Google
 

Trailing-Edge - PDP-10 Archives - decuslib10-09 - 43,50466/bmd.doc
There are 2 other files named bmd.doc in the archive. Click here to see a list.
                         WESTERN MICHIGAN UNIVERSITY
                               COMPUTER CENTER
LIBRARY PROGRAM #1.1.5
PROGRAMMED BY:  *
PREPARED BY:  RUSSELL R. BARR III & SAM ANEMA
APPROVED BY:  JACK R. MEAGHER
DATE:  FEBRUARY, 1977 (VERSION 2)
 
                           UCLA BIOMEDICAL PACKAGE
 
THE COMPUTER CENTER HAS ADAPTED THE FOLLOWING BIOMEDICAL PROGRAMS FOR USE
ON THE PDP-10:
 
   BIOMEDICAL PROGRAM
          NAME                     PROGRAM DESCRIPTION
 
         BMD01D            SIMPLE DATA DESCRIPTION
         BMD02D            CORRELATION WITH TRANSGENERATION
         BMD03D            CORRELATION WITH ITEM DELETION
         BMD05D            GENERAL PLOT INCLUDING HISTOGRAM
         BMD06D            DESCRIPTION OF STRATA
         BMD07D            DESCRIPTION OF STRATA WITH HISTOGRAM
         BMD08D            CROSS-TABULATION WITH VARIABLE STACKING
         BMD09D            CROSS-TABULATION, INCOMPLETE DATA
         BMD10D            DATA PATTERNS FOR DICHOTOMIES
         BMD11D            DATA PATTERNS FOR POLYCHOTOMIES
         BMD01M            PRINCIPAL COMPONENT ANALYSIS
         BMD02M            REGRESSION ON PRINCIPAL COMPONENTS
         BMD03M            FACTOR ANALYSIS
         BMD04M            DISCRIMINANT ANALYSIS FOR TWO GROUPS
         BMD05M            DISCRIMINANT ANALYSIS FOR SEVERAL GROUPS
         BMD06M            CANONICAL ANALYSIS
         BMD07M            STEPWISE DISCRIMINANT ANALYSIS
         BMD01R            SIMPLE LINEAR REGRESSION
         BMD02R            STEPWISE REGRESSION
         BMD03R            MULTIPLE REGRESSION WITH CASE COMBINATIONS
         BMD04R            PERIODIC REGRESSION AND HARMONIC ANALYSIS
         BMD05R            POLYNOMIAL REGRESSION
         BMD06R            ASYMPTOTIC REGRESSION
         BMD01S            LIFE TABLE AND SURVIVAL RATE
         BMD02S            CONTINGENCY TABLE ANALYSIS
         BMD03S            BIOLOGICAL ASSAY: PROBIT ANALYSIS
         BMD04S            GUTTMAN SCALE PREPROCESSOR
         BMD05S            GUTTMAN SCALE #1
         BMD06S            GUTTMAN SCALE #2, PART 1
         BMD07S            GUTTMAN SCALE #2, PART 2
         BMD08S            GUTTMAN SCALE #2, PART 3
         BMD09S            TRANSGENERATION
         BMD10S            TRANSPOSITION OF LARGE MATRICES
         BMD01T            AMPLITUDE AND PHASE ANALYSIS
         BMD02T            AUTOCOVARIANCE AND POWER SPECTRAL ANALYSIS
         BMD01V            ANALYSIS OF VARIANCE FOR ONE-WAY DESIGN
         BMD02V            ANALYSIS OF VARIANCE FOR FACTORIAL DESIGN
 
-------------------
 
*  SOME OF THE OPERATING PROCEDURES WERE WORKED OUT BY MR. GEORGE KOHRMAN
AND RUSSELL R. BARR III.  THE CONVERSION OF THE BMD PROGRAMS TO THE PDP-10 WAS
DONE BY MISS GAN AND MR. ANEMA.
 
         BMD03V            ANALYSIS OF COVARIANCE FOR FACTORIAL DESIGN
         BMD04V            ANALYSIS OF COVARIANCE WITH MULTIPLE COVARIANCE
         BMD05V            GENERAL LINEAR HYPOTHESIS
         BMD06V            GENERAL LINEAR HYPOTHESIS WITH CONTRASTS
         BMD08V            ANALYSIS OF VARIANCE
         BMDX63            MULTIVARIATE GENERAL LINEAR HYPOTHESIS
         BMDX64            GENERAL LINEAR HYPOTHESIS
         BMDX68            MULTIPLE TIME SERIES SPECTRAL ANALYSIS
         BMDX69            MULTIVARIATE ANALYSIS OF VARIANCE AND COVARIANCE
         BMDX70            T-PROGRAM
         BMDX72            FACTOR ANALYSIS
         BMDX75            CANONICAL ANALYSIS
         BMDX77            TRANSGENERATION
         BMDX82            ANALYSIS OF COVARIANCE
         BMDX84            ASYMMETRICAL CORRELATION WITH MISSING DATA
         BMDX85            NON-LINEAR LEAST SQUARES
         BMDX92            TIME SERIES SPECTRUM ESTIMATION
         BMDX93            TIME-LOCKED AVERAGING PROGRAM
         BMDX94            MULTIPASS TRANSGENERATION
 
FOR A COMPLETE DESCRIPTION OF THESE PROGRAMS AND FOR INSTRUCTIONS IN THEIR USE,
SEE THE FOLLOWING BOOKS:
 
     FOR PROGRAMS BMD01D - BMD08V, SEE:
 
         BMD
         BIOMEDICAL COMPUTER PROGRAMS
         W. J. DIXON, EDITOR
         UNIVERSITY OF CALIFORNIA PRESS
         BERKELEY, LOS ANGELES, LONDON
         1970
 
     FOR PROGRAMS BMDX63 - BMDX94, SEE:
 
         BMD
         BIOMEDICAL COMPUTER PROGRAMS
         X-SERIES SUPPLEMENT
         W. J. DIXON, EDITOR
         UNIVERSITY OF CALIFORNIA PRESS
         BERKELEY, LOS ANGELES, LONDON
         1970
 
A COPY OF EACH OF THESE BOOKS HAVE BEEN ORDERED FOR:  (1) PHYSICAL SCIENCE
LIBRARY, (2) WALDO LIBRARY, AND (3) BUSINESS LIBRARY.
 
THE BIOMEDICAL PROGRAMS ARE DESIGNED TO INPUT DATA FROM CARDS, DISK, OR
MAGNETIC TAPE.  THE OUTPUT IS DESIGNED FOR THE LINE PRINTER.  BECAUSE OF THIS,
THE PROGRAMS CAN BEST BE RUN FROM BATCH.
 
THE GENERAL CARD SETUP IS AS FOLLOWS:
 
     FOR DATA ON CARDS:
 
         $JOB [PROJ,PROG]
         $PASSWORD PSWD
         $DECK FILE.F4
         [AT THIS POINT ENTER FORTRAN CARDS.] #
         $DATA
         [AT THIS POINT ENTER ALL OTHER CARDS DESCRIBED IN THE
         BMD MANAUL EXCEPT SYSTEM CARDS.]  @
         $EOD
         $COPY BMD--- (THE SIX CHARACTERS FOLLOWING THE WORD "COPY" MUST
                       BE THE BMD PROGRAM NAME.)
         (EOF)
 
     FOR DATA ON A DISK FILE CALLED FILE.DAT:
 
         $JOB [PROJ,PROG]
         $PASSWORD PSWD
         .AS DSK  7
         .R PIP
         *FOR07.DAT  FILE.DAT   (WHERE FILE.DAT IS THE NAME OF THE DATA FILE)
         $DECK FILE.F4
         [AT THIS POINT ENTER FORTRAN CARDS.]  #
         $DATA
         [AT THIS POINT ENTER ALL OTHER CARDS DESCRIBED IN THE BMD MANUAL
         EXCEPT SYSTEM CARDS.  @ YOU MUST SPECIFY 7 AS AN AUXILIARY
         INPUT DEVICE NUMBER ON PROBLM CARD.]
         $EOD
         $COPY BMD--- (THE SIX CHARACTERS FOLLOWING THE WORD "COPY" MUST
                       BE THE BMD PROGRAM NAME.)
         (EOF)
 
     FOR DATA ON MAGNETIC TAPE:
 
         $JOB [PROJ,PROG]
         $PASSWORD PSWD
         .MOUNT MTA 7 /REE:---- (---- IS THE TAPE IDENTIFICATION NUMBER)
         $DECK FILE.F4
         [AT THIS POINT ENTER ALL OTHER CARDS DESCRIBED IN THE BMD MANUAL
         EXCEPT SYSTEM CARDS.  @ YOU MUST SPECIFY 7 AS AN AUXILIARY
         INPUT DEVICE NUMBER ON PROBLM CARD]
         $EOD
         $COPY BMD--- (THE SIX CHARACTERS FOLLOWING THE WORD "COPY" MUST
                       BE THE BMD PROGRAM NAME.)
         .DISMOUNT 7: /R
         (EOF)
 
 
@ A SYSTEM CARD IS ANY CARD WHICH BEGINS WITH //, /* OR $ID
# THESE CARDS AND THE "DECK" CARD ARE ONLY USED WITH BMDX77, BMDX85, BMDX94.
 
RUNNING BMD PROGRAMS WITHOUT USING CARDS
 
THE BMD INVOCATION SEQUENCE IS COMPLEX ENOUGH THAT BATCH ENTRY IS THE MOST
PRACTICAL APPROACH.  HOWEVER, ACCESS TO BATCH IS POSSIBLE FROM A TERMINAL,
WITHOUT THE USE OF PUNCHED CARDS, BY THE USE OF THE SUBMIT COMMAND.
 
THIS METHOD OF SUBMITTING A BATCH COMMAND FILE IS ONLY APPLICABLE FOR DATA
STORED ON DISK.
 
PROCEDURE:
 
1)  MAKE A FILE CONTAINING THE PARAMETER STATEMENTS DESCRIBED IN THE BMD MANUAL,
EXCEPT SYSTEM STATEMENTS.  YOU MUST SPECIFY "7" AS AN AUXILIARY INPUT DEVICE
NUMBER ON THE PROBLM STATEMENT IF THE DATA IS NOT INCLUDED IN THE PARAMETER
FILE.
 
2)  IF FORTRAN STATEMENTS ARE NEEDED, MAKE A FILE CALLED FILE.F4 CONTAINING
THESE STATEMENTS.
 
3)  IF YOUR DATA IS NOT INCLUDED IN THE PARAMETER FILE, CREATE A FILE ON DISK
CONTAINING YOUR DATA.
 
4)  RUN THE PROGRAM BMDSMT TO CREATE THE CONTROL FILE NEEDED.
 
5)  TYPE:  SUBMIT BMD
 
6)  YOUR OUTPUT WILL GO TO THE LINE PRINTER.
 
EXAMPLE RUN
 
.R BMDSMT
 
ENTER NAME OF BMD PROGRAM (E.G. BMD01D)
 
BMDX72
 
ENTER NAME OF THE PARAMETER FILE
 
PRM.DAT
 
ENTER THE NAME OF YOUR DATA FILE
JUST ENTER A <CR> IF DATA IS INCLUDED IN PARAMETER FILE
 
INPUT.DAT
 
THE NEXT STEP IS TO "SUBMIT BMD"
 
END OF EXECUTION
CPU TIME:  0.55  ELAPSED TIME:  31.33
EXIT
 
.SUBMIT BMD
 
.
 
NOTE:
 
1)  THE NORMAL TIME ALLOWED FOR THE PROCESSING OF A BATCH JOB, (SUBMITTED OR
WITH CARDS) IS 3 MINUTES.  IF THE JOB IS LARGE YOU MAY WISH TO INCREASE THIS.
THE COMMAND TO INCREASE THE TIME OF THE ABOVE JOB TO 6 MINUTES WOULD BE:
SUBMIT BMD/TIME:6:00.
 
2)  IF YOU WISH TO DEFER THE START OF A JOB TO A DIFFERENT TIME OF DAY,
PERHAPS TO TAKE ADVANTAGE OF A LIGHTER SYSTEM LOAD OR CHEAPER RATES, YOU MAY
USE THE /AFTER SWITCH.  SUPPOSE YOU WISH TO DEFER THE START OF THE JOB
13-1/2 HOURS, YOU COULD SUBMIT THE JOB LIKE THIS:  SUBMIT BMD/AFTER:+13:30.
 
3)  BOTH SWITCHES MAY BE USED IF DESIRED OR YOU MAY USE ANY OF THE SWITCHES
LISTED IN THE BATCH SYSTEM COMMANDS SECTION OF THE DECSYSTEM-10 USERS HAND-
BOOK.
 
THE PDP-10 VERSION OF THE BIOMEDICAL PROGRAMS ARE LIMITED IN PROBLEM SIZE
BY THE LIMITATIONS DESCRIBED IN THE BMD MANUALS.  IN ADDITION TO THESE RESTRIC-
TIONS, THE FOLLOWING RESTRICTIONS APPLY:
 
     I.  IN EACH OF THE BIOMEDICAL PROGRAMS, A BLANK IS READ IN AS A ZERO. THIS
         IS TRUE IN SPITE OF WHAT MAY BE SAID ABOUT THIS IN THE BMD MANUAL.
 
    II.  WHENEVER PUNCHED CARD OUTPUT IS EXPECTED,
         1.  INSERT AN ADDITIONAL CARD BEFORE THE $COPY CARD WHICH CONTAINS:
             .MOUNT CDP PTP /FORMS:XXX
             "XXX" SHOULD BE REPLACED WITH THE APPROXIMATE NUMBER OF CARDS
             TO BE PUNCHED.
 
   III.  THE FOLLOWING IS A LIST OF ADDITIONAL RESTRICTIONS AS THEY APPLY
         TO THE INDIVIDUAL PROGRAM INVOLVED:
 
         BMD02D  (1)  P, NUMBER OF ORIGINAL VARIABLES (2<P<50)
                 (2)  Q, NUMBER OF VARIABLES ADDED TO THE ORIGINAL SET AFTER
                      TRANSGENERATION (-48<=Q<=48)
 
         BMD03D  (1)  P, NUMBER OF VARIABLES (1<=P<=62)
 
         BMD05D  (1)  N, NUMBER OF CASES (2<=N<=5000)
 
         BMD06D  (1)  P, NUMBER OF ORIGINAL VARIABLES (2<=P<=25)
                 (2)  N, SAMPLE SIZES OR NUMBER OF CASES (3<=N<=400)
                 (3)  Q, NUMBER OF VARIABLES ADDED TO THE ORIGINAL SET AFTER
                      TRANSGENERATION (-9<=Q<=23)  (2<=P+Q<=25)
         BMD07D  (1)  THE PRODUCT OF THE NUMBER OF VARIABLES AFTER TRANS-
                      GENERATION AND THE NUMBER OF CASES MAY NOT EXCEED 4000.
 
         BMD08D  (1)  THE TOTAL NUMBER OF VARIABLES AFTER STACKING AND TRANS-
                      GENERATION MULTIPLIED BY THE NUMBER OF CASES MUST BE LESS
                      THAN 8000.
 
         BMD09D  (1)  THE TOTAL NUMBER OF VARIABLES AFTER TRANSGENERATION
                      MULTIPLIED BY THE NUMBER OF CASES MUST BE LESS THAN
                      6000.  IF PRINTING OF THE IDENTIFICATION NUMBER IN THE
                      DATA MATRIX IS DESIRED, IT MUST BE LESS THAN 4000.
 
         BMD10D  (1)  P, NUMBER OF VARIABLES (1<=P<=20)
                 (2)  N, NUMBER OF ITEMS OR CASES (1<=N<=400)
 
         BMD11D  (1)  P, NUMBER OF VARIABLES (1<=P<=20)
                 (2)  N, NUMBER OF ITEMS OR CASES (1<=N<=300)
 
         BMD01M  (1)  P, NUMBER OF ORIGINAL VARIABLES (2<=P<=20)
                 (2)  N, NUMBER OF CASES (3<=N<=250)
                 (3)  Q, NUMBER OF VARIABLES ADDED BY TRANSGENERATION (-18<=Q
                      <=18)
                 (4)  P+Q, TOTAL NUMBER OF VARIABLES (2<=P+Q<=20)
 
         BMD02M  (1)  P, NUMBER OF INDEPENDENT VARIABLES (1<=P<=15)
                 (2)  D, NUMBER OF DEPENDENT VARIABLES (1<=D<=15)
                 (3)  T, TOTAL NUMBER OF ORIGINAL VARIABLES (2<=T<=30)
 
         BMD03M  (1)  NUMBER OF VARIABLES (2<=P<=60)
 
         BMD04M  (1)  N(I), SAMPLE SIZE IN I-TH GROUP (P+Q<=N(I)<=250)
                 (2)  P, NUMBER OF ORIGINAL VARIABLES (2<=P<=20)
                 (3)  Q, NUMBER OF VARIABLES ADDED TO THE ORIGINAL SET OF
                      VARIABLES (-9<=Q<=18), (2<=P+Q<=20)
 
         BMD05M  (1)  G, NUMBER OF GROUPS (2<=G<=4)
                 (2)  P, NUMBER OF ORIGINAL VARIABLES (G<=P<=15)
                 (3)  Q, NUMBER OF VARIABLES ADDED TO THE ORIGINAL SET AFTER
                      TRANSGENERATION (-9<=Q<=15), (G<=P+Q<=15)
 
         BMD01R  (1)  T, NUMBER OF TREATMENT GROUPS (1<=T<=599)
                 (2)  N(I), NUMBER OF CASES IN THE I-TH TREATMENT GROUP (1<=N(I)
                      <=599)
 
         BMD02R  (1)  P, NUMBER OF ORIGINAL VARIABLES (2<=P<=50)
                 (2)  Q, NUMBER OF VARIABLES ADDED BY TRANSGENERATION (-9<=Q
                      <=48)
                 (3)  P+Q, TOTAL NUMBER OF VARIABLES (2<=P+Q<=50)
 
         BMD03R  (1)  P, ORIGINAL NUMBER OF VARIABLES (2<=P<=40)
                 (2)  Q, NUMBER OF VARIABLES ADDED TO THE ORIGINAL SET AFTER
                      TRANSGENERATION (-9<=Q<=38), (2<=P+Q<=40)
                 (3)  T, NUMBER OF TRANSGENERATION CARDS (0<=T<=40)
 
         BMD04R  (1)  N, NUMBER OF HARMONICS (1<=N<=9)
                 (2)  K, NUMBER OF TIME VALUES (2N+1<=K<=300)
                 (3)  M, NUMBER OF APPLICATIONS (FOR ANALYSIS OF VARIANCE)
                      (2<=M<=15)
                      M, NUMBER OF REPLICATIONS (NO ANALYSIS OF VARIANCE)
                      (1<=M<=15)
 
         BMD05R  (1)  K, FINAL DEGREE POLYNOMIAL REGRESSION (K<=5)
                 (2)  N, NUMBER OF CASES (K+1<=N<=400)
 
         BMD06R  (1)  K, NUMBER OF VALUES OF THE INDEPENDENT VARIABLE (4<=K<=
                      399)
                 (2)  M(I), NUMBER OF VARIABLES OF THE DEPENDENT VARIABLE CO-
                      RRESPONDING TO THE I-TH VALUE OF THE INDEPENDENT VARIABLE
                      (1<=M(I)<=399)
                 (3)  N, TOTAL NUMBER OF THE DEPENDENT VARIABLE M(1)+...+M(K)=N
                      (K<=N<=1700)
 
         BMD01S  (1)  (FOR ALL FORMS OF INPUT) T, NUMBER OF TIME PERIODS
                      (T<=31)
 
         BMD02S  (1)  K, NUMBER OF INTERVALS IN EACH CATEGORIZATION (K<=14)
                 (2)  S, TOTAL NUMBER OF DIFFERENT CATEGORIZATIONS FORMED FROM
                      ALL VARIABLES (S<350)
                 (3)  Q, THE TOTAL NUMBER OF INTERVALS FOR ALL VARIABLES (Q<=
                      2000)
 
                                                      8000
         BMD04S  (1)  N, NUMBER OF RESPONDENTS (1<=N<=----)
                                                      P+4
 
                                                      8000
         BMD05S  (1)  N, NUMBER OF RESPONDENTS (1<=N<=----)
                                                      P+4
 
                                                      8000
         BMD06S  (1)  N, NUMBER OF RESPONDENTS (1<=N<=----)
                                                      P+4
 
                                                      8000
         BMD07S  (1)  N, NUMBER OF RESPONDENTS (1<=N<=----)
                                                      P+4

 
                                                      8000
         BMD08S  (1)  N, NUMBER OF RESPONDENTS (1<=N<=----)
                                                      P+4
 
         BMD09S  (1)  P, NUMBER OF ORIGINAL VARIABLES (1<=P<=99)
                 (2)  Q, NUMBER OF VARIABLES ADDED AFTER TRANSGENERATION
                      (-98<=Q<=98) (1<=P+Q<=99)
                 (3)  M, NUMBER OF VARIABLES DESIRED FOR PUNCHED CARD OUTPUT
                      (0<=M<=99)
                 (4)  T, TOTAL NUMBER OF TRANSGENERATION CARDS (0<=T<=99)
 
         BMD10S  (1)  R,C THE NUMBER OF ROWS AND COLUMNS IN THE ENTERING
                      MATRIX 1<=R<=10,000
                      1<=C<=10,000
                      1<=R*C<=10,000
                 (2)  F(1) AND F(2), THE NUMBER OF FORMAT CARDS FOR THE INPUT
                      AND OUTPUT FORMAT (1<=F(1)<=10), (1<=F(2)<=10)
                 (3)  T, THE NUMBER OF TAPE DRIVES TO BE USED IN THE TRANS-
                      POSITION (T=1)
                 (4)  R' , THE NUMBER OF ROWS GOVEREND BY THE INPUT FORMAT
                      (1<=R' <=10,000/C)
                 (5)  C' , THE NUMBER OF ROWS GOVERNED BY THE OUTPUT FORMAT
                      (1<=C' <=10,000/R)
 
         BMD01T  (1)  N, NUMBER OF POINTS IN THE SERIES (50<=N<=800)
                 (2)  M, NUMBER OF RESOLUTIONS (M<=350)
                 (3)  L, NUMBER OF POINTS WHERE THE FILTER IS EVALUATED (50<=L
                        L<=80)
                 (4)  K, NUMBER OF VARIABLE FORMAT CARDS (1<=K<=2)
 
         BMD02T  (1)  N, NUMBER OF DISCRETE DATA POINTS PER SERIES (1<=N<=100)
                 (2)  IN NO PROBLEM MAY THE PRODUCT OF N AND S (NXS) BE
                      GREATER THAN 2000
                 (3)  M, NUMBER OF LOGS (1<=M<=29)
                 (4)  Q, NUMBER OF SELECTION CARDS (1<=G<=99)
                 (5)  K, NUMBER OF VARIABLE FORMAT CARDS (1<=K<=2)
 
         BMD01V  (1)  N(I), THE NUMBER OF OBSERVATIONS OR CASES FOR THE I-TH
                      TREATMENT GROUP OR CATEGORY (1<=N(I)<=9000)
                 (2)  N, THE TOTAL NUMBER OF OBSERVATIONS IN ALL GROUPS OR
                      CATEGORIES COMBINED N=SUM(N(I))  (N<=45,000,000)
 
         BMD02V  (1)  L(I), NUMBER OF CATEGORIES OR LEVELS OF ANY ONE VARI-
                      ABLE (L(I)<=999) AND L(1) X L(2) X L(3) X ... X L(W)<=
                      7000)
                 (2)  K, NUMBER OF VARIABLE FORMAT CARDS (1<=K<=2)
 
         BMD03V  (1)  L(I), NUMBER OF CATEGORIES OR LEVELS OF ANY ONE ANALYSIS
                      OF VARIANCE CLASSIFICATION (L(I)<=500), (L(1) X L(2) X
                      L(3) X ... X L(W)<=500)
 
         BMD05V  (1)  P, NUMBER OF VARIABLES USED TO SPECIFY ANALYSIS OF
                      VARIANCE DESIGN (1<=P<=40)
                 (2)  Q, NUMBER OF COVARIATES (1<=P+Q<=40)
                 (3)  M, NUMBER OF TRANSGENERATION CARDS (0<=M<=40)
 
         BMD06V  (1)  M, NUMBER OF TRANSGENERATION CARDS (0<=M<=60)
 
         BMD08V  (1)  C, NUMBER OF COMPONENTS IN THE ANALYSIS OF VARIANCE
                      TABLE (C<=100)
                 (2)  Q, NUMBER OF DEPENDENT VARIABLES
                 (3)  M, NUMBER OF MEANS WHICH MUST BE STORED.  THIS MAY BE
                      AS LARGE AS HALF THE MAOUNT OF DATA, DEPENDING ON THE
                      DESIGN AND INCLUDES AT LEAST ALL THE MEANS WHICH ARE TO
                      BE PRINTED.  CORE IS TAKEN TO STORE AS FEW MEANS AS
                      POSSIBLE.  (C X Q + M<=5000)
 
         BMDX63  WITH P INDPEPNDENT VARIABLES AND Q DEPENDENT VARIABLES THE
                 FOLLOWING RESTRICTION MUST BE SATISFIED FOR EACH HYPOTHESIS
                 BEING TESTED. (P+Q)*(P+Q)+[R,Q]P+[P,Q]R+[R,S]Q+QS<=5000--WHERE
                 R IS THE NUMBER OF ROWS IN A, S IS THE NUMBER OF ROWS IN C,
                 AND [X,Y] DENOTES THE LARGER OF X AND Y.
 
         BMDX69  CP(P+1)/2+M<=5000 WHERE
                 C = NUMBER OF COMPONENTS IN THE ANALYSIS OF VARIANCE PART OF
                 MODEL
                 P = TOTAL NUMBER OF VARIABLES
                 M = NUMBER OF MEANS WHICH MUST BE STORED.  THIS MAY BE AS
                 LARGE AS HALF THE AMOUNT OF DATA, DEPENDING ON THE DESIGN
                 AND INCLUDES AT LEAST THE MEANS TO BE PRINTED.
 
         BMDX70  (1)  P, NUMBER OF ORIGINAL VARIABLES (1<=P<=199)
                 (2)  (NOTE:  ONLY VARIABLES 1-70 CAN BE ANALYZED; THE REMAIN-
                      ING VARIABLES ARE AVAILABLE FOR TRANSGENERATION.)
                 (3)  Q, NUMBER OF VARIABLES ADDED TO THE ORIGINAL SET AFTER
                      TRANSGENERATION (-198<=Q<=69)
                 (4)  P+Q, TOTAL VARIABLES OUTPUT (1<=P+Q<=70)
                 (5)  M, NUMBER OF TRANSGENERATION CARDS (0<=M<=70)
                 (6)  D, NUMBER OF MISSING VALUE CARDS (0<=D<=70)
                 (7)  S, NUMBER OF SUBPROBLEMS (1<=S<=15)
 
         BMDX72  (1)  P(P+1)/2+H<=6700
                 (2)  RP+H+R*R<=6700
                      WHERE P = NUMBER OF VARIABLES
                            R = NUMBER OF FACTORS ROTATED
                            H = 10P + 18F
                            F = NUMBER OF VARIABLE FORMAT CARDS
 
         BMDX77  (1)  AN END OF FILE CARD (EOF) MUST BE INSERTED BETWEEN THE
                      USER SUPPLIED FORTRAN STATEMENT CARD AND THE PROBLM CARD.
 
         BMDX75  (1)  TOTAL NUMBER OF VARIABLES (T<=49)
 
         BMDX82  (1)  APPROXIMATELY HALF OF THE TABLED VALUES FOR T AND G GIVEN
                      IN THE BIOMED MANUAL.
 
         BMDX84  (1)  G<=12000
                      WHERE G = 3P+4N(C) IF MISSING DATA ARE NOT SPECIFIED
                            G = 3P+10N(C) IF MISSING DATA ARE SPECIFIED
                            P = NUMBER OF VARIABLES AFTER TRANSGENERATION
                            N = NUMBER OF COLUMNS PER SUBPROBLEM
 
         BMDX85  (1)  TOTAL AMOUNT OF DATA N(T+1)<=10000
                      WHERE N = NUMBER OF CASES
                            T = NUMBER OF INDEPENDENT VARIABLES
                 (2)  AN END OF FILE CARD (EOF) MUST BE INSERTED BETWEEN THE
                      LAST USER SUPPLIED FORTRAN STATEMENT CARD AND THE PROBLM
                      CARD.
 
         BMDX92  (1)  TOTAL AMOUNT OF SPACE (W<=5000)
 
         BMDX93  (1)  2.0*N*NV*NP<=10000
                      WHERE N = NUMBER OF SCANS; NV = NUMBER OF VARIABLES ANA-
                      LYZED;
                            NP = NUMBER OF PROBLEMS PROCESSING SIMULTANEOUSLY;
                                 IF THE METHOD OF ANALYSIS IS SUBTRACTION OF
                                 PAIRED PULSES, USE: 3.0*N*NV*NP<=10000