Computer Centre Bulletin, Volume 5 Number 8&9, 4 September 1972

User Manual: Computer Centre Bulletin, Volume 5 Number 8&9, 4 September 1972

Open the PDF directly: View PDF .
Page Count: 22

UNIVERSITY OF QUEENSLAND

COMPUTER

CENTRE

COMPUTER

CENTRE

BULLETIN

Vol.

5

Nos

8 & 9 Editor:

4 September

1972

Sarah Barry

(

ADDITIONAL

CORE

MEMORY

FOR

THE

PDP-10

B5-8

&

~

26Aug72

In

its

submission

to

the

Australian

Universities

commission

for

caf'ital

tunds

for

com~uting,

the

University

nominated

additional

core

storage

as

the

item

of

highest

vriority.

In

anticiJ:.'ation

of

sU.i:>i-J0rt

from

the

AUC,

and

to

ensure

the

earliest

~ossible

installation

should

funds

become

available

early

if!

1973,

the

Centre

has

~laced

a

Letter

of

Intent

for

the

.t-'urchase

of

64K

words

of

650

nsec

core

memory.

Although

confirmation

or

othervase

of

this

intention

will

be

known

by

::lk:J

Set-t.ember,

we

are

advising

clients

of

this

intention

now

as

the

~i~ely

installation

of

significantly

more

core

may

affect

clients'

forwara

~~annin~.

une

outcor.,e

would

be

imiJroved

terminal

rest-0nse

and

batch

turn

aroun~

t.imes.

At

t.he

moment

the

PuP-lie:)

is

running

at

its

limit

of

24

active

jODS

tor

the

greater

l--art

of

its

time.

The

main

advanta~e

of

more

core

will

be

that

more

jobs

will

reside

in

cere,

thereby

redUCing

swa~f,;ing

overheads.

It

will

als0

Le

):-ossible

to

increase

the

user's

core

limit,

but

if

t.ni

,j

were

increased

substantially,

the

number

of

core

resident

jOLS

would

Qecrease

and

swa):-~ing

overheads

would

again

increase

•

...

01

this

rea~un,

should

the

memory

be

installed,

the

maximum

li~ely

increase

in

the

user's

core

limit

will

only

be

to

32K.

DATA

PREPA:{ATION

WORK

Witb

the

recent

reoryanization

of

the

COIntluter

centre's

Secretary's

ortice,

many

user-s

are

confused

as

to

where

data

):-Iel--aration

worK

is

to

be

hanaed

in

or

collected.

All

data

.t-'ref--aration

work

should

be

collected

from,

or

handed

into,

the

Receil--t/Dis):-atch

window

in

the

foyer

of

the

centre,

not

at

the

Enquiries

window.

Data

fJret->aration

forms

are

available

in

the

Clients'

room.

87

l.:.5-8 &

'9

2G

i-;w-j

12

Pl.iP-llll

FORTRA~

1

DEFI~ITION

OF

ARITHMETIC STATEMENT FUNCTIONS

[WN-95]

'l'he Comi:Juter

~ent.re

FOk'l'RAN

manual

MNT-5

does

not

state

that

the

definition

of

arithmetic

statement

functions

should

i:-'recede

the

first

executable

statement

in

the

~rogram.

This

omission

will

be

corrected

with

the

next

revision

of

the

manual.

[WN-96]

2

DO

LOOPS

The

pr.esent

version

of

the

comf;iler

does

not

create

proper

code

when

a

function

is

used

as

one

of

a

DO

100i:J's

indices.

example:

DO

2830

J=l

,MIN~(b,NN)

~o

diagnostic

message

is

i:-'roduced

during

compilation

but

an

improper

relocatable

binary

file

is

produced,

and

during

loading

the

message

'!

ILL.

l'

ORMAT

BLOCK

TYPE

NNN

PROG.

name

1

FILE

name

~/REL

will

ue

procuced.

This

error

has

been

remedied

in

a

later

version

of

the

compiler

but

until

this

version

is

adopted,

this

construction

sbould

be

avoided.

3

UNARY

MINUSES

Use

of

the

unary

minuses

in

logical

eXi:Jressions

or

mixed

logical

and

arithmetic

eXi:Jressions

do

not

always

produce

the

correct

results,

tOL

example,

J = ( J 1 • f'.N D

...

777

) • OR. ( ( - ( N + 1 ) ) •

AN

lJ

...

7 7 7 0

010

)

J=(Jl.ANlJ.

"777)+(

(-(1'-<+1))

.ANlJ."77700flJ)

'l'his

has

been

corrected

in

a

later

version

of

the

compiler

which

will

be

imi:-'lemented

in

due

course.

Until

it

is

available,

it

is

suggested

that

this

construction

should

not

be

used

and

that

the

expression

be

split

into

two

as;

JJ=-(N+l

)

J = ( J 1 •

AN

D

...

77

7 ) •

OR

• ( J J •

AN

D

...

7 7 7 0

flJ

0 )

88

(

PDPUN

B5-8

& 9

26Aug72

[WN-99]

~DPUN

may

not

correctly

~unch

the

identification

code

of

other

than

the

first

of

a

series

of

Fortran

files

with

supplied

identification.

T'his

f..'roblem

will

be

corrected

in

the

near

future.

ALGOL

V2

[WN-94]

Version

2

of

the

Algol

system

does

not

allow

the

use

of

external

FORTRAN

sUb~rograms.

This

feature

is

discussed

in

various

manuals,

and

information

from

Digital

indicates

that

this

facility

will

be

available

with

version

3.

COBOL

V3

ERROR

[WN-94]

In

use

of

the

sort

features

of

Cobol,

it

af.Jpears

that

under

some

circumstanced

wrong

code

is

created

by

a

Return

statement

of

the

form

RETURN

filename

INTO

identifier

•

':"his

can

be

uvercome

by

omitting

the

INTO

option,

but

achieving

the

same

effect

by

moving

the

data

to

the

required

destination

after

the

keturn.

OUTSTANDING

PROBLEMS

IN

COBOL

V3

[WN-96]

The

following

note,

J:-,re~ared

on

3

July

1972,

gives

a

summary

of

known

~roblem

areas

with

Cobol.

The

large

majority

of

these

are

corrected

by

patches

issued

by

Digital

and

these

will

be

89

BS-8

& 9

.::.6Aug72

incor~orated

as

effort

becomes

available.

THE

COBOL

COMPILER

(a)

'l'his

version

does

not

allow

qualification

of

condition

names

(to

be

corrected

with

version

3A).

(b)

"RETURN

namel

INTO

name2"

as

part

of

a

sort

does

not

work

(patch

exists).

(c)

Comparison

for

equality

of

6-character

display-6

fields

may

not

give

correct

results

(patch

exists).

(d)

Some

external

references

from

non-resident

segments

are

not

set

u~

pro~erly.

The

reported

situations

were

SORT

verb

in

resident

code

RETURN

and/or

RELEASE

in

non-resident

TALLY

in

examine

generator

TOD1,Y

in

the

USING

generator

Size

error

and

display

of

a COMP-l

variable

(patches

exist).

(e)

"MULTIPLY

namel

BY

fraction

GIVING

name2"

where

fraction

is

less

than

1.0,

e.g.

0.0125;

will

return

the

value

zero

to

name2

in

some

Circumstances,

depending

on

the

typing

of

namel

and

name2

(patch

exists).

(f)

A

~,.l.meric

literal

in

a "VALUE

OF

DATE

WRITTEN"

clause

fails

(patch

exists).

(g)

The

com~iler

allows

a

quoted

literal

for

a PROGRAM-ID,

but

if

the

literal

is

less

than

6

characters,

the

remainder

of

the

name

is

random

(patch

exists).

(h)

The

COlh}iiler

does

not

successfully

restart

catastrOf\he

dump

if

it

was

using

a

command

file.

after

a

(i)

(j)

(k)

l'he

compiler

will

not

accept

lower

case

for

character

of

a

reserved

word,

although

lower

case

are

sat;isfactory

in

all

other

l;Jositions

(a

patch

the

first

character

to

be

lower

case

exists).

the

first

characters

to

allow

Under

seme

Circumstances,

a

spurious

warning

message

"REDEFINITON

NOT

THE

SAME

SIZE

AS

REDEFINED ITEM" may

be

given.

This

can

occur

when

the

two

fields

occupy

an

integral

number

of

words

(patch

exists).

Under

some

Circumstances,

Cobol

comfiiler

tables

expanded

correctly.

This

error

may

manifest

compilation

or

execution

error

for

which

no

90

are

not

itself

as

simfile

---,

~

-

-----

~

--'----,

B5-8

& 9

26Aug72

ex~lanation

may

be

found.

A

patch

e~ists

for

one

situation

such

as

this~

but

at

least

one

other

may

exist.

(1)

"PP,RITY

IS

EVEN"

does

not

compile

properly

(patch

exists)

~

(m)

Improper

line

numbers

are

given

for

the

procedure

division

map

(patch

exists).

(n)

(0

)

2

(a)

(b)

(c)

(d)

(e

)

Compilation

using

the

IF

switch

will

produce

a

catastrophe

dump

if

..

TRACE

Ol'il/OFF"

is

used

(pa

tch~

exi

sts)

•

It

appears

as

if

construction

"OCCURS

nl

TO

n2

TIMES

DEPENDING

ON

namel

will

not

create

an

output

record

of

the

size

s~ecified

by

narnel

but

rather

the

maximum

size

n2.

'I'HE COBOL EXECUTION

PACKAGE

-

LIBOL

When a

random

file

is

opened

for

output

only,

any

attempt

to

write

to

it

will

terminate

the

run

with

an

error

message

"IT

IS

ILLf.GAL

TO

CHANGE

ThE

RECORD

SIZE

WHEN

THE

ACCESS

MODE

IS

RANDOM

(patch

exists).

When a

Cobol

~rogram

attempts

to

of>en

two

files

which

share

trIa

same

butter

area

theb"esultant

error

message

is

garbled

(pOl

tch

exists)

0

When

an

10

file

encounters

an

EOF%

a

data

location

is

not

reE,-=t,

,,.;rith

the

result

that

the

next

sequential

file

input

will

fail

(patch

exists).

Multi--file

magnetic

tapes

arE!

not

positioned

prof,.'erly

(patch

exists)

•

Some

advi:->urv

messc!ges

not

conform

to

the

specifications

for

uigitald~

~

series

monitors

(patch

exists).

(f)

10

access

to

an

indexed

sixbit

file

will

fail

at

the

first

read

(~atch

exists).

(g)

Code

created

for

"USE

AF'rER

ERROR"

procedure

is

incorrect

(pa

tch

exi

sts)

•

(h)

File

record

areas

are

improperly

cleared

circumstances

when

adjacent

files

are

opened.

under

some

(i)

Some

o~er

cheCKS

are

made

when

error

situations

arise,

tor~

examf.Jle

9

when

a f i

is

not

found

(some

patches

exi

st)

•

(j)

Under

some

circumstances

the

wrong

quantity

is

used

to

calculate

record

size

when

a

blocking

factor

is

specified.

91

BS-8

& 9

:O::6Aug'l

:.:.:

This

will

cause

an

undeserved

error

message

at

execution

(patch

exists).

(k)

o~en

immediately

after

close

on

magnetic

tape

fails

(~atch

exists)

•

(1)

Sort

of

an

indexed

file

may

fail

(patch

exists).

(m)

Some

problems

ASCII

key

circumstances

closed

(patch

exist

with

indexed

files,

e.g.

the

size

of

an

is

incorrectly

computed

and

under

some

the

channel

used

for

the

index

tile

is

not

exists)

•

(n)

under

some

circumstances,

there

may

be

an

interaction

between

sort

and

the

use

of

indexed

files

which

will

cause

the

program

to

fail

(patch

exists).

(0)

"NEX'r

GROUP

NEXT

PAGE"

may

cause

the

output

of

a

blank

,f)age

(patch

exists).

(f.')

An

EOF

is

found

when

attempting

to

read

an

indexed

file

sequentially

(patch

exists).

(q

).

A

sort

which

reads

a

file

created

by

the

same

program

may

get

an

error

(patch

exists).

(r)

When a

null

record

is

found

in

a

random

file,

subsequent

in,k-'uts

or

outputs

may

fail

(patch

exists).

3

THE

SORT

PACKAGE

(a)

Under

some

circumstances,

a command

file

is

not

accepted

(patch

exists).

'*

RERUN

Rerun

has

several

problems,

some

of

which

are

related

to

the

version

of

the

operating

system

currently

being

used

and

it

is

doubtful

if

satisfactory

results

will

be

achieved.

A

number

of

patches

exist

and

when

these

have

been

incorporated

consideration

will

be

given

to

further

testing.

5

ISAM

The

Isam

utility

has

some

problems

in

handling

ASCII

input

files

and

in

conversion

from

ASCII

to

sixbit

files

(patches

exist).

92

B5-8

& 9

(

26Aug72

6

GENERAL

b.l

Su~plement

to

Digital's

Cobol

Manual

Attention

is

drawn

to

a

note

in

Volume

XII

no

5

of

the

DECSYSTEM10

Bulletin

which

states

that

a

supplement

to

the

cobol

manual

is

now

available.

This

supplement

describes

ISAM,

COBDDT,

and

COBOL

table

handling

and

is

available

on

request

from

Digital.

6.2

Factors

Affecting

the

Size

of

a

Compilation

In

response

to

a

query

regarding

the

size

of

the

largest

program

that

can

be

com~iled

within

the

present

core

limits,

the

following

suggestions

were

offered

by

Digital.

(a)

Identifiers

longer

than

6

characters

in

length

require

additional

storage.

(b)

Condition

names

(level

88)

should

be

avoided

(c)

Keep

paragraph

names

(number

and

size)

to

a minimum

(d)

Do

not

use

values

as

data

items

(e)

Keep

literals

as

short

as

possible

(f)

Avoid

ALTERs

(g)

~o

not

use

the

report

writer

It

is

not

likely

that

great

advantage

can

be

taken

of

these

suggestions,

but

it

is

worth

noting

that

version

3

of

Cobol

has

fJroved

to

be

able

to

compile

substantially

longer

source

progr3ms

than

any

of

its

~redecessors.

b.3

Recoml-'ilation

of

Existing

programs

A

note

in

the

DECSYSTEM10

Bulleting

on

April

1

1972

indicates

the

next

release

of

Cobol

will

retain

the

existing

version

of

the

execution

time

package

for

those

programs

that

presently

use

it.

Programs

com~iled

with

later

versions

of

Cobol

will

compile

a

version

of

Libol

identified

by

its

version

number.

In

this

way.

eXisting

compiled

programs

will

continue

to

run

without

recompilation.

It

is,

however,

recommended

that

programs

~reviously

compiled

with

versions

earlier

than

version

3

be

recom};)iled.

93

B5-8

& 9

~6Aug72

NEW

COMMAND

DECODER

[WN-95 ]

A new

version

of

the

command

decoder

(version

2E(36)-3)

was

im~lemented

on

the

PDP-l

0

on

Wednesday

12

July.

A

number

of

important

changes

were

made

with

this

decoder.

The

changes

documented

in

section

1

and

2

below,

foreshadow

the

implementation

of

a

general

program

library

and

extended

command

capabili

ty.

Director

ie

s

Library

oirectories

on

the

PDP-10

were

renamed

and

in

some

cases

there

was

some

reorganization

of

the

actual

files

contained

on

these

areas.

Libr~ry

directories

are

referenced

by

name,

e.g.

PLOT, STATS,

MATRIX. The

names

are

not

preceded

by

a

dollar

sign;

a

dollar

sign

is

used

to

indicate

a

device

or

pseudo

device,

e.g.

$DSK,

$I>SR.

User

directories

are

referenced

by

the

project

number

of

the

area

belonging

to

that

user,

e.g.

279,

531.

Files

from

any

directory

can

be

obtained

by

referencing

the

directory

nan,e

and

the

filename

For

exan.t-'le:

(i)

RUN

l>'lYPROG

PLOT.CALCMP(LIB)

runs

the

file

MYPROG

(understood

to

be

on

the

user's

own

directory)

with

the

library

file

CALCMP

from

the

PLOT

directory.

(ii)

STATS.bMD~~R

executes

the

f-rogram

BMD02R

on

the

STATS

directory.

(iii)

COpy TUTOR.MYPROG/F4 TO=374.MYPROG/F4

cOt"ie£

the

FORTRAN

J:-'rogram

l-lYPROG

from

the

TUTOR

directory

onto

t"roject

area

J74.

on

the

12

July,

the

old

library

directories

($BMD,

$MATH,

$MATRIX, $LEARN, $PLOT, $STATS)

were

removed

and

reorganized

and

t.he

new

directories

are

as

follows:

.uirectory

File;:;

contents

GAMES

CliESS/Sr\V

.L.JOTS/SAV

94

hELP

N.h.TRIX

PLOT

STATS

TUTOR

ELIZA/SAV

LUNAR/BAS

MOO/SAV

NIM/SAV

QED/SAV

HELP/,aLP

MANUAL/HLP

NE1.<JS/

HLP

SSP/REL

MhTINV/REL

SSP/REL

SMIS/SAV

lJEC/REL

CALClv"IP/REL

DRIVH/SAV

FLOCT/SAV

FORGN/SAV

BMDIiJ1D/SAV

JJhD02D/SAV

BMDI08D/SAV

BMD01l:i/

St-'.

V

LMDkJ4f'!/

SJ-l.

V

BMDkJiM/SnV

l3MDkl2R/SAV

BMDI61V/SAV

B1"1004

V /

SA

V

SSP/REL

MYPROG/F4

MYSUBR/F4

B5-8

& 9

26Aug72

These

were

all

removed

from

the

system

directory

to

GAMES.

eX1Jlains

how

to

use

help

gives

details

of

current

status

computer

centre

manuals

current

newsletter

a

library

file

containing

all

mathematical

routines

from

scientific

subroutine

package

matrix

inversion

of

the

the

a

library

file

of

matrix

subroutines

from

the

scientific

subroutine

package

Symbolic

l-1atrix

Interpretive

System

old

Digital

~lotter

routines

basic

calComp

software

routines

pr

0

read

cards

for

CRVPT

owchart

Generation

utility

program

Forms

Generation

program

Sim,pIe

Data

De

ser

iption

cor.re

la

tion

wi

th

Transgenera

tioD

cross

Tabulation

with

Variable

stacking

principal

Com~onent

Analysis

Discrj.minant

Analysis

for

Two

Groups

Stepwise

Discriminant

Analysis

Stepwise

Regression

Analysis

of

Variance

for

one

way

Design

AnalYSis

of

covariance

with

Multiple

coordinates

[Note

that

these

programs

are

now

referenced

by

their

complete

names]

statistical

routines

from

the

scientific

£HJ

br

ou

ti

ne

s

package

a

program

and

subroutine

teaching

DDT

to

users

used

in

Initial

documentation

on

the

library

.l':'outines

w:i.,ll

be

available

in

the

clients'

room

y

for

reference

only.

95

B5-8

& 9

~6Aug72

6.

Numeric

!"ilenames

rilenames

can

consist

of

6

alphanumeric

characters.

The

restriction

that

the

first

character

had

to

be

alphabetic

has

been

removed.

The

first

character

of

the

processor

program

name

may

also

be

numeric.

The

editor

will

presently

not

recognize

all

numeric

filenames

or

all

numeric

processor

program

names.

3 The

DIRECTORY

command

A

DIRECTORY

command

in

which

no

argument

is

specified

assumes

a

default

argument

of

ALL/ALL.

That

is

DIR

and

DIR ALL/ALL

both

list

the

user's

complete

directory.

With

the

implementation

of

the

changes

detailed

in

section

1

and

2

above,

the

default

of

ALL/ALL

for

filename

where

none

is

s~ecified,

continues

to

be

assumed

with

the

following

results.

(i)

DIR

PLOT

(ii)

DIR PLOT.CALCMP

(iii)

VIR

PLOT.

(iV)

DIR

379

(v)

DIR

379.TEST/F4

(vi)

DIR

379.

4 IDENT Command

lists

the

directory

entry

for

the

file

PLOT

on

the

user's

area

lists

the

directory

entry

for

the

file

CALCMP

on

the

PLOT

library

assumes

ALL/ALL

for

the

filename

and

lists

the

complete

directory

of

the

PLOT

library

lists

the

directory

entry

of

the

file

379

on

the

user's

area

lists

the

directory

entry

of

the

file

TEST/F4

on

project

379

assumes

the

default

option

for

filename

and

lists

the

directory

of

project

379

The

decoder

will

now

recognize

the

correct

spelling

of

'exercise'

for

the

argument

EXERCISE=

exercise-number.

EX

is

an

allowable

abbreviation

for

the

assignment.

96

(

ALGOL

command

B5-8

& 9

26Aug72

An

ALGOL

command,

similar

to

the

FORTRAN,

COBOL

and

MACRO

commands,

has

been

implemented.

The

only

options

available

are

LIST

and

NOLIST, BIN

and

NOBIN.

ALGOL(LIST

,BIN

)

NOLIST

NOBIN

6

OVERLAY

fIN=!filename-l

fBIN=!

f1lename-2

lLST=}f

ilename-3

An

OVERLAY(0) command

will

now

clear

any

files

on

the

loadlist

and

commence

a new

list

for

the

overlay

command

and

those

that

follow.

7

Error

Messages

uecode

now

sives

the

correct

error

messages

in

all

situations.

SCIENTIFIC

SUBROUTINE

PACKAGE

[WN-98]

The

SCi€Htific

Subroutine

package

comfJrises

a

set

of

over

250

Fortran

subroutines

covering

many

areas

of

mathematics

and

statistics.

The

following

pages

list

the

name

and

a

very

brief

description

of

each

of

the

routines.

As

indicated,

these

routines

are

now

available

in

a

number

of

library

files

on

the

PDP-10

system

in

the

tollowing

directories.

(a)

STATS

directory

The

library

file

SSP/REL

contains

all

the

statistical

routines

from

the

ssp

fJackage.

Any

of

these

can

be

loaded

wi

th

a

user's

fJrogram

by

a command

of

the

form

HUN

prog-name

STATS.SSP(LIB)

(b)

MATRIX

directory

Again,

the

library

file

SSP/REL

on

the

MATRIX

directory

contains

.all

the

matrix

routines

of

the

SSP

f.iackage.

Any

of

these

can

be

loaded

with

a

user's

program

by

a command

of

the

form

97

.05-8

& 9

",6Aug72

RUN

~rog-name

MATRIX.SSP(LIB)

(c)

MATH

directory

There

are

three

library

files

of

SSP

routines

on

the

MATH

directory.

(i)

SSPP/REL

contains

polynomial

routines

(ii)

SSPF/REL

contains

other

mathematical

functions

(iii)

SSPM/REL

contains

miscellaneous

routines.

The

listing

of

the

routines

gives

the

particular

library

file

in

which

any

given

routine

is

to

be

found.

The

user

can

load

a

routine

from

any

library

with

an

ap.kJro,l:

...

riate

RUN

command.

For

example

~

if

a

program

called

routines

from

both

the

library

files

SSPP/REL

and

SSPM/REL

the

command

would

be

HUN

~rog-name

MAThoSSPP(LIB)

~~TH.SSPM(LIB)

The

scientific

subroutine

package

is

classified

as

type

4

software

and

is

therefore

made

available

on

an

'as

is'

basis.

Interim

documentation

is

now

available

in

the

clients'

room

at

tIle

centre

for

reference

only

@

98

(

+

CA

TEGORIAL GIDDE TO SUBROUTINES AND

SAMPLE

PROGRAMS

(Subroutines

added

in

Version

III

are

marked

with

an

asterisk)

STATISTICS

Data

Screening

in

STATS.

SSP

(LIB)

on

PDP-IO

T

ALLY

--totals,

means,

standard

deviations,

minimums,

and

maximums

BOUND--selection

of

observations

within

bounds

SUBST--subset

selection

from

observation

matrix

ABSNT--detection

of

missing

data

27

27

28

28

TAB1--tabulation

of

data

(one

variable)

29

TAB2--tabulation

of

data

(two

variables)

30

SUBMX--building

of

subset

matrix

31

Correlation

and

Regression

(See

Smoothing,

Factorization)

in

STATS.

SSP

(LIB)

on

PDP-IO

CORRE--means,.

standard

deviations,

and

corrdations

*MISR--means,

standard

deviations,

third

and

fourth

moments,

correlations,

simple

regression

coefficients

and

their

standard

errors;

considers

that

data

may

be

missing

ORDER--rearrangement

of

intercorrela-

tions

MULTR--multiple

linear

regression

GDA

TA--data

matrix

generation

for

polynomial

regression

*STPRG--stepwise

multiple

linear

regression

*PROBT--probit

analysis

CANOR--canonical

correlation

32

33

36

37

39

41

44

47

B5-B & 9

26Aug72

Design

Analysis

(See

Smoothing,

Regression,

Factorization)

in

STATS.

SSP

(LIB)

on

PDP-IO

AVDA

T--data

storage

allocation

AVCAL--L

and

~operation

MEANQ--mean

square

operation

Discriminant

Analysis

in

STATS.

SSP

(LIB)

on

PDP-IO.

DMA

TX--means

and

dispersion

matrix

DISCR--discriminant

functions

Factor

Analysis

(See

Eigenanaly.sis)

in

STATS.

SSP

(LIB)

on

PDP-IO

TRACE--cumulative

percentage

of

.

eigenvalues

LOAD--factor

loading

V ARMX

--val'

imax

rotation

Time

Series

(See Smoothing)

in

STATS.

SSP

(LIB)

on

PDP-IO.

A

UTO--autocovariances

CROSS--cross

covariances

SMO--application

of

filter

coefficients

(weights)

EXSMO--triple

exponential

smoothing

Nonparametric

Statistics

in

STATS.SSP(LIB)

on

PDP-lao

*KOLMO--Kolmogorov-Smirnov

one-

sample

test

*KOLM2--Kolmogorov-Smirnov

two-

sample

test

*SMIRN--Kolmogorov-Smirnov

limiting

distribution

values

CHISQ--

X 2

test

for

contingency

tables

KRANK--Kendall

rank

correlation

*MPAIR--

Wilcoxin I s

signed

ranks

test

+

THIS

SUBROUTINE

ALSO

LOCATED

IN

MATRIX.SSP(LIB)

ON

PDP-IO.

99

49

50

57

52

53

55

56

56

59

60

61

62

63

65

66

68

69

70

B5~8

& 9

26Aug72

QTEST--Cochran

Q-test

RANK--rank

observations

*SIGNT--sign

test

SRANK--Spearman

rank

correlation

TIE--calculation

of

ties

in

ranked

obser-

vations

TWOAV--Friedman

two-way

analysis

of

variance

statistic

UTEST--Mann-Whitney

U-test

WTEST--Kendall

coefficient

of

concordance

Generation

of

Random

Variates

-

in

STAT

S .

Distribution

Functions

SSP

(LIB)

on

PDP-IO.

RANDU--uniform

random

deviates

GA

USS-

-normal

deviates

71

71

72

73

74

74

75

76

77

77

*NDTR--normal

distribution

function

78

*BDTR--beta

distribution

function

*CDTR--X

2

distribution

function

*NDTRI--inverse

of

normal

distribution

function

78

81

83

RCPY--copy

row

of

matrix

into

vector

CCPY

--copy

column

of

matrix

into

vector

DCPY

--copy

diagonal

of

matrix

into

vector

XCPY

--copy

submatrix

from

given

matrix

MSTR--storage

conversion

LOC~-location

in

compressed-stored

matrix

94

95

95

96

96

97

C

ONVT--single-precision/

double-precision

97

conversion

ARRA Y

--vector

storage/double-dimensioned

98

storage

conversion

Matrices:

Operations

in

MATRIX. SSP

(LIB)

on

PDP-IO.

GMADD--add

two

general

matrices

98

GMSUB-':'subtract two

general

matrices

99

GMPRD--product

of

two

general

matrices

99

GMTRA--transpose

of

a

general

matrix

100

GTPRD--transpose

product

of

two

general

matrices

MADD--add

two

matrices

100

101

MSUB--subtract

two

matrices

101

Elementary

Statistics

and

Miscellany

in

STATS. SSP

(LIB)

on

PDP.,..IO.

MPRD--matrix

product

(row

into

column)

102

MOMEN--first

four

moments

85

TTEST--test

on

population

means

86

*BISER--biserial

correlation

coefficient

87

*PHI--phi

coefficient

*POINT--point-biserial

correlation

coefficient

*TETRA--tetrachoric

correlation

coefficient

88

89

90

*SRA

TE--survival

rates

92

Matrices:

storage

in

MATRIX. SSP

(LIB)

on

PDP"",IO.

MCPY--matrix

copy

94

100

MTRA--transpose

a

matrix

102

TPRD--transpose

product

MATA--transpose

product

of

matrix

by

itself

SADD--add

scalar

to

matrix

103

103

104

SSUB--subtract

scalar

from

a

matrix

104

SMPY

--matrix

multiplied

by

a

scalar

105

SDIV--matrix

divided

by

a

scalar

105

SCLA--matrix

clear

and add

scalar

106

DCLA--replace

diagonal

with

scalar

106

(

RADD--add

row

of

one

matrix

to

row

of

another

matrix

107

CADD--add

column

of one

matrix

to

column

107

of

another

matrix

SRMA--scalar

multiply

row

and

add

to

another

row

108

SCMA--scalar

multiply

column

and

add

to

108

another

column

.

RINT

--interchange

two

rows

109

CINT--interchange

two

columns

109

RSUM--sum

the

rows

of

a

matrix

110

CSUM--sum

the

columns

of

a

matrix

110

RTAB--tabulate

the

rows

of

a

matrix

III

CTAB--tabulate

the

columns

of

a

matrix

112

RSRT--sort

matrix

rows

CSRT--sort

matrix

columns

RCUT--partition

by

row

CCUT--partition

by

column

RTIE--adjoin

two

matrices

by

row

CTIE--adjoin

two

matrices

by

column

*MPRC,

DMPRC--permute

rows

or

columns

MFUN--matrix

transformation

by

a

function

REC

P--reciprocal

function

for

MFUN

Matrices:

Inversion,

Systems

of

Linear

Equations

and

Related

Topics

in

MATRIX.SSP(LIB)

on

PDP-lO.

+ MINV

--matrix

inversion

*SINV, DSINV -

-invert

a

symmetric

positive·

definite

matrix

SIMQ--solution

of

simultaneous

linear,

algebraic

equations

++

GELG,

DGELG--system

of

general

simul-

taneous

linear

equations

by

Gauss

elimination

+

ALSO

IN

STATS.SSP(LIB)

on

PDP-lO

++

ALSO

IN

MATH.

SSPM(LIB)

on

PDP-lO

112

113

113

114

114

115

115

117

117

118

119

120

121

I

~I

B5~d

& 9

26Aug72

*RSLMC--solution

of

simultaneous

linear

equations

with

iterative

refine-

ment

*FACTR--triangular

factorization

of

a

nonsingular

matrix

MFGR,

DMFGR--matrix

factorization

and

rank

determination

GELS,

DGELS--system

of

general

simul-

taneous

linear

equations

with

symmetric

coeffi-

cients

GELB,

DGELl;l--system

of

general

simul-

taneous

linear

equations

with

band-structured

coefficients

*MTDS,

DMTDS--divide

a

matrix

by

a

triangular

matrix

124

126

127

133

137

142

*MLSS,

DMLSS--solution

of

simultaneous

145

linear

equations

with

symmetric

positive

semidefinite

matrix

*MCHB,

DMCHB--triangular

factorization

148

of a

symmetric

positive

definite

band

matrix

*MFSS,

DMFSS--triangular

factorization

152

and

rank

determination

of

a

symmetric

positive

semidefinite

matrix

*MFSD,

DMFSD--triangular

factorization

158

of

a

symmetric

positive

definite

matrix

LLSQ,

DLLSQ--solution

of

linear

least-

160

squares

problems

Matrices:

Eigenanalysis

and

Related

Topics

in

MATRIX.SSP(LIB)

on

PDP-lO.

+

EIGEN--eigenvalues

and

eigenvectors

of

a 164

real,

symmetric

matrix

+

NROOT--eigenvalues

and

eigenvectors

of

a 166

speCial

nonsymmetric

matrix

*ATEIG--eigenvalues

of

a

real

almost

triangular

matrix

167

*HSBG--reduction

of

a

real

matrix

to

almost

triangular

form

169

BS-8

& 9

..::6Aug72

Polynomials:

Operations

in

MATH.

SSPP

(LIB)

on

PDP-IO.

PADD--add

two

polynomials

PSUB--subtract

one

polynomIal

from

another

PMPY--multiply

two

polynomials

PDIV

--divide

one

polynomial

by

another

PCLA--replace

one

polynomial

by

another

PADDM-~multiply

polynomial

by

constant

and

add

to

another

polynomial

PV

AL--value

of

a

polynomial

PVSUB--substitute

variable

of

polynomial

by

another

polynomial

PILD--evaluate

polynomial

and

its

first

derivative

PDER--derivative

of

a

polymonial

PINT--integral

of

a

polynomial

PQSD--quadratic

synthetic

division

of

a

polynomial

171

171

172

172

173

173

174

174

175

175

176

176

PCLD--complete

linear

synthetic

division

177

PGCD--greatest

common

divisor

of

two

polynomials

PNORM--normalize

coefficient

vector

of

polynomial

PECN,

DPECN--economization

of

a

poly-

nomial

for

symmetric

range

PECS,

DPECS--economization

of

a

poly-

nomial

for

unsymmetric

range

Polynomials:

Roots

in

MATH.

SSPM(LIB)

on

PDP-IO

177

178

178

180

+

POLRT--real

and

complex

roots

of

a

real

181

polynomial

PRQD,

DPRQD--roots

of

a

real

polynomial

183

by

QD

algorithm

with

dis-

placement

+

ALSO

IN

STATS.SSP(LIB)

on

PDP-IO

*PRBM,

DPRBM--roots

of

a

real

poly-

nomial

by

Bairstow's

algorithm

*PQFB,

DPQFB--determine

a

quadratic

factor

of

'a·

real

poly-

nomial

189

193

Polynomials:

Special

Types

in

MATH.

SSPP

(LIB)

on

PDP-IO.

++

CNP,

DCNP--value

of

Nth

Chebyshev

pol~omial

++

CNPS,

DCNPS--value

of

series

expansion

in

Chebyshev

polynomials

198

199

TCNP,

DTCNP--transform

series

expansion

200

in

Chebyshev

polynomials

to

a

polynomial

CSP,

DCSP--value

of Nth

shifted

Chebyshev

201

polynomial

CSPS,

DCSPS--value

of

series

expansionin

202

shifted

Chebyshev

poly-

nomials

TCSP,

DTCSP--transform

series

expansion

203

in

shifted

Chebyshev

poly-

nomials

to

a

polynomial

HEP,

DHEP--value

of

Hermite

polynomial

205

HEPS,

DHEPS--value

of

series

expansion

in

206

Hermite

polynomials

THEP,

DTHEp--transform

series

expan-

207

sion

in

Hermite

polynomials

to

a

polynomial

LAP,

DLAP--value

of

Laguerre

polynomial

208

LAPS,

DLAPS--value

of

series

expansion

in

Laguerre

polynomials

209

TLAP,

DTLAp--transform

series

expan-

210

sion

in

Laguerre

poly-

nomials

to

a

polynomial

LEP,

DLEP--value

of

Legendre

polynomial

212

LEPS,

DLEPS--value

of

series

expansion

in

Legendre

polynomials

213

++

THIS

SUBROUTINE

in

MATH.

SSPM(LIB)

on

PDP-IO.

102

(

TLEP,

DTLEp--transform

a

series

expan-

214

sion

in

Legendre

poly-

nomials

to

a

polynomial

Roots

of

Nonlinear

Equations

in

MATH.

SSPM(LIB)

on

PDP-IO.

RTWI,

DRTWI--refine

estimate

of

root

by

215

Wegstein's

iteration

RTMI,

DRTMI--determine

root

within

a 217

range

by

Mueller's

iteration

RTNI,

DRTNI--refine

estimate

of

root

by

220

Newton's

iteration

Extremum

of

Functions

in

YlATH·

SSPM(LIB)

on

PDP-IO

FMFP,

DFMFP--unconstrained

minimum

221

of

a

function

of

several

variables--

Davidon

method

FMCG,

DFMCG--unconstrained

minimum

225

of

a

function

of

several

variables

--

conjugate

.

gradiant

method

Permutations

in

MATH.

SSPM(LIB)

on

PDP-IO.

*PPRCN--composition

of

permutations

231

*PERM--operations

with

permutations

and

transpositions

232

Sequences:

Sums

and

Limits

in

MATH.

SSPM(LIB)

on

PDP-IO.

TEAS,

DTEAS--limit

of

a

given

sequence

TEUL,

DTEUL--sum

of

a

given

function

sequence

Interpolation

t

Approximation,

and

234

238

Smoothing

in

MATH·

SSPH(LIB)

on

PDP-IO.

ALI,

DALI--Aitken-Lagrange

interpolation

241

AID,

DAID--Aitken-Hermite

interpolation

243

ACFI,

DACFI--continued

fraction

inter-

polation

ATSG,

DATSG--table

selection

out

of

a

general

table

ATSM,

DATSM--table

selection

out

of

a

monotonic

table

245

248

250

1103

B5-8

&.

9

26Aug72

ATSE,

DATSE--table

selection

out

of

an

equidistant

table

*SG13,

DSG13--local

least-squares

smoothing

of

tabulated

functions

*SE13, DSE13

*SE15, DSE15

*SE35,

DSE35--local

least-squares

smoothing

of

eqUidistantly

tabulated

functions

*APFS,

DAPFS--solve

normal

equations

for

least-squares

fit

*APCH,

DAPCH--Ieast-squares

polynomial

approximation

*ARA

T,

DARA T

*FRA

T,

DFRA

T--rational

least-squares

approximation

*APLL,

DAPLL--linear

least-squares

approximation

251

253

255

260

263

265

271

FORIF--Fourier

analysis

of

a

given

function

274

FORIT--

Fourier

analysis

of

a

tabulated

function

275

HARM,

DHARM--complex

three-dimensional

276

Fourier

analysis

RHARM,

DRHARM--realone-dimensional

Fourier

analysis

*APMM,

--linear

Chebyshev

approximation

over

a

discrete

range

281

283

Numerical

Quadrature

in

MATH.

SSPF

(LIB)

on

PDP-IO.

QTFG,

DQTFG--integration

of

monotonically

289

tabulated

function

by

trapezoidal

rule

QTFE,

DQTFE--integration

of

equidistantly

290

tabulated

function

by

trapezoidal

rule

QSF,

DQSF-)...integration

of

equidistantly

291

tabulated

function

by

Simpson's

rule

QHFG,

DQHFG--integration

of

monotonically

293

tabulated

function

with

first

derivative

by

Hermitian

formula

of

first

order

B5-8

&.

9

~6Aug72

QHFE,

DQHFE--integration

of

equidistantly

294

tabulated

function

with

first

derivative

by

Hermitian

formula

of

first

order

QHSG,

DQHSG--integration

of

monotonically

295

tabulated

function

with

first

and

second

derivatives

by

Hermitian

formula

of

first

order

QHSE,

DQHSE--integration

of

equidistantly

296

tabulated

function

with

first

and

second

derivatives

by

Hermitian

formula

of

second

order

QATR,

DQATR--integration

of

a

given

297

function

by

trapezoidal

rule

together

with

Romberg's

extrapolation

method

QG2-QG10,

DQG4-DQG32--integration

of

a 299

given

function

by

Gaussian

quad-

rature

formulas

QL2-QL10,

DQL4-DQL32--integration

of

a 303

given

function

by

Gaussian-

Laguerre

quad-

rature

formulas

QH2-QH10,

DQHS-DQH64--integration

of

a 30S

given

function

by

Gaussian-

Hermite

quad-

rature

formulas

QA2-QA10,

DQA4-DQA32--integration

of

a 314

given

function

by

associated

Gaussian-

Laguerre

quad-

rature

formulas

Numerical

Differentiation

in

MATH.

S

SPF

(L

IB)

on

PDP-lO

*DGT3,

DDGT3--differentiation

of

a

tabu-

319

*DET3,

DDET3

lated

function

by

parabolic

interpolation

*DET5,

DDET5--differentiation

of

an

320

equidistantly

tabulated

function

*DCAR,

DDCAR--derivative

of

a

function

324

at

the

center

of

an

inter-

val

*DBAR,

DDBAR--derivative

of

a

function

327

at

the

border

of

an

interval

Ordinary

Differential

Equations

in

MATH.

SSPF

(LIB)

on

PDP-lO.

RK1--solution

of

first-order

differential

331

equation

by

Runge-Kutta

method

RK2--tabulated

solution

of

first-order

differential

equation

by

Runge-Kutta

method

332

RKGS,

DRKGS--solution

of

system

of

first-

333

order

ordinary

differential

equations

with

given

initial

values

by

the

Runge-Kutta

method

HPCG,

DHPCG--solution

of

general

system

337

of

first-order

ordinary

differential

equations

with

given

initial

values

by

Hamming's

modified

predictor-corrector

method

HPCL,

DHPCL--solution

of

linear

system

343

+

of

first-order

ordinary

differential

equations

with

given

initial

values

by

Hamming's

modified

predictor-corrector

method

LBVP,

DLBVP--solution

of

system

of

350

linear

first-order

ordinary

differential

equations

with

linear

boundary

conditions

by

method

of

adjoint

equations

Special

Functions

in

MATH.

SSPF(LIB)

on

PDP-lO.

GMMMA--gamma

function

++

*DLGAM--Iog

of

gamma

function

BESJ--J

Bessel

function

104

361

362

:363

+ LBVP ALSO

IN

MATH.

SSPM(LIB)

on

PDP-lO.

++

THIS

SUBROUTINE

IN

STATS.SSP(LIB)

on

PDP-lO.

EXPI

--exponential

integral

368

SICI--sine

cosine

integral

370

CS--Fresnel

integrals

372

CELl,

DCELI--complete

elliptic

integral

of

374

the

first

kind

CEL2,

DCEL2--complete

elliptic

integral

of

376

the

second

kind

ELIl,

DELIl--generalized

elliptic

integral

378

of

the

first

kind

ELI2,

DELI2--generalized

eUiptic

integral

380

of

the

second

kind

JELF.

DJELF--Jacobian

elliptic

functions

382

105

B5-8

& 9

26Aug72