
LOWVLU = subscript of the IVALUE array location which
contains the first value which is to be
FASP, FORTRAN Alphameric Subroutine Package Page 288
DAPATH, Routine to Represent Integer Sequence of Form 1.2.3
represented. If LOWVLU is greater than KNTVLU, or
else if LOWVLU equals KNTVLU and IVALUE(LOWVLU) is
less than zero, then no printing characters will be
inserted into the IBUFFR array. If IVALUE(LOWVLU)
is less than zero and KNTVLU is greater than
LOWVLU, then the first printing character returned
in the IBUFFR array will be a period.
KNTVLU = subscript of the IVALUE array location which
contains the final value which is to be
represented. If IVALUE(KNTVLU) is less than zero
and KNTVLU is greater than LOWVLU, then the final
printing character returned in the IBUFFR array
will be a period.
IVALUE = array containing in locations IVALUE(LOWVLU)
through and including IVALUE(KNTVLU) the integer
sequence which is to be represented. Values equal
to or greater than zero are represented directly
and are separated by periods. Values less than
zero are not represented, but the same separating
periods appear as if the values were represented.
JSTIFY = -1, left justify the representation of the integer
sequence in the field consisting of
IBUFFR(LFTCOL+1) through IBUFFR(LFTCOL+IWIDTH) (or
through IBUFFR(MAXBFR) if MAXBFR is less than
LFTCOL+IWIDTH).
= 0, center the representation of the integer
sequence in the field consisting of
IBUFFR(LFTCOL+1) through IBUFFR(LFTCOL+IWIDTH) (or
through IBUFFR(MAXBFR) if MAXBFR is less than
LFTCOL+IWIDTH). IBUFFR array locations which are
to the left of the representation of the integer
sequence are filled with spaces.
= 1, right justify the representation of the integer
sequence in the field consisting of
IBUFFR(LFTCOL+1) through IBUFFR(LFTCOL+IWIDTH) (or
through IBUFFR(MAXBFR) if MAXBFR is less than
LFTCOL+IWIDTH). IBUFFR array locations which are
to the left of the representation of the integer
sequence are filled with spaces.
IFILL = 0, do not fill the portion of the field which is to
the right of the representation of the integer
sequence with spaces. The value of IFILL has no
effect on the printing characters in the
representation of the integer sequence. KOUNT will
be returned pointing to the rightmost printing
character in the representation of the integer
sequence.
= 1, fill with spaces the portion of the field which
is to the right of the representation of the
integer sequence and extending through
FASP, FORTRAN Alphameric Subroutine Package Page 289
DAPATH, Routine to Represent Integer Sequence of Form 1.2.3
IBUFFR(LFTCOL+IWIDTH) (or through IBUFFR(MAXBFR) if
MAXBFR is less than LFTCOL+IWIDTH). KOUNT will be
returned set equal to LFTCOL+IWIDTH or MAXBFR,
whichever is the smaller.
IWIDTH = width, stated as the number of columns or of IBUFFR
array locations, of the field in which the integer
sequence is to be represented. The rightmost
IBUFFR array location in the field has the
subscript LFTCOL+IWIDTH or else MAXBFR, whichever
is the smaller.
LFTCOL = subscript of the IBUFFR array location which is to
the immediate left of the leftmost IBUFFR array
location into which either a space or a character
of the representation of the integer sequence can
be placed.
MAXBFR = subscript of the rightmost IBUFFR array location
into which could be placed a space or a character
of the representation of the integer sequence if
IWIDTH is large enough.
The following argument is returned by this routine
containing the representation of the integer sequence.
IBUFFR = array in which the representation of the integer
sequence is returned, 1 character per array
location as though read by a multiple of an A1
format.
The following arguments are used only for output from this
routine. Their input values are ignored.
KOUNT = returned containing the subscript of the rightmost
IBUFFR array location into which a character has
been placed by this routine.
IERR = -1 returned if the representation of the entire
integer sequence would not fit into the field
indicated by LFTCOL and by either IWIDTH or MAXBFR,
whichever indicates the smaller field. If the
representation would not fit and MAXBFR is greater
than or equal to LFTCOL+IWIDTH, then
IBUFFR(LFTCOL+1) through and including
IBUFFR(LFTCOL+IWIDTH) are returned containing
asterisks, and KOUNT is returned set equal to
LFTCOL+IWIDTH. If the representation would not fit
and MAXBFR is less than LFTCOL+IWIDTH, then the
asterisks extend through IBUFFR(MAXBFR) and KOUNT
is returned set equal to MAXBFR.
= 0 returned if the entire integer sequence could be
represented in the field.
FASP, FORTRAN Alphameric Subroutine Package Page 290
DAPATH, Routine to Represent Integer Sequence of Form 1.2.3
DAPATH Demonstration Program, and Dialog Between It and User
------ ------------- ------- --- ------ ------- -- --- ----
DIMENSION IBUFFR(60),IVALUE(20)
DATA ISLASH/1H//
DATA ITTY,JTTY,IWIDTH/5,5,18/
C
C GET NEXT INTEGER SEQUENCE TO BE REPRESENTED
1 WRITE(ITTY,2)
2 FORMAT(2H *,$)
READ(JTTY,3)IBUFFR
3 FORMAT(60A1)
LOWBFR=1
KNTVLU=0
CALL DATEAM(0,0,1,20,60,
1IBUFFR,LOWBFR,KNTVLU,KIND,IVALUE,IVALUE)
C
C REPRESENT THE INTEGER SEQUENCE
DO 4 I=1,60
4 IBUFFR(I)=ISLASH
LFTCOL=1
DO 5 KOLUMN=1,3
JSTIFY=KOLUMN-2
CALL DAPATH(1,KNTVLU,IVALUE,JSTIFY,1,
1IWIDTH,LFTCOL,60,IBUFFR,KOUNT,IERR)
LFTCOL=KOUNT+1
5 CONTINUE
WRITE(ITTY,6)(IBUFFR(I),I=1,LFTCOL)
6 FORMAT(1X,60A1)
GO TO 1
END
The program prompts the user with an asterisk, then reads a
series of up to 20 numbers. The left justified, centered
and right justified representations of the series are then
reported to the user.
*0
/0 / 0 / 0/
*-1
/ / / /
*0 2
/0.2 / 0.2 / 0.2/
*-1 -1
/. / . / ./
*0 2 4
/0.2.4 / 0.2.4 / 0.2.4/
*-1 2 -1
/.2. / .2. / .2./
*0 2 4 8 16 32 64
/0.2.4.8.16.32.64 / 0.2.4.8.16.32.64 / 0.2.4.8.16.32.64/
*-1 2 4 -1 16 32 -1
/.2.4..16.32. / .2.4..16.32. / .2.4..16.32./
FASP, FORTRAN Alphameric Subroutine Package Page 291
DAPICK, Routine to Interpret Array Name and Subscript Ranges
DDDDD AAA PPPPPP IIIIII CCCCC KK KK
DD DD AAAA PP PP II CC KK KK
DD DD AA AA PP PP II CC KK KK
DD DD AA AA PPPPPP II CC KKKKK
DD DD AAAAAAA PP II CC KKK KK
DD DD AA AA PP II CC KK KK
DDDDD AA AA PP IIIIII CCCCC KK KK
DAPICK, Routine to Interpret Array Name and Subscript Ranges
------ ------- -- --------- ----- ---- --- --------- ------
DAPICK is used along with several other routines in the FASP
package for the purpose of specifying by name, examining and
modifying the values of multiply subscripted arrays
equivalenced with or otherwise loaded into a singly
subscripted buffer. The interaction between these routines
is described at the start of the DALOAD documentation. The
sample program at the end of the DALOAD documentation
illustrates the use of most of these routines.
DAPICK identifies the array name and subscript ranges typed
by the user and read by the calling program with a multiple
of an A1 format. Spaces and/or tab characters can appear
anywhere in the text interpreted by DAPICK, but are ignored.
The printing characters in the buffer array are matched
against the words in the dictionary constructed by the
DALOAD routine. The array name typed by the user must match
a name in the dictionary exactly. Abbreviations are not
recognized. The match must include all printing characters
typed by the user through the end of the buffer or up to but
not including a left parenthesis, equal sign, semicolon,
exclamation point or ampersand. Any other unmatched
printing characters will cause the comparison to fail.
Numbers indicating subscript ranges follow the array name
and are enclosed between left and right parentheses. The
subscript range specifications are separated by commas and
consist of the starting value, a slash character or a colon
(the 2 characters are equivalent), and the ending value.
The starting value of a range can be less than, equal to, or
greater than the ending value. The numbers indicating the
subscript ranges can be signed, but cannot contain an
exponent, neither the E of scientific notation nor K nor M
being accepted. If a slash or a colon is present but either
number is missing, then the missing number is assumed to be
the corresponding limit for the subscript as specified by
the dictionary. If a single number appears without a slash
and without a colon, then the range is assumed to consist
only of the indicated number. If nothing or just a slash or
just a colon appears between the parentheses and/or commas,
then the range is assumed to extend across all possible
values of the subscript as specified by the dictionary. For
FASP, FORTRAN Alphameric Subroutine Package Page 292
DAPICK, Routine to Interpret Array Name and Subscript Ranges
example, if the input buffer contains the text
R2A(2/,,3,12/14)
and if the dictionary indicates that R2A is dimensioned
(3,4,3,18), then DAPICK would specify that
a. the first subscript should be varied from 2 through the
upper limit of 3
b. the second subscript should be varied from the lower
limit of 1 through the upper limit of 4
c. the third subscript should have the single value 3
d. the fourth subscript should be varied from 12 through
14.
After being evaluated, the subscript ranges are compared
with the limits stored in the dictionary. If the dictionary
indicates that the name is not subscripted or that it can
take only the single subscript 1, then in either of these
cases the complete lack of any parenthetical subscript
notation, or else the single subscript 1 typed by the user,
will be accepted.
If an equal sign is found to the right of the array name or
to the right of the array name and of its subscripts, then
the location of the first character to the right of the
equal sign is also identified to the calling program.
Semicolons preceding the array name are ignored, except that
at least one semicolon can be required to precede an
additional array name in a line of text in which a previous
call to this routine has already found and processed one or
more other array names.
If an exclamation point or an ampersand is found anywhere in
the text typed by the user, then the exclamation or
ampersand and all characters to its right are taken to be a
comment and are otherwise ignored.
FASP, FORTRAN Alphameric Subroutine Package Page 293
DAPICK, Routine to Interpret Array Name and Subscript Ranges
The DAPICK Argument List
--- ------ -------- ----
The argument list of routine DAPICK IS
SUBROUTINE DAPICK(MAXBFR,IBUFFR,LTRLOW,LTRUSD,LTRSTR,
1 NUMLOW,NUMUSD,NUMSTR,MAXSUB,LOWBFR,KIND ,LRGLTR,
2 LRGNUM,LRGKNT,INITAL,KOUNT ,LTRINI,NUMINI,KNTSUB,
3 INISUB,LMTSUB)
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR),LTRSTR(LTRUSD),
1NUMSTR(NUMUSD),INISUB(MAXSUB),LMTSUB(MAXSUB)
The following are input arguments left unchanged.
MAXBFR = maximum subscript of the locations in the IBUFFR
array containing characters read by the the calling
program and which are to be interpreted by this
routine.
IBUFFR = input buffer array containing characters typed by
the user and read by a multiple of an A1 format,
which are to be searched for array names and
subscript values. IBUFFR then contains 1 character
per computer storage location.
LTRLOW = lowest subscript of the locations in the LTRSTR
array containing the characters of the names in the
dictionary as originally read by DALOAD using a
multiple of an A1 format. LTRSTR(LTRLOW) contains
either the first letter of the name of the first
logical group of names in the dictionary or else
(if the first group itself isn't named) the first
letter of the first name within the first logical
group in the dictionary.
LTRUSD = highest subscript of the locations in the LTRSTR
array containing the characters of the names in the
dictionary as originally read by DALOAD using a
multiple of an A1 format. LTRSTR(LTRUSD) contains
the last character of the last name in the
dictionary.
LTRSTR = array containing the characters forming the names
in the dictionary, 1 character per array location
as originally read by DALOAD using a multiple of an
A1 format.
NUMLOW = lowest subscript of the locations in the NUMSTR
array containing the numeric information
corresponding to the names stored in the LTRSTR
array. NUMSTR(NUMLOW) must contain the start of
FASP, FORTRAN Alphameric Subroutine Package Page 294
DAPICK, Routine to Interpret Array Name and Subscript Ranges
the description of a logical group of names, not
the start of the description of an individual name.
NUMUSD = highest subscript of the locations in the NUMSTR
array containing the numeric information
corresponding to the names stored in the LTRSTR
array.
NUMSTR = array containing the numeric information
corresponding to the names stored in the LTRSTR
array. The construction of the NUMSTR array is
described in detail in the DALOAD documentation.
For each name in the dictionary, the NUMSTR array
contains
a. the number of characters in the name
b. an indication of the associated data type
c. the number of subscript ranges
d. pairs of starting and ending values of these
ranges.
If the number of characters is instead zero or
negative, then its absolute value is the number of
characters in the name of a logical group of names,
and the next location, rather than indicating the
data type, contains the number of locations within
a singly subscripted buffer which would be needed
to store the values of the multiply subscripted
arrays which are within the logical group and
equivalenced with or otherwise loaded into such a
singly subscripted buffer.
MAXSUB = highest subscript of the locations in the INISUB
and LMTSUB arrays which can be used by this routine
to store the subscript ranges typed by the user.
The start of the first range is stored in INISUB(1)
and the end of the first range in LMTSUB(1). If
the array names stored in the dictionary are
actually the names of FORTRAN language arrays, then
the FORTRAN limit of 7 subscripts would apply,
suggesting that MAXSUB should be 7 and that both
INISUB and LMTSUB should be dimensioned to at least
7. If more than MAXSUB subscript ranges are typed
by the user, then the excess ranges to the right of
those accepted are ignored (that is, KNTSUB is not
incremented, the values of the bounds are not
stored in the INISUB and LMTSUB arrays and the
appearance of the excess ranges is not taken to be
an error), except for such interpretation as is
necessary to determine the location within the
buffer of the closing parenthesis.
FASP, FORTRAN Alphameric Subroutine Package Page 295
DAPICK, Routine to Interpret Array Name and Subscript Ranges
The following arguments are used both for input to the
DAPICK routine, and for output to the calling program.
LOWBFR = input containing the subscript within the IBUFFR
array of the first (leftmost) character which can
be scanned for an array name. If the characters in
the buffer do not match a name in the dictionary,
then LOWBFR will be returned pointing to the
leftmost printing character which is not a
semicolon nor an exclamation point nor an ampersand
and which is not to the right of an exclamation
point and which is not to the right of an
ampersand, or else will be returned containing
MAXBFR+1 if no such printing characters are found
in the buffer. If the text in the buffer does
match a word in the dictionary, then LOWBFR will be
returned pointing to the first character to the
right of the name, or to the first character to the
right of the subscript range if present, or to the
first character to the right of the equal sign
either following the subscript range or following
the name directly. If a match is found, then
LOWBFR is returned containing the location of the
first character of the text in the IBUFFR array
which should be interpreted by the DAVARY routine
to obtain the numeric values, if any, typed by the
user to the right of the equal sign.
KIND = input to this routine describing the previous
processing, if any, of the line of text typed by
the user and contained in the input buffer array
IBUFFR. KIND is returned to the calling program
describing the type of information found in the
buffer.
KIND is input to this routine with its value
specifying the following information.
= -1 input, none of the characters currently in the
IBUFFR array have been processed by previous calls
to DAPICK or DAVARY. An initial semicolon is not
required since the current call to DAPICK will
evaluate the first character typed by the user.
KIND will be returned containing the value 3 if the
input buffer is empty or merely contains semicolons
and/or a comment indicated by a leading exclamation
point.
= 0 input, a semicolon was typed by the user when the
user was asked to define the new value of the
previously selected array. DAVARY returned control
to DAPICK to continue processing of the contents of
the IBUFFR array after having first advanced LOWBFR
beyond the location of the semicolon. If the
remainder of the buffer is empty or contains merely
FASP, FORTRAN Alphameric Subroutine Package Page 296
DAPICK, Routine to Interpret Array Name and Subscript Ranges
semicolons and/or a comment indicated by a leading
exclamation point or by a leading ampersand, then
KIND will be returned with the value 4.
= 1 (or greater) input, the subscripts of the
previously specified simulated array were varied
through the entire range selected by the user.
DAROLL (or DALOOP) has then returned control back
to DAPICK to continue processing of the remaining
characters in the input buffer. If the remainder
of the buffer is empty or contains merely
semicolons and/or a comment indicated by a leading
exclamation point or by a leading ampersand, then
KIND will be returned with the value 4. If an new
array specification appears in the buffer, it must
follow (appear to the right of) at least one
semicolon. KIND is returned containing the value 8
if the first printing character at or to the right
of LOWBFR is not a semicolon nor an exclamation
point nor an ampersand.
KIND is returned by this routine with its value
specifying the following information.
= 1 returned, the text typed by the user matched a
name in the dictionary, but no equal sign followed
the name or subscript ranges in the IBUFFR array.
LOWBFR is returned pointing to the first character
to the right of the right parenthesis or if
subscripts were not indicated to the first
character (which would have to be a space, a tab, a
semicolon, an exclamation point, an ampersand or
else be beyond the end of the buffer) to the right
of the name.
= 2 returned, the text typed by the user matched a
name in the dictionary, and an equal sign follows
the name or the subscript ranges. LOWBFR is
returned pointing to the first character to the
right of the equal sign.
= 3 returned, the input buffer contained no printing
characters other than semicolons and/or the
characters of a comment indicated by a leading
exclamation point, and KIND was input containing -1
indicating that this routine has not previously
been called to process any of the characters
currently in the buffer. LOWBFR is returned
pointing to the next character beyond the end of
the buffer.
= 4 returned, same as when KIND is returned
containing 3, except that KIND was input greater
than or equal to zero. The characters, if any,
evaluated by this call to DAPICK were those
remaining unprocessed after the previous return
from DAVARY, and were not within an unprocessed
buffer just read by the calling program. KIND is
FASP, FORTRAN Alphameric Subroutine Package Page 297
DAPICK, Routine to Interpret Array Name and Subscript Ranges
also returned containing 4, regardless of the input
value of KIND, if the input buffer contained the
characters of a comment indicated by a leading
ampersand, but contained no printing characters
other than possible semicolons to the left of the
ampersand.
Returned values of 3 and 4 for KIND can be treated
by the calling program as identical. If these
values are treated as different, the suggested
interpretation is that KIND=3 indicates that the
user did not type anything on the line, while
KIND=4 indicates that the contents of a non-empty
line have been processed so the user should be
asked to type a new line. Comments indicated by
exclamation points and by ampersands are handled
differently by this routine only when they are the
leftmost printing characters (except perhaps for
semicolons) in the entire line typed by the user.
In this case, a comment indicated by a leading
ampersand returns KIND=4 since by typing the
ampersand the user has indicated that he wishes to
type another line.
= 5 returned, the text typed by the user matched a
name in the dictionary, but was followed by an
incorrect number of subscript ranges.
= 6 returned, the text typed by the user matched a
name in the dictionary and was followed by the
correct number of subscripts ranges, but one or
more of the ranges extended outside of the limits
indicated for these ranges by the dictionary.
= 7 returned, the input buffer contained printing
characters (other than semicolons and/or the
characters of a comment indicated by a leading
exclamation point or by a leading ampersand), but
no match with a name in the dictionary was
obtained. LOWBFR is returned pointing to the first
(leftmost) printing character which is not a
semicolon.
= 8 returned, KIND was input greater than zero, but
the first printing character in the buffer was not
a semicolon nor an exclamation point nor an
ampersand. The initial semicolon is required to
insure that the user really meant that the text
remaining in the input buffer after the completion
of the processing of the previously selected
simulated array is to be acted upon by DAPICK.
The following arguments are used only for output. Their
input values are ignored.
LRGLTR = returned containing the subscript of the LTRSTR
array location which contains the first letter of
the name associated with the logical group of names
FASP, FORTRAN Alphameric Subroutine Package Page 298
DAPICK, Routine to Interpret Array Name and Subscript Ranges
in the dictionary which includes the name which
matched that typed by the user.
LRGNUM = returned containing the subscript of the NUMSTR
array location which contains the first of the
numeric information associated with the logical
group of names in the dictionary which includes the
name which matched that typed by the user.
NUMSTR(LRGNUM) contains as its absolute value the
number of characters starting at LTRSTR(LRGLTR)
which are contained in the name, if any, of the
logical group of names. NUMSTR(LRGNUM+1) contains
the number of locations within a singly subscripted
buffer which would be needed to store the values of
the multiply subscripted arrays which are within
the logical group and equivalenced with or
otherwise loaded into such a singly subscripted
buffer.
LRGKNT = sequence number within the dictionary of the
logical group of names containing the name matching
that typed by the user. If the name typed by the
user is contained in the third logical group in the
dictionary, then LRGKNT would be returned
containing the value 3.
INITAL = returned containing the location of the start of
the array specified by the user relative to the
singly subscripted buffer containing the arrays in
the logical group and with which the specified
array is equivalenced. If the logical groups
represent the contents of individual records in a
file, and if arrays ABC(10), DEF(10), GHI(10) etc.
are contained in a particular record, and if the
user typed the array name GHI (regardless of what
subscripts within the range 1 through 10 he
happened to type), then INITAL would be returned
with the value 10+10+1 or 21.
KOUNT = sequence number of the name selected by the user
within the logical group of names containing it in
the dictionary. This number does not include the
name, if any, of the logical group itself. If the
user types the name of the third array described in
the logical group, then KOUNT is returned with the
value 3.
LTRINI = returned containing the subscript of the location
within the LTRSTR array which contains the first
letter of the name matching that typed by the user.
NUMINI = returned containing the subscript of the location
within the NUMSTR array which contains the start of
the numeric information describing the name
FASP, FORTRAN Alphameric Subroutine Package Page 299
DAPICK, Routine to Interpret Array Name and Subscript Ranges
matching that typed by the user. NUMSTR(NUMINI)
contains the number of characters in the name,
NUMSTR(NUMINI+1) the indication of the associated
data type (-1 for octal integer, 0 for decimal
integer, and +1 for real), and NUMSTR(NUMINI+2) the
number of subscript ranges allowed (and required).
KNTSUB = returned containing the number of subscripts ranges
found.
INISUB = array returned containing the starting values of
the subscript ranges typed by the user. The start
of the first subscript range typed by the user is
returned in INISUB(1) and of the final in
INISUB(KNTSUB).
LMTSUB = array returned containing the ending values of the
subscript ranges typed by the user. The end of the
first subscript range typed by the user is returned
in LMTSUB(1) and of the final in LMTSUB(KNTSUB).
An Example of the Use of DAPICK
-- ------- -- --- --- -- ------
The sample program listed on the following pages reads a
line of text from the user's terminal, then compares the
contents of the line with a dictionary which is defined by
DATA statements at the start of the program, but which could
have been constructed instead by having DALOAD read the
COMMON statements listed as comment cards at the start of
the program. The program summarizes the evaluation before
asking for the next array name specification.
A sample dialog between the program and user is presented
following the listing of the program.
To demonstrate that neither the input text nor the
dictionary need to start at the beginning of the respective
arrays, all of these arrays contain nonsense values ahead of
the valid information.
FASP, FORTRAN Alphameric Subroutine Package Page 300
DAPICK, Routine to Interpret Array Name and Subscript Ranges
C PROGRAM TO DEMONSTRATE DAPICK ROUTINE
C
C FOLLOWING DATA STATEMENTS SIMULATE DICTIONARY WHICH
C DALOAD WOULD CONSTRUCT IF IT READ FOLLOWING AS DATA
C
C COMMON/FIRST/A(10,10),BC(10,10),DEF(10,10)
C COMMON/SECOND/GHI(10),JKL,MNO(10),PQR(1)
C COMMON/THIRD/STU(10,10),VW(10,10/1),X(10/1,10/1)
C
DIMENSION NUMSTR(72),LTRSTR(46),IBUFFR(60),JBUFFR(6),
1KBUFFR(6),INISUB(7),LMTSUB(7)
C
DATA NUMSTR/
1100,100,100,
2 -5,300, 0,
3 1, 1, 2, 1, 10, 1, 10,
4 2, 1, 2, 1, 10, 1, 10,
5 3, 1, 2, 1, 10, 1, 10,
6 -6, 22, 0,
7 3, 1, 1, 1, 10,
8 3, 0, 0,
9 3, 0, 1, 1, 10,
1 3, 1, 1, 1, 1,
2 -5,300, 0,
3 3, 1, 2, 1, 10, 1, 10,
4 2, 1, 2, 1, 10, 10, 1,
5 1, 1, 2, 10, 1, 10, 1/
C
DATA LTRSTR/
11H*,1H*,1H*,1H*,1H*,1H*,
21HF,1HI,1HR,1HS,1HT,
31HA,1HB,1HC,1HD,1HE,1HF,
41HS,1HE,1HC,1HO,1HN,1HD,
51HG,1HH,1HI,1HJ,1HK,1HL,1HM,1HN,1HO,1HP,1HQ,1HR,
61HT,1HH,1HI,1HR,1HD,
71HS,1HT,1HU,1HV,1HW,1HX/
C
DATA NUMLOW,NUMUSD,LTRLOW,LTRUSD,MAXBFR,MAXSUB,ITTY/
14,72,7,46,60,7,5/
DATA ISPACE/1H /
C
C READ USER TYPED ARRAY SPECIFICATION
1 WRITE(ITTY,2)
2 FORMAT(' **********'/' ARRAY NAME:'$)
LOWBFR=11
KIND=-1
READ(ITTY,3)(IBUFFR(I),I=LOWBFR,MAXBFR)
3 FORMAT(60A1)
C
C EVALUATE USER SUPPLIED TEXT
4 LAST=KIND
CALL DAPICK (MAXBFR,IBUFFR,LTRLOW,LTRUSD,LTRSTR,
1 NUMLOW,NUMUSD,NUMSTR,MAXSUB,LOWBFR,KIND ,LRGLTR,
2 LRGNUM,LRGKNT,INITAL,KOUNT ,LTRINI,NUMINI,KNTSUB,
FASP, FORTRAN Alphameric Subroutine Package Page 301
DAPICK, Routine to Interpret Array Name and Subscript Ranges
3 INISUB,LMTSUB)
C
C DETERMINE UNPROCESSED TEXT IF ANY
LIMIT=MAXBFR
GO TO 6
5 LIMIT=LIMIT-1
6 IF(LIMIT.LT.LOWBFR)GO TO 7
IF(IBUFFR(LIMIT).EQ.ISPACE)GO TO 5
C
C REPORT TYPE OF INFORMATION FOUND
7 IF((LAST.GT.0).AND.(KIND.NE.4))WRITE(ITTY,8)
8 FORMAT(' ')
GO TO(21,9,11,1,13,15,17,19),KIND
9 WRITE(ITTY,10)
10 FORMAT(' EQUALS SIGN FOLLOWS NAME')
GO TO 21
11 WRITE(ITTY,12)
12 FORMAT(' EMPTY')
GO TO 1
13 WRITE(ITTY,14)
14 FORMAT(' WRONG NUMBER OF SUBSCRIPTS')
GO TO 21
15 WRITE(ITTY,16)
16 FORMAT(' SUBSCRIPTS OUT OF RANGE')
GO TO 21
17 WRITE(ITTY,18)(IBUFFR(I),I=LOWBFR,LIMIT)
18 FORMAT(' NO MATCH FOUND WITH ',60A1)
GO TO 1
19 WRITE(ITTY,20)(IBUFFR(I),I=LOWBFR,LIMIT)
20 FORMAT(' SEMICOLON REQUIRED BEFORE ',60A1)
GO TO 1
C
C REPORT ARRAY NAME, SUBSCRIPTS AND LOCATION IN BUFFER
21 IF(KNTSUB.GT.0)WRITE(ITTY,22)(INISUB(I),LMTSUB(I),
1I=1,KNTSUB)
22 FORMAT(' SUBSCRIPTS FOUND',7(I3,1H/,I3,1H,))
J=LTRINI+NUMSTR(NUMINI)-1
K=LRGLTR-NUMSTR(LRGNUM)-1
DO 23 I=1,6
JBUFFR(I)=ISPACE
IF(LTRINI.LE.J)JBUFFR(I)=LTRSTR(LTRINI)
KBUFFR(I)=ISPACE
IF(LRGLTR.LE.K)KBUFFR(I)=LTRSTR(LRGLTR)
LRGLTR=LRGLTR+1
23 LTRINI=LTRINI+1
WRITE(ITTY,24)LRGKNT,KBUFFR,INITAL,KOUNT,JBUFFR
24 FORMAT(' GROUP',I2,' NAMED ',6A1,' LOCATION',I4,
1' STARTS ARRAY',I2,' NAMED ',6A1)
GO TO 4
END
FASP, FORTRAN Alphameric Subroutine Package Page 302
DAPICK, Routine to Interpret Array Name and Subscript Ranges
Typical Dialog Between DAPICK Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
**********
ARRAY NAME:DEF(4,);DEF(,7);DEF(4,7)!A COMMENT
SUBSCRIPTS FOUND 4/ 4, 1/ 10,
GROUP 1 NAMED FIRST LOCATION 201 STARTS ARRAY 3 NAMED DEF
SUBSCRIPTS FOUND 1/ 10, 7/ 7,
GROUP 1 NAMED FIRST LOCATION 201 STARTS ARRAY 3 NAMED DEF
SUBSCRIPTS FOUND 4/ 4, 7/ 7,
GROUP 1 NAMED FIRST LOCATION 201 STARTS ARRAY 3 NAMED DEF
**********
ARRAY NAME:DEF(1/2,3/4,5/6,7/8);DEF(4,12)DEF
WRONG NUMBER OF SUBSCRIPTS
SUBSCRIPTS FOUND 1/ 2, 3/ 4, 5/ 6, 7/ 8,
GROUP 1 NAMED FIRST LOCATION 201 STARTS ARRAY 3 NAMED DEF
SUBSCRIPTS OUT OF RANGE
SUBSCRIPTS FOUND 4/ 4, 12/ 12,
GROUP 1 NAMED FIRST LOCATION 201 STARTS ARRAY 3 NAMED DEF
SEMICOLON REQUIRED BEFORE DEF
**********
ARRAY NAME:X(,);X(3/,);X(,/7);X(3/,/7);X(6/9,2/4);X(9/6,4/2)
SUBSCRIPTS FOUND 10/ 1, 10/ 1,
GROUP 3 NAMED THIRD LOCATION 201 STARTS ARRAY 3 NAMED X
SUBSCRIPTS FOUND 3/ 1, 10/ 1,
GROUP 3 NAMED THIRD LOCATION 201 STARTS ARRAY 3 NAMED X
SUBSCRIPTS FOUND 10/ 1, 10/ 7,
GROUP 3 NAMED THIRD LOCATION 201 STARTS ARRAY 3 NAMED X
SUBSCRIPTS FOUND 3/ 1, 10/ 7,
GROUP 3 NAMED THIRD LOCATION 201 STARTS ARRAY 3 NAMED X
SUBSCRIPTS FOUND 6/ 9, 2/ 4,
GROUP 3 NAMED THIRD LOCATION 201 STARTS ARRAY 3 NAMED X
SUBSCRIPTS FOUND 9/ 6, 4/ 2,
GROUP 3 NAMED THIRD LOCATION 201 STARTS ARRAY 3 NAMED X
**********
ARRAY NAME:UNKNOWN
NO MATCH FOUND WITH UNKNOWN
**********
ARRAY NAME: ;;;!SEMICOLONS AND A COMMENT
EMPTY
FASP, FORTRAN Alphameric Subroutine Package Page 303
DAPICK, Routine to Interpret Array Name and Subscript Ranges
**********
ARRAY NAME:STU;STU();STU(,);STU(,)=
WRONG NUMBER OF SUBSCRIPTS
GROUP 3 NAMED THIRD LOCATION 1 STARTS ARRAY 1 NAMED STU
WRONG NUMBER OF SUBSCRIPTS
SUBSCRIPTS FOUND 1/ 10,
GROUP 3 NAMED THIRD LOCATION 1 STARTS ARRAY 1 NAMED STU
SUBSCRIPTS FOUND 1/ 10, 1/ 10,
GROUP 3 NAMED THIRD LOCATION 1 STARTS ARRAY 1 NAMED STU
EQUALS SIGN FOLLOWS NAME
SUBSCRIPTS FOUND 1/ 10, 1/ 10,
GROUP 3 NAMED THIRD LOCATION 1 STARTS ARRAY 1 NAMED STU
**********
ARRAY NAME:JKL;JKL=;JKL(1);JKL(1)=!A COMMENT
SUBSCRIPTS FOUND 1/ 1,
GROUP 2 NAMED SECOND LOCATION 11 STARTS ARRAY 2 NAMED JKL
EQUALS SIGN FOLLOWS NAME
SUBSCRIPTS FOUND 1/ 1,
GROUP 2 NAMED SECOND LOCATION 11 STARTS ARRAY 2 NAMED JKL
SUBSCRIPTS FOUND 1/ 1,
GROUP 2 NAMED SECOND LOCATION 11 STARTS ARRAY 2 NAMED JKL
EQUALS SIGN FOLLOWS NAME
SUBSCRIPTS FOUND 1/ 1,
GROUP 2 NAMED SECOND LOCATION 11 STARTS ARRAY 2 NAMED JKL
**********
ARRAY NAME:PQR;PQR=;PQR(1);PQR(1)=PQR
SUBSCRIPTS FOUND 1/ 1,
GROUP 2 NAMED SECOND LOCATION 22 STARTS ARRAY 4 NAMED PQR
EQUALS SIGN FOLLOWS NAME
SUBSCRIPTS FOUND 1/ 1,
GROUP 2 NAMED SECOND LOCATION 22 STARTS ARRAY 4 NAMED PQR
SUBSCRIPTS FOUND 1/ 1,
GROUP 2 NAMED SECOND LOCATION 22 STARTS ARRAY 4 NAMED PQR
EQUALS SIGN FOLLOWS NAME
SUBSCRIPTS FOUND 1/ 1,
GROUP 2 NAMED SECOND LOCATION 22 STARTS ARRAY 4 NAMED PQR
SEMICOLON REQUIRED BEFORE PQR
FASP, FORTRAN Alphameric Subroutine Package Page 304
DAPLAT, Scatter Plot Routine for Printer
DDDDD AAA PPPPPP LL AAA TTTTTTTT
DD DD AAAA PP PP LL AAAA TT
DD DD AA AA PP PP LL AA AA TT
DD DD AA AA PPPPPP LL AA AA TT
DD DD AAAAAAA PP LL AAAAAAA TT
DD DD AA AA PP LL AA AA TT
DDDDD AA AA PP LLLLLLLL AA AA TT
DAPLAT, Scatter Plot Routine for Printer
------- ------- ---- ------- --- -------
DAPLAT is a FORTRAN subroutine which constructs printable
plots having proportions which are selected by the calling
program. Each plot can contain one or more curves and each
curve can be represented by its own alphabetic letter.
Points on a curve can optionally be connected either by
asterisks or by some alphabetic letter which need not be the
same as that used to represent the points themselves. If
plotting points not connected by lines, a pin map can be
simulated by using the closest empty location if the
location representing the point is already in use.
The plot can be ruled with grid lines at evenly spaced
intervals across the width and height of the plot. Numbers
printed alongside the plot are represented with 5
significant digits, but with suppression of rightmost zeroes
beyond the decimal point. If the number cannot be printed
in floating point form, the number is represented in
scientific notation instead.
If two or more different alphabetic letters are to appear in
the same location in the printed plot, an ampersand will
appear instead. However, if the line segments connecting
points are to be formed from a different letter than the
letter representing the points themselves, then both of
these letters, where used for previous curves or where used
elsewhere on the current curve, are not changed to
ampersands by coincidence with the present curve. Points
and lines which are formed from asterisks can be selected
either to be superseded by (be dominated by) any other
characters, or else to supersede (be dominant over) any
other characters. The character set and dominance rules can
be changed by the calling program if necessary.
When line segments are formed from different letters than
are used to represent points, and when these letters vary
from one curve to the next, then the order in which
overlapping curves are specified can change which locations
of the plot contain ampersands. For example, if we
superimpose the horizontal and vertical line patterns shown
below (assuming that the asterisks are dominated by all
other characters)
FASP, FORTRAN Alphameric Subroutine Package Page 305
DAPLAT, Scatter Plot Routine for Printer
horizontal line pattern vertical line pattern
10 -EDDDDDDDDDDDDDDDDDDDE 10 -B---C---D-+-E---*---A
! ! A B C D * *
DCCCCCCCCCCCCCCCCCCCD A B C D * *
! ! A B C D * *
CBBBBBBBBBBBBBBBBBBBC A B C D * *
5 -+ + + 5 -A B C + D * *
BAAAAAAAAAAAAAAAAAAAB A B C D * *
! ! A B C D * *
A*******************A A B C D * *
! ! A B C D * *
0 -********************* 0 -B---C---D-+-E---*---A
! ! ! ! ! !
0 5 10 0 5 10
then we get the 2 plots shown below depending upon whether
the horizontal or vertical lines are specified first.
horizontal lines first vertical lines first
10 -&DDD&DDDDDDDEDDDDDDD& 10 -&DDD&DDDDDDDEDDDDDDD&
A B C D * * A B C D * *
&CCCCCCCCCCC&CCCCCCCD &CCC&CCCCCCCDCCCCCCCD
A B C D * * A B C D * *
&BBBBBBB&BBB&BBBBBBBC &BBBBBBBCBBB&BBBBBBBC
5 -A B C + D * * 5 -A B C + D * *
BAAA&AAA&AAA&AAAAAAAB BAAABAAA&AAA&AAAAAAAB
A B C D * * A B C D * *
A***B***C***D*******A A***B***C***D*******A
A B C D * * A B C D * *
0 -B***C***D***E*******A 0 -B***C***D***E*******A
! ! ! ! ! !
0 5 10 0 5 10
The important differences between the above 2 plots are at
X=2, Y=8 and at X=4, Y=6. The choice of whether or not to
place an ampersand at a particular location along the
representation of a curve must be made by comparing the 2
letters being used for the curve with the letter if any
already occupying that location. In the left plot above,
the horizontal line segment formed of B's was specified
before the vertical line of C's. When the point of
intersection was reached, the B already on the plot was not
found to match either the D's being used for the current
points or the C's being used for the connecting line
segments so the B was changed to an ampersand. In the right
plot above, the vertical line of C's was specified before
the horizontal line of B's. When the intersection was
reached, the C already on the plot was found to match one of
the letters B and C being used to represent the current
curve. Since a line segment rather than a point was being
constructed at the intersection, the matching character
already on the plot was left unchanged. If a point was
being constructed at the intersection instead of a line
segment, then the character on the plot would have been
FASP, FORTRAN Alphameric Subroutine Package Page 306
DAPLAT, Scatter Plot Routine for Printer
changed to the character being used to represent the point.
The DAPLAT Argument List
--- ------ -------- ----
The argument list of routine DAPLAT is
SUBROUTINE DAPLAT( IPLOT,MAXWID,MAXHIH,LETTER,KONECT,
1 XPOINT,YPOINT,MINSUB,MAXSUB, XLEFT,YLOWER,XRIGHT,
2 YUPPER, IGRID, IEDGE,MARGIN,MSHWID,MSHHIH,LTROFF,
3 LINOFF,LINPRT, IDISK,IRESET,LTRERR,LINERR)
with the associated DIMENSION statement
DIMENSION XPOINT(MAXSUB),YPOINT(MAXSUB)
Arguments Used to Input Values to DAPLAT
--------- ---- -- ----- ------ -- ------
The following input arguments are used by each call to
DAPLAT whether or not a plot is generated.
IPLOT = -1, add the points and/or lines represented by the
coordinates in the XPOINT and YPOINT arrays to the
plot storage, then return to the calling program
without generating the plot. Additional DAPLAT
calls will be executed before the plot is to be
printed.
= 0, add the points and/or lines in the XPOINT and
YPOINT arrays to the plot storage, then generate
the plot and clear the plot storage after the plot
is generated. Additional calls to DAPLAT, if any,
will store upon a blank plot surface.
= 1, add the points and/or lines in the XPOINT and
YPOINT arrays to the plot storage, then generate
the plot but do not clear the plot storage after
the plot is generated. The currently stored plot
will be added to by additional calls to DAPLAT.
MAXWID = width of the plot stated as the number of columns
of characters forming the plotting area upon which
data can be plotted. MAXWID is normally 1 plus a
multiple of MSHWID. The recommended maximum value
of MAXWID is 101. It should be noted that the plot
usually has an additional margin of 12 characters
along the left side in which scale numbers and the
carriage control character are printed, and an
additional 4 characters can appear to the right of
the plot in the lower-right scale number. If the
output is being written to a terminal so that the
character in column 1 is interpreted as a carriage
control, and if wraparound at the right margin of
FASP, FORTRAN Alphameric Subroutine Package Page 307
DAPLAT, Scatter Plot Routine for Printer
the portion of the rightmost scale number extending
beyond the right edge of the plot is acceptable,
then MAXWID can be set to 61 on terminals which can
display 72 characters per line, and can be set to
121 on terminals which can display 132 characters
per line. The largest accepted value of MAXWID is
131, which can lead to a lower scale 147 characters
wide which is too large to output on most
line-printers. If the value of IEDGE turns off
both left scale numbers and lower scale numbers
(this option is described after the sample program,
not in the ordinary argument descriptions) then
only the carriage control character, either a space
or an asterisk depending upon the value of LINPRT,
will appear to the left of the plot (providing that
the value of MARGIN does not also request
additional spaces) so the maximum width of a
printed line would then be merely MAXWID+MARGIN+1.
= 0, use the last nonzero value specified for MAXWID
as the width of the plot, or use the value 101 if a
nonzero value of MAXWID has not been specified.
MAXHIH = height of the plot stated as the number of lines
forming the plotting area upon which data can be
plotted. Normally MAXHIH would be 1 plus a
multiple of MSHHIH. The plot has an additional
lower margin of 2 lines (or 3 lines if the numbers
below the lower scale are in scientific notation).
= 0, use the last nonzero value specified for MAXHIH
as the height of the plot, or use the value 51 if a
nonzero value of MAXHIH has not been specified.
LETTER = 0, plot the points having coordinates in the XPOINT
and YPOINT arrays with nondominant asterisks. If
one (or more) of these points is to appear at the
same location on the plot as some other point or
portion of a line segment which is represented by
an alphabetic character or by an ampersand, then
the other point or portion of a line segment is
shown instead of the current point.
= within the range 1 through 26, LETTER is the serial
number within the alphabet of the letter used to
plot the points having coordinates in the XPOINT
and YPOINT arrays. LETTER=3 would cause the
plotted points to be represented with C's.
= 27 (or greater), plot the points having coordinates
in the XPOINT and YPOINT arrays with dominant
asterisks. If one (or more) of these points is to
appear at the same location on the plot as some
other point or portion of a line segment which is
represented by an alphabetic character or by an
ampersand, then the current point is shown instead
of the other point or portion of a line segment.
FASP, FORTRAN Alphameric Subroutine Package Page 308
DAPLAT, Scatter Plot Routine for Printer
KONECT = less than -1, simulate a pin map. Store the points
having coordinates in the XPOINT and YPOINT arrays
for plotting (either by the current call to DAPLAT
if IPLOT is greater than or equal to zero or by a
subsequent call to DAPLAT if IPLOT is not equal to
zero). If LETTER is greater than zero, and if the
location representing the point already contains a
printing character other than a nondominant
asterisk (selected either by LETTER=0 or by
KONECT=0), then search for the closest location
which is empty or which contains a nondominant
asterisk. If LETTER is zero and if the location
representing the point already contains any
printing character, then search for the closest
location which contains a space. If, after having
searched a number of concentric boxes equal to the
absolute value of KONECT, no available location has
yet been found, then abandon the attempt. Note
that if several points coincide, it is much more
efficient to have them adjacent in the XPOINT and
YPOINT arrays if the pin map option is used since
then the locations already found to be in use do
not have to be searched again for each subsequent
point. The following are examples of the sizes of
pin groups possible for various values of KONECT.
KONECT=-2 KONECT=-3 KONECT=-4 KONECT=-5
555555555
4444444 544444445
33333 4333334 543333345
222 32223 4322234 543222345
212 32123 4321234 543212345
222 32223 4322234 543222345
33333 4333334 543333345
4444444 544444445
555555555
The plot shown below is typical of a pin map
simulation. The 7 appearances of the letter A were
plotted first, then the letter B was requested to
be plotted at the same locations as the A's, and so
on through the letter Y. KONECT had the value -4
but the search patterns did not fill the allowed 4
boxes. The rectangle drawn with asterisks was
plotted in curve mode with both LETTER and KONECT
being zero. The rectangle could have been plotted
either before or after the points of the pin map
since nondominant asterisks are considered to be
the same as spaces when a pin map is simulated, and
since nondominant asterisks cannot supersede a
letter already in the plot. In a real application,
such lines of asterisks could be used to represent
a geographical boundary or a network of streets and
FASP, FORTRAN Alphameric Subroutine Package Page 309
DAPLAT, Scatter Plot Routine for Printer
roads.
15 -+---------+---------+---------+
! UTSRQ TSRQPY !
! *VGFEP**********UGFEOX !
!UTSRQHADO VHADNTSRQ !
!VGFEPIBCN RQPIBCMGFEP !
10 -+WHADOJKLM+ SGFEJKLHADO +
!XIBCNWXY THADOWUIBCN !
!YJKLM UIBCN VJKLM !
! * VJKLM WXY !
! * WXY * !
5 -+ UTSRQ +UTSRQ + * +
! VGFEP VGFEP * !
! WHADO****WHADO********* !
! XIBCN XIBCN !
! YJKLM YJKLM !
0 -+---------+---------+---------+
! ! ! !
0 10 20 30
Many of the points in the 3 pin groups in the upper
right corner of the above plot have been shifted
further out in the search patterns due to their
intended destinations being already in use. This
can be verified by tracing the spiral but broken
paths along which the letters A through Y have been
positioned. The 2 pin groups in the upper left
corner have similarly spread slightly due to
overlap of the search patterns.
Some distortion cannot be avoided when a pin map is
simulated by DAPLAT. The concentric boxes of
characters do not print as squares due to the
differing numbers of rows and columns per inch.
Also, the pin map simulation does not provide a
true window (that is, the portion displayed is not
necessarily the same as if displayed on a plot
showing a larger area) since points which are
outside the plot are not represented even if a
search pattern centered at them would spill over
onto the plot, while points at the borders must be
shown within the plot since the search pattern does
not include the area outside the plot.
The plot shown below demonstrates these edge
distortions. This plot was generated by plotting
the same points as in the pin map shown earlier,
but the window size was reduced to include only
those coordinates which were on and within the
rectangle of asterisks in the earlier plot. The
plot size was reduced to maintain the same ratios
of data coordinates to printed columns and lines.
The left of the pin groups in the original upper
FASP, FORTRAN Alphameric Subroutine Package Page 310
DAPLAT, Scatter Plot Routine for Printer
left corner does not appear in the smaller plot
since the center of this group is outside the
smaller window. The pin group in the lower left
corner was only partially plotted since the 4 box
limit was reached before all of the requested
points could be represented. The pin group in the
lower center is both wider and extends to higher
coordinates than in the larger plot, but all points
were included without exhausting the search
pattern.
13 -+QGFEP----+----TPGFEO
!RHADO UQHADN
!SIBCN RQPIBCM
!TJKLM SGFEJKL
!UVWXY THADORS
8 -+ + UIBCN +
! VJKLM !
PONM YXWVUTS WXY !
IHGL MLKJIR !
DCFK NEDCHQ !
3 -ABEJ---OFABGP-------+
! ! !
4 14 24
= -1, store the input data as unconnected points.
Store the points having coordinates in the XPOINT
and YPOINT arrays for plotting (either by the
current call to DAPLAT if IPLOT is greater than or
equal to zero or by a subsequent call to DAPLAT if
IPLOT is not equal to zero). If LETTER is greater
than zero, and if there already is a printing
character other than a nondominant asterisk at the
location representing the point, then place an
ampersand at the location instead.
= equal to or greater than zero, store the input data
as a curve.
= 0, store the points having coordinates in the
XPOINT and YPOINT arrays for plotting, and connect
these with lines formed of nondominant asterisks.
If some other character already is in the plot, or
is later added to the plot at the locations along
the line, then this other character is seen instead
of the asterisk.
= within the range 1 through 26, KONECT is the serial
number within the alphabet of the letter used to
connect the points having coordinates in the XPOINT
and YPOINT arrays.
= 27 (or greater), store the points having
coordinates in the XPOINT and YPOINT arrays for
plotting, and connect these with lines formed of
dominant asterisks. If some other character
already is in the plot, or if a later attempt is
made to add some other character to the plot at the
FASP, FORTRAN Alphameric Subroutine Package Page 311
DAPLAT, Scatter Plot Routine for Printer
locations along the line, then the asterisk is seen
instead of this other character.
XPOINT = the array of abscissae (X or horizontal
coordinates) of the data points to be plotted.
YPOINT = the array of ordinates (Y or vertical coordinates)
of the data points to be plotted.
MINSUB = the subscript in the XPOINT and YPOINT arrays of
the coordinates of the first point to be plotted.
MAXSUB = the subscript in the XPOINT and YPOINT arrays of
the coordinates of the final point to be plotted.
If it is desired that the current call to DAPLAT
generate the previously stored plot but not add
points or lines to this plot, then MAXSUB should be
less than MINSUB.
XLEFT = the horizontal data coordinate to be placed in the
center of the character column at the left edge of
the plot. If a scale number is printed below the
left column of the plot, then this scale number
will have a value equal to that of XLEFT. Only the
portion of the curve inside the window defined by
XLEFT, YLOWER, XRIGHT and YUPPER will be shown in
the plot. Points shown on the plot must have
horizontal coordinates in the range
XLEFT-((XRIGHT-XLEFT)/(2*(MAXWID-1))) to XRIGHT+
((XRIGHT-XLEFT)/(2*(MAXWID-1))). The vertical
coordinate range would be expressed similarly. If
a line segment crosses the window, the portion of
the line segment which is outside the window will
not be represented.
The data coordinates can either increase or
decrease from left to right or from bottom to top.
In other words, XRIGHT can be either greater than
or less than XLEFT, and YUPPER can be either
greater than or less than YLOWER. If multiple
curves are being superimposed on a single plot, the
curves need not all have the same data coordinate
limits, however, the scale numbers shown on the
plot will correspond to the data coordinate limits
of the final curve.
YLOWER = the vertical data coordinate to be placed in the
center of the character line at the lower edge of
the plot. If a scale number is printed to the left
of the bottom line of the plot, then this scale
number will have a value equal to that of YLOWER.
XRIGHT = the horizontal data coordinate to be placed in the
center of the character column at the right edge of
FASP, FORTRAN Alphameric Subroutine Package Page 312
DAPLAT, Scatter Plot Routine for Printer
the plot. If a scale number is printed below the
right column of the plot, then this scale number
will have a value equal to that of XRIGHT.
YUPPER = the vertical data coordinate to be placed in the
center of the character line at the upper edge of
the plot. If a scale number is printed to the left
of the top line of the plot, then this scale number
will have a value equal to that of YUPPER.
Input Arguments Used Only if a Plot is Generated
----- --------- ---- ---- -- - ---- -- ---------
The following arguments are ignored unless a plot is
generated by the current call to DAPLAT. However, these
arguments and in particular IDISK should always be given
reasonable values since the stored plot will be generated
and the plot storage cleared if the size of the plot (as
indicated by MAXWID and MAXHIH) is changed while points and
lines are still stored from the previous call to DAPLAT.
IGRID = 0, if the plot is generated, the plot will show an
internal grid. This internal grid will be formed
of vertical lines spaced every MSHWID characters
across the width of the plot and of horizontal
lines spaced every MSHHIH lines across the height
of the plot.
= 1, if the plot is generated, the plot will show the
intersections of the lines of an internal grid, but
will not show the grid lines themselves. The
internal grid would, if shown, be formed of
vertical lines spaced every MSHWID characters
across the width of the plot and of horizontal
lines spaced every MSHHIH lines across the height
of the plot.
= 2, if the plot is generated, the plot will not show
an internal grid. However, regardless of the value
of IGRID, scale numbers will be printed along the
scales at the intervals indicated by MSHWID and
MSHHIH.
IEDGE = 0, place numbers both to the left of and below the
plot to identify the coordinate ranges.
= 1, place numbers to the left of the plot, but do
not place numbers below the plot.
= 2, place numbers below the plot, but not to the
left of the plot. Unless prevented by the value of
MARGIN, the distance between the carriage control
character in column 1 and the left edge of the plot
will be just large enough to allow a scale number
immediately below the left edge of the plot.
= 3, do not place numbers either to the left of or
below the plot. Unless prevented by the value of
FASP, FORTRAN Alphameric Subroutine Package Page 313
DAPLAT, Scatter Plot Routine for Printer
MARGIN, no characters will appear between the
carriage control character in column 1 and the left
edge of the plot.
MARGIN = the lower limit to the number of characters which
must appear in the output between the carriage
control character in column 1 and the left edge of
the plot. MARGIN is used to force the plot to
remain a fixed distance from the carriage control
character in column 1 even if a nonzero value of
IEDGE has deselected scale numbers so that this
spacing would not be necessary. MARGIN is assumed
to be at least 11 if IEDGE is zero or 1.
MSHWID = greater than zero but not greater than the width of
the plot, use MSHWID as the width of the grid
divisions stated as the number of columns of
characters. MSHWID=10 would give vertical grid
lines spaced every 10 characters across the width
of the plot area.
= 0, use the last nonzero value specified for MSHWID
as the grid division width, or use the value 10 if
a nonzero value of MSHWID has not been specified by
a previous call to this routine.
= greater than the plot width, do not plot any
vertical grid lines and do not represent scale
numbers below the plot.
The examples shown below illustrate the combination
of MAXWID=11 with various values of MSHWID.
MSHWID=4 MSHWID=10
100 -+---+A--+-- 100 -+----A----+
! * * ! ! * * !
A*********A A*********A
0 -+ **+ ** 0 -+ ** ** +
! * *** * ! * *** * !
! *** *** ! *** *** !
-100 -+A*-+---*A- -100 -+A*-----*A+
! ! ! ! !
-100 -20 60 -100 100
MSHWID=11 MSHWID=12
100 -+----A----- 100 ------A-----
! * * * *
A*********A A*********A
0 -+ ** ** 0 - ** **
! * *** * * *** *
! *** *** *** ***
-100 -+A*-----*A- -100 --A*-----*A-
!
-100
MSHHIH = greater than zero but not greater than the height
FASP, FORTRAN Alphameric Subroutine Package Page 314
DAPLAT, Scatter Plot Routine for Printer
of the plot, use MSHHIH as the height of the grid
divisions stated as the number of lines. MSHHIH=5
would give horizontal grid lines spaced every 5
lines across the height of the plot area.
= 0, use the last nonzero value specified for MSHHIH
as the grid division height, or use the value 5 if
a nonzero value of MSHHIH has not been specified by
a previous call to this routine.
= greater than the plot height, do not plot any
horizontal grid lines and do not represent scale
numbers to the left of the plot. This, unlike
deselection of the left scale numbers by IEDGE,
does not change the horizontal location of the
plot.
The examples shown below illustrate the combination
of MAXHIH=7 with various values of MSHHIH.
MSHHIH=4 MSHHIH=6
! A ! 100 -+----A----+
! * * ! ! * * !
33.333-A*********A A*********A
! ** ** ! ! ** ** !
! * *** * ! ! * *** * !
! *** *** ! ! *** *** !
-100 -+A*--+--*A+ -100 -+A*--+--*A+
! ! ! ! ! !
-100 0 100 -100 0 100
MSHHIH=7 MSHHIH=8
! A ! ! A !
! * * ! ! * * !
A*********A A*********A
! ** ** ! ! ** ** !
! * *** * ! ! * *** * !
! *** *** ! ! *** *** !
-100 -+A*--+--*A+ !A* *A!
! ! ! ! ! !
-100 0 100 -100 0 100
LTROFF = number of columns of characters by which the
leftmost grid line is offset from the left border
of the plot. LTROFF can be in the range zero up to
but not including MSHWID. If LTROFF is negative,
then it is assumed to have the value MSHWID+LTROFF
instead. If LTROFF is nonzero, then the left
border of the plot will be ruled with exclamation
points. The right border is similarly ruled if it
does not bear a grid line.
LINOFF = number of lines of characters by which the lowest
grid line is offset from the lower border of the
plot. LINOFF can be in the range zero up to but
not including MSHHIH. If LINOFF is negative, then
FASP, FORTRAN Alphameric Subroutine Package Page 315
DAPLAT, Scatter Plot Routine for Printer
it is assumed to have the value MSHHIH+LINOFF
instead. If LINOFF is nonzero, then the lower
border of the plot will be ruled with minus signs.
The upper border is similarly marked if it does not
bear a grid line.
LINPRT = -1, do not include a carriage control character to
the left of each line of the plot. Since the minus
sign of a negative scale number can then appear in
column 1, the resulting output must not be treated
as though the left column contains carriage control
characters.
= 0, the plot will be viewed by the user on a
terminal, either typed directly with IDISK being
given the terminal unit number, or typed by the
user after this routine has written the plot into a
file on the unit the number of which is contained
in IDISK. A blank or space will be used as
carriage control character to give single spacing.
= 1, the plot will be printed on the line printer by
the user after the program has written the plot
into a file. An asterisk will be used as carriage
control character to give single spacing with
suppression of skipping extra lines at the page
boundaries. On the PDP-10, an asterisk as the
carriage control character gives overprinting on
the terminal as opposed to single spacing.
IDISK = the unit number of the device onto which the plots
are to be written. This routine will only generate
the plot. It is the responsibility of the calling
program to open the output file and to write the
captions, the form feeds and/or the separating
lines. IDISK should always be given a legal unit
number since DAPLAT will generate the stored plot,
if any, using the new value of IDISK before storing
the new curve if the new plot size differs from
that used to store the previous curves even if
IPLOT=-1.
Argument Used Initially for Input, but then Returned Changed
-------- ---- --------- --- ----- --- ---- -------- -------
The following argument must be defined by the calling
program before this routine is first called. This argument
is returned by this routine set to zero, and this zero value
should be sent unchanged to any subsequent calls to this
routine.
IRESET = 0, the plot storage has been initialized by a
previous call to this routine. If IPLOT was
nonzero during the previous call to this routine,
then the current call to this routine will add to
FASP, FORTRAN Alphameric Subroutine Package Page 316
DAPLAT, Scatter Plot Routine for Printer
the plot already started by the previous call.
= (greater than zero), used when this routine is
first called to initialize the arrays and variables
which will be used to store the plot, or to cause
the current call to this routine to insert the
points or lines into a blank plot even if IPLOT was
nonzero during the previous call to this routine so
that the plot storage was not blanked out after the
plot was generated. IRESET is returned set to
zero.
= 1, plot storage is to be initialized before any
points or lines are inserted into the plot by the
current call to this routine. However, unless this
routine is used upon a computer which does not
allow the testing of the value of a variable which
has not yet been defined, having IRESET always set
to zero will produce the expected results since one
of the variables set within the routine is tested
at the start of the routine to determine whether
the plot storage has been initialized.
= 2, 3, 4 or 5, plot storage is to be initialized
except for the character set, the dominance rules
and/or the size of the plot storage which can be
predefined by the calling program. These options
are described later in this documentation.
Arguments Returned by DAPLAT Containing Error Counts
--------- -------- -- ------ ---------- ----- ------
LTRERR = returned containing the number of points which were
outside the current plot.
LINERR = returned containing the number of line segments
which were completely outside the current plot. If
a pin map is being simulated, LINERR also
accumulates the number of points which should have
been displayed on the plot, but for which there was
no empty location within the allowed distance.
LTRERR and LINERR are returned containing the error
totals accumulated during all calls to DAPLAT which
added curves or points to the same plot. If DAPLAT
is called with argument IPLOT equal to zero causing
the plot storage to be zeroed after printing the
currently stored plot, then it is the subsequent
call to DAPLAT which zeroes the error totals.
FASP, FORTRAN Alphameric Subroutine Package Page 317
DAPLAT, Scatter Plot Routine for Printer
Typical Scales Produced Using MSHWID Values in Range 1 to 10
------- ------ -------- ----- ------ ------ -- ----- - -- --
XRIGHT=-XLEFT=150 and YUPPER=-YLOWER=1234500/(10**MSHWID)
1.2345E5 -+++++++++++++++++++++++++++++++++++++++++
-1.2345E5 -+++++++++++++++++++++++++++++++++++++++++
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
*****************************************
12345 -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
-12345 -+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! ! !
-2-1-1-1-9-8-6-5-3-2 0 2 3 5 6 8 9 1 1 1 2
E2E2E2E2E1E1E1E1E1E1 E1E1E1E1E1E1E2E2E2E2
1234.5 ---+--+--+--+--+--+--+--+--+--+--+--+--+--
-1234.5 ---+--+--+--+--+--+--+--+--+--+--+--+--+--
! ! ! ! ! ! ! ! ! ! ! ! !
-1 -1 -90-68-45-23 0 23 45 68 90 1 1
E2 E2 E2 E2
123.45 -+---+---+---+---+---+---+---+---+---+---+
-123.45 -+---+---+---+---+---+---+---+---+---+---+
! ! ! ! ! ! ! ! ! ! !
-150-120 -90 -60 -30 0 30 60 90 120 150
12.345 -+----+----+----+----+----+----+----+----+
-12.345 -+----+----+----+----+----+----+----+----+
! ! ! ! ! ! ! ! !
-150 -113 -75 -37.5 0 37.5 75 113 150
1.2345 ---+-----+-----+-----+-----+-----+-----+--
-1.2345 ---+-----+-----+-----+-----+-----+-----+--
! ! ! ! ! ! !
-135 -90 -45 0 45 90 135
0.12345 -------+------+------+------+------+------
-0.12345 -------+------+------+------+------+------
! ! ! ! !
-105 -52.5 0 52.5 105
0.012345 -----+-------+-------+-------+-------+----
-0.012345 -----+-------+-------+-------+-------+----
! ! ! ! !
-120 -60 0 60 120
1.235E-3 ---+--------+--------+--------+--------+--
-1.235E-3 ---+--------+--------+--------+--------+--
! ! ! ! !
-135 -67.5 0 67.5 135
FASP, FORTRAN Alphameric Subroutine Package Page 318
DAPLAT, Scatter Plot Routine for Printer
An Example of the use of DAPLAT
-- ------- -- --- --- -- ------
As an example of the use of this routine, the following
program was used to generate the 3 plots shown below the
program.
DIMENSION XPOINT(10),YPOINT(10)
DATA XPOINT/0. ,1. ,2. ,3. ,4. ,0. ,1. ,2. ,3. ,4. /
DATA YPOINT/1. ,1.5,2.5,2.5,1.5,2. ,2.3,3. ,2.5,2. /
DATA IRESET,IDISK/1,1/
DO 2 JGRID=1,3
IGRID=JGRID-1
WRITE(IDISK,1)IGRID
1 FORMAT(7H IGRID=,I2)
C
C IPLOT=-1 MAXWID=41 MAXHIH=16 LETTER=1 KONECT=2
C MINSUB=1 MAXSUB=5 XLEFT=0 YLOWER=.5 XRIGHT=4
C YUPPER=3.5 IEDGE=0 MARGIN=0 MSHWID=0 MSHHIH=0 LTROFF=0
C LINOFF=0 LINPRT=1
CALL DAPLAT(-1,41,16,1,2,
1XPOINT,YPOINT,1,5,0.,.5,4.,
23.5,IGRID,0,0,0,0,0,
30,1,IDISK,IRESET,LTRERR,LINERR)
CHANGING IPLOT=0, LETTER=3, KONECT=4, MINSUB=6, MAXSUB=10
CALL DAPLAT(0,41,16,3,4,
1XPOINT,YPOINT,6,10,0.,.5,4.,
23.5,IGRID,0,0,0,0,0,
30,1,IDISK,IRESET,LTRERR,LINERR)
2 WRITE(IDISK,3)
3 FORMAT(//)
STOP
END
IGRID= 0
* 3.5 -+---------+---------+---------+---------+
* ! ! ! ! !
* ! ! ! ! !
* ! ! DCDD ! !
* ! ! DDDD ! DDDDD ! !
* 2.5 -+---------+-DDD----BABBBBBBB&&&D--------+
* ! DDCD BB ! !B&DD !
* ! DDDDD ! BB ! ! BBDDDD !
* CDD ! BB ! ! BB DC
* ! !BB ! ! BB !
* 1.5 -+--------BA---------+---------+--------BA
* ! BBBB ! ! ! !
* ! BBB ! ! ! !
* AB ! ! ! !
* ! ! ! ! !
* 0.5 -+---------+---------+---------+---------+
* ! ! ! ! !
* 0 1 2 3 4
FASP, FORTRAN Alphameric Subroutine Package Page 319
DAPLAT, Scatter Plot Routine for Printer
IGRID= 1
* 3.5 -+---------+---------+---------+---------+
* ! !
* ! !
* ! DCDD !
* ! DDDD DDDDD !
* 2.5 -+ + DDD BABBBBBBB&&&D +
* ! DDCD BB B&DD !
* ! DDDDD BB BBDDDD !
* CDD BB BB DC
* ! BB BB !
* 1.5 -+ BA + + BA
* ! BBBB !
* ! BBB !
* AB !
* ! !
* 0.5 -+---------+---------+---------+---------+
* ! ! ! ! !
* 0 1 2 3 4
IGRID= 2
* 3.5 -+---------+---------+---------+---------+
* ! !
* ! !
* ! DCDD !
* ! DDDD DDDDD !
* 2.5 -+ DDD BABBBBBBB&&&D +
* ! DDCD BB B&DD !
* ! DDDDD BB BBDDDD !
* CDD BB BB DC
* ! BB BB !
* 1.5 -+ BA BA
* ! BBBB !
* ! BBB !
* AB !
* ! !
* 0.5 -+---------+---------+---------+---------+
* ! ! ! ! !
* 0 1 2 3 4
FASP, FORTRAN Alphameric Subroutine Package Page 320
DAPLAT, Scatter Plot Routine for Printer
Modification of Plot Format
------------ -- ---- ------
If an internal grid is not generated, then at any edge
either the outermost grid line or the border, but not both,
can be ruled. If an internal grid is not generated, then
the digits in the tens and hundreds position in the decimal
integer value of IGRID can be used to suppress the ruling of
specific outermost grid lines. If an internal grid is not
being generated but a particular outermost grid line is
requested, then the corresponding border is not ruled. Any
border can be ruled if an internal grid is generated. If an
internal grid is generated, or if neither an internal grid
nor a particular outermost grid line is being generated,
then the value of the tens or hundreds digits of the decimal
integer value of IEDGE can direct whether the corresponding
border is to be ruled. The meanings assigned to each digit
in the value of IGRID are as follow
ONES DIGIT OF IGRID
0 The plot will show an internal grid. The values of the
tens and hundreds digits of IGRID are ignored.
1 The plot will show the intersections of the lines of an
internal grid, but will not include the grid lines
themselves. The ruling of the outermost grid lines is
controlled by the values in the tens and hundreds
digits of IGRID.
2 The plot will not include an internal grid. The ruling
of the outermost grid lines is controlled by the values
in the tens and hundreds digits of IGRID.
TENS DIGIT OF IGRID
0 Show both the leftmost and rightmost grid lines.
1 Show leftmost grid line, but do not show rightmost grid
line. Note that if the effective value of MSHWID
equals the effective value of MAXWID then a single
vertical grid line is ruled and this grid line is both
the leftmost and rightmost grid line and will be
plotted except for a tens digit value of 3.
2 Show rightmost grid line, but do not show leftmost grid
line.
3 Do not show either leftmost or rightmost grid lines.
HUNDREDS DIGIT OF IGRID
0 Show both the lower and upper grid lines.
1 Show lower grid line, but do not show upper grid line.
Note that if the effective value of MSHHIH equals the
effective value of MAXHIH then a single horizontal grid
line is ruled and this grid line is both the lower and
upper grid line and will be plotted except for a
hundreds digit value of 3.
FASP, FORTRAN Alphameric Subroutine Package Page 321
DAPLAT, Scatter Plot Routine for Printer
2 Show upper grid line, but do not show lower grid line.
3 Do not show either lower or upper grid lines.
The tens and hundreds digits in the decimal integer value of
IEDGE are used to suppress the ruling of specific edges of
the plot which do not happen to coincide with visible grid
lines. The meanings assigned to each digit in the value of
IEDGE are as follow
ONES DIGIT OF IEDGE
0 Show numbers both to the left of and below the plot.
1 Show numbers to the left of the plot, but not below the
plot.
2 Show numbers below the plot, but not to the left of the
plot.
3 Do not show numbers either to the left of or below the
plot.
TENS DIGIT OF IEDGE
0 Show both the left and right border lines.
1 Show left border line, but do not show the right border
line.
2 Show right border line, but do not show left border
line.
3 Do not show either left or right border lines.
HUNDREDS DIGIT OF IEDGE
0 Show both the lower and upper border lines.
1 Show lower border line, but do not show the upper
border line.
2 Show upper border line, but do not show lower border
line.
3 Do not show either lower or upper border lines.
Erasure of Points, Lines and Pin Clusters
------- -- ------ ----- --- --- --------
A LETTER argument value of -1 will cause the erasure of the
specified points, lines or selected levels in pin clusters.
If LETTER is -1 and KONECT is zero (or greater), all
printing characters will be changed to spaces along the
lines specified by values in the XPOINT and YPOINT arrays.
If LETTER is -1 and KONECT is less than zero, spaces will
replace the printing characters on the edges of the square
selected by taking the central point as square 1 and
counting the concentric squares around this point until the
absolute value of KONECT is reached.
FASP, FORTRAN Alphameric Subroutine Package Page 322
DAPLAT, Scatter Plot Routine for Printer
The illustrations shown below demonstrate erasure of edges
of pin clusters. At the left is shown the starting plot
completely filled with the letter A. At the right is shown
the plot after it has been modified by calling DAPLAT with
LETTER being -1 while KONECT was consecutively -1 (revealing
the center point at 1,1), -3, -5, -7, -9 and -11 (revealing
the columns containing the left and right grid lines). The
erasure of each box required a separate call to DAPLAT with
XPOINT(1)=1.0, YPOINT(1)=1.0, MINSUB=1, and MAXSUB=1.
2 -AAAAAAAAAAAAAAAAAAAAA 2 -+A-A-AAAAAAAAAAA-A-A+
AAAAAAAAAAAAAAAAAAAAA !A A A ! A A A!
AAAAAAAAAAAAAAAAAAAAA !A A A AAAAAAA A A A!
AAAAAAAAAAAAAAAAAAAAA !A A A A ! A A A A!
AAAAAAAAAAAAAAAAAAAAA !A A A A AAA A A A A!
1 -AAAAAAAAAAAAAAAAAAAAA 1 -+A-A-A-A-A+A-A-A-A-A+
AAAAAAAAAAAAAAAAAAAAA !A A A A AAA A A A A!
AAAAAAAAAAAAAAAAAAAAA !A A A A ! A A A A!
AAAAAAAAAAAAAAAAAAAAA !A A A AAAAAAA A A A!
AAAAAAAAAAAAAAAAAAAAA !A A A ! A A A!
0 -AAAAAAAAAAAAAAAAAAAAA 0 -+A-A-AAAAAAAAAAA-A-A+
! ! ! ! ! !
0 1 2 0 1 2
If LETTER is -2 and KONECT is -2 or less, then spaces will
replace the printing characters either on the edges of, or
within, the square selected by KONECT. LETTER values of -1
and -2 are equivalent if KONECT is -1 or greater. As an
example of the erasure of entire pin clusters, a starting
plot completely filled with the letter A was modified to
produce the plot shown below by specifying the X and Y
coordinates listed to its left with LETTER=-2 and KONECT=-3.
2 -AAAAAAAAAAA-----AA--+
X Y ! AAAAAA AA !
! AAAAAA AA !
0.0 0.0 ! AAA AAAAA
0.2 1.4 ! AAA AAAAAAAA
0.7 0.2 1 -+ AAA + AAA +
1.0 1.0 AAAAAAAA AAA !
1.3 1.8 AAAAA AAA !
1.8 0.6 ! AA AAAAAA !
2.0 2.0 ! AA AAAAAA !
0 -+--AA-----AAAAAAAAAAA
! ! !
0 1 2
Increasing Maximum Plot Size and Restoring Plot Storage
---------- ------- ---- ---- --- --------- ---- -------
As supplied, DAPLAT contains a storage array large enough to
hold 5151 characters, this being the number of characters
needed for a plot of the default size of 51 lines each of
FASP, FORTRAN Alphameric Subroutine Package Page 323
DAPLAT, Scatter Plot Routine for Printer
101 characters. This plot storage is contained in the
labeled block COMMON/FASPA/ which must contain at least 20
words more than the words needed to store MAXWID*MAXHIH
characters, six characters being held in the right 30 bits
of each word. The 20 additional words store information
about plot size and scaling. The labeled common block
COMMON/FASPA/ contains all the information necessary to
generate the plot, with the exception of the character list
in COMMON/FASPB/, the character dominance rules in
COMMON/FASPC/ (needed if more points or lines are to be
added to the same plot), and the arguments LINPRT and IDISK.
If a plot is requested which is larger than can be held in
the labeled common block, then the effective value of MAXHIH
is reduced so that the plot will fit. The labeled common
block can be made larger than the 879 locations (calculated
as 20+(((51*101)+5)/6) words) provided in this routine if
a. The larger specification of the labeled common block
COMMON/FASPA/ is loaded by the calling program or by a
block data routine before DAPLAT is loaded. On the
DECsystem10 computer at least, the specification of a
labeled common block which is loaded first establishes
its length, but must be at least as large as the
specification of the same common block in any routine
which is loaded later.
b. The first three locations in the labeled common block
COMMON/FASPA/ are integer.
c. Prior to the first call to DAPLAT, the first 3
locations in the labeled common block COMMON/FASPA/ are
loaded with the following information either by a block
data routine or by the main program.
1st location (initialized to have the value 859 if
this routine is called with IRESET=1 or 4) must
contain a value which is 20 less than the total
number of locations in this common block. This is
the number of locations in this common block which
can be used to store values identifying the
characters which are to appear upon the plot
surface.
2nd location (initialized to have the value 30 if this
routine is called with IRESET=1 or 4) must contain
a value which is 3 more than the maximum value of
the arguments LETTER and KONECT. This is the
number of different characters by which data
points and lines can be represented, including in
this count the space indicating an empty position,
and the ampersand indicating superimposed data.
If the second location in COMMON/FASPA/ is set to
a value greater than 30, then the calling program
FASP, FORTRAN Alphameric Subroutine Package Page 324
DAPLAT, Scatter Plot Routine for Printer
must also allocate a correspondingly larger
COMMON/FASPB/ and COMMON/FASPC/, must itself
define the contents of these common blocks and
must call this routine initially with IRESET
having the value 5 to prevent the redefinition by
this routine of the contents of these common
blocks. If the second location in COMMON/FASPA/
is set to the value 3, then the space, ampersand
and nondominant asterisk (or whatever happens to
be selected when LETTER or KONECT is set to zero)
would be the only characters allowed, other than
those forming the grid lines and grid line
intersections, in the plot.
3rd location (initialized to have the value 6 if this
routine is called with IRESET=1 or 4) must contain
a value which is 1 more than the number of times
the value in the second location in this common
block can be multiplied by itself and still yield
as the result a value which can be stored as a
positive single precision integer by the computer
upon which this routine is being run. This is the
number of characters, each of which can be
identified by a value in the range zero through 1
less than the value in the 2nd location in this
common block, which can be stored in each of the
locations in an array the dimension of which is
indicated by the value in the first location in
this common block. The value in this 3rd location
cannot exceed 22 unless the size of COMMON/FASPD/
is increased. If LETTER or KONECT can take on
values through 27, then the values representing 6
characters can be packed into a single positive 32
bit number. If LETTER and KONECT cannot exceed
zero, then the values representing 22 characters
can be packed into a single positive 36 bit number
although then one of the 3 possible values in each
byte is wasted since the ampersand would not be
necessary.
d. The argument named IRESET has either of the values 2 or
5 when this routine is first called. These are the
only values of IRESET which initialize the rest of plot
storage in labeled common block COMMON/FASPA/, but do
not modify the values of the first 3 locations. IRESET
controls the following selective initialization of the
plot storage in COMMON/FASPA/, of the character list in
COMMON/FASPB/ and of the character dominance rules in
COMMON/FASPC/.
IRESET = 0, the entire plot storage, the character
list, and the character dominance rules are
not changed.
= 1, the entire plot storage, the character list
FASP, FORTRAN Alphameric Subroutine Package Page 325
DAPLAT, Scatter Plot Routine for Printer
and the dominance rules are to be initialized.
= 2, the character list, the character dominance
rules, and all but the first three locations
in the plot storage are to be initialized.
The first three locations in the plot storage
are not changed.
= 3, the character list and the character
dominance rules are to be initialized. The
entire plot storage is not changed.
= 4, the entire plot storage is to be
initialized. The character list and character
dominance rules are not changed.
= 5, all but the first three locations in the
plot storage are to be initialized. The
character list, the character dominance rules
and the first three locations in the plot
storage are not changed.
A partially constructed plot can be added to and/or be
displayed by another program if
a. DAPLAT has been called with its argument IPLOT being
given either of the values -1 or 1 so that a
description of the plot is preserved in the labeled
block COMMON/FASPA/.
b. The entire contents of COMMON/FASPA/ are written into a
binary file after DAPLAT returns control to the calling
program. If the character set or the dominance rules
have been changed, then the contents of COMMON/FASPB/
and of COMMON/FASPC/ should similarly be written into
the file.
c. The subsequent program restores the contents of the
common block or blocks, then defines IRESET to be 3
before DAPLAT is first called if only the contents of
COMMON/FASPA/ have been restored, or else defines
IRESET to be zero if the contents of COMMON/FASPB/ and
of COMMON/FASPC/ have similarly been restored.
Changing the Character Set Used for Points and Curves
-------- --- --------- --- ---- --- ------ --- ------
The labeled block COMMON/FASPB/ contains the characters with
which DAPLAT represents points and curves. This common
block is dimensioned at 35, and contains the following
characters
FASP, FORTRAN Alphameric Subroutine Package Page 326
DAPLAT, Scatter Plot Routine for Printer
COMMON/FASPB/LTRALL(35)
DIMENSION LTRBGN(35)
DATA LTRBGN/1H ,1H*,1H!,1H-,1H+,1H ,1H&,1H*,1HA,1HB,
1 1HC,1HD,1HE,1HF,1HG,1HH,1HI,1HJ,1HK,1HL,
2 1HM,1HN,1HO,1HP,1HQ,1HR,1HS,1HT,1HU,1HV,
3 1HW,1HX,1HY,1HZ,1H*/
DO 1 I=1,35
1 LTRALL(I)=LTRBGN(I)
LTRALL(1) and LTRBGN(2) contain the carriage control
characters used for LINPRT=0 and 1 respectively.
LTRALL(3) through LTRALL(5) contain the characters used for
vertical grid lines, for horizontal grid lines, and for grid
line intersections respectively.
LTRALL(6) contains the space character used for the
background of the plot. Changing this character does not
change the space character appearing around the scale
numbers in the margins.
LTRALL(7) contains the ampersand used where curves and/or
points coincide.
LTRALL(8) through LTRALL(35) contain the characters used for
LETTER or KONECT=0 through LETTER or KONECT=27 respectively.
The dominance rules for the characters in LTRALL(7) through
and including LTRALL(35) are contained in COMMON/FASPC/.
COMMON/FASPC/MASTER(29)
MASTER(1)=1 !AMPERSAND DOMINANT OVER ALL BUT 2ND *
MASTER(2)=-1 !NONDOMINANT ASTERISK
DO 2 I=3,28
2 MASTER(I)=0 !ALPHABETIC LETTERS ALL OF SAME DOMINANCE
MASTER(29)=2 !DOMINANT ASTERISK
When characters having different values in the MASTER array
coincide, the character having the larger value in the
MASTER array is dominant. If the values are equal, then the
character at that position is converted to an ampersand (or
to whatever happens to be in LTRALL(7)).
If either the character list or the dominance rules are to
be changed, then either a block data routine or the main
program must define both the LTRALL and MASTER arrays and
DAPLAT must be called the first time with IRESET set to
either 4 or 5 depending upon whether or not DAPLAT is to
initialize the first three locations in COMMON/FASPA/.
If LETTER and KONECT are to be allowed to take on values
greater than 27, then the calling program must allocate a
correspondingly larger COMMON/FASPB/ and COMMON/FASPC/, must
itself define the first three locations in COMMON/FASPA/ as
FASP, FORTRAN Alphameric Subroutine Package Page 327
DAPLAT, Scatter Plot Routine for Printer
well as the entire contents of COMMON/FASPB/ and of
COMMON/FASPC/ before this routine is first called, and must
call this routine initially with IRESET having the value 5
to prevent the redefinition by this routine of any of this
information.
Plots are constructed using the characters in COMMON/FASPB/
when DAPLAT is called having the argument IPLOT greater than
or equal to zero. Only the position of the letters within
the common block as indicated by the values of LETTER and/or
KONECT are used when storing the plot, so the same stored
curves and/or points can be plotted with different character
sets by different calls to DAPLAT.
Hidden Switch used to Enable Time Series Plots
------ ------ ---- -- ------ ---- ------ -----
This section describes a feature of DAPLAT which allows its
use for the construction of tall plots formed of segments
generated one per each DAPLAT call. This feature is not
meant for the casual user, but is described so that the
total interface presented by DAPLAT to all other programs is
documented.
The hidden switch is merely a 2 state, on/off condition
enabled by the sign of the argument IGRID which would not
logically have a negative value. If IGRID is less than
zero, then 1 less than its absolute value is used as
described in the preceding portion of this manual, and the
bottom line of the plot is not represented in the plot.
Instead, after the plot has been generated, the characters
stored for the bottom line replace those stored for the top
line and the rest of the plot storage is left empty.
The following sample plots demonstrate the results obtained
by the use of the hidden switch. The double lines at the
tops of the left two plots and at the bottom of the third
plot are margin lines which are visible since they do not
coincide with grid lines.
IGRID=1 IGRID=-2 IGRID=1 IGRID=-2
LINOFF=0 LINOFF=0 LINOFF=1 LINOFF=1
! ! ! ! 2.3 -+---+---+ 2.3 -+---+---+
2 -+---+---+ 2 -+---+---+ ! ! ! !
! ! ! ! ! ! ! !
! ! ! ! 1.3 -+ + + 1.3 -+ + +
1 -+ + + 1 -+ + + ! ! ! !
! ! ! ! ! ! ! !
! ! ! ! 0.3 -+---+---+ 0.3 -+---+---+
0 -+---+---+ ! !
! ! ! ! ! !
100 200 300 100 200 300
FASP, FORTRAN Alphameric Subroutine Package Page 328
DAPLAT, Scatter Plot Routine for Printer
Rationalization of Plot Scales
--------------- -- ---- ------
The proportions of the grid superimposed upon the plot area
by the DAPLAT routine must be specified by the calling
program. The numbers which are printed beside and below the
plot merely represent the data unit coordinates plotted at
the centers of the columns and lines which bear grid lines.
The range of data unit coordinates represented by these
scales can be difficult to interpret if the calling program
holds the grid proportions constant while attempting to
obtain the maximum resolution by adjusting the minimum and
maximum coordinates to spread the curves across the full
width or height of the plot. DAPLAT cannot itself calculate
the grid proportions which would produce a minimum number of
nonzero digits in each scale number. If predetermined grid
spacings and offsets are not necessary, then the routine
DAGRID can be called once for each scale before the data is
plotted to select the grid proportions and to expand the
coordinate ranges slightly to give neater scale numbers.
If DAGRID is used to rationalize the horizontal scale, then
the argument list can be stated as
SUBROUTINE DAGRID(JSTIFY,MAXWID,XLEFT ,XRIGHT,MSHWID,
1 LTROFF)
or, if the routine is used to rationalize the vertical
scale, then the argument list can instead by stated as
SUBROUTINE DAGRID(JSTIFY,MAXHIH,YLOWER,YUPPER,MSHHIH,
1 LINOFF)
The first two arguments, JSTIFY and MAXWID, are used only
for input and are returned unchanged. The next two
arguments, XLEFT and XRIGHT (or YLOWER and YUPPER), are used
both for input of the original range and for returning the
modified range. The final two arguments, MSHWID and LTROFF
(or MSHHIH and LINOFF), are used only for returning a
description of the grid to the calling program and their
input values are ignored. These arguments are defined as
follow.
JSTIFY = -2, a grid line is to be included in the left
column (or bottom line) of the plot which would be
produced by DAPLAT.
= -1, the data unit coordinate represented by the
value of XLEFT (or YLOWER) is to be shown in the
left column (or bottom line) of the plot which
would be produced by DAPLAT.
= 0, the data unit coordinate range represented by
the values of XLEFT and XRIGHT (or YLOWER and
YUPPER) is to be centered in the plot which would
be produced by DAPLAT.
FASP, FORTRAN Alphameric Subroutine Package Page 329
DAPLAT, Scatter Plot Routine for Printer
= 1, the data unit coordinate represented by the
value of XRIGHT (or YUPPER) is to be shown in the
right column (or top line) of the plot which would
be produced by DAPLAT.
= 2, a grid line is to be included in the right
column (or top line) of the plot which would be
produced by DAPLAT.
MAXWID = width (or height) of the plot stated as the number
of columns (or lines). MAXWID should not be input
containing a zero or negative value.
XLEFT = input containing the data unit coordinate to be
represented by the center of the left column (or
bottom line) of plot. This should be the smallest
data unit coordinate to be represented if the scale
increases toward the right (or top) since DAGRID
allows computation errors which place the returned
limiting data unit coordinates into the outer
halves of the outermost columns (or lines).
= returned containing the data unit coordinate which
would give a rational scale if represented by the
center of the left column (or of the bottom line)
of the plot.
XRIGHT = input containing the data unit coordinate to be
represented by the center of the right column (or
top line) of the plot. This should be the largest
data unit coordinate to be represented if the scale
increases toward the right (or top).
= returned containing the data unit coordinate which
would give a rational scale if represented by the
center of the right column (or of the top line) of
the plot.
MSHWID = returned containing the grid division width (or
height) stated as the number of columns (or lines)
which would give a rational scale. MSHWID is
returned containing 1 more than the number of
columns (or lines) between successive columns (or
lines) which should bear grid lines.
= 0, returned if this routine could not rationalize
the scale. This could happen if either the data
unit coordinate range or the width of the plot is
too small.
LTROFF = returned containing the number of columns (or
lines) by which the left (or bottom) grid line
should be offset from the left (or bottom) edge of
the plot to give a rational scale.
FASP, FORTRAN Alphameric Subroutine Package Page 330
DAPLAT, Scatter Plot Routine for Printer
An Example of the Use of DAGRID
-- ------- -- --- --- -- ------
The following demonstration program asks the user to type
the plot width and the limiting data unit coordinates, then
calls DAPLAT to display scales for each of the possible
values of JSTIFY. A typical dialog between the program and
its user is shown after the listing of the program.
C PROGRAM TO DEMONSTRATE DAGRID ROUTINE
DIMENSION XPOINT(2),YPOINT(2)
DATA YPOINT/0.0,0.0/
DATA ITTY,JTTY/5,5/
DIMENSION EXAMPL(20)
IRESET=1
1 WRITE(JTTY,2)
2 FORMAT(7H WIDTH ,$)
READ(ITTY,3)MAXWID
3 FORMAT(I)
IF(MAXWID.LE.0)MAXWID=51
WRITE(JTTY,4)
4 FORMAT(7H RANGE ,$)
READ(ITTY,5)XLEFT,XRIGHT
5 FORMAT(2F)
XPOINT(1)=XLEFT
XPOINT(2)=XRIGHT
JSTIFY=-2
6 CALL DAGRID(JSTIFY,MAXWID,XLEFT,XRIGHT,MSHWID,LTROFF)
IF(MSHWID.LE.0)GO TO 1
IF(JSTIFY.EQ.-2)WRITE(JTTY,7)XLEFT,XRIGHT,MSHWID,
1LTROFF
IF(JSTIFY.EQ.-1)WRITE(JTTY,8)XLEFT,XRIGHT,MSHWID,
1LTROFF
IF(JSTIFY.EQ.0)WRITE(JTTY,9)XLEFT,XRIGHT,MSHWID,
1LTROFF
IF(JSTIFY.EQ.1)WRITE(JTTY,10)XLEFT,XRIGHT,MSHWID,
1LTROFF
IF(JSTIFY.EQ.2)WRITE(JTTY,11)XLEFT,XRIGHT,MSHWID,
1LTROFF
7 FORMAT(1X/22H LEFT JUSTIFIED GRID ,2E14.6,2I3)
8 FORMAT(1X/22H LEFT JUSTIFIED RANGE ,2E14.6,2I3)
9 FORMAT(1X/22H CENTERED RANGE ,2E14.6,2I3)
10 FORMAT(1X/22H RIGHT JUSTIFIED RANGE,2E14.6,2I3)
11 FORMAT(1X/22H RIGHT JUSTIFIED GRID ,2E14.6,2I3)
CALL DAPLAT(0,MAXWID,1,0,0,
1XPOINT,YPOINT,1,2,XLEFT,0.0,XRIGHT,
21.0,0,2,5,MSHWID,1,LTROFF,
30,0,5,IRESET,LTRERR,LINERR)
IF(JSTIFY.GT.1)GO TO 1
XLEFT=XPOINT(1)
XRIGHT=XPOINT(2)
JSTIFY=JSTIFY+1
GO TO 6
END
FASP, FORTRAN Alphameric Subroutine Package Page 331
DAPLAT, Scatter Plot Routine for Printer
WIDTH 51
RANGE -27 79
LEFT JUSTIFIED GRID -0.500000E+02 0.888889E+02 9 0
+-------***************************************---+
! ! ! ! ! !
-50 -25 0 25 50 75
LEFT JUSTIFIED RANGE -0.272727E+02 0.863636E+02 11 1
************************************************--+
! ! ! ! !
-25 0 25 50 75
CENTERED RANGE -0.318182E+02 0.818182E+02 11 3
+-************************************************+
! ! ! ! !
-25 0 25 50 75
RIGHT JUSTIFIED RANGE -0.340909E+02 0.795455E+02 11 4
+--************************************************
! ! ! ! !
-25 0 25 50 75
RIGHT JUSTIFIED GRID -0.388889E+02 0.100000E+03 9 5
+---***************************************-------+
! ! ! ! ! !
-25 0 25 50 75 100
WIDTH 51
RANGE 112 18
LEFT JUSTIFIED GRID 0.125000E+03 0.113636E+02 11 0
+-----******************************************--+
! ! ! ! !
125 100 75 50 25
LEFT JUSTIFIED RANGE 0.112000E+03 0.120000E+02 5 1
************************************************--+
! ! ! ! ! ! ! ! ! !
110 100 90 80 70 60 50 40 30 20
CENTERED RANGE 0.116000E+03 0.160000E+02 5 3
+-************************************************+
! ! ! ! ! ! ! ! ! !
110 100 90 80 70 60 50 40 30 20
RIGHT JUSTIFIED RANGE 0.118000E+03 0.180000E+02 5 4
+--************************************************
! ! ! ! ! ! ! ! ! !
110 100 90 80 70 60 50 40 30 20
RIGHT JUSTIFIED GRID 0.113636E+03 0.000000E+00 11 6
+******************************************-------+
! ! ! ! !
100 75 50 25 0
FASP, FORTRAN Alphameric Subroutine Package Page 332
DARANK, Returns Sorted Integers Excluding Duplicates
DDDDD AAA RRRRRR AAA NN NN KK KK
DD DD AAAA RR RR AAAA NNN NN KK KK
DD DD AA AA RR RR AA AA NNNN NN KK KK
DD DD AA AA RRRRRR AA AA NN NN NN KKKKK
DD DD AAAAAAA RR RR AAAAAAA NN NNNN KKK KK
DD DD AA AA RR RR AA AA NN NNN KK KK
DDDDD AA AA RR RR AA AA NN NN KK KK
DARANK, Returns Sorted Integers Excluding Duplicates
------ ------- ------ -------- --------- ----------
A single call to DARANK interprets an array read by the
calling program with a multiple of an A1 format and returns
the decimal integer values represented by this array. The
returned values are sorted into either increasing or
decreasing order and do not include duplicate values. If
more values are represented than can be stored in the array
provided by the calling program for returning these values,
then either the smallest or the largest values can be
discarded.
Numbers can be separated by spaces, by tab characters and/or
by commas. Excess commas are ignored and do not indicate
either missing or zero values. The evaluation is terminated
when a semicolon is found within the contents of the input
text buffer or else when all of the characters within the
input text buffer have been interpreted. An exclamation
point and any characters to its right are taken to form a
comment and are otherwise ignored. An ampersand and any
characters to its right are similarly ignored, but the
calling program is informed that an ampersand was found so
that the calling program can read new text into the input
buffer before calling this routine again to continue the
evaluation.
The representation of a number can contain leading sign,
embedded decimal point and/or trailing letter E with
possibly signed exponent. A percent sign following the
number implies E-2, trailing letter K implies E3 and
trailing letter M implies E6. To be recognized as part of a
number, the percent sign or letter K or M or E must follow
either a sign or a digit or a decimal point, otherwise the
percent sign or letter K or M or E will be an unknown
character.
Two versions of this routine are supplied. The longer
version, named DARANK, requires that the routines DANEXT,
DASPAN and DAHEFT be loaded to allow the evaluation of
ranges specified in either slash, colon or asterisk
notations. The shorter version, named DAIRNK, treats the
slash, colon and asterisk characters the same as any other
unknown characters, but only requires that the additional
routines DAMISS and DAHEFT be loaded.
FASP, FORTRAN Alphameric Subroutine Package Page 333
DARANK, Returns Sorted Integers Excluding Duplicates
The range specifications accepted by the longer version of
this routine are written as a lower bound, increment and
upper bound separated by slashes or by colons (the 2
characters are equivalent). If the increment is to be one,
then the lower and upper bounds need be separated only by a
single slash or by a single colon. If either bound is
missing, it is assumed to be zero. The lower bound is the
first number of the series. The series can then either
increase or decrease depending upon whether the upper bound
is less than, equal to, or greater than the lower bound.
The sign of the increment is changed if it does not conform
to the relative values of the bounds. If a range is
specified which contains more values than can be returned to
the calling program, then the portion of the range which
could contain acceptable values is calculated, so that all
values within the originally specified range do not have to
be individually tested.
The longer version of this routine will also accept a single
value preceded by a number and an asterisk. The asterisk
notation is used by other routines in FASP to indicate
multiple appearances of a single value, but the initial
number is ignored by DARANK other than to determine that it
is greater than zero. If the initial number is not greater
than zero, then the value following the separating asterisk
is not returned to the calling program.
The DARANK and DAIRNK Argument Lists
--- ------ --- ------ -------- -----
The argument lists of routines DARANK and DAIRNK are
SUBROUTINE DARANK(INCRES,IFTEST,MINMUM,MAXMUM,MINSTR,
1 MAXSTR,IBUFFR,MAXBFR,LOWBFR,KIND ,MAXUSD,ISTORE)
and
SUBROUTINE DAIRNK(INCRES,IFTEST,MINMUM,MAXMUM,MINSTR,
1 MAXSTR,IBUFFR,MAXBFR,LOWBFR,KIND ,MAXUSD,ISTORE)
both with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR),ISTORE(MAXSTR)
The argument lists of the 2 routines are identical, with the
exception that DAIRNK does not recognize range
specifications and so will not return KIND containing the
value 5 which would indicate an illegal range specification.
The following arguments are used for input and are returned
unchanged.
INCRES = specifies whether the returned values are to be
FASP, FORTRAN Alphameric Subroutine Package Page 334
DARANK, Returns Sorted Integers Excluding Duplicates
sorted into decreasing or increasing order, and
whether it is the smaller or the larger values
which are discarded when more values are found than
can be returned in ISTORE(MINSTR) through and
including ISTORE(MAXSTR).
= 1, the returned values are to be sorted into
decreasing order such that ISTORE(MINSTR) contains
a value greater than ISTORE(MINSTR+1). If more
values are found than can be returned in the
available portion of the ISTORE array, then the
smaller values are discarded.
= 2, the returned values are to be sorted into
decreasing order such that ISTORE(MINSTR) contains
a value greater than ISTORE(MINSTR+1). If more
values are found than can be returned in the
available portion of the ISTORE array, then the
larger values are discarded.
= 3, the returned values are to be sorted into
increasing order such that ISTORE(MINSTR) contains
a value smaller than ISTORE(MINSTR+1). If more
values are found than can be returned in the
available portion of the ISTORE array, then the
smaller values are discarded.
= 4, the returned values are to be sorted into
increasing order such that ISTORE(MINSTR) contains
a value smaller than ISTORE(MINSTR+1). If more
values are found than can be returned in the
available portion of the ISTORE array, then the
larger values are discarded.
IFTEST = specifies whether the calling program has specified
minimum and/or maximum limits to the range of
acceptable values. Values which are smaller than
the minimum limit or which are greater than the
maximum limit are discarded if these limits are
specified.
= -2 or 2, there are no minimum and maximum limits to
the range of acceptable values.
= -1, values which are smaller than the value of the
argument named MINMUM are discarded.
= 0, values which are smaller than the value of the
argument named MINMUM or which are greater than the
value of the argument named MAXMUM are discarded.
If MINMUM is greater than MAXMUM, then no values
whatever will be accepted.
= 1, values which are greater than the value of the
argument named MAXMUM are discarded.
MINMUM = lower limit of the range of acceptable values if
IFTEST has either of the values -1 or 0. MINMUM is
ignored if IFTEST has any value other than -1 or 0.
MAXMUM = upper limit of the range of acceptable values if
IFTEST has either of the values 0 or 1. MAXMUM is
FASP, FORTRAN Alphameric Subroutine Package Page 335
DARANK, Returns Sorted Integers Excluding Duplicates
ignored if IFTEST has any value other than 0 or 1.
MINSTR = subscript of the first location in the ISTORE array
into which a value can be stored if the argument
named KIND is input containing a value less than 3.
If KIND is input containing a value of 3 or
greater, then ISTORE(MINSTR) through and including
ISTORE(MAXUSD) are assumed to contain a group of
integer values which were returned by a previous
call to this routine sorted in the same direction
as currently requested and which are to be merged
with the values found by the current call to this
routine.
MAXSTR = subscript of the final location in the ISTORE array
into which a value can be stored. This routine
does not report whether more values were found than
can be returned in the available portion of the
ISTORE array. If the user is to be warned by the
calling program that some values were discarded,
then the calling program should allocate at least
one more location in the ISTORE array than is
expected to be necessary. The calling program
could then warn the user that values have been
discarded if MAXUSD is returned set equal to
MAXSTR.
IBUFFR = input buffer array containing characters typed by
the user, read by a multiple of an A1 format, which
is to be searched for the representations of
numbers. IBUFFR then contains 1 character per
computer storage location.
MAXBFR = subscript of the final (rightmost) location in the
IBUFFR array which can be searched for the
representation of a number.
The following arguments are used both for input to this
routine and for output to the calling program.
LOWBFR = input containing the subscript within the IBUFFR
array of the first (leftmost) character which can
be scanned for the representation of a number.
LOWBFR is returned pointing beyond the end of the
buffer if no printing characters other than commas
appear in addition to the numbers in the buffer, or
if either an exclamation point or an ampersand is
encountered. If a semicolon is found, then LOWBFR
is returned pointing to the character to the right
of the semicolon. If an unknown character is
found, then LOWBFR is returned pointing to the
unknown character and must be incremented by the
calling program before this routine is called again
to continue the evaluation of the remaining
FASP, FORTRAN Alphameric Subroutine Package Page 336
DARANK, Returns Sorted Integers Excluding Duplicates
contents of the IBUFFR array. The calling program
must reset LOWBFR to point to the first character
in the buffer each time a new line of text is read.
KIND = should be set to zero by the calling program before
this routine is first called to return a new group
of sorted integers. KIND is returned describing
the reason for which control has been transferred
back to the calling program. KIND should be sent
to the subsequent call to this routine unchanged if
additional values are to be searched for following
an error, or if an ampersand has directed that the
calling program read new text into the input
buffer. This routine will reset MAXUSD to have the
value MINSTR-1 if KIND is input containing a value
less than 3. MAXUSD will be left unchanged, and
new values will be merged with the previously
sorted values in ISTORE(MINSTR) through and
including ISTORE(MAXUSD) if KIND is input
containing a value of 3 or greater.
= 1, returned if no printing characters other than
commas appear in addition to the numbers in the
input buffer, or if an exclamation point is found
to the right of the numbers and/or commas. If an
exclamation point is found, then the characters to
the right of the exclamation point are assumed to
form a comment and are not evaluated and LOWBFR is
returned containing MAXBFR+1. The calling program
should read a new line of text into the input
buffer and reset LOWBFR to point to the first
character in the new text before again calling this
or any other routine in FASP. If this routine is
called again without the calling program having
first changed the value of KIND to have a value of
3 (or greater), then MAXUSD will be reset to
MINSTR-1 before any new values are searched for,
and any values already in the ISTORE array will be
discarded. If the subsequent call to this routine
is instead to continue the merging of the new
values with those previously found, then the
calling program should rest KIND to have the value
3 (or greater) before this routine is again called.
= 2, returned if a semicolon is found to the right of
the numbers and/or commas in the input text buffer.
LOWBFR is returned pointing to the character to the
right of the semicolon. If this routine is called
again without the calling program having first
changed the value of KIND to have a value of 3 (or
greater), then MAXUSD will be reset to MINSTR-1
before any new values are searched for, and any
values already in the ISTORE array will be
discarded. If the subsequent call to this routine
is instead to continue the merging of the new
values with those previously found, then the
FASP, FORTRAN Alphameric Subroutine Package Page 337
DARANK, Returns Sorted Integers Excluding Duplicates
calling program should rest KIND to have the value
3 (or greater) before this routine is again called.
= 3, returned if an ampersand was found to the right
of the numbers and/or commas. The characters to
the right of the ampersand are assumed to form a
comment and are not evaluated. LOWBFR is returned
containing MAXBFR+1. The calling program should
read new text into the input buffer and reset
LOWBFR to point to the first character of the new
text. If this routine is called again without the
value of KIND having been changed, then any new
values which are found will be merged with those
already in ISTORE(MINSTR) through and including
ISTORE(MAXUSD).
= 4, returned if an unknown character was found in
the input text buffer. LOWBFR is returned pointing
to this unknown character and must be incremented
by the calling program before this routine is
called again. The letters E, K and M and the
percent sign are unknown if these do not appear
within numbers. The asterisk, slash and colon
characters are unknown to DAIRNK. If this routine
is called again without the value of KIND having
been changed, then any new values which are found
will be merged with those already in ISTORE(MINSTR)
through and including ISTORE(MAXUSD).
= 5, returned if a range specification was found
which contains too many asterisks, too many slashes
or too many colons. LOWBFR is returned pointing to
the character to the immediate right of the range
specification. DAIRNK would instead return KIND
containing the value 4 to indicate an unknown
character if an asterisk, slash or colon was found.
If this routine is called again without the value
of KIND having been changed, then any new values
which are found will be merged with those already
in ISTORE(MINSTR) through and including
ISTORE(MAXUSD).
MAXUSD = returned containing the subscript of the highest
location in the ISTORE array used for returning the
sorted values. The first value found by this
routine is placed into ISTORE(MINSTR) if KIND is
input containing a value less than 3, or will be
merged into the sorted numbers already in
ISTORE(MINSTR) through and including ISTORE(MAXUSD)
if KIND is input containing a value of 3 or greater
indicating that the previous call to this routine
has already begun the evaluation of the group of
sorted numbers.
ISTORE = array in which the sorted decimal integer values
are returned.
FASP, FORTRAN Alphameric Subroutine Package Page 338
DARANK, Returns Sorted Integers Excluding Duplicates
An Example of the Use of DARANK
-- ------- -- --- --- -- ------
The following demonstration program reports the values
represented by the text typed by the user. The values of
the arguments INCRES, IFTEST, MINMUM and MAXMUM can be
specified by the user when the program is started or
whenever a semicolon is encountered in the text being
interpreted. If a zero is typed for the value of INCRES,
then the former values of these arguments are all retained
unchanged.
DIMENSION ISTORE(100),IBUFFR(72),JBUFFR(72)
DATA INCRES,IFTEST,MINMUM,MAXMUM,KIND,MAXBFR,MINSTR,
1ITTY,JTTY/1,2,0,0,0,72,2,5,5/
WRITE(JTTY,1)
1 FORMAT(15H SIZE OF ARRAY ,$)
READ(ITTY,4)MAXSTR
MAXSTR=MINSTR+MAXSTR-1
IF(MAXSTR.GT.99)MAXSTR=99
ISTORE(MINSTR-1)=-999
ISTORE(MAXSTR+1)=-999
LOWBFR=MAXBFR+1
2 WRITE(JTTY,3)
3 FORMAT(29H INCRES,IFTEST,MINMUM,MAXMUM ,$)
READ(ITTY,4)ISAVE,JSAVE,KSAVE,LSAVE
4 FORMAT(4I)
IF(ISAVE.LE.0)GO TO 11
INCRES=ISAVE
IFTEST=JSAVE
MINMUM=KSAVE
MAXMUM=LSAVE
GO TO 11
C OBTAIN NEW TEXT TO BE PROCESSED
5 WRITE(JTTY,6)
6 FORMAT(2H *,$)
GO TO 9
7 WRITE(JTTY,8)
8 FORMAT(2H &,$)
9 READ(ITTY,10)IBUFFR
10 FORMAT(72A1)
LOWBFR=1
C EVALUATE TEXT AND REPORT RESULTS
11 CALL DARANK(INCRES,IFTEST,MINMUM,MAXMUM,MINSTR,
1MAXSTR,IBUFFR,MAXBFR,LOWBFR,KIND ,MAXUSD,ISTORE)
IF(ISTORE(MINSTR-1).NE.-999)WRITE(JTTY,12)
IF(ISTORE(MAXSTR+1).NE.-999)WRITE(JTTY,12)
12 FORMAT(6H ERROR)
IF(MAXUSD.LT.MINSTR)GO TO 15
KOUNT=0
DO 13 INDEX=MINSTR,MAXUSD
13 CALL DANUMB(0,ISTORE(INDEX),10,JBUFFR,KOUNT,0,72)
WRITE(JTTY,14)(JBUFFR(I),I=1,KOUNT)
14 FORMAT(1X,72A1)
FASP, FORTRAN Alphameric Subroutine Package Page 339
DARANK, Returns Sorted Integers Excluding Duplicates
15 GO TO(5,2,7,16,18),KIND
16 WRITE(JTTY,17)IBUFFR(LOWBFR)
17 FORMAT(19H ILLEGAL CHARACTER ,1A1)
LOWBFR=LOWBFR+1
GO TO 11
18 WRITE(JTTY,19)
19 FORMAT(14H ILLEGAL RANGE)
GO TO 11
END
Typical Dialog Between User and DARANK Demonstration Program
------- ------ ------- ---- --- ------ ------------- -------
SIZE OF ARRAY 10
INCRES,IFTEST,MINMUM,MAXMUM 3 0 -10 10
*-1001/5/1000
-6 -1 4 9
*-1000/5/1000
-10 -5 0 5 10
*-999/5/1000
-9 -4 1 6
*-998/5/1000
-8 -3 2 7
*-997/5/1000
-7 -2 3 8
*-996/5/1000
-6 -1 4 9
*-995/5/1000
-10 -5 0 5 10
*1001/5/-1000
-9 -4 1 6
*1000/5/-1000
-10 -5 0 5 10
*999/5/-1000
-6 -1 4 9
*998/5/-1000
-7 -2 3 8
*997/5/-1000
-8 -3 2 7
*996/5/-10000
-9 -4 1 6
*995/5/-1000;-50/10/50&SEMICOLON INTERUPTS PROCESSING
-10 -5 0 5 10
INCRES,IFTEST,MINMUM,MAXMUM 3 2
-40 -30 -20 -10 0 10 20 30 40 50
&-37-20+10-30,,-33&CONTINUES GROUP AFTER SEMICOLON ABOVE
-33 -30 -20 -10 0 10 20 30 40 50
&45+55&
-20 -10 0 10 20 30 40 45 50 55
&35 1000*5 *6 -3*6!LATTER 2 RANGES DON'T SPECIFY ANYTHING
0 5 10 20 30 35 40 45 50 55
FASP, FORTRAN Alphameric Subroutine Package Page 340
DARITE, Free Format FORTRAN Output Routine
DDDDDD AAAA RRRRRRR IIIIII TTTTTTTTTT EEEEEEEEE
DD DD AA AA RR RR II TT EE
DD DD AA AA RR RR II TT EE
DD DD AA AA RRRRRRR II TT EEEEE
DD DD AAAAAAAA RR RR II TT EE
DD DD AA AA RR RR II TT EE
DDDDDD AA AA RR RR IIIIII TT EEEEEEEEE
DARITE, Free Format FORTRAN Output Routine
------ ---- ------ ------- ------ -------
DARITE represents a value so that it can be written with a
FORTRAN format statement containing a multiple A1 alphameric
specification. The value can be represented either with a
specified number of digits right of the decimal point and
with a maximum number of significant digits, or in
scientific notation as a number in the range 1.0 to 9.999...
with a following exponent. If floating point notation has
been requested, but the number cannot fit into the supplied
field with the specified number of digits right of the
decimal point, then the number of digits right of the
decimal point will be decreased, and if the number will
still not fit, then scientific notation will be used. If
the number will not fit into the field even in scientific
notation, then the field will be filled with asterisks.
The representation of the number is rounded based upon the
digit which would be to the right of the rightmost displayed
digit. If the digit to the right of the rightmost displayed
digit is 0 through 4, then the displayed digits are left
unchanged. If the digit to the right of the rightmost
displayed digit is 5 through 9, then the absolute value of
the displayed value is increased.
The DARITE Argument List
--- ------ -------- ----
The argument list of routine DARITE is
SUBROUTINE DARITE(VALUE ,JSTIFY,NOTATN,IPART ,ISIGN ,
1 KLIP ,IFILL ,IWIDTH,MINDEC,MAXDEC,MINSIG,MAXSIG,
2 INIZRO,MARGIN,IDECML,IEXPNT,IFORMT,IZERO ,LFTCOL,
3 MAXBFR,IBUFFR,KOUNT ,IERR )
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The following are input arguments left unchanged
VALUE = the number to be represented
FASP, FORTRAN Alphameric Subroutine Package Page 341
DARITE, Free Format FORTRAN Output Routine
JSTIFY = -1, left justify the representation of the value in
a field of width IWIDTH.
= 0, center the representation of the value in a
field of width IWIDTH.
= 1, right justify the representation of the value in
a field of width IWIDTH.
NOTATN = -1, represent in array IBUFFR the value in
scientific notation. The value 1234 would be
represented in IBUFFR as 1.234E3
= 0, represent in IBUFFR the value in floating point
form. The value 1234.56 would be displayed as it
is written if the number of digits requested right
of the decimal point is 2 and if at least 6
significant digits in a field of at least 7
characters are allowed.
= 1, multiply value by 100 and insert the percent
sign following digits of number. IWIDTH must
include room for this percent sign. MINDEC and
MAXDEC refer to the displayed decimal point. To
print tenths of a percent, MAXDEC would be given
the value 1, and MINDEC would be 1 or less.
= 2, if the value is in range 1000 to 1000000, divide
the value by 1000 and insert the letter K following
the digits. If the value is 1000000 or greater,
divide the value by 1000000 and insert the letter M
following the digits. IWIDTH must include room for
the letter K or M. MINDEC and MAXDEC refer to the
displayed decimal point. MINDEC and MAXDEC both
set at 2 would represent the value 1234 as 1.23K.
= 3, same as NOTATN=2 except that MINDEC and MAXDEC
refer to the decimal point in the original value,
not to the decimal point in the displayed number.
If a K or M is displayed right of the number, then
MINDEC and MAXDEC values of -1 are equivalent to
values of 0. MINDEC=-2 indicates that there is no
lower limit to the number of digits to the right of
the displayed decimal point, and is probably the
appropriate value unless it is absolutely necessary
to display the digits which would be to the right
of the decimal point in the original value, as for
example if the amounts are dollars and must be
displayed always including the cents digits. The
value 1234.56 would be represented as 1.23456K for
NOTATN=3 and MAXDEC=2.
IPART = -1, if the value is represented in floating point
form, completely represent this value. If the
value is represented in scientific notation,
represent only the normalized portion of the value
without the exponent, it being this normalized
value (in the range of 1.0 to 9.999...) which is
left justified, centered or right justified
according to the value of JSTIFY. IWIDTH must,
FASP, FORTRAN Alphameric Subroutine Package Page 342
DARITE, Free Format FORTRAN Output Routine
however, contain sufficient room for either the
normalized value or the exponent, whichever
requires the most characters to represent.
= 0, completely represent the value regardless of
whether in floating point or scientific notation.
= 1, if the value is represented in floating point
form, insert nothing (except the possible trailing
spaces indicated by IFILL) into IBUFFR. If the
value is represented in scientific notation,
represent only the exponent portion, it being this
exponent portion which is left justified, centered
or right justified according to the value of
JSTIFY. IWIDTH must, however, contain sufficient
room for either the normalized value or the
exponent, whichever requires the most characters to
represent.
ISIGN = -1, if the value is greater than zero, and if IPART
is less than or equal to zero, then place a plus
sign to the left of the representation of the
value. If the value is equal to zero, do not add
an extra space where the plus or minus sign would
otherwise be.
= 0, if the value is greater than or equal to zero,
do not add an extra space where the minus sign
would otherwise be.
= 1, if the value is greater than or equal to zero,
and if IPART is less than or equal to zero, then a
space will be placed at the left end of the
representation of the value, where the minus sign
would otherwise be, even if the number of digits in
the representation must be reduced to make room for
this space.
= 2, if the value is greater than zero, and if IPART
is less than or equal to zero, then place a plus
sign to the left of the representation of the
value. If the value is equal to zero, and if IPART
is less than or equal to zero, then place at least
1 space at the left end of the representation of
the value even if the number of digits in the
representation must be reduced to make room for
this space.
KLIP = -1, suppress representation of all zeroes which are
beyond the first digit after the decimal point and
which do not have a non-zero digit to their right.
If centering or right justifying, it is the number
after removal of the zeroes which is centered or
right justified.
= 0, represent as zeroes all zeroes which are to the
right of the decimal point.
= 1, convert to spaces all zeroes which are beyond
the first digit after the decimal point and which
do not have a non-zero digit to their right.
FASP, FORTRAN Alphameric Subroutine Package Page 343
DARITE, Free Format FORTRAN Output Routine
Remaining characters will be in the same positions
as if the zeroes where nonzeroes. If NOTATN is
greater than zero, then spaces can appear between
the nonspace characters of the number and the
following percent sign or K or M since the location
of these printing characters is not changed by the
conversion of the trailing zeroes to spaces.
= -2, same as KLIP=-1, except that if no non-zero
digit would be represented to the right of the
decimal point, then the zero which would be
immediately to the right of the decimal point is
not represented. If centering or right justifying,
it is the number after removal of these zeroes
which is centered or right justified.
= -3, same as KLIP=-2, except that if no non-zero
digit would be represented to the right of the
decimal point, then the decimal point is not
represented either. If centering or right
justifying, it is the number after removal of the
decimal point and the zeroes to its right which is
centered or right justified.
= 2, same as KLIP=1, except that if no non-zero digit
would be represented to the right of the right of
the decimal point, then the zero which would be
immediately to the right of the decimal point is
also converted to a space. Remaining characters
will be in the same positions as if the zeroes were
non-zeroes.
= 3, same as KLIP=2, except that if no non-zero digit
would be represented to the right of the decimal
point, then the decimal point is also converted to
a space. Characters which would be to the left of
the decimal point will be in the same positions as
if a non-zero digit appeared to the right of the
decimal point.
For example, the representations of the value 12.34
in floating point with MAXDEC=4 and in scientific
notation with IDECML=5 would be
for KLIP.LT.0 12.34 1.23E1
KLIP=0 12.3400 1.23400E1
KLIP=.GT.0 12.34 1.234 E1
The value 123000 would be represented as shown
below if NOTATN=2
for KLIP=-3 123K 1.23E5
KLIP=-2 123.K 1.23E5
KLIP=-1 123.0K 1.23E5
KLIP=0 123.0000K 1.23000E5
KLIP=1 123.0 K 1.23 E5
KLIP=2 123. K 1.23 E5
KLIP=3 123 K 1.23 E5
FASP, FORTRAN Alphameric Subroutine Package Page 344
DARITE, Free Format FORTRAN Output Routine
IFILL = 0, do not fill the portion of the field right of
the rightmost nonspace character in the
representation of the value with spaces. KOUNT
will be left pointing at the rightmost nonspace
character in the representation of the value. The
value of IFILL has no effect on the nonspace
characters in the representation of the value. If
IFILL=0, and if IPART is less than or equal to
zero, then the characters originally in the IBUFFR
array to the right of the nonspace portion of the
representation of the value are left intact. (The
reason why IPART when greater than zero is an
exception is that the normalized portion of the
number must be temporarily stored in IBUFFR even if
it is not going to be represented since possible
exponent change due to rounding must be checked
for.)
= 1, fill the field right of the nonspace portion of
the representation of the value with spaces. KOUNT
will be left pointing at LFTCOL+IWIDTH.
IWIDTH = the number of characters to be in the field in
which the value is represented. IWIDTH must
include room for exponent, sign and decimal point
if these are necessary to represent the value.
The following arguments control the format of numbers
displayed in floating point form (without E exponent), or
displayed with following K, M or percent sign. These
arguments do not control the format of numbers represented
in scientific notation (either with displayed exponent or
with suppressed zero exponent).
MINDEC = minimum number of digits right of the displayed
decimal point in numbers which are displayed in
floating point form. If it would be necessary to
represent less digits right of the decimal point
than indicated by MINDEC, then the value will be
represented in scientific notation instead. The
maximum number of digits right of the decimal point
is determined by the combination of MAXDEC and
MAXSIG.
= -2, there is no lower limit to the number of digits
which must be represented right of the decimal
point, and even the decimal point itself need not
be represented. MINDEC=-1 is equivalent to
MINDEC=-2 unless NOTATN=3 while VALUE is 1000 or
greater.
= -1, if NOTATN is not 3, or if NOTATN is 3 but the
value being represented is less than 1000, then
MINDEC=-1 is equivalent to MINDEC=-2 such that
there is no lower limit to the number of digits
which must be represented right of the decimal
point, and even the decimal point itself need not
FASP, FORTRAN Alphameric Subroutine Package Page 345
DARITE, Free Format FORTRAN Output Routine
be represented.
If NOTATN=3 and VALUE=1000 or greater, then
MINDEC=-1 would have the same effect as MINDEC=0
such that all digits to the left of the decimal
point in the original unshifted value (as
represented without K or M) must be included in the
actual representation of the value even when the K
or M is included. MINDEC should have the value -2
if, in order to fit the representation into the
available field width while NOTATN=3, it is to be
possible that a value which is equal to or greater
than 1000 be represented without all of the digits
which would be to the left of the decimal point in
the original value, or even without the decimal
point itself.
= equal or greater than zero, if NOTATN is not 3, or
if NOTATN is 3 but the value being represented is
less than 1000, then MINDEC is the minimum number
of digits which can be displayed right of the
decimal point in a floating point number. If less
than MINDEC digits would be displayed right of the
decimal point, then the value will be represented
in scientific notation instead. If MINDEC=0, then
it is not necessary that any digits be displayed
right of the decimal point, but the decimal point
itself must be displayed.
If NOTATN=3 and VALUE=1000 or greater, then MINDEC
is the minimum number of digits to the right of the
decimal point in the original value which must be
displayed. If VALUE is between 1000 and 1000000,
then at least MINDEC+3 digits must be displayed
right of the decimal point. If VALUE is 1000000 or
greater, then at least MINDEC+6 digits must be
displayed right of the decimal point.
For example, the value 1234.56 would be represented
as follows for various values of MINDEC and NOTATN.
It should be noted that when the digit 5 right of
the decimal point in the original value is not
represented, then the digit 4 left of the decimal
point is rounded upwards to 5.
NOTATN=0 NOTATN=2 NOTATN=3
for MINDEC=MAXDEC=-1 1235 1K 1.235K
MINDEC=MAXDEC=0 1235. 1.K 1.235K
MINDEC=MAXDEC=1 1234.6 1.2K 1.2346K
MINDEC=MAXDEC=2 1234.56 1.23K 1.23456K
MAXDEC = maximum number of digits right of the displayed
decimal point in numbers displayed in floating
point form.
= -2, represent as many digits right of the decimal
FASP, FORTRAN Alphameric Subroutine Package Page 346
DARITE, Free Format FORTRAN Output Routine
point as the field will hold (up to a maximum total
number of digits specified by MAXSIG).
= -1, if NOTATN is not 3, or if NOTATN is 3 but the
value being represented is less than 1000,
represent only digits left of the decimal point.
The decimal point itself will not be represented.
If more than IWIDTH or MAXSIG digits would appear
left of the decimal point, then the number will be
represented in scientific notation.
If NOTATN=3 and VALUE=1000 or greater, then
MAXDEC=-1 would have the same effect as MAXDEC=0
such that all digits which would be to the left of
the decimal point in the original value (as
represented without K or M) will if possible be
included in the representation of the value shifted
for the K or M notation. If MINDEC is greater than
-2, and if not all of the digits to the left of the
decimal point in the original unshifted value can
be included in the supplied field width, then the
value will instead be represented in scientific
notation.
= greater than or equal to zero, if NOTATN is not 3,
or if NOTATN is 3 but the value being represented
is less than 1000, represent the value with MAXDEC
digits right of decimal point. If this
representation of the value will not fit into the
field size indicated by IWIDTH or MAXSIG, then
reduce the number of digits represented right of
the decimal point so that the representation will
fit, or represent the value in scientific notation
if there are more than IWIDTH or MAXSIG digits left
of the decimal point.
If NOTATN=3 and VALUE=1000 or greater, then MAXDEC
is the number of digits to the right of the decimal
point in the original unshifted value which can be
displayed if there is room for these in the
supplied field width. If VALUE is between 1000 and
1000000, then at most MAXDEC+3 digits can be
displayed right of the displayed decimal point. If
VALUE is 1000000 or greater, then at most MAXDEC+6
digits can be displayed right of the decimal point.
MINSIG = minimum number of significant digits in the
floating point representation of the value. If the
floating point representation of the value would
contain less than MINSIG significant digits, then
the value will be represented in scientific
notation. If the value being displayed has the
value zero, then MINSIG is ignored.
MAXSIG = selects the maximum number of significant digits in
FASP, FORTRAN Alphameric Subroutine Package Page 347
DARITE, Free Format FORTRAN Output Routine
the floating point representation of the value.
= less than or equal to zero, allow as many digits as
the field will hold
= greater than zero, MAXSIG is the maximum number of
digits which can be displayed starting with the
leftmost nonzero digit, counting it and all digits
to its right. MAXSIG does not include the decimal
point, does not include the minus sign if the value
is negative, and does not include the percent sign,
K or M if NOTATN is greater than zero. The number
of digits displayed right of the decimal point is
reduced if necessary so that the number of digits
starting at the leftmost nonzero displayed digit
and counting it and all digits displayed to its
right does not exceed MAXSIG. If MAXSIG would be
less than the number of digits left of the decimal
point in the representation of the value, then the
value will be represented in scientific notation.
INIZRO = -1, if the number is represented in floating point
form and has an absolute value less than 1.0, then
a zero is displayed left of the decimal point if
and only if no digits would otherwise be displayed.
= 0, if the number is represented in floating point
form and has an absolute value less than 1.0, then
a zero is displayed to the left of the decimal
point.
= 1, if the number is represented in floating point
form and has an absolute value less than 1.0, then
a zero is not displayed to the left of the decimal
point. If no digits appear in the representation
of the number and IFILL equals one, then neither
decimal point nor sign are represented and the
field will be filled with spaces. If no digits
appear in the representation of the number and
IFILL is zero indicating that the field is not to
be filled with spaces right of the number, then
KOUNT will be returned equal to the input value of
LFTCOL.
For example, values greater than or equal to zero
but less than 0.005 would be represented as shown
below if MAXDEC equals 2
KLIP=0 KLIP=1 KLIP=2 KLIP=3
INIZRO=-1 .00 .0 0. 0
= 0 0.00 0.0 0. 0
= 1 .00 .0
Likewise, values between -0.005 and zero would be
represented as follow if MAXDEC equals 2
FASP, FORTRAN Alphameric Subroutine Package Page 348
DARITE, Free Format FORTRAN Output Routine
KLIP=0 KLIP=1 KLIP=2 KLIP=3
INIZRO=-1 -.00 -.0 -0. -0
= 0 -0.00 -0.0 -0. -0
= 1 -.00 -.0
MARGIN = 0, if the number is represented in floating point
form, and if a K, M or percent sign does not follow
the representation of the value, then do not add an
extra space where the K, M or percent sign would
otherwise be.
= greater than zero, MARGIN is the number of
characters in the field containing spaces and/or
the K, M or percent sign to the right of a floating
point number. MARGIN can be used to force a
floating point number to have its rightmost digit
at the same position as the rightmost digit of the
normalized portion of a value represented in
scientific notation. For this purpose, MARGIN
would be 1 greater than IEXPNT. If MARGIN is
greater than 1 and a K, M or percent sign must be
placed to the right of the number, then the K, M or
percent sign is placed at the left of the field of
spaces where the E of a scientific notation number
would appear, and the field then contains MARGIN-1
spaces to the right of the K, M or percent sign.
MARGIN would have the value 1 if it is merely
desired to force a single space to the right of the
representation of the number if a K, M or percent
sign does not appear to its right.
The following arguments control the format of values
represented in scientific notation either with displayed
exponent or with suppressed zero exponent (IZERO less than
or equal to zero).
IDECML = same as MAXDEC except that IDECML applies only to
values displayed in scientific notation. Note that
if the value is being represented in scientific
notation, then a nonzero digit will be used left of
the decimal point unless the value is itself zero.
Therefore, if IDECML is greater than or equal to
zero, then the maximum number of significant digits
which can be displayed in scientific notation is
IDECML+1.
IEXPNT = minimum number of digits in the exponent if the
value is represented in scientific notation. If
fewer than IEXPNT digits are needed in the
exponent, these are justified in an exponent field
of IEXPNT width with either spaces or zeroes as
fill according to the value of IFORMT. IEXPNT must
include room for the sign (if the exponent is
negative) but does not include room for the initial
FASP, FORTRAN Alphameric Subroutine Package Page 349
DARITE, Free Format FORTRAN Output Routine
letter E of the exponent.
IFORMT = 0, if IEXPNT is greater than the number of
characters needed to represent the exponent, then
left justify the exponent within the exponent
field.
= 1, if IEXPNT is greater than the number of
characters needed to represent the exponent, then
right justify the digits of the exponent within the
exponent field. If the exponent is negative, place
the sign right of the E at the start of the
exponent. Zeroes, not spaces, are used to fill the
rest of the exponent field.
= 2, same as IFORMT=1 except that spaces, not zeroes
are used to fill between the sign if any and the
digits of the exponent.
= 3, same as IFORMT=2 except that the sign if any is
placed immediately to the left of the digits of the
exponent instead of to the right of the E at the
start of the exponent.
For example, if IEXPNT is 4, then the value 1.2E-3
would be represented
for IFORMT=0 1.2E-3
IFORMT=1 1.2E-003
IFORMT=2 1.2E- 3
IFORMT=3 1.2E -3
IZERO = -1, if the value is being represented in scientific
notation and has a zero exponent (value of the
number is zero or is in either range -9.99... to
-1.00... or range 1.00... to 9.99...), then the
representation of the number will not include an
exponent field.
= 0, if the value is being represented in scientific
notation and has a zero exponent (value of the
number is zero or is in either range -9.99... to
-1.00... or range 1.00... to 9.99...), then the
exponent field is of the same size as if the
exponent was one but the exponent field is filled
with spaces.
= 1, if the value is being represented in scientific
notation, then the exponent will be displayed even
if this exponent is zero.
For example, if IEXPNT is 4, then the value 1.2
would be represented
for IZERO=-1 1.2
IZERO=0 1.2
IZERO=1 1.2E 0
The following arguments, together with IWIDTH described
FASP, FORTRAN Alphameric Subroutine Package Page 350
DARITE, Free Format FORTRAN Output Routine
earlier, define the usable portion of the IBUFFR array, and
are returned unchanged.
LFTCOL = the subscript of the IBUFFR array entry to the
immediate left of the field in which the number is
to be represented. Although LFTCOL will usually be
equal to the number of characters already in the
IBUFFR array, differently named arguments should be
used for LFTCOL and KOUNT since, if the compiler
transfers these by value rather than by address,
then the return from this routine could copy out
the new value of KOUNT before copying out the
unchanged value of LFTCOL.
MAXBFR = the maximum dimension of the IBUFFR array. If
LFTCOL+IWIDTH is greater than MAXBFR, then the
effective value of IWIDTH is reduced to
MAXBFR-LFTCOL.
The following are output arguments
IBUFFR = array in which the number is to be represented and
which can then be printed by the calling program
using a multiple of an A1 format.
KOUNT = returned pointing to the rightmost character
inserted into the IBUFFR array. This will always
equal LFTCOL+IWIDTH if IFILL=1.
IERR = -1 returned if the field was filled with asterisks
due to field overflow even in scientific notation.
= 0 returned if the value was represented in floating
point form with or without following percent sign,
K or M.
= 1 or greater returned if value was represented in
scientific notation. IERR is the number of digits
which are shown left of the exponent prior to the
suppression of rightmost zeroes which might be
requested by nonzero values of KLIP. If IPART is 1
causing only the exponent to be shown, then IERR is
the number of digits which would shown if IPART
instead had the value -1 (again prior to the
suppression of rightmost zeroes). If this routine
is being called to append the representation of a
number to a line of text which already contains
other information, then the calling program can
test the returned value of IERR to determine
whether there was enough room at the end of the
line for the number to be represented with
sufficient accuracy.
FASP, FORTRAN Alphameric Subroutine Package Page 351
DARITE, Free Format FORTRAN Output Routine
An Example of the Use of DARITE
-- ------- -- --- --- -- ------
The following sample program demonstrates the manner in
which DARITE is called.
DIMENSION IBUFFR(12),VALUE(2)
DATA VALUE/1.2345678E4,1.2345678E-6/
DATA IBUFFR(1)/1H(/
DO 4 ITEM=1,2
DO 3 MAXSIG=1,8
C VALUE=VALUE(ITEM) IWIDTH=10 IDECML=MAXSIG-1
C JSTIFY=1 MINDEC=-2 IEXPNT=0
C NOTATN=0 MAXDEC=-2 IFORMT=0
C IPART =0 MINSIG=MAXSIG IZERO =1
C ISIGN =0 MAXSIG LFTCOL=1
C KLIP =0 INIZRO=0 MAXBFR=12
C IFILL =1 MARGIN=0 IBUFFR
CALL DARITE(VALUE(ITEM),1,0,0,0,
10,1,10,-2,-2,MAXSIG,MAXSIG,0,0,
2MAXSIG-1,0,0,1,1,12,IBUFFR,KOUNT,IERR)
KOUNT=KOUNT+1
IBUFFR(KOUNT)=1H)
IF(MAXSIG.EQ.1)WRITE(1,1)MAXSIG,
1(IBUFFR(I),I=1,KOUNT),VALUE(ITEM)
1 FORMAT(1X/1X,'MAXSIG=',1I1,1X,12A1,
1' VALUE =',1E16.8)
IF(MAXSIG.GT.1)WRITE(1,2)MAXSIG,
1(IBUFFR(I),I=1,KOUNT)
2 FORMAT(1X,'MAXSIG=',1I1,1X,12A1)
3 CONTINUE
4 CONTINUE
END
The output generated by the above program is shown below
MAXSIG=1 ( 1.E4) VALUE = 0.12345678E+05
MAXSIG=2 ( 1.2E4)
MAXSIG=3 ( 1.23E4)
MAXSIG=4 ( 1.235E4)
MAXSIG=5 ( 12346.)
MAXSIG=6 ( 12345.7)
MAXSIG=7 ( 12345.68)
MAXSIG=8 ( 12345.678)
MAXSIG=1 ( 0.000001) VALUE = 0.12345678E-05
MAXSIG=2 ( 0.0000012)
MAXSIG=3 (0.00000123)
MAXSIG=4 ( 1.235E-6)
MAXSIG=5 ( 1.2346E-6)
MAXSIG=6 (1.23457E-6)
MAXSIG=7 (1.23457E-6)
MAXSIG=8 (1.23457E-6)
FASP, FORTRAN Alphameric Subroutine Package Page 352
DARITE, Free Format FORTRAN Output Routine
Demonstration Program to Interactively Test DARITE Arguments
------------- ------- -- ------------- ---- ------ ---------
The user of the program listed on the following pages
defines by name and value each argument which is to be
tested. When an empty line is read, the program will report
enclosed in parentheses the results of calling DARITE with
the specified argument values, for all possible values of
JSTIFY and of IFILL. A sample dialog between the program
and the user is presented following the listing of the
program.
C PROGRAM TO DEMONSTRATE DARITE ROUTINE
C
C DEFINE DICTIONARY CONSISTING OF ARGUMENT NAMES
DIMENSION ITYPE(23),KNTLTR(23),IWORD(128)
DATA MAXWRD,MAXKNT,MAXINP,MAXOUT/128,23,60,132/
DATA IWORD/ 1HV,1HA,1HL,1HU,1HE,
1 1HJ,1HS,1HT,1HI,1HF,1HY, 1HN,1HO,1HT,1HA,1HT,1HN,
2 1HI,1HP,1HA,1HR,1HT, 1HI,1HS,1HI,1HG,1HN,
3 1HK,1HL,1HI,1HP, 1HI,1HF,1HI,1HL,1HL,
4 1HI,1HW,1HI,1HD,1HT,1HH, 1HM,1HI,1HN,1HD,1HE,1HC,
5 1HM,1HA,1HX,1HD,1HE,1HC, 1HM,1HI,1HN,1HS,1HI,1HG,
6 1HM,1HA,1HX,1HS,1HI,1HG, 1HI,1HN,1HI,1HZ,1HR,1HO,
7 1HM,1HA,1HR,1HG,1HI,1HN, 1HI,1HD,1HE,1HC,1HM,1HL,
8 1HI,1HE,1HX,1HP,1HN,1HT, 1HI,1HF,1HO,1HR,1HM,1HT,
9 1HI,1HZ,1HE,1HR,1HO, 1HL,1HF,1HT,1HC,1HO,1HL,
1 1HM,1HA,1HX,1HB,1HF,1HR, 1HI,1HB,1HU,1HF,1HF,1HR,
2 1HK,1HO,1HU,1HN,1HT, 1HI,1HE,1HR,1HR/
DATA ITYPE/4,-1,0,0,0,0,-1,0,0,0,
10,0,0,0,0,0,0,0,-1,0,
2-1,-1,-1/
DATA KNTLTR/5,6,6,5,5,4,5,6,6,6,
16,6,6,6,6,6,6,5,6,6,
26,5,4/
C
C DEFINE INITIAL VALUES OF ARGUMENTS
DATA NOTATN, IPART, ISIGN, KLIP, IFILL,IWIDTH,MINDEC,
1 MAXDEC,MINSIG,MAXSIG,INIZRO,MARGIN,IDECML,IEXPNT,
2 IFORMT, IZERO,MAXBFR/5*0,10,10*0,99/
DATA VALUE/0.0/
C
C DEFINE NEEDED CHARACTERS
DATA IWHAT,IBLANK,ITAB/1H?,1H ,1H /
C
C INPUT AND OUTPUT BUFFERS
DIMENSION IBUFFR(132),JBUFFR(60)
C
C INPUT AND OUTPUT UNIT NUMBERS
DATA ITTY,JTTY/5,5/
C
WRITE(JTTY,1)
FASP, FORTRAN Alphameric Subroutine Package Page 353
DARITE, Free Format FORTRAN Output Routine
1 FORMAT(1X,37HPROGRAM TO DEMONSTRATE DARITE ROUTINE/
11X,47HTYPE ARGUMENT NAME AND VALUE, 1 OR MORE ON LINE/
21X,41HEMPTY LINE CALLS DARITE WITH THESE VALUES/
31X,43HSPACE IN BUFFER PRINTS AS <, UNDEFINED AS >)
C
C READ USER SPECIFICATION OF ARGUMENT VALUES
2 WRITE(JTTY,3)
3 FORMAT(1X,1H*,$)
READ(ITTY,4)JBUFFR
4 FORMAT(60A1)
C
C IDENTIFY THE ARGUMENT
LOWBFR=0
5 LOWBFR=LOWBFR+1
6 INIBFR=LOWBFR
CALL DAVERB(1,MAXWRD,IWORD,1,MAXKNT,
1KNTLTR,JBUFFR,MAXINP,LOWBFR,KIND,MATCH,LCNWRD,
2LCNKNT,LCNBFR)
GO TO(21,9,8,8,17),KIND
C
C OBTAIN VALUE OF ARGUMENT
7 LOWBFR=LOWBFR+1
8 KONTRL=ITYPE(MATCH)
IF(KONTRL.LT.0)KONTRL=1
CALL DAHEFT(KONTRL,1,0,JBUFFR,MAXINP,LOWBFR,KIND,
1ISHIFT,JSHIFT,KSHIFT,LSHIFT,IVALUE,AVALUE)
GO TO(2,10,12),KIND
C
C ALLOW COMMA OR EQUAL SIGN
9 IF(JBUFFR(LOWBFR).EQ.1H=)GO TO 13
GO TO 11
10 IF(JBUFFR(LOWBFR).EQ.1H=)GO TO 7
11 IF(JBUFFR(LOWBFR).EQ.1H,)GO TO 5
IF(JBUFFR(LOWBFR).EQ.1H;)GO TO 5
IF(JBUFFR(LOWBFR).EQ.1H!)GO TO 2
LCNBFR=LOWBFR
GO TO 18
C
C SET ARGUMENT IF VALUE FOUND
12 IF(ITYPE(MATCH).LT.0)GO TO 15
IF(MATCH.EQ.1)VALUE=AVALUE
IF(MATCH.EQ.2)JSTIFY=IVALUE
IF(MATCH.EQ.3)NOTATN=IVALUE
IF(MATCH.EQ.4)IPART=IVALUE
IF(MATCH.EQ.5)ISIGN=IVALUE
IF(MATCH.EQ.6)KLIP=IVALUE
IF(MATCH.EQ.7)IFILL=IVALUE
IF(MATCH.EQ.8)IWIDTH=IVALUE
IF(MATCH.EQ.9)MINDEC=IVALUE
IF(MATCH.EQ.10)MAXDEC=IVALUE
IF(MATCH.EQ.11)MINSIG=IVALUE
IF(MATCH.EQ.12)MAXSIG=IVALUE
IF(MATCH.EQ.13)INIZRO=IVALUE
IF(MATCH.EQ.14)MARGIN=IVALUE
FASP, FORTRAN Alphameric Subroutine Package Page 354
DARITE, Free Format FORTRAN Output Routine
IF(MATCH.EQ.15)IDECML=IVALUE
IF(MATCH.EQ.16)IEXPNT=IVALUE
IF(MATCH.EQ.17)IFORMT=IVALUE
IF(MATCH.EQ.18)IZERO=IVALUE
IF(MATCH.EQ.20)MAXBFR=IVALUE
GO TO 6
C
C ARGUMENT TYPED BY USER FOUND TO BE IN ERROR
13 WRITE(JTTY,14)
14 FORMAT(1X,22HARGUMENT NOT SPECIFIED)
GO TO 2
15 J=LCNWRD+KNTLTR(LCNKNT)-1
WRITE(JTTY,16)(IWORD(I),I=LCNWRD,J)
16 FORMAT(1X,16HUSER CANNOT SET ,100A1)
GO TO 2
17 LOWBFR=LOWBFR-1
18 IF(LOWBFR.GE.MAXINP)GO TO 19
IF(JBUFFR(LOWBFR+1).EQ.IBLANK)GO TO 19
IF(JBUFFR(LOWBFR+1).EQ.ITAB)GO TO 19
IF(JBUFFR(LOWBFR+1).EQ.1H,)GO TO 19
IF(JBUFFR(LOWBFR+1).EQ.1H=)GO TO 19
LOWBFR=LOWBFR+1
GO TO 18
19 WRITE(JTTY,20)(JBUFFR(I),I=LCNBFR,LOWBFR),IWHAT
20 FORMAT(1X,100A1)
GO TO 2
C
C TEST IF ANYTHING WAS TYPED WHEN END OF LINE FOUND
21 IF(INIBFR.GT.1)GO TO 2
C
C REPORT ARGUMENT VALUES TO USER
WRITE(JTTY,22)NOTATN, IPART, ISIGN, KLIP,IWIDTH,
1MINDEC,MAXDEC,MINSIG,MAXSIG,INIZRO,MARGIN,IDECML,
2IEXPNT,IFORMT, IZERO,MAXBFR, VALUE
22 FORMAT(7H NOTATN,1I3,7H IPART,1I3,7H ISIGN,1I3,
1 7H KLIP,1I3,7H IWIDTH,1I3,7H MINDEC,1I3/
2 7H MAXDEC,1I3,7H MINSIG,1I3,7H MAXSIG,1I3,
3 7H INIZRO,1I3,7H MARGIN,1I3,7H IDECML,1I3/
4 7H IEXPNT,1I3,7H IFORMT,1I3,7H IZERO,1I3,
5 7H MAXBFR,1I3,7H VALUE,1E13.5)
C
C LOOP FOR IFILL=0 AND 1 FOR JSTIFY=-1, 0 AND 1
DO 31 IJSTIF=1,3
JSTIFY=IJSTIF-2
LFTCOL=0
DO 29 IIFILL=1,2
IFILL=IIFILL-1
C
C INSERT VALUE OF IFILL INTO OUTPUT TEXT
IBUFFR(LFTCOL+1)=IBLANK
IBUFFR(LFTCOL+2)=1HF
IBUFFR(LFTCOL+3)=1HI
IBUFFR(LFTCOL+4)=1HL
LFTCOL=LFTCOL+4
FASP, FORTRAN Alphameric Subroutine Package Page 355
DARITE, Free Format FORTRAN Output Routine
AVALUE=IFILL
CALL DARITE(AVALUE,1,0,0,0,
10,0,2,-1,-1,0,2,
20,0,0,0,0,0,LFTCOL,
3MAXOUT,IBUFFR,LFTCOL,IERR)
C
C INSERT TEXT LEFT OF VALUE, LEAVING ROOM FOR IERR
LFTERR=LFTCOL+4
IBUFFR(LFTCOL+1)=IBLANK
IBUFFR(LFTCOL+2)=1HE
IBUFFR(LFTCOL+3)=1HR
IBUFFR(LFTCOL+4)=1HR
IBUFFR(LFTCOL+7)=IBLANK
IBUFFR(LFTCOL+8)=1H(
LFTCOL=LFTCOL+8
IBUFFR(LFTCOL+IWIDTH+1)=1H)
C
C CALL DARITE TO REPRESENT VALUE WITH USER SET FORMAT
LIMIT=LFTCOL+MAXBFR
CALL DARITE( VALUE,JSTIFY,NOTATN, IPART, ISIGN,
1 KLIP, IFILL,IWIDTH,MINDEC,MAXDEC,MINSIG,MAXSIG,
2INIZRO,MARGIN,IDECML,IEXPNT,IFORMT, IZERO,LFTCOL,
3 LIMIT,IBUFFR, KOUNT, IERR)
IF(IBUFFR(LFTCOL).NE.1H()TYPE 23
IF(IBUFFR(LFTCOL+IWIDTH+1).NE.1H))TYPE 24
23 FORMAT(1X,33HCHARACTER LEFT OF FIELD DESTROYED)
24 FORMAT(1X,34HCHARACTER RIGHT OF FIELD DESTROYED)
C
C INSERT VALUE OF IERR INTO TEXT LEFT OF VALUE
AVALUE=IERR
CALL DARITE(AVALUE,1,0,0,0,
10,0,2,-1,-1,0,2,
20,0,0,0,0,0,LFTERR,
3MAXOUT,IBUFFR,I,IERR)
C
C MARK SPACES AND EMPTY LOCATIONS RETURNED BY DARITE
I=LFTCOL+1
25 IF(I.GT.KOUNT)GO TO 26
IF(IBUFFR(I).EQ.IBLANK)IBUFFR(I)=1H<
I=I+1
GO TO 25
26 KOUNT=KOUNT+1
IF(KOUNT.GT.LFTCOL+IWIDTH)GO TO 27
IBUFFR(KOUNT)=1H>
GO TO 26
27 IF(KOUNT.GE.LFTCOL+IWIDTH+IEXTRA)GO TO 28
KOUNT=KOUNT+1
IBUFFR(KOUNT)=IBLANK
GO TO 27
28 LFTCOL=KOUNT
29 CONTINUE
C
C DISPLAY THE RESULT TO USER
WRITE(JTTY,30)JSTIFY,(IBUFFR(I),I=1,KOUNT)
FASP, FORTRAN Alphameric Subroutine Package Page 356
DARITE, Free Format FORTRAN Output Routine
30 FORMAT(1X,3HJST,I2,120A1)
31 CONTINUE
C
C GO BACK TO ASK USER FOR NEXT ARGUMENT TO CHANGE
GO TO 2
END
Typical Dialog Between DARITE Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
PROGRAM TO DEMONSTRATE DARITE ROUTINE
TYPE ARGUMENT NAME AND VALUE, 1 OR MORE ON LINE
EMPTY LINE CALLS DARITE WITH THESE VALUES
SPACE IN BUFFER PRINTS AS <, UNDEFINED AS >
*MAXD -2 MAXS 6 VALUE 123.45678
*
NOTATN 0 IPART 0 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 0
MAXDEC -2 MINSIG 0 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 0
IEXPNT 0 IFORMT 0 IZERO 0 MAXBFR 99 VALUE 0.12346E+03
JST-1 FIL 0 ERR 0 (123.457>>>) FIL 1 ERR 0 (123.457<<<)
JST 0 FIL 0 ERR 0 (<123.457>>) FIL 1 ERR 0 (<123.457<<)
JST 1 FIL 0 ERR 0 (<<<123.457) FIL 1 ERR 0 (<<<123.457)
*MINS 6 MIND 4 !IMPOSSIBLE CONDITIONS TO FORCE E NOTATION
*
NOTATN 0 IPART 0 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 0
IEXPNT 0 IFORMT 0 IZERO 0 MAXBFR 99 VALUE 0.12346E+03
JST-1 FIL 0 ERR 1 (1.E2>>>>>>) FIL 1 ERR 1 (1.E2<<<<<<)
JST 0 FIL 0 ERR 1 (<<<1.E2>>>) FIL 1 ERR 1 (<<<1.E2<<<)
JST 1 FIL 0 ERR 1 (<<<<<<1.E2) FIL 1 ERR 1 (<<<<<<1.E2)
*IDEC 5 IEXP 3 IFOR 1 !MAKE DISPLAY LOOK BETTER
*
NOTATN 0 IPART 0 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 5
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 99 VALUE 0.12346E+03
JST-1 FIL 0 ERR 5 (1.2346E002) FIL 1 ERR 5 (1.2346E002)
JST 0 FIL 0 ERR 5 (1.2346E002) FIL 1 ERR 5 (1.2346E002)
JST 1 FIL 0 ERR 5 (1.2346E002) FIL 1 ERR 5 (1.2346E002)
*VAL 9.085E-7 !TRY A NEGATIVE EXPONENT
*
NOTATN 0 IPART 0 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 5
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 99 VALUE 0.90850E-06
JST-1 FIL 0 ERR 5 (9.0850E-07) FIL 1 ERR 5 (9.0850E-07)
JST 0 FIL 0 ERR 5 (9.0850E-07) FIL 1 ERR 5 (9.0850E-07)
JST 1 FIL 0 ERR 5 (9.0850E-07) FIL 1 ERR 5 (9.0850E-07)
FASP, FORTRAN Alphameric Subroutine Package Page 357
DARITE, Free Format FORTRAN Output Routine
*IPAR -1 !LOOK AT NORMALIZED PART ONLY
*
NOTATN 0 IPART -1 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 5
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 99 VALUE 0.90850E-06
JST-1 FIL 0 ERR 6 (9.08500>>>) FIL 1 ERR 6 (9.08500<<<)
JST 0 FIL 0 ERR 6 (<9.08500>>) FIL 1 ERR 6 (<9.08500<<)
JST 1 FIL 0 ERR 6 (<<<9.08500) FIL 1 ERR 6 (<<<9.08500)
*IPAR 1 !LOOK AT EXPONENT
*
NOTATN 0 IPART 1 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 5
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 99 VALUE 0.90850E-06
JST-1 FIL 0 ERR 6 (E-07>>>>>>) FIL 1 ERR 6 (E-07<<<<<<)
JST 0 FIL 0 ERR 6 (<<<E-07>>>) FIL 1 ERR 6 (<<<E-07<<<)
JST 1 FIL 0 ERR 6 (<<<<<<E-07) FIL 1 ERR 6 (<<<<<<E-07)
*VAL 9.99996E-7 IPA 0 !TEST ROUNDING
*
NOTATN 0 IPART 0 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 5
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 99 VALUE 0.10000E-05
JST-1 FIL 0 ERR 5 (1.0000E-06) FIL 1 ERR 5 (1.0000E-06)
JST 0 FIL 0 ERR 5 (1.0000E-06) FIL 1 ERR 5 (1.0000E-06)
JST 1 FIL 0 ERR 5 (1.0000E-06) FIL 1 ERR 5 (1.0000E-06)
*IPA -1 !LOOK AT NORMALIZED PART
*
NOTATN 0 IPART -1 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 5
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 99 VALUE 0.10000E-05
JST-1 FIL 0 ERR 6 (9.99996>>>) FIL 1 ERR 6 (9.99996<<<)
JST 0 FIL 0 ERR 6 (<9.99996>>) FIL 1 ERR 6 (<9.99996<<)
JST 1 FIL 0 ERR 6 (<<<9.99996) FIL 1 ERR 6 (<<<9.99996)
*IPA 1 !LOOK AT EXPONENT
*
NOTATN 0 IPART 1 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 5
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 99 VALUE 0.10000E-05
JST-1 FIL 0 ERR 6 (E-07>>>>>>) FIL 1 ERR 6 (E-07<<<<<<)
JST 0 FIL 0 ERR 6 (<<<E-07>>>) FIL 1 ERR 6 (<<<E-07<<<)
JST 1 FIL 0 ERR 6 (<<<<<<E-07) FIL 1 ERR 6 (<<<<<<E-07)
*IDEC 4 IPA -1 !REDUCE DIGITS TO SAME FOR ALL OR PART
*
NOTATN 0 IPART -1 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 4
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 99 VALUE 0.10000E-05
JST-1 FIL 0 ERR 5 (1.0000>>>>) FIL 1 ERR 5 (1.0000<<<<)
JST 0 FIL 0 ERR 5 (<<1.0000>>) FIL 1 ERR 5 (<<1.0000<<)
JST 1 FIL 0 ERR 5 (<<<<1.0000) FIL 1 ERR 5 (<<<<1.0000)
FASP, FORTRAN Alphameric Subroutine Package Page 358
DARITE, Free Format FORTRAN Output Routine
*IPA 1 !MAKE SURE THAT EXPONENT SHOWS RESULTS OF ROUNDING
*
NOTATN 0 IPART 1 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 4
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 99 VALUE 0.10000E-05
JST-1 FIL 0 ERR 5 (E-06>>>>>>) FIL 1 ERR 5 (E-06<<<<<<)
JST 0 FIL 0 ERR 5 (<<<E-06>>>) FIL 1 ERR 5 (<<<E-06<<<)
JST 1 FIL 0 ERR 5 (<<<<<<E-06) FIL 1 ERR 5 (<<<<<<E-06)
*MAXB 5 IPA 0 !REDUCE BUFFER LENGTH TO MIDDLE OF FIELD
*
NOTATN 0 IPART 0 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 4
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 5 VALUE 0.10000E-05
JST-1 FIL 0 ERR 1 (1E-06>>>>>) FIL 1 ERR 1 (1E-06>>>>>)
JST 0 FIL 0 ERR 1 (1E-06>>>>>) FIL 1 ERR 1 (1E-06>>>>>)
JST 1 FIL 0 ERR 1 (1E-06>>>>>) FIL 1 ERR 1 (1E-06>>>>>)
*IPA -1
*
NOTATN 0 IPART -1 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 4
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 5 VALUE 0.10000E-05
JST-1 FIL 0 ERR 4 (1.000>>>>>) FIL 1 ERR 4 (1.000>>>>>)
JST 0 FIL 0 ERR 4 (1.000>>>>>) FIL 1 ERR 4 (1.000>>>>>)
JST 1 FIL 0 ERR 4 (1.000>>>>>) FIL 1 ERR 4 (1.000>>>>>)
*IPA 1
*
NOTATN 0 IPART 1 ISIGN 0 KLIP 0 IWIDTH 10 MINDEC 4
MAXDEC -2 MINSIG 6 MAXSIG 6 INIZRO 0 MARGIN 0 IDECML 4
IEXPNT 3 IFORMT 1 IZERO 0 MAXBFR 5 VALUE 0.10000E-05
JST-1 FIL 0 ERR 4 (E-06>>>>>>) FIL 1 ERR 4 (E-06<>>>>>)
JST 0 FIL 0 ERR 4 (E-06>>>>>>) FIL 1 ERR 4 (E-06<>>>>>)
JST 1 FIL 0 ERR 4 (<E-06>>>>>) FIL 1 ERR 4 (<E-06>>>>>)
FASP, FORTRAN Alphameric Subroutine Package Page 359
DAROME, Roman Numeral Generator
DDDDDDD AAA RRRRRRRRR OOOOOOOOO MMM MMM EEEEEEEE
DDD DDD AAA AAA RRR RRR OOO OOO MMMM MMMM EEE
DDD DDD AAA AAA RRRRRRR OOO OOO MMM M M MMM EEEEEE
DDD DDD AAAAAAAAA RRR RRR OOO OOO MMM M MMM EEE
DDDDDDD AAA AAA RRR RRR OOOOOOOOO MMM MMM EEEEEEEE
DAROME, Roman Numeral Generator
------ ----- ------- ---------
DAROME represents an integer value in Roman numeral notation
so that it can be written with a FORTRAN format statement
containing a multiple of an A1 alphameric format
specification. The argument list of DAROME is identical to
that of DANUMB except for the DANUMB argument IRADIX which
is not applicable to the construction of Roman numerals.
Values over 3999 cannot be represented in Roman numeral
notation since the the largest number having a letter
equivalent is 1000 represented by M. The routine could
represent larger numbers if letter assignments for larger
powers of ten and for halves of powers of ten are decided
upon and incorporated into the array which specifies these
assignments in DAROME. Values of zero or less cannot be
represented by DAROME since Roman numeral notation cannot be
used to represent these values.
DAROME has no output argument used to signal an error
condition. Instead, the character count KOUNT is returned
unchanged if the representation of the number will not fit
into the buffer provided or if the value is not within the
range which can be represented.
FASP, FORTRAN Alphameric Subroutine Package Page 360
DAROME, Roman Numeral Generator
The DAROME Argument List
--- ------ -------- ----
The argument list of routine DAROME is
SUBROUTINE DAROME(KONTRL,NUMBER,LETTER,KOUNT,LFTCOL,
1MAX)
with the associated DIMENSION statement
DIMENSION LETTER(MAX)
The following are input arguments left unchanged
KONTRL = 0, left justify the representation of the number.
If KOUNT is less than LFTCOL, then the output
buffer starting at LETTER(KOUNT+1) and extending
through LETTER(LFTCOL) will be filled with spaces.
If KOUNT is less than or equal to LFTCOL, then the
leftmost character of the Roman numeral
representation of the number will be placed in
LETTER(LFTCOL+1).
If KOUNT is greater than LFTCOL, meaning that
characters are already in use to the right of
LFTCOL, then a space will be placed in
LETTER(KOUNT+1), and the leftmost letter of the
representation of the number will be placed in
LETTER(KOUNT+2).
If the representation of the number will fit into
the output buffer, then the character count KOUNT
will be left pointing to the location of the
rightmost letter of the representation of the
number. If the representation of the number would
extend to the right of LETTER(MAX), then the
character count KOUNT is returned unchanged.
= 1, right justify the representation of the number
so that the rightmost letter of the representation
of the number is placed in LETTER(LFTCOL).
If the representation of the number requires less
than LFTCOL-KOUNT characters, then fill the output
buffer with spaces starting at LETTER(KOUNT+1) and
extending up to the leftmost character of the
representation of the number.
If the representation of the number will fit into
the output buffer, then the character count KOUNT
will be returned equal to LFTCOL. If LFTCOL is
greater than the dimension of the buffer, or if the
FASP, FORTRAN Alphameric Subroutine Package Page 361
DAROME, Roman Numeral Generator
number is too large to be represented starting at
LETTER(KOUNT+1) and ending at LETTER(LFTCOL), then
the character count KOUNT is returned unchanged.
NUMBER = the integer value to be represented.
The following argument is an array to which the
representation of the value is appended.
LETTER = the output buffer into which DAROME will place the
representation of the number and which can be
written by a multiple of an A1 format by the
calling program.
The following argument is used for both input to, and output
from this routine.
KOUNT = is input containing the number of characters
already in use in the output buffer LETTER array.
If the representation of the value will fit into
the letter array, then KOUNT will be returned
containing the location within the LETTER array of
the rightmost character of the representation of
the value. KOUNT is returned unchanged either if
the representation of the value will not fit into
the LETTER array or if the value cannot be
represented using Roman numeral notation.
The following are input arguments which describe the size of
the output text buffer. These arguments are returned
unchanged.
LFTCOL = number of characters in the buffer which are to be
to the left of the representation of the number if
KONTRL equals 0.
= position in the buffer at which is to be placed the
right letter of the representation of the number if
KONTRL equals 1.
MAX = the dimension of the output buffer LETTER array.
The character count KOUNT will be returned
unchanged if the representation of the value would
extend beyond LETTER(MAX).
FASP, FORTRAN Alphameric Subroutine Package Page 362
DAROME, Roman Numeral Generator
Program to Demonstrate the Use of DAROME
------- -- ----------- --- --- -- ------
The program listed below accepts a value from the user,
twice generates the left justified Roman numeral
representation of the number, with the first starting in
column 6, then generates the right justified representation
ending in column 59. The asterisks originally in the buffer
are retained in column 1 by setting KOUNT=1 before
generating the first representation, to the right of the
second representation by incrementing KOUNT by 1, and in
column 60 by setting LFTCOL=59 for the right justification.
A sample dialog between the program and the user is
presented below the listing of the program.
DIMENSION LETTER(60)
DATA ITTY,JTTY,MAX/5,5,60/
C
C REQUEST AND READ NUMBER TO BE CONVERTED
WRITE(JTTY,1)
1 FORMAT(1X,37HPROGRAM TO DEMONSTRATE DAROME ROUTINE)
2 WRITE(JTTY,3)
3 FORMAT(7H VALUE ,$)
READ(ITTY,4)NUMBER
4 FORMAT(I)
DO 5 I=1,MAX
5 LETTER(I)=1H*
C
C CONVERT NUMBER TO ROMAN NUMERAL FORM
KOUNT=1
LFTCOL=5
KONTRL=0
IPASS=-2
6 CALL DAROME(KONTRL,NUMBER,LETTER,KOUNT,LFTCOL,
1MAX)
IPASS=IPASS+1
IF(IPASS.LT.0)GO TO 6
IF(IPASS.GT.0)GO TO 7
KONTRL=1
KOUNT=KOUNT+1
LFTCOL=MAX-1
GO TO 6
C
C REPORT RESULTS TO USER
7 WRITE(JTTY,8)LETTER
8 FORMAT(1X,60A1)
GO TO 2
END
PROGRAM TO DEMONSTRATE DAROME ROUTINE
VALUE 3846
* MMMDCCCXLVI MMMDCCCXLVI* MMMDCCCXLVI*
VALUE 29
* XXIX XXIX* XXIX*
FASP, FORTRAN Alphameric Subroutine Package Page 363
DASAVE, Integer & Character FORTRAN DATA Statement Generator
DDDDD AAA SSSSSS AAA VV VV EEEEEEEE
DD DD AAAA SS AAAA VV VV EE
DD DD AA AA SS AA AA VV VV EE
DD DD AA AA SSSS AA AA VV VV EEEEE
DD DD AAAAAAA SS AAAAAAA VVVV EE
DD DD AA AA SS AA AA VVV EE
DDDDD AA AA SSSSSS AA AA VV EEEEEEEE
DASAVE, Integer & Character FORTRAN DATA Statement Generator
------ ------- - --------- ------- ---- --------- ---------
The FORTRAN DATA statement generator routine named DASAVE
allows the compilation into a FORTRAN program of integer
data which has been calculated by another program or of
alphameric data which another program has read using a
multiple of an A1 format.
The DASAVE routine accepts as input an integer or alphameric
array which can be longer than can be represented by a
single DATA statement. Output consists of the DIMENSION
statements for the original array and for component arrays
each small enough to be represented by a single DATA
statement, of the EQUIVALENCE statements necessary to
reconstruct the original array from the component arrays,
and of the DATA statements representing the component
arrays. A maximum of 200 component arrays can be
represented, but this limit can easily be raised by
increasing the dimension of a single array internal to the
DASAVE routine. In addition to the length limit imposed by
the maximum number of component arrays, the manner in which
the names of the component arrays are constructed limits the
length of the original array to 99999 locations.
The width of the statements and the number of lines per
statement are specified by the calling program. If an
integer array is being represented, the DATA statements can
have the numbers arranged in columns or else compressed to
produce a shorter output file. Alphameric arrays are always
represented with the characters arranged in columns.
A single call to DASAVE either can produce the complete set
of DIMENSION, EQUIVALENCE and DATA statements necessary to
represent the contents of the input array, or can produce
only those statements of a particular type which would be
combined with statements generated by other calls to
represent the contents of the input array. If several
arrays are being represented in a single output file by a
sequence of calls to this routine, then separate calls to
this routine should be used to generate all of the
statements of a particular type necessary to represent all
of the arrays before generating all of the statements of the
next type since some compilers require that all DIMENSION
FASP, FORTRAN Alphameric Subroutine Package Page 364
DASAVE, Integer & Character FORTRAN DATA Statement Generator
and EQUIVALENCE statements appear prior to any DATA
statements.
The DASAVE Routine Argument List
--- ------ ------- -------- ----
The argument list of routine DASAVE is
SUBROUTINE DASAVE(IPART ,IFORMT,MAXCLM,MAXLIN,IDATA ,
1 KNTDAT,LETTER,KNTLTR,NAME ,KNTNAM,IOUT ,IERR )
with the associated DIMENSION statement
DIMENSION IDATA(KNTDAT),LETTER(KNTLTR),NAME(KNTNAM)
All arguments except IERR are used for input only and are
returned unchanged.
IPART = specifies whether the current call to this routine
is to generate all of the statements necessary to
represent the contents of the input array, or
whether the current call is to generate only those
statements of a particular type which could be
combined with the output from other calls to this
routine to represent the contents of the input
array.
= -1, construct DIMENSION and EQUIVALENCE statements
but not DATA statements
= 0, construct DIMENSION, EQUIVALENCE and DATA
statements
= 1, construct DIMENSION statements only
= 2, construct EQUIVALENCE statements only
= 3, construct DATA statements only
= -4, -3 or -2, identical to IPART=-1, 0 or 1
respectively, except that the resulting DIMENSION
statements specify the component arrays necessary
to construct the original array, but do not include
the name and dimension of the original array.
IFORMT = specifies whether the characters in the LETTER
array or the integer values in the IDATA array are
to be represented. If integer values are being
represented, then IFORMT also specifies whether the
DATA statements are to be written with spaces
compacted out, or with the numbers arranged in
columns.
= -1, represents the characters in LETTER(1) through
and including LETTER(KNTLTR) in a column form.
Each character is preceded by 1H in the DATA
statements.
= 0, represents the integer values in IDATA(1)
through and including IDATA(KNTDAT) in a compact
FASP, FORTRAN Alphameric Subroutine Package Page 365
DASAVE, Integer & Character FORTRAN DATA Statement Generator
form. Although this compact form is acceptable to
a FORTRAN compiler, a person would have
considerable difficulty locating the value
associated with a particular subscript when
visually scanning DATA statements written in this
compact form.
= 1 or greater, represents the integers values in
IDATA(1) through and including IDATA(KNTDAT) in a
column form. If IFORMT is less than the number of
characters needed to represent the value requiring
the most characters, then the columns will be of
the same width as the representation of the value
requiring the most characters. If IFORMT is equal
to or greater than the number of characters needed
to represent the value requiring the most
characters, then the width of the columns will
equal IFORMT.
MAXCLM = the maximum number of characters to be in the
statement field of each line in the output. The
maximum value of MAXCLM is 66 which is the 72
character total width of a FORTRAN statement minus
the 6 characters in the statement number and
continuation fields.
MAXLIN = the maximum number of lines onto which a FORTRAN
statement can be written. MAXLIN includes both the
initial line of the statement and all possible
continuation lines. There is no limit imposed by
this routine, but the FORTRAN compiler will
probably restrict the number of lines to a number
such as 10 or 20.
IDATA = the array of integer values which are to be
represented in the output DATA statements if IFORMT
has the value zero or greater.
KNTDAT = the number of values in the IDATA array which are
to be represented if IFORMT has the value zero or
greater. All values starting with IDATA(1) and
extending through IDATA(KNTDAT) will be
represented.
LETTER = the array of characters which are to be represented
in the output DATA statements if IFORMT has the
value -1. Each location in the LETTER array holds
a single character read by an A1 format or defined
by a 1H field.
KNTLTR = the number of locations in the LETTER array which
contain characters to be represented if IFORMT has
the value -1. LETTER(1) through and including
LETTER(KNTLTR) each contain a single character
which is to be represented.
FASP, FORTRAN Alphameric Subroutine Package Page 366
DASAVE, Integer & Character FORTRAN DATA Statement Generator
NAME = an array containing the letters forming the name of
the array as it is to be written into the output
DATA statements. Each letter in the NAME array
must be read by an A1 format or defined by a 1H
field in a DATA statement in the calling program.
The name of a component array is constructed by
placing the digits of the first subscript
represented by the component array to the right of
the basic name, or if this would produce a
component array name longer than 6 characters by
replacing the rightmost letters of the basic name
with the digits of the first subscript represented
by the component array. Therefore, names of the
following types result.
basic name names of typical component arrays
A A001 A210 A419
ABC ABC001 ABC210 ABC419
ABCDEF ABCDE1 ABC210 ABC419
The names of all component arrays are made to be of
the same length by the insertion of zeroes, if
necessary, between the characters of the basic name
and the character representations of the smaller
corresponding IDATA array subscripts.
KNTNAM = the number of letters in the name contained in the
NAME array. KNTNAM must be 6 or less. The name is
contained in NAME(1) through NAME(KNTNAM).
IOUT = the unit number onto which the DATA statements are
to be written. This routine will only generate the
FORTRAN statements necessary to represent the data.
It is the responsibility of the calling program to
open the output file and to write any other FORTRAN
statements and/or FORTRAN comment lines which,
depending upon the application, might be necessary
either before or after the output from this
routine.
IERR = returned describing whether the values or
characters could be represented. The input value
of IERR is ignored.
= 0, returned if the values or characters could be
represented.
= 1, returned if the line width specified by MAXCLM
was too short.
= 2, returned if more than 200 individual DATA
statements would be required.
FASP, FORTRAN Alphameric Subroutine Package Page 367
DASAVE, Integer & Character FORTRAN DATA Statement Generator
Sample Output from the DASAVE routine
------ ------ ---- --- ---- -------
The following short main program produced the output printed
below it. The only difference between the 2 DASAVE routine
calls is in the value of IFORMT which is 0 in the first call
and 1 in the second.
DIMENSION IDATA(115),NAME(6)
DATA NAME/1HI,1HV,1HA,1HL,1HU,1HE/
DATA IOUT,KNTDAT/1,115/
DO 1 I=1,KNTDAT
1 IDATA(I)=KNTDAT-I+1
WRITE(IOUT,2)
2 FORMAT(13H COMPACT FORM)
CALL DASAVE(0,0,54,5,IDATA,
1KNTDAT,LETTER,1,NAME,6,IOUT,IERR)
WRITE(IOUT,3)
3 FORMAT(12H COLUMN FORM)
CALL DASAVE(0,1,54,5,IDATA,
1KNTDAT,LETTER,1,NAME,6,IOUT,IERR)
STOP
END
COMPACT FORM
DIMENSION IVALUE(115),IVALU1(80),IVAL81(35)
EQUIVALENCE (IVALU1(1),IVALUE(1)),
1(IVAL81(1),IVALUE(81))
DATA IVALU1/115,114,113,112,111,110,109,108,107,106,
1105,104,103,102,101,100,99,98,97,96,95,94,93,92,91,90,
289,88,87,86,85,84,83,82,81,80,79,78,77,76,75,74,73,72,
371,70,69,68,67,66,65,64,63,62,61,60,59,58,57,56,55,54,
453,52,51,50,49,48,47,46,45,44,43,42,41,40,39,38,37,36/
DATA IVAL81/35,34,33,32,31,30,29,28,27,26,25,24,23,22,
121,20,19,18,17,16,15,14,13,12,11,10,9,8,7,6,5,4,3,2,1/
COLUMN FORM
DIMENSION IVALUE(115),IVALU1( 50),IVAL51( 50),
1 IVA101( 15)
EQUIVALENCE (IVALU1(1),IVALUE( 1)),
1 (IVAL51(1),IVALUE( 51)),
2 (IVA101(1),IVALUE(101))
DATA IVALU1/115,114,113,112,111,110,109,108,107,106,
1 105,104,103,102,101,100, 99, 98, 97, 96,
2 95, 94, 93, 92, 91, 90, 89, 88, 87, 86,
3 85, 84, 83, 82, 81, 80, 79, 78, 77, 76,
4 75, 74, 73, 72, 71, 70, 69, 68, 67, 66/
DATA IVAL51/ 65, 64, 63, 62, 61, 60, 59, 58, 57, 56,
1 55, 54, 53, 52, 51, 50, 49, 48, 47, 46,
2 45, 44, 43, 42, 41, 40, 39, 38, 37, 36,
3 35, 34, 33, 32, 31, 30, 29, 28, 27, 26,
4 25, 24, 23, 22, 21, 20, 19, 18, 17, 16/
DATA IVA101/ 15, 14, 13, 12, 11, 10, 9, 8, 7, 6,
1 5, 4, 3, 2, 1/
FASP, FORTRAN Alphameric Subroutine Package Page 368
DASHOW, Displays Integer or Real Number in Narrowest Field
DDDDD AAA SSSSSS HH HH OOOOO WW WW
DD DD AAAA SS HH HH OO OO WW WW
DD DD AA AA SS HH HH OO OO WW WW WW
DD DD AA AA SSSS HHHHHHHH OO OO WW WWWW WW
DD DD AAAAAAA SS HH HH OO OO WWWW WWWW
DD DD AA AA SS HH HH OO OO WWW WWW
DDDDD AA AA SSSSSS HH HH OOOOO WW WW
DASHOW, Displays Integer or Real Number in Narrowest Field
------ -------- ------- -- ---- ------ -- --------- -----
DASHOW represents an integer value or a real value so that
it can be written with a FORTRAN format statement containing
a multiple A1 alphameric specification. The representation
is left justified and rightmost fill with spaces (blanks) is
not provided. If the array which is supplied to receive the
representation of the number is too short, then this array
is instead filled with asterisks.
Integers can be represented as either octal or decimal
numbers. Real values can be represented without an
exponent, or can be represented in a scientific notation
consisting of the representation of a value in the range 1.0
through 9.999... followed by the letter E and a
representation of the power of ten by which the original
value was multiplied to obtain the value within the range
1.0 through 9.999... If a real value is being represented,
then any zero which is to the right of the decimal point and
which does not have a nonzero value to its right is
suppressed. A real value which is represented without an
exponent and which has an absolute value less than one is
represented without a zero to the left of the decimal point
unless this zero would be the only digit shown.
DASHOW calls DANUMB to represent integers and DARITE to
represent real values. DARITE must be called directly if
more flexibility in the representation of real values is
required.
FASP, FORTRAN Alphameric Subroutine Package Page 369
DASHOW, Displays Integer or Real Number in Narrowest Field
The DASHOW Argument List
--- ------ -------- ----
The argument list of routine DASHOW is
SUBROUTINE DASHOW(KONTRL,MINDEC,MAXDEC,MINSIG,MAXSIG,
1 IDECML,IVALUE,VALUE ,MAXBFR,KOUNT ,IBUFFR,IERR )
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The following argument is used only for input and is
returned unchanged.
KONTRL = selects whether the value being represented is
input as an integer or as a real number.
= -1, represent the integer value input as the
argument named IVALUE as an octal number. It
should be noted that it is the value which is
represented, not the particular bit sequence by
which this value happens to be stored in the
computer. The integer having the decimal value -8
would be represented as the octal number -10.
= 0, represent the integer value input as the
argument named IVALUE as a decimal number.
= 1 or greater, represent the real value input as the
argument named VALUE with no more than MAXDEC nor
less than MINDEC digits to the right of the decimal
point and with no more than MAXSIG nor less than
MINSIG significant digits. If the value cannot be
represented using the specified numbers of digits
in floating point form (without an E exponent),
then the value is instead represented in scientific
notation (with an E exponent) and with IDECML
digits to the right of the decimal point in the
representation of the value which either has the
value zero, or which has been modified though
multiplication by a power of ten so as to fall
within the range 1.0 through 9.999...
The following arguments control the representation of those
real values which are shown in floating point form (without
an E exponent). These arguments are returned unchanged, and
do not control the representation of integers or of those
real values which are shown in scientific notation (with an
E exponent). However, if MAXSIG is set to zero, then the
real value will be shown in scientific notation.
MINDEC = 0, there is no lower limit to the number of digits
which must be shown to the right of the decimal
point of a number represented in floating point
form, and the field into which the representation
of the number is inserted does not even need to
FASP, FORTRAN Alphameric Subroutine Package Page 370
DASHOW, Displays Integer or Real Number in Narrowest Field
include room for the decimal point.
= greater than zero, MINDEC is the minimum number of
digits which can be shown to the right of the
decimal point of a number represented in floating
point form if the rightmost of these digits is
nonzero. If it would not be possible to include at
least this number of digits to the right of the
decimal point, then the value will be represented
in scientific notation instead. After determining
that it would be possible to represent at least
MINDEC digits to the right of the decimal point,
then each digit which is to the right of the
decimal point and which does not have a nonzero
digit to its right is suppressed and even the
decimal point will not be shown if there are no
nonzero digits to the right of the decimal point.
MAXDEC = 0, represent only the digits which are to the left
of the decimal point of a number represented in
floating point form. The decimal point will not be
shown.
= greater than zero, MAXDEC is the maximum number of
digits which can be shown to the right of the
decimal point of a number represented in floating
point form. If MAXSIG (the maximum number of
significant digits) minus the number of digits to
the left of the decimal point is less than MAXDEC
but not less than MINDEC, then the number of digits
to the right of the decimal point is instead set to
MAXSIG minus the number of digits to the left of
the decimal point. If MAXSIG minus the number of
digits to the left of the decimal point is less
than MINDEC, then the number is represented in
scientific notation instead.
MINSIG = minimum number of significant digits which can be
shown in a number represented in floating point
form if the rightmost of these digits is nonzero.
If the floating point representation of the value
would contain less than MINSIG significant digits,
then the value will be represented in scientific
notation. If the value being displayed has the
value zero, then MINSIG is ignored.
MAXSIG = 0, represent the value input as the argument named
VALUE in scientific notation (with an E exponent)
rather than in floating point form (without an E
exponent).
= greater than zero, MAXSIG is the maximum number of
digits which can be shown in a number represented
in floating point form, starting with the leftmost
nonzero digit, counting it and all digits to its
right. MAXSIG does not include the decimal point
and does not include the minus sign if the value is
FASP, FORTRAN Alphameric Subroutine Package Page 371
DASHOW, Displays Integer or Real Number in Narrowest Field
negative. The number of digits displayed to the
right of the decimal point is reduced if necessary
so that the number of digits starting at the
leftmost nonzero displayed digit and counting it
and all digits displayed to its right does not
exceed MAXSIG. If MAXSIG would be less than the
number of digits left of the decimal point in the
representation of the value, then the value will be
represented in scientific notation.
The following argument controls the representation of those
values which are shown in scientific notation (with an E
exponent). This argument is returned unchanged, and does
not control the representation of integers or of those real
values which are represented in floating point form (without
an E exponent).
IDECML = same as MAXDEC except that IDECML applies only to
values displayed in scientific notation. Note that
if the value is being represented in scientific
notation, then a nonzero digit will be used left of
the decimal point unless the value is itself zero.
Therefore, the maximum number of significant digits
which can be displayed in scientific notation is
IDECML+1.
The following arguments are used only for input and are
returned unchanged.
IVALUE = input containing the integer value to be
represented if KONTRL is less than or equal to
zero.
VALUE = input containing the real value to be represented
if KONTRL is greater than zero.
MAXBFR = subscript of the rightmost location in the IBUFFR
array which can be used by this routine for the
return of characters in the representation of the
number. The representation of the number will use
fewer locations in the IBUFFR array if possible.
The following argument is used both for input to this
routine and for output to the calling program.
KOUNT = input containing the subscript of the rightmost
(highest subscript value) IBUFFR array location
which is currently in use and which must be
preserved. The representation of the number will
be returned starting in IBUFFR(KOUNT+1) and
extending through IBUFFR(MAXBFR) if necessary.
= returned containing the subscript of the rightmost
IBUFFR array location in which a character of the
representation of the number is returned. The
FASP, FORTRAN Alphameric Subroutine Package Page 372
DASHOW, Displays Integer or Real Number in Narrowest Field
portion of the IBUFFR array to the right of the
representation and extending through IBUFFR(MAXBFR)
is returned undefined, but possibly changed.
The following arguments are used only for output. Their
input values are ignored.
IBUFFR = array in which the characters which form the
representation of the number are returned as though
read by a multiple of an A1 format specification or
as though defined by several 1H fields. The
leftmost character of the representation is
returned in IBUFFR(KOUNT+1). KOUNT is returned
containing the subscript of the IBUFFR array
location containing the rightmost character of the
representation of the number.
IERR = -1, returned if the field starting with
IBUFFR(KOUNT+1) and extending through
IBUFFR(MAXBFR) was too short to contain the
representation of the number. This field is
returned containing asterisks, and KOUNT is
returned set equal to MAXBFR. If KONTRL is greater
than zero, then the field was too short to contain
even the representation of the real number in
scientific notation.
= 0, returned if KONTRL is less than or equal to zero
and the integer value could be shown, or if KONTRL
is greater than zero and the real value was
represented in floating point form.
= 1 or greater, returned if KONTRL is input greater
than zero selecting the representation of the real
number which was input as the argument named VALUE,
and this value was represented in scientific
notation. IERR is the number of digits which would
be shown to the left of the exponent if the
rightmost of these digits is nonzero. Zeroes which
are to the right of the decimal point and which do
not have a nonzero digit to their right are not
shown but are included in the value of IERR. If
this routine is being called to append the
representation of a number to a line of text which
already contains other information, then the
calling program can test the returned value of IERR
to determine whether there was enough room at the
end of the line for the number to be represented
with sufficient accuracy.
FASP, FORTRAN Alphameric Subroutine Package Page 373
DASHOW, Displays Integer or Real Number in Narrowest Field
Demonstration Program to Interactively Test DASHOW Arguments
------------- ------- -- ------------- ---- ------ ---------
The user of the program listed below types a pair or pairs
of input argument names and values. DASHOW is called at the
end of each input line, or whenever a semicolon is found.
DIMENSION IBUFFR(60),JBUFFR(60),KNTLTR(9),IWORD(53),
1NUMTYP(9)
DATA IWORD/ 1HK,1HO,1HN,1HT,1HR,1HL,
1 1HM,1HI,1HN,1HD,1HE,1HC, 1HM,1HA,1HX,1HD,1HE,1HC,
2 1HM,1HI,1HN,1HS,1HI,1HG, 1HM,1HA,1HX,1HS,1HI,1HG,
3 1HI,1HD,1HE,1HC,1HM,1HL, 1HI,1HV,1HA,1HL,1HU,1HE,
4 1HV,1HA,1HL,1HU,1HE, 1HM,1HA,1HX,1HB,1HF,1HR/
DATA KONTRL,MINDEC,MAXDEC,MINSIG,MAXSIG,IDECML,IVALUE,
1VALUE,MAXBFR/0,0,0,0,0,0,0,0.0,60/,ITTY,JTTY/5,5/,
2KNTLTR/6,6,6,6,6,6,6,5,6/,NUMTYP/0,0,0,0,0,0,0,1,0/
KIND=0
1 WRITE(JTTY,2)KONTRL,MINDEC,MAXDEC,MINSIG,MAXSIG,
1IDECML,MAXBFR,IVALUE,VALUE
2 FORMAT(8H KONTRL,I3,7H MINDEC,I3,7H MAXDEC,I3,
17H MINSIG,I3,7H MAXSIG,I3,7H IDECML,I3/8H MAXBFR,I3,
27H IVALUE,I13,7H VALUE,1E13.4)
3 WRITE(JTTY,4)
4 FORMAT(2H *,$)
READ(ITTY,5)JBUFFR
5 FORMAT(60A1)
LOWBFR=1
C OBTAIN NEXT USER SPECIFICATION OF ARGUMENT VALUE
6 CALL DALEAD(-2,-3,1,1,53,
1IWORD,1,9,KNTLTR,NUMTYP,2,1,
2IWORD,2,1,KNTLTR,JBUFFR,60,LOWBFR,
3KIND,KOMAND,KWRDID,KCNTID,LOCAL,LWRDID,LCNTID,
4INITAL,JVALUE,AVALUE,IFLOAT)
GO TO(8,8,3,6,7,6,6,6,8,3,6,7,6,6,6),KIND
7 IF(KOMAND.EQ.1)KONTRL=JVALUE
IF(KOMAND.EQ.2)MINDEC=JVALUE
IF(KOMAND.EQ.3)MAXDEC=JVALUE
IF(KOMAND.EQ.4)MINSIG=JVALUE
IF(KOMAND.EQ.5)MAXSIG=JVALUE
IF(KOMAND.EQ.6)IDECML=JVALUE
IF(KOMAND.EQ.7)IVALUE=JVALUE
IF(KOMAND.EQ.8)VALUE =AVALUE
IF(KOMAND.EQ.9)MAXBFR=JVALUE
GO TO 6
C REPORT REPRESENTATION OF NUMBER IN SPECIFIED FORMAT
8 KOUNT=0
CALL DASHOW(KONTRL,MINDEC,MAXDEC,MINSIG,MAXSIG,
1 IDECML,IVALUE,VALUE ,MAXBFR,KOUNT ,IBUFFR,IERR )
IF(KOUNT.GT.0)WRITE(JTTY,9)IERR,(IBUFFR(I),I=1,KOUNT)
9 FORMAT(3H (,1I2,2H) ,60A1)
IF(KIND.EQ.1)GO TO 1
GO TO 6
END
FASP, FORTRAN Alphameric Subroutine Package Page 374
DASHOW, Displays Integer or Real Number in Narrowest Field
Typical Dialog Between DASHOW Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
KONTRL 0 MINDEC 0 MAXDEC 0 MINSIG 0 MAXSIG 0 IDECML 0
MAXBFR 60 IVALUE 0 VALUE 0.0000E+00
*KONTRL 1 MAXD 2 MAXS 6 VAL 12.03;VAL 12.003;VAL 12.006
( 0) 12.03
( 0) 12
( 0) 12.01
KONTRL 1 MINDEC 0 MAXDEC 2 MINSIG 0 MAXSIG 6 IDECML 0
MAXBFR 60 IVALUE 0 VALUE 0.1201E+02
*VAL -0.03;VAL -0.003;VAL -.006;MINS 4;IDEC 3
( 0) -.03
( 0) -0
( 0) -.01
( 1) -6E-3
( 4) -6E-3
KONTRL 1 MINDEC 0 MAXDEC 2 MINSIG 4 MAXSIG 6 IDECML 3
MAXBFR 60 IVALUE 0 VALUE -0.6000E-02
*VAL 12345.6;VAL 1234.56;VAL 123.456;VAL 12.3456;VAL 1.23456
( 0) 12345.6
( 0) 1234.56
( 0) 123.46
( 0) 12.35
( 4) 1.235
KONTRL 1 MINDEC 0 MAXDEC 2 MINSIG 4 MAXSIG 6 IDECML 3
MAXBFR 60 IVALUE 0 VALUE 0.1235E+01
*MAXS 4 MINS 4 MIND 2 VA 9.994;VA 9.996;VA 99.994;VA 99.996
( 4) 9.994
( 0) 10
( 0) 99.99
( 4) 1E2
KONTRL 1 MINDEC 2 MAXDEC 2 MINSIG 4 MAXSIG 4 IDECML 3
MAXBFR 60 IVALUE 0 VALUE 0.1000E+03
*MAXS 0 MAXB 7 VAL 0.001234;MAXB 6;MAXB 5;MAXB 4;MAXB 3
( 3) 1.23E-3
( 2) 1.2E-3
( 1) 1E-3
( 1) 1E-3
(-1) ***
KONTRL 1 MINDEC 2 MAXDEC 2 MINSIG 4 MAXSIG 0 IDECML 3
MAXBFR 3 IVALUE 0 VALUE 0.1234E-02
*MAXS 0 MAXB 7 VAL -0.001234;MAXB 6;MAXB 5;MAXB 4;MAXB 3
( 2) -1.2E-3
( 1) -1E-3
( 1) -1E-3
(-1) ****
(-1) ***
KONTRL 1 MINDEC 2 MAXDEC 2 MINSIG 4 MAXSIG 0 IDECML 3
MAXBFR 3 IVALUE 0 VALUE -0.1234E-02
FASP, FORTRAN Alphameric Subroutine Package Page 375
DASITE, Routine to Locate Simulated Array Item in Buffer
DDDDD AAA SSSSSS IIIIII TTTTTTTT EEEEEEEE
DD DD AAAA SS II TT EE
DD DD AA AA SS II TT EE
DD DD AA AA SSSS II TT EEEEE
DD DD AAAAAAA SS II TT EE
DD DD AA AA SS II TT EE
DDDDD AA AA SSSSSS IIIIII TT EEEEEEEE
DASITE, Routine to Locate Simulated Array Item in Buffer
------ ------- -- ------ --------- ----- ---- -- ------
DASITE returns the position within a singly subscripted
buffer of a single item of a possibly multiply subscripted
array equivalenced with or otherwise loaded into part or all
of the singly subscripted buffer (as defined perhaps by the
dictionary constructed by the DALOAD routine). There is no
upper limit to the number of subscripts of the arrays
simulated in the buffer (other than the obvious restrictions
imposed by the lengths of the NOWSUB and NUMSTR arrays and
by the length of the buffer itself). The range of values of
any subscript can start at any value and can be either
increasing or decreasing. This conversion is the opposite
of that performed by DANAME.
For example, if the first statement read during the
construction of the dictionary by DALOAD was
DIMENSION FIRST(10),SECOND(10,10),THIRD(10,10,10)
and if it is assumed that the first (left) subscript of each
simulated array varies the most rapidly (the usual FORTRAN
convention for array names appearing without subscripts in
READ and WRITE statements) then the simulated array location
THIRD(3,2,1) would be equivalent to the singly subscripted
buffer location having the subscript 123. The array name
and its subscripts are converted by DASITE to the buffer
location subscript by determining that the locations below
the desired location include
a. all 10 locations of the FIRST array
b. all 100 locations of the SECOND array
c. the 10 locations THIRD(1,1,1) through THIRD(10,1,1)
d. the 2 locations THIRD(1,2,1) and THIRD(2,2,1)
FASP, FORTRAN Alphameric Subroutine Package Page 376
DASITE, Routine to Locate Simulated Array Item in Buffer
The DASITE Argument List
--- ------ -------- ----
The argument list of routine DASITE is
SUBROUTINE DASITE(IRAPID,KOUNT ,LOWSUB,KNTSUB,NOWSUB,
1 IEXTRA,LRGNUM,NUMUSD,NUMSTR,LSTKNT,NUMINI,INITAL,
2 LOCATN)
with the associated DIMENSION statement
DIMENSION NOWSUB(KNTSUB),NUMSTR(NUMUSD)
The following arguments are used as input only and are
returned unchanged
IRAPID = 0, the simulated arrays, if multiply subscripted,
have their left subscripts varying the most
rapidly. This is the normal FORTRAN convention for
READs or WRITEs in which the name of the array is
used without any subscripts.
= 1, the simulated arrays, if multiply subscripted,
have their right subscripts varying the most
rapidly.
KOUNT = sequence number of the desired array among all
arrays in the buffer. The first array would be
selected by KOUNT=1, the second by KOUNT=2 and so
on.
LOWSUB = subscript of the NOWSUB array location containing
the first simulated subscript of the possibly
multiply subscripted array location being found
within the singly subscripted buffer.
KNTSUB = subscript of the NOWSUB array location containing
the final simulated subscript of the possibly
multiply subscripted array location being found
within the singly subscripted buffer. If the lower
portion of the NOWSUB array is used, then LOWSUB
will have the value 1 and KNTSUB will be the number
of simulated subscripts of the possibly multiply
subscripted array.
NOWSUB = array containing the simulated subscripts of the
possibly multiply subscripted array location being
found within the singly subscripted buffer.
NOWSUB(LOWSUB) through NOWSUB(KNTSUB) contain the
values of the simulated subscripts.
IEXTRA = 0, for each simulated array, the NUMSTR array
contains only the number of subscripts and the
subscript limits.
= greater than zero, NUMSTR(LRGNUM) contains the
FASP, FORTRAN Alphameric Subroutine Package Page 377
DASITE, Routine to Locate Simulated Array Item in Buffer
first of IEXTRA words which appear before the first
subscript description. Thereafter, IEXTRA extra
words are to be ignored between descriptions of
consecutive simulated arrays.
= -1, the description of the subscript ranges for
each of the simulated arrays is preceded by a
variable number of words to be ignored. Each
section to be ignored starts with a word containing
the number of words exclusive of itself which are
to be ignored before the next subscript count is
found. NUMSTR(LRGNUM) then contains the number of
words exclusive of itself which are to be ignored
before the first subscript count and range
descriptions.
= -2, the contents of the NUMSTR array were defined
by the DALOAD routine. The description of the
subscript ranges for each of the simulated arrays
is preceded by 2 extra words. LRGNUM can contain
the subscript of the start of the description of a
logical group of arrays, or can contain the
subscript of the start of the description of the
first array in the logical group. If
NUMSTR(LRGNUM) is less than or equal to zero then
the subsequent array description is the first which
can be matched by this routine, and the dictionary
is effectively terminated prior to the next array
description which does not start with a value
greater than zero. If NUMSTR(LRGNUM) is greater
than zero, then the array description starting with
NUMSTR(LRGNUM) is the first which can be matched by
this routine, and the dictionary is effectively
terminated prior to the first array description
which does not start with a value greater than
zero.
LRGNUM = if IEXTRA is greater than -2, or else if IEXTRA=-2
and NUMSTR(LRGNUM) is greater than zero, then
LRGNUM is the subscript of the NUMSTR array
location containing the start of the description of
the simulated array which would be selected if
KOUNT=1.
= if IEXTRA=-2 and if NUMSTR(LRGNUM) is less than or
equal to zero, then NUMSTR(LRGNUM) is the first
location in the description of the logical group of
arrays containing the array to be identified, and
the subsequent description is of the array which
would be selected if KOUNT=1.
NUMUSD = subscript of the NUMSTR array location containing
the end of the description of the final simulated
array.
NUMSTR = array describing the subscript limits of the
possibly multiply subscripted arrays simulated in
FASP, FORTRAN Alphameric Subroutine Package Page 378
DASITE, Routine to Locate Simulated Array Item in Buffer
the singly subscripted buffer. The contents of the
NUMSTR array are, for each array simulated in the
buffer, the number of subscripts of the simulated
array followed by left and right limiting values of
these subscripts (values which the subscripts would
have if the arrays simulated in the buffer were
actually included in DIMENSION statements). If the
item in the buffer would be subscripted at 1, or
would not subscripted, then a single 0 can be used
in place of the sequence 1,1,1. it should be noted
that the right limit can be either greater than,
equal to or less than the left limit.
The following arguments are used as both input and output
LSTKNT = should be set to zero by the calling program before
DASITE is first called and whenever the dictionary
corresponding to the buffer changes.
= returned containing the input value of KOUNT. The
subsequent call to this routine will use the input
values of LSTKNT, NUMINI and INITAL rather than
recompute these for the simulated arrays earlier in
the dictionary if the new value of KOUNT is greater
than or equal to LSTKNT.
NUMINI = input value is ignored if LSTKNT is zero or if
LSTKNT is greater than KOUNT. NUMINI is set by
each call to DASITE and should never be set by the
calling program.
= if LSTKNT is input greater than zero, then the
value of NUMINI, as returned by the previous call
to this routine, is the subscript of the NUMSTR
array location containing the start of the
description of the array having as its sequence
number the value of LSTKNT. If IEXTRA=0, then
NUMSTR(NUMINI) contains the subscript count at the
start of the description.
= returned containing the subscript of the NUMSTR
array location containing the start of the
description of the array having as its sequence
number the input value of KOUNT.
INITAL = input value is ignored if LSTKNT is input
containing zero or if LSTKNT is input greater than
KOUNT. INITAL is set by each call to DASITE and
should never be set by the calling program.
= if LSTKNT is input greater than zero, then the
value of INITAL, as returned by the previous call
to this routine, is the subscript within the singly
subscripted buffer of the location containing the
start of the simulated array having as its sequence
number the input value of LSTKNT.
= returned containing the subscript within the singly
subscripted buffer of the location containing the
FASP, FORTRAN Alphameric Subroutine Package Page 379
DASITE, Routine to Locate Simulated Array Item in Buffer
start of the simulated array having as its sequence
number the input value of KOUNT.
The following argument is returned as output. Its input
value is ignored.
LOCATN = if returned greater than zero, LOCATN is the
subscript of the location within the singly
subscripted buffer which corresponds to the
simulated array location which is selected by the
input value of KOUNT and by the subscripts
contained in the NOWSUB array.
= 0, returned if one or more of the the subscripts
contained in the NOWSUB array are not in the ranges
predicted by the NUMSTR array.
= -1, returned if the NOWSUB array contains a
different number of subscripts than indicated by
the NUMSTR array.
= -2, returned if the NUMSTR array contains the
descriptions of less simulated arrays than the
input value of KOUNT. This value of LOCATN is also
returned if an error is detected in the NUMSTR
array.
For example, if the singly subscripted buffer contains the
values of the arrays A, B and E subscripted
A(1/3,1/5), B(1/5,1/6), E(1/10,1/10)
(this specification means that the left subscript of array A
can range from 1 through 3 and that the second subscript can
range from 1 through 5) and contains nonsubscripted items C
and D in the order
A,B,C,D,E
then the contents of the NUMSTR array would be (for
IEXTRA=0)
2,1,3,1,5,2,1,5,1,6,0,0,2,1,10,1,10
or
2,1,3,1,5,2,1,5,1,6,1,1,1,1,1,1,2,1,10,1,10
If KOUNT has the value 2 and if the NOWSUB array contains
the values 4 and 3 selecting B(4,3) and if IRAPID has the
value 0, so that all of array A, B(1,1) through B(5,1),
B(1,2) through B(5,2), and B(1,3) through B(3,3) would be
below B(4,3) then LOCATN would be returned as 15+5+5+3+1=29.
If IRAPID=0, then the order of the A array in the singly
subscripted buffer would be (reading across each line from
left to right)
FASP, FORTRAN Alphameric Subroutine Package Page 380
DASITE, Routine to Locate Simulated Array Item in Buffer
A(1,1),A(2,1),A(3,1),A(1,2),A(2,2),A(3,2),
A(1,3),A(2,3),A(3,3),A(1,4),A(2,4),A(3,4),
A(1,5),A(2,5),A(3,5)
If IRAPID=1, then the order of the A array in the singly
subscripted buffer would be
A(1,1),A(1,2),A(1,3),A(1,4),A(1,5)
A(2,1),A(2,2),A(2,3),A(2,4),A(2,5)
A(3,1),A(3,2),A(3,3),A(3,4),A(3,5)
If the Z array is effectively subscripted Z(4/3,-1/1) then
its representation in the NUMSTR array would be
2,4,3,-1,1
and if IRAPID=0, then its order in the singly subscripted
buffer would be
Z(4,-1),Z(3,-1),Z(4,0),Z(3,0),Z(4,1),Z(3,1)
If instead IRAPID=1, then the order of the Z array in the
singly subscripted buffer would be
Z(4,-1),Z(4,0),Z(4,1),Z(3,-1),Z(3,0),Z(3,1)
Use of DASITE with the Named Array Manipulation Routines
--- -- ------ ---- --- ----- ----- ------------ --------
If the DALOAD routine was used to construct the dictionary
the numeric portion of which is contained in the NUMSTR
array, and if the DAPICK routine has been called prior to
DASITE to identify the location within the buffer
corresponding to the start of the possibly multiply
subscripted array equivalenced with or otherwise loaded into
part or all of the buffer, then DASITE does not need to
calculate the sizes of the arrays below the identified
array, but DASITE must calculate the additional offset
required in the buffer to obtain the location corresponding
to a particular set of subscripts. This set of subscripts
will have been constructed by the DAROLL or DALOOP routine,
and will have been based upon the subscript ranges returned
by DAPICK. When DASITE is called subsequent to DAPICK and
DAROLL (or DALOOP), the arguments KOUNT, KNTSUB, LRGNUM,
NUMINI and INITAL should all be input to DASITE containing
the values assigned to the arguments of the same names by
DAPICK, and LSTKNT should be set equal to the value of KOUNT
returned by DAPICK. IEXTRA should be set to -2, LOWSUB
should be set to 1, and the NOWSUB array should be as
produced by the DAROLL or DALOOP routine.
FASP, FORTRAN Alphameric Subroutine Package Page 381
DASPAN, Routine to Evaluate Range Specifications
DDDDD AAA SSSSSS PPPPPP AAA NN NN
DD DD AAAA SS PP PP AAAA NNN NN
DD DD AA AA SS PP PP AA AA NNNN NN
DD DD AA AA SSSS PPPPPP AA AA NN NN NN
DD DD AAAAAAA SS PP AAAAAAA NN NNNN
DD DD AA AA SS PP AA AA NN NNN
DDDDD AA AA SSSSSS PP AA AA NN NN
DASPAN, Routine to Evaluate Range Specifications
------ ------- -- -------- ----- --------------
DASPAN evaluates range specifications typed by the user and
read by the calling program with a multiple of an A1 format.
Such range specifications can consist of a single number, or
of 2 numbers separated by an asterisk, or of either 2 or 3
numbers separated by slashes or by colons (the 2 characters
are equivalent). The asterisk notation is meant to indicate
that the value to the right of the asterisk is to be
repeated the number of times indicated by the number to the
left of the asterisk. Numbers specified in the slash or
colon notation can be interpreted as the lower and upper
bounds of a range if 2 numbers are present, or as the lower
bound, the increment, and the upper bound, if 3 numbers are
included.
DASPAN reports to the calling program whether the numbers
are present in the series specification, as well as the
values of those numbers which are present. The calling
program will have to assign default values for any numbers
which are missing. The range specification /2/10 or :2:10
indicates that the range is to extend from the default lower
bound with an increment of 2 through the upper bound of 10.
Each of the range specifications 1/10 or 1:10 or 1//10 or
1::10 is taken to mean that the range is to extend from 1
through 10 with the default increment. If the numbers
specify the values of a subscript of an array, then a
missing lower bound might be taken to indicate that the
subscript being varied starts at its minimum possible value
(usually 1), while a missing upper bound might indicate that
the subscript is to terminate at its maximum possible value.
The only printing delimiter character allowed between series
specifications is the comma. Two commas with no other
printing characters between them are taken to indicate a
missing series specification. The calling program is
informed if a semicolon is encountered in the text being
evaluated by this routine. The semicolon might be typed by
the user to signal that the calling program is to terminate
its requests for data from the user. An exclamation point
and any characters to its right are taken to be a comment
and are otherwise ignored. An ampersand and the characters
to its right are similarly considered to form a comment, but
FASP, FORTRAN Alphameric Subroutine Package Page 382
DASPAN, Routine to Evaluate Range Specifications
the calling program is informed that an ampersand was found.
Such an ampersand might be typed by the user to indicate
that the input text is to be continued on the following
line.
Two versions of the routine are supplied. DASPAN can
evaluate real numbers as well as integers including E, K and
M notations for specifying either of these. If the calling
program does not otherwise reference the free format input
routine DAHEFT, if the evaluation of real and octal numbers
is not needed, and if integers can be specified without
resorting to the E, K and M notations, then the routine
DAISPN should be used instead of DASPAN. Numbers evaluated
by DAISPN must consist only of digits following the optional
sign. DAISPN treats the characters ., %, K and M as illegal
characters.
The DASPAN and DAISPN Argument Lists
--- ------ --- ------ -------- -----
The argument lists of routines DASPAN and DAISPN are
SUBROUTINE DASPAN(KONTRL,KONECT,IBUFFR,MAXBFR,LOWBFR,
1 MANY ,KIND ,INIGOT,INCGOT,LMTGOT,INIVAL,INCVAL,
2 LMTVAL,VALINI,VALINC,VALLMT)
and
SUBROUTINE DAISPN (KONECT,IBUFFR,MAXBFR,LOWBFR,
1 MANY ,KIND ,INIGOT,INCGOT,LMTGOT,INIVAL,INCVAL,
2 LMTVAL)
both with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR)
The ordering of the arguments is the same for both routines,
but the first argument of DASPAN, and the last 3 arguments
of DASPAN are not included in the DAISPN argument list. The
argument definitions for DAISPN are identical to those of
DASPAN with the exception that the range specification is
always returned in integer form for DAISPN.
The following arguments are used for input only and are
returned unchanged.
KONTRL = -1, the number in the IBUFFR array is an octal
number. The number can be typed with a decimal
point and/or with an exponent. However, the number
following the letter E of the exponent is evaluated
in decimal. The value of the octal number is
returned as one of the arguments INIVAL, INCVAL or
FASP, FORTRAN Alphameric Subroutine Package Page 383
DASPAN, Routine to Evaluate Range Specifications
LMTVAL. It must be noted that numbers evaluated as
negative octal integers have the negative octal
integer as their value, not as their bit
representation in computer storage. For example,
on a 36 bit twos complement computer, the octal
number -400000000000 (which could also be typed as
-4E11 or -4E+11 where the 11 after the E is in
decimal) is represented as bit pattern having octal
notation 400000000000 and the octal number
-377777777777 is represented by the bit pattern
400000000001.
= 0, the number in the IBUFFR array is a decimal
integer. The number can be typed with a decimal
point (for example 1.23K or 1.23E3 equals 1230),
but is stored as an integer in DAHEFT, and is
output as one of the arguments INIVAL, INCVAL or
LMTVAL. Any decimal integer which the computer can
represent can be evaluated. This includes, on twos
complement computers, the largest negative number
the absolute value of which cannot be stored. On
the PDP-10, a 36 bit computer with twos complement
notation, the range of decimal integers is
-34359738368 through 34359738367 (the octal
notation of the bit patterns being 400000000000
through 377777777777).
= 1 or greater, the number in the IBUFFR array is a
real number. If possible, the real number will be
accumulated as an integer, then be converted to
real and shifted if necessary. KONTRL is then the
maximum number of digits in the integer. The value
is output as one of the arguments VALINI, VALINC or
VALLMT. If the item has more than KONTRL digits,
then the entire evaluation is done in real mode.
The advantage of calculating the real values in
integer as long as the precision of the computer is
not overflowed is that the calculation of the
portion of the number right of the decimal point is
more exact. As an example, if KONTRL is greater
than or equal to 4, then the number 33.33 can be
stored as the integer 3333, then be converted to
the real value 3333.0 and divided by 100.0 to
obtain the final answer. If it makes no difference
whether the number typed as 33.33 has value 33.33
or 33.32999... then KONTRL can be given the value
1.
KONECT = -1, once a slash, a colon, an asterisk or a number
is found in the IBUFFR array, the series
specification will extend to the next space or tab
character, or to any character other than a slash
or a colon or an asterisk which immediately follows
a number in the series specification. If
KONECT=-1, then the text
FASP, FORTRAN Alphameric Subroutine Package Page 384
DASPAN, Routine to Evaluate Range Specifications
1/2+3/4 5/6+7/+8
would contain the range specifications 1 to 2, 3 to
4, 5 to 6, and 7 to 8.
= 0, spaces and/or tab characters can appear between
the numbers of a series in addition to slashes or
colons or asterisk. The slashes or colons or the
asterisk are, however, required.
= 1, spaces and/or tab characters can appear between
the numbers of a series in addition to or in place
of slashes or colons or asterisks. If KONECT=1,
then the text
1 2 3,4/5 6,7 8/9,10 / 11 12
is exactly equivalent, except for the value of KIND
returned for the first series, to the text
1/2/3,4/5/6,7/8/9,10/11/12
The first series (1 2 3) in the first example would
return KIND=5 indicating a series in which neither
a slash nor a colon nor an asterisk appeared, while
the corresponding series specification in the
second example would return KIND=6 indicating that
at least 1 slash or at least 1 colon was
encountered in the series specification.
IBUFFR = input buffer array containing characters typed by
the user, as read by a multiple of an A1 format,
which is to be searched for series specifications.
IBUFFR then contains one character per computer
storage location.
MAXBFR = maximum subscript of the IBUFFR array locations to
be searched
The following arguments are used as both input and output.
LOWBFR = input containing the subscript within the IBUFFR
array of the first (leftmost) character which can
be scanned for a series specification. LOWBFR will
be returned pointing to the next character beyond
the series specification. LOWBFR and MANY must be
set by the calling program before anything is
processed in the current contents of the IBUFFR
array, but then should not be modified by the
calling program until the entire contents of the
IBUFFR array have been processed.
If KONECT is less than or equal to 0, forcing a
series specification to contain either a slash or a
colon or an asterisk between the numbers of a
series, and if a second number immediately follows
FASP, FORTRAN Alphameric Subroutine Package Page 385
DASPAN, Routine to Evaluate Range Specifications
a first without a separating slash or colon or
asterisk, then LOWBFR will be returned pointing to
the first character of the second number. If
KONECT is -1 and if either a space or a tab
character follows a series specification, then
LOWBFR will be returned pointing to the space or
tab character. If KONECT is greater than or equal
to 0, allowing spaces and tab characters to appear
within a series specification, then LOWBFR will be
returned pointing to the first character to the
right of the series and which is not itself a
space, a tab character, a slash, a colon or an
asterisk and which cannot appear within a number.
If there are no printing characters at or to right
of LOWBFR, then LOWBFR will be returned containing
MAXBFR+1 and KIND will be returned containing one.
MANY = should be input containing zero each time this
routine is called to begin processing of a new
logical section of text, as for example when
beginning processing of a line of text not tied to
the previous line by an ampersand at the end of the
previous line, or when processing the text to the
right of a semicolon. The initial zeroing of this
argument must be done by the calling program, but
thereafter the value returned by the previous call
to this routine can usually be used. MANY is
returned set to zero each time a semicolon (KIND=2)
is found, and each time an end of line not tied to
the following line by an ampersand (KIND=1) is
found. MANY is returned containing one plus its
input absolute value each time a description of a
series specification is returned by this routine,
each time an erroneous series specification is
found, each time an unknown character is found, or
each time an indication of a missing series
specification is found. KIND is returned
containing the value 3 and MANY is returned
containing the negative of the number of items
found if the next printing character following a
comma is an ampersand. MANY should not be changed
by the calling program if an ampersand (KIND being
returned=3) is found indicating that the subsequent
call to this routine is to process text which is to
be treated as though it appeared in place of the
ampersand and the characters to its right. The
effect is not quite the same as if the user had
typed all of the text on a single line since a
single series specification cannot be split across
the line boundary.
If MANY is input containing zero, then an initial
comma in the input text buffer is taken to indicate
an initial missing item, and MANY is then returned
FASP, FORTRAN Alphameric Subroutine Package Page 386
DASPAN, Routine to Evaluate Range Specifications
containing 1. If MANY is input greater than zero,
then an initial comma is ignored if followed by a
series specification. If MANY is input greater
than zero, then an initial comma followed by no
other printing characters, by a semicolon, or by an
exclamation point indicates a missing item. If
MANY is input greater than zero, then an initial
comma followed by an ampersand will cause the
remaining characters in the buffer to be ignored,
and MANY will be returned containing the negative
of its input value. If MANY is input negative,
then it is assumed that the contents of the current
buffer continue a previous line which terminated
with a comma followed by an ampersand, and MANY is
returned greater than zero.
The following arguments are used only for output.
KIND = returned describing the kind of item located in the
IBUFFR array.
= 1, either no printing characters or else an
exclamation point as the first printing character
(marking the rest of the line as a comment) was
found at or to the right of IBUFFR(LOWBFR). The
calling program should read a new line into IBUFFR.
MANY, INIGOT, INCGOT and LMTGOT are each returned
with the value zero.
= 2, a semicolon was found as the first printing
character at or to the right of IBUFFR(LOWBFR).
LOWBFR is returned pointing to the next character
beyond the location of the semicolon. It is
assumed the calling program will treat the
appearance of the semicolon as marking the end of a
statement. MANY is returned set to zero.
= 3, an ampersand was found as the first printing
character at or to the right of IBUFFR(LOWBFR).
The text to the right of the ampersand is taken as
a comment so LOWBFR is returned pointing beyond the
right end of the buffer. It is assumed that the
calling program will read in the contents of a new
buffer, then again request a new series evaluation
from this routine. The value of MANY must not be
changed by the calling program prior to this
following call. The effect is not quite the same
as if the user had typed all of the text on a
single line since a series specification cannot be
split across a line boundary.
= 4, a number or series specification was not found,
but a comma was found indicating a missing series
specification. INIGOT, INCGOT and LMTGOT are each
returned with the value zero so KIND=4 can be
considered equivalent to KIND=5 if such is
appropriate to the application for which this
routine is being used.
FASP, FORTRAN Alphameric Subroutine Package Page 387
DASPAN, Routine to Evaluate Range Specifications
= 5, a single number with neither a slash nor a colon
nor an asterisk was found in the input buffer.
Both INIGOT and LMTGOT are returned containing 1,
and the value of the number is returned either in
both INIVAL and LMTVAL or in both VALINI and
VALLMT, whichever is appropriate. INCGOT is
returned containing zero.
= 6, either 2 or 3 numbers were found, but without
slashes or colons or asterisks. The value of the
left number is returned in either INIVAL or VALINI,
of the right in either LMTVAL or VALLMT, and of the
middle, if present, in either INCVAL or VALINC,
whichever is appropriate. INIGOT and LMTGOT are
each returned containing 1. INCGOT is returned
containing 1 only if 3 numbers were found. KONECT
would have to be input as 1 for KIND to be returned
as 6.
= 7, a series specification containing one or more
slashes and/or colons was found. The location of
the slash or slashes or colon or colons relative to
the numbers, if any, in the series specification is
indicated by the returned values of INIGOT, INCGOT
and LMTGOT.
= 8, a series specification containing a single
asterisk was found. INCGOT is returned containing
zero. The location of the asterisk relative to the
numbers, if any, in the series specification is
indicated by the returned values of INIGOT and
LMTGOT.
= 9, a series specification was found which included
too many numbers, too many slashes, too many colons
or too many asterisks. INIGOT, INCGOT and LMTGOT
are each returned containing zero.
= 10, the first printing character (other than a
possible comma if MANY was input greater than zero)
in or to right of IBUFFR(LOWBFR) was not a
character which could appear in a number or number
range, and was not a comma, semicolon, exclamation
point or ampersand. LOWBFR is returned pointing to
the next character beyond this initial printing
character. The calling program must decrement
LOWBFR by 1 if the unknown character is to be
identified by other routines in this package.
INIGOT = 0, returned if the characters in the buffer do not
represent a series specification which includes an
initial number. INIVAL or VALINI, whichever is
appropriate, is returned undefined, but probably
changed.
= 1, returned if the characters in the buffer
represent a series specification which includes an
initial number. This initial number would probably
be interpreted as the start of a range if preceding
a slash or a colon, or as the number of times the
FASP, FORTRAN Alphameric Subroutine Package Page 388
DASPAN, Routine to Evaluate Range Specifications
following number is to be repeated if preceding an
asterisk. INIVAL or VALINI, whichever is
appropriate, is returned containing this initial
number.
INCGOT = 0, returned if the characters in the buffer do not
represent a series specification which includes a
middle number. INCGOT is always returned
containing zero for a series specified in asterisk
notation, since such a series specification cannot
include a middle number. INCVAL or VALINC,
whichever is appropriate, is returned undefined,
but probably changed.
= 1, returned if the characters in the buffer
represent a series specification which includes a
middle number. This middle number would probably
be interpreted as the increment by which the
initial number is to be varied until it reaches the
final number. INCVAL or VALINC, whichever is
appropriate, is returned containing this increment.
LMTGOT = 0, returned if the characters in the buffer do not
represent a series specification which includes a
final number. LMTVAL or VALLMT, whichever is
appropriate, is returned undefined, but probably
changed.
= 1, returned if the characters in the buffer
represent a series specification which includes a
final number. This final number would probably be
interpreted as the end of a range if following a
slash or a colon, or if following an asterisk as
the number which is to be repeated the number of
times indicated by the number preceding the
asterisk. LMTVAL or VALLMT, whichever is
appropriate, is returned containing this final
number. If merely a single number is found with
neither asterisk nor slash nor colon, then INCGOT
is returned containing zero, both INIGOT and LMTGOT
are returned containing 1, and the number is
returned either in both INIVAL and LMTVAL or in
both VALINI and VALLMT.
The following illustration shows the returned values
of INIGOT, INCGOT and LMTGOT for various series
specifications. Colons could appear in place of any
or all slashes shown in the examples. Those series
specifications containing spaces in place of slashes
or asterisks would require KONECT=1 for their
evaluation as single range specifications. It is
assumed that KONTRL=0 so that the evaluated numbers
are returned in INIVAL, INCVAL and LMTVAL. The
letter u in place of a returned value indicates that
the corresponding argument is returned undefined but
probably changed.
FASP, FORTRAN Alphameric Subroutine Package Page 389
DASPAN, Routine to Evaluate Range Specifications
LMTVAL
INCVAL !
INIVAL ! !
LMTGOT ! ! !
INCGOT ! ! ! !
INIGOT ! ! ! ! !
empty or / or // or * 0 0 0 u u u
/// or 5///9 or 5 2 2 9 0 0 0 u u u
** or 5*2*9 or 5*2/9 or 5/2*9 0 0 0 u u u
/9 or //9 or *9 0 0 1 u u 9
/2/ 0 1 0 u 2 u
/2/9 or /2 9 0 1 1 u 2 9
5/ or 5// or 5* 1 0 0 5 u u
5 1 0 1 5 u 5
5/9 or 5//9 or 5 9 or 5*9 1 0 1 5 u 9
5/2/ or 5 2/ 1 1 0 5 2 u
5/2/9 or 5 2/9 or 5/2 9 or 5 2 9 1 1 1 5 2 9
INIVAL = returned containing the number at the start of the
range specification if this is evaluated as an
octal or decimal integer (KONTRL input less than or
equal to zero and INIGOT returned = 1).
INCVAL = returned containing the middle number in the range
specification if this is evaluated as an octal or
decimal integer (KONTRL input less than or equal to
zero and INCGOT returned = 1).
LMTVAL = returned containing the number at the end of the
range specification if this is evaluated as an
octal or decimal integer (KONTRL input less than or
equal to zero and LMTGOT returned = 1).
VALINI = returned containing the number at the start of the
range specification if this is evaluated as a real
number (KONTRL input greater than zero and INIGOT
returned = 1).
VALINC = returned containing the middle number in the range
specification if this is evaluated as a real number
(KONTRL input greater than zero and INCGOT returned
= 1).
VALLMT = returned containing the number at the end of the
range specification if this is evaluated as a real
number (KONTRL input greater than zero and LMTGOT
returned = 1).
FASP, FORTRAN Alphameric Subroutine Package Page 390
DASPAN, Routine to Evaluate Range Specifications
Demonstration Program to Interactively Test DASPAN
------------- ------- -- ------------- ---- ------
The program listed on the following pages accepts a line of
text from the user, then reports each punctuation mark, each
single value, and each range together with the portion of
the text by which these are specified. A sample dialog
between the program and the user is presented following the
listing of the program.
C PROGRAM TO DEMONSTRATE DASPAN ROUTINE
C
DIMENSION IBUFFR(60),JBUFFR(60),LEGEND(28)
DATA LEGEND/1HI,1HN,1HI,1HT,1HI,1HA,1HL,1H ,
11H,,1H ,1HI,1HN,1HC,1HR,1HE,1HM,1HE,1HN,1HT,1H ,
21H,,1H ,1HL,1HI,1HM,1HI,1HT,1H /
DATA ITTY,JTTY/5,5/
DATA MAXBFR,MAXOUT/60,60/
DATA IGREAT,IBLANK/1H>,1H /
WRITE(JTTY,1)
1 FORMAT(1X,37HPROGRAM TO DEMONSTRATE DASPAN ROUTINE/
124H TEST DAISPN (Y OR N) = ,$)
READ(ITTY,2)IANS
2 FORMAT(1A1)
KONTRL=0
IF(IANS.EQ.1HY)GO TO 4
WRITE(JTTY,3)
3 FORMAT(10H KONTRL = ,$)
READ(ITTY,6)KONTRL
4 IRADIX=10
IF(KONTRL.LT.0)IRADIX=8
WRITE(JTTY,5)
5 FORMAT(10H KONECT = ,$)
READ(ITTY,6)KONECT
6 FORMAT(1I)
MANY=0
C
C READ TEXT TYPED BY USER
7 WRITE(JTTY,8)
8 FORMAT(1X,1H*,$)
READ(ITTY,9)IBUFFR
9 FORMAT(60A1)
LOWBFR=1
10 INIBFR=LOWBFR
IF(IANS.EQ.1HY)
1 CALL DAISPN( KONECT,IBUFFR,MAXBFR,LOWBFR,
2MANY ,KIND ,INIGOT,INCGOT,LMTGOT,INIVAL,INCVAL,
3LMTVAL)
IF(IANS.NE.1HY)
1 CALL DASPAN(KONTRL,KONECT,IBUFFR,MAXBFR,LOWBFR,
2MANY ,KIND ,INIGOT,INCGOT,LMTGOT,INIVAL,INCVAL,
3LMTVAL,VALINI,VALINC,VALLMT)
IF(LOWBFR.LE.INIBFR)GO TO 12
J=LOWBFR-1
FASP, FORTRAN Alphameric Subroutine Package Page 391
DASPAN, Routine to Evaluate Range Specifications
WRITE(JTTY,11)(IBUFFR(I),I=INIBFR,J),IGREAT
11 FORMAT(1X,1H<,100A1)
12 GO TO(13,15,17,19,25,25,25,25,21,23),KIND
C
C REPORT IF OTHER THAN SINGLE NUMBER OR CORRECT SERIES
13 WRITE(JTTY,14)
14 FORMAT(6H EMPTY)
GO TO 7
15 WRITE(JTTY,16)
16 FORMAT(17H END OF STATEMENT)
GO TO 10
17 WRITE(JTTY,18)
18 FORMAT(10H AMPERSAND)
GO TO 7
19 WRITE(JTTY,20)MANY
20 FORMAT(8H MISSING,I3)
GO TO 10
21 WRITE(JTTY,22)
22 FORMAT(15H TOO MANY ITEMS)
GO TO 10
23 WRITE(JTTY,24)
24 FORMAT(18H ILLEGAL DELIMITER)
GO TO 10
C
C REPORT SINGLE NUMBER OR CORRECT SERIES
25 KOUNT=1
JBUFFR(1)=IBLANK
DO 37 IPASS=1,3
GO TO(26,29,32),IPASS
26 IF(INIGOT.EQ.0)GO TO 37
DO 27 I=1,8
KOUNT=KOUNT+1
27 JBUFFR(KOUNT)=LEGEND(I)
IF(KONTRL.GT.0)GO TO 28
NUMBER=INIVAL
GO TO 34
28 VALUE=VALINI
GO TO 36
29 IF(INCGOT.EQ.0)GO TO 37
J=11
IF(KOUNT.GT.1)J=9
DO 30 I=J,20
KOUNT=KOUNT+1
30 JBUFFR(KOUNT)=LEGEND(I)
IF(KONTRL.GT.0)GO TO 31
NUMBER=INCVAL
GO TO 34
31 VALUE=VALINC
GO TO 36
32 IF(LMTGOT.EQ.0)GO TO 37
J=23
IF(KOUNT.GT.1)J=21
DO 33 I=J,28
KOUNT=KOUNT+1
FASP, FORTRAN Alphameric Subroutine Package Page 392
DASPAN, Routine to Evaluate Range Specifications
33 JBUFFR(KOUNT)=LEGEND(I)
IF(KONTRL.GT.0)GO TO 35
NUMBER=LMTVAL
34 LFTCOL=KOUNT
CALL DANUMB(0,NUMBER,IRADIX,JBUFFR,
1KOUNT,LFTCOL,MAXOUT)
GO TO 37
35 VALUE=VALLMT
36 LFTCOL=KOUNT
CALL DARITE(VALUE,-1,0,0,0,
1-3,0,10,-1,-2,6,6,
2-1,0,5,0,0,0,LFTCOL,
3MAXOUT,JBUFFR,KOUNT,IERR)
37 CONTINUE
IF(KIND.EQ.5)WRITE(JTTY,38)MANY,(JBUFFR(I),I=1,KOUNT)
38 FORMAT(8H NUMBER ,I3,1H ,100A1)
IF(KIND.EQ.6)WRITE(JTTY,39)MANY,(JBUFFR(I),I=1,KOUNT)
39 FORMAT(8H SERIES ,I3,1H ,100A1)
IF(KIND.EQ.7)WRITE(JTTY,40)MANY,(JBUFFR(I),I=1,KOUNT)
40 FORMAT(8H SERIES ,I3,1H/,100A1)
IF(KIND.EQ.8)WRITE(JTTY,41)MANY,(JBUFFR(I),I=1,KOUNT)
41 FORMAT(8H SERIES ,I3,1H*,100A1)
GO TO 10
END
Typical Dialog Between DASPAN Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
PROGRAM TO DEMONSTRATE DASPAN ROUTINE
TEST DAISPN (Y OR N) = N
KONTRL = 4
KONECT = -1
*10-20 30 -40,50 , -60 , 70 , &COMMA ALSO STARTS NEXT LINE
<10>
NUMBER 1 INITIAL 10, LIMIT 10
<-20>
NUMBER 2 INITIAL -20, LIMIT -20
< 30>
NUMBER 3 INITIAL 30, LIMIT 30
< -40>
NUMBER 4 INITIAL -40, LIMIT -40
<,50>
NUMBER 5 INITIAL 50, LIMIT 50
< , -60>
NUMBER 6 INITIAL -60, LIMIT -60
< , 70>
NUMBER 7 INITIAL 70, LIMIT 70
< , &COMMA ALSO STARTS NEXT LINE >
AMPERSAND
*,10/20-30/40 -50/60 , -70/80,&COMMA ONLY BEFORE AMPERSAND
MISSING 8
<,10/20>
FASP, FORTRAN Alphameric Subroutine Package Page 393
DASPAN, Routine to Evaluate Range Specifications
SERIES 9/ INITIAL 10, LIMIT 20
<-30/40>
SERIES 10/ INITIAL -30, LIMIT 40
< -50/60>
SERIES 11/ INITIAL -50, LIMIT 60
< , -70/80>
SERIES 12/ INITIAL -70, LIMIT 80
<,&COMMA ONLY BEFORE AMPERSAND >
AMPERSAND
*3*-16.01 , , //77K ; , /87.2/ &COMMA ON NEXT LINE
<3*-16.01>
SERIES 13* INITIAL 3, LIMIT -16.01
< , >
MISSING 14
<, //77K>
SERIES 15/ LIMIT 77000
< ;>
END OF STATEMENT
< >
MISSING 1
<, /87.2/>
SERIES 2/ INCREMENT 87.2
< &COMMA ON NEXT LINE >
AMPERSAND
*, / , // , * &COMMA ON NEITHER LINE
<, />
SERIES 3/
< , //>
SERIES 4/
< , *>
SERIES 5*
< &COMMA ON NEITHER LINE >
AMPERSAND
*/9 , //9 , *9 !COMMENT NOT MARKING A CONTINUATION
</9>
SERIES 6/ LIMIT 9
< , //9>
SERIES 7/ LIMIT 9
< , *9>
SERIES 8* LIMIT 9
< !COMMENT NOT MARKING A CONTINUATION >
EMPTY
*/2/
</2/>
SERIES 1/ INCREMENT 2
< >
EMPTY
*/2/9
</2/9>
SERIES 1/ INCREMENT 2, LIMIT 9
< >
EMPTY
*5/ , 5// , 5*
<5/>
FASP, FORTRAN Alphameric Subroutine Package Page 394
DASPAN, Routine to Evaluate Range Specifications
SERIES 1/ INITIAL 5
< , 5//>
SERIES 2/ INITIAL 5
< , 5*>
SERIES 3* INITIAL 5
< >
EMPTY
* 5
< 5>
NUMBER 1 INITIAL 5, LIMIT 5
< >
EMPTY
*5/9 , 5//9 , 5*9
<5/9>
SERIES 1/ INITIAL 5, LIMIT 9
< , 5//9>
SERIES 2/ INITIAL 5, LIMIT 9
< , 5*9>
SERIES 3* INITIAL 5, LIMIT 9
< >
EMPTY
*5/2/
<5/2/>
SERIES 1/ INITIAL 5, INCREMENT 2
< >
EMPTY
*5/2/9
<5/2/9>
SERIES 1/ INITIAL 5, INCREMENT 2, LIMIT 9
< >
EMPTY
*/// , 5///9 , /2/2/ , 5//2/ , ** , 5*2*9 , 5*2/9 , 5/2*9
<///>
TOO MANY ITEMS
< , 5///9>
TOO MANY ITEMS
< , /2/2/>
TOO MANY ITEMS
< , 5//2/>
TOO MANY ITEMS
< , **>
TOO MANY ITEMS
< , 5*2*9>
TOO MANY ITEMS
< , 5*2/9>
TOO MANY ITEMS
< , 5/2*9>
TOO MANY ITEMS
< >
EMPTY
FASP, FORTRAN Alphameric Subroutine Package Page 395
DASWAP, Swaps Adjacent Regions in Buffer
DDDDDDD AAA SSSSSSS WWW WWW AAA PPPPPPPPP
DDD DDD AAA AAA SSS WWW WWW AAA AAA PPP PPP
DDD DDD AAA AAA SSSSSSS WWW WW WWW AAA AAA PPPPPPPPP
DDD DDD AAAAAAAAA SSS WWWW WWWW AAAAAAAAA PPP
DDDDDDD AAA AAA SSSSSSSS WWW WWW AAA AAA PPP
DASWAP, Swaps Adjacent Regions in Buffer
------ ----- -------- ------- -- ------
DASWAP is a FORTRAN routine which interchanges 2 adjacent
regions in the input array without the use of a temporary
storage array. The swap is performed by moving the values
directly to the locations which they are to occupy in the
result. This operation is often useful for the manipulation
of characters read 1 to a computer storage location by a
multiple of an A1 format or else defined by other routines
within this package.
The DASWAP Argument List
--- ------ -------- ----
The argument list of routine DASWAP is
SUBROUTINE DASWAP(IARRAY,LOW,MID,MAX)
with the associated DIMENSION statement
DIMENSION IARRAY(MAX)
IARRAY is used for both input to and output from this
routine. The rest of the arguments are used for input only
and are returned unchanged.
IARRAY = the array which is input containing in locations
IARRAY(LOW) through IARRAY(MID) the lower section
of text which is to be swapped with the upper
section of text in locations IARRAY(MID+1) through
IARRAY(MAX). Each IARRAY location contains a
single character as though read by an A1 format or
defined by a 1H field.
LOW = the subscript within the IARRAY array of the lowest
location within the lower region.
MID = the subscript within the IARRAY array of the
highest location within the lower region.
MAX = the subscript within the IARRAY array of the
highest location within the higher region.
FASP, FORTRAN Alphameric Subroutine Package Page 396
DASWAP, Swaps Adjacent Regions in Buffer
Demonstration of Manner in which Regions are Swapped
------------- -- ------ -- ----- ------- --- -------
A B C D 1 2 Illustration at left demonstrates how regions
. <-------< of 4 items, letters A through D, and of 2
. >---> . . items, digits 1 and 2, are swapped. 2 is
. . . >---> moved into location holding B, which is moved
<-------< . into location holding D, which is finally
>---> . . . moved into location originally holding 2.
. . >---> . This swapped half of items so same sequence
1 2 A B C D is performed with items at 1, A and C.
On this and the following page are shown all pairs of
regions for which the sums of the region sizes vary from 2
through 8, except for those cases in which 1 region has a
zero length.
sum=2 sum=3 sum=4
A 1 A 1 2 A B 1 A 1 2 3 A B 1 2 A B C 1
<-< . <-< <---< . . <-< . <---< <-----<
>-> <-< . >-> . . <-< . . >---> >-> . .
1 A >---> . >-> <-< . . <---< . . >-> .
1 2 A 1 A B >-----> >---> . . . >->
1 2 3 A 1 2 A B 1 A B C
sum=5
A 1 2 3 4 A B 1 2 3 A B C 1 2 A B C D 1
. . . <-< . . <---< . <-----< <-------<
. . <-< . <---< . . . >---> . >-> . . .
. <-< . . >-----> . <-----< . . >-> . .
<-< . . . . <---< . >---> . . . . >-> .
>-------> . >-----> . . >---> . . . >->
1 2 3 4 A 1 2 3 A B 1 2 A B C 1 A B C D
sum=6
A 1 2 3 4 5 A B 1 2 3 4 A B C 1 2 3 A B C D 1 2
. . . . <-< . . . <---< . . <-----< . <-------<
. . . <-< . . <---< . . . . >-----> . >---> . .
. . <-< . . . >-------> . <-----< . . . . >--->
. <-< . . . . . <---< . . >-----> . <-------< .
<-< . . . . <---< . . . <-----< . . >---> . . .
>---------> >-------> . >-----> . . . . >---> .
1 2 3 4 5 A 1 2 3 4 A B 1 2 3 A B C 1 2 A B C D
A B C D E 1
<---------<
>-> . . . .
. >-> . . .
. . >-> . .
. . . >-> .
. . . . >->
1 A B C D E
FASP, FORTRAN Alphameric Subroutine Package Page 397
DASWAP, Swaps Adjacent Regions in Buffer
Illustrations shown below demonstrate sums of 7 and 8.
A 1 2 3 4 5 6 A B 1 2 3 4 5 A B C 1 2 3 4
. . . . . <-< . . . . <---< . . . <-----<
. . . . <-< . . . <---< . . <-----< . . .
. . . <-< . . <---< . . . . >-------> . .
. . <-< . . . >---------> . . <-----< . .
. <-< . . . . . . . <---< . . >-------> .
<-< . . . . . . <---< . . . . . <-----< .
>-----------> . >---------> . . >------->
1 2 3 4 5 6 A 1 2 3 4 5 A B 1 2 3 4 A B C
A B C D 1 2 3 A B C D E 1 2 A B C D E F 1
. . <-------< . <---------< <-----------<
. . >-----> . . >---> . . . >-> . . . . .
. <-------< . . . . >---> . . >-> . . . .
. >-----> . . <---------< . . . >-> . . .
<-------< . . >---> . . . . . . . >-> . .
>-----> . . . . . >---> . . . . . . >-> .
. . . >-----> . . . . >---> . . . . . >->
1 2 3 A B C D 1 2 A B C D E 1 A B C D E F
A 1 2 3 4 5 6 7 A B 1 2 3 4 5 6 A B C 1 2 3 4 5
. . . . . . <-< . . . . . <---< . . . . <-----<
. . . . . <-< . . . . <---< . . . <-----< . . .
. . . . <-< . . . <---< . . . . . >---------> .
. . . <-< . . . . >-----------> . . . <-----< .
. . <-< . . . . . . . . <---< . <-----< . . . .
. <-< . . . . . . . <---< . . . >---------> . .
<-< . . . . . . <---< . . . . . . . <-----< . .
>-------------> >-----------> . . . >--------->
1 2 3 4 5 6 7 A 1 2 3 4 5 6 A B 1 2 3 4 5 A B C
A B C D 1 2 3 4 A B C D E 1 2 3 A B C D E F 1 2
. . . <-------< . . <---------< . <-----------<
. . . >-------> . . >-----> . . . >---> . . . .
. . <-------< . <---------< . . . . . >---> . .
. . >-------> . >-----> . . . . . . . . . >--->
. <-------< . . . . . >-----> . <-----------< .
. >-------> . . . <---------< . >---> . . . . .
<-------< . . . . >-----> . . . . . >---> . . .
>-------> . . . . . . . >-----> . . . . >---> .
1 2 3 4 A B C D 1 2 3 A B C D E 1 2 A B C D E F
A B C D E F G 1
<-------------<
>-> . . . . . .
. >-> . . . . .
. . >-> . . . .
. . . >-> . . .
. . . . >-> . .
. . . . . >-> .
. . . . . . >->
1 A B C D E F G
FASP, FORTRAN Alphameric Subroutine Package Page 398
DATALL, Time Series Plot Routine for Printer
DDDDD AAA TTTTTTTT AAA LL LL
DD DD AAAA TT AAAA LL LL
DD DD AA AA TT AA AA LL LL
DD DD AA AA TT AA AA LL LL
DD DD AAAAAAA TT AAAAAAA LL LL
DD DD AA AA TT AA AA LL LL
DDDDD AA AA TT AA AA LLLLLLLL LLLLLLLL
DATALL, Time Series Plot Routine for Printer
------ ---- ------ ---- ------- --- -------
DATALL is a FORTRAN subroutine which constructs printable
plots with a vertical axis representing time (or any other
variable which has a constant increment between samples)
extending onto as many lines and pages as are necessary to
represent the data. The points which represent a particular
data item in consecutive time periods can be connected with
line segments to form a curve. A curve does not need to
extend across all time periods. The sections of a curve
which are outside the plot area are not represented. A
maximum of 26 curves are distinguishable by being plotted
with different letters, but there is no limit to the total
number of curves which can be included in the plot of a
particular time period or in the plot of all time periods.
Overlapping sections of curves represented by different
letters are indicated by ampersands.
DATALL is called once for each time period to append a
representation of the data for that time period to the plot
of the data for the previous time period. Arrays input to
DATALL specify the horizontal or non-time coordinate of each
point for the time period, the letters with which these
points are to be plotted, and the letters with which the
points are to be connected with the points for the previous
time period. The segment of the plot representing a single
time period can contain several points and/or lines. It is
optional whether the plot segment representing a time period
is ruled with a grid line and is identified by a scale
number.
DATALL is a relatively short routine which relies upon
DAPLAT for its plotting capabilities. The routines DARITE
and PLTCUT must also be loaded as these are called by
DAPLAT. DAPLAT must not be called by any other program
until the plotting of all time periods has been completed
since DAPLAT has internal storage which is used by DATALL.
If the user's program calls DAPLAT to do other plotting
after the plotting of the data for all time periods has been
completed, then this subsequent call to DAPLAT must use
non-zero values for the arguments MAXWID, MAXHIH, MSHWID and
MSHHIH since the default values of these DAPLAT arguments
are changed by DATALL. The character set used for plots
FASP, FORTRAN Alphameric Subroutine Package Page 399
DATALL, Time Series Plot Routine for Printer
produced by DATALL can be manipulated as described in the
instruction manual for DAPLAT. The maximum plot size
limitation for DAPLAT applies to each time period, not to
the entire plot of all time periods, and so should be of no
concern to the user.
The DATALL Argument List
--- ------ -------- ----
The argument list of routine DATALL is
SUBROUTINE DATALL(LSTLIN,MRKLIN,MAXWID,IFLTTR,LETTER,
1 IFCNCT,KONECT,XPOINT,MINSUB,MAXSUB, XLEFT,YVALUE,
2 XRIGHT, IGRID, IEDGE,MARGIN,MSHWID,MSHHIH,LTROFF,
3 LINPRT, IDISK,IRESET,LSTORE,XSTORE)
with the associated DIMENSION statement
DIMENSION LETTER(MAXSUB),KONECT(MAXSUB),
1XPOINT(MAXSUB),LSTORE(MAXSUB),XSTORE(MAXSUB)
Arguments Used to Input Values to DATALL
--------- ---- -- ----- ------ -- ------
The following arguments are used to input values to DATALL.
LSTLIN = 0, the current time period is the final time period
to be represented in the plot. Scale numbers
ranging in value from that of XLEFT through that of
XRIGHT are to be written below the representation
of the current time period, and the internal
storage in DATALL and DAPLAT is then to be cleared.
= greater than zero, the current time period is not
the final time period in the plot. Subsequent
calls to DATALL will add additional segments to the
current plot. The value of LSTLIN is ignored other
than to determine whether it is greater than zero.
If the main program knows the total number of time
periods, it can count LSTLIN down to zero. If the
main program does not know the total number of time
periods, it is sufficient to set LSTLIN to 1 until
the final time period.
MRKLIN = 0, the current value of YVALUE can be printed to
the left of the bottom line in the representation
of the current time period and a line can be ruled
through this bottom line. Whether the value of
YVALUE will actually be printed to the left of the
the bottom line will depend upon the value of
IEDGE. Whether a line will actually be ruled
through the bottom line will depend upon the values
FASP, FORTRAN Alphameric Subroutine Package Page 400
DATALL, Time Series Plot Routine for Printer
of IGRID and of IEDGE, and upon whether the current
line is at either the top or the bottom of the
entire plot.
= greater than zero, a scale number is not be be
placed beside the bottom line in the representation
of the current time period and this bottom line
cannot be ruled as a grid line. The value of
MRKLIN is ignored other than to determine whether
it is greater than zero. If the value of YVALUE is
to be printed beside some but not all time periods
(for example, perhaps beside every third time
period), then MRKLIN can be counted down to zero by
the main program, and when equal to zero can be
reset back to its initial value after DATALL has
been called.
MAXWID = width of the plot stated as the number of columns
of characters forming the plotting area upon which
data can be plotted. MAXWID is normally 1 plus a
multiple of MSHWID. The maximum effective value of
MAXWID is 131. It should be noted that an
additional 12 characters along the left side of the
plot are used for the scale numbers and the
carriage control character.
= 0, use the last nonzero value specified for MAXWID
either to DATALL or to DAPLAT as the width of the
plot, or use the value 101 if a nonzero value of
MAXWID has not been specified.
IFLTTR = selects whether the points are to be represented
using alphabetic letters which are identified by
the LETTER array or which are identified by the
locations of the coordinates of the points in the
XPOINT array.
= 0, the LETTER array identifies the characters to be
plotted at the points.
= 1, the points are plotted using the letters having
the same sequence numbers in the alphabet as the
subscripts of the locations in the XPOINT array
containing the coordinates. The point having its
horizontal coordinate in XPOINT(3) would be
represented by the letter C, the 3rd letter in the
alphabet. Points having subscripts greater than 26
are represented by asterisks which will replace any
other characters already in the plot at these
locations. The contents of the LETTER array are
ignored.
LETTER = an array identifying the letters to be used to
represent the points having their horizontal
coordinates in the XPOINT array locations having
the same subscripts as the LETTER array. LETTER
array values of -1 indicate that the corresponding
points are not to be plotted. LETTER array values
FASP, FORTRAN Alphameric Subroutine Package Page 401
DATALL, Time Series Plot Routine for Printer
of zero indicate that the corresponding points are
to be plotted with asterisks. Values greater than
zero are the sequence numbers within the alphabet
of the letters to be used to represent the points.
If an alphabetic letter and an asterisk selected by
a zero value in the LETTER array are to occupy the
same printing character position on the plot, then
the letter appears. If different letters are to
occupy the same printing character position on the
plot, then an ampersand appears instead. Values in
the LETTER array which are greater than 26 select
asterisks which will replace any other characters
already in the plot at these locations. If IFLTTR
is non-zero, then the LETTER array is not used and
need not be dimensioned.
IFCNCT = selects whether the characters to be plotted along
the line segments connecting points are to be
identified by the KONECT array or are to be the
same as those used to represent the points.
= -1, points in the current time period are not to be
connected to the points of the previous time
period. The contents of the KONECT array are
ignored.
= 0, the KONECT array identifies the characters to be
plotted along the line segments.
= 1, the line segments are to be formed of the same
characters as are used to plot the points in the
current period. The contents of the KONECT array
are ignored.
KONECT = an array identifying the characters with which to
connect points of the current time period with
points of the previous time period which had their
coordinates in XPOINT array locations with the same
subscripts. If XPOINT array locations having the
same subscripts did not specify coordinates to be
plotted both in the current time period and in the
previous time period, then the value in the KONECT
array is ignored. Letters are selected by the
KONECT array in the same manner as by the LETTER
array, with the exception that a KONECT array value
of -1 causes the corresponding points, if any, to
not be connected. If IFCNCT is non-zero, then the
KONECT array is not used and need not be
dimensioned.
XPOINT = an array containing the horizontal or non-time
coordinates of the points to be plotted for the
current time period. The coordinate system used
for the XPOINT array must be the same as that used
for the arguments XLEFT and XRIGHT which select the
coordinates to be placed at the left edge and at
the right edge of the plot respectively. Only
FASP, FORTRAN Alphameric Subroutine Package Page 402
DATALL, Time Series Plot Routine for Printer
values of LETTER, KONECT and XPOINT having
subscripts in the range starting with MINSUB and
extending through MAXSUB are used.
MINSUB = subscript of the LETTER, KONECT and XPOINT array
locations containing the information about the
first point to be plotted. The subscript ranges
can vary from one time period to the next, and need
not even overlap.
MAXSUB = subscript of the LETTER, KONECT and XPOINT array
locations containing the information about the
final point to be plotted. The subscript ranges
can vary from one time period to the next, and need
not even overlap. If no data is to be plotted for
the current time period, then MAXSUB can be less
than MINSUB. It should be noted that empty time
periods at the start of the time series plot are
discarded, but that once a non-empty time period
has been encountered, then all remaining time
periods are plotted whether empty or not. This
does not, of course, insure that something will
actually be plotted in the first time period shown,
since all of the points might be outside the window
defined by XLEFT and XRIGHT, but only that an
attempt is made to plot something in the first time
period shown. The arrays XSTORE and LSTORE must
also be dimensioned to at least the maximum value
of MAXSUB.
XLEFT = the horizontal or non-time data coordinate to be
placed in the center of the character column at the
left edge of the plot. If a scale number is
printed below the left column of the plot, then
this scale number will have a value equal to that
of XLEFT. Only the portion of the curves in the
XPOINT array having values in the range
XLEFT-((XRIGHT-XLEFT)/(2*(MAXWID-1))) to XRIGHT+
((XRIGHT-XLEFT)/(2*(MAXWID-1))) will be shown on
the plot. If a line segment crosses the plot, then
the portion of the line segment which is outside
the plot area will not be represented. The data
coordinates can either increase or decrease from
left to right.
YVALUE = number identifying the current time period. This
value must change from one call to DATALL to the
next, but can either increase or decrease. If the
current value of MRKLIN is zero, then the value of
YVALUE can be printed to the left of the lowest
line of the current plot segment.
XRIGHT = the horizontal or non-time data coordinate to be
placed in the center of the character column at the
FASP, FORTRAN Alphameric Subroutine Package Page 403
DATALL, Time Series Plot Routine for Printer
right edge of the plot. If a scale number is
printed below the right column of the plot, then
this scale number will have a value equal to that
of XRIGHT.
IGRID = 0, the plot will show an internal grid. This
internal grid will be ruled vertically every MSHWID
characters and, if MRKLIN is equal to zero,
horizontally every MSHHIH lines.
= 1, the plot will show the intersections of the
lines of an internal grid, but will not show the
grid lines themselves. The internal grid would, if
shown, have vertical lines every MSHWID characters
across the width of the plot and would, if MRKLIN
is equal to zero, have horizontal lines on the
bottom line of each plot segment. Each plot
segment consists of MSHHIH lines, so grid
intersections can be shown every MSHHIH lines if
MRKLIN is always zero.
= 2, the plot will not show an internal grid.
= greater than 2, the decimal digits forming the
value of IGRID select modification of the grid
format as described in the documentation of the
DAPLAT routine.
IEDGE = place numbers both to the left of and below the
plot to identify the coordinate ranges.
= 1, place numbers to the left of the plot, but do
not place numbers below the plot.
= 2, place numbers below the plot, but not to the
left of the plot. Unless increased by the MARGIN
argument, the distance between the carriage control
character in column 1 and the left edge of the plot
will be just large enough to allow a scale number
immediately below the left edge of the plot.
= 3, do not place numbers either to the left of or
below the plot. Unless prevented by the MARGIN
argument, no characters will appear between the
carriage control character in column 1 and the left
edge of the plot.
MARGIN = the lower limit to the number of characters which
must appear in the output between the carriage
control character in column 1 and the left edge of
the plot. MARGIN is used to force the plot to
remain a fixed distance from the carriage control
character in column 1 even if a nonzero value of
IEDGE has deselected scale numbers. MARGIN is
assumed to be at least 11 if IEDGE is zero or 1.
MSHWID = width of the grid divisions stated as the number of
columns of characters. MSHWID=10 would give the
vertical grid lines every 10 characters across the
width of the plot area.
FASP, FORTRAN Alphameric Subroutine Package Page 404
DATALL, Time Series Plot Routine for Printer
= 0, use the last nonzero value specified for MSHWID
either to DATALL or to DAPLAT as the grid division
width, or use the value 10 if a nonzero value of
MSHWID has not been specified.
MSHHIH = the number of lines of printing to be included in
the plot segment representing the current time
period. If the points for successive time periods
are being connected by line segments, then these
line segments will extend across these lines of
printing. The points themselves are represented in
the bottom line of the plot segment.
= 0, assume that MSHHIH=1 is meant. Each time period
will be represented by a single line in the time
series plot.
LTROFF = number of columns of characters by which the
leftmost grid line is offset from the left border
of the plot. LTROFF can be in the range zero up to
but not including MSHWID. If LTROFF is negative,
then it is assumed to have the value MSHWID+LTROFF
instead. If LTROFF is nonzero, then the left
border of the plot will be ruled with exclamation
points. The right border is similarly ruled if it
does not bear a grid line.
LINPRT = -1, do not include a carriage control character to
the left of each line of the plot. Since the minus
sign of a negative scale number can then appear in
column 1, the resulting output must not be treated
as though the left column contains carriage control
characters.
= 0, the plot will be viewed by the user on a
terminal, either typed directly with IDISK being
given the terminal unit number, or typed by the
user after this routine has written the plot into a
file on the unit the number of which is contained
in IDISK. A blank or space will be used as
carriage control character to give single spacing.
= 1, the plot will be printed on the line printer by
the user after the program has written the plot
into a file. An asterisk will be used as carriage
control character to give single spacing with
suppression of skipping extra lines at the page
boundaries. On the PDP-10, an asterisk as the
carriage control character gives overprinting on
the terminal as opposed to single spacing.
IDISK = the unit number of the device onto which the plots
are to be written. This routine will only generate
the plot. It is the responsibility of the calling
program to open the output file and to write the
captions, the form feeds and/or the separating
lines.
FASP, FORTRAN Alphameric Subroutine Package Page 405
DATALL, Time Series Plot Routine for Printer
Argument Used Initially for Input, but then Returned Changed
-------- ---- --------- --- ----- --- ---- -------- -------
The following argument must be defined by the calling
program before this routine is first called. This argument
is returned by this routine set to -1, and this -1 value
should be sent unchanged to any subsequent calls to this
routine.
IRESET = used to identify the first call to DATALL so that
storage inside DATALL can be properly initialized.
However, unless DATALL is used upon a computer
which does not allow the testing of the value of a
variable which has not yet been defined, having
IRESET always set to -1 will produce the expected
results since one of the variables set within the
routine is tested at the start of the routine to
determine whether the internal storage has been
initialized.
= -1, this is not the first call to DATALL.
= 0, DAPLAT has already been called, but DATALL has
not previously been called. IRESET is returned set
to -1.
= 1 (or greater), neither DAPLAT nor DATALL have
previously been called. IRESET values greater than
1, as described in the documentation of the DAPLAT
argument also named IRESET, cause DAPLAT to
preserve selected values in its own internal
storage which have been initialized instead by the
calling program. IRESET is returned set to -1.
Arguments Used Only by DATALL For Intermediate Data Storage
--------- ---- ---- -- ------ --- ------------ ---- -------
The values initially in the array arguments LSTORE and
XSTORE are ignored and are destroyed. These arrays are used
by DATALL to store the numbers identifying the letters used
to plot the points and to store the horizontal coordinates
of these points so that the subsequent call to DATALL can
extend line segments from the points of the time period
previous to it. The calling program must not modify the
contents of these arrays while a time series plot is being
generated. Both arrays must be dimensioned to at least the
maximum value of MAXSUB. This array space can be used for
other purposes by the calling program after DATALL has been
called with LSTLIN equal to zero.
LSTORE = used to store the contents of the LETTER array.
The dimension of the LSTORE array must at least be
equal to the maximum value of MAXSUB encountered
while IFLTTR is zero. If IFLTTR is always 1, then
LSTORE is never used and need not be dimensioned.
XSTORE = used to store the contents of the XPOINT array.
FASP, FORTRAN Alphameric Subroutine Package Page 406
DATALL, Time Series Plot Routine for Printer
an Example of the Use of DATALL
-- ------- -- --- --- -- ------
As an example of the use of this routine, the following
program generated the plots shown on the following pages.
DIMENSION XMATRX(11,3),XPOINT(3),LETTER(3),
1KONECT(3),XSTORE(3),LSTORE(3)
DATA ((XMATRX(I,J),I=1,11),J=1,3)/
112.,14.,15.,16.,16.,16.,15.,14.,14.,11.,10.,
214.,16.,20.,20.,24.,20.,18.,17., 0., 0., 0.,
3 0., 0.,12.,13.,14.,16.,20.,22.,22.,21.,20./
DATA IDISK/1/
IRESET=1
DO 7 KPASS=1,4
IGRID=(100*KPASS)-99
WRITE(IDISK,1)IGRID
1 FORMAT(7H IGRID=,1I4)
MSHHIH=2
DO 6 JPASS=1,2
DO 4 IPASS=1,4
IF(IPASS.EQ.2)MRKLIN=1
IF(IPASS.EQ.3)MRKLIN=0
LSTLIN=10
IF(IPASS.GE.3)LSTLIN=6
IF(KPASS.GE.2)LSTLIN=2
IF(JPASS.EQ.2)LSTLIN=1
INDEX=0
2 INDEX=INDEX+1
YVALUE=INDEX
IF(IPASS.EQ.1)MRKLIN=1
IF(IPASS.EQ.4)MRKLIN=0
C TRANSFER MATRIX INTO SINGLE DIMENSION ARRAY
C X COORDINATE OF 0 IS TAKEN AS NO POINT
DO 3 I=1,3
XPOINT(I)=XMATRX(INDEX,I)
LETTER(I)=I
IF(XPOINT(I).EQ.0.0)LETTER(I)=-1
3 KONECT(I)=I
C MAXWID=31 IFLTTR=0 IFCNCT=0 MINSUB=1 MAXSUB=3
C XLEFT=10. XRIGHT=25. IGRID=1 IEDGE=300 MARGIN=0
C MSHWID=10 LTROFF=0 LINPRT=1
CALL DATALL(LSTLIN,MRKLIN,31,0,LETTER,
10,KONECT,XPOINT,1,3,10.0,YVALUE,
225.0,IGRID,300,0,10,MSHHIH,0,
31,IDISK,IRESET,LSTORE,XSTORE)
MRKLIN=1-MRKLIN
LSTLIN=LSTLIN-1
IF(LSTLIN.GE.0)GO TO 2
4 WRITE(IDISK,5)
5 FORMAT(1X)
6 MSHHIH=1
7 CONTINUE
STOP
FASP, FORTRAN Alphameric Subroutine Package Page 407
DATALL, Time Series Plot Routine for Printer
END
IGRID= 1
* ! A B !
* ! AA BB !
* ! AA BBB !
* ! A BBBB !
* ! C A BBB !
* ! C A B !
* ! C A BB !
* ! C A BBBB !
* ! C A BBB !
* ! CC A BBBB !
* ! C&C BBB !
* ! A CCCCBB !
* ! A BBCCC !
* ! A B CC !
* ! A B CC !
* ! A C !
* ! AA C !
* ! AAA C !
* ! AA C !
* !A C !
* A C !
* ! ! ! !
* 10 15 20 25
* ! A B !
* ! AA BB !
* 2 -+ AA +BBB + +
* ! A BBBB !
* ! C A BBB !
* ! C A B !
* 4 -+ C + A BB +
* ! C A BBBB !
* ! C A BBB !
* ! CC A BBBB !
* 6 -+ +C&C BBB +
* ! A CCCCBB !
* ! A BBCCC !
* ! A B CC !
* 8 -+ A + B + CC +
* ! A C !
* ! AA C !
* ! AAA C !
* 10 -+ AA + + C +
* !A C !
* A C !
* ! ! ! !
* 10 15 20 25
FASP, FORTRAN Alphameric Subroutine Package Page 408
DATALL, Time Series Plot Routine for Printer
* 1 -+---A---B-+---------+---------+
* ! AA BB !
* ! AA BBB !
* ! A BBBB !
* 3 -+ C A BBB +
* ! C A B !
* ! C A BB !
* ! C A BBBB !
* 5 -+ C + A + BBB +
* ! CC A BBBB !
* ! C&C BBB !
* ! A CCCCBB !
* 7 -+---------A-----BBCCC---------+
* ! ! ! !
* 10 15 20 25
* 1 -+---A---B-+---------+---------+
* ! AA BB !
* 2 -+ AA +BBB + +
* ! A BBBB !
* 3 -+ C A BBB +
* ! C A B !
* 4 -+ C + A BB +
* ! C A BBBB !
* 5 -+ C + A + BBB +
* ! CC A BBBB !
* 6 -+ +C&C BBB +
* ! A CCCCBB !
* 7 -+---------A-----BBCCC---------+
* ! ! ! !
* 10 15 20 25
* ! AA BB !
* ! AAA BBB !
* ! ! ! !
* 10 15 20 25
* ! AA BB !
* 2 -+-----AAA-BBB-------+---------+
* ! ! ! !
* 10 15 20 25
* 1 -+---AA--BB+---------+---------+
* ! AAA BBB !
* ! ! ! !
* 10 15 20 25
* 1 -+---AA--BB+---------+---------+
* 2 -+-----AAA-BBB-------+---------+
* ! ! ! !
* 10 15 20 25
FASP, FORTRAN Alphameric Subroutine Package Page 409
DATALL, Time Series Plot Routine for Printer
IGRID= 101
* ! A B !
* ! AA BB !
* ! AA BBB !
* ! A BBBB !
* ! C A BBB !
* ! ! ! !
* 10 15 20 25
* ! A B !
* ! AA BB !
* 2 -+ AA +BBB + +
* ! A BBBB !
* ! C A BBB !
* ! ! ! !
* 10 15 20 25
* 1 -+ A B + + +
* ! AA BB !
* ! AA BBB !
* ! A BBBB !
* 3 -+---C-----A-------BBB---------+
* ! ! ! !
* 10 15 20 25
* 1 -+ A B + + +
* ! AA BB !
* 2 -+ AA +BBB + +
* ! A BBBB !
* 3 -+---C-----A-------BBB---------+
* ! ! ! !
* 10 15 20 25
* ! AA BB !
* ! AAA BBB !
* ! ! ! !
* 10 15 20 25
* ! AA BB !
* 2 -+-----AAA-BBB-------+---------+
* ! ! ! !
* 10 15 20 25
* 1 -+ AA BB+ + +
* ! AAA BBB !
* ! ! ! !
* 10 15 20 25
* 1 -+ AA BB+ + +
* 2 -+-----AAA-BBB-------+---------+
* ! ! ! !
* 10 15 20 25
FASP, FORTRAN Alphameric Subroutine Package Page 410
DATALL, Time Series Plot Routine for Printer
IGRID= 201
* ! A B !
* ! AA BB !
* ! AA BBB !
* ! A BBBB !
* ! C A BBB !
* ! ! ! !
* 10 15 20 25
* ! A B !
* ! AA BB !
* 2 -+ AA +BBB + +
* ! A BBBB !
* ! C A BBB !
* ! ! ! !
* 10 15 20 25
* 1 -+---A---B-+---------+---------+
* ! AA BB !
* ! AA BBB !
* ! A BBBB !
* 3 -+ C A BBB +
* ! ! ! !
* 10 15 20 25
* 1 -+---A---B-+---------+---------+
* ! AA BB !
* 2 -+ AA +BBB + +
* ! A BBBB !
* 3 -+ C A BBB +
* ! ! ! !
* 10 15 20 25
* ! AA BB !
* ! AAA BBB !
* ! ! ! !
* 10 15 20 25
* ! AA BB !
* 2 -+ AAA BBB + +
* ! ! ! !
* 10 15 20 25
* 1 -+---AA--BB+---------+---------+
* ! AAA BBB !
* ! ! ! !
* 10 15 20 25
* 1 -+---AA--BB+---------+---------+
* 2 -+ AAA BBB + +
* ! ! ! !
* 10 15 20 25
FASP, FORTRAN Alphameric Subroutine Package Page 411
DATALL, Time Series Plot Routine for Printer
IGRID= 301
* ! A B !
* ! AA BB !
* ! AA BBB !
* ! A BBBB !
* ! C A BBB !
* ! ! ! !
* 10 15 20 25
* ! A B !
* ! AA BB !
* 2 -+ AA +BBB + +
* ! A BBBB !
* ! C A BBB !
* ! ! ! !
* 10 15 20 25
* 1 -+ A B + + +
* ! AA BB !
* ! AA BBB !
* ! A BBBB !
* 3 -+ C A BBB +
* ! ! ! !
* 10 15 20 25
* 1 -+ A B + + +
* ! AA BB !
* 2 -+ AA +BBB + +
* ! A BBBB !
* 3 -+ C A BBB +
* ! ! ! !
* 10 15 20 25
* ! AA BB !
* ! AAA BBB !
* ! ! ! !
* 10 15 20 25
* ! AA BB !
* 2 -+ AAA BBB + +
* ! ! ! !
* 10 15 20 25
* 1 -+ AA BB+ + +
* ! AAA BBB !
* ! ! ! !
* 10 15 20 25
* 1 -+ AA BB+ + +
* 2 -+ AAA BBB + +
* ! ! ! !
* 10 15 20 25
FASP, FORTRAN Alphameric Subroutine Package Page 412
DATEAM, Evaluates Several Numbers in a Single Line of Text
DDDDD AAA TTTTTTTT EEEEEEEE AAA MM MM
DD DD AAAA TT EE AAAA MMM MMM
DD DD AA AA TT EE AA AA MMMM MMMM
DD DD AA AA TT EEEEE AA AA MM MMMM MM
DD DD AAAAAAA TT EE AAAAAAA MM MM MM
DD DD AA AA TT EE AA AA MM MM
DDDDD AA AA TT EEEEEEEE AA AA MM MM
DATEAM, Evaluates Several Numbers in a Single Line of Text
------ --------- ------- ------- -- - ------ ---- -- ----
A single call to DATEAM interprets an array read by the
calling program with a multiple of an A1 format and returns
all of the values represented in this array. If more values
are found than can be stored in the array provided for
returning these values to the calling program, then DATEAM
can indicate the first character of the first extra number,
or can scan across and possibly count the excess numbers.
Numbers can be separated by spaces, by tab characters and/or
by commas. Excess commas are ignored and do not indicate
either missing or zero values. The evaluation is terminated
when a semicolon is found within the contents of the input
text buffer or else when all of the characters within the
input text buffer have been interpreted. An exclamation
point and any characters to its right are taken to form a
comment and are otherwise ignored. An ampersand and any
characters to its right are similarly ignored, but the
calling program is informed that an ampersand was found so
that the calling program can read new text into the input
buffer before calling this routine again to continue the
evaluation.
The DATEAM Argument List
--- ------ -------- ----
The argument list of routine DATEAM is
SUBROUTINE DATEAM(KONTNU,KONTRL,ITRAIL,NUMMAX,MAXBFR,
1 IBUFFR,LOWBFR,NUMKNT,KIND ,NUMVAL,VALNUM)
with the associated DIMENSION statement
DIMENSION NUMVAL(NUMMAX),VALNUM(NUMMAX),
1IBUFFR(MAXBFR)
The following arguments are used for input only and are
returned unchanged.
KONTNU = -1, if more values are found than can be returned
in the NUMVAL or VALNUM array, then KIND is
returned containing 5 and LOWBFR is returned
FASP, FORTRAN Alphameric Subroutine Package Page 413
DATEAM, Evaluates Several Numbers in a Single Line of Text
pointing to the left character of the first excess
number. If KONTNU contains -1 and too many values
are found, then it is expected that this routine
will be called to continue processing of the
contents of the input text buffer after the calling
program has processed the values returned in the
NUMVAL or VALNUM array and has reset NUMKNT. KIND
cannot be returned containing 5 if KONTNU is
greater than or equal to zero.
= 0, if more values are found than can be returned in
the NUMVAL or VALNUM array, then the excess values
are interpreted and LOWBFR is returned pointing
beyond the final number, but NUMKNT is not
incremented for these excess values and the excess
values are not returned to the calling program.
= 1, if more values are found than can be returned in
the NUMVAL or VALNUM array, then the excess values
are interpreted, LOWBFR is returned pointing beyond
the final number, and NUMKNT is incremented for
each value found, but the excess values are not
returned to the calling program.
KONTRL = if the representation of a number is found, KONTRL
specifies whether the value is to be returned in
the integer array which is named NUMVAL or in the
real array which is named VALNUM. The number can
be typed with a decimal point and/or an exponent
regardless of the value of KONTRL.
= -1, the value is calculated as an octal integer and
is returned in the NUMVAL array. However, the
number following the letter E of an exponent is
evaluated in decimal.
= 0, the value is calculated as a decimal integer and
is returned in the NUMVAL array.
= 1 or greater, the value is returned in the VALNUM
array. If possible, the real number will be
accumulated as an integer, then be converted to
real and shifted as necessary. KONTRL is the
maximum number of digits in the integer.
ITRAIL = selects whether exponents are to be recognized. If
exponents are not to be recognized but an exponent
is found, then the evaluation of the contents of
the input text buffer will be terminated prior to
the exponent and the first character of the
exponent will be treated the same as any other
unknown alphabetic character. When such an unknown
character is found, KIND is returned containing 4
and LOWBFR is returned pointing to the unknown
character.
= -1, exponents expressed in E notation are to be
recognized, but the percent sign and the letters K
and M are to be treated the same as any other
alphabetic characters.
FASP, FORTRAN Alphameric Subroutine Package Page 414
DATEAM, Evaluates Several Numbers in a Single Line of Text
= 0, no exponents are to be recognized. The
evaluation will be terminated prior to percent
signs or to the letters E or K or M.
= 1, percent signs, the letters K and M, and
exponents expressed in E notation are all to be
recognized.
NUMMAX = highest subscript of the NUMVAL or VALNUM array
locations into which can be placed the values
represented by the characters in the IBUFFR array.
The first value found is returned in
NUMVAL(NUMKNT+1) or VALNUM(NUMKNT+1). If the
available portion of the NUMVAL or VALNUM array is
full and if an additional value is encountered,
then KIND is returned set to 5 if KONTNU is -1, or
else the evaluation of additional numbers continues
until a semicolon or the end of line is reached if
KONTNU is greater than or equal to zero. If KONTNU
is input greater than zero, then NUMKNT can be
returned greater than NUMMAX, and the value of
NUMKNT merely indicates the maximum subscript of
the NUMVAL or VALNUM array locations which would
have been used if available but no locations above
NUMVAL(NUMMAX) or VALNUM(NUMMAX) are actually used
by this routine.
MAXBFR = subscript of the IBUFFR array location containing
the rightmost (highest subscript) character in the
line of text being interpreted. MAXBFR would
normally be the dimension of the IBUFFR array.
IBUFFR = the input buffer array containing the characters of
the line of text to be interpreted, one character
per array location, as read by a multiple of an A1
format. The text to be interpreted begins with
IBUFFR(LOWBFR) and extends up to the next unknown
character or through the next semicolon or through
the end of the line if an ampersand or an
exclamation point is found.
The following arguments are used both for input to this
routine and for output to the calling program.
LOWBFR = input containing the subscript of the IBUFFR array
location which contains the leftmost (lowest
subscript) character which is to be interpreted by
this routine. LOWBFR is returned pointing to the
leftmost character not yet identified by this
routine. LOWBFR is returned containing the
subscript of the IBUFFR array location containing
an unknown character (KIND being returned
containing 4) or containing the character to the
right of a semicolon (KIND being returned
containing 2). It is expected that some other
FASP, FORTRAN Alphameric Subroutine Package Page 415
DATEAM, Evaluates Several Numbers in a Single Line of Text
routine will be called to evaluate an unknown
character, but if this routine is instead called to
evaluate the text to the right of the unknown
character then the calling program must first
increment LOWBFR by one. If KONTNU is set to -1
and if more values are found than can be stored in
the available portion of the NUMVAL or VALNUM
array, then LOWBFR is returned containing the
subscript of the IBUFFR array location which
contains the first character of the first value
which could not be stored. If an ampersand or an
exclamation point is found or if all characters in
the input text buffer have been interpreted, then
LOWBFR is returned pointing beyond the right end of
the buffer.
NUMKNT = input containing the subscript of the highest
location in the NUMVAL or VALNUM array which is
currently in use and which must therefore be
returned unchanged. The first value found by this
routine will be stored in NUMVAL(NUMKNT+1) or in
VALNUM(NUMKNT+1). If KONTNU is less than or equal
to zero, or if KONTNU is greater than zero but no
more than NUMMAX-NUMKNT values are found, then
NUMKNT is returned containing the subscript of the
highest location in the NUMVAL or VALNUM array
which was used by this routine for storage of
values represented by the text in the IBUFFR array.
If KONTNU is greater than zero, but more than
NUMMAX-NUMKNT values are found, then the locations
above NUMVAL(NUMMAX) or VALNUM(NUMMAX) are returned
unchanged, but NUMKNT is returned incremented as
though these excess values had been stored.
The following arguments are used only for output to the
calling program. Their input values are ignored. However,
the portion of the NUMVAL or VALNUM array having subscripts
less than or equal to the input value of NUMKNT and the
portion having subscripts greater than NUMMAX are returned
unchanged.
KIND = returned describing the reason for transfer of
control back to the calling program. KIND does not
indicate whether any values have been stored in the
NUMVAL or VALNUM array. The calling program must
compare the returned value of NUMKNT against its
original value to determine whether any values were
found by this routine. If KONTNU is greater than
zero, then the returned value of NUMKNT must
similarly be tested against NUMMAX to determine
whether any excess values were found but not
returned.
= 1, all characters currently within the IBUFFR array
have been interpreted. If an exclamation point was
FASP, FORTRAN Alphameric Subroutine Package Page 416
DATEAM, Evaluates Several Numbers in a Single Line of Text
found, then the characters to the right of the
exclamation point have been ignored and LOWBFR is
returned containing MAXBFR+1.
= 2, a semicolon was found. LOWBFR is returned
pointing to the character to the right of the
semicolon. If semicolons are to be considered as
equivalent to spaces, then the calling program
should again call this routine without first
changing the values of any of the arguments.
= 3, an ampersand was found. The characters to the
right of the ampersand have been ignored and LOWBFR
is returned containing MAXBFR+1. If the ampersand
indicates that text representing additional values
is to be read by the calling program, then LOWBFR
should be reset to point to the start of the new
text before this routine is called again.
= 4, an unknown character was found. LOWBFR is
returned containing the subscript of the IBUFFR
array location containing this unknown character.
If the unknown character is to be considered as
equivalent to a space, then LOWBFR must be
incremented by one before this routine is called
again.
= 5, KONTNU contains -1 and a value was found which
could not be stored in the available portion of the
NUMVAL or VALNUM array. LOWBFR is returned
pointing to the leftmost character in the
representation of the number. The calling program
must supply additional space in the NUMVAL or
VALNUM array or else must reset KONTNU to be zero
or greater before again calling this routine to
process the remaining text in the IBUFFR array.
NUMVAL = array into which are stored the values represented
by the text in the IBUFFR array if KONTRL is less
than or equal to zero. The locations starting with
NUMVAL(NUMKNT+1) and extending through
NUMVAL(NUMMAX) can be used for returning integer
values to the calling program. NUMKNT is returned
pointing to the highest location in the NUMVAL
array which is used to return these values, or, if
KONTNU is greater than zero, which would be used if
available.
VALNUM = array into which are stored the values represented
by the text in the IBUFFR array if KONTRL is
greater than zero. The locations starting with
VALNUM(NUMKNT+1) and extending through
VALNUM(NUMMAX) can be used for returning real
values to the calling program. NUMKNT is returned
pointing to the highest location in the VALNUM
array which is used to return these values, or, if
KONTNU is greater than zero, which would be used if
available.
FASP, FORTRAN Alphameric Subroutine Package Page 417
DATEAM, Evaluates Several Numbers in a Single Line of Text
An Example of the Use of DATEAM
-- ------- -- --- --- -- ------
The following program calls DATEAM to store values into
NUMVAL(11) through NUMVAL(20) or into VALNUM(11) through
VALNUM(20). The user is asked to supply the values of the
arguments KONTNU, KONTRL and ITRAIL, then is prompted with
an asterisk to type the first line of text and with an
ampersand for each continuation line if any.
DIMENSION NUMVAL(20),VALNUM(20),IBUFFR(60)
DATA ITTY,JTTY,NUMLOW,NUMMAX,MAXBFR/5,5,10,20,60/
1 WRITE(ITTY,2)
2 FORMAT(24H KONTNU,KONTRL,ITRAIL = ,$)
READ(JTTY,3)KONTNU,KONTRL,ITRAIL
3 FORMAT(3I)
C
C OBTAIN NEXT LINE OF TEXT TO BE EVALUATED
WRITE(ITTY,4)
4 FORMAT(2H *,$)
NUMKNT=NUMLOW
GO TO 7
5 WRITE(ITTY,6)
6 FORMAT(2H &,$)
7 READ(JTTY,8)IBUFFR
8 FORMAT(60A1)
LOWBFR=1
C
C SEARCH FOR VALUES
9 CALL DATEAM(KONTNU,KONTRL,ITRAIL,NUMMAX,MAXBFR,
1IBUFFR,LOWBFR,NUMKNT,KIND,NUMVAL,VALNUM)
IFOUND=NUMKNT-NUMLOW
GO TO(10,12,5,20,14),KIND
C
C REPORT RESULTS
10 WRITE(ITTY,11)IFOUND
11 FORMAT(14H END OF LINE,,I3,14H NUMBERS FOUND)
GO TO 16
12 WRITE(ITTY,13)IFOUND
13 FORMAT(14H SEMICOLON, ,I3,14H NUMBERS FOUND)
GO TO 16
14 WRITE(ITTY,15)IFOUND
15 FORMAT(14H OVERFLOW, ,I3,14H NUMBERS FOUND)
16 IF(IFOUND.LE.0)GO TO 19
IF(NUMKNT.GT.NUMMAX)NUMKNT=NUMMAX
J=NUMLOW+1
IF(KONTRL.LE.0)WRITE(ITTY,17)(NUMVAL(I),I=J,NUMKNT)
IF(KONTRL.GT.0)WRITE(ITTY,18)(VALNUM(I),I=J,NUMKNT)
17 FORMAT(1X,5I12)
18 FORMAT(1X,5E12.4)
19 IF(KIND.EQ.1)GO TO 1
NUMKNT=NUMLOW
GO TO 9
20 WRITE(ITTY,21)IBUFFR(LOWBFR)
FASP, FORTRAN Alphameric Subroutine Package Page 418
DATEAM, Evaluates Several Numbers in a Single Line of Text
21 FORMAT(3H ?,1A1,1H?)
LOWBFR=LOWBFR+1
GO TO 9
END
Typical Dialog Between User and DATEAM Demonstration Program
------- ------ ------- ---- --- ------ ------------- -------
KONTNU,KONTRL,ITRAIL = -1 0 1
*1K 2K 3K 4K 5K 6K 7K 8K 9K 10K 11K 12K&MORE THAN 10 VALUES
OVERFLOW, 10 NUMBERS FOUND
1000 2000 3000 4000 5000
6000 7000 8000 9000 10000
&13K;14K 15K 16K;;!2 GROUPS OF 3 FOLLOWED BY 2 EMPTY GROUPS
SEMICOLON, 3 NUMBERS FOUND
11000 12000 13000
SEMICOLON, 3 NUMBERS FOUND
14000 15000 16000
SEMICOLON, 0 NUMBERS FOUND
END OF LINE, 0 NUMBERS FOUND
KONTNU,KONTRL,ITRAIL = 0 1 1
*1.01 2.02 3.03 4.04 5.05 6.06 7.07 8.08 9.09 10.10 11.11&
&12.12 13.13;1 2 3 4 5 6 7 8 9 10 11 12 13;14
SEMICOLON, 10 NUMBERS FOUND
0.1010E+01 0.2020E+01 0.3030E+01 0.4040E+01 0.5050E+01
0.6060E+01 0.7070E+01 0.8080E+01 0.9090E+01 0.1010E+02
SEMICOLON, 10 NUMBERS FOUND
0.1000E+01 0.2000E+01 0.3000E+01 0.4000E+01 0.5000E+01
0.6000E+01 0.7000E+01 0.8000E+01 0.9000E+01 0.1000E+02
END OF LINE, 1 NUMBERS FOUND
0.1400E+02
KONTNU,KONTRL,ITRAIL = 1 1 1
*1.01 2.02 3.03 4.04 5.05 6.06 7.07 8.08 9.09 10.10 11.11&
&12.12 13.13;1 2 3 4 5 6 7 8 9 10 11 12 13;14
SEMICOLON, 13 NUMBERS FOUND
0.1010E+01 0.2020E+01 0.3030E+01 0.4040E+01 0.5050E+01
0.6060E+01 0.7070E+01 0.8080E+01 0.9090E+01 0.1010E+02
SEMICOLON, 13 NUMBERS FOUND
0.1000E+01 0.2000E+01 0.3000E+01 0.4000E+01 0.5000E+01
0.6000E+01 0.7000E+01 0.8000E+01 0.9000E+01 0.1000E+02
END OF LINE, 1 NUMBERS FOUND
0.1400E+02
FASP, FORTRAN Alphameric Subroutine Package Page 419
DATEXT, FORTRAN Routine for Large Printable Characters
DDDDD AAA TTTTTTTT EEEEEEEE XX XX TTTTTTTT
DD DD AAAA TT EE XX XX TT
DD DD AA AA TT EE XXXX TT
DD DD AA AA TT EEEEE XX TT
DD DD AAAAAAA TT EE XXXX TT
DD DD AA AA TT EE XX XX TT
DDDDD AA AA TT EEEEEEEE XX XX TT
DATEXT, FORTRAN Routine for Large Printable Characters
------ ------- ------- --- ----- --------- ----------
DATEXT is a FORTRAN subroutine which enables the calling
program to print large multiple line lettering similar to
that used for the above title. The user's program calls
DATEXT once for each line, specifying the characters to be
represented, and supplying a buffer array to be defined by
DATEXT which can then be printed by the calling program with
a FORTRAN FORMAT statement containing a multiple of an
alphameric A1 field. The spacing between letters can be
adjusted to equalize white space.
DATEXT letters across the width of a page. Each call to
DATEXT generates a portion of all of the characters to be
represented. DATEXT returns the line height of the loaded
font, so the calling program can determine how many times
DATEXT must be called. A second version of the routine,
named DATURN, is also supplied which turns the lettering 90
degrees, lettering from the top to the bottom of the page,
and onto subsequent pages. Each call to DATURN constructs a
small portion of a single character. DATURN signals the
calling program when the final character has been completely
represented. Since neither routine has a argument list
which is a subset of the other, the differences in the
DATURN argument list are described after the DATEXT
description.
These routines must be used with a font created by the
program DAFONT. DAFONT can produce the fonts either as
SUBROUTINEs or as BLOCK DATA routines. The SUBROUTINE forms
of the fonts are supplied but can easily be converted to
BLOCK DATA routines if necessary. The font SUBROUTINEs have
names such as TEXT1 and TEXT2 and were produced when DAFONT
processed the data files with names such as TEXT1.DAT and
TEXT2.DAT. These data files are easily modified by the user
to define new character shapes or to change the shapes
currently available.
Only 1 of the fonts should be called from the user's program
since it is the presence in the user's program of a call to
the font SUBROUTINE to force the loading of the SUBROUTINE
which is important, not the execution of the call. If the
font is loaded as a separate file, rather than in library
search mode, then it need not be called.
FASP, FORTRAN Alphameric Subroutine Package Page 420
DATEXT, FORTRAN Routine for Large Printable Characters
A dollar sign appearing in the text to be represented is
taken to be a control character and is not itself
represented in the output. If the character following the
initial dollar sign is also a dollar sign, then a single
dollar sign is represented. A second character other than a
dollar sign selects some option and neither character is
represented. If the font contains multiple shape
specifications for some or all characters, then a dollar
sign followed by a digit causes the subsequent characters to
be represented by the shape selected by the digit. If $2
appears in the text being represented, then the subsequent
characters would be represented by the second specifications
of their shapes. If the digit following the dollar sign
selects a number greater than the number of shapes for a
particular character, then the last (highest valued) shape
for that character is used. Either $1 or $= selects the
first shape of each of the subsequent characters. $= also
removes any other modifications of the lettering which might
have been selected by previous $-character pairs. Several
$-letter pairs are recognized and are listed below. If a
leading dollar sign is followed by any character other than
those which are listed here, then both the dollar sign and
the following character are ignored.
$H and $V are used to reflect and invert the character
representations. The lettering can be read from the other
side of the paper if the sentence is spelled backwards and
if both $H and $V are applied.
$H (Horizontal reflection) reflects the subsequent
character representations horizontally so that the
normal left edge is at the right. If $H is already in
effect, then a second $H is ignored. A subsequent $=
would return the character representation to normal.
$V (Vertical inversion) inverts the subsequent character
representations vertically so that the normal lower
edge is at the top. If $V is already in effect, then
a second $V is ignored. A subsequent $= would return
the character representation to normal.
$A and $F specify whether the distance between adjacent
character representations is to be kept constant or whether
the distance between character representations is to be
adjusted to approximately equalize white space.
$A (Adjust), provided that the argument named MOVE either
selects that the distance between adjacent character
representations is to be kept constant (MOVE=0) or
else selects that the distance between character
representations is to be adjusted to approximately
equalize white space (MOVE=1), then the distances to
the next character representation and between
subsequent character representations are to be
FASP, FORTRAN Alphameric Subroutine Package Page 421
DATEXT, FORTRAN Routine for Large Printable Characters
adjusted to include approximately equal white space
between the characters. A subsequent $F, or a
subsequent $= if MOVE has the value 0, will cause the
distances between adjacent character representations
to be kept constant. If MOVE has the value 1, then
the inclusion of the $A in the text being represented
is not necessary unless a $F has been encountered. If
MOVE is less than zero, selecting that narrow
characters are centered in a wider field, then both $A
and $F are ignored.
$F (Fixed), provided that the argument named MOVE either
selects that the distance between adjacent character
representations is to be kept constant (MOVE=0) or
else selects that the distance between character
representations is to be adjusted to approximately
equalize white space (MOVE=1), then the distances to
the next character representation and between
subsequent character representations are to be kept
constant. A subsequent $A, or a subsequent $= if MOVE
has the value 1, will cause the distances between
adjacent character representations to be adjusted to
approximately equalize white space. If MOVE has the
value 0, then the inclusion of the $F in the text
being represented is not necessary unless a $A has
been encountered. If MOVE is less than zero,
selecting that narrow characters are centered in a
wider field, then both $A and $F are ignored.
$N and $W modify the number of empty columns (or lines for
routine DATURN) used to represent space characters.
$N (Narrow), space characters are to be represented by
half of the usual number of empty columns (or lines
for routine DATURN) whether or not a preceding $W has
been encountered. If the argument named MOVE has the
value -2, then the width of spaces is to be half of
the width of the widest printing character in the
font. If the argument named MOVE has a value greater
than -2, then the width of spaces is to be half of the
most common width of printing characters in the font.
A subsequent $= would cause subsequent spaces to have
their normal width. A subsequent $W would cause
subsequent spaces to have one and one half times their
normal width whether or not a preceding $N has been
encountered.
$W (Wide), space characters are to be represented by one
and one half times the usual number of empty columns
(or lines for routine DATURN) whether or not a
preceding $N has been encountered. If the argument
named MOVE has the value -2, then the width of spaces
is to be one and one half times the width of the
widest printing character in the font. If the
FASP, FORTRAN Alphameric Subroutine Package Page 422
DATEXT, FORTRAN Routine for Large Printable Characters
argument named MOVE has a value greater than -2, then
the width of spaces is to be one and one half times
the most common width of printing characters in the
font. A subsequent $= would cause subsequent spaces
to have their normal width. A subsequent $N would
cause subsequent spaces to have half of their normal
width whether or not a preceding $W has been
encountered.
$L, $U and $E control the automatic capitalization of the
initial letters of words.
$L (Lower case), if the font describes more than a single
shape for any character, then for each character the
pair of shapes identified by consecutive odd and even
numbers, counting the first shape specified for the
character as shape number one, is to be taken as the
pair of the upper and lower (or lower and upper) case
shapes of the character. Providing than a $U or $L
has not already been encountered in the text being
represented, any printing character which follows a
space or spaces but which is not itself a member of a
$-character command pair is to be represented by the
shape (upper case) which would have been selected for
that character if the $L had not been found, and all
other characters are to be represented by the other
shape (lower case) of the odd-even pair of shapes. If
the $L is within the range of a $U or of another $L,
then the next character which is not itself a member
of a $-character command pair is to be represented by
its lower case shape whether or not it would have
otherwise been represented by its upper case shape.
The range of the $L is terminated if either a $E, or a
$= or a $ followed by a non-zero digit is found.
For example, if the text being represented contains
$3A$LN EXAMPLE $LOF ITS USE
and if the font contains 4 shapes for each of the
alphabetic letters (this would require increasing the
sizes of some of the arrays in DAFONT, DATEXT and
DATURN), then the first letters of the words An,
Example, Its and Use would be represented by their
third shapes in the font and all other letters would
be represented by their fourth shapes in the font.
$U (Upper case), if the font describes more than a single
shape for any character, then for each character the
pair of shapes identified by consecutive odd and even
numbers, counting the first shape specified for the
character as shape number one, is to be taken as the
pair of the upper and lower (or lower and upper) case
shapes of the character. Providing than a $U or $L
FASP, FORTRAN Alphameric Subroutine Package Page 423
DATEXT, FORTRAN Routine for Large Printable Characters
has not already been encountered in the text being
represented, the next printing character which is not
itself a member of a $-character command pair and any
printing character which follows a space or spaces but
is not itself a member of a $-character command pair
is to be represented by the shape (upper case) which
would have been selected for that character if the $U
had not been found, and all other characters are to be
represented by the other shape (lower case) of the
odd-even pair of shapes. If the $U is within the
range of a $L or of another $U, then the next
character which is not itself a member of a
$-character command pair is to be represented by its
upper case shape whether or not it would have
otherwise been represented by its lower case shape.
The range of the $U is terminated if either a $E, or a
$= or a $ followed by a non-zero digit is found.
For example, if the text being represented contains
$4$UTHE F$UA$US$UP PACKAGE
and if the font contains 4 shapes for each of the
alphabetic letters, then the name FASP and the first
letters of the words The and Package would be
represented by their fourth shapes in the font and all
other letters would be represented by their third
shapes in the font.
$E (End case), terminates the range of a $L or $U.
Subsequent characters for which the font describes
several shapes will be represented by the same shapes
as would have been selected before the $L or $U was
encountered. The $E can also be used to capitalize a
complete word. The $E is ignored if neither a $L nor
$U is in effect.
For example, the preceding example could be rewritten
$4$UTHE $EFASP $UPACKAGE
THE DATEXT Argument List
--- ------ -------- ----
The argument list of routine DATEXT is
SUBROUTINE DATEXT(LINE ,JSTIFY,IFILL ,INTRVL,MOVE ,
1 ISPACE,LTTR ,LTRBGN,LTREND,LFTCOL,IWIDTH,MAXBFR,
2 IBUFFR,MAXUSD,MAXLIN,LTRNXT)
with the associated DIMENSION statement
DIMENSION LTTR(LTREND),IBUFFR(MAXBFR)
FASP, FORTRAN Alphameric Subroutine Package Page 424
DATEXT, FORTRAN Routine for Large Printable Characters
The following arguments are used for input only, and are
returned unchanged.
LINE = the line within the letter representation which is
to be placed into the IBUFFR array. Line number 1
is the top line of the representation of the
characters. Characters are a total of MAXLIN (an
argument returned by each call to this routine)
lines high. To print a letter or letters, it is
necessary to call DATEXT MAXLIN times with LINE
being assigned the values 1 through MAXLIN, with
the calling program printing the IBUFFR array after
each return from DATEXT. This allows the insertion
of the constructed letters into other text or other
forms.
= 0, place the highest valued line (that containing
the bottom line of each character) into the IBUFFR
array. The calling program should set the value of
LINE to one less than the returned value of MAXLIN
prior to the subsequent call to this routine.
The manner in which line numbers are used is illustrated by
the following program which generated the large letters used
as the title at the beginning of the documentation of this
routine.
DIMENSION LTTR(6),IBUFFR(60)
DATA LTTR/1HD,1HA,1HT,1HE,1HX,1HT/
CALL TEXT4
LINE=0
1 LINE=LINE+1
C
C JSTIFY = 0 (CENTER LETTER REPRESENTATIONS IN IWIDTH)
C IFILL = 0 (DON'T FILL OUT END WITH SPACES)
C INTRVL = 2 (SPACING BETWEEN LETTER REPRESENTATIONS)
C MOVE = 1 (ADJUST FOR EQUAL WHITE SPACES)
C ISPACE = 0 (REPRESENT INTIAL SPACES IF ANY)
C LTRBGN = 1 (FIRST LETTER TO REPRESENT IS LTTR(1))
C LTREND = 6 (FINAL LETTER TO REPRESENT IS LTTR(6))
C LFTCOL = 0 (START REPRESENTATION IN IBUFFR(1))
C IWIDTH = 60 (WIDTH OF FIELD WHICH CAN USE)
C MAXBFR = 60 (DIMENSION OF IBUFFR ARRAY)
C
CALL DATEXT(LINE,0,0,2,1,
1 0,LTTR,1,6,0,60,60,
1IBUFFR,MAXUSD,MAXLIN,LTRNXT)
IF(MAXUSD.LE.0)GO TO 3
WRITE(1,2)(IBUFFR(I),I=1,MAXUSD)
2 FORMAT(100A1)
IF(LINE.LT.MAXLIN)GO TO 1
3 STOP
END
JSTIFY = -1, left justify the letter representations in a
FASP, FORTRAN Alphameric Subroutine Package Page 425
DATEXT, FORTRAN Routine for Large Printable Characters
field of width IWIDTH.
= 0, center the letter representations in a field of
width IWIDTH.
= 1, right justify the letter representations in a
field of width IWIDTH.
IFILL = 0, if left justifying or centering the letter
representations, do not fill the unused portion of
the field right of the letter representations with
spaces. MAXUSD will be returned pointing to the
right end of the rightmost letter representation.
= 1, if left justifying or centering the letter
representations, do fill the unused portion of the
field right of the letter representations with
spaces. MAXUSD will be returned containing
LFTCOL+IWIDTH or MAXBFR, whichever is the smaller.
INTRVL = the number of space (blank) characters to be
inserted between represented characters. If MOVE
is given the value 1 to select white space
adjustment, then INTRVL is the apparent distance
between the profiles of adjacent characters.
MOVE = -2, characters which are narrower than the widest
character are centered within the width of the
widest character. The width of spaces is also the
width of the widest character. No white space
adjustment of positions is to be made. If MOVE is
either -2 or -1 and if a character is narrow, then
the leading and trailing portions of the width in
which the character is centered will be filled with
spaces regardless of the value of IFILL or of
ISPACE.
= -1, characters which are narrower than most common
width will be centered within the most common
width. No white space adjustment of positions is
to be made. If MOVE is -1 or greater, then the
most common character width is always used as the
width of the representation of the space character.
= 0, normal inter-character spacing is acceptable
without white space adjustment.
= 1, adjust space between characters to equalize
white spaces.
ISPACE = -1, represent both initial and final spaces which
appear in the LTTR array.
= 0, represent initial spaces which appear in LTTR
array. Suppress final spaces which appear in LTTR
array.
= 1, suppress both initial and final spaces which
appear in LTTR array.
LTTR = array containing the letters to be represented, 1
letter per word, as read by multiple of an A1
FASP, FORTRAN Alphameric Subroutine Package Page 426
DATEXT, FORTRAN Routine for Large Printable Characters
format. Since the lettering produced by this
routine is large, rightmost spaces are ignored
unless ISPACE=-1. Note that MAXUSD is returned
equal to LFTCOL if LTTR contains only spaces, and
ISPACE is greater than or equal to zero, and IFILL
is equal to zero.
LTRBGN = subscript of the LTTR array location containing the
first character which is to be represented.
LTREND = subscript of the LTTR array location containing the
final character which is to be represented.
LFTCOL = the subscript of the IBUFFR array location to the
immediate left of the field in which the characters
are to be represented.
IWIDTH = the width of the field into which the letter
representations can be placed. The maximum value
with which MAXUSD can then be returned is
LFTCOL+IWIDTH or MAXBFR, whichever is the smaller.
MAXBFR = the dimension of the IBUFFR array into which the
representation is placed. If LFTCOL+IWIDTH is
greater than MAXBFR, then the effective value of
IWIDTH is reduced to MAXBFR-LFTCOL.
The following arguments are used only for output. Their
input values (with the exception of the portions of the
IBUFFR array which are not needed and so are returned
unchanged) are ignored and destroyed.
IBUFFR = the array into which are to be placed the
representations of the letters in the LTTR array.
MAXUSD = returned by DATEXT containing the subscript of the
IBUFFR array location containing the rightmost
character of the representation. If nothing has
been placed into IBUFFR by the current call to this
routine, then MAXUSD is returned set equal to the
value of LFTCOL.
MAXLIN = returned containing the maximum value which the
line count LINE can attain. This will depend on
which font has been loaded. MAXLIN is returned
containing zero if a font has not been loaded.
LTRNXT = returned containing the subscript within the LTTR
array of the first letter which was not
represented. If all letters requested were
represented, then LTRNXT will be returned equal to
LTREND+1. If IWIDTH is too small for all of the
letters to be represented, then LTRNXT will point
to the first letter which would not fit.
FASP, FORTRAN Alphameric Subroutine Package Page 427
DATEXT, FORTRAN Routine for Large Printable Characters
The lettering samples shown below demonstrate white space
adjustment. The printing characters which represent the
shapes of the letters in the word "FAULT" have been changed
to spaces. The spaces around the letter shapes have been
changed to plus signs where due to the normal shape of the
letters or to X's where added to obtain the desired spacing.
intercharacter spacing 1 with white space adjustment
++++++++++++++++++X++++++++X+++++++++++++++++
+ ++++++ X ++++ X ++++++ +
+ +++++++++++ X ++++ X +++++++++ ++++
+ ++++++++++ + X ++++ X +++++++++ ++++
+ ++++++ ++ X ++++ X +++++++++ ++++
+ ++++++++ X ++++ X +++++++++ ++++
+ +++++++ ++++ X+ ++ +X +++++++++ ++++
+ ++++++ +++++ X++ ++X +++ ++++
++++++++++++++++++X++++++++X+++++++++++++++++
intercharacter spacing 2 with white space adjustment
++++++++++++++++++XX++++++++XX+++++++++++++++++
+ ++++++ XX ++++ XX ++++++ +
+ +++++++++++ XX ++++ XX +++++++++ ++++
+ ++++++++++ + XX ++++ XX +++++++++ ++++
+ ++++++ ++ XX ++++ XX +++++++++ ++++
+ ++++++++ XX ++++ XX +++++++++ ++++
+ +++++++ ++++ XX+ ++ +XX +++++++++ ++++
+ ++++++ +++++ XX++ ++XX +++ ++++
++++++++++++++++++XX++++++++XX+++++++++++++++++
intercharacter spacing 3 with white space adjustment
++++++++++++++++++XXX++++++++XXX+++++++++++++++++
+ ++++++ XXX ++++ XXX ++++++ +
+ +++++++++++ XXX ++++ XXX +++++++++ ++++
+ ++++++++++ + XXX ++++ XXX +++++++++ ++++
+ ++++++ ++ XXX ++++ XXX +++++++++ ++++
+ ++++++++ XXX ++++ XXX +++++++++ ++++
+ +++++++ ++++ XXX+ ++ +XXX +++++++++ ++++
+ ++++++ +++++ XXX++ ++XXX +++ ++++
++++++++++++++++++XXX++++++++XXX+++++++++++++++++
intercharacter spacing 4 with white space adjustment
++++++++++++++++++XXXX++++++++XXXX++++++++X+++++++++
+ ++++++ XXXX ++++ XXXX ++++++X +
+ +++++++++++ XXXX ++++ XXXX ++++++X+++ ++++
+ ++++++++++ + XXXX ++++ XXXX ++++++X+++ ++++
+ ++++++ ++ XXXX ++++ XXXX ++++++X+++ ++++
+ ++++++++ XXXX ++++ XXXX ++++++X+++ ++++
+ +++++++ ++++ XXXX+ ++ +XXXX ++++++X+++ ++++
+ ++++++ +++++ XXXX++ ++XXXX X+++ ++++
++++++++++++++++++XXXX++++++++XXXX++++++++X+++++++++
FASP, FORTRAN Alphameric Subroutine Package Page 428
DATEXT, FORTRAN Routine for Large Printable Characters
A Program Demonstrating the Use of the DATEXT Routine
- ------- ------------- --- --- -- --- ------ -------
The following program attempts to center in a desired column
width the representation of a text string typed by the user.
Narrower intervals are tried if the representation will not
fit with the desired interval. If the representation of the
text string will not fit with an interval of 1, then larger
output widths are tried until either all characters typed by
the user are shown or the entire output buffer is used.
DATA IDISK,ITTY,JTTY,MAXLTR,MAXBFR/1,5,5,20,132/
DIMENSION LTTR(20),IBUFFR(132)
CALL TEXT4
WRITE(JTTY,1)
1 FORMAT(32H CENTER IN WHAT WIDTH (MAX 132) ,$)
READ(ITTY,2)IDEAL
2 FORMAT(I)
WRITE(JTTY,3)
3 FORMAT(33H DESIRED SPACING BETWEEN LETTERS ,$)
READ(ITTY,2)INITAL
4 WRITE(JTTY,5)
5 FORMAT(8H STRING ,$)
READ(ITTY,11)LTTR
LTREND=MAXLTR+1
6 LTREND=LTREND-1
IF(LTREND.EQ.0)GO TO 12
IF(LTTR(LTREND).EQ.1H )GO TO 6
LINE=1
INTRVL=INITAL
IWIDTH=IDEAL
7 CALL DATEXT(LINE,0,0,INTRVL,1,
10,LTTR,1,LTREND,0,IWIDTH,MAXBFR,
2IBUFFR,MAXUSD,MAXLIN,LTRNXT)
IF(LINE.GT.1)GO TO 9
IF(LTRNXT.GT.LTREND)GO TO 9
IF(INTRVL.LE.1)GO TO 8
INTRVL=INTRVL-1
GO TO 7
8 IF(IWIDTH.GE.MAXBFR)GO TO 9
IWIDTH=IWIDTH+1
GO TO 7
9 WRITE(JTTY,10)(IBUFFR(I),I=1,MAXUSD)
10 FORMAT(1X,132A1)
WRITE(IDISK,11)(IBUFFR(I),I=1,MAXUSD)
11 FORMAT(132A1)
LINE=LINE+1
IF(LINE.LE.MAXLIN)GO TO 7
WRITE(JTTY,10)
WRITE(IDISK,10)
GO TO 4
12 STOP
END
FASP, FORTRAN Alphameric Subroutine Package Page 429
DATEXT, FORTRAN Routine for Large Printable Characters
The DATURN Argument List
--- ------ -------- ----
The argument list of routine DATURN is
SUBROUTINE DATURN(INTRVL,MOVE ,ISPACE,LTTR ,LTRBGN,
1 LTREND,LFTCOL,MAXBFR,IBUFFR,MAXUSD,MAGNFY,INISTR,
2 MAXSTR,KIND ,ISTORE)
with the associated DIMENSION statement
DIMENSION LTTR(LTREND),IBUFFR(MAXBFR),ISTORE(MAXSTR)
The DATURN argument list does not include the following
DATEXT arguments
LINE which selects the line number,
JSTIFY which selects the justification,
IFILL which selects whether the buffer is to be filled with
spaces to the right of the lettering,
IWIDTH which specifies the field width,
MAXLIN which returns the maximum line count, and
LTRNXT which identifies the first character which could not
be represented.
Also, since the lettering produced by DATURN tends to be
rather large, rightmost spaces are ignored if found in the
text being represented so that ISPACE values of -1 and 0 are
equivalent. Instead, if two or more parallel strips of
lettering are being generated, then a line of spaces will be
returned if DATURN is called after all of the printing
characters in a particular strip have been represented.
The final 5 DATURN arguments are not included in the DATEXT
argument list. MAGNFY, INISTR and MAXSTR are used only for
input and are returned unchanged. KIND must be set to zero
by the calling program before this routine is first called
to represent a particular line of text, and the value of
KIND which is returned is then supplied to the subsequent
call of this routine which continues the representation of
the same line of text. ISTORE is an array the values of
which need not be defined by the calling program. ISTORE is
used by each call of this routine to transfer information
about the current state of the lettering to the subsequent
call to this routine which is continuing the representation
of the same line of text. These arguments are defined as
follows.
MAGNFY = input containing a magnification factor which is
FASP, FORTRAN Alphameric Subroutine Package Page 430
DATEXT, FORTRAN Routine for Large Printable Characters
applied to the height of the lettering. If MAGNFY
has the value 2, then each character which would be
generated to form the shape is repeated twice.
Since the fonts are designed for a 6 lines per inch
and 10 characters per inch format, the lettering,
when turned 90 degrees, will appear extremely
elongated unless the height is magnified. A MAGNFY
value of 3 would produce approximately normally
proportioned lettering. If each line returned by
DATURN is printed twice, then extremely large
lettering can be produced by setting MAGNFY to 5.
It is of course necessary that the IBUFFR array be
large enough to contain the magnified image of the
lettering.
INISTR = input containing the subscript of the first
location in the ISTORE array which can be used to
transfer information about the current state of the
lettering process to the subsequent call of this
routine which is to continue the representation of
the same line of text.
MAXSTR = input containing the subscript of the final
location in the ISTORE array which can be used to
transfer information about the current state of the
lettering process to the subsequent call of this
routine which is to continue the representation of
the same line of text. At least 18 locations in
the ISTORE array are needed for this purpose, but
it is requested that at least 24 locations be
reserved to allow for future enhancement of the
routine.
KIND = must be input containing zero when this routine is
first called to represent a particular line of
text. Thereafter, the value of KIND returned by
this routine should be supplied to the subsequent
call which is continuing the representation of the
same line of text. KIND is returned containing one
of the following values.
= 1, returned if the line of text has been completely
represented. IBUFFR(LFTCOL+1) through and
including IBUFFR(MAXUSD) is returned containing
spaces. This portion of the IBUFFR array will
again be returned containing spaces if DATURN is
subsequently called with the value of KIND being
unchanged.
= 2, returned if IBUFFR(LFTCOL+1) through and
including IBUFFR(MAXUSD) is returned containing a
portion of the representation of a single
character.
= 3, returned if the available portion of the IBUFFR
array was insufficient to contain the
representation of the lettering. MAXBFR-LFTCOL is
FASP, FORTRAN Alphameric Subroutine Package Page 431
DATEXT, FORTRAN Routine for Large Printable Characters
less than MAGNFY times the character height.
= 4, returned if the available portion of the ISTORE
array was insufficient to contain the description
of the current state of the lettering process for
transfer to the subsequent call to this routine
which is to continue the lettering of the same line
of text.
= 5, returned if a font was not loaded.
ISTORE = array used to transfer a description of the current
state of the lettering process to the subsequent
call of this routine which is to continue the
representation of the same line of text. The
original contents of the ISTORE array are ignored
and are destroyed.
A Program Demonstrating the Use of the DATURN Routine
- ------- ------------- --- --- -- --- ------ -------
The program listed below constructs a large multi-page
banner which can be printed on the line-printer. Each
banner can contain up to 10 parallel lines of lettering.
Each line in the input file should contain the height
magnification factor, the width magnification factor, the
intercharacter spacing (before application of the width
magnification factor), and the number of extra blank lines
to be included in the banner prior lettering for possible
centering, followed on the same line by a single character
which is ignored and then by the text to be represented as a
single line in the banner. The numbers are specified as
integers. A line containing a single zero terminates the
description of a single banner. Additional groups of lines
containing numbers and text can follow. two consecutive
lines each containing a single zero terminate execution.
The contents of a typical input file describing a first
banner containing 3 lines, and a second containing 2 lines
are shown below
3 1 2 0 TOP LINE, 1ST BANNER
3 1 2 0 MIDDLE LINE, 1ST BANNER
3 1 2 0 BOTTOM LINE, 1ST BANNER
3 1 2 0 LINE REJECTED BECAUSE WON'T FIT
0
5 2 2 0 TOP LINE, 2ND BANNER
5 2 2 0 BOTTOM LINE, 2ND BANNER
0
0
The characters produced by this program are turned 90
degrees from the orientation for which they were designed.
If TEXT5 is used, then height and width magnification
factors of 1 will give letters which are 9 columns high and
FASP, FORTRAN Alphameric Subroutine Package Page 432
DATEXT, FORTRAN Routine for Large Printable Characters
14 lines wide. The lettering can extend across the full 132
column width of the line-printer paper. Two parallel lines
of lettering can be generated if a height magnification
factor of 5 and width magnification factor of 2 are used.
Three parallel lines can be generated if a height factor of
3 and width factor of 1 are used. The intercharacter
spacing should be about 2.
C RENBR(BANNER/CONSTRUCT LETTERING ALONG FANFOLD PAPER)
DIMENSION LTTR(1000),IBUFFR(132),ISTORE(240),
1LINKND(10),LINLNG(10),LINHIH(10),NEEDED(10),
2LINWID(10),LINUSD(10),LINSPC(10),INITAL(10),
3MULTPL(4)
DATA IDISK,JDISK,ITTY,JTTY,MAXBFR,MAXSTR/
11,20,5,5,132,240/
DATA IBLANK,IONE,ISTAR,IPLUS/1H ,1H1,1H*,1H+/
DATA MOVE,ISPACE/1,0/
DATA MULTPL/1HH,1HI,1HO,1HX/
CALL TEXT5
LONGST=0
WRITE(JTTY,1)
1 FORMAT(39H NUMBER OF IMPRESSIONS (- FOR DARKEST) ,$)
READ(ITTY,2)KOPIES
2 FORMAT(I)
IF(KOPIES.LT.-4)KOPIES=-4
IF(KOPIES.GT.8)KOPIES=8
C
C READ HEIGHT, WIDTH, SPACING AND TEXT TO REPRESENT
IEOF=0
3 KNTLIN=0
LTREND=0
4 KNTLIN=KNTLIN+1
IF(KNTLIN.GT.10)GO TO 12
READ(IDISK,5,END=11)IBUFFR
5 FORMAT(132A1)
LOWBFR=1
LINHIH(KNTLIN)=3
LINWID(KNTLIN)=1
LINSPC(KNTLIN)=2
INITAL(KNTLIN)=0
NEEDED(KNTLIN)=0
LINLNG(KNTLIN)=-1
LINKND(KNTLIN)=0
DO 7 INDEX=1,4
CALL DAIHFT(0,0,0,IBUFFR,MAXBFR,
1LOWBFR,KIND,ISHIFT,JSHIFT,KSHIFT,LSHIFT,IVALUE)
GO TO(12,8,6),KIND
6 IF(INDEX.EQ.1)LINHIH(KNTLIN)=IVALUE
IF(INDEX.EQ.2)LINWID(KNTLIN)=IVALUE
IF(INDEX.EQ.3)LINSPC(KNTLIN)=IVALUE
7 CONTINUE
IF(IVALUE.LE.0)GO TO 9
INITAL(KNTLIN)=IVALUE
LINKND(KNTLIN)=1
FASP, FORTRAN Alphameric Subroutine Package Page 433
DATEXT, FORTRAN Routine for Large Printable Characters
GO TO 9
8 LOWBFR=LOWBFR-1
9 LTRBGN=LTREND
LTREND=LTREND+100
10 LTRBGN=LTRBGN+1
IF(LTRBGN.GT.LTREND)GO TO 4
LOWBFR=LOWBFR+1
LTTR(LTRBGN)=IBLANK
IF(LOWBFR.LE.MAXBFR)LTTR(LTRBGN)=IBUFFR(LOWBFR)
GO TO 10
11 IEOF=1
12 KNTLIN=KNTLIN-1
IF(KNTLIN.LE.0)GO TO 45
C
C PRODUCE EACH LINE OF LETTERING UNTIL ALL LINES DONE
IF(LONGST.GT.0)WRITE(JDISK,13)
13 FORMAT(1X/1X/1X/1X/1X/1X/1X/1X)
LONGST=0
KONTRL=IONE
14 NOWLIN=KNTLIN
MAXUSD=0
GO TO 17
C
C INSERT SEPARATION BETWEEN COLUMNS OF TEXT
15 IF(MAGNFY.LT.LINHIH(NOWLIN))MAGNFY=LINHIH(NOWLIN)
16 IF(MAXUSD.GE.MAXBFR)GO TO 37
MAGNFY=MAGNFY-1
MAXUSD=MAXUSD+1
IBUFFR(MAXUSD)=IBLANK
IF(MAGNFY.GT.0)GO TO 16
C
C REPRESENT CURRENT LINE OF CURRENT COLUMN OF TEXT
17 MAGNFY=LINHIH(NOWLIN)
NEEDED(NOWLIN)=NEEDED(NOWLIN)-1
IF(NEEDED(NOWLIN).GT.0)GO TO 21
NEEDED(NOWLIN)=LINWID(NOWLIN)
LFTCOL=MAXUSD
LTREND=100*NOWLIN
LTRBGN=LTREND-99
INISTR=(24*NOWLIN)-23
KIND=LINKND(NOWLIN)
INTRVL=LINSPC(NOWLIN)
CALL DATURN(INTRVL,MOVE ,ISPACE,LTTR ,LTRBGN,
1LTREND,LFTCOL,MAXBFR,IBUFFR,MAXUSD,MAGNFY,INISTR,
2MAXSTR,KIND ,ISTORE)
IF(INITAL(NOWLIN).LE.0)GO TO 18
IF(KIND.NE.1)GO TO 20
NEEDED(NOWLIN)=INITAL(NOWLIN)
INITAL(NOWLIN)=-INITAL(NOWLIN)
LINKND(NOWLIN)=0
KIND=2
GO TO 19
18 LINKND(NOWLIN)=KIND
19 LINUSD(NOWLIN)=MAXUSD
FASP, FORTRAN Alphameric Subroutine Package Page 434
DATEXT, FORTRAN Routine for Large Printable Characters
20 GO TO(23,22,37,41,43),KIND
21 MAXUSD=LINUSD(NOWLIN)
22 NOWLIN=NOWLIN-1
IF(NOWLIN.GT.0)GO TO 15
GO TO 25
C
C DONE WITH THIS COLUMN OF TEXT, CHECK IF DONE WITH ALL
23 IF(LINLNG(NOWLIN).LT.0)LINLNG(NOWLIN)=LONGST
NOWLIN=NOWLIN-1
IF(NOWLIN.GT.0)GO TO 15
NOWLIN=KNTLIN
24 IF(NOWLIN.LE.0)GO TO 33
IF(LINLNG(NOWLIN).LT.0)GO TO 25
NOWLIN=NOWLIN-1
GO TO 24
C
C OUTPUT CURRRENT LINE
25 LONGST=LONGST+1
IF(KOPIES.LT.0)GO TO 28
J=KOPIES
26 WRITE(JDISK,27)KONTRL,(IBUFFR(I),I=1,MAXUSD)
27 FORMAT(133A1)
KONTRL=IPLUS
J=J-1
IF(J.GT.0)GO TO 26
GO TO 32
28 DO 31 K=1,4
LETTER=MULTPL(K)
DO 29 I=1,MAXUSD
IF(IBUFFR(I).NE.IBLANK)IBUFFR(I)=LETTER
29 CONTINUE
J=KOPIES
30 WRITE(JDISK,27)KONTRL,(IBUFFR(I),I=1,MAXUSD)
KONTRL=IPLUS
J=J+1
IF(J.LT.0)GO TO 30
31 CONTINUE
32 KONTRL=ISTAR
GO TO 14
C
C ERROR MESSAGES AND SUMMATION
33 WRITE(JTTY,34)(LINLNG(I),I=1,KNTLIN)
34 FORMAT(18H LENGTHS (TOP 1ST),10I5)
DO 35 I=1,KNTLIN
35 LINLNG(I)=(LONGST-LINLNG(I)-INITAL(I))/2
WRITE(JTTY,36)(LINLNG(I),I=1,KNTLIN)
36 FORMAT(18H NEEDED TO CENTER ,10I5)
GO TO 40
37 WRITE(JTTY,38),KNTLIN
38 FORMAT(36H LETTERS TOO HIGH TO REPRESENT LINE ,I3)
DO 39 I=1,KNTLIN
NEEDED(I)=0
LINKND(I)=0
IF(INITAL(I).EQ.0)GO TO 39
FASP, FORTRAN Alphameric Subroutine Package Page 435
DATEXT, FORTRAN Routine for Large Printable Characters
IF(INITAL(I).LT.0)INITAL(I)=-INITAL(I)
LINKND(I)=1
39 CONTINUE
IF(KNTLIN.GT.1)GO TO 12
40 IF(IEOF.EQ.0)GO TO 3
GO TO 45
C
C SERIOUS ERROR CONDITIONS
41 WRITE(JTTY,42)
42 FORMAT(21H INSUFFICIENT STORAGE)
GO TO 45
43 WRITE(JTTY,44)
44 FORMAT(16H FONT NOT LOADED)
45 ENDFILE JDISK
STOP
C251676520779$
END
Font Style Samples
---- ----- -------
The designs of the numerals shown below are typical of the
designs of the characters in the corresponding fonts TEXT1
through TEXT5 respectively.
333333
33 33 55555555555555
33 33 4444 5555
111 2222 33 44 44 5555
11111 2 33 44 44 5555555555
111 222 33333 44 44 55555
111 2 33 44444444 5555
111 22222 33 44 5555
33 33 44 55555
33 33 5555555555
333333
FASP, FORTRAN Alphameric Subroutine Package Page 436
DATEXT, FORTRAN Routine for Large Printable Characters
Typical Lettering using the TEXT1 Font
------- --------- ----- --- ----- ----
The only punctuation marks in this font are ()*/+-=:., and '
AAA BBBBBBBBB CCCCCCCCC DDDDDDD EEEEEEEEE
AAA AAA BBB BBB CCC DDD DDD EEE
AAA AAA BBBBBBB CCC DDD DDD EEEEEE
AAAAAAAAA BBB BBB CCC DDD DDD EEE
AAA AAA BBBBBBBBB CCCCCCCCC DDDDDDD EEEEEEEEE
FFFFFFFFF GGGGGGGGG HHH HHH IIIIIII JJJ
FFF GGG HHH HHH III JJJ
FFFFFF GGG GGGG HHHHHHHHH III JJJ
FFF GGG GGG HHH HHH III JJJ JJJ
FFF GGGGGGGGG HHH HHH IIIIIII JJJJJJJJJ
KKK KKK LLL MMM MMM NNN NNN OOOOOOOOO
KKK KKK LLL MMMM MMMM NNNN NNN OOO OOO
KKKKKKK LLL MMM M M MMM NNN N NNN OOO OOO
KKK KKK LLL MMM M MMM NNN NNNN OOO OOO
KKK KKK LLLLLLLLL MMM MMM NNN NNN OOOOOOOOO
PPPPPPPPP QQQQQQQQQ RRRRRRRRR SSSSSSS TTTTTTTTT
PPP PPP QQQ QQQ RRR RRR SSS TTT
PPPPPPPPP QQQ QQQ RRRRRRR SSSSSSS TTT
PPP QQQ QQQ RRR RRR SSS TTT
PPP QQQQQQQQQQQ RRR RRR SSSSSSSS TTT
UUU UUU VVV VVV WWW WWW XXX XXX YYY YYY
UUU UUU VVV VVV WWW W WWW XXX XXX YYY YYY
UUU UUU VVV VVV WWW W W WWW XXX YYY
UUU UUU VVVVV WWWW WWWW XXX XXX YYY
UUUUUUUUU VVV WWW WWW XXX XXX YYY
ZZZZZZZZZ 000000000 111 222222222 333333333
ZZZ 000 000 11111 222 333
ZZZ 000 000 111 2222222 333
ZZZ 000 000 111 222 333
ZZZZZZZZZ 000000000 111 222222222 333333333
444 444 555555555 666 777777777 888888888
444 444 555 666 777 888 888
444444444 5555555 666666666 777 88888
444 555 666 666 777 888 888
444 555555555 666666666 777 888888888
999999999 ((( *** *** /// +++
999 999 ((( *** /// +++
999999999 ((( ********* /// +++++++++
999 ((( *** /// +++
999 ((( *** *** /// +++
FASP, FORTRAN Alphameric Subroutine Package Page 437
DATEXT, FORTRAN Routine for Large Printable Characters
Typical Lettering using the TEXT2 Font
------- --------- ----- --- ----- ----
A BBBB CCCC DDDD EEEEE FFFFF GGGG H H
A A B B C D D E F G H H
A A BBBB C D D EEEE FFFF G GG HHHHH
AAAAA B B C D D E F G G H H
A A BBBB CCCC DDDD EEEEE F GGGG H H
III J K K L M M N N OOO PPPP
I J K K L MM MM NN N O O P P
I J KKK L M M M N N N O O PPPP
I J J K K L M M N NN O O P
III JJJ K K LLLLL M M N N OOO P
QQQ RRRR SSSS TTTTT U U V V W W X X
Q Q R R S T U U V V W W X X
Q Q Q RRRR SSS T U U V V W W W X
Q Q R R S T U U V V WW WW X X
QQ Q R R SSSS T UUU V W W X X
Y Y ZZZZZ 000 11 2222 3333 4 4 55555
Y Y Z 0 0 1 2 3 4 4 5
Y Z 0 0 1 222 33 44444 5555
Y Z 0 0 1 2 3 4 5
Y ZZZZZ 000 111 22222 3333 4 5555
6 77777 888 999 (( ))
6 7 8 8 9 9 ( ) +
6666 7 888 9999 ( ) +++++ -----
6 6 7 8 8 9 ( ) +
666 7 888 9 (( ))
/ '' :: ^ "" ""
*** / '' :: ^ ^ "" ""
***** / ,, ' ^ ^ " "
*** / .. ,, ::
/ .. , ::
# # $$$$ %% % &&& < > ????
## ## $ $ %% % & ===== < > ?
$$$ % & & & < > ??
## ## $ $ % %% & & ===== < >
# # $$$$ % %% && & < > ?
@@@ !
@ @ !
@ @@ !
@ @@
@ !
FASP, FORTRAN Alphameric Subroutine Package Page 438
DATEXT, FORTRAN Routine for Large Printable Characters
Typical Lettering using the TEXT3 Font
------- --------- ----- --- ----- ----
AA BBBBBBBBB CCCCCC DDDDDDDD
AAAA BB BB CC CC DD DD
AA AA BB BB CC CC DD DD
AA AA BB BB CC DD DD
AA AA BB BB CC DD DD
AA AA BBBBBBBBB CC DD DD
AA AA BB BB CC DD DD
AAAAAAAAAAAA BB BB CC DD DD
AA AA BB BB CC CC DD DD
AA AA BB BB CC CC DD DD
AA AA BBBBBBBBB CCCCCC DDDDDDDD
EEEEEEEEEEEE FFFFFFFFFFFF GGGGGG HH HH
EE FF GG GG HH HH
EE FF GG GG HH HH
EE FF GG HH HH
EE FF GG HH HH
EEEEEEEE FFFFFFFF GG HHHHHHHHHHHH
EE FF GG GGGGG HH HH
EE FF GG GG HH HH
EE FF GG GG HH HH
EE FF GG GG HH HH
EEEEEEEEEEEE FF GGGGGGGGG HH HH
IIIIII JJ KK KK LL
II JJ KK KK LL
II JJ KK KK LL
II JJ KK KK LL
II JJ KK KK LL
II JJ KK KKKK LL
II JJ KKKK KK LL
II JJ KKK KK LL
II JJ JJ KK KK LL
II JJ JJ KK KK LL
IIIIII JJJJJJ KK KK LLLLLLLLLLLL
MM MM NN NN OOOOOO PPPPPPPPP
MMM MMM NNN NN OO OO PP PP
MMMM MMMM NNNN NN OO OO PP PP
MM MM MM MM NN NN NN OO OO PP PP
MM MMMM MM NN NN NN OO OO PP PP
MM MM MM NN NN NN OO OO PPPPPPPPP
MM MM NN NN NN OO OO PP
MM MM NN NN NN OO OO PP
MM MM NN NNNN OO OO PP
MM MM NN NNN OO OO PP
MM MM NN NN OOOOOO PP
FASP, FORTRAN Alphameric Subroutine Package Page 439
DATEXT, FORTRAN Routine for Large Printable Characters
TEXT3 Font (continued)
----- ---- ---------
QQQQQQ RRRRRRRRR SSSSSS TTTTTTTTTTTT
QQ QQ RR RR SS SS TT
QQ QQ RR RR SS SS TT
QQ QQ RR RR SS TT
QQ QQ RR RR SS TT
QQ QQ RRRRRRRRR SSSSSS TT
QQ QQ RR RR SS TT
QQ QQ Q RR RR SS TT
QQ QQ RR RR SS SS TT
QQ QQ RR RR SS SS TT
QQQQQ QQ RR RR SSSSSS TT
UU UU VV VV WW WW XX XX
UU UU VV VV WW WW XX XX
UU UU VV VV WW WW XX XX
UU UU VV VV WW WW XX XX
UU UU VV VV WW WW XXXX
UU UU VV VV WW WW WW XX
UU UU VV VV WW WWWW WW XXXX
UU UU VV VV WW WW WW WW XX XX
UU UU VV VV WWWW WWWW XX XX
UU UU VVVV WWW WWW XX XX
UUUUUU VV WW WW XX XX
YY YY ZZZZZZZZZZZZ 000000 11
YY YY ZZ 00 00 111
YY YY ZZ 00 00 1111
YY YY ZZ 00 00 11
YYYY ZZ 00 00 11
YY ZZ 00 00 11
YY ZZ 00 00 11
YY ZZ 00 00 11
YY ZZ 00 00 11
YY ZZ 00 00 11
YY ZZZZZZZZZZZZ 000000 111111
222222 333333 44 555555555555
22 22 33 33 44 55
22 22 33 33 44 55
22 33 44 44 55
22 33 44 44 55
22 33333 44 44 555555555
22 33 444444444444 55
22 33 44 55
22 33 33 44 55
22 33 33 44 55
222222222222 333333 44 555555555
FASP, FORTRAN Alphameric Subroutine Package Page 440
DATEXT, FORTRAN Routine for Large Printable Characters
TEXT3 Font (continued)
----- ---- ---------
66 777777777777 888888 999999
66 77 88 88 99 99
66 77 88 88 99 99
66 77 88 88 99 99
66 77 88 88 99 99
6666666 77 888888 9999999
66 66 77 88 88 99
66 66 77 88 88 99
66 66 77 88 88 99
66 66 77 88 88 99
666666 77 888888 99
%%% %% $$
<< >> % % %% $$
<< >> % % %% $$$$$$$$$$
<< >> %%% %% $$ $$
<< >> %% $$ $$
<< >> %% $$$$$$$$
<< >> %% $$ $$
<< >> %% %%% $$ $$
<< >> %% % % $$$$$$$$$$
<< >> %% % % $$
%% %%% $$
????????? ^^ !! ## ##
?? ^^^^ !! ## ##
?? ^^ ^^ !! ## ##
?? ^^ ^^ !! ##### #####
?? ^^ ^^ !!
?????? ^^ ^^ !!
?? !!
##### #####
?? !! ## ##
???? !!!! ## ##
?? !! ## ##
&&&& @@@@@@ //
&& && @@ @@ //
&& && @@ @@ // ++
&&&& @@ @@ @@ // ++
&& @@ @ @ @@ // ++
&&&& @@ @ @ @@ // ++++++++++++
&& && @@ @ @ @@ // ++
&& && && @@ @@@@@@ // ++
&& && @@ // ++
&& && && @@ //
&&&&& && @@@@@@@@@ //
FASP, FORTRAN Alphameric Subroutine Package Page 441
DATEXT, FORTRAN Routine for Large Printable Characters
TEXT3 Font (continued)
----- ---- ---------
((((
(((
** ** ((
** ** ============ ((
**** ((
************ ------------ ((
**** ((
** ** ============ ((
** ** ((
(((
((((
'' :: "" "" ;;
''' :::: """ """ ;;;;
'' :: "" "" ;;
'' "" ""
;;
:: ;;;
:::: ;;
:: ;;
))))
)))
))
))
))
))
))
,, ))
.. ,,, ))
.... ,, )))
.. ,, ))))
FASP, FORTRAN Alphameric Subroutine Package Page 442
DATEXT, FORTRAN Routine for Large Printable Characters
Typical Lettering using the TEXT4 Font
------- --------- ----- --- ----- ----
AAA BBBBBB CCCCC DDDDD EEEEEEEE FFFFFFFF
AAAA BB BB CC DD DD EE FF
AA AA BB BB CC DD DD EE FF
AA AA BBBBBB CC DD DD EEEEE FFFFF
AAAAAAA BB BB CC DD DD EE FF
AA AA BB BB CC DD DD EE FF
AA AA BBBBBB CCCCC DDDDD EEEEEEEE FF
GGGGG HH HH IIIIII JJ KK KK LL
GG HH HH II JJ KK KK LL
GG HH HH II JJ KK KK LL
GG GGGG HHHHHHHH II JJ KKKKK LL
GG GG HH HH II JJ JJ KKK KK LL
GG GG HH HH II JJ JJ KK KK LL
GGGGG HH HH IIIIII JJJJ KK KK LLLLLLLL
MM MM NN NN OOOOO PPPPPP QQQQQ RRRRRR
MMM MMM NNN NN OO OO PP PP QQ QQ RR RR
MMMM MMMM NNNN NN OO OO PP PP QQ QQ RR RR
MM MMMM MM NN NN NN OO OO PPPPPP QQ QQ QQ RRRRRR
MM MM MM NN NNNN OO OO PP QQ QQ Q RR RR
MM MM NN NNN OO OO PP QQ QQ RR RR
MM MM NN NN OOOOO PP QQQQ QQ RR RR
SSSSSS TTTTTTTT UU UU VV VV WW WW YY YY
SS TT UU UU VV VV WW WW YY YY
SS TT UU UU VV VV WW WW WW YYYY
SSSS TT UU UU VV VV WW WWWW WW YY
SS TT UU UU VVVV WWWW WWWW YY
SS TT UU UU VVV WWW WWW YY
SSSSSS TT UUUU VV WW WW YY
XX XX ZZZZZZZZ 0000 11 222222 333333
XX XX ZZ 00 00 1111 22 33
XXXX ZZ 00 00 11 22 33
XX ZZ 00 00 11 2222 3333
XXXX ZZ 00 00 11 22 33
XX XX ZZ 00 00 11 22 33
XX XX ZZZZZZZZ 0000 11 22222222 333333
4444 55555555 66 77777777 8888 9999
44 44 55 66 77 88 88 99 99
44 44 55 66 77 88 88 99 99
44 44 555555 66666 77 8888 99999
44444444 55 66 66 77 88 88 99
44 55 66 66 77 88 88 99
44 555555 6666 77 8888 99
FASP, FORTRAN Alphameric Subroutine Package Page 443
DATEXT, FORTRAN Routine for Large Printable Characters
TEXT4 Font (continued)
----- ---- ---------
(( //
(( ++ ** ** //
(( ++ **** // ========
(( ++++++++ -------- ******** //
(( ++ **** // ========
(( ++ ** ** //
(( //
''' )) ^^ ??????
'' )) ^^^^ ??
'' )) ^^ ^^ ??
)) ^^ ^^ ???
,,, ))
.. ,, )) ??
.. ,, )) ??
## ## :: ;; !!
<< >> ## ## :: ;; !!
<< >> ### ### !!
<< >> !!
<< >> ### ### ;;;
<< >> ## ## :: ;; !!
## ## :: ;; !!
FASP, FORTRAN Alphameric Subroutine Package Page 444
DATEXT, FORTRAN Routine for Large Printable Characters
Typical Lettering using the TEXT5 Font
------- --------- ----- --- ----- ----
AA BBBBBBBBBB CCCCCCCCCC DDDDDDDD
AAAA BBBB BBBBB CCCCC DDDD DDDDD
AAAAAA BBBB BBBB CCCC DDDD DDDD
AA AAAA BBBB BBBBB CCCC DDDD DDDD
AA AAAA BBBBBBBBBB CCCC DDDD DDDD
AA AAAA BBBB BBBBB CCCC DDDD DDDD
AAAAAAAAAAAAAA BBBB BBBB CCCC DDDD DDDD
AA AAAA BBBB BBBBB CCCCC DDDD DDDDD
AA AAAA BBBBBBBBBB CCCCCCCCCC DDDDDDDD
EEEEEEEEEEEEEE FFFFFFFFFFFFFF GGGGGGGGGG HHHH HHHH
EEEE FFFF GGGGG HHHH HHHH
EEEE FFFF GGGG HHHH HHHH
EEEE FFFF GGGG HHHH HHHH
EEEEEEEEE FFFFFFFFF GGGG HHHHHHHHHHHHHH
EEEE FFFF GGGG GGGGGG HHHH HHHH
EEEE FFFF GGGG GGGG HHHH HHHH
EEEE FFFF GGGGG GGGG HHHH HHHH
EEEEEEEEEEEEEE FFFF GGGGGGGGGG HHHH HHHH
IIIIIIII JJJJ KKKK KKK LLLL
IIII JJJJ KKKK KKK LLLL
IIII JJJJ KKKK KKK LLLL
IIII JJJJ KKKK KKK LLLL
IIII JJJJ KKKKKK LLLL
IIII JJJJ JJJJ KKKKKKKK LLLL
IIII JJJJ JJJJ KKKK KKKK LLLL
IIII JJJJJ JJJJJ KKKK KKKK LLLL
IIIIIIII JJJJJJJ KKKK KKKK LLLLLLLLLLLLLL
MM MMMM NNNNN NN OOOOOO PPPPPPPPPP
MMM MMMMM NNNNNN NN OOOOO OOOOO PPPP PPPPP
MMMM MMMMMM NN NNNN NN OOOO OOOO PPPP PPPP
MMMMM MM MMMM NN NNNN NN OOOO OOOO PPPP PPPP
MM MMMMM MMMM NN NNNN NN OOOO OOOO PPPP PPPPP
MM MMM MMMM NN NNNN NN OOOO OOOO PPPPPPPPPP
MM M MMMM NN NNNN NN OOOO OOOO PPPP
MM MMMM NN NNNNNN OOOOO OOOOO PPPP
MM MMMM NN NNNNN OOOOOO PPPP
QQQQQQ RRRRRRRRRR SSSSSSSSSS TTTTTTTTTTTTTT
QQQQQ QQQQQ RRRR RRRRR SSSSS TTTT
QQQQ QQQQ RRRR RRRR SSSS TTTT
QQQQ QQQQ RRRR RRRR SSSSS TTTT
QQQQ QQQQ RRRR RRRRR SSSSSS TTTT
QQQQ QQQ QQQ RRRRRRRRRRR SSSSS TTTT
QQQQ QQQ Q RRRR RRRR SSSS TTTT
QQQQQ QQQ RRRR RRRR SSSSS TTTT
QQQQQ QQQ RRRR RRRR SSSSSSSSSS TTTT
FASP, FORTRAN Alphameric Subroutine Package Page 445
DATEXT, FORTRAN Routine for Large Printable Characters
TEXT5 Font (continued)
----- ---- ---------
UUUU UU VVVV VV WWWW WW XXXX XX
UUUU UU VVVV VV WWWW WW XXXX XX
UUUU UU VVVV VV WWWW W WW XXXX XX
UUUU UU VVVV VV WWWW WWW WW XXXX XX
UUUU UU VVVV VV WWWW WWWWW WW XXXXXX
UUUU UU VVVV VV WWWW WW WWWWW XX XXXX
UUUU UU VVVVVV WWWWWW WWWW XX XXXX
UUUUU UUUU VVVV WWWWW WWW XX XXXX
UUUUUUU VV WWWW WW XX XXXX
YYYY YY ZZZZZZZZZZZZZZ BBBB
YYYY YY ZZZZ BBBB
YYYY YY ZZZZ BBBB
YYYY YY ZZZZ AAAAAAAAA BBBBBBBBB
YYYYYY ZZZZ AAAA BBBB BBBB
YYYY ZZZZ AAAAAAAAAA BBBB BBBB
YYYY ZZZZ AAAA AAAA BBBB BBBB
YYYY ZZZZ AAAA AAAA BBBB BBBB
YYYY ZZZZZZZZZZZZZZ AAAAAAAAAA BBBBBBBBB
DDDD FFFFF
DDDD FFFF
DDDD FFFF
CCCCCCCCC DDDDDDDDD EEEEEE FFFFFFFFFFFF
CCCC DDDD DDDD EEEE EEEE FFFF
CCCC DDDD DDDD EEEE EEEE FFFF
CCCC DDDD DDDD EEEEEEEEEEEE FFFF
CCCC DDDD DDDD EEEE FFFF
CCCCCCCCC DDDDDDDDD EEEEEEEEE FFFF
HHHH II JJ KKKK LLLL
HHHH IIII JJJJ KKKK LLLL
HHHH II JJ KKKK LLLL
GGGGGGGGG HHHH HHHH KKKK KKK LLLL
GGGG GGGG HHHH HHHHHH IIII JJJJ KKKK KKK LLLL
GGGG GGGG HHHHHH HHHH IIII JJJJ KKKKKK LLLL
GGGG GGGG HHHHH HHHH IIII JJJJ KKKKKKKK LLLL
GGGG GGGG HHHH HHHH IIII JJJJ KKKK KKKK LLLL
GGGGGGGGG HHHH HHHH IIII JJJJ KKKK KKKK LLLL
GGGG JJJJ
GGGG JJJJ
GGGGGGGG JJJJJ
MMMM MMMM MMMM NNNN NNNN OOOOOO PPPPPPPPP
MMMM MMMMMM MMMMMM NNNN NNNNNN OOOO OOOO PPPP PPPP
MMMMMM MMMMMM MMMM NNNNNN NNNN OOOO OOOO PPPP PPPP
MMMMM MMMMM MMMM NNNNN NNNN OOOO OOOO PPPP PPPP
MMMM MMMM MMMM NNNN NNNN OOOO OOOO PPPP PPPP
MMMM MMMM MMMM NNNN NNNN OOOOOO PPPPPPPPP
PPPP
PPPP
PPPP
FASP, FORTRAN Alphameric Subroutine Package Page 446
DATEXT, FORTRAN Routine for Large Printable Characters
TEXT5 Font (continued)
----- ---- ---------
TTTT
TTTT
TTTT
QQQQQQQQQ RRRR RRRRR SSSSSSSSSS TTTTTTTTTTTT
QQQQ QQQQ RRRR RRRR SSSS TTTT
QQQQ QQQQ RRRRRR SSSSSSS TTTT
QQQQ QQQQ RRRR SSSSSSS TTTT
QQQQ QQQQ RRRR SSSS TTTT
QQQQQQQQQ RRRR SSSSSSSSSS TTTT
QQQQ
QQQQ
QQQQ
UUUU UUUU VVVV VV WWWW WW WW XXXX XXX
UUUU UUUU VVVV VV WWWW WWWW WW XXXX XXX
UUUU UUUUU VVVV VV WWWW WWWWWW WW XXXXX
UUUU UUUUUU VVVVVV WWWWWW WWWWWW XXXXX
UUUUUU UUUU VVVV WWWW WWWW XXX XXXX
UUUU UUUU VV WW WW XXX XXXX
000000 11
00000 00000 1111
0000 0000 1111111
YYYY YY ZZZZZZZZZZZZ 0000 0000 1111
YYYY YY ZZZZ 0000 0000 1111
YYYY YY ZZZZ 0000 0000 1111
YYYYYY ZZZZ 0000 0000 1111
YYYY ZZZZ 00000 00000 1111
YY ZZZZZZZZZZZZ 000000 1111
YY
YYY
YYY
2222222222 3333333333 4444444 55555555555555
22222 33333 44 4444 5555
2222 3333 44 4444 5555
22222 33333 44 4444 5555555555
222222 333333 44 4444 55555
22222 33333 44444444444444 5555
2222 3333 44444444444444 5555
2222 33333 4444 55555
22222222222222 3333333333 4444 5555555555
6666 77777777777777 88888888 999999
6666 7777 8888 888 99999 99999
6666 7777 8888 888 9999 9999
6666 7777 88888 888 99999 9999
6666666666 7777 888888 9999999999
6666 66666 7777 888 8888 9999
6666 6666 7777 888 8888 9999
66666 66666 7777 888 8888 9999
666666 7777 88888888 9999
FASP, FORTRAN Alphameric Subroutine Package Page 447
DATEXT, FORTRAN Routine for Large Printable Characters
TEXT5 Font (continued)
----- ---- ---------
;; :: !!!! "" "" ////
;;;; :::: !!!! """ """ //// ++++
;; :: !!!! "" "" //// ++++
!!!! "" "" //// ++++
!!!! //// ++++++++++++++
;; ,, //// ++++++++++++++
;;; .. :: !! ,,, //// ++++
;; .... :::: !!!! ,, //// ++++
;; .. :: !! ,, //// ++++
'' ))))) (((((
*** *** ''' ))))) (((((
*** *** '' )))) ((((
****** '' )))) ((((
************** -------------- )))) ((((
************** -------------- )))) ((((
****** )))) ((((
*** *** ))))) (((((
*** *** ))))) (((((
### ### ^^
============== ### ### ^^^^^^ <<<<
============== ############## ^^^^ ^^^^ <<<<
### ### ^^^^ ^^^^ <<<<
### ### <<<<
### ### <<<<
============== ############## <<<<
============== ### ### <<<<
### ###
?????????? &&&&&&& %%% %%%%
>>>> ????? &&&& %% %% %%%%
>>>> ???? &&&& %% %% %%%%
>>>> ????? &&&&&& %%% %%%%
>>>> ??????? &&&& %%%%
>>>> &&&& &&&&&& %%%% %%%
>>>> ?? &&&& &&& %%%% %% %%
>>>> ???? &&&& &&& %%%% %% %%
?? &&&&&&& %%%% %%%
$$ @@@@@@@@
$$ @@@ @@@
$$$$$$$$$$$$ @@ @@@@ @@
$$$$ @@ @@ @@ @@
$$$$$$$$$$ @@ @@ @@ @@
$$$$ @@ @@ @@ @@
$$$$$$$$$$$$ @@ @@@@@@@@
$$ @@@
$$ @@@@@@@@@@@
FASP, FORTRAN Alphameric Subroutine Package Page 448
DATEXT, FORTRAN Routine for Large Printable Characters
User Specification of New Character Fonts
---- ------------- -- --- --------- -----
The SUBROUTINE or the BLOCK DATA routine which defines the
character shapes constructed by DATEXT is itself written by
another program named DAFONT which reads a file written by
the user containing the characters drawn as they would be
constructed by DATEXT. DAFONT asks the user to specify the
SUBROUTINE name which is to be included in the SUBROUTINE
statement and will generate a BLOCK DATA routine instead if
no name is supplied.
The rules governing the construction of the template file
which DAFONT reads as input are as follow.
1. Each character shape is constructed only of the
character which it is to represent. Since DATEXT can
select the N'th shape by which a particular letter is
represented, both upper and lower case characters can
be constructed if it is not objectionable that the
lower case shapes are constructed of, and selected by,
the corresponding upper case characters. The only
precaution which must be observed if both upper and
lower case shapes are to be available but constructed
of the upper case letters is that both the lower and
upper case shapes for a particular letter cannot be
defined on the same row of shapes in the template file.
Also, if the second row of character shapes in the
template file defines the alternate case shape of a
character in the first row, then the first and second
rows of character shapes should be separated by one or
more blank lines. Of course, if the FORTRAN compiler
and runtime system both allow lower case letters, and
neither translates such to upper case, then the lower
case letters can be selected by, and be formed, of the
lower case letters being represented.
2. The first row of character shapes in the template file
begins on the next nonblank line following the first
line, and extends to either the next blank line in the
file, or to the next line which contains no character
already found in the previous lines. The first row of
character shape specifications must be of maximum
height. The contents of the first line of the input
file is written directly into the output as a comment
line. The letter C is placed at the start of this
output line to cause it to be ignored by the FORTRAN
compiler. It is suggested that this comment line
contain a short description of the matrix upon which
the letters are designed.
3. Subsequent rows of character shapes start at the next
nonblank line and extend through the number of lines
established by the first row as being the character
shape height, or extend to but do not include the next
line which begins the specification of the shape of a
FASP, FORTRAN Alphameric Subroutine Package Page 449
DATEXT, FORTRAN Routine for Large Printable Characters
different character or characters but which does not
contain any of the characters being defined in the
current row of shapes. A row of character shapes is
top justified if the row (as opposed to the shapes
within the row) is shorter than the height established
by the first row of shapes.
4. Character shapes can contain horizontal rows and
vertical columns of spaces. Character shapes specified
on the same row can be separated by spaces or can
touch, but cannot overlap.
5. If the shape of a character has already been found on
the current row of character shape specifications, the
new (right) definition of the shape is ignored.
6. If the FORTRAN system does not allow end-of-file tests
in READ statements, then the template file must be
terminated by the letters END on a final line not part
of the preceding row of character shape specifications.
7. Unless the limits and dimensions in the DAFONT program
are changed, only the first 80 characters on a line are
read, a character shape can be no more than 30 lines
high, and at most 96 shapes can be defined.
8. The maximum width of any character is the number of
bits, excluding the sign bit, in a single computer
word. This means that letter shapes up to 31 bits wide
can be represented on an IBM 360 or 370 computer.
A Sample Template File
- ------ -------- ----
On the following page is shown a sample template file. The
shapes of only a few of the letters are defined in this
template file. As long as the character shapes do not
overlap, the horizontal spacing between shapes is optional.
Similarly, extra blank lines can separate the rows of
shapes. On the PDP-10 computer, a blank line must contain
at least 1 space to be recognized by a FORTRAN program.
This means that the blank line in the definition of the
shape of the number sign contains at least one space.
It should be noted that in this sample template file, a
second specification of the shape of the character S on the
same row of shapes as the first has been used to force the
correct vertical position of the lower case shapes of the
letters S and A. The second specification of the letter S,
in this case a vertical line, is ignored for other than
determination of the first line of the row of shapes.
However, if the vertical line formed of the letter S was not
separated from the first shape of the letter S by the shape
of some other character, then the 2 parts would have been
taken together as the specification of the shape of the
letter. This is necessary if characters are to be allowed
to be formed from more than 1 group of characters, and is
FASP, FORTRAN Alphameric Subroutine Package Page 450
DATEXT, FORTRAN Routine for Large Printable Characters
necessary for specification of the shape of the character #.
The sample template file defines the lower case letter
shapes g l i s and a, the character shapes : and #, and the
upper case letter shapes S A I and L. The first line
contains a comment to be written into the output file on a
line containing the letter C in column 1. The final line
containing the word END signals the end of the file and is
otherwise ignored.
SAMPLE TEMPLATE FILE, THIS LINE CONTAINS A COMMENT
LLL II
L II
G L
GGG L III
G G L I
G G L I
GGG LLLLL IIIII
G
GGGG
S
S
SSS AAAA S
S A S
SSS AAAA S
SA A S
SSSS AAA A S
S
S
:: # #
:: # #
### ###
### ###
:: # #
:: # #
SSSSS A IIIIIIIL
S A A I L
S A A I L
SSSSS A A I L
SAAAAAAA I L
SA A I L
SSSSSS A AIIIIIIILLLLLLL
END
FASP, FORTRAN Alphameric Subroutine Package Page 451
DATEXT, FORTRAN Routine for Large Printable Characters
A Sample BLOCK DATA Routine Defining a Short Font
- ------ ----- ---- ------- -------- - ----- ----
DAFONT produced the following FORTRAN BLOCK DATA routine as
output when it read the template file shown on the preceding
page.
BLOCK DATA
CSAMPLE TEMPLATE FILE, THIS LINE CONTAINS A COMMENT
COMMON/FASPG/KNTLTR,IHIGH,IWIDE,JWIDE,LOCK,
1LETTER(96),LENGTH(96),IPACKD( 672)
DIMENSION JLETTR( 11)
DATA JLETTR/
11HG,1HL,1HI,1HS,1HA,1H:,1H#,1HS,1HA,1HI,
11HL/
DIMENSION JLNGTH(11),JLNGT1(11)
EQUIVALENCE (JLNGT1(1),JLNGTH(1))
DATA JLNGT1/5,5,5,5,6,2,7,7,7,7,7/
DIMENSION JPACKD(99),JPACK1(99)
EQUIVALENCE (JPACK1(1),JPACKD(1))
DATA JPACK1/ 0, 0, 1, 14, 17, 17, 14, 1, 30, 28,
1 4, 4, 4, 4, 4, 31, 0, 0, 12, 12,
2 0, 28, 4, 4, 31, 0, 0, 0, 0, 14,
3 16, 14, 1, 30, 0, 0, 0, 0, 60, 2,
4 30, 34, 29, 0, 0, 3, 3, 0, 0, 0,
5 3, 3, 0, 0, 20, 20,119, 0,119, 20,
6 20, 0, 0, 62, 64, 64, 62, 1, 1,126,
7 0, 0, 8, 20, 34, 65,127, 65, 65, 0,
8 0,127, 8, 8, 8, 8, 8,127, 0, 0,
9 64, 64, 64, 64, 64, 64,127, 0, 0/
DATA KNTLTR,IHIGH,IWIDE,JWIDE,LOCK/ 11, 9, 7, 7,999/
EQUIVALENCE (LETTER(1),JLETTR(1)),
1(LENGTH(1),JLNGTH(1)),(IPACKD(1),JPACKD(1))
END
FASP, FORTRAN Alphameric Subroutine Package Page 452
DATEXT, FORTRAN Routine for Large Printable Characters
Demonstration Program to Display Entire Contents of a Font
------------- ------- -- ------- ------ -------- -- - ----
The program listed below was used to generate most of the
font illustrations presented earlier in this document.
Since all of the shapes for each character must be shown,
yet each line of constructed character shapes can contain
only a small portion of the total available character
shapes, the program first inserts a dollar sign and case
identification digit in front of each character so that each
of the alternative shapes for any character can be selected
independently. An extra dollar sign is also inserted before
each dollar sign for which the shape is described in the
font so that DATEXT does not incorrectly interpret the
original dollar sign as the start of a command character
pair. Once a line of character shapes has been constructed,
the following dollar sign-digit pair will have been
processed, so the program backs up to the previous dollar
sign-digit pair prior to constructing the next line of
character shapes. The program uses the largest
intercharacter spacing which will not reduce the number of
character shapes included on the line. A zero
intercharacter spacing is allowed only if the characters on
the line are extremely large.
C RENBR(LTRALL/DISPLAY ALL CHARACTERS IN A FONT)
DIMENSION IBUFFR(60),LTTR(300),KASE(300),IDIGIT(10)
COMMON/FASPG/KNTLTR,IHIGH,IWIDE,JWIDE,LOCK,
1LETTER(96),LENGTH(96),IPACKD(672)
C
C KOLUMN = 0, PACK AS MANY CHARACTERS AS POSSIBLE ON
C EACH LINE. THIS OPTION IS BEST IF MOST
C COMMON WIDTH IS CONSIDERABLY NARROWER THAN
C WIDEST CHARACTER.
C = 1, ARRANGE CHARACTERS IN COLUMNS AS WIDE AS
C WIDEST CHARACTER. THIS OPTION IS BEST IF
C MOST COMMON WIDTH IS ONLY SLIGHTLY LESS THAN
C WIDEST CHARACTER.
DATA KOLUMN/0/
DATA IDIGIT/1H1,1H2,1H3,1H4,1H5,1H6,1H7,1H8,1H9,1H0/
DATA IDISK/5/
DATA KOMAND,ISPACE/1H$,1H /
IF(LOCK.NE.999)GO TO 23
C
C KASE ARRAY HOLDS WIDTHS AND MARKS CHARACTER LOCATIONS
DO 1 I=1,300
1 KASE(I)=0
C
C INSERT CASE NUMBER BEFORE EACH LETTER AND $ BEFORE $
LTREND=0
DO 3 INDEX=1,KNTLTR
ISAME=0
DO 2 I=1,INDEX
IF(LETTER(INDEX).EQ.LETTER(I))ISAME=ISAME+1
FASP, FORTRAN Alphameric Subroutine Package Page 453
DATEXT, FORTRAN Routine for Large Printable Characters
2 CONTINUE
IF(ISAME.GT.10)ISAME=10
LTREND=LTREND+3
KASE(LTREND-2)=LENGTH(INDEX)
LTTR(LTREND-2)=KOMAND
LTTR(LTREND-1)=IDIGIT(ISAME)
LTTR(LTREND)=LETTER(INDEX)
IF(LETTER(INDEX).NE.KOMAND)GO TO 3
LTREND=LTREND+1
LTTR(LTREND)=KOMAND
3 CONTINUE
LTRBGN=1
MAXMUM=0
C
C DETERMINE INTERCHARACTER SPACING AND IF TAILS USED.
C ALLOW 0 SPACING IF NEXT 3 CHARACTERS USE FULL WIDTH
4 LINE=IHIGH
IDONE=-1
JSTIFY=-1
IF(KOLUMN.NE.0)GO TO 7
MOVE=1
J=0
K=-1
DO 5 I=LTRBGN,LTREND
IF(KASE(I).EQ.0)GO TO 5
J=J+1
K=K+KASE(I)+1
IF(K.GT.60)GO TO 8
IF(J.GE.3)GO TO 6
5 CONTINUE
6 INTRVL=1
GO TO 9
7 MOVE=-2
IF(IWIDE.NE.JWIDE)GO TO 8
IF(JWIDE.LT.20)GO TO 6
8 INTRVL=0
9 MAXUSD=0
CALL DATEXT(LINE,JSTIFY,0,INTRVL,MOVE,
11,LTTR,LTRBGN,LTREND,0,60,60,
2IBUFFR,MAXUSD,MAXLIN,LTRNXT)
IF(IDONE.EQ.0)GO TO 12
IF(IDONE.GT.0)GO TO 15
DO 10 I=1,MAXUSD
IF(IBUFFR(I).NE.ISPACE)GO TO 11
10 CONTINUE
LINE=LINE-1
IF(LINE.GT.1)GO TO 9
11 IDONE=0
NEEDED=LTRNXT
LSTUSD=MAXUSD
GO TO 13
12 IF(MAXUSD.LE.LSTUSD)GO TO 14
IF(NEEDED.NE.LTRNXT)GO TO 14
13 INTRVL=INTRVL+1
FASP, FORTRAN Alphameric Subroutine Package Page 454
DATEXT, FORTRAN Routine for Large Printable Characters
IF(LTRNXT.LE.LTREND)GO TO 9
IF(INTRVL.LE.MAXMUM)GO TO 9
C
C GENERATE LETTERS WITHOUT WHITE ADJUSTMENT IF POSSIBLE
14 INTRVL=INTRVL-1
IF(MAXMUM.LT.INTRVL)MAXMUM=INTRVL
IDONE=1
LIMIT=LINE
LINE=1
IF(KOLUMN.NE.0)GO TO 9
MOVE=0
JSTIFY=0
GO TO 9
15 IF(LTRNXT.EQ.NEEDED)GO TO 17
MOVE=1
GO TO 9
16 IF(IBUFFR(MAXUSD).NE.ISPACE)GO TO 19
MAXUSD=MAXUSD-1
17 IF(MAXUSD.GT.0)GO TO 16
IF(IDONE.EQ.1)GO TO 21
WRITE(IDISK,18)
18 FORMAT(1X)
GO TO 21
19 IDONE=2
WRITE(IDISK,20)(IBUFFR(I),I=1,MAXUSD)
20 FORMAT(1X,132A1)
21 LINE=LINE+1
IF(LINE.LE.LIMIT)GO TO 9
IF(LTRNXT.GT.LTREND)GO TO 25
WRITE(IDISK,18)
LTRBGN=LTRNXT
22 IF(KASE(LTRBGN).NE.0)GO TO 4
LTRBGN=LTRBGN-1
GO TO 22
23 WRITE(IDISK,24)
24 FORMAT(16H FONT NOT LOADED)
25 STOP
END
FASP, FORTRAN Alphameric Subroutine Package Page 455
DATREE Returns Next Line in Representation of Tree Structure
DDDDD AAA TTTTTTTT RRRRRR EEEEEEEE EEEEEEEE
DD DD AAAA TT RR RR EE EE
DD DD AA AA TT RR RR EE EE
DD DD AA AA TT RRRRRR EEEEE EEEEE
DD DD AAAAAAA TT RR RR EE EE
DD DD AA AA TT RR RR EE EE
DDDDD AA AA TT RR RR EEEEEEEE EEEEEEEE
DATREE Returns Next Line in Representation of Tree Structure
------ ------- ---- ---- -- -------------- -- ---- ---------
Simple tree structures, such as the calling sequences of the
subroutines within a computer program, can be represented by
a typed (or printed) notation in which the root is placed in
the left column, those nodes which lie immediately above the
root are placed in the second column, those which are above
the nodes in the second column are placed in the third
column, and so on. As a typical example, the calling
hierarchy of the demonstration program presented in the
DALOAD documentation in this manual can be represented by
the following conventionally drawn representation in which
references to routines not in FASP have been excluded
DAHEFT
*
*
DASPAN
DANUMB *
* *
* DANEXT DANUMB DARITE
DALONE DANUMB ** * **
* * ** * **
* * ** * **
DALIST DALOAD DALONE DAPICK DAROLL DASITE DAVARY
*** ** * * * ** ***
*** ** * * * ** ***
*** ** * * * ** ***
*** ** * * * ** ***
DEMONSTRATION PROGRAM
If this tree is rotated clockwise 90 degrees, the calling
hierarchy of the program can be represented as shown below
MAIN DALIST DALONE DANUMB
MAIN DALOAD
MAIN DALONE DANUMB
MAIN DAPICK
MAIN DAROLL
MAIN DASITE
MAIN DAVARY DANEXT DASPAN DAHEFT
MAIN DAVARY DANUMB
MAIN DAVARY DARITE
FASP, FORTRAN Alphameric Subroutine Package Page 456
DATREE Returns Next Line in Representation of Tree Structure
An array prepared by the calling program and input to DATREE
as an argument describes the tree structure either in terms
of which items are connected to and lie above which items,
or else in terms of which items are connected to and lie
below which items. Each different item in the tree must be
identified in this array by a unique number. For the tree
shown on the previous page, each appearance of DALONE would
be identified by the same number, perhaps 4, but this number
would be different from that used for any other item in the
tree. Whether the numbers employed represent packed
character codes or are subscripts in some other array, or
whatever, will depend upon how these are used by the main
program. It is only necessary that DATREE, treating these
numbers as single precision integers, be able to compare the
item identifiers to find which refer to the same item.
For each item which extends the tree, the array describing
the tree contains the item identifier followed by the
identifiers of all those items which are either connected to
it and lie above it in the tree or which are connected to it
and lie below it in the tree. Regardless of which notation
is employed, each group of item identifiers is preceded by
the number of items included within the group. For example,
if the following numeric assignments are made
0=MAIN
1=DAHEFT 2=DALIST 3=DALOAD 4=DALONE 5=DANEXT 6=DANUMB
7=DAPICK 8=DARITE 9=DAROLL 10=DASITE 11=DASPAN 12=DAVARY
then the input array might contain the following if the
first item in each group is below (to the left of) the rest
8, 0(MAIN ), 2(DALIST), 3(DALOAD), 4(DALONE), 7(DAPICK),
9(DAROLL),10(DASITE),12(DAVARY)
2, 2(DALIST), 4(DALONE)
2, 4(DALONE), 6(DANUMB)
4,12(DAVARY), 5(DANEXT), 6(DANUMB), 8(DARITE)
2, 5(DANEXT),11(DASPAN)
2,11(DASPAN), 1(DAHEFT)
or else the input array might contain the following if the
first item in each group is above (to the right of) the rest
2, 1(DAHEFT),11(DASPAN)
2, 2(DALIST), 0(MAIN )
2, 3(DALOAD), 0(MAIN )
3, 4(DALONE), 0(MAIN ), 2(DALIST)
2, 5(DANEXT),12(DAVARY)
3, 6(DANUMB), 4(DALONE),12(DAVARY)
2, 7(DAPICK), 0(MAIN )
2, 8(DARITE),12(DAVARY)
2, 9(DAROLL), 0(MAIN )
2,10(DASITE), 0(MAIN )
2,11(DASPAN), 5(DANEXT)
2,12(DAVARY), 0(MAIN )
FASP, FORTRAN Alphameric Subroutine Package Page 457
DATREE Returns Next Line in Representation of Tree Structure
Each successive call to DATREE identifies to the calling
program which items are to appear upon the next line of the
typed representation of the tree. DATREE also specifies
which item within the line is the leftmost item which did
not appear within the previous line of the representation of
the tree. The tree display is more legible if the duplicate
appearances of the items at the left side of the tree are
removed as in the illustration shown below
MAIN DALIST DALONE DANUMB
. DALOAD
. DALONE DANUMB
. DAPICK
. DAROLL
. DASITE
. DAVARY DANEXT DASPAN DAHEFT
. . DANUMB
. . DARITE
The DATREE Argument List
--- ------ -------- ----
The argument list of routine DATREE is
SUBROUTINE DATREE(KLIMB ,KOMPAR,ITYPE ,MINNOD,MAXNOD,
1 NODES ,MINCLM,MAXCLM,NOWCLM,KOLUMN,INITAL,KIND ,
2 NEWCLM)
with the associated DIMENSION statement
DIMENSION NODES(MAXNOD),KOLUMN(MAXCLM),INITAL(MAXCLM)
The following arguments are used for input to this routine
and are returned unchanged.
KLIMB = 0, the entire tree is to be represented by the
current series of calls to this routine.
= 1, only the portion of the tree starting at the
item having an identification number in the NODES
array which is equal to the input value of KOMPAR
is to be represented.
KOMPAR = if KLIMB=1, then KOMPAR is equal to the number in
the NODES array which identifies the item at the
bottom (left end) of the portion of the tree which
is to be represented. Any items which are below
(to the left of) this item will not be included in
the representation.
ITYPE = 0, each group in the NODES array consists of the
number of items which are identified in the group
followed by a number which uniquely identifies an
item and then by numbers which uniquely identify
FASP, FORTRAN Alphameric Subroutine Package Page 458
DATREE Returns Next Line in Representation of Tree Structure
some or all of the items above (to the right of)
and connected to the first item. The NODES array
is terminated by a group containing only a single
zero. If item 10 is below (to the left of) and
connected to 11 and 12, and item 11 is below (to
the left of) and connected to 12 and 13, so that
the tree might be represented as
13
*
13 12 *
* * *
* * 11
* * * *
11 12 or * *
* * * *
* * 10 12
* * *
10 *
*
12
then the NODES array would contain
3, 10, 11, 12, 3, 11, 12, 13, 0
or even
2, 10, 11, 2, 11, 12, 2, 10, 12, 2, 11, 13, 0
and the KOLUMN array would then be returned
containing the subscripts of the NODES array
locations which in turn contain the following
values
10, 11, 12 after the 1st call to this routine
10, 11, 13 after the 2nd call to this routine
10, 12 after the 3rd call to this routine
It should be noted that if the applications of
ITYPE=0 and 1 are reversed, then this routine
returns the routes by which the tips of the
branches of the tree are reached, rather than the
routes extending from the root. For example, if
the above NODES array contents were interpreted
with ITYPE=1 rather than 0, then the following
values would be returned instead
12, 10 after the 1st call to this routine
12, 11, 10 after the 2nd call to this routine
13, 11, 10 after the 3rd call to this routine
= 1, each group in the NODES array consists of the
number of items identified in the group followed by
FASP, FORTRAN Alphameric Subroutine Package Page 459
DATREE Returns Next Line in Representation of Tree Structure
a number which uniquely identifies an item and then
by numbers which uniquely identify some or all of
items below (to the left of) and connected to the
first item. The NODES array is terminated by a
group containing only a single zero. For the above
example in which 12 is above and connected to both
10 and 11, in which 11 is above and connected to 10
and in which 13 above and connected to 11, the
nodes array would contain
3, 12, 10, 11, 2, 11, 10, 2, 13, 11, 0
MINNOD = subscript of the location in the NODES array which
contains the start of the description of the tree
or trees which are to be represented.
MAXNOD = dimension of the NODES array.
NODES = array containing the description of which items are
linked to which in the tree.
MINCLM = subscript of the first of the locations in the
KOLUMN array in which are to be returned the
subscripts of the NODES array locations which in
turn contain the numbers which uniquely identify
the items in the next line in the representation of
the tree.
MAXCLM = subscript of the final location in the KOLUMN array
which is available for use.
The following argument must be set by the calling program
before this routine is first called, but thereafter should
not be changed by the calling program since this argument is
used to transfer information from one call of this routine
to the next.
NOWCLM = must be set to less than the value of MINCLM (the
actual value doesn't matter since it is reset
inside the routine to MINCLM-1) by the calling
program before this routine is first called to
represent the tree or trees described by the NODES
array. If NOWCLM is input set to less than the
value of MINCLM, then the first line in the
description of the first tree (or of the tree
starting at a item having an identification number
equal to the value of KOMPAR if KLIMB=1) is to be
returned to the calling program. NOWCLM is
returned containing the highest subscript used in
the KOLUMN array to represent the current line and
must be sent to the subsequent call of this routine
unchanged. If NOWCLM is input set greater than or
equal to MINCLM, then the next line in the
description of the tree currently being represented
FASP, FORTRAN Alphameric Subroutine Package Page 460
DATREE Returns Next Line in Representation of Tree Structure
is to be returned to the calling program, or else,
if KLIMB is equal to zero and if the previous call
to this routine has completed the representation of
the current tree, then the first line of the
representation of the next tree is to be returned
to the calling program.
The first of the following arrays is used for returning
information to the calling program. The information
returned in both of these arrays must be supplied unchanged
to the subsequent call of this routine. The values in these
arrays are ignored when this routine is first called to
represent the trees described in the NODES array.
KOLUMN = array returned containing the subscripts of the
locations in the NODES array which identify those
items on the current line. The contents of this
array must not be changed between calls to this
routine until KIND is returned containing 1
indicating that the representation of the tree has
been completed. The contents of the KOLUMN array
are ignored when this routine is called with NOWCLM
less than MINCLM.
INITAL = array dimensioned the same as the KOLUMN array, but
which is used only for the transfer of values from
one call of this routine to the next. The contents
of this array must not be changed between calls to
this routine until KIND is returned containing 1
indicating that the representation of the tree has
been completed.
The following arguments are used only for returning
information to the calling program. Their input values are
ignored.
KIND = 1, returned if the representation of the tree had
been finished by the previous call to this routine.
= 2, returned if a line in the representation is
being returned in KOLUMN(MINCLM) through and
including KOLUMN(NOWCLM).
= 3, same as KIND=2 except that the representation is
terminated after the second appearance of a item
which has already appeared in the line.
= 4, same as KIND=2 except that there was not enough
space in the KOLUMN array to represent all of the
items which should be within the current line.
= 5, KLIMB was input containing 1 and NOWCLM
containing less than the value of MINCLM indicating
that the representation of a partial tree was
desired, but the item identified by KOMPAR could
not be found in the NODES array. No items are
being returned in KOLUMN array, and NOWCLM is
returned containing MINCLM-1.
FASP, FORTRAN Alphameric Subroutine Package Page 461
DATREE Returns Next Line in Representation of Tree Structure
NEWCLM = returned containing the subscript of the lowest
location in the KOLUMN array which has been
returned changed by the current call to this
routine. The input value of NEWCLM is ignored.
Demonstration Program to Interactively Test DATREE
------------- ------- -- ------------- ---- ------
The program listed below interprets each line of text typed
by the user as a list of one or more single-letter names of
those items which are to be included within a single group
in the description of the tree structure. The typing of an
empty line, rather than a list of the names of the items to
be placed into the tree, causes the program to display the
trees which would result if the tree description is
interpreted first with ITYPE=0, then with ITYPE=1. The user
can then request the display of sub-trees, again for both
values of ITYPE, by specifying, one per line, the letters
identifying the nodes to be used as roots of the sub-trees.
Typing of an empty line, instead of the letter identifying a
root, signals that the user is going to specify an entirely
new tree structure.
C PROGRAM TO DEMONSTRATE DATREE ROUTINE
DIMENSION IBUFFR(72),NODES(500),KODES(100),
1INITAL(100),KOLUMN(100)
DATA ITTY,JTTY/5,5/
C MAXKNO = DIMENSION OF KODES ARRAY AND MAXIMUM OF
C DIFFERENT LETTERS WHICH CAN BE RECOGNIZED.
DATA MAXKNO/100/
DATA MINCLM,MAXCLM,MINNOD,MAXNOD/12,100,20,500/
DATA IDOT,ISPACE/1H.,1H /
C
C ASK USER FOR DESCRIPTION OF TREE
NOWCLM=MINCLM-1
1 NODUSD=MINNOD-1
KNOWN=0
2 WRITE(JTTY,3)
3 FORMAT(2H *,$)
READ(ITTY,4)IBUFFR
4 FORMAT(100A1)
C
C INSERT SINGLE GROUP INTO NODES ARRAY
LOCAL=0
DO 8 I=1,72
IF(IBUFFR(I).EQ.ISPACE)GO TO 8
KOMPAR=0
5 KOMPAR=KOMPAR+1
IF(KOMPAR.GT.KNOWN)GO TO 6
IF(IBUFFR(I).NE.KODES(KOMPAR))GO TO 5
GO TO 7
6 IF(KNOWN.LT.MAXKNO)KNOWN=KNOWN+1
KOMPAR=KNOWN
FASP, FORTRAN Alphameric Subroutine Package Page 462
DATREE Returns Next Line in Representation of Tree Structure
KODES(KNOWN)=IBUFFR(I)
7 IF(LOCAL.EQ.0)NODUSD=NODUSD+1
LOCAL=LOCAL+1
NODUSD=NODUSD+1
NODES(NODUSD)=KOMPAR
8 CONTINUE
IF(LOCAL.LE.0)GO TO 9
I=NODUSD-LOCAL
NODES(I)=LOCAL
GO TO 2
C
C ADD TERMINAL ZERO TO NODES ARRAY
9 NODUSD=NODUSD+1
NODES(NODUSD)=0
C
C CYCLE THROUGH DISPLAY FOR ITYPE=0 AND 1
KLIMB=0
10 DO 15 JTYPE=1,2
NOWCLM=MINCLM-1
ITYPE=JTYPE-1
11 CALL DATREE(KLIMB,KOMPAR,ITYPE,MINNOD,MAXNOD,
1NODES,MINCLM,MAXCLM,NOWCLM,KOLUMN,INITAL,KIND,
2NEWCLM)
IF(KIND.EQ.1)GO TO 14
IF(KIND.EQ.5)GO TO 14
K=0
DO 12 I=MINCLM,NOWCLM
K=K+1
IBUFFR(K)=IDOT
J=KOLUMN(I)
J=NODES(J)
12 IF(I.GE.NEWCLM)IBUFFR(K)=KODES(J)
WRITE(JTTY,13)KIND,NEWCLM,NOWCLM,(IBUFFR(I),I=1,K)
13 FORMAT(1X,I1,2I3,1X,100A1)
GO TO 11
14 WRITE(JTTY,13)KIND,NEWCLM,NOWCLM
15 CONTINUE
C
C ASK USER FOR ORIGIN OF PARTIAL DISPLAY
WRITE(JTTY,16)
16 FORMAT(11H DETAIL OF ,$)
READ(ITTY,4)LETTER
IF(LETTER.EQ.ISPACE)GO TO 1
KLIMB=1
KOMPAR=0
17 KOMPAR=KOMPAR+1
IF(KOMPAR.GT.KNOWN)GO TO 10
IF(LETTER.EQ.KODES(KOMPAR))GO TO 10
GO TO 17
END
FASP, FORTRAN Alphameric Subroutine Package Page 463
DATREE Returns Next Line in Representation of Tree Structure
Typical Dialog Between DATREE Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
*ABCD
*BGI
*CIB
*DM
*MN
*IJK
*GHI
*
2 12 15 ABGH The values of KIND, NEWCLM and NOWCLM
2 15 16 ...IJ are displayed to the left of the
2 16 16 ....K representation of the contents of the
2 14 15 ..IJ tree itself. KIND is returned twice set
2 15 15 ...K to 1, first when the representation with
2 13 15 .CIJ ITYPE=0 has been completed, then when
2 15 15 ...K the representation with ITYPE=1 has been
2 14 16 ..BGH completed.
2 16 17 ....IJ
2 17 17 .....K
2 15 16 ...IJ
2 16 16 ....K
2 13 15 .DMN
1 12 11
2 12 15 NMDA
2 12 15 JIBA
2 15 16 ...CA
2 14 15 ..CA
2 14 16 ..GBA
2 16 17 ....CA
2 12 15 KIBA
2 15 16 ...CA
2 14 15 ..CA
2 14 16 ..GBA
2 16 17 ....CA
2 12 15 HGBA
2 15 16 ...CA
1 12 11
DETAIL OF C
2 12 14 CIJ
2 14 14 ..K
2 13 15 .BGH
2 15 16 ...IJ
2 16 16 ....K
2 14 15 ..IJ
2 15 15 ...K
1 12 11
2 12 13 CA
1 12 11
DETAIL OF
FASP, FORTRAN Alphameric Subroutine Package Page 464
DATREE Returns Next Line in Representation of Tree Structure
*JI This is the same tree structure as is
*KI shown above, but with the linkages
*GB reversed so that now ITYPE=1 produces
*NM the same representation as was generated
*MD above for ITYPE=0.
*CA
*BAC
*DA
*IBCG
*HG
*
2 12 15 JIBA
2 15 16 ...CA
2 14 15 ..CA
2 14 16 ..GBA
2 16 17 ....CA
2 12 15 KIBA
2 15 16 ...CA
2 14 15 ..CA
2 14 16 ..GBA
2 16 17 ....CA
2 12 15 NMDA
2 12 15 HGBA
2 15 16 ...CA
1 12 11
2 12 17 ACBGIJ
2 17 17 .....K
2 16 16 ....H
2 15 16 ...IJ
2 16 16 ....K
2 14 15 ..IJ
2 15 15 ...K
2 13 16 .BGIJ
2 16 16 ....K
2 15 15 ...H
2 14 15 ..IJ
2 15 15 ...K
2 13 15 .DMN
1 12 11
DETAIL OF C
2 12 13 CA
1 12 11
2 12 16 CBGIJ
2 16 16 ....K
2 15 15 ...H
2 14 15 ..IJ
2 15 15 ...K
2 13 14 .IJ
2 14 14 ..K
1 12 11
DETAIL OF
FASP, FORTRAN Alphameric Subroutine Package Page 465
DATREE Returns Next Line in Representation of Tree Structure
*AB This cyclic structure has no root for
*BC ITYPE=0, but nodes O and N are both
*CDH roots for ITYPE=1. N O
*DE < >
*EF < >
*FA A > > B I > > J
*HI > > > >
*IJN > > > >
*JKO F C > > H K
*KL < < < <
*LM < < < <
*MH E < < D M < < L
*
1 12 11 Effective structure for ITYPE=0 and node
3 12 21 NIHCBAFEDC F as root is shown below. Rings are
3 15 19 ...MLKJI terminated when they return to starting
3 12 22 OJIHCBAFEDC nodes F and H. N O
3 16 19 ....MLKJ < >
1 12 11 < >
DETAIL OF F A > > B I > > J
3 12 18 FABCDEF > > > >
3 16 22 ....HIJKLMH *** > > >
2 19 19 .......O *F* C > > H K
2 18 18 ......N *** < <
1 12 11 < <
3 12 18 FEDCBAF F < < E < < D H < < M < < L
1 12 11
DETAIL OF C F > > A > > B Right ring is not
3 12 18 CDEFABC > included when ITYPE=1
3 13 19 .HIJKLMH *** > and root is in left
2 16 16 ....O *F* C ring.
2 15 15 ...N *** <
1 12 11 < <
3 12 18 CBAFEDC E < < D
1 12 11
DETAIL OF H Both rings are included if ITYPE=0 and
3 12 18 HIJKLMH node C is root. N O
2 15 15 ...O < >
2 14 14 ..N < >
1 12 11 A > > B > > C I > > J
3 12 19 HCBAFEDC > > >
3 13 18 .MLKJIH > *** > >
1 12 11 F *C*> > H K
DETAIL OF K < *** <
3 12 18 KLMHIJK < < <
2 18 18 ......O E < < D H < < M < < L
2 17 17 .....N
1 12 11 A > > B Right ring is not
3 12 22 KJIHCBAFEDC > > included if ITYPE=1
3 16 18 ....MLK > *** and root is in left
1 12 11 F *C* ring.
DETAIL OF < ***
<
E < < D < < C
FASP, FORTRAN Alphameric Subroutine Package Page 466
DATREK, Routine to Evaluate Integer Sequence of Form 1.2.3
DDDDD AAA TTTTTTTT RRRRRR EEEEEEEE KK KK
DD DD AAAA TT RR RR EE KK KK
DD DD AA AA TT RR RR EE KK KK
DD DD AA AA TT RRRRRR EEEEE KKKKK
DD DD AAAAAAA TT RR RR EE KKK KK
DD DD AA AA TT RR RR EE KK KK
DDDDD AA AA TT RR RR EEEEEEEE KK KK
DATREK, Routine to Evaluate Integer Sequence of Form 1.2.3
------ ------- -- -------- ------- -------- -- ---- -----
DATREK evaluates a sequence of unsigned integers,
represented by groups of digits separated by periods, typed
by the user and read by the calling program with a multiple
of an A1 format. Each integer returns a value of zero or
larger. The value -1 is returned for each integer indicated
as missing by an initial period, by a trailing period, or by
2 adjacent periods. Signs and exponents are not recognized.
The DATREK Argument List
--- ------ -------- ----
The argument list of routine DATREK is
SUBROUTINE DATREK(LOWVLU,MAXVLU,MAXBFR,IBUFFR,LOWBFR,
1 KIND ,IVALUE,KNTVLU)
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR),IVALUE(MAXVLU)
The following arguments are used for input only and are
returned unchanged.
LOWVLU = subscript of the lowest location in the IVALUE
array which can be used to return the values of the
integers in the sequence. If the input text buffer
is empty or if the first printing character at or
to the right of IBUFFR(LOWBFR) is neither a digit
nor a period, then the IVALUE array is not used and
KNTVLU which is returned pointing to the highest
location used in the IVALUE array is instead
returned containing the value LOWVLU-1.
MAXVLU = subscript of the highest location in the IVALUE
array which can be used to return the values of the
integers in the sequence. If the integer sequence
contains more values than can be stored in
IVALUE(LOWVLU) through IVALUE(MAXVLU), then the
evaluation of the sequence continues through the
FASP, FORTRAN Alphameric Subroutine Package Page 467
DATREK, Routine to Evaluate Integer Sequence of Form 1.2.3
terminal digit or period, but the rightmost values
are not stored in the IVALUE array, and KNTVLU is
returned containing MAXVLU.
MAXBFR = subscript of the final (rightmost) location in the
IBUFFR array which can be searched for the
characters which represent the integer sequence.
IBUFFR = input buffer containing characters typed by the
user, read by a multiple of an A1 format, which are
to be searched for the representation of the
integer sequence. IBUFFR then contains 1 character
per computer storage location.
The following argument must be set by the calling program
before this routine is first called, then is returned by
this routine describing the location of the first character
not yet processed in the input buffer.
LOWBFR = subscript within the IBUFFR array of the first
(leftmost) character which can be scanned for an
integer sequence. If an integer sequence is found,
then LOWBFR will be returned pointing to the first
character which is beyond the sequence, and which
could not be part of the sequence. If an integer
sequence is not found, then LOWBFR will be returned
pointing to the first printing character which
would be neither a digit nor a period, or will be
returned pointing beyond the end of the buffer if
the buffer does not contain any printing characters
at or to the right of IBUFFR(LOWBFR). LOWBFR must
be set by the calling program before this routine
is first called to process the contents of the
buffer, and should not be changed by the calling
program if this routine locates an integer
sequence. If an initial printing character is
found which is neither a digit nor a period, then
the contents of the buffer should be processed by
some other FASP routine, or at least the calling
program should increment the value of LOWBFR by 1
before again calling this routine.
The following arguments are used for returning information
to the calling program. Their input values are ignored.
KIND = returned describing the kind of item located in the
IBUFFR array.
= 1, nothing was found at or to the right of
IBUFFR(LOWBFR). The calling program should read a
new line into the IBUFFR array and reset LOWBFR to
point to the first character in this array. KNTVLU
is returned containing LOWVLU-1.
= 2, an integer sequence was not found, but a
printing character which is neither a digit nor a
FASP, FORTRAN Alphameric Subroutine Package Page 468
DATREK, Routine to Evaluate Integer Sequence of Form 1.2.3
period was found at or to the right of
IBUFFR(LOWBFR). LOWBFR will be returned pointing
to this printing character. The calling program
must increment the value of LOWBFR before again
calling this routine since supplying the same
initial character would produce identical results.
KNTVLU is returned containing LOWVLU-1.
= 3, an integer sequence was found which did not
contain more values, or indications of missing
values, than could be returned in the IVALUE array.
LOWBFR is returned pointing to the next character
beyond the end of the sequence.
= 4, an integer sequence was found which contained
more values, or indications of missing values, than
could be returned in the IVALUE array. The
evaluation of the sequence continues through the
terminal digit or period, but the rightmost values
are not stored in the IVALUE array, and KNTVLU is
returned containing MAXVLU. LOWBFR is returned
pointing to the next character beyond the end of
the sequence.
IVALUE = array in which the values of the integers are
returned. If an integer is indicated as being
missing by an initial period, by a trailing period,
or by 2 adjacent periods, then the value -1 is
instead stored in the corresponding location in the
IVALUE array.
KNTVLU = returned containing the subscript of the highest
location in the IVALUE array used either to store
the value of an integer in the sequence, or to
indicate a missing integer. If the first printing
character in the input buffer is neither a digit
nor a period, then KNTVLU is returned containing
LOWVLU-1. If the integer sequence contains more
values than can be stored in IVALUE(LOWVLU) through
IVALUE(MAXVLU), then the evaluation of the sequence
continues through the terminal digit or period, but
the rightmost values are not stored in the IVALUE
array, and KNTVLU is returned containing MAXVLU.
FASP, FORTRAN Alphameric Subroutine Package Page 469
DATREK, Routine to Evaluate Integer Sequence of Form 1.2.3
An Example of the Use of DATREK
-- ------- -- --- --- -- ------
C PROGRAM TO DEMONSTRATE DATREK ROUTINE
C
DATA ITTY,JTTY/5,5/
DATA LOWVLU,MAXVLU,MAXBFR/1,5,60/
DIMENSION IBUFFR(60),IVALUE(5)
C
C ASK FOR AND ACCEPT USER RESPONSE
1 LOWBFR=1
WRITE(ITTY,2)
2 FORMAT(2H *$)
READ(JTTY,3)IBUFFR
3 FORMAT(60A1)
C
C EVALUATE NEXT INTEGER SEQUENCE IN USER RESPONSE
4 CALL DATREK(LOWVLU,MAXVLU,MAXBFR,IBUFFR,LOWBFR,
1 KIND ,IVALUE,KNTVLU)
IF(KIND.EQ.1)GO TO 1
IF(KIND.EQ.2)GO TO 7
C
C REPORT VALUES IN SEQUENCE
IF(KIND.EQ.3)WRITE(ITTY,5)(IVALUE(I),I=1,KNTVLU)
5 FORMAT(9H SEQUENCE,10I10)
IF(KIND.EQ.4)WRITE(ITTY,6)(IVALUE(I),I=1,KNTVLU)
6 FORMAT(9H OVERFLOW,10I10)
GO TO 4
C
C REPORT UNKNOWN CHARACTER
7 WRITE(ITTY,8)IBUFFR(LOWBFR)
8 FORMAT(11H UNKNOWN ',1A1,1H')
LOWBFR=LOWBFR+1
GO TO 4
END
Typical Dialog Between User and DATREK Demonstration Program
------- ------ ------- ---- --- ------ ------------- -------
*12 12. 12.34 12.0034 12.00340
SEQUENCE 12
SEQUENCE 12 -1
SEQUENCE 12 34
SEQUENCE 12 34
SEQUENCE 12 340
*12..56..90 . .. .34.-0.1.2.3.4.5.6.7.8.9
SEQUENCE 12 -1 56 -1 90
SEQUENCE -1 -1
SEQUENCE -1 -1 -1
SEQUENCE -1 34 -1
UNKNOWN '-'
OVERFLOW 0 1 2 3 4
FASP, FORTRAN Alphameric Subroutine Package Page 470
DAVARY, Routine to Display and Assign Values to Named Arrays
DDDDD AAA VV VV AAA RRRRRR YY YY
DD DD AAAA VV VV AAAA RR RR YY YY
DD DD AA AA VV VV AA AA RR RR YYYY
DD DD AA AA VV VV AA AA RRRRRR YY
DD DD AAAAAAA VVVV AAAAAAA RR RR YY
DD DD AA AA VVV AA AA RR RR YY
DDDDD AA AA VV AA AA RR RR YY
DAVARY, Routine to Display and Assign Values to Named Arrays
------ ------- -- ------- --- ------ ------ -- ----- ------
DAVARY is used along with several other routines in the FASP
package for the purpose of specifying by name, examining and
modifying the values of multiply subscripted arrays
equivalenced with or otherwise loaded into a singly
subscripted buffer. The interaction between these routines
is described at the start of the DALOAD documentation. The
sample program at the end of the DALOAD documentation
illustrates the use of most of these routines.
Prior to passing control to DAVARY, the calling program will
have requested from the user a specification of the array
locations which the user wants to examine and/or to modify.
The text typed by the user in response to this request is
evaluated by the DAPICK routine to identify the array name
and its subscript ranges, if any. DAVARY is called
separately to display each of the values in the array
locations within the range of array locations indicated by
the user. If DAPICK found that the user typed an equals
sign to the right of the original array specification, then
the calling program should request that DAVARY is to
evaluate the numbers and/or number range specifications
typed to the right of the equals sign and assign these as
the new values of the array locations within the range of
array locations indicated by the user. If too few values
were typed right of the equals sign to define new values for
all of the array locations within the total range of array
locations indicated by the user, then, when the supply of
typed values has been exhausted, DAVARY will ask the user to
specify the value of each of the remaining array locations
as the calling program steps through the range of array
locations. When the user is asked to specify the new value,
the user has the option of specifying no value in which case
the former value will be retained, of specifying a range of
values in which case DAVARY will upon each successive call
select values from the range until the range is completed,
of specifying that some number of array locations are to be
skipped over retaining their former values in the process,
or of abandoning the interaction altogether for the current
range of array locations.
The text typed by the user, either to the right of the
FASP, FORTRAN Alphameric Subroutine Package Page 471
DAVARY, Routine to Display and Assign Values to Named Arrays
equals sign in the buffer evaluated by DAPICK, or in the
text read directly by the DAVARY routine, consists of
numbers and/or number range specifications. If the user has
typed several numbers and/or range specifications on the
same line, then DAVARY will evaluate each of these in turn
when the previous calls to DAVARY have used the previous
value or have used all of the values within the previous
range. Any number of spaces and/or tab characters, but at
most a single comma, can separate two adjacent numbers
and/or number range specifications. Two adjacent commas,
possibly separated by spaces and/or tabs, indicate that the
array location being assigned a value when the
interpretation of the buffer has proceeded to the evaluation
of the missing number between the commas is to instead
retain its former value. If a comma is the first printing
character to the right of the equals sign in the text
evaluated by DAPICK, then the first array location in the
range of array locations indicated by the user, is to retain
its former value. Similarly, if a comma is the first
printing character in the buffer read by DAVARY itself, then
the array location currently being defined is to retain its
former value. A comma which is the rightmost printing
character in the buffer (except possibly for semicolons
and/or an exclamation point and its following comment or an
ampersand and its following comment) indicates that the
array location being assigned a value when the
interpretation of the buffer has proceeded to the evaluation
of the text to the right of the comma is instead to retain
its former value. A single comma with no other printing
characters to its left (except possibly the array
specification and the equals sign found by DAPICK) and with
no printing characters to its right (except possibly for
semicolons and/or an exclamation point and its following
comment or an ampersand and its following comment),
indicates two array locations which are to retain their
former values.
Numbers, whether used to specify single values or ranges,
can be typed by the user in either floating point form or
scientific notation. If scientific notation is used for the
specification of octal numbers, then the exponent to the
right of the letter E (which must be typed in upper case) is
evaluated as the decimal number of octal digits to be
shifted. If an integer (either octal or decimal) is
specified with a decimal point and/or with an exponent, then
the value is first obtained as though neither the decimal
point nor exponent appeared, and this value is shifted only
once by the amount indicated by the combination of decimal
point location and exponent. If the number 1.2345E3 is
typed by the user and is evaluated as a decimal integer,
then the value is calculated to be 12345 before being
divided by 10 to obtain the returned value of integer 1234.
A range of values can be specified as the limits between
FASP, FORTRAN Alphameric Subroutine Package Page 472
DAVARY, Routine to Display and Assign Values to Named Arrays
which and increment by which the value is to be varied, or
can be specified as the repetition of a single value. If
the value is to be varied between limits, then the range is
written as the starting value, the increment, and the ending
value separated by slashes or by colons (the 2 characters
are equivalent). If the increment is to be one, then the
starting and ending values need be separated only by a
single slash or by a single colon. If either limit is
missing, it is assumed to have the value zero. The
effective sign of the increment is changed if it does not
agree with the relative values of the limits.
If the user wants to specify that a single value is to be
assigned to several array locations within the range of
array locations being defined, then the range of values is
written as the number of times the value is to be used
followed immediately by an asterisk and the value. If a
value does not follow the asterisk, then the user is
requesting that the stated number of consecutive locations
starting with the location currently being defined are
instead to have their former values kept intact.
If an exclamation point or an ampersand is found anywhere in
the text typed by the user, then the exclamation point or
the ampersand and all characters to its right are taken to
be a comment and are otherwise ignored. If in response to a
request for a new value, the user types either a completely
blank line or a line in which the leftmost printing
character is an exclamation point, then the former value of
the array location will be retained, and the next call to
DAVARY will request the value of the next array location.
If in response to a request for a new value, the user
instead types a line in which the leftmost printing
character is an ampersand, then the remainder of the line is
similarly taken as a comment, but the user is again asked to
specify the value. Use of the ampersand as the leftmost
printing character allows multiple line comments for
annotation of batch command files. The exclamation point
and the ampersand are equivalent if any printing character
is typed to their left.
If the user types a semicolon, then DAVARY signals to the
calling program that the user has requested that the
stepping through the array locations within the range of
array locations is to be terminated.
The text
5.3E3,-0.00728,4*,1E6/1E5/2E6,,12*.7;
would, if typed by the user and interpreted by DAVARY,
specify
A. that the first array location is to be assigned the
FASP, FORTRAN Alphameric Subroutine Package Page 473
DAVARY, Routine to Display and Assign Values to Named Arrays
value 5.3E3
B. that the second array location is to be assigned the
value -7.28E-3
C. that the next 4 array locations are to keep their
former values
D. that the next eleven array locations are to be assigned
the values 1.0E6, 1.1E6, 1.2E6 through 2.0E6
E. that the next array location is retain its former value
F. that the next 12 array locations are to all be assigned
the value 0.7
G. and that the stepping through array locations within
the current range of array locations is then to be
terminated.
Spaces and/or tab characters could appear in addition to the
commas in the example. The only commas in the example which
could not be replaced by spaces and/or tab characters are
the 2 adjacent commas which indicate that the corresponding
array location is to retain its former value.
The DAVARY Argument List
--- ------ -------- ----
The argument list of routine DAVARY is
SUBROUTINE DAVARY(KONTRL,ITTY ,JTTY ,LOCATN,NAMUSD,
1 NAMMAX,MAXBFR,NOTATN,MINDEC,MAXDEC,MINSIG,MAXSIG,
2 IDECML,AARRAY,IARRAY,NAME ,IBUFFR,LOWBFR,KIND ,
3 MODIFY)
with the associated DIMENSION statement
DIMENSION NAME(NAMMAX),AARRAY(LOCATN),IARRAY(LOCATN),
1IBUFFR(MAXBFR)
The following are input arguments left unchanged.
KONTRL = indicates the type (octal, integer or real) of the
value to be displayed and possibly modified.
= -1, this routine is to display in octal and
possibly modify the integer value in
IARRAY(LOCATN).
= 0, this routine is to display in decimal and
possibly modify the integer value in
IARRAY(LOCATN).
= greater than zero, this routine is to display in
decimal, using either floating point form or
FASP, FORTRAN Alphameric Subroutine Package Page 474
DAVARY, Routine to Display and Assign Values to Named Arrays
scientific notation as specified by the other
arguments or as necessary, and possibly modify the
real value in AARRAY(LOCATN).
During interpretation of the values typed by the
user, KONTRL is identical to the argument of the
same name for the DAHEFT routine, and is defined as
follows.
= -1, the item typed by the user is an octal number.
The number can be typed with a decimal point and/or
with an exponent. However, the number following
the letter E of the exponent is evaluated in
decimal. It must be noted that numbers evaluated
as negative octal integers have the negative octal
integer as their value, not as their bit
representation in computer storage. For example,
on a 36 bit twos complement computer, the octal
number -400000000000 (which could also be typed as
-4E11 or -4E+11 where the 11 after the E is in
decimal) is represented as bit pattern having octal
notation 400000000000 and the octal number
-377777777777 is represented by the bit pattern
400000000001.
= 0, the item typed by the user is an integer decimal
number. The number can be typed with a decimal
point (for example 1.23K or 1.23E3 equals 1230),
but is stored as an integer. Any decimal integer
which the computer can represent can be evaluated.
This includes, on twos complement computers, the
largest negative number the absolute value of which
cannot be stored. On the PDP-10, a 36 bit computer
with twos complement notation, the range of decimal
integers is -34359738368 through 34359738367 (octal
notation of bit patterns being 400000000000 through
377777777777).
= 1 or greater, the item typed by the user is a real
number. If possible, the real number will be
accumulated as an integer, then be converted to a
real number and shifted if necessary. KONTRL is
then the maximum number of digits in the integer.
If the item has more than KONTRL digits (not
counting leftmost zeroes), then the entire
evaluation is done as a real number. The advantage
of calculating the real values in integer as long
as the precision of the computer is not overflowed
is that the calculation of the portion of the
number right of the decimal point is more exact.
As an example, if KONTRL is greater than or equal
to 4, then the number 33.33 can be stored as the
integer 3333, then be converted to the real value
3333.0 and divided by 100.0 to obtain the final
answer. If it makes no difference whether the
number typed as 33.33 has value 33.33 or
33.32999... then KONTRL can be given the value 1.
FASP, FORTRAN Alphameric Subroutine Package Page 475
DAVARY, Routine to Display and Assign Values to Named Arrays
ITTY = number of the unit from which this routine is to
read the user's specification of new values. On
the PDP-10 computer, ITTY should have the value 5.
JTTY = number of the unit to which this routine is to
write the representation of the value currently in
AARRAY(LOCATN) or IARRAY(LOCATN), and to which
error messages are to be sent. On the PDP-10
computer, JTTY should have the value 5.
LOCATN = subscript of the array location (within either the
AARRAY or IARRAY arrays, whichever is selected by
the value of KONTRL) of which the value is to be
shown to the user and possibly modified.
NAMUSD = subscript of the NAME array location containing the
rightmost character of the representation of the
name and associated subscripts of the multiply
subscripted array being simulated in the singly
subscripted buffer AARRAY or IARRAY. The
representation of the name starts at NAME(1). If
the current call to DAVARY is displaying and
possibly modifying the contents of the doubly
subscripted simulated array named ABC, and if the
subscripts of the simulated array ABC which would
correspond to the subscript LOCATN in the singly
subscripted buffer would be 3 and 7, then the
contents of the NAME(1) through and including
NAME(NAMUSD) where NAMUSD is 8 would be identical
to that assigned by the statement
DATA (NAME(I),I=1,8)/1HA,1HB,1HC,
11H(,1H3,1H,,1H7,1H)/
When DAVARY is used with the other routines in the
FASP package, the contents of the NAME array would
be defined by calling the DALONE routine prior to
each call to DAVARY. The dimension of the NAME
array, or at least the maximum usable area within
this array, is given by the argument NAMMAX. The
NAME array locations starting at subscript
NAMUSD+1, and extending through subscript NAMMAX,
are used for storing the representation of the
value in the currently selected location within the
AARRAY or IARRAY array, and for storing the
representation of the new value if this new value
is based upon the user's response read prior to the
current call to DAVARY. NAMMAX should be at least
30 to 40 greater than NAMUSD.
NAMMAX = subscript of the highest location within the NAME
array which can be used as a scratch area for the
construction of the character representation of the
value in the currently selected location within the
FASP, FORTRAN Alphameric Subroutine Package Page 476
DAVARY, Routine to Display and Assign Values to Named Arrays
AARRAY or IARRAY array, and for the construction of
the representation of the new value if this new
value is based upon the user's response read prior
to the current call to DAVARY. NAMMAX should be at
least 30 to 40 greater than NAMUSD.
MAXBFR = subscript of the highest location within the IBUFFR
array into which the calling program or a previous
call to this routine has read characters typed by
the user. If this routine must read a new response
from the user, then the characters of the response
are read into IBUFFR(1) through IBUFFR(MAXBFR).
Since the format used by DAVARY to read the
response is 100A1, MAXBFR should not exceed 100.
The following arguments are sent to the DARITE routine to
control the display of the former value in AARRAY(LOCATN) or
IARRAY(LOCATN) and of the new value if this new value is
based upon the user's response read prior to the current
call to DAVARY. All of these arguments correspond to the
DARITE arguments of the same name, and all are returned
unchanged. These arguments are ignored if KONTRL is less
than or equal to zero. It should be noted that where the
following descriptions speak of the display of a particular
number of characters, this is the number of characters which
could be displayed if all are nonzero. After rounding and
prior to actual display, rightmost zeroes to the right of
the decimal point are discarded as is the decimal point
itself if no nonzero digits appear to its right.
NOTATN = -1, represent the value in scientific notation.
The value 1234 would be represented as 1.234E3
= 0, represent the value in floating point form. The
value 1234.56 would be displayed as it is written
if the number of digits requested right of the
decimal point is 2 and if at least 6 significant
digits are allowed.
= 1, multiply the value by 100 and insert the percent
sign following the digits of the number. MINDEC
and MAXDEC refer to the displayed decimal point.
To print tenths of a percent, MAXDEC would be given
the value 1, and MINDEC would be 1 or less.
= 2, if the value is in the range 1000 to 1000000,
then divide the value by 1000 and insert the letter
K following the digits. If the value is 1000000 or
greater, then divide the value by 1000000 and
insert the letter M following the digits. MINDEC
and MAXDEC refer to the displayed decimal point.
MINDEC and MAXDEC both set at 2 would represent the
value 1234 as 1.23K.
= 3, same as NOTATN=2 except that MINDEC and MAXDEC
refer to the decimal point in the original value,
not to the decimal point in the displayed number.
If a K or M is displayed right of the number, then
FASP, FORTRAN Alphameric Subroutine Package Page 477
DAVARY, Routine to Display and Assign Values to Named Arrays
MINDEC and MAXDEC values of -1 are equivalent to
values of 0. MINDEC=-2 indicates that there is no
lower limit to the number of digits to the right of
the displayed decimal point, and is probably the
appropriate value unless it is absolutely necessary
to display the digits which would be to the right
of the decimal point in the original value, as for
example if the amounts are dollars and must be
displayed always including the cents digits. The
value 1234.56 would be represented as 1.23456K for
NOTATN=3 and MAXDEC=2.
MINDEC = minimum number of digits right of the displayed
decimal point in numbers which are displayed in
floating point form. If it would be necessary to
represent less digits right of the decimal point
than indicated by MINDEC, then the value will be
represented in scientific notation instead. The
maximum number of digits right of the decimal point
is determined by the combination of MAXDEC and
MAXSIG.
= -2, there is no lower limit to the number of digits
which must be represented right of the decimal
point, and even the decimal point itself need not
be represented. MINDEC=-1 is equivalent to
MINDEC=-2 unless NOTATN=3 while the value is 1000
or greater.
= -1, if NOTATN is not 3, or if NOTATN is 3 but the
value being represented is less than 1000, then
MINDEC=-1 is equivalent to MINDEC=-2 such that
there is no lower limit to the number of digits
which must be represented right of the decimal
point, and even the decimal point itself need not
be represented.
If NOTATN=3 and the value is 1000 or greater, then
MINDEC=-1 would have the same effect as MINDEC=0
such that all digits to the left of the decimal
point in the original unshifted value (as
represented without K or M) must be included in the
actual representation of the value even when the K
or M is included. MINDEC should have the value -2
if, in order to fit the representation into the
available field width while NOTATN=3, it is to be
possible that a value which is equal to or greater
than 1000 be represented without all of the digits
which would be to the left of the decimal point in
the original value, or even without the decimal
point.
= equal or greater than zero, if NOTATN is not 3, or
if NOTATN is 3 but the value being represented is
less than 1000, then MINDEC is the minimum number
of digits which can be displayed right of the
decimal point in a floating point number. If less
FASP, FORTRAN Alphameric Subroutine Package Page 478
DAVARY, Routine to Display and Assign Values to Named Arrays
than MINDEC digits would be displayed right of the
decimal point, then the value will be represented
in scientific notation instead. If MINDEC=0, then
it is not necessary that any digits be displayed
right of the decimal point, but the decimal point
itself must be displayed.
If NOTATN=3 and the value is 1000 or greater, then
MINDEC is the minimum number of digits to the right
of the decimal point in the original value which
must be displayed. If the value is between 1000
and 1000000, then at least MINDEC+3 digits must be
displayed right of the decimal point. If the value
is 1000000 or greater, then at least MINDEC+6
digits must be displayed right of the decimal
point.
For example, the value 1234.56 would be represented
as follows for various values of MINDEC and NOTATN.
It should be noted that when the digit 5 right of
the decimal point in the original value is not
represented, then the digit 4 left of the decimal
point is rounded upwards to 5.
NOTATN=0 NOTATN=2 NOTATN=3
for MINDEC=MAXDEC=0 1235 1K 1.235K
MINDEC=MAXDEC=1 1234.6 1.2K 1.2346K
MINDEC=MAXDEC=2 1234.56 1.23K 1.23456K
MAXDEC = maximum number of digits right of the displayed
decimal point in numbers displayed in floating
point form.
= -2, represent as many digits right of the decimal
point as the field will hold (up to a maximum total
number of digits specified by MAXSIG).
= -1, if NOTATN is not 3, or if NOTATN is 3 but the
value being represented is less than 1000,
represent only digits left of the decimal point.
The decimal point itself will not be represented.
If more than MAXSIG digits would appear left of the
decimal point, then the number will be represented
in scientific notation.
If NOTATN=3 and the value is 1000 or greater, then
MAXDEC=-1 would have the same effect as MAXDEC=0
such that all digits which would be to the left of
the decimal point in the original value (as
represented without K or M) will if possible be
included in the representation of the value shifted
for the K or M notation. If MINDEC is greater than
-2, and if not all of the digits to the left of the
decimal point in the original unshifted value can
be included in the supplied field width, then the
value will instead be represented in scientific
FASP, FORTRAN Alphameric Subroutine Package Page 479
DAVARY, Routine to Display and Assign Values to Named Arrays
notation.
= greater than or equal to zero, if NOTATN is not 3,
or if NOTATN is 3 but the value being represented
is less than 1000, represent the value with MAXDEC
digits right of decimal point. If this
representation of the value will not fit into the
field size indicated by MAXSIG, then reduce the
number of digits represented right of the decimal
point so that the representation will fit, or
represent the value in scientific notation if there
are more than MAXSIG digits left of the decimal
point.
If NOTATN=3 and the value is 1000 or greater, then
MAXDEC is the number of digits to the right of the
decimal point in the original unshifted value which
can be displayed if there is room for these in the
supplied field width. If the value is between 1000
and 1000000, then at most MAXDEC+3 digits can be
displayed right of the displayed decimal point. If
the value is 1000000 or greater, then at most
MAXDEC+6 digits can be displayed right of the
decimal point.
MINSIG = minimum number of significant digits in the
floating point representation of the value. If the
floating point representation of the value would
contain less than MINSIG significant digits, then
the value will be represented in scientific
notation. If the value being displayed has the
value zero, then MINSIG is ignored.
MAXSIG = maximum number of significant digits in the
floating point representation of the value. MAXSIG
is the maximum number of digits which can be
displayed starting with the leftmost nonzero digit,
counting it and all digits to its right. MAXSIG
does not include the decimal point, does not
include the minus sign if the value is negative,
and does not include the percent sign, K or M if
NOTATN is greater than zero. The number of digits
displayed right of the decimal point is reduced if
necessary so that the number of digits starting at
the leftmost nonzero displayed digit and counting
it and all digits displayed to its right does not
exceed MAXSIG. If MAXSIG would be less than the
number of digits left of the decimal point in the
representation of the value, then the value will be
represented in scientific notation.
IDECML = same as MAXDEC except that IDECML applies only to
values displayed in scientific notation. Note that
if the value is being represented in scientific
notation, then a nonzero digit will be used left of
FASP, FORTRAN Alphameric Subroutine Package Page 480
DAVARY, Routine to Display and Assign Values to Named Arrays
the decimal point unless the value is itself zero.
Therefore, if IDECML is greater than or equal to
zero, then the maximum number of significant digits
which can be displayed in scientific notation is
IDECML+1.
The following arguments are used for both input to and
output from this routine.
AARRAY = singly subscripted array of real numbers with which
the simulated multiply subscripted array is
equivalenced and which contains the values to be
displayed and possibly modified. This array is
used only if KONTRL is input greater than zero. It
is assumed that the calling program will have
equivalenced AARRAY and IARRAY so that no extra
space is required to allow the manipulation of both
real and integer values.
IARRAY = singly subscripted array of integer numbers with
which the simulated multiply subscripted array is
equivalenced and which contains the values to be
displayed and possibly modified. This array is
used only if KONTRL is input less than or equal to
zero.
NAME = array which is input containing the characters
which form the name and the representation of the
current subscripts of the simulated array. This
name representation and subscript representation
are contained in NAME(1) through NAME(NAMUSD), one
character per array location as read by a multiple
of an A1 format or defined by several 1H fields.
The dimension of the NAME array, stated as the
argument NAMMAX, should be 30 to 40 greater than
NAMUSD since the area above NAME(MAXUSD) is used
for construction of the representation of the
former value and, if necessary, of the new value.
IBUFFR = array originally input containing the characters
typed by the user as read by a multiple of an A1
format, and which are to the right of the name and
subscript specification evaluated by DAPICK. If
the user typed an equal sign to the right of the
name and subscripts, then the values represented at
or to the right of IBUFFR(LOWBFR) are evaluated, or
if more values are needed than are represented in
the remainder of the IBUFFR array, then the
representations of additional values are requested
from the user and read into IBUFFR(1) through
IBUFFR(MAXBFR).
LOWBFR = input containing the location in the IBUFFR array
of the first character to be interpreted by this
FASP, FORTRAN Alphameric Subroutine Package Page 481
DAVARY, Routine to Display and Assign Values to Named Arrays
routine to determine the next user specified value.
LOWBFR is ignored if KIND is not originally input
containing the value 2. If KIND is originally
input containing the value 2, then LOWBFR is
returned pointing to the leftmost character in the
buffer which has not yet been interpreted.
KIND = when this routine is called to process the first
location within the range of array locations
specified by the user, KIND should be input
containing the value of the argument of the same
name as defined by the DAPICK routine.
= initially input containing 1, an equals sign was
not found by the DAPICK routine. DAVARY is to
report the value currently in AARRAY(LOCATN) or
IARRAY(LOCATN), but is not to modify the contents
of this array location. Since DAVARY does not
evaluate the contents of the buffer, KIND is
returned unchanged.
= initially input containing 2, an equals sign was
found by the DAPICK routine. DAVARY is to report
the value currently in AARRAY(LOCATN) or
IARRAY(LOCATN), then is to obtain the next number
represented by the characters in the IBUFFR array
and is to assign this value to the array location.
If the value cannot be obtained as part of a
previously specified range, and if the character
buffer is empty (except for possibly a comment),
then DAVARY is to ask the user for a specification
of the new value of the array location.
Each call to DAVARY will use KIND to specify to the
subsequent call to DAVARY whether the generation of
values within a range of values should be
continued, or whether the next number represented
within the buffer should be evaluated. KIND will
be returned greater than zero unless the user has
typed a semicolon which is found instead of a value
by the current call to this routine, in which case
KIND is returned containing zero to signal to the
calling program that the user has requested that
the stepping through the current range of locations
be terminated. Except for the case in which KIND
is returned zeroed, the calling program must not
modify the returned value of KIND during the
processing of the locations which are within the
range of locations indicated by the user.
MODIFY = when this routine is called to process new contents
of the singly subscripted buffer, MODIFY should
first have been set to zero by the calling program.
MODIFY is incremented by one each time a value
within any simulated array is changed by this
routine. MODIFY might then be checked at the end
FASP, FORTRAN Alphameric Subroutine Package Page 482
DAVARY, Routine to Display and Assign Values to Named Arrays
of processing of the contents of the buffer to
determine whether the calling program needs to
write out the new values in the buffer.
Use of DAVARY With Overlays
--- -- ------ ---- --------
The six values in the labeled COMMON block named FASPH are
required by the subsequent call to DAVARY for the generation
of the next value within a range of values, if any. If an
equals sign was found by the DAPICK routine, then these six
values in FASPH must be preserved during the processing of
locations within the range of array locations specified by
the user. The values in FASPH are not necessary if an
equals sign was not found, and are not needed after all
locations within the range of array locations specified by
the user have been processed.
An Example of the Use of DAVARY
-- ------- -- --- --- -- ------
The sample program listed on the following page demonstrates
both DAPICK and DAVARY since these 2 routines are designed
to cooperate in the evaluation of the text typed by the
user. DATA statements at the start of the demonstration
program define a dictionary describing the simulated arrays
named REAL, OCTAL and INTEGER, each simulated array
containing values of the type indicated by its name. To
allow construction of the alphameric representation of the
subscripts of the simulated arrays without having to take
into consideration the number of digits in the numbers, but
yet without calling either the DALONE or DANUMB routines,
each simulated subscript can only have the values 1 through
9. The simulated arrays named OCTAL and INTEGER are
contained in a buffer separate from that containing the
simulated array named REAL. Initially, each simulated array
location has the corresponding buffer subscript as its
value.
FASP, FORTRAN Alphameric Subroutine Package Page 483
DAVARY, Routine to Display and Assign Values to Named Arrays
DIMENSION NAME(60),IBUFFR(72),AARRAY(9),IARRAY(18),
1LTRSTR(16),NUMSTR(21),INISUB(7),LMTSUB(7),IDIGIT(9)
DATA IDIGIT/1H1,1H2,1H3,1H4,1H5,1H6,1H7,1H8,1H9/
DATA LTRSTR,ILEFT,IRIGHT/1HR,1HE,1HA,1HL,1HO,1HC,1HT,
11HA,1HL,1HI,1HN,1HT,1HE,1HG,1HE,1HR,1H(,1H)/
DATA MAXSUB,NAMMAX,MAXBFR,ITTY,JTTY/7,60,72,5,5/,
1NUMSTR/0,9,0,4,1,1,1,9,0,18,0,5,-1,1,1,9,7,0,1,1,9/,
2AARRAY/1.,2.,3.,4.,5.,6.,7.,8.,9./,
3IARRAY/1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18/
1 WRITE(JTTY,2)
2 FORMAT(' *'$)
READ(ITTY,3)IBUFFR
3 FORMAT(72A1)
LOWBFR=1
KIND=-1
MODIFY=0
C EVALUATE NAME AND SUBSCRIPTS
4 CALL DAPICK(MAXBFR,IBUFFR,1,16,LTRSTR,
11,21,NUMSTR,MAXSUB,LOWBFR,KIND,LRGLTR,
2LRGNUM,LRGKNT,INITAL,KOUNT,LTRINI,NUMINI,KNTSUB,
3INISUB,LMTSUB)
GO TO(13,13,1,1,5,7,9,11),KIND
5 WRITE(JTTY,6)
6 FORMAT(' INCORRECT NUMBER OF SUBSCRIPTS')
GO TO 1
7 WRITE(JTTY,8)
8 FORMAT(' SUBSCRIPTS OUT OF RANGE')
GO TO 1
9 WRITE(JTTY,10)
10 FORMAT(' UNKNOWN NAME')
GO TO 1
11 WRITE(JTTY,12)
12 FORMAT(' SEMICOLON REQUIRED')
GO TO 1
C LOOP THROUGH ARRAY LOCATIONS
13 NAMUSD=NUMSTR(NUMINI)+3
DO 14 I=4,NAMUSD
NAME(I-3)=LTRSTR(LTRINI)
14 LTRINI=LTRINI+1
KONTRL=NUMSTR(NUMINI+1)
INCRMT=1
INDEX=INISUB(1)
IF(INDEX.GT.LMTSUB(1))INCRMT=-1
NAME(NAMUSD-2)=ILEFT
NAME(NAMUSD)=IRIGHT
15 NAME(NAMUSD-1)=IDIGIT(INDEX)
LOCATN=INDEX+INITAL-1
CALL DAVARY(KONTRL,ITTY,JTTY,LOCATN,NAMUSD,
1NAMMAX,MAXBFR,0,-1,8,6,6,
25,AARRAY,IARRAY,NAME,IBUFFR,LOWBFR,KIND,MODIFY)
IF((KIND.EQ.0).OR.(INDEX.EQ.LMTSUB(1)))GO TO 4
INDEX=INDEX+INCRMT
GO TO 15
END
FASP, FORTRAN Alphameric Subroutine Package Page 484
DAVARY, Routine to Display and Assign Values to Named Arrays
Portion of Dialog Text Shown Below Which Was Typed by User
------- -- ------ ---- ----- ----- ----- --- ----- -- ----
In the dialog, the user typed all text to the right of a
leading asterisk. A few of the lines typed by DAVARY are
also requests for additional data. The text typed by the
user is shown immediately below without any of the prompts
generated by the demonstration program.
OCTAL();INTEGER(9/1)!INITIAL VALUES
REAL(/3);REAL(4/6);REAL(7/)!VARIOUS SUBSCRIPT FORMS
REAL(1)=1.5;REAL(2/4)=2.5E-9/.5E-12/10;REAL(5/)=
,6.5,,8.5;REAL(/)
OCTAL(3/)=-5K 2*-0.005M!VARIOUS FORMS OF -5*8*8*8
&DAVARY WILL REPEAT REQUEST
&WHEN IT FINDS LEADING AMPERSAND
-500000E-2-0.05E5;!TERMINATE THIS EARLY
OCTAL()
INTEGER()=20/10/40,3*-22,2*,1234;INTEGER()
Typical Dialog Between DAVARY Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
*OCTAL();INTEGER(9/1)!INITIAL VALUES
OCTAL(1) = 1
OCTAL(2) = 2
OCTAL(3) = 3
OCTAL(4) = 4
OCTAL(5) = 5
OCTAL(6) = 6
OCTAL(7) = 7
OCTAL(8) = 10
OCTAL(9) = 11
INTEGER(9) = 18
INTEGER(8) = 17
INTEGER(7) = 16
INTEGER(6) = 15
INTEGER(5) = 14
INTEGER(4) = 13
INTEGER(3) = 12
INTEGER(2) = 11
INTEGER(1) = 10
*REAL(/3);REAL(4/6);REAL(7/)!VARIOUS SUBSCRIPT FORMS
REAL(1) = 1
REAL(2) = 2
REAL(3) = 3
REAL(4) = 4
REAL(5) = 5
REAL(6) = 6
REAL(7) = 7
REAL(8) = 8
REAL(9) = 9
FASP, FORTRAN Alphameric Subroutine Package Page 485
DAVARY, Routine to Display and Assign Values to Named Arrays
*REAL(1)=1.5;REAL(2/4)=2.5E-9/.5E-12/10;REAL(5/)=
REAL(1) = 1 = 1.5
REAL(2) = 2 = 2.5E-9
REAL(3) = 3 = 2.5005E-9
REAL(4) = 4 = 2.501E-9
REAL(5) = 5 = ,6.5,,8.5;REAL(/)
REAL(6) = 6 = 6.5
REAL(8) = 8 = 8.5
REAL(1) = 1.5
REAL(2) = 2.5E-9
REAL(3) = 2.5005E-9
REAL(4) = 2.501E-9
REAL(5) = 5
REAL(6) = 6.5
REAL(7) = 7
REAL(8) = 8.5
REAL(9) = 9
*OCTAL(3/)=-5K 2*-0.005M!VARIOUS FORMS OF -5*8*8*8
OCTAL(3) = 3 = -5000
OCTAL(4) = 4 = -5000
OCTAL(5) = 5 = -5000
OCTAL(6) = 6 =
OCTAL(7) = 7 = &DAVARY WILL REPEAT REQUEST
OCTAL(7) = 7 = &WHEN IT FINDS LEADING AMPERSAND
OCTAL(7) = 7 = -500000E-2-0.05E5;!TERMINATE THIS EARLY
OCTAL(8) = 10 = -5000
*OCTAL()
OCTAL(1) = 1
OCTAL(2) = 2
OCTAL(3) = -5000
OCTAL(4) = -5000
OCTAL(5) = -5000
OCTAL(6) = 6
OCTAL(7) = -5000
OCTAL(8) = -5000
OCTAL(9) = 11
*INTEGER()=20/10/40,3*-22,2*,1234;INTEGER()
INTEGER(1) = 10 = 20
INTEGER(2) = 11 = 30
INTEGER(3) = 12 = 40
INTEGER(4) = 13 = -22
INTEGER(5) = 14 = -22
INTEGER(6) = 15 = -22
INTEGER(9) = 18 = 1234
INTEGER(1) = 20
INTEGER(2) = 30
INTEGER(3) = 40
INTEGER(4) = -22
INTEGER(5) = -22
INTEGER(6) = -22
INTEGER(7) = 16
INTEGER(8) = 17
INTEGER(9) = 1234
FASP, FORTRAN Alphameric Subroutine Package Page 486
DAVERB, Routine to Identify Words and Word Abbreviations
DDDDD AAA VV VV EEEEEEEE RRRRRR BBBBBB
DD DD AAAA VV VV EE RR RR BB BB
DD DD AA AA VV VV EE RR RR BB BB
DD DD AA AA VV VV EEEEE RRRRRR BBBBBB
DD DD AAAAAAA VVVV EE RR RR BB BB
DD DD AA AA VVV EE RR RR BB BB
DDDDD AA AA VV EEEEEEEE RR RR BBBBBB
DAVERB, Routine to Identify Words and Word Abbreviations
------ ------- -- -------- ----- --- ---- -------------
DAVERB interprets an array read by the calling program with
a multiple of an A1 format and identifies the words and word
abbreviations contained in this array. The words are
recognized by being matched against a user defined
dictionary. If the array contains abbreviations of words in
the dictionary, then DAVERB also specifies whether the
abbreviations are ambiguous.
All alphabetic letters within the dictionary supplied by the
calling program must be specified in upper case. However,
this routine treats as equivalent the upper and lower case
forms of the same alphabetic letter in the array read by the
calling program. DAVERB contains a list of the lower case
letters and a list of the corresponding upper case letters.
The lower case letters are defined by a DATA statement to be
in the order 1Ha through 1Hz which on the PDP10 computer
results in the associated integer values being sorted into
an increasing order. Each lower case letter in the input
text buffer is identified by a ternary search for a match
within the sorted list of lower case letters. If this
routine is used upon a computer system in which the
alphabetical order of the letters 1Ha through 1Hz does not
result in an increasing order for the associated integer
values, then the DATA statement which defines the array
which contains the lower case letters should be rewritten so
that the values associated with the letters will be in
increasing numerical order, and then the DATA statement
which defines the corresponding upper case letters must be
rewritten so that the lower and upper case versions of each
letter appear in locations in the respective arrays having
the same subscripts. If the computer upon which this
routine is used does not support lower case letters, then
the array which would otherwise contain the list of lower
case letters as well as the array which contains the list of
upper case letters can each contain the upper case letters
1HA through 1HZ in alphabetical order even if this order is
not the numerically sorted order.
FASP, FORTRAN Alphameric Subroutine Package Page 487
DAVERB, Routine to Identify Words and Word Abbreviations
The DAVERB Argument List
--- ------ -------- ----
The argument list of routine DAVERB is
SUBROUTINE DAVERB(LOWWRD,MAXWRD,IWORD ,LOWKNT,MAXKNT,
1 KNTLTR,IBUFFR,MAXBFR,LOWBFR,KIND ,MATCH ,LCNWRD,
2 LCNKNT,LCNBFR)
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR),IWORD(MAXWRD),KNTLTR(MAXKNT)
Original values of LOWWRD, MAXWRD, IWORD, LOWKNT, MAXKNT,
KNTLTR, IBUFFR and MAXBFR are not changed.
LOWBFR, KIND, MATCH, LCNWRD, LCNKNT and LCNBFR are returned
describing the word or word abbreviation.
LOWBFR is used for both input to and output from this
routine.
LOWWRD = subscript of the location in the IWORD array which
contains the first character of the first word in
the dictionary. Note that if KNTLTR(LOWKNT) is
less than or equal to zero, then the first
character of the first word is instead contained in
IWORD(LOWWRD-KNTLTR(LOWKNT)).
MAXWRD = maximum dimension of the IWORD array.
IWORD = dictionary array containing the characters of the
words which are to be recognized, 1 character per
array location as read by an A1 format or else
defined by 1H field. All alphabetic letters within
the IWORD array must be supplied in upper case.
The word typed by the user and read into the input
buffer array IBUFFR can be split into 2 or more
portions any of which can be abbreviated and/or
separated by spaces or tabs if the word in the
IWORD array contains a single space at the location
at which the split is allowed, and if the length
stored in the KNTLTR array is 100 more than the
number of characters including the space or spaces
which form the word in the IWORD array. For
example, if the IWORD array contains
1HN,1HO,1H ,1HL,1HI,1HM,1HI,1HT
and if the KNTLTR array contains the corresponding
length of 108, then this word could be selected by
any of the letter sequences
N, NO, N L, NL, NO L, NOL, N LI or NLI
FASP, FORTRAN Alphameric Subroutine Package Page 488
DAVERB, Routine to Identify Words and Word Abbreviations
providing in each case that the sequence is not
ambiguous.
LOWKNT = subscript of the KNTLTR array location containing
the length of the first word which can be matched
in the IWORD array. This first word will start at
IWORD(LOWWRD). If no words are to be recognized,
then either MAXKNT should be less than LOWKNT, or
else both LOWKNT and MAXKNT can point to the same
zero entry in the KNTLTR array.
MAXKNT = subscript of the KNTLTR array location containing
the length of the final word which can be matched
in the IWORD array.
KNTLTR = array containing the numbers of characters in the
words in the IWORD array. A zero or negative value
in the KNTLTR array offsets the next possible word
which can be matched in the IWORD array by the
number of letters given by the absolute value of
the negative number in the KNTLTR array. The
dimension of KNTLTR must at least be equal to
MAXKNT. For example to recognize the words
YES, NO, MAYBE
the contents of the IWORD array would be
1HY,1HE,1HS,1HN,1HO,1HM,1HA,1HY,1HB,1HE
and contents of the KNTLTR array would be
3,2,5
If the corresponding word stored in the IWORD array
contains one or more spaces, then the length stored
in the KNTLTR array must be 100 more than the
actual number of characters in the word including
the spaces. The values 101 through 199 in the
KNTLTR array thus indicate words of length 1
through 99 which could contain spaces. The value
100 in the KNTLTR array is treated the same as a
zero. Although word and word abbreviation
recognition will proceed properly even if 100 is
added to all of the lengths stored in the KNTLTR
array, more effort is required to recognize an
abbreviation of a word which is marked as
containing a space.
IBUFFR = input buffer array containing characters typed by
the user, read by a multiple of an A1 format, which
is to be searched for words and word abbreviations.
IBUFFR then contains 1 letter per computer storage
location. All lower case alphabetic letters in the
FASP, FORTRAN Alphameric Subroutine Package Page 489
DAVERB, Routine to Identify Words and Word Abbreviations
IBUFFR array are treated as being exactly
equivalent to the upper case forms of these
letters.
MAXBFR = maximum subscript of IBUFFR array to be searched
LOWBFR = subscript within the IBUFFR array of the first
(leftmost) character which can be scanned for words
or word abbreviations. LOWBFR will be returned
pointing to the next character beyond a matched
word if a word is found. If there are no printing
characters at or to right of the location indicated
by LOWBFR, then LOWBFR will be returned containing
MAXBFR+1 and KIND will be returned containing one.
LOWBFR must be set by the calling program before
anything is processed in the current contents of
the IBUFFR array, but then should not be modified
by the calling program until the entire contents of
the IBUFFR array have been processed.
KIND = returned describing the kind of item located in the
IBUFFR array.
= 1, no printing character was found at or to the
right of the location indicated by LOWBFR. The
calling program should read a new line into IBUFFR
and should reset LOWBFR to point to the first
character in the new contents of the buffer.
= 2, a word or word abbreviation was not found, but a
non-space non-tab printing character which does not
start a word in the dictionary was found. LOWBFR
is returned pointing to this printing character.
It is expected that the calling program will
otherwise process this printing character since
DAVERB would return the same results if called
again with the same value of LOWBFR, with the same
dictionary and with the same buffer contents.
= returned containing 3, 4 or 5 if a word in the
dictionary was matched even partially. For
example, if the dictionary contained both of the
words NO and NONE, then
A) initial letter N in the buffer followed by some
character other than the letter O would be an
ambiguous abbreviation and the pointer named
MATCH would be returned pointing to (containing
the sequence number within the dictionary of)
whichever word NO or NONE appeared first in the
dictionary.
B) initial letters N and O followed by some
character other than the letter N would be an
exact match with the word NO.
C) initial letters N and O and N would be a partial
but nonambiguous match with the word NONE.
FASP, FORTRAN Alphameric Subroutine Package Page 490
DAVERB, Routine to Identify Words and Word Abbreviations
Leading spaces and/or tabs are ignored. A string
of characters containing embedded spaces and/or
tabs can match a word in the dictionary only if the
word in the dictionary contains a single space at
the position at which the spaces and/or tabs are
allowed (but not necessary), and only if the
corresponding value in the KNTLTR array is 100 more
than the number of characters including the
embedded spaces which are stored in the IWORD
array.
= 3, a word in the IWORD array was matched exactly.
MATCH is returned containing the sequence number of
the word matched in the IWORD array.
= 4, a nonambiguous abbreviation of a word in the
IWORD array was found. MATCH is returned
containing the sequence number of the word in the
IWORD array.
= 5, an ambiguous abbreviation of a word was found.
MATCH is returned containing the sequence number of
the first word matched in the IWORD array.
MATCH = returned containing the sequence number of a word
matched in the IWORD array if KIND is returned
containing 3, 4 or 5. For example, if the second
word is matched, then MATCH would be returned
containing 2. The sequence number of the word in
the IWORD array does not include the letters
skipped over by the value of LOWWRD being greater
than 1, and does not include the letters skipped
over by negative values encountered in the KNTLTR
array. If a word in the IWORD array is matched,
KIND being returned with one of the values 3, 4 or
5, then MATCH is the number of values in the KNTLTR
array which are greater than zero starting at
KNTLTR(LOWKNT) up to and including the KNTLTR
location which contains the number of letters in
the word which is successfully matched.
MATCH is returned containing KIND-2 if KIND is
returned less than or equal to 2 indicating that no
word in the IWORD array could be matched even
partially. This means that if the calling program
tests for KIND=5 after the return from DAVERB, and
if KIND=4 is to be taken as equivalent to KIND=3,
then the calling program can add 2 to the value of
MATCH and use this sum as the index for a computed
GO TO statement.
LCNWRD = if KIND is returned containing 3 or greater, LCNWRD
is returned containing the subscript of the IWORD
array location which contains the first character
of the word matched in the dictionary.
LCNKNT = if KIND is returned containing 3 or greater, LCNKNT
FASP, FORTRAN Alphameric Subroutine Package Page 491
DAVERB, Routine to Identify Words and Word Abbreviations
is returned containing the subscript of the KNTLTR
array location which contains the number of
characters in the word matched in the dictionary.
If there are no zero or negative entries in the
KNTLTR array, then LCNKNT is returned equal to
LOWKNT+MATCH-1.
LCNBFR = if KIND is returned containing 3 or greater, LCNBFR
is returned containing the subscript of the IBUFFR
array location which contains the first character
of the word or its abbreviation.
If KIND is returned containing 2, then both LCNBFR
and LOWBFR are returned containing the subscript of
the IBUFFR array location which contains the
printing character which could not be identified.
An Example of the Use of DAVERB
-- ------- -- --- --- -- ------
The dictionary in the sample program listed below contains
both of the character sequences 'CANNOT' and 'CAN NOT'.
Since a space within a word in the dictionary indicates a
location at which the buffer can optionally contain a space,
the buffer contents 'CANNOT' could match either entry in the
dictionary and are ambiguous. The buffer contents
'CAN NOT', or the abbreviation 'C N', would not be ambiguous
since only one of the dictionary entries permits a space
between the adjacent appearances of the letter N.
A sample dialog between the program and the user is
presented following the listing of the program.
C PROGRAM TO DEMONSTRATE DAVERB ROUTINE
C
C FOLLOWING LINES LIST WORDS IN DICTIONARY.
C WORDS NOT PRECEDED BY NUMBERS CANNOT BE MATCHED.
C
C 'DUMMY' 'IGNORE' 12 'NONE'
C 1 'NO' 6 'YES' 13 'NOT LONG'
C 2 'NO LIMIT' 7 'CAN' 14 'NOT LARGE'
C 3 'MAYBE' 8 'CAN NOT' 15 'NO LONGER'
C 4 'MAY BE' 9 'CANNOT' 16 'HAS MANY SPACES'
C 5 'MAY' 10 'NO NEED' 17 'HASNOSPACES'
C 'EMPTY' 11 'NO LARGER'
C
DIMENSION IBUFFR(72),IWORD(131),KNTLTR(22)
DATA IWORD/1HD,1HU,1HM,1HM,1HY,1HN,1HO,1HN,1HO,1H ,
1 1HL,1HI,1HM,1HI,1HT,1HM,1HA,1HY,1HB,1HE,
2 1HM,1HA,1HY,1H ,1HB,1HE,1HM,1HA,1HY,1HE,
3 1HM,1HP,1HT,1HY,1HI,1HG,1HN,1HO,1HR,1HE,
FASP, FORTRAN Alphameric Subroutine Package Page 492
DAVERB, Routine to Identify Words and Word Abbreviations
4 1HY,1HE,1HS,1HC,1HA,1HN,1HC,1HA,1HN,1H ,
5 1HN,1HO,1HT,1HC,1HA,1HN,1HN,1HO,1HT,1HN,
6 1HO,1H ,1HN,1HE,1HE,1HD,1HN,1HO,1H ,1HL,
7 1HA,1HR,1HG,1HE,1HR,1HN,1HO,1HN,1HE,1HN,
8 1HO,1HT,1H ,1HL,1HO,1HN,1HG,1HN,1HO,1HT,
9 1H ,1HL,1HA,1HR,1HG,1HE,1HN,1HO,1H ,1HL,
1 1HO,1HN,1HG,1HE,1HR,1HH,1HA,1HS,1H ,1HM,
2 1HA,1HN,1HY,1H ,1HS,1HP,1HA,1HC,1HE,1HS,
3 1HH,1HA,1HS,1HN,1HO,1HS,1HP,1HA,1HC,1HE,
4 1HS/
DATA KNTLTR/ 99, 99, -3, 2,108, 5,106, 3, -5, -6,
1 3, 3,107, 6,107,109, 4,108,109,109,
2 115, 11/
DATA LOWWRD,MAXWRD,LOWKNT,MAXKNT,MAXBFR/3,131,3,22,72/
DATA IGREAT,IBLANK/1H>,1H /
C
C ASK FOR AND ACCEPT USER RESPONSE
1 TYPE 2
2 FORMAT(1X,1H*,$)
ACCEPT 3,IBUFFR
3 FORMAT(100A1)
LOWBFR=1
C
C IDENTIFY NEXT WORD IN USER RESPONSE
4 CALL DAVERB(LOWWRD,MAXWRD,IWORD ,LOWKNT,MAXKNT,
1 KNTLTR,IBUFFR,MAXBFR,LOWBFR,KIND ,MATCH ,LCNWRD,
2 LCNKNT,LCNBFR)
IF(KIND.EQ.1)GO TO 1
IF(KIND.EQ.2)GO TO 8
C
C REPORT USER RESPONSE AND DICTIONARY WORD MATCHED
IUSED=LOWBFR-1
IEND=KNTLTR(LCNKNT)
IF(IEND.GT.100)IEND=IEND-100
IEND=LCNWRD+IEND-1
IF(KIND.EQ.3)TYPE 5,(IBUFFR(I),I=LCNBFR,IUSED),
1IGREAT,IBLANK,(IWORD(I),I=LCNWRD,IEND)
5 FORMAT(' EXACT <',100A1)
IF(KIND.EQ.4)TYPE 6,(IBUFFR(I),I=LCNBFR,IUSED),
1IGREAT,IBLANK,(IWORD(I),I=LCNWRD,IEND)
6 FORMAT(' UNIQUE <',100A1)
IF(KIND.EQ.5)TYPE 7,(IBUFFR(I),I=LCNBFR,IUSED),
1IGREAT,IBLANK,(IWORD(I),I=LCNWRD,IEND)
7 FORMAT(' AMBIGUOUS <',100A1)
GO TO 4
C
C REPORT IF NO MATCH
8 TYPE 9,IBUFFR(LOWBFR)
9 FORMAT(' UNKNOWN <',1A1,1H>)
LOWBFR=LOWBFR+1
GO TO 4
END
FASP, FORTRAN Alphameric Subroutine Package Page 493
DAVERB, Routine to Identify Words and Word Abbreviations
Typical Dialog Between DAVERB Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
*N Y NO YE NON YES NONE
AMBIGUOUS <N> NO
UNIQUE <Y> YES
EXACT <NO> NO
UNIQUE <YE> YES
AMBIGUOUS <NON> NO NEED
EXACT <YES> YES
EXACT <NONE> NONE
*N N NN N NE NNE NO NE NONEE
UNIQUE <N N> NO NEED
UNIQUE <NN> NO NEED
UNIQUE <N NE> NO NEED
UNIQUE <NNE> NO NEED
UNIQUE <NO NE> NO NEED
UNIQUE <NONEE> NO NEED
*M MA MAY MAYB MAYBE
AMBIGUOUS <M> MAYBE
AMBIGUOUS <MA> MAYBE
EXACT <MAY> MAY
AMBIGUOUS <MAYB> MAYBE
AMBIGUOUS <MAYBE> MAYBE
*M B MB MA B MAB MAY B MA BE MABE MAY BE
UNIQUE <M B> MAY BE
UNIQUE <MB> MAY BE
UNIQUE <MA B> MAY BE
UNIQUE <MAB> MAY BE
UNIQUE <MAY B> MAY BE
UNIQUE <MA BE> MAY BE
UNIQUE <MABE> MAY BE
EXACT <MAY BE> MAY BE
*C CA CAN CANN CANNO CANNOT
AMBIGUOUS <C> CAN
AMBIGUOUS <CA> CAN
EXACT <CAN> CAN
AMBIGUOUS <CANN> CAN NOT
AMBIGUOUS <CANNO> CAN NOT
AMBIGUOUS <CANNOT> CAN NOT
*C N CN CA N CAN N C NO CNO CA NO CAN NO CAN NOT
UNIQUE <C N> CAN NOT
UNIQUE <CN> CAN NOT
UNIQUE <CA N> CAN NOT
UNIQUE <CAN N> CAN NOT
UNIQUE <C NO> CAN NOT
UNIQUE <CNO> CAN NOT
UNIQUE <CA NO> CAN NOT
UNIQUE <CAN NO> CAN NOT
EXACT <CAN NOT> CAN NOT
FASP, FORTRAN Alphameric Subroutine Package Page 494
DAVERB, Routine to Identify Words and Word Abbreviations
*NO LA NOLA NOTLA NOT LA NO LARGE NOT LARGE NO LARGER
AMBIGUOUS <NO LA> NO LARGER
AMBIGUOUS <NOLA> NO LARGER
UNIQUE <NOTLA> NOT LARGE
UNIQUE <NOT LA> NOT LARGE
AMBIGUOUS <NO LARGE> NO LARGER
EXACT <NOT LARGE> NOT LARGE
EXACT <NO LARGER> NO LARGER
*NO LO NOLO NOTLO NOT LO NO LONG NOT LONG NO LONGE NO LONGER
AMBIGUOUS <NO LO> NOT LONG
AMBIGUOUS <NOLO> NOT LONG
UNIQUE <NOTLO> NOT LONG
UNIQUE <NOT LO> NOT LONG
AMBIGUOUS <NO LONG> NOT LONG
EXACT <NOT LONG> NOT LONG
UNIQUE <NO LONGE> NO LONGER
EXACT <NO LONGER> NO LONGER
*NOTLONGNOT LONG NOT LONG NOT LONG
EXACT <NOTLONG> NOT LONG
EXACT <NOT LONG> NOT LONG
EXACT <NOT LONG> NOT LONG
EXACT <NOT LONG> NOT LONG
*HASMANYSPACES HAS MANY SPACES HAS MANY SPACES
EXACT <HASMANYSPACES> HAS MANY SPACES
EXACT <HAS MANY SPACES> HAS MANY SPACES
EXACT <HAS MANY SPACES> HAS MANY SPACES
*HMS H M S H M S H M S
UNIQUE <HMS> HAS MANY SPACES
UNIQUE <H M S> HAS MANY SPACES
UNIQUE <H M S> HAS MANY SPACES
UNIQUE <H M S> HAS MANY SPACES
*hamasp ha ma sp ha ma sp
UNIQUE <hamasp> HAS MANY SPACES
UNIQUE <ha ma sp> HAS MANY SPACES
UNIQUE <ha ma sp> HAS MANY SPACES
*HaMaSp hA mA sP HA ma SP
UNIQUE <HaMaSp> HAS MANY SPACES
UNIQUE <hA mA sP> HAS MANY SPACES
UNIQUE <HA ma SP> HAS MANY SPACES
*HAMASP HA MA SP HA MA SP
UNIQUE <HAMASP> HAS MANY SPACES
UNIQUE <HA MA SP> HAS MANY SPACES
UNIQUE <HA MA SP> HAS MANY SPACES
*HHAHASHASNHASMHASMSPACES
AMBIGUOUS <H> HAS MANY SPACES
AMBIGUOUS <HA> HAS MANY SPACES
AMBIGUOUS <HAS> HAS MANY SPACES
UNIQUE <HASN> HASNOSPACES
UNIQUE <HASM> HAS MANY SPACES
UNIQUE <HASMSPACES> HAS MANY SPACES
FASP, FORTRAN Alphameric Subroutine Package Page 495
DAVERB, Routine to Identify Words and Word Abbreviations
DALOSS, Extends DAVERB to Allow Comments and Missing Words
------ ------- ------ -- ----- -------- --- ------- -----
If the first printing character found by DAVERB in the line
of input text does not appear at the start of any word in
the dictionary supplied by the calling program, then the
character is considered to be unknown causing DAVERB to
return control to the calling program. In particular,
DAVERB does not recognize the punctuation characters allowed
by many of the other routines in the FORTRAN Alphameric
Subroutine Package. DALOSS, a short subroutine which calls
upon DAVERB for word recognition, allows commas between
words, identifies missing items indicated by extra commas,
skips over any text which is to the right of either an
exclamation point or an ampersand, and reports any
semicolons found in the text being evaluated. In addition,
DALOSS reports whether the character to the right of a word
or its abbreviation is a character other than a space, a tab
character or an allowed punctuation character.
The argument list of routine DALOSS is
SUBROUTINE DALOSS(LOWWRD,MAXWRD,IWORD ,LOWKNT,MAXKNT,
1 KNTLTR,IBUFFR,MAXBFR,LOWBFR,KIND ,MATCH ,LCNWRD,
1 LCNKNT,LCNBFR,MANY ,LCNERR)
with the associated DIMENSION statement
DIMENSION IBUFFR(MAXBFR),IWORD(MAXWRD),
1KNTLTR(MAXKNT)
The argument lists of DALOSS and DAVERB are identical except
for the DALOSS arguments MANY and LCNERR which do not appear
in the DAVERB argument list, and except that DALOSS can
return the argument named KIND containing the additional
values 6 through 11. The argument named MANY must be set to
zero by the calling program before calling either this
routine or any of the other routines in the FASP package
(such as DAMISS, DANEXT and DASPAN) which define this
argument in a similar manner. The arguments named KIND and
LCNERR are used only for output to the calling program and
their input values are ignored. These arguments are
described below. The documentation of DAVERB should be
consulted for descriptions of the remaining arguments.
KIND = 1, nothing, except perhaps a comment indicated by a
leading exclamation point, was found at or to the
right of IBUFFR(LOWBFR). The calling program
should read a new line into the IBUFFR array before
again calling this routine if additional words are
required. LOWBFR is returned pointing beyond the
end of the buffer. MANY is returned set to zero.
MATCH is returned undefined.
= 2, the first printing character (other than a
FASP, FORTRAN Alphameric Subroutine Package Page 496
DAVERB, Routine to Identify Words and Word Abbreviations
possible comma if MANY was input greater than zero)
in or to right of IBUFFR(LOWBFR) did not match the
first character of any word in the dictionary and
was not a comma, semicolon, ampersand or
exclamation point. Both LCNBFR and LOWBFR are
returned pointing to this printing character. It
is expected that the calling program will otherwise
process this printing character since DALOSS would
return the same results if called again with the
same value of LOWBFR, with the same dictionary and
with the same buffer contents. LCNERR is returned
pointing to the next space, tab character, comma,
semicolon, exclamation point or ampersand to the
right of the unknown character, or is returned set
to MAXBFR+1 if none of these characters is found.
MANY is returned containing one plus its input
absolute value. Match is returned undefined.
= 3 or 4 or 5, same as when DAVERB returns these
values, except that if there are additional
characters to the right of the word or its
abbreviation, then the character to the immediate
right of the word or its abbreviation is either a
space, tab character, comma, semicolon, exclamation
point or ampersand. MANY is returned containing
one plus its input absolute value. LCNBFR is
returned pointing in the buffer to the first
character of the word or its abbreviation. LCNERR
and LOWBFR are returned pointing to the character
to the right of the word or its abbreviation.
= 3, a word in the IWORD array was matched exactly.
MATCH is returned containing the sequence number of
the word matched in the IWORD array.
= 4, a nonambiguous abbreviation of a word in the
IWORD array was found. MATCH is returned
containing the sequence number of the word in the
IWORD array.
= 5, an ambiguous abbreviation of a word was found.
MATCH is returned containing the sequence number of
the first word matched in the IWORD array.
= 6 or 7 or 8, same as when KIND is returned
containing 3 or 4 or 5 respectively, except that a
character other than a space, tab character, comma,
semicolon, exclamation point or ampersand appeared
to the immediate right of the word or its
abbreviation. LCNBFR is returned pointing in the
buffer to the first character of the word or its
abbreviation. LOWBFR is returned pointing in the
buffer to the character to the right of the word or
its abbreviation. LCNERR is returned pointing in
the buffer to the next space, tab character, comma,
semicolon, exclamation point or ampersand to the
right of the word or its abbreviation, or is
returned pointing beyond the end of the buffer if
no space, tab character, comma, semicolon,
FASP, FORTRAN Alphameric Subroutine Package Page 497
DAVERB, Routine to Identify Words and Word Abbreviations
exclamation point or ampersand is found to the
right of the word or its abbreviation. MANY is
returned set to one plus its input absolute value.
= 9, a semicolon was found as the first printing
character at or to the right of IBUFFR(LOWBFR).
LOWBFR is returned pointing to the next character
beyond the semicolon. It is assumed that the
calling program will treat the appearance of the
semicolon as marking the end of a statement. MANY
is returned set to zero. MATCH is returned
undefined.
= 10, an ampersand was found as the first printing
character at or to the right of LOWBFR. The text
to the right of the ampersand is taken as a comment
so LOWBFR is returned pointing beyond the right end
of the buffer. It is assumed that the calling
program will read in the contents of a new buffer,
then again request a new word identification from
this routine. The value of MANY must not be
changed by the calling program prior to this
following call. The effect is not quite the same
as if the user had typed all of the text on a
single line since a single word cannot be split
across a line boundary. MATCH is returned
undefined.
= 11, a word was not found, but an extra comma was
found indicating a missing word. MANY is returned
containing one plus its input absolute value.
MATCH is returned undefined.
MANY = should be input containing zero each time this
routine is called to begin processing of a new
logical section of text, as for example when
beginning processing of a line of text not tied to
the previous line by an ampersand at the end of the
previous line, or when processing the text to the
right of a semicolon. The initial zeroing of this
argument must be done by the calling program, but
thereafter the value returned by the previous call
to this routine can usually be used. MANY is
returned set to zero each time a semicolon (KIND=9)
is found, and each time an end of line not tied to
the following line by an ampersand (KIND=1) is
found. MANY is returned containing one plus its
input absolute value each time a word is found,
each time an unknown character is found, or each
time an indication of a missing word is found.
KIND is returned containing the value 10 and MANY
is returned containing the negative of the number
of items found if the next printing character
following a comma is an ampersand. MANY should not
be changed by the calling program if an ampersand
(KIND being returned=10) is found indicating that
the subsequent call to this routine is to process
FASP, FORTRAN Alphameric Subroutine Package Page 498
DAVERB, Routine to Identify Words and Word Abbreviations
text which is to be treated as though it appeared
in place of the ampersand and the characters to its
right. The effect is not quite the same as if the
user had typed all of the text on a single line
since a single word cannot be split across the line
boundary.
If MANY is input containing zero, then an initial
comma in the input text buffer is taken to indicate
an initial missing item, and MANY is then returned
containing 1. If MANY is input greater than zero,
then an initial comma is ignored if followed by a
word. If MANY is input greater than zero, then an
initial comma followed by no other printing
characters, by a semicolon, or by an exclamation
point indicates a missing item. If MANY is input
greater than zero, then an initial comma followed
by an ampersand will cause the remaining characters
in the buffer to be ignored, and MANY will be
returned containing the negative of its input
value. If MANY is input negative, then it is
assumed that the contents of the current buffer
continue a previous line which terminated with a
comma followed by an ampersand, and MANY is
returned greater than zero.
LCNERR = if KIND is returned containing a number in the
range 3 through 8 indicating that a word or its
abbreviation was found, then LCNERR is returned
containing the subscript of the IBUFFR array
location which contains the next space, tab
character, comma, semicolon, exclamation point or
ampersand to the right of the word or its
abbreviation, or else LCNERR is returned containing
MAXBFR+1 if none of these characters appear
anywhere to the right of the word or its
abbreviation. If KIND is returned containing 3, 4
or 5, then LCNERR and LOWBFR are both returned
pointing to the character to the immediate right of
the word or its abbreviation. If KIND is returned
containing 6, 7 or 8, then LOWBFR is returned
pointing to the unexpected printing character to
the immediate right of the word or its
abbreviation.
If KIND is returned containing 2, indicating that a
printing character was found which did not begin
any word in the dictionary and which was not one of
the allowed punctuation marks, then LCNERR is
returned pointing to the next space, tab character,
comma, semicolon, exclamation point or ampersand,
or else LCNERR is returned containing MAXBFR+1 if
none of these characters appear anywhere to the
right of the initial unexpected character.
FASP, FORTRAN Alphameric Subroutine Package Page 499
DAVERB, Routine to Identify Words and Word Abbreviations
An Example of the Use of DALOSS
-- ------- -- --- --- -- ------
The program listed on the following pages is based upon the
program listed earlier as a demonstration the DAVERB
routine. The major modification is that in the program
listed here, unexpected characters which follow a word or
its abbreviation are typed to the right of a question mark
which is added to the right of the characters which are
recognized, and LOWBFR is then advanced to have the value of
LCNERR before DALOSS is again called.
C PROGRAM TO DEMONSTRATE DALOSS ROUTINE
C
C FOLLOWING LINES LIST WORDS IN DICTIONARY.
C WORDS NOT PRECEDED BY NUMBERS CANNOT BE MATCHED.
C
C 'DUMMY' 'IGNORE' 12 'NONE'
C 1 'NO' 6 'YES' 13 'NOT LONG'
C 2 'NO LIMIT' 7 'CAN' 14 'NOT LARGE'
C 3 'MAYBE' 8 'CAN NOT' 15 'NO LONGER'
C 4 'MAY BE' 9 'CANNOT' 16 'HAS MANY SPACES'
C 5 'MAY' 10 'NO NEED' 17 'HASNOSPACES'
C 'EMPTY' 11 'NO LARGER'
C
DIMENSION IBUFFR(72),IWORD(131),KNTLTR(22)
DATA IWORD/1HD,1HU,1HM,1HM,1HY,1HN,1HO,1HN,1HO,1H ,
1 1HL,1HI,1HM,1HI,1HT,1HM,1HA,1HY,1HB,1HE,
2 1HM,1HA,1HY,1H ,1HB,1HE,1HM,1HA,1HY,1HE,
3 1HM,1HP,1HT,1HY,1HI,1HG,1HN,1HO,1HR,1HE,
4 1HY,1HE,1HS,1HC,1HA,1HN,1HC,1HA,1HN,1H ,
5 1HN,1HO,1HT,1HC,1HA,1HN,1HN,1HO,1HT,1HN,
6 1HO,1H ,1HN,1HE,1HE,1HD,1HN,1HO,1H ,1HL,
7 1HA,1HR,1HG,1HE,1HR,1HN,1HO,1HN,1HE,1HN,
8 1HO,1HT,1H ,1HL,1HO,1HN,1HG,1HN,1HO,1HT,
9 1H ,1HL,1HA,1HR,1HG,1HE,1HN,1HO,1H ,1HL,
1 1HO,1HN,1HG,1HE,1HR,1HH,1HA,1HS,1H ,1HM,
2 1HA,1HN,1HY,1H ,1HS,1HP,1HA,1HC,1HE,1HS,
3 1HH,1HA,1HS,1HN,1HO,1HS,1HP,1HA,1HC,1HE,
4 1HS/
DATA KNTLTR/ 99, 99, -3, 2,108, 5,106, 3, -5, -6,
1 3, 3,107, 6,107,109, 4,108,109,109,
2 115, 11/
DATA LOWWRD,MAXWRD,LOWKNT,MAXKNT,MAXBFR/3,131,3,22,72/
DATA IGREAT,IWHAT,IBLANK/1H>,1H?,1H /
C
C ASK FOR AND ACCEPT USER RESPONSE
MANY=0
1 TYPE 2
2 FORMAT(1X,1H*,$)
ACCEPT 3,IBUFFR
3 FORMAT(100A1)
LOWBFR=1
C
FASP, FORTRAN Alphameric Subroutine Package Page 500
DAVERB, Routine to Identify Words and Word Abbreviations
C IDENTIFY NEXT WORD IN USER RESPONSE
4 CALL DALOSS(LOWWRD,MAXWRD,IWORD ,LOWKNT,MAXKNT,
1 KNTLTR,IBUFFR,MAXBFR,LOWBFR,KIND ,MATCH ,LCNWRD,
2 LCNKNT,LCNBFR,MANY,LCNERR)
IF(KIND.EQ.1)GO TO 1
IF(KIND.EQ.2)GO TO 12
IF(KIND.LT.9)GO TO 8
IF(KIND.EQ.9)TYPE 5
IF(KIND.EQ.10)TYPE 6
IF(KIND.EQ.11)TYPE 7
5 FORMAT(' SEMICOLON')
6 FORMAT(' AMPERSAND')
7 FORMAT(' MISSING')
IF(KIND.EQ.10)GO TO 1
GO TO 4
C
C REPORT USER RESPONSE AND DICTIONARY WORD MATCHED
8 IUSED=LOWBFR-1
JUSED=LCNERR-1
IEND=KNTLTR(LCNKNT)
IF(IEND.GT.100)IEND=IEND-100
IEND=LCNWRD+IEND-1
IF(KIND.EQ.3)TYPE 9,(IBUFFR(I),I=LCNBFR,IUSED),
1IGREAT,IBLANK,(IWORD(I),I=LCNWRD,IEND)
9 FORMAT(' EXACT <',100A1)
IF(KIND.EQ.4)TYPE 10,(IBUFFR(I),I=LCNBFR,IUSED),
1IGREAT,IBLANK,(IWORD(I),I=LCNWRD,IEND)
10 FORMAT(' UNIQUE <',100A1)
IF(KIND.EQ.5)TYPE 11,(IBUFFR(I),I=LCNBFR,IUSED),
1IGREAT,IBLANK,(IWORD(I),I=LCNWRD,IEND)
IF(KIND.EQ.6)TYPE 9,(IBUFFR(I),I=LCNBFR,IUSED),
1IWHAT,(IBUFFR(I),I=LOWBFR,JUSED),
2IGREAT,IBLANK,(IWORD(I),I=LCNWRD,IEND)
IF(KIND.EQ.7)TYPE 10,(IBUFFR(I),I=LCNBFR,IUSED),
1IWHAT,(IBUFFR(I),I=LOWBFR,JUSED),
2IGREAT,IBLANK,(IWORD(I),I=LCNWRD,IEND)
IF(KIND.EQ.8)TYPE 11,(IBUFFR(I),I=LCNBFR,IUSED),
1IWHAT,(IBUFFR(I),I=LOWBFR,JUSED),
2IGREAT,IBLANK,(IWORD(I),I=LCNWRD,IEND)
11 FORMAT(' AMBIGUOUS <',100A1)
IF(KIND.GT.5)LOWBFR=LCNERR
GO TO 4
C
C REPORT IF NO MATCH
12 JUSED=LCNERR-1
TYPE 13,(IBUFFR(I),I=LCNBFR,JUSED),IGREAT
13 FORMAT(' UNKNOWN <',100A1)
LOWBFR=LCNERR
GO TO 4
END
FASP, FORTRAN Alphameric Subroutine Package Page 501
DAVERB, Routine to Identify Words and Word Abbreviations
Typical Dialog Between DALOSS Demonstration Program and User
------- ------ ------- ------ ------------- ------- --- ----
*NO YES,NO,,YELLOW , NO , ; , YES!SPACES AND COMMAS
EXACT <NO> NO
EXACT <YES> YES
EXACT <NO> NO
MISSING
UNIQUE <YE?LLOW> YES
EXACT <NO> NO
MISSING
SEMICOLON
MISSING
EXACT <YES> YES
*,NO,
MISSING
EXACT <NO> NO
MISSING
*NOMATCH;WITH NOTHING!ERRORS
EXACT <NO?MATCH> NO
SEMICOLON
UNKNOWN <WITH>
AMBIGUOUS <NOT?HING> NOT LONG
*NO&DEMONSTRATIONS OF COMMAS WITH AMPERSANDS
EXACT <NO> NO
AMPERSAND
*YES
EXACT <YES> YES
*NO&
EXACT <NO> NO
AMPERSAND
*,YES
EXACT <YES> YES
*NO,&
EXACT <NO> NO
AMPERSAND
*YES
EXACT <YES> YES
*NO,&
EXACT <NO> NO
AMPERSAND
*,YES
MISSING
EXACT <YES> YES
*NO,&
EXACT <NO> NO
AMPERSAND
*&
AMPERSAND
*,YES
MISSING
EXACT <YES> YES
FASP, FORTRAN Alphameric Subroutine Package Page 502
DAVERB, Routine to Identify Words and Word Abbreviations
Simple but Machine Dependent Method of Dictionary Definition
------ --- ------- --------- ------ -- ---------- ----------
The numeric codes by which characters are stored in a
computer are much smaller than the maximum size number which
the computer can represent, so several codes are often
packed within a single computer storage location.
Unfortunately, this leads to difficulties in the recognition
of these characters unless the character to column alignment
remains fixed. The character strings 'THE ', ' THE ' and
' THE' are stored as different numbers, and can only be
matched as containing the same word if the individual
characters are extracted and compared.
On the PDP-10 which is a 36 bit computer with 7 bit
character codes, 5 characters can be packed in a single
computer storage location, leading to the use of A5 format
specifications, and 2 storage locations would be needed to
contain the character string represented either as
10HABCDEFGHIJ or 'ABCDEFGHIJ'. If a long string is used in
a DATA statement, then the programmer must still calculate
the number of computer storage locations corresponding to
the string and explicitly allocate exactly the correct array
length to contain the string. However, the compiler
performs the allocation of sufficient space when a long
string is used directly within a CALL statement. The called
routine could then convert the input text string for which
storage was allocated by the compiler into a second array
which is returned to the calling program containing the
characters of the original string, but separated one
character to an array location as though read by a multiple
of an A1 format.
The conversion of a string of 20 characters from A5 to A1
format could be done by writing the characters to a scratch
file with a 4A5 format, then reading them back with a 20A1
format. The writing and reading of a scratch file is
simulated on the PDP-10 in the sample routine listed below
by the DECODE statement. This routine restructures a
dictionary, specified as a string of words separated by
spaces, into the array of single characters (IWORD) and the
array of characters counts per word (KNTLTR) required by the
DAVERB routine. Since the programmer does not want to be
bothered with counting the number of characters within the
string, and certainly does not want to update such a count
when the string is later modified, the end of the string is
instead delimited by a second appearance of the initial
printing character, both appearances being otherwise
ignored. This delimiting character is actually a part of
the string sent to the routine and is in addition to the
number and letter H or to the apostrophes which delimit the
string to the compiler. The argument LMTPKD does place an
upper limit upon the length of the string, but is used only
if the programmer forgets to delimit the ends of the string.
FASP, FORTRAN Alphameric Subroutine Package Page 503
DAVERB, Routine to Identify Words and Word Abbreviations
The arguments of the routine are defined as follow. The
arguments IPACKD, LMTPKD, LMTWRD and LMTKNT are used only
for input and are returned unchanged.
IPACKD = array containing several characters in each array
location. The number of characters in each array
location is indicated by the variable named INEACH,
which is defined to have the value 5 by a DATA
statement near the start of the program. If
INEACH=5, then the array is treated as though it
was read with a multiple of an A5 format. If
INEACH is equal to the default for the computer on
which this routine is used, then the IPACKD array
can be defined in the call to this routine as a
single long Hollerith string, or as a text string
quoted with apostrophes.
LMTPKD = maximum number of characters which can be contained
in the IPACKD array. It is expected that this will
be set somewhat greater than the number of
characters actually contained in the array.
LMTWRD = the dimension of the IWORD array in which the
characters of words in the dictionary are returned.
LMTKNT = the dimension of the KNTLTR array in which the
numbers of characters in each word in the
dictionary are returned.
MAXWRD = input containing the number of locations in the
IWORD array which are currently in use and which
must be preserved. The characters added to the
dictionary are stored in IWORD(MAXWRD+1) through
IWORD(LMTWRD).
= returned containing the number of locations in the
IWORD array in use after this routine has added the
words in the IPACKD array to the dictionary. If
more words must be added to the dictionary than can
be stored in a single Hollerith string or string
quoted with apostrophes, then additional calls to
this routine can be used to extend the dictionary.
MAXKNT = input containing the number of locations in the
KNTLTR array which are currently in used and which
must be preserved. The number of characters in the
words added to the dictionary are stored in
KNTLTR(MAXKNT+1) through KNTLTR(LMTKNT).
= returned containing the number of locations in the
KNTLTR array in use after this call has added the
words in the IPACKD array to the dictionary.
IWORD = array into which the characters of the words in the
IPACKD array are stored as though these characters
had been read by a multiple of an A1 format.
KNTLTR = array into which the numbers of characters in each
of the words in the IPACKD array are stored.
This routine is not included in the FASP package since it is
machine dependent both in the assumed character packing (5
per location) and in the syntax of the DECODE statement.
FASP, FORTRAN Alphameric Subroutine Package Page 504
DAVERB, Routine to Identify Words and Word Abbreviations
SUBROUTINE A5TOA1(IPACKD,LMTPKD,LMTWRD,LMTKNT,MAXWRD,
1 MAXKNT,IWORD ,KNTLTR)
C ROUTINE TO RESTRUCTURE DICTIONARY, SPECIFIED AS WORDS
C SEPARATED BY SPACES CONTAINED IN SINGLE LONG
C HOLLERITH STRING OR IN SINGLE LONG TEXT STRING QUOTED
C WITH APOSTROPHES, INTO IWORD ARRAY OF SINGLE
C CHARACTERS AND INTO KNTLTR ARRAY OF CHARACTER COUNTS
C PER WORD NEEDED BY DAVERB. SECOND APPEARANCE OF
C FIRST PRINTING CHARACTER MARKS END OF STRING.
C
C INEACH IS NUMBER OF CHARACTERS IN EACH HOLLERITH WORD
DATA INEACH/5/,IBLANK/1H /
DIMENSION IPACKD(LMTPKD),KNTLTR(LMTKNT),IWORD(LMTWRD)
KNTPKD=0
LOCAL=-1
LIMIT=LMTPKD
C
C GET NEXT GROUP OF CHARACTERS
1 NEEDED=LMTWRD-MAXWRD
IF(NEEDED.GT.INEACH)NEEDED=INEACH
IF(NEEDED.GT.LIMIT)NEEDED=LIMIT
IF(NEEDED.LE.0)GO TO 8
KNTPKD=KNTPKD+1
ISTART=MAXWRD+1
IEND=MAXWRD+NEEDED
LIMIT=LIMIT-NEEDED
C READ ALPHAMERIC INFORMATION FROM IPACKD(KNTPKD) INTO
C IWORD(ISTART) THROUGH IWORD(IEND) VIA THE INDICATED
C FORMAT. NEEDED IS THE NUMBER OF CHARACTERS TO READ.
DECODE(NEEDED,2,IPACKD(KNTPKD))
1(IWORD(I),I=ISTART,IEND)
2 FORMAT(100A1)
C
C FIND WORDS OR DELIMITER CHARACTERS
3 IF(ISTART.GT.IEND)GO TO 1
IF(LOCAL.LT.0)GO TO 5
IF(IWORD(ISTART).EQ.IFINAL)GO TO 8
IF(IWORD(ISTART).EQ.IBLANK)GO TO 6
IF(LOCAL.GT.0)GO TO 4
IF(MAXKNT.GE.LMTKNT)GO TO 8
MAXKNT=MAXKNT+1
KNTLTR(MAXKNT)=0
LOCAL=1
4 KNTLTR(MAXKNT)=KNTLTR(MAXKNT)+1
MAXWRD=MAXWRD+1
IWORD(MAXWRD)=IWORD(ISTART)
GO TO 7
5 IF(IWORD(ISTART).EQ.IBLANK)GO TO 7
IFINAL=IWORD(ISTART)
6 LOCAL=0
7 ISTART=ISTART+1
GO TO 3
8 RETURN
END
FASP, FORTRAN Alphameric Subroutine Package Page 505
DAVERB, Routine to Identify Words and Word Abbreviations
A Typical Wrapper Routine for the DALOSS and A5TOA1 Routines
- ------- ------- ------- --- --- ------ --- ------ --------
If the calling program requires only a few calls to any one
of the routines in FASP, then these routines can be called
directly. However, if the calling program needs to evaluate
the answers typed by the user in a lengthy dialog, it is
much more effective to identify the characteristics common
to the evaluation of several answers and then to write an
intermediate level routine having the minimal argument list
which can in turn call the routines in FASP and perhaps do
some answer verification and issue prompts when an error is
detected. The GETWRD routine listed below is such a wrapper
for the DALOSS routine, but is machine dependent both in the
use of unit 5 for error messages and in the calling of the
A5TOA1 routine. The GETWRD routine evaluates the words
and/or abbreviations typed on a single line. Any characters
to the right of an exclamation point are ignored. An
ampersand, if encountered, is treated exactly the same as an
exclamation point. Likewise, a semicolon, if found, is
treated the same as a comma. The routine issues an error
message and then lists for the user the currently allowed
answers if an error is detected in the user's response.
The following are input arguments returned unchanged.
IPACKD = same as the array of the same name described in the
A5TOA1 documentation. This is an array of
characters specified in the call to this routine
either as a Hollerith string or as a text string
quoted with apostrophes. The first printing
character in the string is ignored, and the string
is assumed to continue through the next appearance
of this delimiting character. Between the two
appearances of this delimiting character should be
specified, separated by spaces, all of the words
which are to be allowed. At most 20 words
containing a total of no more than 100 characters
can be defined in this manner. For example, the
statement
CALL GETWRD('/YES NO MAYBE/',IBUFFR,MAXBFR,
1LOWBFR,KIND)
or
CALL GETWRD(14H/YES NO MAYBE/,IBUFFR,MAXBFR,
1LOWBFR,KIND)
would define a dictionary containing the 3 words
YES, NO and MAYBE.
IBUFFR = array containing the text to be evaluated as read
by a multiple of an A1 format. IBUFFR then
contains 1 character in each array location.
MAXBFR = subscript of the highest (rightmost) location in
the IBUFFR array which contains a character to be
FASP, FORTRAN Alphameric Subroutine Package Page 506
DAVERB, Routine to Identify Words and Word Abbreviations
evaluated.
The following argument is used both for input to and output
from this routine.
LOWBFR = should be input set to zero whenever this routine
is first called to evaluate the contents of a
particular line of text. Thereafter, LOWBFR is
returned containing the subscript of the IBUFFR
array location which contains the leftmost
character which has not yet been evaluated.
Setting the initial value of LOWBFR to zero instead
of to one signals to this routine that if the first
printing character within the buffer is found to be
either a comma or a semicolon then an initial
missing word is to be assumed. The returned value
of LOWBFR should be transmitted to the subsequent
call to this routine unchanged until the line of
text has been completely evaluated.
The following argument is used only for output.
KIND = 1, returned if the input text buffer is empty or
contains only a comment indicated by a leading
exclamation point or by a leading ampersand.
= 2, returned if this routine finds printing
characters which are not allowed punctuation
characters and which either do not uniquely select
a word from the dictionary, or else which select
such a word but are followed by an unexpected
character. This routine informs the user both of
the text containing the error and of the list of
currently allowed words. LOWBFR is returned
pointing to the character to the immediate right of
the character sequence containing the error.
= 3, an extra comma or semicolon indicates a missing
item.
= 4 or greater, KIND is 3 plus the sequence number of
the matched word in the dictionary. If the
dictionary contains only the words YES, NO and
MAYBE, and if the word NO was detected, then KIND
would be returned containing 3+2 or 5.
The GETWRD routine is listed below.
SUBROUTINE GETWRD(IPACKD,IBUFFR,MAXBFR,LOWBFR,KIND )
DIMENSION IPACKD(100),IBUFFR(MAXBFR),KNTLTR(20),
1IWORD(100)
DATA ITTY,LMTWRD,LMTKNT/5,100,20/
DATA IWHAT/1H?/
C
C CONSTRUCT A1 FORMAT DICTIONARY
MAXWRD=0
MAXKNT=0
FASP, FORTRAN Alphameric Subroutine Package Page 507
DAVERB, Routine to Identify Words and Word Abbreviations
CALL A5TOA1(IPACKD,100,LMTWRD,LMTKNT,MAXWRD,MAXKNT,
1IWORD,KNTLTR)
C
C MATCH USER TYPED TEXT AGAINST A1 FORMAT DICTIONARY
MANY=1
IF(LOWBFR.GT.0)GO TO 1
LOWBFR=1
MANY=0
1 LOCK=MANY
CALL DALOSS(1,MAXWRD,IWORD,1,MAXKNT,
1KNTLTR,IBUFFR,MAXBFR,LOWBFR,KIND,MATCH,LCNWRD,
2LCNKNT,LCNBFR,MANY,LCNERR)
GO TO(5,6,12,12,6,6,6,6,2,4,11),KIND
C
C TREAT SEMICOLON LIKE COMMA
2 IF(LOCK.EQ.0)GO TO 3
MANY=-1
GO TO 1
3 LOWBFR=LOWBFR-1
GO TO 11
C
C BUFFER IS EMPTY
4 IF(MANY.LT.0)GO TO 11
5 KIND=1
GO TO 13
C
C UNKNOWN INITIAL CHARACTER
6 LOWBFR=LCNERR
LCNERR=LCNERR-1
KIND=2
WRITE(ITTY,7)IWHAT,(IBUFFR(I),I=LCNBFR,LCNERR),IWHAT
7 FORMAT(' ILLEGAL RESPONSE ',132A1)
WRITE(ITTY,8)
8 FORMAT(' RESPOND WITH ONE OF FOLLOWING')
LTREND=0
KNTEND=0
9 KNTEND=KNTEND+1
IF(KNTEND.GT.MAXKNT)GO TO 13
LTRBGN=LTREND+1
LTREND=LTREND+KNTLTR(KNTEND)
WRITE(ITTY,10)(IWORD(I),I=LTRBGN,LTREND)
10 FORMAT(3X,132A1)
GO TO 9
C
C MISSING WORD
11 KIND=3
GO TO 13
C
C CORRECT MATCH FOUND
12 KIND=MATCH+3
C
C RETURN TO CALLING PROGRAM
13 RETURN
END
FASP, FORTRAN Alphameric Subroutine Package Page 508
Appendix A: Program to Verify Routines Copied from Listings
Appendix A: Program to Verify Routines Copied from Listings
-------- - ------- -- ------ -------- ------ ---- --------
CHKSUM is a FORTRAN program which lists FORTRAN programs and
routines, generating a checksum for each noncomment line and
for each column excluding the characters in the comment
lines. The checksums are the least significant 5 digits in
the sums of the products of the line or column counts times
numbers identifying each character. The characters are
identified by the numbers 0 for the space (and tab
character), 1 through 26 for the letters A through Z, 27
through 36 for the digits 0 through 9, and 37 through 45 for
the characters + - * / = ( ) . and , respectively, with any
additional characters being assigned successively higher
numbers in the order in which these characters are
encountered. The file being read can contain several
programs and/or routines which must be written in standard
card format. If the end of file test (END= transfer in the
READ statement) is not available, then the END= must be
removed from the CHKSUM program and each input file should
then be terminated by an additional END statement which will
not be copied into the output file. It is expected that tab
characters will appear rarely if ever. Tab characters are
treated as single spaces in forming the checksums and are
converted to single spaces in the listing, but an asterisk
is displayed in the listing to the left of each line and
below each column containing tab characters.
Comment lines in which the only printing character is the
initial letter C are not included in the listing. At the
start of the program or routine, only the first 2 comments
with printing characters to the right of the initial letter
C are copied into the listing since initial comments often
duplicate the instruction manual. A comment line which
contains a printing character in column 2 and which appears
just before the END statement is treated specially. When a
program or routine is first read by this program, a comment
line which contains a printing character in column 2 and
which specifies a cumulative line checksum, a cumulative
column checksum, and a list of characters which did not have
predefined codes is inserted into the listing just prior to
the END statement. The cumulative checksums each contain 6
digits and do not include a representation of the END
statement. When the CHKSUM program processes a program or
routine which has been typed from the listing and which
includes the final comment line defining the expected
checksums, then the contents of this comment line are
compared with those which are expected based upon the
characters then in the program or routine. If the contents
of this final comment line are not as expected, then a new
final comment line specifying the checksums actually found
is written just before the END statement. If the cumulative
checksums are incorrect, then "CHKSM ERROR" is written at
the lower left corner of the listing, and the user should
FASP, FORTRAN Alphameric Subroutine Package Page 509
Appendix A: Program to Verify Routines Copied from Listings
compare the individual line and column checksums in the new
listing with those in the original listing to locate the
lines and columns which contain the errors. If the
checksums are correct but the list of characters is
incorrect (as could happen if a character which does not
have a predefined code appears once in the FORTRAN
statements but is copied incorrectly either there or on the
final comment line), then "SYMBL ERROR" is written instead.
The word VALID is written at the lower left corner of the
listing if no errors are found.
A listing produced by the CHKSUM program is shown below.
The routine did not include a checksum comment line before
the END statement, so neither of the words VALID nor ERROR
appears at the lower left corner of the listing.
16802 SUBROUTINE LENGTH(IBUFFR,MAXBFR,MAXPRT)
C ROUTINE TO FIND RIGHTMOST PRINTING CHARACTER
6377 DIMENSION IBUFFR(MAXBFR)
7662* DATA ISPACE,ITAB/1H ,1H /
3839 MAXPRT=MAXBFR+1
4491 1 MAXPRT=MAXPRT-1
7156 IF(MAXPRT.LE.0)GO TO 2
13905 IF(IBUFFR(MAXPRT).EQ.ISPACE)GO TO 1
12998 IF(IBUFFR(MAXPRT).EQ.ITAB)GO TO 1
1067 2 RETURN
C074297017153
183 END
*
000040521548644133888498238432139452113121124
0 023705504181720041392152872417384 26803
1 72673102428006247205019488 5 1852 0
6 4
If the routine is retyped from the listing and reprocessed
by CHKSUM, but with the letter O replacing the digit 0 in
the first IF statement, then the new listing would be
16802 SUBROUTINE LENGTH(IBUFFR,MAXBFR,MAXPRT)
C ROUTINE TO FIND RIGHTMOST PRINTING CHARACTER
6377 DIMENSION IBUFFR(MAXBFR)
7662* DATA ISPACE,ITAB/1H ,1H /
3839 MAXPRT=MAXBFR+1
4491 1 MAXPRT=MAXPRT-1
6916 IF(MAXPRT.LE.O)GO TO 2
13905 IF(IBUFFR(MAXPRT).EQ.ISPACE)GO TO 1
12998 IF(IBUFFR(MAXPRT).EQ.ITAB)GO TO 1
1067 2 RETURN
C074297017153
C074057017081
183 END
*
CHKSM 000040521548644133878498238432139452113121124
ERROR 0 023705504181750041392152872417384 26803
1 72673102428004247205019488 5 1852 0
6 4
FASP, FORTRAN Alphameric Subroutine Package Page 510
Appendix A: Program to Verify Routines Copied from Listings
Input/output formats are selected by the variables named
KMPR, KOPY, LEAD and NULL within the CHKSUM program. The
supplied values of KMPR=0, KOPY=2, LEAD=2 and NULL=1 specify
listings of the sort described and shown on the preceding
pages. Checksum comment lines are always inserted at the
end of the routines and programs which do not contain
correct checksum comment lines, but if KMPR, KOPY, LEAD and
NULL are all set to zero, then the routines and programs in
the input file will be copied with no other changes. The
variables which can be changed are described below.
IDSK = number of the input device from which the programs
and routines to be processed are read.
ILPT = unit number to which the copies of or listings of the
programs and routines are written.
ITAB = defined in a DATA statement in the program as a
single Hollerith tab character. If the tab character
is not available, then ITAB should instead be defined
as a single Hollerith space character.
KMPR = -1, each non-comment line in the input file has to
its right a 5-digit checksum in columns 73 through
77. If this line checksum is incorrect, then the
output file will contain 5 asterisks to the left of
the line containing the error if KOPY is greater than
zero, or to its right in columns 73 through 77 if
KOPY is less than or equal to zero.
= 0, the input file does not contain a checksum for
each non-comment line.
= 1, each non-comment line in the input file has to its
left a 5-digit checksum followed by a character which
is ignored. If this line checksum is incorrect, then
the output file will contain 5 asterisks to the left
of the line containing the error if KOPY is greater
than zero, or to its right in columns 73 through 77
if KOPY is less than or equal to zero. Each comment
line in the input file has to its left 6 characters
which are ignored. This input format results when
routines and/or programs are typed from listings
which were produced by the CHKSUM program. The
column checksums which appear in the listing
following the END statements need not be included in
the input file and are ignored if found.
= 2, similar to KMPR=1, except that the first character
is ignored in each line which is read. A listing
file produced by this program using KOPY=2 can be
processed using this value of KMPR.
KOPY = -1, a checksum is to be included in columns 73
through 77 to the right of each non-comment line in
the output file. All tab characters are copied
intact. Column checksums are not generated following
FASP, FORTRAN Alphameric Subroutine Package Page 511
Appendix A: Program to Verify Routines Copied from Listings
the END statements. The output file conforms to
FORTRAN format standards and can be compiled.
= 0, the individual lines of the programs and routines
are not to bear checksums in the output file. All
tab characters are copied intact. Column checksums
are not generated following the END statements. The
output file conforms to FORTRAN format standards and
can be compiled.
= 1, a 5-digit checksum and a separating space are to
be inserted to the left of each non-comment line in
the output file. Six spaces are inserted to the left
of each comment line in the output file. All tab
characters are copied intact. Column checksums are
not generated following the END statements.
= 2, a listing of the programs and routines in the
input file is to be created. Each line in the
listing will begin with a space as the carriage
control character. All tab characters in FORTRAN
statements in the input file are changed to spaces.
Line checksums appear to the left of each non-comment
line, and column checksums are generated following
the END statements.
LEAD = greater than zero, LEAD is the maximum number of
comment lines at the start of each routine or program
in the input file which are to be copied into the
output file. Comment lines are considered to be at
the start of a routine or program if they precede the
first FORTRAN statement, or, when no comment lines
appear before the first FORTRAN statement, if they
appear between the first and second FORTRAN
statements. If NULL has the value 0, then this also
includes comment lines in which the initial letter C
is the only printing character.
= 0, all comment lines which are at the start of the
routine or program and which contain at least one
printing character other than the initial letter C
are to be copied into the output file. If NULL also
has the value 0, then all comment lines are copied.
NULL = 0, comment lines which contain no printing characters
other than the initial letter C are to be treated in
the same manner as any other comment lines.
= 1, comment lines which contain no printing characters
other than the initial letter C are not copied.
FASP, FORTRAN Alphameric Subroutine Package Page 512
Appendix A: Program to Verify Routines Copied from Listings
The CHKSUM program, as listed by itself, is shown below.
C RENBR(CHKSUM/CHECKSUM LISTER AND VERIFIER)
C DONALD BARTH, HARVARD BUSINESS SCHOOL
35697 DIMENSION KODE(100),IBFR(80),JBFR(72),KLMSUM(72)
29164 1,KLMTAB(72),LTREND(3),IDIGIT(10),KOMMON(45)
14403 EQUIVALENCE(KODE(1),KOMMON(1))
20624 DATA KMPR,KOPY,LEAD,NULL,IDSK,ILPT,INITAL/
8427 10,2,2,1,1,20,45/
30835 DATA KOMMON/1HA,1HB,1HC,1HD,1HE,1HF,1HG,1HH,1HI,
36410 11HJ,1HK,1HL,1HM,1HN,1HO,1HP,1HQ,1HR,1HS,1HT,1HU,
40880 21HV,1HW,1HX,1HY,1HZ,1H0,1H1,1H2,1H3,1H4,1H5,1H6,
45080 31H7,1H8,1H9,1H+,1H-,1H*,1H/,1H=,1H(,1H),1H.,1H,/
8408 DATA LTREND/1HE,1HN,1HD/
39186 DATA IDIGIT/1H0,1H1,1H2,1H3,1H4,1H5,1H6,1H7,1H8,
1152 11H9/
25822* DATA ISTAR,KOMENT,ISPACE,ITAB/1H*,1HC,1H ,1H /
C GENERAL INPUT AND OUTPUT FORMATS
3552 1 FORMAT(80A1)
5939 2 FORMAT(72A1,1I5)
5525 3 FORMAT(6X,72A1)
5544 4 FORMAT(7X,72A1)
8197 5 FORMAT(1I5,1X,72A1)
7480 6 FORMAT(1X,I5,73A1)
9500 7 FORMAT(1X,6HVALID ,72A1)
10102 8 FORMAT(1X,6HERROR ,72A1)
9984 9 FORMAT(1X,6HSYMBL ,72A1)
9762 10 FORMAT(1X,6HCHKSM ,72A1)
12526 11 FORMAT(1X,5H*****,73A1)
12441 12 FORMAT(5H*****,1X,72A1)
10097 13 FORMAT(72A1,5H*****)
6588 14 FORMAT(1X/1X/1X)
22827 15 FORMAT(7X,1HC,1I6,24H ROUTINES CONTAIN ERRORS)
C PREPARE FOR FIRST ROUTINE
1586 KIND=-1
1744 KNTERR=0
7283 IF(KMPR.LT.0)GO TO 18
7103 IF(KMPR.EQ.0)GO TO 17
7094 IF(KMPR.EQ.1)GO TO 16
1529 KMPBGN=2
1572 KMPEND=6
1570 LINBGN=8
2083 LINEND=79
2187 INPUT=79
1402 GO TO 19
1792 16 KMPBGN=1
1558 KMPEND=5
1556 LINBGN=7
2068 LINEND=78
2173 INPUT=78
1402 GO TO 19
1754 17 LINBGN=1
1978 LINEND=72
2089 INPUT=72
FASP, FORTRAN Alphameric Subroutine Package Page 513
Appendix A: Program to Verify Routines Copied from Listings
1402 GO TO 19
1759 18 LINBGN=1
1978 LINEND=72
2049 KMPBGN=73
2096 KMPEND=77
2159 INPUT=77
C PREPARE FOR NEXT ROUTINE TO BE LISTED
2493 19 KNOWN=INITAL
1357 MOST=0
1697 LINKNT=0
1175 LONG=0
1443 LINCHK=0
1451 KLMCHK=0
4810 DO 20 KLMN=1,72
4286 KLMSUM(KLMN)=0
4806 20 KLMTAB(KLMN)=ISPACE
C READ NEXT LINE AND FIND RIGHT PRINTING CHARACTER
23853 21 READ(IDSK,1,END=46)(IBFR(I),I=1,INPUT)
2896 MORE=LINEND+1
2897 22 MORE=MORE-1
8501 IF(MORE.LT.LINBGN)GO TO 21
12244 IF(IBFR(MORE).EQ.ISPACE)GO TO 22
2682 LESS=MORE+1
3046 23 LESS=LESS-1
8575 IF(LESS.LT.LINBGN)GO TO 21
12376 IF(IBFR(LESS).EQ.ISPACE)GO TO 23
11491 IF(IBFR(LESS).EQ.ITAB)GO TO 23
14121 IF(IBFR(LINBGN).NE.KOMENT)GO TO 38
8517 IF(LESS.EQ.LINBGN)GO TO 33
6578 IF(KIND.LT.0)KIND=1
15398 IF(IBFR(LINBGN+1).EQ.ISPACE)GO TO 34
14425 IF(IBFR(LINBGN+1).EQ.ITAB)GO TO 34
C TEST IF CHECKSUMMED COMMENT IS CORRECT
1718 KLMN=LINBGN
2123 IVALUE=LINCHK
3138 DO 26 J=1,2
1577 JVALUE=0
3182 DO 25 I=1,6
2629 KLMN=KLMN+1
8020 IF(KLMN.GT.LESS)GO TO 30
3598 DO 24 K=1,10
15909 IF(IBFR(KLMN).NE.IDIGIT(K))GO TO 24
8802 JVALUE=(10*JVALUE)+K-1
1359 GO TO 25
1348 24 CONTINUE
1302 GO TO 30
1353 25 CONTINUE
9515 IF(IVALUE.NE.JVALUE)GO TO 30
2419 26 IVALUE=KLMCHK
6674 ITEST=KLMN+KNOWN-INITAL
10915 IF(ITEST.GT.LINEND)ITEST=LINEND
8721 IF(LESS.NE.ITEST)GO TO 28
1132 I=INITAL
1436 27 I=I+1
FASP, FORTRAN Alphameric Subroutine Package Page 514
Appendix A: Program to Verify Routines Copied from Listings
7161 IF(I.GT.KNOWN)GO TO 32
2629 KLMN=KLMN+1
14838 IF(IBFR(KLMN).EQ.KODE(I))GO TO 27
6947 28 IF(LEAD.LE.0)GO TO 29
7195 IF(KIND.GT.LEAD)GO TO 21
1558 29 LAST=2
1386 GO TO 36
6721 30 IF(LEAD.LE.0)GO TO 31
7195 IF(KIND.GT.LEAD)GO TO 21
1534 31 LAST=3
1386 GO TO 36
1503 32 LAST=0
1400 GO TO 37
C COMMENT HAVING SPACE OR TAB IN COLUMN 2
7307 33 IF(NULL.GT.0)GO TO 21
6578 IF(KIND.LT.0)KIND=1
6849 34 IF(LEAD.LE.0)GO TO 35
7195 IF(KIND.GT.LEAD)GO TO 21
2014 35 LAST=-1
9000 36 IF(KIND.GT.0)KIND=KIND+1
C OUTPUT COMMENT LINE
27860 37 IF(KOPY.GT.1)WRITE(ILPT,4)(IBFR(I),I=LINBGN,
1126 1LESS)
27461 IF(KOPY.EQ.1)WRITE(ILPT,3)(IBFR(I),I=LINBGN,
1126 1LESS)
27294 IF(KOPY.LE.0)WRITE(ILPT,1)(IBFR(I),I=LINBGN,
1126 1LESS)
1303 GO TO 21
C CHECK FOR COLUMN CHECKSUMS IN INPUT FILE
6902 38 IF(KIND.GE.0)GO TO 41
5922 DO 40 KLMN=LINBGN,LESS
4641 LETTER=IBFR(KLMN)
9319 IF(LETTER.EQ.ISPACE)GO TO 40
9484 IF(LETTER.EQ.ISTAR)GO TO 40
3637 DO 39 I=1,10
12520 IF(LETTER.EQ.IDIGIT(I))GO TO 40
1377 39 CONTINUE
1329 GO TO 41
1336 40 CONTINUE
1303 GO TO 21
C TEST FOR END STATEMENT
3256 41 KLMN=LINBGN+5
12317 IF(IBFR(KLMN).EQ.ISPACE)GO TO 42
11430 IF(IBFR(KLMN).EQ.ITAB)GO TO 42
16509 IF(IBFR(KLMN).EQ.IDIGIT(1))GO TO 42
6714 IF(KIND.GE.0)GO TO 54
1371 GO TO 44
2235 42 LTRTST=0
2903 43 KLMN=KLMN+1
8169 IF(KLMN.GT.LESS)GO TO 44
12355 IF(IBFR(KLMN).EQ.ISPACE)GO TO 43
11466 IF(IBFR(KLMN).EQ.ITAB)GO TO 43
4734 LTRTST=LTRTST+1
20672 IF(IBFR(KLMN).NE.LTREND(LTRTST))GO TO 44
FASP, FORTRAN Alphameric Subroutine Package Page 515
Appendix A: Program to Verify Routines Copied from Listings
8605 IF(LTRTST.LT.3)GO TO 43
8121 IF(KLMN.EQ.LESS)GO TO 45
6966 44 IF(KIND.GE.0)GO TO 53
1102 KIND=1
1384 GO TO 54
C CONSTRUCT CHECKSUMMED COMMENT IF NONE OR WRONG
1870 45 KIND=-1
8534 46 IF(LINKNT.EQ.0)GO TO 78
7012 IF(LAST.EQ.0)GO TO 52
3633 JBFR(1)=KOMENT
1217 KLMN=1
2123 IVALUE=LINCHK
3191 DO 49 J=1,2
1577 JVALUE=0
3224 DO 47 I=1,6
4584 LFTOVR=IVALUE/10
20157 JVALUE=(10*JVALUE)+IVALUE-(10*LFTOVR)
3057 47 IVALUE=LFTOVR
3235 DO 48 I=1,6
4598 LFTOVR=JVALUE/10
12138 JDIGIT=JVALUE-(10*LFTOVR)+1
2770 JVALUE=LFTOVR
2629 KLMN=KLMN+1
8364 48 JBFR(KLMN)=IDIGIT(JDIGIT)
2442 49 IVALUE=KLMCHK
1132 I=INITAL
1413 50 I=I+1
7187 IF(I.GT.KNOWN)GO TO 51
7621 IF(KLMN.GE.72)GO TO 51
2629 KLMN=KLMN+1
5678 JBFR(KLMN)=KODE(I)
1328 GO TO 50
30716 51 IF(KOPY.GT.1)WRITE(ILPT,4)(JBFR(I),I=1,KLMN)
30339 IF(KOPY.EQ.1)WRITE(ILPT,3)(JBFR(I),I=1,KLMN)
30172 IF(KOPY.LE.0)WRITE(ILPT,1)(JBFR(I),I=1,KLMN)
8070 52 IF(KIND.EQ.-1)GO TO 55
1354 GO TO 70
C IDENTIFY CHARACTERS AND CONSTRUCT CHECK SUMS
1368 53 KIND=0
2017 54 LAST=-1
4281 55 LINKNT=LINKNT+1
1883 LINTAB=ISPACE
1770 LINSUM=0
1705 KLMKNT=0
5830 DO 62 KLMN=LINBGN,MORE
4016 KLMKNT=KLMKNT+1
4641 LETTER=IBFR(KLMN)
9453 IF(LETTER.EQ.ISPACE)GO TO 62
8614 IF(LETTER.EQ.ITAB)GO TO 61
1812 NEWLTR=0
4600 56 NEWLTR=NEWLTR+1
10232 IF(NEWLTR.GT.KNOWN)GO TO 57
14770 IF(KODE(NEWLTR).NE.LETTER)GO TO 56
1440 GO TO 58
FASP, FORTRAN Alphameric Subroutine Package Page 516
Appendix A: Program to Verify Routines Copied from Listings
3866 57 KNOWN=KNOWN+1
5042 KODE(KNOWN)=LETTER
12323 58 LINSUM=LINSUM+(KLMKNT*NEWLTR)
7001 IF(KIND.LT.0)GO TO 62
26412 KLMSUM(KLMKNT)=KLMSUM(KLMKNT)+(LINKNT*NEWLTR)
11149 KLMCHK=KLMCHK+(LINKNT*NEWLTR)
19308 59 IF(KLMSUM(KLMKNT).LT.100000)GO TO 60
19758 KLMSUM(KLMKNT)=KLMSUM(KLMKNT)-100000
1454 GO TO 59
12970 60 IF(KLMCHK.LT.1000000)GO TO 62
7336 KLMCHK=KLMCHK-1000000
1341 GO TO 60
11980 61 IF(KOPY.GT.1)IBFR(KLMN)=ISPACE
5761 KLMTAB(KLMKNT)=ISTAR
10724 IF(LONG.LT.KLMKNT)LONG=KLMKNT
2122 LINTAB=ISTAR
1354 62 CONTINUE
11487 IF(MOST.LT.KLMKNT)MOST=KLMKNT
4827 LINCHK=LINCHK+LINSUM
13044 63 IF(LINCHK.LT.1000000)GO TO 64
7313 LINCHK=LINCHK-1000000
1383 GO TO 63
12637 64 IF(LINSUM.LT.100000)GO TO 65
7455 LINSUM=LINSUM-100000
1397 GO TO 64
C DETERMINE IF CHECKSUM ON LINE IS CORRECT
7552 65 IF(KMPR.EQ.0)GO TO 68
1570 IVALUE=0
6353 DO 67 KLMN=KMPBGN,KMPEND
3634 DO 66 I=1,10
16089 IF(IBFR(KLMN).NE.IDIGIT(I))GO TO 66
8725 IVALUE=(10*IVALUE)+I-1
1439 GO TO 67
1374 66 CONTINUE
1379 67 CONTINUE
10363 IF(IVALUE.EQ.LINSUM)GO TO 68
8060 IF(IVALUE.EQ.0)GO TO 68
26943 IF(KOPY.GT.1)WRITE(ILPT,11)LINTAB,(IBFR(I),
3639 1I=LINBGN,MORE)
28739 IF(KOPY.EQ.1)WRITE(ILPT,12)(IBFR(I),I=LINBGN,
1064 1MORE)
28666 IF(KOPY.LE.0)WRITE(ILPT,13)(IBFR(I),I=LINBGN,
1261 1LINEND)
1467 GO TO 69
C OUTPUT THIS LINE OF FORTRAN TEXT
22102 68 IF(KOPY.GT.1)WRITE(ILPT,6)LINSUM,LINTAB,
8402 1(IBFR(I),I=LINBGN,MORE)
26853 IF(KOPY.EQ.1)WRITE(ILPT,5)LINSUM,(IBFR(I),
3639 1I=LINBGN,MORE)
27381 IF(KOPY.EQ.0)WRITE(ILPT,1)(IBFR(I),I=LINBGN,
1064 1MORE)
27565 IF(KOPY.LT.0)WRITE(ILPT,2)(IBFR(I),I=LINBGN,
3454 1LINEND),LINSUM
6867 69 IF(KIND.GE.0)GO TO 21
FASP, FORTRAN Alphameric Subroutine Package Page 517
Appendix A: Program to Verify Routines Copied from Listings
C CONSTRUCT CHECKSUM TOTALS BELOW COLUMNS
12214 70 IF(LAST.GT.0)KNTERR=KNTERR+1
7303 IF(KOPY.LE.1)GO TO 77
31606 IF(LONG.GT.0)WRITE(ILPT,4)(KLMTAB(I),I=1,LONG)
11986 IF(LONG.EQ.0)WRITE(ILPT,4)
5046 DO 72 KLMN=1,MOST
5727 IVALUE=KLMSUM(KLMN)
1591 JVALUE=1
4860 71 LFTOVR=IVALUE/10
15885 JVALUE=(10*(JVALUE-LFTOVR))+IVALUE
2763 IVALUE=LFTOVR
7973 IF(IVALUE.GT.0)GO TO 71
5616 72 KLMSUM(KLMN)=JVALUE
1461 73 LONG=0
5079 DO 75 KLMN=1,MOST
5660 JDIGIT=KLMSUM(KLMN)
7980 IF(JDIGIT.GT.1)GO TO 74
4555 KLMTAB(KLMN)=ISPACE
1424 GO TO 75
4731 74 LFTOVR=JDIGIT/10
11980 JDIGIT=JDIGIT-(10*LFTOVR)+1
5875 KLMSUM(KLMN)=LFTOVR
9026 KLMTAB(KLMN)=IDIGIT(JDIGIT)
1531 LONG=KLMN
1373 75 CONTINUE
7010 IF(LONG.LE.0)GO TO 76
31756 IF(LAST.LT.0)WRITE(ILPT,4)(KLMTAB(I),I=1,LONG)
31696 IF(LAST.EQ.0)WRITE(ILPT,7)(KLMTAB(I),I=1,LONG)
31745 IF(LAST.EQ.1)WRITE(ILPT,8)(KLMTAB(I),I=1,LONG)
31794 IF(LAST.EQ.2)WRITE(ILPT,9)(KLMTAB(I),I=1,LONG)
32925 IF(LAST.EQ.3)WRITE(ILPT,10)(KLMTAB(I),I=1,LONG)
7854 IF(LAST.LE.1)LAST=-1
7023 IF(LAST.GT.0)LAST=1
1396 GO TO 73
5009 76 WRITE(ILPT,14)
8107 77 IF(KIND.EQ.-1)GO TO 19
8553 78 IF(KNTERR.LE.0)GO TO 79
18075 IF(KOPY.GT.1)WRITE(ILPT,15),KNTERR
1826 79 ENDFILE ILPT
588 STOP
C398993428599
183 END
*
VALID 000445663413667234416441951158843631332962582346694611
689475201587047869826639125427572846518981628485338
796924233646001894585241200520045176521484938178344
511398734598041905024462442983817228963367565077253
170 12 4 65 84570916 544 9242 9297 700843 512813