COMMON_Meeting_Proceedings_196809 COMMON Meeting Proceedings 196809

COMMON_Meeting_Proceedings_196809 COMMON_Meeting_Proceedings_196809

User Manual: COMMON_Meeting_Proceedings_196809

Open the PDF directly: View PDF .
Page Count: 685

Download
Open PDF In Browser	View PDF

COMMON
Meeting Proc eedings
Philadelphia, Benjamin Franklin Hotel
September 9-11, 1968-

TABLE OF CONTENTS
1. General Session/Open Board Meeting
2. AMTRAM for the IBM 1130 with 8k core
3. "Relic" - Remote Job Processing on a 1620
4.

Job Control under the 1620 Monitor I

5. Card System Modifications, SPS and PDQ
6. Conversion of FORTRAN II to FORTRAN IV

7. Date Table for the Monitor System
8.

System for Segmenting Programs

9. SCAT 4: A Binary Synchronous Communication Subroutine for Conversational Use
10. High Precision Integer Subprograms for the 1130 Commercial Applications

0

11. Computer Selection and Justification of Computer Applications
12. A Selective Dissemination of Information System for Medical Literature
13. Project Control Systems
14. Computerized Organization and Maintenance of Files in the A.utomated
Clinical Laboratory

15. Data Reduction Techniques in a Clinical Laboratory
16. Professionalism in Programming

17. Numerical Control Languages for the Small Computer
18.

Job Cost Accounting and Student Monitor System

19. ' An Iterative Information Retrieval System

20. Programs of Potenial use to High School 1620 Users
21. The Pre-Compiler (1620)
"

'.

22. High School use of a 1620 Computer

1
,"""",,MW44U\¢U,u¢,4¥IM ..I?4QAN . .&44. «

M-

;:aa±$L&:&i&E

iIJW.ZSiiJ£JhP.r'

o

TABLE OF CONTENTS/cont.

o

23. A. Machine Utilization Procedure for a Small University Computer Center
24. Documentation - Quality and Uses
25. An Accounting System for Small College/University Computing Centers
26.

An Easy Method of Utilizing Mathematical Operations

27. The IBM Palo Alto Laboratory System
28. Computerized Catalog System
29. Small Computers and Simplex Optimization Technique for Mixtures
30. DYSTAL: A Dynamic Storage Allocation Language in FORTRAN
31. PL/I - Funny Things Happen
32. POWER: !:riority Qutput ~riters ~xecution Processors and Input !3.eaders

c

33. DOS Physical IOCS and FORTRAN
34. Engineering Spooling Program
35. Data Acquisition System under TSX
36. Interrupt Level and Service Subprograms
37. Description of papers presented but not submitted for publication in the
Proceedings

.'
2
,I

Philadelphia COMMON Meeting
September 9, 1968
September 10, 1968
Benjamin Franklin Hotel
Philadelphia, Pa.

o

General Opening Sessi on ••••••.•.•.•.•••.•••• Page 2
Open Board Meeting
Report on proposal for providing clerical
services ••••••.•..•.•••.•••••••••.•••.• Page 20,61
Local user organizations becoming affi1iate s 0 f COMM ON • • . • • • • • • . • • • . • . • . • • • . • . . P ag e 2 7 , 38
Distribution of CAST on subscription basis.Page 31
Addenda to CAST •.•••.•••.•.•.•••.••••••••• Page 34
Subject index for newsletters •.••••••••••. Page 41
Val ue of ne·w sle t ter s •••••.•.•.•.•.•.•.•••• Page 42
Scheduling of sessions •••.•••••.•.•••••.•. Page 50
Spacing of COMMON meetings •••••.•.•••.•.•. Page 52

•

Bonding of Secretary-Treasurer •••••••••••• Page 59
Installation registration •••••••••••••.••• Page 63

... -----

')

C~

MONDAY, SEPTEMBER 9, 1968, 8:30 A.M.
JAMES C. STANSBURY, President, presiding.

PRESIDENT STANSBURY:
Philadelphia.

Welcome to COMMON and

I suspect that among the older members here

at any rate the big question is the result of the elections.
I am not going to introduce the officers until later, but
I'll tell you the results now.
The new President is Paul Bickford and the
Secretary-Treasurer is Chuck Maudlin.
unique position

or

Chuck enjoyed the

being unopposed, so we knew that one.

The new Board is:

Dave Dunsmore, Frank

Maskiell, Norm Goldman, Brian Swain; Wade Norton and Hugh
Kerr.
I'll introduce thea later; theylre not all
here yet.
We've tried to give you a good meeting this
time.

•
I received compliments
on the Agenda which I

deserve; that's Joe Aicher, Mike Schilder,

don't

theAiv~sion

managers.
I've also received complaints.
I don't deserve those.
everybody.

I won't say

We recognize th.at we can't please

All we can do is try.

c

-o----~------------------------------------~-We have a rather good turnout this time.

I

am told the registration is in excess of 600 already, and
from the looks of the room I can well believe it.
This is not a keynote address, as you know.
It's sort of an introduction to COMMON, to the meeting.
We try to tell you any specific problems we might have.
There is no planned speech or anything of that sort.
result of that,

As a

I'll take this opportunity, as usual, to

ask whether anybody has questions, suggestions, requests,
things they want to bring up before the open Board tomorrow

o

night and would like to have people think about before then
-- anything of that sort.
(No response)
I don't believe it!
We'll have changes in the agenda posted in
the coffee break area, registration area, as they occur.
There will be changes.

We'll try to keep that up to date.

We'll try to announce them during the morning and afternoon
coffee breaks, but we'll post them so that people can see
them as well.
Mike, any last minute announcements you

•

want to make?
MR. SCHILDER:

No,

I don't think so.

- -----._ ..._._._._-------_._ .. -.". - _._ ... -

-_._-_ ..__ .----_.._----

- ----_._.

__._-----_._- --_._- -_..__._.-------_._._-_ .._._...

- ------- --

- - _._ .. __.. __ ._.. _.._--_ •.. _...__ ._----_... _.. _- ....•.•...•...•.......-- ...•

:_ _----IJ-_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _4
_-+--_

C'~

PRESIDENT STANSBURY: What were all those
sessions last night?

What were all of the planning ses-

sions doing last night?
MR. SCHILDER: They didn't talk to me.
PRESIDENT STANSBURY:

If you do have changes

-- section chairmen, anybody -- please see that Mike or
Joe Aicher gets them.
MR. SCHILDER:
announce any changes.

At the coffee breaks we will

At the present time there are no

changes in sessions, but that's because we haven't met yet.
I expect to see changes.
The latest wording on the 1130 group is

c

General and Technical, rather than Commercial and NonCommercial.
is.

At the moment I don't know which way the break

Whether you'll be in the Crystal Ballroom or the Val-

ley Forge Room depends on which is the bigger group; I'm
sure Don Kiel will take care of letting you know about that.
That's all I have now.
PRESIDENT STANSBURY: For once we have a
meeting that seems to be well enotigh organized that there
is nothing to say.

But right now 1 am going to introduce

four people:
Don Kiel, the 1130 Project Chairman.

For

c'

~ ______~__________________________________________5__~_ _ •
you new members, if you're l130s Don will be chairing the
first session probably that you attend.

In any case, he

can inform you about l130s.
Dick Gabriel, of the 1620 project.
Al Ragsdale, 360 Project Chairman.
Bob Fostrom, 1800 Project Chairman.
Bill Hill, Model 44 Project Chairman.
These are the people you should be looking
for if you need information about specific machine oriented
activities.

o

What about our 360-20?

Is there any contact

here?
(Everett Sylvester)
Gentlemen and ladies, we do need some response from you -- questions, comments, suggestions.

I

can't believe that no one has any.
PARTICIPANT:

Who are you, Mr.

Chairman'~'

You introduced everyone else and we don't know who you are.
PRESIDENT STANSBURY:

I'm Jim Stansbury.

Officially I suppose I am not the Past President of COMMON.
Somehow or other I never get around to introducing myself.

•

PARTICIPANT:

Jim, there is one thing you

could consider to be brought up at the gener al Board

meetin~,

6

----------~--------------------------------------------------------------------~-----

c

and that is the question of whether or not we are going to
change the organization to allow other user groups to join
us.
PRESIDENT STANSBURY:
gestion.

I'll repeat the sug-

COMMON at present has a rule that we cannot have

chapters per see

The question was what we would do about

local user groups who wish to become members of COMMON.
The existing policy is that they may become members as individual installations, but not as a group.

We welcome

people fr an such groups. We'll be glad to send them copies
of CAST.

We'll be glad to publish their news in CAST.
Let me give you a specific example and it

will, I think, make it clear.
HOUSTON 1130 USERS GROUP.

There is a group called the

Under the bylaws as they now

stand this group cannot become a member of COMMON as a group.
They have to become members as individuals.

The reason

for this is that meeting attendance requiremen totours.
We can't have one member from a 35-man 1130 user group, a
local man, who represents the user group rather than the
installation and quali:fies everybody for membership in
COMMON on that basis.

That t s the problem basically.

Again, we welcome them as individuals as we
do the members of special user groups, but so f at we have

c

"2 YWW

rw

,

7

-o----~------------------------------------~-not considered incorporating them as groups.
The suggestion from the floor was simply
th~t

we ought to think about this policy, see whether there

should be some change s in bylaws to perm it it, und

Er

what

circumstances, and that sort of thing.
This one I think you would need more than
two days to think about, but i f you have any suggestions
bring them up at the open Board meeting which will be tomorrow afternoon at five-thirty.
MRS. LAURA B. AUSTIN (Mid-West Region):

o

1

would like to make announcements of coming meetings. This
information is in your COMMON reference manual.
under Section 2, COMMON organization.

It's

When you get back

to your installation and want to look ahead for future meetings, we try to keep this updated for future meetings; but
I will announce the next two here so that you can make
not es of them.
The next meeting will be in Houston, Texas,
December 9-11 at the Rice Hotel.
The following meeting will be April 14-16,
1969, in Los Angeles at, we think,

•

the Ambassador Hotel

(though there may be a change in the hotel) •
In the COMMON reference manual is the list

8

of meetings through 1970, so when you're planning budgets
ahead take this into consideration when you're looking at
attending COMMON meetings.
The one after that will be September 17-19
in Pittsburgh.
I

wanted to announce these future meetings

because Wednesday at the general session of the Administrative Division we would like to have sane people come forward as volunteers for program chairmen, local arrangements
chairmen for these coming meetings.

Houston is taken care

of, but we do need to start now to plan for the Los Angeles
meeting and for the Pittsburgh meeting; so we would like to
know if you are available to help work in COMMON to make a
successful meeting in those two cities.

We would like to

have your names and some indication that you could help us
and how you could help us at the Wednesday session of the
Administrative Division.
PRESIDENT STANSBURY:
have very much else to do I

Since we don't

seem to

will introduce the new members

of the Executive Board and then our local Arrangements
Chairman, Joe Aicher, wh

0

has a few announcements he wishes

Ito make.
On my right here is Paul Bickford, the new

~

M"·

"ft

-mn ··-r-__

TlP

__

__.

__

_~

___~_. _ _ _ .

_~

___. __________ . _

_~~_~~

_____ _

_() ______~------------------------------------------9---r-President of COMMON.
The gentleman on my left is Chuck Maudlin,
secretary-Treasurer.

Since he is a prominent target and is

not only our new Secretary-Treasurer but our old one, you
mn talk to him with a clear conscience.

He's not only re-

sponsible for what was, but what will be.
The gentleman on his left is Dave Dunsmore,
member of the Executive Board.
The next one is Brian Swain, member of the
Executive Board.

o

On my right is Mrs. Laura Austin, who is now
the Vice President of COMMON but who decided she could not
run again this time.

This is her last meeting.

Next, Frank Maskiell, a member of the Executive Board.
Next, Hugh Kerr, a member of the Executive
Board.
There are two other people who are not present:

Wade Norton and Norm Goldman.

both of them, I think.

Most of you know

They've both been fairly active.

We'll introduce them at the open Board meeting tomorrow

•

night •
I

should say that this meeting is officially

10

the responsibility or myselr and Chuck, the President and
the Secretary-Treasurer.

If you have questions or comments

or suggestions about any meeting after this, then Paul is
your man.

But comments about this meeting should come to

me.
Joe Aicher, who is our 10cal.Arrangements
Chairman, has some announcements he wishes to make and I
~vant

to int roduce you to him and thank him f or a job well

done.

MR. AICHER: Thank you, Jim.
I guess this is the rirst time and the only
time I'll be able to say this, but as of now I
program changes.

know or no

From here on in I'm sure I'll be talking

about more of them.
I

would like to call your attention to a

few things that are in the program just as a reminder.

Pert-

haps Mike has even mentioned them already -- I wasn't here
-- and if he has, the repetition won't be too bad anyway.
Demonstrations of IBM equipment will be
available at the IBM Data Center, which is at 18th and
Kennedy Boulevard, some twelve blocks fram here.

We will

have both on Monday and Tuesday time blocked out at this
Data Center on a variety of equipment:

1130, Mod 20, 30,

c

_ClL'____~~-------------------------------------------I-I--r_40.

From 2:00 to 4:00 Monday and Tuesday there will be

time for some demonstrations, but there is the possibility
of your utilizing these machines.
To transport you from here to there we will
have at 1:45 a normal Philadelphia Transportation Company
bus (which will probably say "chartered") leaving the Sansom Street exit of the hotel.

There will be a sign down

there indicating exactly where this is.
floor in the rear.

It's on the grOlna

The traffic pattern is such that they

mU$t load and unload in the rear.

o

There will be on the board a sheet for anyone wishing to sign up, but even if you do not sign up you
can get this bus at 1:45.

It will be shuttling back and

forth, and there will be chances to get on at the JFK Data
Center, also at the IBM Building at 1700 Market Street,
and, of course, at the Ben Franklin Hotel.
a round trip.

So you can make

The hours, again, are 2:00 to 4:00.
One other thing I wanted to talk about was

this late registration.

The hotel is jammed with conven-

tions and we have permission to hold seventy-five late
registrations; this means Wednesday at six o'clock.

•

We'll

put a sheet on the board and the first seventy-five names
on it will be entitled to stay 'til 6:00 on Wednesday.

At

C

12

the coffee break this sheet will probably be

available.~

If

there are more than seventy-five all I can suggest is that
we double up or you check out and leave your luggage in
care of the Bell Captain.
These birds-of-a-feather sessions will have
sheets on the board.

We'd like you to fill out the

subject~

There will be space for other interested parties to sign
their names.

We'd also like you to put down the day and

the time you'd like these to occur.
be in the evening.

They'll

mostl~kely

We'll assign the rooms, and you can

c

check back to see what room it is.
That's all I have.
PRESIDENT STANSBURY:

There are two

ments I thought of while Joe was speaking.

announce~

The procedure

of getting programs from the 1620 Program Library is to be
changed to conform to the procedure used with the other
machines.

I would suggest that most or the 1620 installa-

tions go to the Program .InforJllation Department presentation
on Wednesday.
your advantage.

I think personally that the change is to
It makes it simpler to get programs, not

harder.
Last night there was some sales information
about a piece of equipment to be used with anll30placed

~, ______

~l

~

13 +-__
__________________________________________

in the registration area.

I had it taken down.

The by-

laws specifically prohibit advertising in the public area
for anything other than IBM equipment.

We have absolutely

no objection to any of you who represent vendors talking
to anybody you want to.

That's your privilege.

We do ask

that you not post notices or not put material on display.
I

don't know who the installation was.

not checked.

1 deliberately have

If they want to discuss it with me or with

Paul they're welcome to do so.
Any other questions?

o

MR. BRIAN SWAIN:
tion of Chuck IvIaudlin.

I would like to ask a

ques~

I am assuming that there are a good

number of people who are here probably for the first time
and, therefore, I wonder if we could spend a minute or two
talking about communications.

We had this brought up to us

last night, that an important lacetot the work of COMMON
is communications bet.een members.
In my experience it took me some time to
find out what was our "method of communication.

Our publi-

cation called CAST was not drawn to my attention very early
and it took me a long time, therefore,

•

to find out how

valuable it is •
I wonder if i t would be of general use if

14

c

Chuck would spend a minute or two talking about what is
CAST, how often it comes out, how you get something published in CAST if necessary.

Would this be of general 1n-

terest?
MR. MAUDLIN: I have something for you first.
The Secretary-Treasurer has a new address and telephone
number; that is, on GS in your book the address is incorrect.

Instead of Texas Woman's University it's Texas

Christian University, and instead ,of Denton, Texas

~t's

Fort Worth, Texas; and instead of 387-1322 it's WA6-246l.
Zip code is 76129.

The user code number if 5130, I think;

I~

(~

I'm not real sure.
CAST is a ptiblication that comes out nine
times a year.

It goes to the printer on the first day of

the month every month during which there is not a meeting.
The letters supposedly stand for ffannouncements from the
Secretary-Treasurer."
the membership.

It

IS

announcements from anybod y to

It consists of correspondence between

th~

Board or members of the organization and IBM, their replies;
in some cases announcements by

IBM~

It contains cross-talk

between members where it is of general

inte~est

to the grou •

Presumably I have some decision to make each time i
ceive something that is for CAST:

re-

Does it go in or not?

-O

IS

~--~--------------------------------------+--

That's the only decisi on.

So far everything that has come to

me has gone into CAST, to the best of my knowledge,

if

such a request was made.
It gets bigger and bigger and bigger.

There

have been issues of CAST as small as six pages and there
have been larger ones; the last one was 92 pages.
th e 1 as t

I think

on e was a ve r y good one.
I don't know wha t el se to say except tha t

it is a very valuable tool and it's something that you
should look at very carefully and something you should con-

o

tribute to.
MR. DON KIEL:

That special information on

the 1130 went out, as far as I know, to all CAST recipients.

If you have no need for that information we in the

1130 project would like to get that back, so we can have
extra copies.
PRESIDENT STANSBURY:

Don is referring to

that 8-page appendix that came out the last time.
PARTI CI PANT:

I wonder if we could continue

to have the notebook binder holes punched in CAST.
MR. MAUDLIN:

•

As long as I'm having them

printed they'll have the three holes •
PARTICIPANT:

It might be helpful if the

16

c

entries in CAST could be categorized, at least with a key
word or subject content on each page, or perhaps something
a Ii ttle fancier, like an index; so then en e could

go throu ~h

these ninety pages without having to read a lot about somebody else's system.
MR. MAUDLIN: To the best of my ability
that's Page 1, I thought.

Ix you can tell me how to im-

prove Page 1 I'll be glad to listen to anything you have to
say.
PARTICIPANT: Would it be possible for the
various projects to get a listing of the membership by machine?

In our planning session last night of the 1130

c

group we found that we could usefully use a listing of the
1130 installations.
general user groups.

It's very difficult to sort out the
I s that possible?

MR. MAUDLIN: It's not something that is
easi1 y arrived -at.

I n the first place, you can't get the

answer by going through_the manual.

If you look at 5130 I

think -you find a 1500 for our machine.
1500 sitting on an 1800.
have a 1620.

It happens to be a

We also have a 360 Model 20.

We

We have a stand-alone 1800 on order and we

have a 65 on order.

You couldn't find all of those by

looking 'at anything in the listing.

c

hI

Nt - -ztrlrittftttibrit. -. jr

_()~.____~~__________________________________________1_7__~_

It is my intention to very shortly after
this meeting send out to the membership the re-registration
form that is required in the bylaws, and to incorporate
that in some sense or other in CAST by having all machine
types listed.

But I don't know any way to do it other. than

to have a page of l130s and just list user code numbers.
SAME PARTICIPANT: That would probably be
good.
MR. MAUDLIN:

And a page of l800s and user
..:

code numbers.

o

With the reference manual itself, that would

give you a list of everything and we would be in there
three or f our times.
PARTICIPANT:

Perhaps you could have

e~.ry­

one put a card in with their re-registration -- what computers they

hav~;

and then just run it on cards.

MR. 'MAUDLIN:

Just the thought of that scares

me -- that many sburce documents.
those cards

mys~~f~

I would rather punch

I'll take it from the forms.

shouldn't take very long.

It

It will probably not be in the

October 1 CAST because that's when the registration form
will go out.

•

The results will probably not be in the

November 1 CAST because I doubt that the replies will be
back in time to put that inf ormation together.

In December

II

18

c

there isn't one, so I would expect it would be in January.
Beyond that

I

don't have any idea how to attack the problem

but just that many cards scares me.
PARTICIPANT:

Isn't there another kind of

ox

newsletter that somebody by the name

Carroll put out?

Is that different?
MRS. AUSTIN:

Thi$will be discussed at
~h. ~~ws~etter

general session Wednesday,

t~

situatiort

That ,one,l"ve D,ot;";heard of.

MR. MAUDLIN:

MRS. AUSTIN: !l'hat'!S'Carroll Hall.
MR. MAUDLIN:

I'misoX'ry.

I was thinking last

names.
Fora period ottime that was incorporated
in CAST.

When CAST got o£t the ground the contributions to
.ini~al,

the newsletter became very

and it died a very

.... ;)..

,\

natural death.
live item.

It may be revived, but it's currently not a

There are two newsletters that I get, one

ox

which has been going into CASt and the other one may go
into CAST.

.

.... ".
,','

PRESI DENTSTANSBURY':"B~:t;b.._k ,you,:";,,", ..

'.,

.1;;':-",.: :,,'

he:r'.:'~'Qr
a sij,~
. of'i,itallds on
.....,;: . ", "
S).····:., ..
>

the people who areplanning<" to.: attend the AM~A~ ,presentatio •
They will try to have Xerox copies at material available

c

"to.

_Cl~I

Tf'TZr"i"

nnw··

____~~___________________________________________l_9__r -__
for all those attending.

(Show of hands)

AMPAC is Auto-

mated Material Procurement and Control.
(President Stansbury repeated the follow i

ng

information given by someone from the floor:)
This system was written for a 1620.

It has

been implemented on a 360 and is widely useful in any area
where this material procurement and control is required.
The fact that i t has been implemented on a 1620 would indicate that it could be implemented on an 1130.
The time for that is 8:30 Tuesday and there

o

are two additional sessions.

There is an abstract of the

presentation on T20.
Let's have that show of hands again, please o
(Show of hands)
adequate.

It looks as if a hundred copies would be

That allows for the fact that people will be

changing their minds, too; a 25% safety margin.
That's all I have, gentlemen.

I

think we

can adjourn until the next session.

(The session recessed at 9:20 a.m.)

•
."

20
OPEN BOARD P,lEETING
TUESDAY,

SEPTEMBER 10, 1968, 5:30 P.M.

JAIvlES C.

STANSBURY, President, presiding.

PRESIDENT STANSBURY: The Iirst thing I

would

like to cover at this open Board meeting, beioreI ask Ior

is a report .from Paul Bickford

questions from the floor,

on a proposal that was made to COMMON by Share and Guide

for providing certain clerical services, in general something corresponding to the paid executive question that

c

came up in Chicago.

After the Chicago meeting Don Madden of the

ACM and I talked about it.

I went out to their meeting

in Chicago with the President of Share and the President
of Guide, and we discussed the propo$a1 that the ACM had
submitted.
I've asked Paul Bick:f'ord, Chuck Maudlin,
Bill Lane and Sam Lynch, whow&$ .at that time a candidate

for Secretary-Treasurer,

to prepare a report on this pro-

posal and tell you what i t is and what they teel should be

done about it.
tive Board I

purpose.

Since this is a decision for tbe new Execu-

will turn the floor over to Paul for that

o

21

~I----~--------------------------------------+-lviR.

BICKFORD:

This summary will also cover

the report by Porzak, Swanson and Blackney of COMMON and
also the ACM Code.

The report of Porzak,

Swanson and Blackney

proposal proposed expenditures (I'm going to briefly summarize what they proposed:)
$25,000 a year;

fice;

cost

of

a secretary,

$5,000 for clerical staff;

$6,000 for travel;

$4,000 for

$15,000-

$1,800 for of-

telephone; and miscel-

laneous expenses -- for a total of between $32,000 and
$42,000 a year.

Also,

they recommended a number of bylaw

changes in their report.
Just about this time the ACM proposal was

o

gi ve n us.

Their proposal considered salaries in the range

of $19,000 for two people -- a $12,000 man with a $7,000
secretary, with taxes and fringe benefits of $2,000; also,
travel, entertainment, phone and other services.

But it

excluded the printing of the Minutes of the meeting.
ACM proposal did not

The

require any bylaw changes; it required

the approval of the Executive Board.

But that expenditure

amounted to approximately $49,000.
The Executive Board discussed it and decided
that the best proposal would be to continue to function

•

under the present system that provides for the SecretaryTreasurer, and a full-time person to assist him in his

. _. ____...__._. . . . . . _______.__.__ __.____
.~

~_._

..,._ ..___.. ,.......___. _._,_. __,.._.".............""..,.. _.. _.. "... '...."., ,,"_.... , ___ .._,., .._,_ .... _"_,,, ___'M_._._ .. ·.".•.•.__..___......_•••__.._............ __ ..•.••_..•__ •__•.• _•.._•..•__"'"" •.•. "m._........ "." .• ".•.,,.,•. _ "...".. ". "." ....., "." ""'.""" ., .._.. "." ..... "_ ••• _,, .. -."_-"- -

-~--

,,~.

.

22

duties.
The reasons we turned down the other propo sal s we r e :
year,

Cu r rent 1 y COf\U·I0N

C

ou 1 d no t a f for d $ 4 9 , 00

°a

or even $42,000 a year, or probably not even $40,000

a year to support such endeavors; that's number one.
with the Porzak, Swanson, Blackney report,

Also,

the time it

would take to implement the changes in the bylaws would probably be at least a year, and judging from the response
we've received on past amendment we probably would not get
it done.
So in order to get something going and to
provide the help that the Secretary-Treasurer needed to

c

get .the CAST out on a regular basis and to perform the other
duties of the Secretary-Treasurer,

such as keeping track

of the membership, balloting, etc., we felt that it would
be better to provide the Secretary-Treasurer with funds to
hire a secretary ,to perform these duties.
In the meantime Share and Guide had decided
to consider strongly this proposal.

I'm not sure yet

whether or not they've accepted it, but 1 feel that if
they do accept it and they do find it workable for them
then possibly in a year's time we might reconsider their
proposal.

Possibly COMMON might be on a more sound financial

o

23

basis or more defini te financ ial basis, and at that time
we can reconsider the issue and if it seemed feasible and
reasonable we could then implement it.

But there seems to

be no question that now it would be premature.
I think that pretty well summarizes.

Are

there any questions?
PRESIDENT STANSBURY:

One of the general ap-

praisals that the three Presidents made at Chicago was that
the probable cost would be knocked down from that quoted
$49,000 to around $30,000-$35,000 a year, since the three

o

user groups would not require three times the facilities,
by any means.
I should like to mention that in addition
to not considering the cost of publishing proceedings
neither of these reports considered the cost of publishing
CAST.

That, is an additional expense.

We'd still need

about a $40,000-$50,000 a year budget to cover, and in
either case we'd be committed to that expenditure and we
would almost certainly have to have, roughly, a year's
backlog as a cash reserve in our treasury before we could
even consider it.

•

MR. BICKFORD:

Also, we might mention ,that

the proposal that we've accepted will be published in the

24

c

next CAST.
PRESIDENT STANSBURY: Is there any discussion
from anyone who would care to express his views on it -endorsement or opposition?
MR. EDWARD J. WOOD (3631):

Are you going

to tell us a money amount that is going to be allotted to
the Secretary-Treasurer to hire this secretary?
PRESIDENT STANSBURY:

We have not stipulated

a money amount, but we intend a clerical type secretary,
someone who would handle the routine duties of the position.
We stipulated specifically that Chuck,or the SecretaryTreasurer, was not to hire such a person, but that COMMON
would pay the salary.

It is expjcted that the hiring

would be handled through the Secretary-Treasurer's installation.

If that wasn't feasible,

then he would obtain the

services of an office temporary on a permanent basis so
that COMMON is not placed in a position of being an employer
and having all the Social Security, FICA, tax reports and
all the other things that have to be submitted in such case.
Does thatclatify what we propose to do?

MR. WADE

~ORTON

(1125):

I balieve that it

was clear to most of you what was accc:!pted am what was
rejected, but as I sat and listened to the details and the

c

~\~,________-4____________________________________________________________________2
__5__-r____

way it unfolded I am not absolutely sure.

As a matter of

reiteration, then, that which was rejected was ACM's proposal to provide and that which was accepted was a report
which detailed what ACM had proposed and recommended that
it be rejected in lieu of the proposal that the SecretaryTreasurer be authorized to hire clerical help.
MR. ROBERT J. SNAILER (1495):
procedure for approving a budget?

Is there a

In other words, would

this have to be submitted to someone?

Is this to be ap-

proved just by the Executive Board or is it open to the

o

general membership?

This is a general type question for

any type of expenditure.
PRESIDENT STANSBURY:

The

Secretary-Treasure~

is authorized by the bylaws to pay the expenses of being a
Secretary-Treasurer.

In our opinion the Executive ~oard,

therefore, can direct him to employ someone to enable him
to perform these functions.

We did not consider that that

particular option needed to have the approval of the membership, but it definitely would normally have the approval
of the Executive Board.
fvlR. SNAILER: Then thi s i s ju s t f or the in-

•

formation of the people here, right?
PRESIDENT STANSBURY:

That's correct.

It's

...

~-.-

..--.--.----.--.--..------- - .

---

26

c

the report that I promised in April at the Chicago meeting.
I'm going to ask Chuck to give us a rinancial
report, not too formal.
MR. MAUDLIN:

I made a hurried trip up to

the room and picked up many things; 1 made a list of things
I should bring down and I absolutely forgot that.
The only thing 1 can remember is the first
four digits of what our current balance is, and they happen to be correct.

We do have in excess of $20,000

$20,8l6+some odd cents.

Most of that came from the Chicago

meeting; $15,700 came in from the Chicago meeting and there
was slightly in excess of $5,000 in the treasury just before that.
I am guessing that there will be $12,000
from this meeting.

That's a pure blind guess, with no

guidance from the program chairman.
The last issue of CAST cost $1,701, by the
way.

We have a backlog to produce CAST for about a year

and a half ir we have no money coming in between now am
then and no meeting expenses that aren't taken care or by
registration fees.
PRESIDENT STANSBURY:

The objective of the

Executive Board has been to establish a balance that would

c'

27

enable us to operate for

appr~imately

went absolutely awry and we lost on

thlngr

a year in case

everyth~ng.

Once we

accumulate that balance we're going to consider publishing
the proceedings of the meetings themselves, and that's the
inducement for belonging to

COMMON.

be available only to members of

The proceedings would

COMMON,

so that would have

some additional sales appeal.
Unless there are other comments on this
specific subject I

am going to open the floor for questions

MR. WM. F. BURGGRABE (3418):

o

At the opening

session I made a few comments about how local users organizations could possibly become affiliates of
something of that sort.
am asking again.

COIviMON,

or

Now that the new Board has met I

fvlost of these people are also new members

Those of us who have been around for a while realize what
the situation is.

I

would like clarification on what could

ha.ppen and perhaps a feeling from the Executive Board of
what t hey would I ike to see.
PRESIDENT STANSBURY:

Because this is a fu-

ture decision actually I'm going to ask Chuck Maudlin to
give the answer to that.

•

He was detailing what he could

now do and what he was now doing to enable things such as
this.

Paul and I have discussed it with him, too.

28

MR. MAUDLIN:

o

I'm arraid I didn't understand

everything that came out of that microphone, so I'm going
to ask Bill to
best

I

go

through that again and then I'll do the

can.
MR. BURGGRABE:

to be pu t in two parts.

The question probably ought

The fir st is: Wha t are the present

possibilities of interphasing local users organizations
with COMMON1

The second part is:

What possibly could we

do

and perhaps what does the Board feel about this sort

of

a si tua tion?
I

think there are obvious pros and cons,

and 1 think the membership is interested.

I

think there

are advantages and disadvantages.
MR. MAUDLIN:

currently there is nothing

clearcut in the bylaws to describe what the requirements are
for membership other than association with one of the small
IBM machines.

After we get out of that category then it's

in the realm of interpretation.

In the past there were

the small regional groups sprouting up and the interpretation was generally· that within a single installation there
would be at most a membership for each computer installation under separate management inside that organization;
that if there were two different computer organizations

c

on- .

-0

29

they were logically entitled to membership.

There were

presented to the organ1zation the ground rules of maintaining membership;

that is,

they had to have enough people at

meetings to satisfy meeting attendance requirements for
two organizations, and with the current bylaws that would
be responding to ballots from both representatives, etc.
There were two things that caused me to interpret a little bit differently in the organization of
the Houston 1130 users group.

I'll talk about that one

first and then defend it on the basis of something else.

o

In the bylaws there is a provision for the
Executive Board doing some things, and when 1 found out
about the Houston 1130 users group the Executive Board
wasn't really around to ask everything,
cisions.

so I

made some de-

The Houston 1130 users group is a honest member

of the organization in the sense that they have a number
and they get a task and they are on IBM's list as an 1130
customer.

There is an IBM branch office number and an IBM

customer number -- it happens to be a customer number of
one of the members

and they are currently maintaining

the meeting attendance requirements by attending the meet-

•

ing •
In addition to that membership which is in

_____.____ _____..... __ __""._._._ . .__ . _."._M_e._. ._. .". __. ____ _
~

~._.

~

".~

_~_.

o

30

the name of the Houston 1130 users group and in the name
of the president, there are several members of the Houston
1130 users group that are additional members of COMMON.

Th~

president is not, and so it means he is only getting one
mailing.

It is my opinion that at least for the time be-

ing they're giving us more than we're giving them by just
putting them on the list.

I'm not sure that will always

be the case because there will be some

organizations that

don't fit in that category.
One of the things that caused me to make
this decision (and I spent well over two weeks in correspondence on it) was an organization that I can't even
ber the name of.

remem~

c

There are several river forecast centers

throughout the country

I think there's one in Cincinnati

one in Fort Worth, one in Atlanta and I don't know where
the rest of

~hem

are --and they're having a hard time

making the meetings.
like many others.

They are small installations, just

I think all of the insta1lations got a

letter about the same time saying that if they don't make
the next meeting they go

out~

. One of the members proposed a joint membership for the entire group and it would be up to them to
distribute information to the rest of the river forecast

o

**

[""j'

'f

H"

rizZ" +rt#

_.5tsb

'11

'K -r"T'TC 'I@

'zr- r..·

.. 'IL"(·····fHIiW

. mgT

31

~-----4---------------------------------------r~
centers.

He gave me a title that was very long.

it had "southwest'"

in it and I

I think

think it had "hydrologic"

in it, but I can't remember the rest o£ it.

They agreed

to distribute to all the river forecast centers the information that was in CAST and in the proceedings that was
pertinent to them.
It seemed to me that this group fit
same category.

in the

As a general rule I would vote th&we

shoul~

continue that policy, but I ' l l stand back now that there's
an opportunity to get a consensus other than mine.

o

That answers one question.

I don't even

remember what the other one was.
f.1R. BURGGRABE:
of it except for one thing.

I bel ieve that answers most
We were talking about the £act

that if an organization actually became a member they might
have a possibility then of also paying a subscription fee
to get CAST mailings for all of the members, whether all
the members are members of COMMON or not.

That was the

point.
MR. MAUDLIN:

There is a currently approved

policy adopted by the Board for distribution of CAST on a

•

subscription basis to IBM installations, that is,

to any-

body in IBM, anybody tha t has an address wh ich has "IBM It in

7

32

it, provided they pay their $15.00 subscription fee.

o

There

have been very few requests for subscriptions other than
from IBM, and so there have been no real opportunities to
make a decision.

I have quoted a price of $15.00, but, to

the best o£ my knowledge, no one has asked to take advantage of it.

I don't think I've had to make a decision on

that.
I

would like to sU9gest that anyone can sub-

Fi.fteen dollars covers the cost of CAST for a year

scribe.

unless i t gets a lot biggertban it is, and I think it would
It's my understanding that Share bas that

be a good mea.
policy

ox

distributing CAST, to ber FFDs on a'subseription

c

basis, and it would be ibyopinion that we should do, that if
anyone wants it .

'La:tgenumbers, I

would like t08\19ge5t,

is a bad idea, but the,n, on the other 'hand, I donttthink
anybody would want litany copies at $15.00.
PRESIDENT STANSBORY:

There wasone·specif ic

clause in the old 'bylaws which prohibited subscriptions' to
other than IBM installations or users.

When we wrote the

new ones, when we drafted the new ones we tookouttha t p'X'o
hibition.

we didn't quote a price for other than IBMih-

stallations,butwe did take olittha:t prohibition so that

it could be done withoUt amending the bylaws.

;~--~~--------------------------------------~--------------------------~----i

o

. HTSr"' [ . . - ""

3'71""'"3

CR-'"

fWrp5nWtt["¥S6 W "i¥I"""-""i"""

W

33

.n
-

~IR.

JAMES MARK (1988):

Are we then from

our Secretary-Treasurer's remarks to understand that local
users that are organization-structured will be permitted
in the future to maintain a relationship with COMMON?
MR. MAUDLIN:

Unless 1 am corrected to the

contrary at this meeting.
MR. BICKFORD:

I

think that the philosophy

behind COMMON is that we will provide a structure wherein
these groups can participate in COMMON.

I

think the struc-

ture and organization is such that we can make it so that

o

they can

~ticipate

in COMMON, carry out their own goals

and still not bother about rules and regulations.
PRESIDENT STANSBURY:

I

would 1 ike to poin t

out my own interpretation of Chuck's rule,

too.

The speci-

fic rule is that the members of a group do not become members because of a group-type membership in COMMON.

We in-

tend it to apply so that individuals may become members
but permit the group to become a member and permit members
of that group who are not members of COMMON and who will
not become members of COMMON to subscribe to CAST and obtain the proceedings through the IBM office.

•

MR. WADE NORTON (1125):

I am sure it's not

the purpose of any Board member here to defeat the good

34

C:

purposes of users, and as surely as there is a precedent
(although we don t t necessarily have to look to precedent)
in that stockholders in a particular company may be either
other corporations or individuals 1 think the same should
apply to membership in COMMON.

I can

think particularly

of an activity within a project that appears very dear to
Dick Gabriel's heart, and that is the 1620 as a tool for
teaching in the high schools.

There is a wealth of infor-

mation around on 1620 and that information needs to get to
these people that are teaching in the high schools.
I submit to you that when they fight the
battle against proration of their salaries that they 1 re not
going to have unlimited funds to attend meetings.

So I

think that the policy that Chuck proposed here serves the
interest of user s.
MR. EDWARD J. WOOD (3631):

For the August 1

CAST 1 sent an a-page booklet to Chuck Maudlin tobe mailed
as a separate mailing.

Incidentally, this was called an

opinion in the opening meeting and it is not; it's handed
out by IBM and I just re-set itso'it would be readable.
Apparently this caused a pretty big increase
in cost of mailing CAST.
You shake your head, Chuck, but $1,700 soundls

C'

-C) ______~-------------------------------------------35--~like an awful lot of money.

Are there that many new mem-

bers?
~IR

• MAUDLIN:

MR. WOOD:

Yes.

Fine.

PRESIDENT STANSBURY: This was a 90 or

92-pag~

CAST, which was rather substantially larger than the last
one.

It was not due to the 1130 information.

Both Chuck

and I felt that this was an appropriate way to distribute
it.

We recognized it was going to people who didn't need

it, but at the same time that the cost of segregating out

c

those who didn't need it from those who did would not be
worth the trouble probably.
MR. WOOD:

The reason I'm really bringing

this up is that the opinion of the few people I've talked
to abou t thi s (and I haven't talked to everybody, of coursE)
is that brand new COMMON members who are 1130 users could
also use this.
or is there

NOW,

somethi~g

is there a policy that you could adop1
you could do to see that this would

get distributed, or something like it in the future if it
came up for 360 or 1620?

Is there a policy you could use

to be sure that this would be included in the membership

•

things that are received from IBM (for instance, the COMMor
Reference Manual) -- that this could be included as a

36
newmembe,r:s handout?
STANSBURY:

swer that question.

I don 't intend to an-

Paul?
If we can find a way in which

it could be reasonably implementable, possibly so.

MR. WOOD:

I think the thing that has to

be taken into consideration is this.

I

specifically re-

quested from Chuck that this be made an 8-page pull-out or
separate piece in this mailing, for the simple reason that
it's atooltbat you would leave at your desk and pick up
whenever you wa.nted to do some quick coding.

I feel this

is the way it should be handled, but, of course, your de-

o

cision would be final.
MR. BICKFORD: It's difficult to handle or to
process all the possible variations of requests that we
get for distributing contents of CAST.

Some people like

them bound and punched with holes, and some people don't.
Some people l.,ike all types of variations.

We don't have an

organiza.tion at the Secretary-Treasurer's installation such
that he

ca~ implemen~

all these variations easily.

1 would

like to think that if it was something of value to be distributed'frOIll now on that we could fairly easily make something that we', could do every time, without any significant

o

'nrJ'f"r' r--C:FS"" W' "rPM

- P" -

-g··Htty-----

-,

"W"·--

__() ______~---------------------------------------------3-7--~-1
additional cost.

MR. WOOD: Every time with CAST?

MR. BICKFORD: The new membership packet,
with specific reference to what you're talking about.

Any

time we try to change CAST in the way it's presented, as I
understand it, it's difficult because it is set up and it
has worked reasonably well.

If Chuck wants to change it

and makes this change effective and continuous, then fine,
if

it's something worthwhile.
Does that sound reasonable, Chuck, that if

o

there is some change you want to make in CAST and the group
wants it that way we'll consider it

MR. WOOD:
should be in CAST.

for change'!

I'm sorry.

I

I don't mean that at

PRESIDENT STANSBURY: I

don't mean that thi$
~ll~

think 1 may have a

satisfactory alternative; I don't know.

As you know, this

information is prepared from masters that are prepared
from the copy you send to Chuck.

How much trouble is it

for the submitter on such items to keep a duplicate set of
the originals, or request the masters from Chuck -- let the
author of such information or, in your case, a person on

•

the 1130 process or possibly a two or three man group be
the custodian and Chuck could simply supply them, I am cer-

r~"·--

38
t~

tain, with the addresses of the new 1130 members and

o

group could take over the responsibility and possibly charg e
for reproduction cost or a handling charge for sending it
to new members.
MR. WOOD:

The author happens to be in New

York because it was authored by IBM I don't know how many
years ago, and what I received was a Xerox copy of a Xerox
copy of a Xerox copy of a Xerox copy and it wasn't very
readable; so what I did on my own, -at my own expense, was
~

have it set so it's readable.

Maybe the whole answer is

to have IBM print them themselves and distribute them to

o

every 1130 installation and just forget about i t .
Is that what you would say -- to attempt to
get IBM to do it?
PRESIDENT STANSBURY:

Obviously.

Then it

doesn't cost us and it doesn't cost you.
M~.

WOOD: Fine.

That's what I'll do

PRESIDENT STANSBURY:

~hen.

Particularly if it

was originally authored by someone in IBM I'd suggest you
try to track down the author because it was probably copyrighted.
MR. WOOD; I know it's not copyrighted.
PRESIDENT STANSBURY:

As to the actions of

c

td

«t

. j Him_f»... t

'tt'

#

tid

't "dbr" tfrihtiWbrid sfrirlti'-"

tidhrib£ bftri tttibt

6'

r ..

-"-"-p-+l f

-t·'T"'··'·'

.-

["

t·"".···'··,·"

39

the Secretary-Treasurer in reference to membership for
local user groups I would like to have a show of hands from
those present who would endorse the Secretary-Treasurer's
action, policy, as he stated it.
(Most hands were raised in favor.)
PRESIDENT STANSBURY:

Are ther e any wh 0 wi sh

to express an opposing viewpoint?
MR. JOHN CRANDALL (3473):

Won't this tend

to decrease our membership if there are people that are now
small users that will join a local users group and tend to

o

drop out and rely on the services of the users group?
PRESIDENT STANSBURY:.
meeting.
out of.
I

You've been at this

You tell me what you get the most information
Meetings?

The proceedings of meetings?

CAST?

think that those people who can attend will become mem-

bers, that this would only apply to those who cannot.

That~s

my personal opinion.
MR. DONALD S. GARDNER (1150):

The problem

that Ed is talking about is certainly as old as the group
is.

Many years ago we had a 1620 users group and we put

together newsletters and had lots of information that was

•

quite applicable from their viewpoint, and as late as a

yea~

ago 1 overheard 1620 people talking about the problems of

40

readers of five years ago.

It seems rather foolish to

have talent go to the trouble of publishing and writing the
newsletters and have that information unavailable to new
people, and these new people will now be spending time and
resources doing what some of us have done many years back.
I think it's a responsibility of COMMON.
PRESIDENT STANSBURY:

statement:
IBM

in

~,vhite

To con£irm Don's

About six months or a yearago I got a call from
Plains wanting to know i.f I had a particular

1963 issue of the 1620 newsletter.

Some IBM installation

in Iran needed a copy of that specific newsletter.
they found out

How

answer to their problem was in that

th~the

newsletter I don't know.

However, there does exist a file

of newsletters and certainly there are complete sets of
proceedings in IBM's file.

There has been serious consid-

eration several times of having someone take over those
files as archivist

ror the group and permitting him to re-

produce at cost specific proceedings, newsletters, papers,
what-have-you.

It has never been implemented because no

one was willing to take the responsibility.
There are now, however, a lot more members
than there were then.

You might wish to farm it out.

But

such a thing could be done if the people in the organizati cln

c

i'""j'"r"f"'WJUUfOe-

~

__________~_________________________________________________________________4__1__~__~

were willing to accept the job of doing it.

I don't think

it can be imposed upon the Secretary-Treasurer.
MR. WM. A. PEASE (1516):

1 think it's im-

possible for any organization to lift the realm that it
operates in without lifting itself, and since the purpose
of COMMON is to free people's time (among other things) so
that they can find out more new things for us I think it
would be.-the proper function of a committee to compile some
sort of subject index to the past newsletters of CAST.
"
They could have representatives, say, to the Administrative

o

Division to operate with the archives, so that they could
have access to the archives at IBM or could communicate
with them; and some sort of lines could be set up to

dis~

tribute, for example, to 1130 people an index of 1130 articles.

You would throwaway the old index once you got a

new one, so you wouldn't be accumulating paper.
PRESIDENT STANSBURY:
pared for the
firs~

procee~ings

three or four years

Such an index was pre-

of the 1620 users group for the
I've forgotten whether it was

everything up to '63 or everything up to f64--and was published in one of the proceedings.

•

In other words, there

do exist partial indices to past proceedings.

I don't be-

lieve that any has ever been compiled for the newsletter.

42

----------~------------------------------------------------------------------------~---

I'm certain about CAST.

C'

The point I was trying to make is

that 1 completely agree that it's a subject for a committee.
You people organize the committee, please.

We'll work with

them.
MR. NORTON:
reference here.

I'd like to get a frame of

It appears to me that one of the things

we are striving for is to insure that new users get information immediately and, as such, we have been asked to put
a lot of things into CAST, to be sure that all of the inrormation that is pertinent to the activities of a project
or even of a committee goes via CAST even though it may not
be of interest to the full membership.

c

The other problem, as I see it, is to get
that essential information about the organization to everyone.

We are a very large organization and we are a very

diverse organization.
These thin; s are the things we're talking
about.

It' s really a communications problem:

How do we

get to those who need to know that which they need to know
and in a form in which they don't have to dig out of it
or dig through a myriad of stu£f' just to find what is pertine.nt.
Now I want to make a comment, having defined

C

43

n
the frame of reference.

It is my personal opinion that

the reason that the newsletter dies is that it was redundant
by the very fact that it was too broad and was trying to
cover all of the machines.

It was trying to perform the

function that had been delegated to CAST in effect.
Now, there is precedent, since we're generically the old 1620 users group, for newsletters that pertain to specific teams or sub-units of the users group.

I

can cite you the Electric Utilities Newsletter, which was
a regular publication, provided information that was needed

o

by the utilities group, and did not clutter up the general
newsletter.
This was not a function of the Executive
Board, though.

It was the function of that particular

group.
As I see it, the only problem that exists on
newsletters

,in projects is finding an editor and then be-

ing sure that some procedure is set up or some means is
made to get this information into the hands of new members.
I don't think we have to mail a volume yea thick each month
in or de r t

•

0

dot h is •
MR. GARDNER:

Wade, I would like to take

issue with just part of what you said.

The newsletter for

i_ _ _ _ _ _

~--------------------------------------------4-4~~-()
any machine group, whether it's an individual onefor a

group or a newsletter to cover all groups, in my opinim
is very important.

benefit by one.

I

I think everybody in this room will
think the reason it died is not that

CAST took it over, but that

i t was allowed to publish

in CAST that belonged to the newsletter.

thing~

Someone allowed

it, perhaps byde£ault because we had no editor.

But CAST

took it over and started publishing verbatim what was sent
in, items relating to a particular machine group, and took
over the duties of the newsletter.

That's why it died.

It had a way to take care of these things.
W e c oul d

indeed make some eff or t

in some

c

committee or by some people -- to revive the idea of a
newsletter, with types of items that are common denominators.
PRESIDENT STANSBURY:

I would think the im-

plication here is that the newsletter should be confined to
the users of a narrow, defined range of interest -,;., one
machine type application, project, or
I

some~hing

of the sort.

would say thatona project basis I wou,ld probably agree.

On a machine type basis I'm' afraid I don't.
MR. ROBERT T. SCULLY, JR. (1957):

It took

us something on the order of nine months to find out that

o

_()______~-----------------------------------------------4-S--+_-COMMON existed as an organization for 1130 users.

If I

may suggest that we can exert any pressure on IBM, I think
they should be at least the unit that informs their new
customers of this organization for this particular type of
equipment.

It's because so many requests have been made

of our sales office that 1 found out about the existence of
this organization.

If in addition to delivering the new

equipment they could include as part of the installation
procedure a back-up file or some sort of index on what has
already been discovered by the users of the equipment prior
to the new customer coming on the line, we could get additional benefit out of that.
MR.

MAUDLIN:

I don't think IBM should be

condemned for their role in thaparticular respect.
do something that we do.

They

Since we do it by bylaws and

thou~t

a long time about it, then I can't condemn somebody else
for doing the same thing.
ship list.

we won't give out our member-

The equivalent in IBM is

"we won' t te 11 you

who has IBM machines."
Now, since they won I t give it out the next
best thing they can do is to inform all purchasers or

•

renters of machines about the existence of users groups •
Corporate IBM frequently (1 don't know how often, but

. - - - - - - ' . -... -, ....-.~.-.-------.,',.~

. ---..- ..........-...

--.~

....

-."...........

,-.-".,~-

........ ..~'........."~"-..... .. ".'.. "",._,"""
,

'

46

significantly more than once a year) sends to branch ottice
and sales reps and everybody else associated with customers
memoranda describing users groups, pushilYj them, naming
the secretaries, publishing that in the DATA PROCBSSOR,
which some of you have seen, I am sure.
The customer, however, has one interface
into IBM and that's the sales rep ,and it's his decision.
If he thinks it's going to help his sale or increase his
sale or increase his relation, then he's going to make
COMMON known; and if he doesntt think so, he may not.
There are several subscription services available rrom IBM

c

and the only way you find out about them is .from your sales
rep, and corporate IBM tells the sales reps that it's a
good deal; then they make it available to tha customers.
The only people we can condemn are the sales

rep~.sentative

wh 0 don't get the word ou t.
We can

cond~mn

us; we don't have

kind of policy to get the, word out, too.

~4e

~igbt

M.ybe we otight

to be advertising in DATA MISSION; I don' tknow.

But

corporate IBM is at least m;J.k.ingtbe word known to their
salesmen and the inx ormationshduld be gotten; to each of
the users.
I would like to comment on therea,l death

o

n-

47

of the newsletter also.

The newsletter died of lack of

items for the newsletter. It was published as part of CAST.
There was an editor doing a good job.

That editor was

forced to resign because of changing installations.

At

that time the editor had had submitted to her five documents that she had had for almost eight months waiting for
enough to make one page.

Those things were put together in

one blob and went into a CAST as they were submitted.

And

there has never been any real move to create a newsletter
sin ce then.

()

I would suggest that if there are people who
want newsletters for any thing that is a function of the
Administrative Division, and we should iron out a way to
get that done.

But until people want to contribute, that's

the situation.
MR. MICHAEL SCHILDER (1557):

I was either

fortunate or unfortunate in not being part of the 1620
group (depending on how you want to look at it)

user~

in that I

can now look at COMMON from a different view than those who
were members.

We have a number of machines here and we

have a number of applications, and the answer you gave be-

•

fore to one of the questions was: How can the secretaryTreasurer or anybody decide whether or not something pertaips

48

to a particular machine or not.

I think the answer really

boils down to the fact that there are some areas which are
of general interest to COMMON members.

There are some area ~

which are of particular interest to application people and
other areas of particular interest to machine people.
1 don't have immediately the answer on how
to separate these out, but perhaps by :finding out from the
members themselves what areas they're interested in then
we could not reorgan ize CAST but perhaps organize a newsletter which would go to the people who want to get the
newsletter.

Then maybe CAST could be separated into an

1130 section and a 360 section, e·tc.

The important thing is that COMMON has to
recognize now that even though we have a common goal (and
I'm not trying to make a pun) we at the same time do have
individual area interests for which we want to be satisried.
And certainly as Program Chairman of
that out.

this meeting I found

But I think we have to consider the fact that

there are two separate and maybe contrary desires on the
part of people:

to be together and also to be separate.

I think we have to recognize this, and I think there are a
lot of areas which we may have to reevaluate in . order to
come up with a solution.

49

I submit that what we

PRESIDENT STANSBURY:

are discussing here is a communications problem within
CAST.

There is a general session of the Administrative

Division tomorrow morning at ten o'clock in this room, and
ways and means or implementing this

think can be

I

discusse~

then.
MR. MAUDLIN:

There have been two other ses-

sions that I've been in where there was a discussion of
thumbing through a large issue of CAST to find those
you were interested in.

o

articl~s

I'm probably not the best in the

world for creating titles, but there is an honest effort
on my part to get "machine type" in the title; and the first
page of that document, whether it's good, bad or what-haveyou, does contain at least some indication of what is in
all of the documents.

I

suggest that you can go down the

first page and if it's machine-oriented then there will be
a machine type in that title and if it's not machineoriented and it's of general interest to the membership
the article will be directed to the membership, or the

titl~

will be such that it will indicate that it's to the membership.

•

PRESIDENT STANSBURY:

I

th ink that

will cl os~

the discussion, and I think the appropriate place to con-

so

:----------~------------------------------------------------------------------------+----

o

tinue it will be in the session tomorrow morning.
1 know that the OS project got stubborn and
decided they were going to continue meeting.

The 1130

projects -- one or two or more -- got stubborn and decided
they were going to continue meeting.

That sounds as if the

Executive Board and Program and Local Arrangements Chairman, and so on, could use some guidance as to why you felt
i.t necessary,

what your objections were.

comments to that

May 1 have some

effect~

MR. WOOD:

at 8:15, going to 9:45.

Tomorrow morning we're starting
The problem was that an original

schedule was sent before the agenda was made up.

The

agenda was made up and scheduled us from 8:30 to 9:30,
10:00 to 11:00 and 11:00 to 12:00.
take an hour and a half.

One of these talks must

Now, we're not going over any-

body's head, but it is considered a tutorial and to have to
break it for a full hour and a half -- going from 9:30 to
10:00

and then through this Administrative Division to

11:00

would just hurt us terribly.
PRESIDENT STANSBURY: I 'm not criticizing.
MR. WOOD: I know you're not.
PRESIDENT STANSBURY:

reason.

I'm merely asking the

o

_()______~---------------------------------------------5-1--~MR. WOOD:

I realize you're not criticizing

and I'm merely telling you the reason.
To attain this we had at first decided that
we would go from 8:30 and hopefully at 9:30 hold everybody

through the coffee break.

Then Mike was kind enough to

suggest that he would get us an early start and then we'd
go fifteen minutes into the coffee break and then still

giv~

the people an opportunity to come to the Administrative
Division meeting.
MR. JOHN FISH (1878):

o

Our meeting was ex-

tended from three sessions to around six, with three more
planned.

We have encountered many areas which we didn't

realize had a common interest, and this meeting has brought
that out.

We'll plan for a much broader session in Houston,

in the area of seven to nine sessions.
PRESIDENT STANSBURY:

Then the general con-

sensus seems to be more or less that the project itself
didn't anticipate all of its requirements and that they
weren't properly communicated to the Program and Local Arrangements Chairmen.
MR. FISH: It's spontaneity.

•

PRESIDENT STANSBURY:
fine, yes.

The spontaneity is

Don't get me wrong; I'm glad it occurred.

I

am

-

--

----

.-... ...•---~-.-----.-.----------..

52

o

I am vitally aware after this meeting of the
problems of scheduling and I know there are bound to be
overlaps where in small installations frequently only one

c

53

person can attend.

This has been true for at least the

pas~

two meetings as well as this one on my own part, where I
would like to attend more than one session at the same

time~

This might give us a little less chance of overlapping too
much.
PRESIDENT STANSBURY:
tive Board's position or decision,

I'll give the Executhe reason for the deci-

sion, and then open the session to comments from the floor.
The meetings, despite Don's feeling, are not
scheduled in quite the intervals he said.

o

They're sched-

uled, roughly, four months apart during more or less nine
months of the year.

If you will note,

there is a meeting

in September (which we had expected to continue), there is
one in December about four months away
MEMBERS:

That's three months.

PRESIDENT STANSBURY:

I

said "abou t" 1

You

can't make it any later in December and you can't make it
early in January.

There are obvious reasons for that.

There is a meeting in April.

We have

found from past ex-

perience that summer meetings of COMMON hurt and hurt badly.
There are too many educational institutions in here, so we

•

deliberately try to avoid the three summer months •
That's the basic reason for the spacing.

In

o

54

addition to that, we have ll30s, which are relatively new
machines, and we have a very large number of new users.
There were something like 160 in the new users meetings
at this session.

If we spread that out and try to get by

with two meetings a year,

I

submit that with the systems

in the state of f1 ux tha t they now are thi s just wou 1dn' t
work.

You'd be missing too much information.
I

might add that while Share and Guide each

meet twice a year those meetings are spaced so that Shareis
Systems Division meets at the same time that Guide does
and Guide's Systems Division meets at

the same time that Shar

c

does, and as a consequence their Systems Divisions meet
four times a year.
That's the reasonf or our decision.

Does

anybody care to comment?
I

MR.

ROLAND MAGEE

(3079):

As to the

schedul-

ing of the meetings in the last year or two, it seems to me
they follow immediately after or immediately before a holiday and this often creates problem with transportation,
and so on.

I'm wondering if that couldn't be taken into

considerati 00 and spacing changed.
MR. MAUDLIN:l have a real
that one.

qu~ck

answer'to

o

There is a great deal of effort that goes into

!t

-C) ______~---------------------------------------------5-5--~~
avoiding holidays and when you avoid all of them you're
either immediately before them or immediately after them,
and usually both.
MRS. LAURA AUSTIN:
that, too.

Jim, I'd like to add to

In working on the planning of future meetings

the Administrative Division has been planning through 1972
for meeting dates, and at this time we have found that in
1971 we could not get the dates we wanted. We were too late
trying to get dates in hotels for 1971.

When you stop to

think that COMMON has not planned this far ahead in the

o

past you can see some of the reason why the meetings have
come at the dates they have.
PRESIDENT STANSBURY: Does anybody else wish
to sympathize with us7
MR. FISH:

Speaking for the OS Committee,

during this session we had approximately twenty-four people
at the average session.

Of these twenty felt they could

not attend the HouSt9n meeting because they couldn't

justif~

to their companies attending another session so soon after
this one.

Of these twenty people many could have con-

tributed a great deal to the Houston meeting and it would

•

have allowed us to plan better for papers and presentations
and to se t up subcommi ttees to work harder.

I feel this

is

56
,

an extreme handicap for our group.

PRESIDENT STANSBURY:

o

I agree because the

Executive Board has to attend every COMMON meeting and we
had a meeting in Philadelphia in July.

It does impose a

burden and I will have to admit that the hard core of

the

project effort, whether it be OS or an applications project or 1130 or 1800 or Model 44

whatever it be -- will

be a relatively few individuals to carry that load; and the
success of your project depends upon how much they can contribute and how much they benefit from these meetings.
From what I have gathered (Dick Zerweck,
John, is Manager of Operations at our installation, and he' s

0

been talking about your meetings) the people who attended
here feel that they acquired a great deal.

1 submit that

even if those same people attended Houston there will be
additional attendees from the midwest and you will probably
find it will be an equally rewarding meeting.

MR. FISH:

I felt that the Houston meetings

would be much better with the addition of people from the
east coast who could plan and present better formal presentations in additi on to those of the Houston people.

1

think this could be done if we had better spacing of meeting5, such as every four months.

G

5W·· ...

. j""Ur"trar·

EP"'·

57

PRESIDENT STANSBURY:

My answer to that is

tha t my boss and Dick t. s boss is here, and he has requested
Dick to attend the Houston meeting.
MR. BERNARD A. SOBEL (1490)
fend,

if 1 may, the Board.

I'd 1 ike t

0

de-

Sometimes it's hard to defend

yourself and I'd like to defend the Board in the light of
my experience with a few other groupso
meetings and I have to detest it.

I detest coming to

Every day I'm away fran

my office means that I have to work two days more when I
get back.

o

And this is true of all of us, and it's even

more true of the members of the Executive Board who must
attend every meeting, who must attend interim meetings.
I am grateful to them for doing this.
are doing something that I

mysel~

They

want to have done and

that I don't have the time to do myself, and I cannot possibly object to their setting up a meeting in Houston tomorrow or, the day after.

The fact that I personally will

not be able to attend cuts no ice.
girls, I hope.

I will send one of my

I will possibly walk in on it -- and this

meeting is only a half hour airplane jaunt from our operation.

•

The simple fact is that there are many organizations in which all of us are interested.

I am inter-

58

ested in

th~s

group; I am intere$ted in Share.

I'm also

c·

a member of and active in possibly ten technical societies
-- in the chemical engineering, chemistry and computing
sciences fields.

If I were to attend

every oned these I
meetings.
that.

I

I

~very

meeting of

would spend my entire year

attendi~g

don't think the Ethyl Corporation pays me for

don't think that's true of anyone of us here.

The fact that there are some people who are lucky enough
and willing enough to spend their time -- God bless them.
(Applause)
MEMBER:

Don' tput the cross on the Execu-

tive Board or on your management.

This cross bE!loDgs on

your own shoulders in th).s,c&$e:for the

simplere~son

that

the management can't knQw . the value or what yougetbere
unless you cOlJlmunicate it to them in a report when you're
still warm and enthu$iastic,when you get home.

That~s

when

you write your report and write it in a concise manner.
Don't try to·go into detail on all tb-ePlpers.

JU$twrite

in a concise manner to whet their interest so that they
can come to you ror the details.

See that it gets distribu-

ted in proper places around your organization.

Then you'll

be in a good position.
One of the best things that happened to me

c

n-

59

was missing the Cincinnati meeting.
budget couldn't af£ord it.
check with his boss.

My boss told me the

What he neglected to do was to

His boss happened to be Director of

Research who is immediately under the corporate Research
Director.

Now my boss has to justify my staying home when

there's a COMM,ON meeting 1
PRESIDENT STANSBURY: Is there any other
discussion?
MR. LARRY ARMBRUSTER (3408):

I 'm

going to

kick a horse that 1 thought had been dead for about a year

o

a couple more times.

As some people will remember,

I

got

up at the open Board meeting in Cincinnati and made a request for information as to the bonding of the SecretaryTreasurer who handles, according to his own statement, some
$20,000 right now, and the bonding of the Local Arrangements Chairman who, according to the Secretary-Treasurer's
report, is handling some $15,000 or so.
1 have the utmost trust for everyone in

room.

If you don't believe me,

thi~

come up and I'll give you

money to buy your dinner tonight if you promise to pay it
back.

•

But this is not the point.

We are an organization

with a rather 1arge membership, as has already been stated,
with a

lot of money relatively speaking; but they are in no

I
f

60

way protected by a bond.
I

would like to get an answer as to whether

this has been discussed by the Executive Committee and what

decision has been made.
PRESIDENT STANSBURY:

There has been no dis-

cussion as yet of the bonding £or the Local Arrangements
or Program Chairman.

There was a resolution passed by the

Executive Board, I believe originally in San Francisco
possibly at Cincinnati, directing the Secretary-Treasurer
to bond himsel.f at COMMON's expense.
not give him a deadline.
as a member

cer,

I

of

Unfortunately, we did

He has not done so yet.

Merely

the new Board and not as the presiding offi-

can say that the Executive Board las t

night passed

a new resolution instructing Mr. Maudlin to bond himself im-

mediately.
MR.

ARJVIBRUSTER:

Thank you.

And 1

had my

hand slightly tarnished a little bit when I was told that i t
\-

was already in the bylawso

Now let's beat this other horse

-- the Local Arrangements Chairman.
MR. MAUDLIN: That will be a part of the
bonding when it takes place.

It will be thrriugh my bond.

He will be an agent of mine.

PRESIDENT STANSBURY: That I didn't know.

o

I'll

t

,tit

"¢Pith

$_

t

tizttrittritiHbtrittf "j"t& Hit"

"tt

±rl&tii#±

iritHrini

u

±WzH¥ ,.,

Iny

-Tn··

WRn

"j'T

t . 'en

Y·WIb'-' ,.. W··'·

61

Thank you, Chuck.

This is about

MR. ROBERT SNAILER (1495):

your secretary, Chuck, and I'm completely in favor of the
idea.

I would like an opinion on this:

What is the bene-

fit going to be by Chuck having this secretary?

Is this

secretary going to serve the existing membership well, or
is this secretary also going to entice many more new members than we would have if we didn't have a secretary?
PRESIDENT STANSBURY:

I hope not.

going to be that attractivel

o

I didn't think she was
I dontt think

Joy would let Chuck get away with it!
Joking aside,

though, I don't think it's go-

ing to assist us in the slightest to obtain new members
except as the prompt and efficient performance of the
duties of the Secretary-Treasurer would encourage new members to join the organization.

We wouldn't have hopefully

too many letters addressed certified mail to the President
of COMMON wanting to ,know why Chuck hadn' t

replied to a

letter which rather obviously he had not received.
It will increase the expenditures of COMMON
in order to give you the service that we think COMMON wants

•

and needs

nOtA7,

but we don't expect it to give us new mem-

bers per se except in such fringe benefits as arise from

!,..,

r

-~~'''-".''''''----,,---,--~-----,

62
the prompt and efficient perf ormance of Chuck. t s du ties.

MRS. AUSTIN:

This new membership problem

is going to be one of the topics for discussion in the
Administrative Division session tomorrow.
PRESIDENT STANSBURY:

of that session.
tw

0

Laura will be Chairman

Laurs has been the Vice President for

years and I couldn't :twe done the job wi thou t her.

(Applause)

She now has to give up her position not only

as member of the Board and Executive Vice president; but
she a1so has to give up her post as

Administra~ive

Division

Manager.

I

think paul has a couple of

announcem~nts

Q

he would like to make.
MR. BICKFORD:

meeting I

Jus t bet ore coming- into the

did contact a new prospect for the position of

Administrative Division Manager, Bill Lane, who was

formerl~

on the Executive Board, formerly Western RegUn President.

He was also a candidate for President.

He did decide to

accept the position of Manager of the Adminis trat ive Division.

So we now have all Division Manager slots filled

and I think we have excellent people in all of them.

We're

going to strive to continue to build COMMON and strengthen
it within.

I

think that with the good managers and good

()

... -

rnf"···r-r

.

-n'fIfTHt.,.'U

-

,.,,"

~.
63
--'I----tl~------------_+__

project leaders thatwe have we'll have as good or better
meetings in the future.
PRESIDENT STANSBURY:

And 1 can also add

that for the first time since we've had division managers
they aren't members of the Executive Board.
try~ng

~o

We have been

for two years to get rid of the feeling of wearing

hats because we felt the jobs were impossible for one

man to perform successfully -- two different jobs.
MR. SCHILDER:

I don't at this point want to

start something new, but I do want to bring up a question

o

that I've been thinking about.
the question.

1 don't know how to phrase

1 know we have a structure which says that

the way in which we collect our "dues" is by attendance at
meetings.

Before this time I have never been aware of a

problem that does exist.
about it.

1 raised it with you.

Chuck

know~

The question was also raised with me from people

locally. This is it:
When an installation sends a number of

peopl~

to a meeting because the meeting happens to be located in
that city or nearby the end up paying the number of people
times the fee.

•

I discussed this with my manager, the one

above me, because I had to get his okay to get the money_
One of the ideas he suggested which I think I'd like to

64
I

~-,-------;~-----------------------------------------------------------------------+----

o

offer for consideration is that maybe at this stage of the
game, or somewhere along the way, we should consider an
installation membership rather than having the individual
who attends pay.
PRESIDENT STANSBURY: I'm going to correct
your phrasing:

an installation registration.

MR. SCHILDER:

Okay.

Actually the installa-

tion does belong, not the individual.

One of the things he

said was,

for example, okaying

ttl would have no difficulty,

$150 a year" -- that's just a figure he picked out of the

air -- "if I

knew that,as a result, whenever there was a

COMMON meeting I could send same reasonable number of

peopl~,

c

paying only for their luncheons or their cocktail parties
(extra things that normally you put on an expense account
anyway)."
1 discussed this with Chuck, and I
ltd b ring it up.

thought

I don't intend to keep the meeting here

all night discussing this, but I

just wonder if we shouldn't

at this point do something to start to consider this as a
possibility, or something like it.
PRESIDENT STANSBURY:

Paul ha. suggested that

the Executive Board will take it under advisement.

I

th ink

I will point out that there are two or three alternatives

c

65

-0
here.

We agree that the registration fee is not handled

equitably for thcs e people who attend every meeting, and
I th in1

particularly for those who have multiple attendees.

the proper procedure here would be for the Executive Board
or the Executive Board plus possibly some ad hoc members
from the group (as an ad hoc committee) to look into the
matter, look over the various alternatives, see what kind
of position they would cost, and have a report at the
Houston meeting.

Then we can bring it up for discussion,

and have some facts and figures when we're talking about it

o

But that is Paul's decision and I'm going
to have to relinquish that, Mike, to him.
MR. SCHILDER:

I

think that's a good

idea~

I wasn't trying to raise it on the floor now, but just
wanted to bring up the question for consideration at some
time soon.

I think it should be considered by the Execu-

ti ve Board.
PRESIDENT STANSBURY:

Okay. 1 don't think we

should try to do it here because while we have considered
several possibilities I couldn't tell you relative cost
or anything of the sort.

o

MR. BICKFORD:

One other announcement.

The

Executive Board unanimously appointed Wade Norton as Vice

66
I

I
I

President o£ COMMON.

o

Hisprimar.y responsibility will be to

have the division managers report

to him.

d~rect1y

(Ap-

plause)
MR. MAUDLIN:

I went up to the headquarters

room to get some Xerox copies made a little earlier today
and I

noticed what I think is an abuse of a privilege.

I

was there while sixty copies of an in-excess-of-forty-page
document was being created and a hundred copies of a nineteen or twenty page document was being created.
think that's what the Xerox machine is for.

I

I

don t t

think when

people prepare papers that a part of the presenting of the
paper is to think about it before you get to the meeting
and reproduce it before you get to the meeting.

I'm sure

it's cheaper.
MR. SCHILDER:
However,

there

w~e

Chuck, I agree with you.

special circumstances there.

The gen-

tleman who asked for the sixty copies came with forty
MEMBER:

T~e

copie~.

19-page document is mine.

cut it down from fifty pages on request.

I

An 1130 insta11a-

tion has to have a lot of money to have a Xerox copier
with it.

If IBM can't get it collated -- they say they

have'a problem getting it collated -- we're going to go up
and collate it ourselves.

We aren't trying toabuse any

G

67

-0"
privilege that we're given.
PRESIDENT STANSBURY:
I

As a personal opinion,

agree with Mike and tend to disagree with Chuck.

I

do

not think it was an abuse.

Okay, gentlemen, 1 move we qui t.

---(Whereupon, the meeting was adj ou rned at
7:10 p.m.

----

0

0

- ....--

..

.. ~-

o

c

c

Alv1TRAN FOR THE IBM 1130 WITH 8k OF CORE

Presented at
PHlLADELPllIA COMMON MEETING
Septelllber 8 - 11, 1968

c

By
T. J. Buntyn

COMPUTER SCIENCES DIVISION
RESEARCH LABORATORIES
BROWN ENGINEERING, A TELEDYNE COMPANY
HUNTS TILLE, ALABAMA

•

c

JI"llP'·-

INTRODUCTION

Automatic Mathematical Translator (AMTRAN) is a conversational
mode, mathematically oriented language developed to as sist in solving a
\

wide variety of mathematical, statistical, and engineering problems with
a minimum of programming effort by the user.

The AMTRAN system,

consisting of the language and special terminals with push- button input
and graphical output via storage scopes, was conceived by Dr. Robert Seitz
pf NASA, Marshall Space Flight Center, Huntsville, Alabama.

In; 1965-

1966, Dr. Seitz developed and implemented the AMTRAN system for
the IBM 1620.
Brown Engineering, a Teledyne Company, has developed an
abbreviated version of AMTRAN for the IBM 1130 under NASA contract
NAS8- 20415.

Although the system is designed to operate with special

terminals, it is felt that 1130 AMTRAN can be of considerable use when

o

operated from the console keyboard with the usual I/O devices.
The 1130 version of AMTRAN is vvritten primarily in FORTRAN
and operates under the disk monitor system.

AMTRAN provides the

following features:
• A declaration free working environment
• A set of fundamental operators with a well. defined syntax
• The capability for array arithmetic
• The ability to construct, store, and recall new operators
denoted as console programs
• Edit capabilities to allow easy modification of previously
defined console programs
• Dynamic core allocation for storage of data and console programs

o

• Choice of operation in either an immediate execution mode or
a suppressed mode.

2

SOFTWARE DESIGN

The AMTRAN software is designed around the operators presented
in Appendix A.

A principal objective in designing the AMTRAN software

for the IBM 1130 has been to optimally divide core between the system
software and the user storage areas, in order to maintain an effective
balance between execution speed and core available to the user.

To meet

this objective the program has been modularized into logical core loads
so that only the module or modules necessary for the corrlpletion of a
specific task need be in core at any particular tilne during execution of
the systen"l.

The modules of the system may be grouped into the classes:

system contol, incremental compiler, execution, and utility.

SYSTEM CONTROL
The primary functions of the control phase of the prograrn are to
initialize tbc system and to perform the bookkeeping necessary to indicate
which portions of the system are needed in core and what work areas are
available.

The control phase also guides the execution of the progralTI

through the various tasks required by each statelncnt.

INCREMEt'f'T AL TRANSLATION

Thi:, process begins with the input of a statelTIent.

The first

step of the proces s is to convert each character from the acceptable
set {O 1 2 3 4 5 6 7 8 9 space ABC D E F G H I JK L M N 0 P
Q R STU V W X Y Z

intege rOt 0 48.

*/

+ - ( ) • ,

&

% # $ } to a corresponding

Next the elements of the input statement are recogni zed

and replaced by a three digit integer which is used to classify the various
types of elements.

This is accomplished in the following manner:

• Console program names are replaced by a number in the range
1 )1 to 110 assigned on the order of first occurrence. The
c )nsole program name is stored in the called programs section
of the program construction area (see Appendix B).
3

c

1

1 '. .

o

n' rnnrtMTItT .

Fr--WEltJi .,.." u"-er .,'wrlo'E

e Delimiters and executable ope rators are replaced by predetermined
numbers in the range 201 through 274.
• Numeric constants are converted to floating point and stored in
the constants section of the program construction area. They
are replaced by a reference nwnber in the range 301 to 354
assigned on the order of first occurrence •
• User defined variable names are replaced by nwnbers from
401 to 429 assigned on the order of first occurrence. Variable
names are temporarily maintained in core for continuity between
statements.
The scanner performs extensive syntax checks as it moves through
the source statement.

In addition to the conversion and checking, system

delimiters are inserted and special formating is performed for several
of the operators.

The system- operators RESET, SUPPRESS, EXECUTE,

LIST, NAME, EDIT, DELETE, and SAVE are recognized and the systern
takes immediate action on these.

o

Parsing
Upon completion of a succes sful scan the resultant string is
released to the parser.

Parsing is the process by which the order of

execution is determined.

The priority of ope rations is as follows:

1.

operations within parenthesis

2.

functions (s uch as SIN, EXP, etc. )

3.

exponentiati on

4.

concatenation

5.

Inultiplication and division

6.

addition and subtraction

7.

replacement (=).

Parsing is carried out by working-with the .input string, a delimiter

•

list, and an output stack.

The procedure works as follows.

input symbol is a:

4

When the

---------------------------------------.~-

•

Vartable - place the variable in the output stack and examine the
next input s ytnbol.

C

Operator - examine the last symbol in the delimiter list and if
the result is
a.

an operator of lower priority

b.

a left parenthesis

c.

the delimjte r list is empty

then place the symbol in the delimiter list and examine the next
input symbol.
When none of the above conditions are satisfied, transfer the last
symbol from the delimiter list to the output stack and repeat
the proces s until a, b, or c is satisfied.
• Left parenthesis - place in the delimiter list and examine the
next input s ytnbol.
• Right parenthesis - transfer the symbols from the delimiter
list (last symbol ente red first) to the output stack until a left
parenthesis is encountered. Delete the left parenthesis and
continue to the next input s YlTlbol.
• Corruna - transfer the contents of the delinliter list (last symbol
entered first) to the output stack until a left parenthesis is
reached or the delimiter list is empty, then continue to the
next input s ytnbol.
• End of statement - transfer the contents of the delimiter list
(last symbol entered first) to the output stack. This will end
the parsing procedure with the results in output stack.
Coding
The output stack from the parser is released to the coder.

The

coder c-ollaspes the stack by cycling through it, replacing operators and
operands. and generating a sequence of LOAD, OPERATE, STORE, or
OPERATE. STORE commands.

This process continues until the entire

stack has been transformed into a macro langu,age.
macro instruction requires one 16- bit word.

In general each

The leading 7 bits specify

the operator and the remainder of the word specifies the variable.

5

c

'PPf""e,

.

"

WIllY",,'"

. WfZfrr

t

EXECUTION
Interpretation
This is the process of accessing a macro instruction, classifying
it and branching to the appropriate routine that performs the prescribed

operation.

There are two Inajor routines that handle the majority of

actual calculations.
Data Access and Execution
Data acces s is the proces s of dete rm.ining if the variable is defined,
its size, its location, and its cOInpatability with the operation requesting
it.

If the variable is defined and compatible, execution is perforITled;

otherwise, an error Inessage is typed out.
Storage Allocation
Storage allocation for data is a continuous process throughout

o

execution of the macro language instructions.

Storage for user variables,

temporary results, and the systeIn accumulator is provided in one data
area.

Storage for any data type is allocated only as needed and in the

exact amounts required during execution, is redimensioned as necessa ry,
and, in the case of teITlporary storage, is made available for further
use as soon as possible.
Access to free storage is maintained by a pointer scheme linking
together available blocks of contiguous storage locatio:1s in the data area.
When new space is required for the storage of data, the pointers are scanned
until the first block containing at least the space requested is encounte red.
The exact number of contiguous data words required is then rem.oved froITl
the free storage linkage.

If a block of the appropriate size is not encoun-

tered, an additional s can is Inade through the pointers, replacing the
sequential linkage to contiguous blocks with a single pointer link.

o

If this

process does not provide a large enough block, all data currently stored
in the data area is packed, providing one large block of available storage.

6

When a block of data is returned to free storage, the pointers are
adjusted and a pointer is added to include the additional block in the free
storage linkage.

o

UTILITY
The utility portion of the system consists of several subroutines
which provide the services of storing, listing, editing, or deleting ()f
console programs, and the deletion or retention of user defined variables.
The services in this catagory are initiated by the operators found under
SYSTEM OPERATORS in Appendix A.

c'
7

ERgi'

SAMPLE PROBLEMS

Presented here are three sample problems selected to illustrate
some of the capabilities of the system.

o

o
8

c

EXAMPLE 1.

ENTER PROC1RAn
1.
X=ARRAY -3,-1,8, Y=ARRAY 0,PI/3,2, N=O.
2•
REP EA T 3, P= Y SU8 rJ,
Z=3.28+X*X*X*COS P+l/COS P * EXPC-X/(X+7.45»,
DEGREES ) , Tf\ B2 CX, Z ) , N= rJ + 1 •

\.J RJ TE( P/

0.0000
-3.0000
-2.7500
-2.5000
-2.2500
-2.0000
-1. 75'00
-1.5000
-1.2500
-1.0000

-21.7576
-15.7217
-10.(;879
-6.5692

-3.27G6
-0.7200
1.1917
2.5502
3.4477

30.0000

-3.0000
-2.7500
-2.5000
-2.2500
-2.0000
-1.7500
-1.5000

-17.8367
-12.6577
-8.3382
8 0 L~ 7
-1.9816
0.2083
-l~ •

-1.~500

1.8429
3.0012

-1.0000

3.7623

60.0000
-3.0000
-2.7500
-2.5000
-2.2500
-2.0000
-1.7500
-1.5000
-1.2500
-1.0000

-6.2952
-3.5281
-1.2184
0.6675

2.1667

3.3190
4.1660

4.7502
5.1154

3.

o
9

'5

·bw.

Its '·srt'w,_·mtibH9

'Tn.".

EXA~1PLE

IT-

2.

THIS EXAMPLE PRESENTS AN OPERATOR PROGRAMMED IN AMTRAN TO PERFORM THE TRANSFORMATION DUE TO BILHARZ ON THE COEFFICIENTS OF A COr~PLEX POLYNOMIAL. THIS
PORCESS IS USEFUL IN STUDYING THE STABILITY PROPERTIES OF A SYSTEM OF LINEAR
DIFFERENTIAL EQUATIONS*.

1. X=Q,Y=R,M=2 INTERVALS X-2,A=O,L=O.
2. TYPEOUT GILHARZ ~1I\TRIX OF cnEFrICIE~!TS ••
3. \' J R I T EX, \.! R I T F. Y, A SUR 0 = Y S 11 r 0, r I == 1. •
4. F= - X StJ 8 0 / Y SUR 0, Z= L E F ( F * Y + X ) , \' I r~ I T r Z,
X=Y,Y=Z,A SUB tl==Y SUR O.
.
5 • I F tJ L E r'1 TH EN N=r1+ 1 , GOT 0 4 F. L S F F=0 , Y == 2 •
6. N=O,P=l.
7. REP E1\ T K, P== P * A SLJ R tJ, N== N+ 1 •
8, L SUR F=P,K=K+2,F=F+1.
9. IF F+1 LE INTERVALS Y THErl GO TO 6,
10. TYPEOUT ALPHA N FOr. N=l, 2, ••••
11. \.J R I TEA.
12. TYPEOUT PRODUCT OF ALPHA N. ~
1 3. ~'J R I TEL •
14. NJ\r1 E 8 I LH Z •

1. Y=X, Y SUB 0=0, Y=SHIFT(-l,Y).
2. NAME LEF.

*FOR THEORY RELATING TO THIS EXAMPLE, SEE "STABILITY THEOREMS FOR LINEAR
MOTIONS BY S. H. LEHNIGK, PRENTICE HALL, 1966.
II

o
10

GIVEN A COMPLEX POLYNOMIAL OF THE FORM

p

L

=

f(s)

(an + bni) sp-n

n=O

TO USE THE BILHZ OPERATOR WRITE OUT THE COEFFICIENTS AS FOLLOWS:

THEN ENTER BILHZ (Al, A2).

A1 = -bO' a1' b2' -a 3, -b 4 , .•.
A2 = aO' b" -a 2, -b 3, a4 , ....
HENCE FOR THE POLYNOMIAL
f(s)

= 2(1 - 3i) s4

+

(9 - 13i) s3 + 3(4 - 3;) 52

<"

+ 2(3 - i) s + ,.

......
......

ENTE r PROG RA~·1
1. LET X=6,9,-9,-6,0.
2. LET Y=2,-13,-12,2,1 •

3. BILHZ(X,Y).
BILHARZ MATRrX OF COEFFICIENTS.

6.0000
2.0000
48.0000
-14.1250
-12.0796
-5.9-121
-8.1691
2.8498
1.2216

9.0000
-13.0000
27.0000
-11.5000
-4.7788
1.6593
-1.6450
1.0000
0.0000

1.0000
0.0000
0.0000
0.0000

-6.0000
2.0000
-3.0000
1.0000
0.0000
0.0000
0.0000
0.0000
0.0000

0.0000
1.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000

-14.1250

-12.0795

-5.9121

-9.0000
-12.0000
-12.0000
2.1250
O~39S2

ALPHA tl FOR N=l, 2, •••

2.0000

48.0000

-8.1501

2.R4og

PRonUCT Or- ALPHA tJ.
0.~6000E

02

0.16379E 05

0.70109[ Of)

0.27539E 07

4.

~

~

C?

1.2216

hi' .. ·tt. ¥nbt#rit#¥ri***#

o

j .. , .... _, ..

-

- -hHH%t

.n

EXAMPLE 3.
THIS EXAMPLE WAS WRITTEN TO STUDY THE SMOOTHING EFFECTS OF THE PARAMETER
THE PROBABILITY DENSITY FUNCTION

0

m

f(x)

=

0(21T) 1/2

x

~

E

exp [-

(\:/i)2 ]

1. J=Ir-ITERVALS S+l,Y=INTERVALS GG+l,H=J*So.RT(2PI),L=O.
2. X=ARRAY -3,3,100,F=O.
3. Y=O*X,r.=GG SUR L,~l::O.
4. RF. PEA TJ, Y= Y+ EXP- { ( X- S S U8 N) * * 2 / ( 2G* G) ) , ~'I = N+ 1 ·
5. Y=Y/(H*G),B=MAX Y.
6.

o

IFF L T R TH EN F =B •

7. A=.5&O&6.5&F&2, PLOTS(X,Y,A),L=L+l.
8. I F L L T K TH EN r, a TO 3.

9. NAt:tF. PTRr-!.

EnTER PROGR/\rt
LET X=-.9G,-.88,-.36,-.40,O,.23,.44,.88
2•
LET S 1Gr1A = 0 • 1, 0 • 2, 0 • 5 , 1 •

1.

3.

NOTE:

·

PTR~!(x,SIGr1A).

OUTPUT FOR THIS EXAMPLE WAS DISPLAYED ON THE SCOPE.
OUTPUT WERE SHOWN AT THE PRESENTATION.

12.

SLIDES OF THE

ON

SYSTEM CONSTRAINTS

In the abbreviated version of AMTRAN currently operational
on the IBM 1130 with 8k of core, 1, 208 words are reserved for user
data storage.
point numbers.

Hence, a user is allowed to work with up to 604 floating

Up to 95 console programs may be defined and stored

on the disk, and 860 words are reserved for internal storage of console
programs.

A block of 450 words is reserved for the keyboard execution

and progralTI construction area.
of up to 209 characters.

An AMTRAN statement may consist

A total of 89 variables are allowed to be active

at anyone time and imbedding of console prograrns through ten levels
is allowed.
When constructing a console program or executing at the keyboard,
up to 45 statements are allowed per program.

However, the total length

of the source must not exceed 1, 140 characters and the macro langu.age
must not exceed 244 words.

Each console progranl ll1ay contain up to

29 user defined variables, 54 distinct constants (excluding the integers

o through

10), and may call up to 10 other prograrns.

13

#t"'

orT

I"T

n-I

CONCLUSION

o

AMTRAN (the IBM 1130 console keyboard version) has been
available on a limited basis to several employees in Brown Engineeri.ng
Research Laboratories since June 1968.

They have found it to be a

reliable and effective tool in evaluating and studying the behavior of
functions and in performing user controlled iterations.

Feedback from

these users has been invaluable in finalizing the 8k AMTRAN system.

o
14 .

APPENDIX A

SUMMARY OF AMTRAN OPERATORS FOR THE IBM 1130

C"

15

c

o

o

()

SUMMARY OF AMTRAN OPERATORS
BINARY OPERATIONS
~.Ql

* or implied

Operation

Operands

Results

Multiplication

All arithmetic operations listed can be performed between
constants, variables, or exoressions in the following
forms:

The results of array arithmetic are the
product, Quotient. etc. between corresoondinn elements of the operands.

I

'Division

+

Addition

1.

2.
3.

Subtraction
Exponentiation

**or POW

Concatenate

&

4.
NOTE:

1.

Two scalars
A scalar and an array with ~ elements
An array with ~ elements and a scalar
Two arrays with the same number of entries

2.

3.
4.

A scalar
An array with
An array with
An array with

~
~
~

elements
elements
elements

An error condition occurs when an arithmetic
operation is attempted between arrays of different
lengths.

Scalar constant(s}, variable(s), or expression(s) and/or
array variable{s) or exoression(s).

An array with the number of elements equal
to the sum of the number of elements in
each arqument.

.....
0'

MATHEMATICAL FUNCTIONS
Symbol

Function Performed

No. of
Arguments

Results 2

~

SIN

Trigonometric sine

Scalar or array (in radians)

COS

Trigonometric cosine

Scalar or array (in radians)

LN

Natural logarithm

Scalar or array

EXP

Argument power of e

Scalar or array

SQRT

Square root

Scalar or array

ATAN

Arctangent

Scalar or array

ASS

Absolute value

Scalar or array

TANH

Hyperbolic tangent

Scalar or array

SQ

Quantity squared

Scalar or array

Scalar or array the same size as argument
u'

" (in radians)

Scalar or array the same size as argument

(used on riqht of argument)
i

The ar~uments for all these functions may be either constants, variables, or expressions.

2

The results in the case of array arithmetic are the results of the function performed on each entry of the argument array.

SPECIAL FUNCTIONS FOR ARRAY MANIPULATION
Symbol

.....

Function Performed

No. of
Arguments

Results

~

ARRAY

Generation of an array

3

Scalar constants, variables or expressions

An array with the first element
equal to the first argument. the
last element equal to the second
argument, and the number of equal
sized increments specified by the
third arQument.

LETl

Generation of an array

n

Scalar constants, variables. and/or
expressions

An array with the n arguments as
elements.

SUB

Subscript modification

Scalar constant. variable, or expression

Scalar

SUB-THRU

Subscript

~odification

2

Scalar constants, variables,or expressions

Array

SUB LAST

Subscript modification

o

MIfI

Selection of the minimum element
of an array

Array variable or expression

Scalar

MAX

Selection of the maximum element
of an array

Array variable or expression

Scalar

SUM

Computation of the running summation of the elements in an array

Array variable or expression

SUMF

Computation of the total sum of the
elements of an array

Array variable or expreSSitg/

Scalar

SHIFT

Cyclic shift of the elements in
an array

Scalar constant, variable~ or expression
and array variable or expression

Array the same dimension as the
second argument with elements
shifted the number of places and
direction specified by the first
argument.

Scalar, 'last element of an array

-J

\1

,:'

I

~

The LET statement takes the form:

LET variable

2

. ":'~h'~"4'

Array of the same dimension as
the argument

argument(s) .

CJ

(1

•

o

o

LOGICAL OPERATORS AND TRANSFER STATEMENT
Symbol

.....

Description

IFl

Begins the IF statement used to control execution based on the relationshtp between two scalar quantities.

EQ, HE. LT, GT, LE GE

Relations available to the IF statement.

THEN

Denotes the beginning of the THEN clause associated with an IF statement.

ELSE

Denotes the beginning of the ELSE clause associated with an IF statement.

REPEAT

Causes the succeeding substatements of the line to be executed the number of times specified by the scalar
argument following the repeat. e.9. Repeat r, •••••

GO TO or GOTO

Causes execution to be transferred to the statement number indicated by the numeric argument following the
GO TO. In the EXECUTE mode, the GO TO may onl'y refer to a previously defined statement.

INOTE:

These represent :, ~,

<,

>, !. and

respectively.

(The ELSE clause is optional.)

The general form of the IF statement is:

00

IF 01 relation Q2 THEN ...•...•... ELSE
When the relation is satisfied the THEN clause is executed and the ELSE clause skipped.
and the ELSE clause executed~

When the relation fails the THEN clause is skipped

INPUT AND OUTPUT OPERATORS
~

SET

Description

AllowS numeric datatGbt read tn ff'OIIlttle cGnso1e printer (sense switch 15 off)

.0'"

the card reader

~senn

switch 15 on). A string of 'numeri.cconstants in free format is read into memory and aSSoCiated wit'hthe
variable name spedftedby the arqument. The variable assumeS the dimension of the input string. Tile numbers
may be either integers or dectmalnumbers in either fixed or floating pointfo""at and are separ.ated .by COflWTlclS
or blanks. The input ,strinc) is terminated by two consecutive slashes (j I).
WRITE

Causes tnevaJue(s} .of· the a.rgument to be -printed on the console printer (sense switch 0 Off) or on the 1132
printer (sense switc.b. Goo) infi~edpoint format (sense switch 1 off) or floating ooint format (sense switch
1 on).

PUNCH

Causes the value(s)::6f .theargumel'ltto.be punched 01'1 cards in fixed oointformatand to be printed on the console
printer (sense .$witCh,Ooff) or 01'\ the 113.2 printer (sense switch 0 on).

TYPEOUT

Causes the alph~iettifonnaHon folloWing the WOrd TYPEOUT to bepr:inted on the-console prtnter(sense
switch o off ) or on,t~ H32pri nter{ seni.e switch 0 on) .

LIST

Causes the sourcesta~ementsofthe specified console program or a descripUonof thespecifie'd intrinsic
operator to be printed ,on the console printer (sense switch 0 Off) Qron the 1132 printer (sense switch 0 on).
When the .cirgument iSilceons,oleprogramname and senSe switch 12 is on. the-console proqramwill also be
punched on. cards.

LIST All

Causes the names ofnvers~on

place while the

o

g

termin~l

previous character.
second.

Si.::·L28

is reading the next character

01'

takes

typing the

The terminal transfer rate is 10 c.hd.:cact ers per

)

---------------------

-----~~-

---1(,

RELIC'S SOFTWARE SYSTEM
In general, the monitor system within the "small computer" supervises
all data transfer and communications to its peripheral devicesQ
monitor system must read and store jops

RELIC'S

entered from the terminals.

send these jobs to the 1620 when they (both 1620 and job) are ready,
receive output from the 1620, and return it to the terminals

0

RELIC - MONITOR I

A.

CAPABILITIES:
Access
The three terminals are able to demand service. "a.lmost" s imulr"
taneously by utilizing the interrupt controlled liD capability
of the "small computer".

This permits three users at three

remote terminals to enter three different jobs

~oncurrently~

Job Storage
Our first approach in the development of the HELIe system was
to proceed as simply as possible and to avoid maior complications
and changes to the 1620 Monitor I systemo

We,

tl~efore,

the easiest and most pragmatic method to implement this
The 8K byte (each byte contains 8 bits)
computer" was partitionedo
(105K) were allocated to

storage.

memo::c'y (If:

One thousand and fi
each of the three

c

t~ae

chose
sy~temo

"small

hundred bytes

t(;~rninals

for job

This would permit seventy-five FORTRAN statements (20

characters per statement) per

terminal~

One thousand bytes were

o

11

(),.;

allocated for a common output area for data being returne.d from
the 16200

Approximately lK bytes were needed for the RELIC monitor

system which remains resident in memory, and an additional lo5K bytes
still remain for a fourth terminal if added o

JOB ROUTING
Jobs are entered remotely from the terminals either through the
keyboard or from punched paper tapeo

The first job to be completely
s~nt

read. including program and data, will be the first job
1620 fur execution, when the 1620 is ready to accept ito
is stored in its own terminal input area o

to the

Each job

When the 1620 has read

and has begun execution of the Job, the two remaining terminals are

o

still able to enter and complete their jobs o
is returned to

d.

Output from the 1620

common output area shared by all t:he ter'minals (one

job at a time ) within the
"waiting" terminal.

1f

small computer"

~

and t'hen back: to the

Upon completion of the job return, the 1620

will be ready to accept the next job when ready i.n the
puter"

0

Iff

small com-

(See Job Routing Diagsll next page)

TURN-AROUND TIME
We expect turn-around time per job to be about one to tv.JO minutes
from the beginning of 1620 execution to the outputting of results
to the terminal"

•
&&&===k-~·~

REHOTE JOB ROUTING AND STORAGE

nSmall Computer»

Terminals

~.c~'_'..~.~~~~~

).t=;:=====- .

~
~

#1

RELIC - Monitor I

~ __
,:"-.

~~-7--=_=====:

Input

Buffe~

for 11

-

1========~~====·~---

Input Buffer for

B2

ti2

Input Buffer for • 3

#3

Green line - Input to amall compu·
RGO line - Output to ter·r,1inals..
IBM 1620 Data Processing System

c·

i" 1""

j"!!!

r """' ", '"" !I" "". WIlt. [I"! 11

T"-

t""B'rtt

/f
B.

TERMINAL JOB
Job Format
Jobs which are sent in the 1620 from the "sT!l.all computer n follow
the same format as any job entered under the IBM 1620 Monitor
I systemo

The first card read by the 1620 fron

"sma.Il

t:h,e

computer" will be a job card, the second will be a 1:1oni tor Con,=
trol Card (e.go FORX or XEQ), followed by the pro gran andlor
data, and an END OF JOB record (##¢#) at the end of the data o
Job Execution
RELIC - Monitor I protects the 1620 by setting up the }620
..

o

Monitor Control Cards internally in the input area for the ter=
minal or terminals being serviced o

If a FORTRAN lID program is

to be executed, each FORTRAN statement is entered in free format,
one line at a time, and a 1620 paper tape end of line character
is placed at the end of each line when the retu't"n key is depressed
at the terminalo

Wh~n

the FORTRAN "END" statCr1lc.nt is read, the

message "ENTER DATA" is typed, and all the

CLlt,J.

rnl);;:;t

be entered o

If the job is to be a 1620 disk-stored program execution (XEQ),
then upon receipt of the program CALL NAME, the
DATA" is typed

0

messa~e

~BNTER

When all the data has been ent:ered ~ "the job is

sent to the 1620 for executiono
The 1620 will read the job one line at a time causing an interrupt
each time it is ready to read a

o

nevJ

line o

\IJrd,lE;

the, 1620 is ex ...

ecuting the job, the terminal being serviced is blocked from entering any data, unless it is requested by the 1620

0

Job Execution continued
Again, while the 1620 is processing each

line~

the Hsmall computer"

c

is free to accept job entry from the terminals not belng serviced by
t.he 1620

0

When the 1620 has completed compilation of the program, the

data, if there is any, will be read o

Output is sent from the 1620 on

interrupt request to the 1620 output area in the "small computer" One
line at a time o

This output is returned to the terminal being serviced

on an interrupt request between job processing and 1620 c.ommunicationQ
Job Permitted Under RELIC - Monitor I
RELIC permits a wide range of job typeso

Jobs may be entered remotely

for execution under the 1620 Monitor I system or for execution in the
"small computer"c
1620 Monitor I JOBS:

FORTRAN II D(FORX)
FORTRAN Program compilation and execution will be the
major types of jobs for 1620 executiono

The proper

Job Card and FORTRAN Control cards will be constructed
by the RELIC - Monitor Systemo The FORTRAN program is
entered in free format from the terminal and

18

held

i1) the memory of the "small computer" until the data is
read o

When the data is read, the job is 'sent to the 1620

for executiono

Any e:r;ror messages which occur "-7i11 b,e

sent by the 1620 back to the terminal

0

'
C
-

'jl

reWT

()

'flit 5¥%¥

EXECUTE JOBS(XEQ)
Disk-storea program execution is possible by entering
the XEQ Job Specification Record and the proper program CALL NAMEo

From this information, the proper

Monitor Control cards are establishedG

The data is

read and the job sent to the 1620 for executiono

SPS II D
1620 Symbolic Programming System jobs will be allowed,
but only by the more competent usero
"Small Comou"[;er"
Job Execution:
t
RELIC will provide the capability for the "small computer" system
programs to be called for execution within the "small computer"

()J

itselfo

These system programs will be stored on the 1311 Disk

Storage Drive of the 1620 by means of a special conversion routine.
An XEQ job will first be sent to the 1620 to write the program into
the "small computer's" memory and turn control to that program.
This will permit the execution of the following programs:

o

10

An Ineractive (or Conversational) FORTRAN

20

Assembly Language (with third generation capability)

30

Production Programs

40

Loaders

5.

Test r.rograms

.-._----_...-.. _..-....._.... _ - -

MODIFICATIONS TO THE 1620 MONITOR I
Modifications to. the 1620 Monitor I system are minor and consist of
prohibiting the terminal job from using the 1620 peripheral. devices
other than the paper tape channel.

This is achieved by overlaying

Supervisor's 1/0 constants in the Monitor 1/0 routines.

When a job

card is read, the digit in column seven (which indicates the input
device) is testedo

If that digit is a three (##JOB 3), the job being

executed is sent from the "small computer"

0

To jnsure that the job

output is returned to the terminal and not to the 1443 ?rinter, 1622
Card Punch, or 1620 Typewriter, their 1/0 constants are overlayed to
make them appear as the paper tape channelo

(eo g. a FORTRAN Print

instrtiction will cause data to be written on to the paper tape channel
and not to the 1443

0

)

The new set of constants designed to permit

only communication between the 1620 and the paper tape channel are
stored in Supervisor on the disk over the normal 1/0 constants for
all terminal jobso

The regular I/O constants are replaced for jobs

entered from the Card Reader cr TypewriterQ
At the completion of a terminal job, control will be returned to
Supervisor to read the next job from paper tape (channel).

Thus,

the call exit status for the 1620 will be an alphameric paper tape
read under Monitor 10

This will make it possible for the "small

computer" to send another terminal job to the 1620 while it is idle
and without an operator's intervention.

c

-r

o

NEW MONITOR LOADER CARD
A new monitor loader card had to be designed to return job entry
to the Card Reader if the 1620 was in an 1/0 hold for the paper
tape channel

0

ThE new monitor loader card transmits the 1/0

constant to enable the Monitor system to read alphamerically from
the Card

Reader~

in Supervisoro)

(A

~7

is changed to a

~7

at location 01761 in

When the job card is read from the card reader,

the proper IIO constants are restoredo

THE NEW MONITOR LOADER CARD FOR RELIC
Location

o

•

00000
00012
0002'4
00036
00044

Instruction
34 00044 00701
36 00044 00702
15 01761 5000b"
49 02402 0
IT9636'f1300102

Comment
Seek to cyl. for Supervisor
Read Supervisor from disk.
Enable Alpha Card Read
Branch to Superviso~.
DDA for Supervisor

.... _ifOOMI ............,...'...'_...... IJojo _

_ _ ~tilo:/I'j!t~\~*.....,·'···

OJ

APPENDIX A

INTERFACE SPECIFICATIONS

B.q. fI::.
(J

J

.....

-8 C

I

X0.1.
p/Jr)

I.J

COHUAND

-=-----IAl -

-

...

.,....

~1Il

o ...X-

-=

r

--...

J

~

iA

FOR 1620 and the SMALL COMPUTER

(/~10

16 ~~ 0 fttpe.
IbtOe L-iNeS

._••• "",,... ,"., . . . . . . . . ,. . . . . ~

"

---~

..
-.
-

-

Lll~ES

COMMPJN 0 oR
S7t9ru.s eyrE"

~\

J.

~

f, tI

H

j.

~,-.--.--,--~--,~
~>-----.~-.----.----

4 - - -..-.------.-.--.~---

__________

~.J

_~

_ _ _ _. _ . _ _ _ • _ _ _ _ •_ _ _ _ R" _ _ _ _ _. _

DATA TRAHSFER RATE - 2,OOOCPS
The 1620 parity checking is utnized by staying within the 1620
character set.

7:" , '"",;i

APPENDIX B
ASCII - 1620 PAPER TAPE CODE CONVERSION TABLES IN HEXADECIMAL

0:

Character

1620 paper
tape code

(7-bit)
ASCII

86
46
CE
26
AE
6E
E6
16
9E
8A
4A
C2
2A
A2
62
EA
1A
92
4C
C4
2C
A4
64
EC,
lC
94

41
42
43
44
45
46
47
48
49
4A
4B
4C
4D
4E
4F
50
51
52
53
54
55
56
57
58

A
B
C
D
E
F
G
H

I
J
K

L
M
N
0

P
Q
R
S
T

0

U

V
W
X

Y
Z

51
5

Character

1620 paper
tape code

0
1
2
3
4
5
6
7
8
9
blank

04
80
40
C8
20
A8
68
EO
10
98
08
D6
)
3E
+
OE
$
DA
~'c
32
02
8C
/
,
DC
(
8C
DO
=
@
38
¢
(recmk) 54
LF
08
CR (EL)
01
end of line

30
31
32
33
34
35
36
37
38
39
20
2E
29
2B
24
2A
2D
2F
2C
28

0

line feed
carriage
return

(7-bit)
ASCII

3D
4'0

""'OA
OD

"CORE-IMAGE/t" NUHERICAL CONVERSION TABLE - (1620 hexadecimal representation)
"Core-Image"(sma11 computer)
Numeric Character(hexadecimal)

o
1
2
3
4
5
6

.,

G

9

A
B

o

C

1620 Paper
Tape Code
04
80
40
C8
20
A8
68
EO
10
98
02
8A
4A

1620 "Core-Image"
Representation

o
1

2
3
4
5
6
7
8
9

Qor
1 or
'1 or
1" or

J
K

L

(

"~~"'-."'"

I

Ii

o

"CORE-IMAGE NUMERICAL CONVERSION TABLE"
1620 Paper
Tape Code

Numeric
Character

C2
2A
A2

D
E

F

1620 Core
Representation

'! or L
'4 or M
~. or N

STORAGE OF "small computer" CORE-IMAGE PROGRAMS ON THE 1311
The "small computer" uses a four-bit character structure which
permits the hexadecimal number set (O,1,2,oov,9,A,B,C,D, and F)o
Each byte of the "small computer" contains two hexadecimal characterso

The 1620 utilizes a six-bit data structure (C,F,8,4,2,l)o

The hex

number set may be represented within the 1620 by using the above
tableo

The decimal numbers remain the same with the 6 highest hex

numbers (A, B, C, D, F) being represented in the 1620 as flagged
digits

(~,Tt~,~,4t~)e

Thus, by means of a simple conversion routine, numerical data can
be transferred between the 1620 and the small computer at a
rapid rateo

vel~

Also, the "small computer" system programs can be

stored on the 1311 Disk Pack and they may be called when needed.

c·

c

_ ........,. .-..c._c""r ........

t

•.;. ......

~'.'~Io.W.... u."C." ..,_:

...._

t ....

'0' . .,./ ..... ; •.•~ ." -.

~

.,h ....

~.,

" . . . . . ,,". :.:, ...J. .'.'

~'.- .,~

·"· .. ·1 • .."... ."'\ ....

t'~ .........

DE"vICE READY
TYPE USER NlJt.1Bf::R 1 70ne '/
TYPE PROJ NUMBEH '/ GIn l ~:

TYPE JOB SPECIFtCATION RECORD

FGl:.:~:

ENTP~ PROGR~I

i., ;:-J
1 Fcn!\\(!~\T{r3)
L. EA D :2,6": A~ ~ : ~ J. t~:! j ~ ~J )
:? FO Hl"if~T ~ Ii' ~ O·J:~sk~

PN.OGRAHHING REQUIRED
Two additional items of information were required in the Job Card,
a user number, which identifies the user, and a project number,
which identifies the job or project being executed
Four programs were written to implement the

Ie

o

system~

A program to further analyze the Job Card,
control job execution, and update tablesa

c

o·

To ca.I.l trlis pX'ogrr3.n modi.fications to H':)ni 1':.0:0 11 s

Supervisor were necessary; however, the Monitor
Contr.ol Record Analyzer (HeRA) in. Supe:('\i'isor undex'-.:·

went only minor modific.at ions
')
,~..

{)

0

A pro gran to interrogate these tables to provide
infor.mation and listings on all 16'20

1)~;;I~

di..:tring

a specified periodu
.3"

A program to edit, de.lete
nu~bers

4

~

ente:t"'~

an.d/ or. List

in the tableJ

A program to enter the date into an aVdilable
."
'.
,oca:~on
2n supervlsoro
t
1

o

A more extensive description of those programs will
follow in this report

0

DESCRIPTION OF USER NUl'1BER AND PROJECT NU11BERS

We are able to monitor rontrol and provide statistics on
use under users and proJects by using a special
numb er

0

.,

•

casslt~catl,on
1

,J...

~nd proje~t

USEr

system~

Every person using the

Hall UniversIty is g :;

~oc~ut~r

fa/::,ilitie~:i

O?H"\
V ... A'

:.1

"'-4

q.:;:
f!~
\'~""'_'

r'

a.-I: th.e Compl.J,ter

i-.:
'Y~'
.~.t;. II.....p·lh,:-,.
l,04"'-~.

:\l~i....(~).

~

Ce~'Jte.c

.=. :;:.,::::.
.; '" np
t..;:.J..~
..•• _ . d

l..,."'\ ......

,:1
...

at Seton

.L1"Jr'-~1
.........J

pet
...... - .

number which authoJ:'-l'l.e.8 tha user to use 1:he c.oril'putef' for his

particular

o

project~

o

USER NUMBER
For each user number there are five

categories~

Faculty (1)5)

Administrative Staff (2), Undergraduate full-time students (3),
Undergraduate part-time (4) students, and Graduate Students (5)0
Each category is represented by a single digit 1 through 5
ivelyo

re~ryer~=

Each user number is a six-digit number and has the structure:

CSDDNN where the first digit
the second digit (S) the

ee)

school~

represents one of the five categories,
the third and fourth digits (DD) the

department under the school, and the last two digits the (NN) sequence
numbero
EXAMPLE - 312801

indicates an undergraduate full-time
student (3), in the School of Arts
and Science (l)~ in Mathematics Department (28), and is the first student
assigned this number (01)0

c

PROJECT NUHBERS
For

eac~

project number there are nine categories: Arts and Science (1),
(5)~

Business (2), Divinity (3), Education (4), Medical

Computer Center (7), Administrative (8), and Other (9)0
is represented by a single digit 1 through 9

ee)

~

(6)~

Each category

respectively~

number is a six-digit number and has the structupe
first digit

Dental

Each project

CDDtHHJ tvhe,re the

is one of the nine categories, the second and third

digits (DD) the department, and the last three digits (NNN) the sequence
number or class number o
EXAMPLE - 128181

indicates the project is in the School of
~ts and Science (1), under the Mathematics
Department (28), for the Mt 181 class (181)0

This user and project classification system provided an easy structure
from which to design tables o

o

hr-- :rtllrll--

o

DESCRIPTION OF TABLES FOR 1620 USE FOR PROJECT AND USER NUMBERS
The files for storing the frequency of 1620 use are separated into
three division (1), USER NUMBERS (non-computer center), (2) PROJECT
NUMBERS, and (3) COMPUTER CENTER USERS WITH PROJECT NUMBERS

0

The files for Divisions 1 and 2 have the capa.city for 500 user numbers,
100 for each of the 5 user divisions and 500 project numbers, 50 for

each of the nine divisionso

The format for the user number and project

number file is a six-digit user or project number followed by a threedigit weekly count, a three-digit monthly count, a four-digit year-todate count, and a four-digit (first four digits of date)o
EXAMPLE: UUuuUUtVWWHMM?'YY1)DDD WHERE VUUUUU is the user or project number
(each high order position is flagged) WWW

is the weekly count
is the monthly count
YYYY is the year-to-date
IDDD is the date of last useo

.l'·IMH

o

10

USER NUMBERS are stored on cylinder 18, sector addresses 09000-09099
Each user number and its related fields occupies 20 core locations,
and the entire file for a particular user number category occupies
20 sectorso

There are 5 categories under which the user numbers

are classified o
USER CATEGORIES
FACULTY
ADMINISTRATIVE
FULL-TIME UNDERGRADo
PART-TINE UNDERGRADo
GRADUATE STUDENT
20

•

FIRST DIGIT
1
2
3
4
5

SECTOR ADDRESSES
09000~09019

09020-09039
09040-=09059
09060=09079

09080-09099

PROJECT NUMBERS are stored between sector addresses 09100 and 09199

0

0

.~~-----.-.-.-,

....-"'-, . .

...

"."'.".,-.,,-~~

---.-.-.",~,~~'.=~

Each project number and its related fields occupes 20 core locations
and the entire file for a project number category occupies 10 sectors o
There are nine categories under which a project number can be classified o

PROJECT CATEGORIES

FIRST DIGIT

ARTS AND SCIENCE
BUSINESS
DIVINITY
EDUCATION
MEDICAL
DENTAL
COMPUTER CENTER
ADMINISTRATIVE
OTHER

SECTOR ADDRESSES
09100-091'19
09110-09119
09120-09129
09130-091J9
09140-09149
09150-09159
09160-09169
09170-09179
09180-09189

1
2
3
4
5
6
7
8
9

Sector addresses 09190-09199 are not used and are availableo

3

0

COMPUTER CENTER USER NUMBERS
All Computer Center user numbers, both academic and administrative
staff, are placed into a different file then those described on
the previous

page~

These tiumbers are stored in cylinder 19

0

c

Both

the academic and administrative user numbers contain space for 10
users with 20 project numbers under each usere

The academic staff

for the Computer Center will have a "17" as the first two digits
of the number and these numbers are stored between sector address
09200 and 09239

0

The file for this category occupies 40 sectorso

The administrative Computer Center user numbers are stored between
sector address 09240 and 09279

0

The user number again is a six-digit number and will appear as the
first number of the file followed by a project number, a three=digit
weekly count, a three-digit monthly, a four-digit yearly and a fourdigit date, followed by another project number, etc o

c

o

PROGRAM DESCRIPTIONS
JOBCD-PROGRAM TO FURTHE-R ANALYZE THE JOB CAED

.....~~:t'\·~~'='·~~Y'!~''''.;:~;--;·J_.'''·L.ftL.)::''·.o:.~:.;!:.:. ..:;._.~' ...'. '~~~~..t.. ...,.; ..... ~

iiodification to the HeRA stores Supervisor on the work cylinders at
sector address 01400 and calls the program to further analyze the
Job Cardo

JOB CARD STRIPPED
The contents of the job card which had been read by Supervisor is found

in memory locations 13000 to

13159~

Supervisorfis Transfer Numeric

Strip Routines at 08146 (version 2) is used to strip the numerical digits
of the user number and project number on the job cardo

The calling

sequence for this routine is described on pagelS of this papero
Once the user and proj ect numbers have been str1i.pped and re"",constructed

o

they are tested for validityo

If the numbers are found to be invalid,

an error message is typed and 10 seconds is allowed for entry of the
correct numbero

If the correct number is not ente.red within. 10 seconds,

the job is abandoned and control is returned to Supervisor ,to read
another job cardo

NON-COMPUTER CENTER USER NUMBER PROCEDURE
User numbers of non-Computer Center Users are processed before the project
number~

-

The user's category of the user file is read from the disk into

memory, and a "straight look-up" method is use.d to search for the user
number in the table called o

If the user number punched on the Job Card

does not match a number in the table, "USER NUl'1BER NOT AUTHORIZED H , will
be typed and the job abandoned o

o

When the user number has been found in

the table, a unit is added to the weekly count and the first four digits
of the date are entered o

The tables are then returned to the disk and

the same procedure is followed for the project number o

o

COMPUTER CENTER STAFF PROCEDURE
Since additional information is required for the Computer Center staff,
their user numbers are processed somewhat differentlyo
number on their job cards is updated firsto

The project

This is necessary to test

for the validity of the project number, since when a particular Computer
Center user's project number is not found under his user number it is
entered automaticallyo

Thus, if we test the project number

first~

we

are assured it is authorized if it is entered under the user number o
Therefore, a unit of use and the date are entered in the project number
field under the Computer Center user

number~

The fields under the project

number in the project number division are also updated to provide a total
use for that project regardless of the usero

ADDITIONAL TIME SPENT
The time

sp~nt

in logging a unit of use and the date under each user

o

number and project number adds approximately one and one=half seconds
to two and one-half seconds to the Job Card Routine o

JOB CARD SUMMARY SHEET OPTION
When a digit is punched on the job card between the user number and the
project number (cole 38), the job card summary sheet option is executed o
This option will print a header sheet identifying the job being run
and contains the Seton Hall University Computer Center heading 9 the date,
user and project numbers, and other descriptive information furnished
by the job

card~

(see pagel7 )

c

SUPERVISOR RETURNED
When the user and project numbers have been updated andlor the Job
Card Summary Sheet option executed, Supervisor is read back from
the work cylinders and control is returned to the Monitor Control
Record Analyzer and the next control card· is read o
OTHER APPLICATIONS
This program will be used to control and inhibit certain user's
programs from using one of the 1620 peripheral devices by overlaying
Monitor I's 1/0 constant~o

Thus, any job which has access to our

1620 from our remote terminal system (RELIC) can only return output
back to the terminals, and not to the printer, typewriter, or card
puncho

o

This program can be furt'her used to control sched~ling of jobs and
select the input device :which is ready to send

o

Lastly, the program provides the possibility of writing messages and
instructions to a certain user or users

•

o

o

,,"11=

DESCRIPTION OF SUPPORTING PROGRAMS
Ao

WEEKLY - MONTHLY - YEARLY - Interrogates frequency tables for
listingso
The purpose of this program is to interrogate the tables Cstored
on the disk and updated by JOBCD each time a job card is read)
and punch a deck of cards from which a weekly,

monthly~

or year-

to-date reports can be listed describing the frequency of use
for the 1620 by users and projects for the period specified o
The report is divided into three divisions:
10

A listing by non-Computer Center user numberso
will appear as the first page of the reporto
use for users under the five categorieso

This lifting
It will contain

If there is no use

for the time specified for a particular user number or

category~

the number or category will not be listed

number~s

0

Only those

c

for which there has been use appears in the listingo
20

A listing of all proj ect numbers

0

This sE;c.tion contains project

numbers und-er the nine categories!p and only those for which
there has been use during the period specifiedo
3

0

A listing of Computer Center user' numbers \.;ith proj ect numbers
under usero
listingo

This report will appear as the last section of the

The user numbers are classified under

Acad(~mic

and

Administrative staff with their projects listed under the user
number o
WEEKLY - lists all weekly use and resets the weekly fields
(run every week~
MONTHLY- lists all monthly use and resets
(run every month)

tl:\(2~

monthly fields

YEARLY - lists all year-to-date use and reset all fields
(run every year)

c

o

=12=

B.

EDTBLE Edit disk-stored frequency tableso
The purpose of this

progra~

is to provide the following four

capabilities:
1.

To enter new user and project numbers into the tables

2~

to delete older user and project numbers from the tables

3.

to edit or correct frequency counts to the part of the
table or tables that need correcting or that get "wiped-outfto

4.

To provide listings of all use.r numbers and project numbers
which are stored in the tables, and to list them under their
proper categoryo

Co

UPDATE

program executed each day to store the date o

The purpose of this program is to pead and store the current

o

date on the disk in the supervisor program so that it is available
for use at any time by other prop.;:rar:ls
date (month, day, and last digit

0

f

0

t ~H~

The five digits of the
year) are stored in the

supervisor at memory locaton 01303 enq 01307
containing 'the date is 19648

0

The sector address

If thi3 sector is read into core

location XXXXX, the date would be foun(1
at XXXXX+lo

0

dt

XXXXX+ 5 with a flag

This program was written by Mrso Jud!th So Heyman,

a former member of our staffo
EXECUTION AND EXPERIENCE
Two years of experience operating under-' this
its usefulness to our installationo

s~em

We are now able to control who

uses and for what purpose our computer is used o

o

has more ,than proved

We are able to document how the computer is being used which

16

o

of "some" importance to the administl'ations VJ:ho provide the fiscal
budgeto

Also by evaluating past use we have been able to better

predict and prepare for increases in future use o

o

c

It"

W'···UIUI"Uf'W

11

**

t

tit

st.

*# ..

r!

.

'····Kt·t±W··'P··-lyJ··]J"·, . ["5e

"Y .... '

SPECIFICATIONS

c

HODIFICATIONS TO SUPERVISOR (!-iONITOR I VERSION 2)
In order to further analyze the Job Card the following modifications
have been made to the Supervisor Prograno
10

A NOP (41) has been placed at core locations 05238 and 05239
to avoid transmitting a digit into 060940
EXAMPLE: 05238 41 06094 02851

20

An area of Supervisor 06086 -

96169 is used when there are more

than one disk storage fileso

Since we have only the single disk

file, this area is not used by the

o

Supervisor~

Therefore, the

instructions listed below are placed in that area o

This routine

is used to temporarily stoY'e the Supervisor on the work cylinders
and call the Job Card progran to further analyze the Job Carda

A

06086 3 Lj. 06142 00701

HDN A

0609c3 38 U6142 00702

TI\

IBMMOD,B

06110 31 OU440 06156

BI

MONOCAL, 19

061L2

46 00796 01900

B

OVERLAY

0613L~

49 00466 00000

HJB 20 SK

o

A

DDA ,1,01400,173,02302

06142 1"0140017702302

B

DDA ,1,SSSSS,CCC,02402

16156 1SSSSSCCCU2402

DC

16170

1,#

~

o
where SSSSS is the sector address of the program "JOB-CP", and CCC
is the number of sectors used by JOBCDo
The

Sect~~

Addresses where this sectiDn of Supervisor is stored is

between 19687-196980
ROUTINE TO EXTRACT NUMERICAL STRIP
The routine to extract numerical strip from the Job Card in Read
area (13000-13159) is at core location 08146

0

The Calling Sequence
TFM 09800
NC1-IAl\
BTM 08146
ADDR
Return from routine TF WORK , 09811
NCHAR are the number of characters
ADDR

is the odd address (high order) of field to be stripped

09811 contains the stripped numerical field o

WI

r·"

_On

..

-16~-

c

JOB CARD FORMAT
Each Job Card entered under the Monitor I System for the IBM l62Q at
The Computer Center, Seton Hall University must contain a user number
and project number according to the specifications belowo

Additional

information may be punched on the job card to identify the job or print
a Job Summary Sheet

SPECIFICATION

COLUMN

()

1

-

2

3

-

5

JOB

6

BLANK

7
37

1 FOR TYPEWRITER, 3 FOR PAPER TAPE OR 5 for
CARD INPUT 1
USER NUMBER

44

45

-

BLAN.K (l wil'l print a Job Card Summary sheet
title page
PROJECT NUMBER

47

BLANK

48

-

72

IDENTIFICATION

32

-

38
39

BLANK

73
7;4-

80

o

0

-

79

JOB TYPE
BLANK

SETON HALL UNIVERSITY

o

COt\1PUTER -CENTER

09/03/68

**JOB CARD

SUM~ARY

c
USER NUMBEP.-

110001

P.ROJ ECT NU.,lBER-

700001

IDE NT! I: I CAT ION.

SUM~1ARV

~JOa

TYP6.

SPSII

SHEET ~AGE OPT ION

.•

!

-;,fifflibtW·l""TY'ftTn"F31"Y!1tihWHWrr· ... -

- 'p"'-'

o
EXAMPLES OF TYPED MESSAGES FOR INVALID USER OR PROJECT NUMBERS.

155014 700Q101

**JOB 5
USER NUMBER NOT AUTHOR I ZED c
END OF JOB

170007 300001

**J08 5

PROJECT NUMBER NOT AUTHORIZEO o
EtID Of JOB

o

**'JOB 5

INVALID USER NUHBER o TO TYPE NUMBER SWl ONo
1 700B7R5
PLEASE TURN OFF SW 1 THANK YOU o

I NVAL ID PROJECT NUHBtR D TO TyPE NUMBER SWl ONo
701011.RS

**JOB 5
iHVAt~O

USER NUMBER o TO TVPE NUMBER SWl ONo

EttD OF JOB

o

- - - ---------

REPORT FOR 1620 USE FOR THE WEEK 08 30 68

USER NUM~~S~--------------------+/~l~

UNDERGRADUATE FULL TIME STUDENT
-----NUMBER
WEEKLY
TOTAL
DATE LAST USED
-------------------------.-----------~-----------------~--~~-------312805
..

WEEKLY TOTAL

---_._--------_. __.__._--

065

00065- - - - - -

0193

08 28 63

REPORT FOR 1620 USE FOR THE WEEK 08 30 68

. --_. ----- --.----.------.----- - . - - - - - - - - - - . - - - - - - - - - - - - - . - . - - - . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 1 1

_ 0 ____... ________...

----.... -- -..-.-.-..-..- - - - - . - - - - - - - - - - - _ . _ - - - - - - - - - - - - - - - - - - - - 1 1

FACULTY
-

----

-

_. -

--. __

._-_

USER
NO.· 170007
..
_- _._- - _._--_._----_.- --

-

...

.

.. _--

__

-_._- ---_._---_._---------- -_._. __ . - - ------- .. ---- -_._----_.._-_._---_._._-_._- _._._- ._---_._..•. _----_._-_. . _ . _ . _ - - - - - - - - - - - -

---_ .. _--_ ..

---_._. __ ._------_.__._---

----------------_._ ... - ._--_.-_._----_. -_._---_.. _-------------_._-------------------

NUMBER
WEEKLY- - - - - TOTAL
USED
---.-.--- . . .-.---.-.--------.---------------------- - - - - - - - -DATE
- - -LAST
=-----1
700001

002

0026

08 30 68

---- -- --- -.-------- -------- -- --TOTal r .- -----·-----------00-5------------------0-01-3-------------·---0[----]
700002

002

0026

:::--~--3-----·---·--···-

08 30 68

----------------- --------------------7uIo-r2-------------------uOS----------lY03-cr-----------OS-30-6S------------------ . ------- -- wFEl(CYTTITAI-on 0 14
_.. _--_.--_ .. - ...- .. -._---_ .. _--_..

_-_ ...-----_._------- . _ - - - - - - - - - _ . _ - - - - - - - - - - - - - - - - - - - _ . _ - - - - - - - - - - - - - - - - - - - -

USER ._--_._
NO._-_._.--_ 170008
---_._--_._-------------------------------------------------_.__

...

..

NO WEEKLY USE
--0 S--ER -N O~--1T(rooT------- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

---0

USER NO.

170002

---.. -.-.-.--..-.. - .. ----....- - - - - - - . - - - . - - - - - - . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 1

NO WEEKLY USE

-.--.-----..-------- ._---_ . _ - - - - - - - _ . _ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 1 1

----------- US ER Nlr.

-_._----_ .. -._------_._---_.. _._------------_.

170010

__. _ - - - - - - - - - - - - - - _ . _ - - - - - - - - - - - - - - - _ . _ - - - - - - - - - - - - - -

USER NO. 170009
--.------.--------------------------------------------1
NO WEEKLY US~

ADMINISTRATIVE STAFF
USER NO.

270001

NO WEEKLY USE
USER NO.

270006

O----------------""'N-,.,-U"7-0M
-=-=B=-=E=-=R=-----------,-W'-=Ec-=E-:-:-K-:--L--:-"":Y----=-rO=-T=-A-:--Lo-----------=O=--A=-=T=-=E::---:-L---=A---=Sc-=T::--U:-::S:-::E:-::D-----;

216434
830001

WEEKLY TOTAL

00011

006
005

0010
0005

08 28 68
08 23 68

,----------~

---.-.---------.-.-------.---------..-------- REPORT

...-.

FOR 1620 USE FOR THE WEEK
30- 68
- - -08- -------2-t:r
PROJ ECT NU-MBERS

I

-

. . - - - - - - -. . - - - - . - .

--

I

I

BUSINESS

...-.._....__•.. -......._ - - , - - -

NUMBER

WEEKLY

216434

006

WEEKLY TOTAL

TOTAL
0010

DATE LAST USED
08 28 68

00006

---- .. _ ....._-- ..... _----_ ...._.. ---_ ... _---_ .... _.. _.._...._ . _ - - - - _ . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

COMPUTER CENTER
NUMBER

WEEKLY

700001

027

------.......·----·---------·------·---,0000

TOTAL
0051
0215
0037
0041
0001
0030

DATE LAST USED
08
08
08
08
08
08

30
30
26
30
22
30

68
68
68
68
68
68

2------0=-6..-9=--------,;;~-;;:;------=--=--=-=---=--=---------

701006
10 lOT!"
702010
101012

006
011
001
005

C--'

ADMINISTRATION
NUMBER
83UUOI
-----------

820005

WEEKLY
b05
018

TOTAL
0005
0018

DATE LAST USED
08 23 68
08 20 68

-------------------------------------------------------------------

TOTAL 00023
_ _---_WEEKLY
..__ ._......- - - - - - - - -

.-- ..._.. __... _-- ..

·----1-JtFKl-V TOT AL

00 148"""

brew",

NUMBER

#1*. tt

t

t t .. b.drtrlt Msht

WEEKLY

"d ;

*riii t'r"" tt#6trif¥t#**#dri

TOTAL

r""z

rT-

tIe",

t r"

u

m-

DATE LAST USED

------------------------,oIo()"6- ---------006------- 0037

0----- - - - - - - - - - - - - - - - - -

.

08 266-8--------------08 28 68
~-

006
0028
- - - - - - - - - - - - - ---------------------------------------------------------------:..."'----

701011

WEEKLY TOTAL

00012

--------------------------------------------------------------------11

uS

E-R-NO-~

---2"1'0012-------- ------ -------------------------------- - - - - - - - - - - - - - - - - - - - -

NO-WE EK[y---USE-------- -- -- ----- -------------------------------------------------------------------------

---------------------------------------------------------II
USER NO. 270013
------------------------- -------------- -----------------------------------------------------------------------------------------------

NUMBER

-

-----_

702010

....

_----------------------

WEEKLY TOTAL

---------

WEEKLY

TOTAL _ _ - - - - - - - - - - - DATE
LAST USED
- - _ _-----------------------

---- ----------- ----------- ------------------ ----- ------------- .. _-----_..---------------------------------- -----

-----------------

-----

001

..

0001

..

08 22 68

00001

--------------------------------- --------------------------- ----------------------------------------------------------------

- ----- - - US ER --NU-;-----Z7UUT6- ---------------------------------------------------------------------------------------------- --------- ---------------------..----- ----------------------..N---..--UTTM. . .BEF:
..

Wl:l:KLY

lurAL

---- ----- -------------- ------------(1JOUUT----------------------------o19--------------------- -----()'OfSf---------

DATE LASl USED
08 28 6 8

~

-_.' .. -.----'---------..----

--- -- -- ------WEEKL y--TDTAL--UDUT9-----------------------------------------------------------------------------------------------

0-------------------------------------------------------1
USER NO. 270007
u

______________________________________________________________________________~ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ -

NUMBER

WEEK LY

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ _

TOTAL

DATE LAST USED

--------------------

820005

018

0018

08 20 68

700001

006

0006

08 28 68

----------------------(nn~OIO~....-2------".O~02---------;0"'0""0:-;;;2.--------0:;::;-8A'---:2...-;8;--';6;--;8:;---------I

----------------------- -- ---- --------- ---------------------------------------------------------- - - - - - - - - - - - - - - - - - - -

WEEKLY TOTAL

00026

- --- ------------------------ --- -------------------------------------------- ---------- ------------------------ - - - - - - - - -..-- - - - - - - - - - - - - - - - - - - - - - - 1

-----------us ER

o

NO.

2 10015

o

c

CABD SY,3TE1Y:

~·lODIFICATICNS,

SP3 AND

pr)(~

John H. Wise
Washington and Lee University
. Lexington, Virginia 24.450

This paper describes some modifications in the card system
currently used at IJashington and Lee. The COiilpu.ter configuration is
a 1620, Illiodel 1, with 1622 card read-punch and 1443 printer, and includes
the special features package for indirect addressing and TNS, TNF.
Memory is 20K, but the modifications would be the same for larger memory
machines. The special feature package is required for the PD:~ modifications.
The PDQ modificationo have gone through a ~umber of stages. Initial
changes to permit. printer output 1t1ere rlade at vlashington and Lee 3.S well
as by loir. John Vi. Holmes, Cooper-BeSSel:ler Corp., Hount Vernon, Ohio.
~1r. Andre Lacerte, while at ;/vashington and Lee, rr.a.de an extensive revision with added subroutine features. About the only Lacerte modification retained here is the E-forrnat style in which the decimal point
nmy be placed wherever desired (however, the tot~l width of an E field
is filled with digits even if!! is greater than 14).

The sequence of modifications is outlined below.
Both compiler and subroutines required
changes. Features include: Trace on printer in a two line fornmt
(address and value); Tl~E statements on console, but requiring
CONTHOl for new lines; an option of 105 pl'Jint position output and
an option for either 72 or BO colulnn card input.

1) Changes for 1443 printer.

2)

Changes in switch settings so that ON calls the featllre into
operation (thus S3l ON calls for printer listing of source).

3)

Changes in loader from the dieit-by-digi.t scheme to a card read
scheme. Loading time is reduced from 40.5 sec to 2;.5 sec with
this modification.

4) Symbol table on printer. To include this, the SS2 option of
punched referenced source statements :las been deleted (S32 is used
below). Listing of the symbol table is under control of 3Sl.
Page skip on channel 12 is included in source and symbol table
listing. The listing is four items p~r line except for dimensionei variables at one line each.
5)

o

A pre-compiler option has been included under control of SS2. If
332 is ON, object punching and symbol table listing is suppressed~
With SS2 OFF, object is punched until an error is detected, and
the symbol table is listed only after a successful compilation
of the program.

· ."._-".

"-.~.--,,-,,---.,"-.~-.-~~

.... ,.""-"'-,,-,-- •.,,-,,... -,,- ..--.. _-'." ......"._ .. "","".",-., ...,," ..... ,,".,, .....,,-,.-,.,,-.,,--.....

..

-~~~~~-~~~

----_._---,---------,,,_.-.-..._...•._-- ---_.,,----

.1

~I

"i

o

c

,H,

-r -,

o

TT'lJIW

- r -,.-.

rrn'KTse-t 77Rt

6)

Subroutine changes are required for printer output and the E-format
mentioned above (the E-format change forces ~he t".lo-line tra.ce output) •

7)

A condensed subroutine deck is used for all programs that do not

require relocatable subroutines. (A program can be checked for
relocatables on the third from last card of the object. This ca.rd
has -0402 in cc 1-5, and will contain 0' s in cc 55 to 70 if :.10
relocatables are needed - othe"~se some l's will be in cc 55 to 70).
This deck has been further compressed by a revised loading program
that uses a 5 column address instead of the 19 colurr~s of the
standard deck. It can not be used if subroutines are included in
the object deck (if SS3 is ON during compilation).
After completing these PDQ modifications, the SPS deck was examined.
By using a loading program 3imilar to that in the PDQ decks and eliminat-

ing some unnecessary or unused instructions (e.g., the test for Hodel 1
or Model 2 machine), the SPS III deck was condensed from 514 cards to
241 cards. This deck size is much more convenient to handle.
The only SPS modification was to remove the halt which is included
in an SPS object deck at the conclusion of loading the object. At
present, a branch to start occurs automatically.
Copies of the modified decks can be obtained from the author.

o

o

c

HJ

.

-K

[f!

-

t

W 'fRan

o

C()NVEkSIUN UF FURTkAN II TU FORTRAN IV
RY
PATklCIA E. KAY
D. W. MATTSON CUMPUTER CENTER

o

o

TENNESSEE TECHNOLOGICAL UNIVERSITY
COOKEVILLE,

TENNESS~E

38501

o

CUNVEkSIUN UF FUkTkAN II TU FORTRAN IV
INTkOOUCT IUN
WHEN FORTRAN II Ok II-I.) PROGRAMS WRITTEN fUR THE IHM 1620
CUMPUTt:R NEtD TU 8E rvIAUl: COMPATIHLE FUR RUNNING UN A CUIVlPUTER IJSING
FURTRAN IV, IT IS NECESSAkY TO CHANGE CERTAIN STATEMENTS.
THE FIL TER PROGRAM, RUNNING ON THE 1620 SYSTEM, CHANGES MANY OF
THESE INCLH~PATIBLE STATEMENTS OR POINTS OUT Tn THE PRUGRAMMER
THE MURE DIFFICULT CHANGES TO BE MADE.
HASIC PRUGkAM DESCRIPTIUN
THE PkIMAkY CHANGES TO Bl: MADE IN A FURTRAN I I SUURCE DECK
AkE FUUND IN THE INPUT/OUTPUT STATEMENTS, READ, PRINT, AND PUNCH
IN PAkTICULAR. SINCE FORTRAN IV REQUIRES A DEVICE CUDE FOR EACH OF
ITS I/U STATEMENTS, THE OPERATOR WILL bE ASKED TO ENTER THE DEVICE
CUDES FUR THE CAkO kEADER, CARD PUNCH, AND PRINTER AT THE BEGIf\INING
OF THIS PKUGRAM.
THUS, THE FURTRAN II STATEMENT •• PRINT 70, A••
WILL HE CHANGED BY THE PRUGRAM TO •• WRITE(N,70)A •• , WHERE N
REPRESENTS THE DEVICE CODE FOR THE PRINTER. CJTHER FORTRAN I I I/(l
STATEMENTS THAT WILL BE DETECTED ARE THE TYPE, ACCEPT, AND IF SENSE
SWITCH STATEMENTS, AND THE KECURO, FETCH, AND UEFINt DISK STAEMENTS
OF FURTkAN II-D. SINCE THESE STATEMENTS RE(JUIRE REPRUGRAMMING THEY
AkE PUINTED OUT ON THE CONSULt: TYPEWRITER. NU CHANGE IS MAnE Tn THE
ORIGINAL STATEMENT.

o

UTHER CHANGES TO A FORTRAN II SUURCE DECK MAY ~E NECESSARY
WITH THE USE OF THE ARIT~ETIC FUNCTIUNS -- SINF, CUSF, ATANF,
S{")kTF, LUGF, EXPF, AND ABSF.
IN ALL CASES THE F MUST HE D.RUPPED.
THIS CAN USUALLY BE DONE BY THE PRUGRAM. THE LUGf- TO ALUG AND
AHSF TO A~S OK lABS CONVERSION MAY REWUIR~ MURE CHANGES THAN THIS
PKOGRAM CAN HANDLE.
IF SO, DIAGNOSTICS WILL BE TYPED.
EACH CAkD IS GIVEN A SEQUENCE NUM~~k THAT IS PLACEO IN THE LAST
FOUk CULUMNS OF THE CARD. THE PROGRAM TER/VIINATES UPUN THE DETECTIUN
OF AN •• END •• CARD WITHIN THE FORTRAN SOURCE DECK. THIS LAST CARD
IS kEPKUIJUCED AND THE PkOGRAM HALTS.
THE OUTPUT
AFTER THE PROGkAM ANALYZES EACH FURTRAN II SUURCE CARD, IT
EITHEK kEPRODUCES THAT CARD IMAGE, MAKES NECESSARY CHANGES AND
THEN PUNCHES THE CARD, OR DETECTS THE rvll)RE COMPL ICATED CHANGES AND
PUNCHES THE CARD WITHOUT THAT CHANGE. THE CHANGES THAT COULD NUT
HE MADE WILL BE LISTED ON THE TYPEWRITER, GIVING THE DIAGNOSTIC AND
THE CAkD NUMBER I N THE LAS T FUUR COLUI'~I\JS UF THE CARO.
THE LISTING

o

AN UPTION UF THIS PROGkAM IS A PRINTED LISTING UF EVERY CARD
. IMAGl:, INCLUDING CHANGES, AND AN ASTERISK TO THE RIGHT OF THE
UNCHANGABLE CARD IMAGES.

/

THE DIAGNUSTICS

l ) t\
~)

;-~ ~ S( H J I\J A S T H t: P k UG K A M [H: TEe T S A l\j l J f\l C UN V E R T I K L t S TAT E tVI E N T
A K 1 T H1"'1 E TIC fUN C T I UN, A (lil E S SAC; r: 1ST YP t: lJ FUR THE P I<. (J G K A/VI /VI E K t\ L UN (;

1 fH

o

I t-H:: CAI-

ir±rlffli "· :-V"" TW

- . W"f . -nm"H""t"trW"'Tf

These wi I I represent the 5 possible entry points to

o

C.

d

particular segment.

Returning to the error I ist, place the entry Doint number
after the seqment number tor each statement number.
should have a segment number and entry point

numbe~

You now
for each

statement number on the error! ist excert rORMAT statements.

D.

Add two fixed

point~variables

to the Common statement (such

as LI NKl AND LI NK2).
E.

To each segment add the fol lowing statements for each missinq
non-format

staem~nt:

.xXXi LI NK 1 = SN

L.I NK2

=

Er

RETURN
Where XXXX is the statement number, SrJ : "the correspond i n'j
segmen-t number, and FP is the c0rn~spon:ji.n(J,
F.

Sf'

""Y p()i'~it number.

Add a computed GO TO statement to each segment unless there is
only one entry poln+ and that is the first executable statement
of the seqment.
The computed GO TO should immediately fol low the COMMON
statement.
LI NK2.

The variable of the computed GO TO wi I I be

The statement numbers wi I I correspon d to the !;', t

made up in part B.
C.

Insert the required

FOR~J,AT

statements in each segment.

Some FORMAT statements may appear" in severa I segments.
H.

Code the mainl ine program.
pattern shown in Figure 3.

o

It wi II generally fol low the

r

-----~

...

_ - - - -.._•............•

-

...

_-------------'
DIMENSION

L! NI<2

o

=

CALL name
90 GO TO ( 1,,2,3,4,5,---------), LlNKl

2 C/\LL name 2
GO TO 90
3 CALL name 3

GO TO 90

END

F i qure 3

Basica! Iy the mainl ine program usues LINK1 to route control
to the proper segment.

LINK2

must be defined before the

first segment is cal led unless the first segment has no
computed GO TO statement.
I.

It may be more:.efficient to place certain statements from
the original program into the mainline program.

Statements

to consider for this are computed GO TOS, STOPs" and
CALL

EXITs.

Each of these statements placed in the main-

line program wi I I have a distinct value of LINK1 associated
with it.
J•

I f lib rary sub rout i nes are needed by any of the segments, it
wi I I be necessary to place a dummy statement in the mainl ine

program to avoid an "L7" error.

As an example, the following

c
/

I

$"

mw RrJ!!XWTlmr

statement would load the LOG, EXP, subscripting, COS, SORT,

o

and ASS subroutines.

A(I+1)

= ABS{COS(~.»+SORT(LOG(1.)+EXP(0.»

This dummy statement should be placed where it can nctbe executed.
It is also possible to cause proper library subroutine loading
by punching the correct indicators into the header record of the
mainltne program.
VI.

Campi Ie all segments as Subroutine Subprograms.

or punch object decks.

Load them on disk

Compi Ie and execute the mainline program.

Be

sure to include the *LOCAL Control Card immediately after the source deck.
Also note that the number of LOCAL records must be given in the FORX

or

XEQS Control Cards.
It is helpful to turn S51

on'wh~n

the:,I'EXECUTION" message is typed.

This wi II cause a core use tabl;e Tp.be .tvped· out as the segments are loaded.
If an "overlap" message appears at this time it wit I be necessary to return
to Step III.

Th's system has been used to segment several programs that require
over 100,000 core positions.

These programs are now running on a 20K

system with no difficulties.

In one instance a program that was written

in three segments for a 60K 1620 was broken down into seventeen segments.
Timings on this program show that it ran slightly faster on the 20K machine,
in spite of the segmenting.

This was probably due to a difference in floating

point hardware.
One difficulty not handled by this system is that of reading data in
with an "H" specification and later printing out the same data from the
same "Hit specification.

o

If the input and output statements are in different

segments, the data wi I I not be transmitted between the segments.
for this is to change the

The remedy

"H" specification to an "A" specification.

o
:!

;',

APPENDIX A
"COMMON" L I 5T I NG

:,

/

\,W'.!TW','ig''f''t''f1''t

f

\$ zt

t'g'WM"WF ..,"'"

t PNI'

* rt

t t t

(ttf*"t±hH#".· &""ms ·"rlw-tMt±i"rribriiiftH

- - - CN

-

r··-

*LU I SI:tsth~lin~f:f()rENQ; upon receIvlng
tr ansmit $: ;AC KO.andre'ceiVesth:,:fil"'st· message.
:~.

o

" . ;".:': ,": :".<: . . ~-:; :.;:~';' ;': '.~. ','" :::: : :. :',;,:' :~'<.': :~::'~:':.,;~:':':; ::": '" .: :~~<

AtitOlna.ticMode.•.. - . sets UP:~~A''I'4.~,b;~utomatic ally tr~nsJnit·.
after '~Ctcliincomin:g>tne$$~g:'.fh~;'~FP~~p:riateacknowledg'ffient'
unle,ss.the'.·user inttlate8·.•. an()~b:~r::'~,e':U~$t.

Serni~.automatic . ·. Mode . ': ··set~~p::~~,~r.p~,~~.r~ve;rse. th!>~~l~eafte.t··
each inc omingmes sagebut'to".aU; ,'for.:fu'rthel' dire'ct*,On,$' frc;rn
theuser~

Receive Continue - Transril!tsthec;orrect positlveaeknowledg . .
rnent for the current message and receives the next'message
when SCA':r4 is in serni-automatic mode.
Receive Repeat- Transm.its NAKfol" thecurreht message and
receives the next.message~,
When performing a Re(H:~iv.:o{)~·t'ati()ti, the firstwo~d

of the

II o are a C ontai? ~,the. ·niaxilnumri~'m.be r (Yfunpacked.char:ac'ter s
(word count) that can be read into that area.
The entire .rnesSagethati':r,~~e:iV'::dfsstt)~ed Inthe .1.10area ,
including the SOH>cha.racte:r:~~ndl or;'1:~'e.·. ST}{ charae~er, •{l)LE
.
$ T X, if. T l"'ansparent te~), .~h~<:~':t"Bc;¥a raete r .(J:jlr~El'll,. if
Trans parent te~},Ol"ithe'E';p"XCh~l'ae~r" (DLEE'TX,>if: Trans- .'
pare nt text) '.Afte: t~~.t:i1.es,,~ age:has-b~~n reee,iV'ed,.-'the nnmbe r
of character$re~ei.ved)'ln.~~udl~~control characters, is stored
in the first. word' of the.lt',O:','~:re·a~': ,",' .• ',

':/-:'~; ~.>

;.'/."

tt:,the·.• ·;:re:~~'td,l~• ,.:.~.~'~:el

" ·"~delete.~,.,• th~ i.,e~.~~~'~(

..:

'stationl·m,~aeh.'p:~i.r'ti:f':"i,.i .

Mter, rece!Vingt~'bi~,~~~~t~~~~~~~~l:~$8~~4,aU~~~tiC~lY

rev~·r··~e·.s· ··tb.e .• litie:--.. . . •. . :F.f.• :l1~-~'ett~!*':~·~e.·.£o~~;.·i~·. •t~~:' . .~e·:8\.~le:.:and
if.' aut()tnatlc:·., •. :tll.()d:•.••.·. ~•.~:~:.,.ef4:~~:~~.~. ·.;~(.;~T:~~r~. p~.l'.~·s . . ,~(f.··. tra.»:s mit
,
_·the· '~PP;l'O~ti~!e·.~e'~:~t~;;';.~~~~~~~~~'e~.. ...'l)Qrirll the · line . tu rn.. ar oun?·.l1\\t,erval, . .···,the;'\'l.·.~:~1~:tlve·r~ide>;~~.•. ·. tr~n~nli.s.$ion.' .of the
~ckn0"fl~dg~nt. '. ·.~ll~'!~~Q~.:. ~~:r.~"roun~,'.antl,'if.the .••.••mi...
a utomaii:ettlodewa$:$.·l.:~t.d; .Sic~·T::4:W"its .ft:Y.tfurthe.:r

c

- T lfITT 'f

n·'

JI"!!I!T . Ir-·

6 -

instructions from. the user. Upon receipt of a message SCAT4
branches to the user's error routine with (0080)16 in the
accumulator. The location of the input buffer is in the accumulator extension.
If the block check character (CRC-16) is found to be incorrect
or if the message overflows the I/O area, SCAT4 transmits

NAK and attempts to receive the me s sage again. Afte r eight
unsuccessful attempts~ SCAT4 branches to the user's error
routine with an error code in the accumulator. (See Postoperation Error Detection.) If the user returns with a positive
accumulator, SCAT4 transmits NAK and attempts to receive
the mes sage again (up to seven ITlore attempts before branching
to the user's error routine). If the user returns with a zero
accumulator, SCAT4 perform.s a Close Operation.

o

If the user returns with a negative accumulator, SCAT4 clears
the Device Routine Busy indicator and stores the num.ber of
characters, including control characters, received in the message
in the first word of the I/O area, allowing the user to initiate a
Receive Repeat or Receive Continue operation.
If a timeout occurs while receiving a message, SCAT4 monitors

for ENQ and, when ENQ is received, transm.its the last acknowledglTIent.
If the EOT character is received, SCAT4 clears the first word

of the I/O area to zero and clears the Device Routine Busy
indicator. The user should then initiate a Receive Initial,
Transmit Initial, TranslTIit EOT, TranslTIit DLE EaT (Dial
line only), or Close operation.
If DLE EaT (Disconnect Signal) is received, SCAT4 sets the

first word of the
operation.

Ii 0 area to FFFF16 and performs a Close

5/6 TransInit Block! Text - There are four types of TranslTIit Block and
TranslTIit Text operations. Digits 2 and 4 of the control paralTIete r
are used to specify which sub-function of TranslTIit Block/ Text is
requested.

o

-

Digit 2

7 -

Digital Value
0
1
3

0

Sub- F'unction
Transmit Initial- Sen'li- Autom.atic
Transmit Continue
Transmit Initial- Automatic

Digit 4
0
non-zero

Normal EBCDIC text
Full- Transparent text

Transmit Initial Block/Text - Transmits ENQ, receives the
acknowledgment (ACKO), transmits the me s sage from the I/O
area, transmits the CRC-16, and receives the acknowledgment
(ACKl) or return lTI.essage. (Serni-autoITlatic or autoITlatic m.ode),
Transmit Initial Transparent Block/Text - Transm.its ENQ,
receives the acknowledgment (ACKO), transtnits the ITlessage
from the I/O area, transIl1its DLE ETB/DLE ETX, transmits
the CRC-16 and receives the acknowledgment (ACKl) or return
rne s sage.
Transrnit Continue Block/Text - Transmits the rnessage frorn
the I/O area, transmits the CRC-16, and receive~ the acknowlEodgrnent or return ITle s sage.
Transmit Continue Transparent Block/ Text - Transrnits the
rnessage from. the I/O area, transmits DLE ETB/DLE ETX,
transrnits the CRC-16, and receives the acknowledgment or
return rnes sage.
Contention exists when the two stations on a line simultaneously bid
for control of the line.
SCAT4 provide s a rneans to break contention. If the user wishes
to be the m.aster station in the event of contention, digit 3 of the
control pararneter must be zero. If the user wishes to be the
slave station, digit 3 of the control parameter must be non- zero.
In a master station, when contention exists, SCAT4 re-transmits
ENQ. After eight attempts, SCAT4 branches to the user's error
routine with an error code (4000 ) in the accumulator. If the
16
user returns froIn the error routine with a non-zero accum.ulator,
SCAT4 attempts to break contention seven m.ore times. If the

c
/

8 -

o
user returns with a zero accum.ulator, SCAT4 perform.s a Close
operation.
In a slave station, when contention exits, SCAT4 branches to the
user's error routine with an error code (4000 1 6) in the accumulator
and, upon return from. the error routine, perforrrls a Close operation,
allowing the use r to initiate a Receive Initial ope ration.
When perform.ing a Transm.it Block/ Text operation, the first word
of the I/O area contains the number of characters in the :rnessage.
The character count includes the control characters in the ITlessage~
All character s in the I/O area are unpacked and left- justified. If
the user wishes to start the message with a heading (optional), he
must supply the SOH character as the first character of the n'lessage.

o

If there is text in the message, the text portion of the m.essage follows
the heading. When digit 4 of the control paran'leter is zero, the text
is Normal EBCDIC text and n'lust begin with STX and end with ETB/
ETX. The user m.ust supply these characters. When digit 4 of the
control parameter is non-zero, the text is Full-Transparent text
and must begin with DLE STX. The user must supply these characters. The ending characters, DLE ETB/ETX, are supplied by SCAT4.
SCAT4 transmits a second DLE character after each DLE that is found.
in the Transparent text.

If a redundancy check of the heading separate from. the text is desired,
the heading must end with ETB. The ETB is supplied by the user.
The I/O area is not checked for misplaced or incorrect control
characters.
SCAT4 transmits the 1 6 ... bit block check character (CRC-16) after
the ETB/ETX is transmitted. The CRC-16 is generated by SCAT4.
When the proper acknowledgment is received, SCAT4 clears the
Device Routine Busy indicator and returns to receive m.ode.

o

If SOH, STX or DLE STX is received in response to a message,
SCAT4 puts the return message into the last mentioned I/O buffer.
If the user does not want the return ITlessage to overlay the output
buffe r, he must initiate a Change Buffer reque st immediately afte r
reque sting a transmit operation.

-

9 -

o
If ACKO is not received in re sponse to the initial ENQ 1 ENQ is

re-transITlitted; except when contention exists and the station
is a slave station.
If NAK is received in response to a message, the ITlessage is retransmitted.
If EaT is received in response to ENQ or to a message, SCAT4
clears the first word of the 1/0 area to zero and clears the Device

Routine Busy indicator, allowing the user to initiate a Receive
Initial, Transm.it Initial, TranSITlit End, or Close operation.
If DLE EOT is received in response to ENQ or to a message,
SCAT4 sets the first word of the 1/0 area to FFFF16 and perforITls

a Close operation.
If anything other than EaT" DLE EOT" NAK, SOH, STX, DLE STX

or a positive acknowledgITlent is received or if a Receive tim.eout
occurred after the message was transITlitted; SCAT4 transITlits ENQ.
If an incorrect acknowledgITlent is received before a receive tiITleout
occurred, the message is re-transITlitted" otherwise an ENQ is transmitted.
after eight attempts, the proper positive acknowledgment is not
received, SCAT4 branches to the user's error routine with an error
code in the accumulator (see Post- operation Error Detection). If
the use r returns froITlthe error routine with a positive accuITlulator,
the transmission is atteITlpted seven ITlore times. If the user ret.urns
with a zero accumulator; SCAT4 performs a Close operation. If the
user returns with a negative accumulator, SCAT4 continues. as if
the proper positive acknowledgment had been received.

If

7

TransITlitEnd -The Transmit End operation can be one of two types.
Digit 2 of the control parameter is used to specify which sub-function
of TransITlit End is requested.
Digit 2

Digital Value

o
I

Sub-function
Tr ansITlitEOT
TransITlit DLE EOT

TranSITlit EOT-- Transmits EOT, then reacts according to the reply
received.
TranSITlit DLE EOT-- TransITlits DLE EOT, then perforITled a Close
operation.

G

10 -

:)i~Jji: -:

control parameter il:) used to specify the action to
un a no response condition during a Trans'mit ECT

,)f

f.i'

.,'11

!_l:_,tlji

-+

-i-he

Digital Value

o
non-zero

Sub-function
Close
Do NOT Close

<->n J Transmit EOT operation, SCAT4 transn-dts EOT and receive s
the response.
If there is no response (i. e., a Receive tiIT1eout
occurs), SCAT4 perfornis a Close operation if digit 4 of the control
:~~cLi: l~.etcr is zero.
If digit 4 of the control pararneter is non-zero,
SCA.T4 clears the Device Routine Busy indicator and starts the 3d(;C .ind hnler 1 allowing the use r to initiate a Receive Initial) Trans'rrdt
Initi,j]; or Close operation.
If the response is DLE EOT, SCAT4 sets the first word of the I/C
area to FFFFl6 and perfor-rns a Close operation.

o

If the response is EOT, SCAT4 stores an EOT character in the
location specified by the I/C area address and clears the Device
Routine Busy indicator, allowing the user to initiate a TransIT1it
Initial, Transmit DLE EOT, or Close operation. If the response
is ENQ, SCAT4 stores ENQ in the location specified by the I/O
area address and clears the Device Routine Busy indicator, allowing the user to initiate a Receive Continue or Receive Repeat
ope ration,

If a response other than DLE EOT, EOT, or ENQ is received,
SCAT4 re,·transIT1its EOT. After eight unsuccessful attelYlpts,
SCAT4 branches to the user's error routine. If the user returns
with a non- zero accunl.'J.lator tranSIT1ission is attempted seven
rnore times. If the user returns with a zero accumulator, SCAT4
perfol'Ins a Close operation.
8

Change Buffer. - Allows the user to separate input and output buffers
or to provide a limited multiple buffering capability for input. This
operation rriay be requested at any time without te sting for the co'm_pletion of the ongoing operation. Any new request involving buffer
allocation will override the previous Change Buffer reque stc

o
----------------------

-

11 -

o
The buffer is changed when the next text me ssage is received.
If a tranSITlit ope ration is in effect when the Change Buffer
request is initiated then the new buffer will be used if a return
text J11essage arrives or if an unsolicated text nl.essage arrives.
If a receive operation is in effect when a Change Buffer request
is initiated the current operation is unaffected but the next inc otning file s sage will be inserted into the new buffer.

Error Handling
For a description of error handling procedures, refer to General Error
Handling Procedures in the publication IBM 1130 Subroutine Library.
Pre-operation Error Detection
The following conditions re suIt in pre- operation error action (accuITlulator
settings are shown in parenthe see):
Invalid function code (8001

16

)

c

Invalid sub-function code for SOllle Translllit or
Receive operation (8001
)
16
Invalid word count, (8001

)

16

Post- ope ration Error Detection
The following conditions result in a branch to the user's error routine
(accutnulator settings are shown in parentheses):
Data set not ready (8000
Contention exists (4000

16

16

)

)

3- second titneout occurred while receiving a J11essage or
lllonitoring for ENO, or ENO 'not received while monitor.
ing for ENO (2000 )
16
I/O area overflow (1000

16

)

c'

·UTT!

.

rr

II T .\1

eerley·ow

- I

12 -

o
Block check character (CRC-16) in error (080016)
Receive tim.eout occurred after transmitting arne ssage or
ENQ, or invalid sequence received in response to a message
orENQ (020016)
NAK received, or the incorrect acknowledgnv~nt received
following a Receive timeout (040° )
16
Incorrect acknowledgment received with no Receive timeout
(01°°16)
Message received (0800

o

o

16

) EXT has location of input buffer.

o

c

tlt/ti,

tt

Ii

trit!tb't5tft.ttf'tttrbHtb#iWf&itt'rtt t

t*

It

irirf:*tftrib #Mtirti*riiMbd#itttttff, -if!

j"I

-:

"H6ri5tflti:ih - ·W

J

)"uP"manw

o
I

I

HIGH PHECISIO:-I INTEGER SUDP110GRA!,ill
F'OH IBM 11)0 CO~'J:'rER8IAL .~PF'IJICNJ.'IOI'~S

by

B. J. SHAIN

o

o

~.lhc ShH'r'TilllGan EnGinec!:i."'lng Cc};)'pc.ny Limit(:Q
Box ;'010, Stflt :i..O~1.B, !.~0nt.rcQ.l

o

c

i:

t to

t

tt

1 $

tt

!!tim

H ,

t'¢,,'W¥!!!Hl'''P

,t

i,t'" trif t ' ttit if 'Hit'#ri frlrir6rlititrli'f t t±±tttHflHktM .

""rlif#i*Wriw**,ttfifP y" .... *fflFb±'" T tTW

o
Page

o

Use of I,'ortran for COI11!:lcrcial Applications

1

Program Philosophy

1

Data Format.s

2

Card Read or Punch

3

Mode Conversion

3

Arithmetic Calculations

3

Overlapped Read

4

Input/Output on Con.sole

4

Comparisons

4

Logice.l Operat ions

4

Record Blocking

5

Comparison \-lith IR\-l Commercial Subroutines Package

5

Probrrarnrning Languages Used

6

Card Read Subprogram

7

QREAD
Data Definition and Mode Conversion Subprograms

o

7

8

PDATA

8

IDATA

8

PC¢NV

9

IC¢NV

9

PEDIT

9

!EDIT

13

PPASS

13

IPASS'

14

PPASX

15

Contmlts

cont'd
Pnge

~tlnnetic

0

Subprograms

FMPY

16

PADD

16

PSUB

16

PDIV

16

PDIVT

16

Pl.f¢D

16

PMINO

16

PMAXO

17

PABS

17

Overlapped. Rend Subrou'cines
R¢PEN

18

XREAD

18

RlI¢LD

18

C'

Overlapped Typewriter InFut/output Subroutines
QTYPE

20

QT"lWR

21

QWRTY

22

I¢ND

23

Compare Subprograms
IPCMP

24

IQCMP

24

IQDLK

25

C

"flaW

PU

..

orne;

o
PLIS'f

26

QPNCH

27

Sh~fting

and Logical Operation Subrrop)."ams

IALS

28

I1illS

28

IRQL

28

lAND

28

29

o

IPIK

29

PACK

29

Zone Recognition Subroutine

30

t

Disk File Record Blocking Subroutines

31
31

o

DGET

32

DPUT

33

DPl'S

33

DFIND

33

DcL¢s

33

DGE111

33

DFLITl

33

DUSE

Content.s
------

cont'd
~~

Page
Other Subpror;rams not Normally Caller! by

0

U~

PSFAC

35

QED IT

35

QEROR

35

QGRAB

35

QSHUV

35

PNMPY

35

PNADD

35

PNDVT

36

XR¢TE

36

TYWR

36

WRTY

36

C

Compatibility rlith COMET Subroutines
QCLER

37

Q,PASS

37

QSTAK

38

c

1.

o

1.

The Use of Fort~~or Cori:!:l,crcinl A.pplicntions
The use of Fortran for cummercial applications on th~ IEM 1130
presents several problems.

These problems have to be solved since

Fortran is the only compiler supported for the IBf-1 1130 and it is
usually necessary to produce n limited amount of corrunercial reports
despite the feet that the typical IBM 1130 installation is intended
for scientific

worl~.

The principal problems presented by the use of

Fortran for commercial work are the following:

o

a)

Manipulation of character strings

b)

Editing problems, for exar.-lple, floating dollar siens,
cheque protection

c)

Round off for cross footing purposes

d)

Recognition of zone over-punches

e)

Selection of the alternate stacker

f)

Input/output speed

Various subroutine pack8.ges have been evolved to eliwinate these

•

problems.

The

one described in this paper was developed by the

Sha'\vinigan Eneineerihg Company Limited in 1967.

In c:)mmon with other

subprogram packages which permit the use of Fortran for commercia.l purposes
it L:i based on FORCOM (IBM l620.library 01.6.051).

HO\>lever , it incorpor-

ates a number of features which have not been publinhed elsewhere.
2.

Program Philosophy
Using a cormncrcial subroutine package the programming philosophy
adopted is as follows:

o

a)

Read card and store its card ima.ge in an alphabetic array

b)

Test the card type

c)

Extract numeric and olphabetic data, convert numeric data
to the form required for procecsing

2.

"c...

Progranu:ling

Pl1i}.o~hy

cont'd

d)

Process the data, build a printer hnaec f;.)r output

e)

Print

o

This philosophy has been followed in the conunercial subroutine
package now being presented.

It is expected that the writer's over-

lapped outpu.t eubroutinc will be used (see IBM 1130 Libre,ry)

4

fare the printer image need only be partially built before printing.

3.

Data Formats
The following data formats have been used in the package:

a)

For alphabetic information a COMET array has been employed.
COMET array is defined as an array ha'!ing a single subscripted

integer ns.me.

It contains 2 alphabetic characters per word,

in EBCDIC coding.

The

*

O}rn; \-lORD INTEGERS

opt ion must be
~,

used to produce maximum packing density. . A COMET array can
be printed by the use of A2 field specification in a
statement.

,*,7

FOmv~T

(This mode of storage is the one used in the

contributed 'program package COMET.
programs described have been made

Where possible the sub~ompatiblc

\-11th the cor·tEll'

package written by J.R.Hurley and others of IBM.)
b)

A standard Fortran integer variable is used for counters and
for data whose value does not exceed the range -32,768 to +32,767.

c)

For numeric data. which cannot be contained in an integer variable,
e.g. amounts of money, the high preCision integer variable has
been defined.
subscript.

This has a real variable nrune, subscript or nonThe use of standard precision is required.

This

permits the use of 2 word's or 32 bits giving a range of
This range is generally
adequ€l,t.e

to express money amounts to one cent or one mil.

c

:(!IU'

y')''''''!''\,

i'!

't:WIhI'Ut""ij'·Wir+i='riWrHillrl:!t'!id#htWW'r"W" [J""!f!P'!

tt

: til

tt

t

#Hht'rtdtrt'±n+

dr

t

dfrlt'urdb6·· ...

i. 'w-

r"

:gr"

!!rrzMlU'" "

3.
J\..

Cnrd Rcrtd

o

01'

1~U1ch

Subroutines hdve been provided to rU1rl a card awl store
its card
card.

or to punch a card i.mRge from storRge into

im8~c
Th(~se

A.

QREAD
QI1\~CH

subroutines permit the cOHlplete EBCDIC charo.ctcr

set to be converted.

5.

Mode Conversion
Subroutines have been provided to convert alphabetic

PDATA

i.nformatiOll contained in a COl.ffiT array to integer or high

IDATA

precision

PC~~NV

inte~er.

The conversion function permit8

version from int.eger to hieh precesion integer
precision integer to integer.

0:':"

COll-

fom high

IC¢NV

The EDIT subrout.ine perrpits

PEDIT

either of these fonDS to be edited into an alphabetic Cm.f8T

lED IT

array into which an EDIT mask has previously been inserted.

PPASS

In t.his way floating dollar signs, cheque prote.ction and

IPASS

credit indication can be handled.

PPASX

8,lphabetic form is rC(luired

If simple conversion to

subroutin/~s

are available to

permit this to be done.

6.

Arithmetic Calculations
In view of the fact that. the data contained in li'o:rtran

PADD

real variables is in fact in the form of integers then Hormal

PSUB

Fortran arithmetic cannot be used.

PDIV

mp.tic functions have been provided.

Inst.ead a series of arithFor example, the statement:

PDIVT

R = FNFY(P,Q) causes Pand Q to be multiplied and the result stored PM¢D
in R.

The use of functions rather than subroutines in this

instance permits a series of arithmetic operations to be

o

p~rformed

in a single Fortra.n statement, for example, the c8.1culation:
D =

A + B
C
. is handled by the single statement: D

= FDIV(PAi)D(A,B) ,C)

PMllIO

PHAXO
PLIS'r
PABS

4.
7.

o

oV'~l~pcd Hc['~

Provided it is not necessary to punch inforJ1'ation into

R¢PEN

cards which have been previously read 'then additional reader

XREAD

speed can be obtalncd.

RH¢LD

into

0,

'l'hl.s is achieved by reading a card

buffer ahead of the time at which the information is

required.

At the time the information is called for in the

users program it is converted and moved to the required
stora.ge in core.

Meanwhile reading of the next card takes

place simultaneously with processlng the data on the previous
card.

8.

Input/Output on Console
Subroutines nre provided to support communication with
the console keyboard alld type\>lri.ter.

Information to be

transmitted to these devices is held in core, in a.lphabetic

QTYPE

QTWYR

c

QWRlty'

form, in COMET arrays.

9.

Comparisons
Comparison functions are provided for comparing 2 high

IPCMP

precisl.on integers, two alphabetic arrays, to determine if an

IQCMP

laphabetic array is blank or if

column of an alphabetic array

IQBLK

The compa.rison function for alphabetic

IQZ~N

contains a zone pllilCh.

~

arrays has blank at the lou end of the collating sequence.
10.

~gical

Operations

Functi.ons are provided to permit 1~gical operation on 16
bit arguments.
8.

The 'use of thelu'ID function i.n conjunction with

mask will pe:cmit the extraction

in a '·lord.
The

RO'l'AT}~

lAND

of part of the data contained

IRQL

The OR function ca."l be used for packing information.

lArs

fu"1ction
..
permits a \-Tord to be shifted until B.ny desired

lARS

c

5.
10.

COll~pn

cont t d

t.'_i::::ons

bit is in the high oi'der or low order position.

o

IPIK

Thus tC0ts on

ind.ividual bits cun be pe:cf.)rmed. or bit p2.tterns treated as arithmetical data.

PACK

A function is provided to extract a sinBle

character of a CObffiT array or to place a single character into
a

11.

CONET array.

Record

Blockin~

D¢PN

It has been found very desirable to block records which are
to be transferred to or from

th~

this 'olay several records

be handled with a single disk \ffiITE

IT'.By

disk pack of the IBM 1130.

or READ with a corresponding increase in speed.

c

In

D¢PEN
DPUT

A set of sub-

DGET

routines has been provided to perform the opeL'ations required to

DC:r..¢S

block several logical records into a longer physical record and

DPUTl

to read or 'Trite the physical record

DGETl

~s

required.

DFIND
12.

Comp:-lrison with
Th~

IB~,1

COliunercial Subroutine Package

principal advantege to be gained in the use of this subroutine

package in cornparion with the lEv! Cormnercial Subroutine Package i;;i the
speed of coding and the length of source program required to perform the
given task.

As an t'!xample, a. progrem involving card reading, disk reading

and writing, arithmetic calculations and output on the printer has been
written using these commercial subroutines and the commercial subroutine
package.

The number of x"ortran

these comml!rcial subroutines.

stat~ments

required is much smaller using

The core storage required is somewhat

la:i'eer but would also be decreased if the program were more complex.
The execution time on an IBM 1130, having 8k core, 1132 printer

o

1442 t-iode1

6 card reader, is ullaffected.

and

The following table shows this

comparison:

_ ..•..

_---------------

6.

o

12.

csp
Me.in Program Stater.lent

}'1ain Program Cor(!
Total Core
Execution Time

III

lIP

64

29·

998

586

4706

5288

1 min.

13

sec.

1 min. I', sec.

fr0grammin.g Language.s Used

Wherever possible, progr8.1lls have been

~lritten

in FORTRAN.

may lead to l)roblems in the use of high speed peripheral devices.

This
The

a.ssistance of Mr W. Baden of tv1a.rshall Communications, Santa Ana, Cali-

fornia is acknowledged.

He has re-writtcn the disk blocking subroutines

in Assembler language, "Tith a resulting increase in speed and decrease
in core requirement.

t"**,r&rtt&''&t

ibt6"*,-

T")"

'""WU"

--k-

n
".",

n" "n," f

J"WW

'(

o

Ca.rd Read

S11.b.J?ro~

CALL QnEAD (IA ,U)
To read a card and to store its card image.
IA.

co~·~

array in which chnracters ere to be

stol."'ed.

=
T'JlC firot

Number of columns .to be stored

}.i \~OlUlunS

this stat<:::ment.

of datp, are tr~n61'errGcl by execution of

Card reading a.nd

convcr~ion

overlf:i.pped by use of this aubprograill.

to EBCDIC a.re

lIo':h:;ve).' control 10 held

in the sul)orutine until reading the cord. is completed..

XRE\D

(see belm~) pcrnuts overlapping of c(u·d rea.ding and processing.
Note:
Since QREAD uses the IBM supplied CARDO subroutine rather than

the

nOrrnB~ CAr~

subroutine, then all card reading in a program

must be done using this subroutine.

Use of this subroutine

alloW's all character codes to be recognized, including
which are no-I; recognized by Fortran READ

The ~I¢CS record should not contain CARD.

o

8-~Q.tements.

~

nod

0

-

8.

o
CALL PDNrA{IA,J,U,R)
'10 conv(;rt tho

ch~racters

high precision

intEge~

=

III

ftppea,ril1g in

COl. IE,T array containing data to be converted

is to

=
=

M

R

C01.lli"r array to a

variable.

Character number

J

0.

str~t

11ithin

IA at 'Whj.ch conversion

(left hand end)

Number of characters to be converted
Variable to contain returned value

In 1utel"prctil1g the COME'£ arra.y lA, bIe,ruts will be takEm as

zeros.
nrl.!lUS

A

n~gntivc

number is designated by either a leading

sign, or by e.n 11 zone overpunch' in the units position,

thus converting that character into an olphabetic character.
The existence of other

charaete~s

vhich nre not numeric or

bla.nk will ca.use an error message to be printed, en d R to be

act to zero.

CALL IDATA(IA,J,N,I)
To convert the characters appearing in a

cor~

nrray to an

integer variable.
IA

=

COl~

a.rray conta.ining characters to be

convezoted
J

=

Column number w:l.thin IA at which conversion

is to stnrt (left hand end)
Nuu~er

of characters to be converted

c

9·

~'p"'efinltion and Hoc!.e Cor~.vcrsion_ Subpr<2.Gl"D.~

CALIJ IDATA

(cont~ r d)

( cont r d)

=

I

Variable to contai.n returned value

BlankB and non-numeric characters

~re

handled in the came

way as described under PDATA.

R

=

~ONV(I)

To convert

n~

integer variable to the high precision integer

mocle.

Previously defined integer variable or

I

constant to be converted.

Returned value

R

o

I

&::

ICONV(P)
To convert a high precision integer variable to integer mode.

P

=

Previously defined high precision inteBer
variable to be converted.

I

::

Returned value.

If the magnitude of P is

too large to be stored in I, then the high
order digits are lost.

No error message is

produced.
CALL PEDl'T(P,IA,J,M)

To edit a high precision. integer va.riable into a COMID:I' array
P

a

Previously defined high precision integer
variable to be combined with the

o

COM1~

array.

10.

o
!!!.~_DefillitiOll anCl. l-io(le ConverDio~l Subpro~m~

CALL PEDlT

(cont'a)

(cont'd)
IA

Previously defined COMEI' e.rra.y, containing e.n
edit mask

J

Column number within lA Llarking the left-most
end of. the edit mask

M

Number of columns comprising the edj.t. mask

The final appearance of the array IA depends upon the edit
mask which existed before calling th:J.s GUbl·outine.

The f'ollo't';ing

is a. table otcharacters used to ma.ke up the nk"\sk, and their
respective functions.

--

c

Control Character

Function

b(blank)

This character is reple,ced by a character

from the high. precision :Integer variable P.
O(zero)

This character indicates zero suppression"

and is replaced by a character from the high
precision illteger variable P.

TIle position

01' this character indicates the right-moat

limit of zero suppression •
• (decimal)

This character remains in the mask field
where placed'l

It is considered an eJ.pba.-

betic character, and may not be zero

suppressed.

G

11.

o
Control

Ch~ra.cter
-----------_.

Function

,( co~~n.)

This character rcnains in the
where pla.ced.

lIo~'Tcver,

is reqt.. . ested;. this

l:1.'.lck

field

j.f zero cnpprese:i.on

ch~".ractcr

'\-rill be rCjovcd
chr:..r().~tcr

if it is to the left of t.he ID,st

to be zero Gupprcr.sed.

en( crcd:tt

)

Thbse tvo chry.re.ctcrn can be plcJccc in t.he
two right-mor.t p::->sitionr'.

or

the

L:.:~sk

field.

They a.re undi,,:turbed if the source field

is negative.

o

If the oourcc field is posj.tive,

the charactors C and R are blanked out.
Whether CR is ble.nkcd out o:r. not, no date. ,\-Till

cn

be edited iuto thes;:; positions, where
present, but rathel· into the

to the left.

clli"~

The letters C and H

is

ch5.ractcrs
trlSy

be used

intbe rell1ainlng of the edit mask where they
"Till be treated as

no~.l

characters, ,{ithout being

alphabetic
3ubjec"~

to sign

control.

-(minus)

. This character is ha.ndled similarly to CR

in the r:tght-most position of t.he mask
field.

o

Otherwise it ic trea.ted

alpha.betic clla.rueter.

D.B

an

12.

o
C/J:,L r)~DIT

( cont • cl)

Control_.9har~~ter

Function

-(minus)

Uote:

cont'd

If neither CR or •. appears a.t the right
hand end of the J1lask field, the negative
sign on negat.ive numbers \'1ill be lost.
(See PPASS and PPASX for other

fOI""u1S

of

nega.tive slgn control)

*(asterisk)

This cha.racter operr.i.tes the srune

P..S

the

O(zero) for zero suppression, except tht'l,t
asterisks are inserted in the i'j.eld

r~ther

tha.n blanl-::s in high order posl tions, providing

c

asterisk check protection.
$(floating dollar

. sign)

This character has the same effect as the
O(zero) for zero suppreosion, except that
a $ is inserted to the left of the first

significant character
• (apostrophe)

fo~~d.

This character 1s removed and reple.ced by
a blank, which then rema.ins as nn alpha.betic
cho.racter.

If insufficient spa.ce is available to insert fl.ll significant
digits from the variable P into the edit mask,

c

wrt!

it"ti

It

i" trit .!!

. tit
t

t

#iiriddBe *rtttddtritW' *btltttdidt¥bMt t t ·itrtilb::1ti Ii'

dtrbt**** '* 'isi#rlW

:n:

o
tht:! complete edit mask wi.ll be l'cplc1ccci. by asterisks.

CALL IEDI! (I,IA,J,M)

I

=

Previously defined irf(:cccr vuriable,
to be cor.lbined wi t!l the CO:·mT array.

IA

=

Previously defined COivIET ar J.~ay,
containinG an edi t nlr.ct;.

left··most character of the edit
M

=

Humber of

column~3 eornpri~.;inG

lr,~.LGK.

the edit

mask.
Except for the mode of the variable tu be combined with the
CO 1·11': T array,

th'~

function of thi s subprogra1.l j.s ident ical to

PEDr.e •

CALL PPASS (P,IA,J·,N)

To convert the high precision integer variable P to its EBCDIC
equivalent, and storE: the resulting character string in the COl.1ET

arrny lA.

Lending zeros are suppressed.

If P is negati.ve, a

leadinG minns sien is added to the resul tine chal~actcr string.
If insufficient. field \lidth has been a.llowed, high order

o

Ih.

o
CAlJ.. PPASS

( COtrt • d)

characters are lost e
llotr:.:

1'10 Cl'":ror intlications 1a g1.vcn.

The identical. effect to PPASS

Cic).n

be oLtrt:tnefl. l>y the uco

of PEDrr, llit.h iniice;tiio!lo of. error.

shortc'r onc1 fnotcr ~ anG. requires

110

liollcvcr PPASS iB
t:::Gk to corrtl"ol the

transfC:rt.' ..

p

Pre"Viou~ly

c.cf'ined hiCh"'l)A"ccinion intccc):' v(I..l1ie.blc,

to be converted to EDCDIC.
IA

:::

CO:illJlI 8Xl"ay to conttJ.:Ln E.t!CDIC €\iuival c?:l1; of P ..

J

ft

Cho.J.'"actc:l.'

l)umber \f:tthin Ili to r:=trk lef."c·hund cud.

of resulting fielo. '"
M

::

Field vid-i;h ill J.A of l'"cthllting EBCDIC cha.racter f;trin6.

CAI.L IPASS (I,IA,J ,ll)

To convert the

integCJ~

va.riable I to 1t.o IIDCDIC

store the result:tng chnrac'ter

stl~1ng

equ1\J'l,~ent,

and

ill the CO:IGT f):r:l'i.).Y V.,

Leading zeros s.re rH.lpprcs5cd.

If P io lv::.[;ative, a

sign 10 added to the restlltiliS

ch~,.re..,ctel"

stl"1ng.

le~cl1nz

tdnus

If innufficient

field width ha.s been allo't1cd, high ol'lldcr chnrnctcrs (i.re lOBi;.
No error inclicat:l.oll in given.

I.

-=

Previously def1ne-d integer 'variable to be converted

to EBCDIC.
IA

a::

COHl1Ir

arrt~to

COlltain

m~CD.IC

cqu1valelltof P.

c
,:;*! /" .

wMttdi

M

'["Me

dtit±±bb

I

CALTJ IPl\SS

(con';;' (1)

J

=
end of resulting field.

M

&:

Ftield l;ic1.t.h ll1 lA of result:i.!1C

EDeDIC clmr<:\cter

CALL PPASX (P,IA,J..,!lU)
To convert the hi.gh precision int.eger

va~~lable

P to

its EDCDIC cqu1.vnlcnt, and s·tore the resulting cLtu'ac'ccr
s-'Gring in the CO:·m7l' c:t-ray IA.

The effect of t.his SUbpl"OgrRl"Il

is :tdcntical to PPASS, except as follC'f';1O:

1)

2)

If' P :ts nega.tive, an J..l

the

10;1

~OU(~

coc.le is

e.cld{~d

to

order chf3.racter of the resulting field.

in IA.

This flt1.broutine w:tll nOrIn-:llly be uocd. to convert hlgh pJ:ec:tf;lion
integers to EBCDIC
rcsu~t:tng

by

o

fo~

punching.

It should not be used if the

CO:rlE.r 6rray is to be printed., a.s

the printer.

0

is not recogn:lzo.ble

t.;

.-'

.

16.

o

Al":t.th·i:'
_
..................... t:t.C S·~'hl)l"O~l'~;.r~~~
... _.c .....
.... _ ......
_~~·,.

---.....~_

To mlltiply P x Q

n~ld

store the resul.t in R

c.,G·e,

hiell

precision intcacr v.· .r1o.ble.
R

c:

PADD

(p,Q)

To add P nni Q nnd ctore tho result'in n

R

n

0.0 Q bi@l

precision

mUB (p,Q)

To subtract Q

f'l~O'1

P aud. store

ther~fj\ll t

:I.n n

e,£:;

n high

precision integer variable.

R

I:

PDIV (p,Q)
To div1d.e P by Q and store the result :tn n
integer variable.

Tb~ re~ult

~.s

'a13 Q

high pl·ecird.on

rounded to the ncarcot

int~Gcr.

R = PDIVT (p,Q,)
To divide P by Q and store the result in n as

integer variable.

0.

h,iBh pl'cc1uio:a

The result is truncnted to the

n~~ast

integer,

and the remainder aftcrd1v1sion is lost.

Sets n to the
R

a:

r~"t3.inder

aftel·d1 vid1ngP by Q,.

Pltmo (p,Q)

Sets R to tho lessor ofP e...'ld. Q.

~ ,'"
:~,)

' ••....

IT! :-'

"e 7

. II'

R

=

.

FNJ'..XO (p, Q)

Sets H to the

R

so:

o

larg~r

of P and Q

PABS (p)
Sets R

o

l!ff,"wg"M'

t~

th8 absolulje value of P

18.

o
The purpone of these Gtl'bprosrc:;::s is to r(tlow overlap of' cru."d
rending end procc3s1ng.

They ohou.ld not be usc-d if cerd

punching is required on the

card as

Gru.",J.~

\tnr;

read.

CALL TIpPEN (D.roW J l,R)

Opens Bead Buffer IRBUF, lmviu3 length L.R 'Wol"da.

The f'irot

card in the rrod hoppO-t· is read and is v·tored in the buffer,
one Ch?l""'dctCl per word, in cSl'd code
1t

appear in

0. Dn.IEI!SIOH&ta:~erclcnt,

should not be uoed for any other
8ele~ted.

in any

loo:·ge enough that 6S

c~'trd

CAIJ, XREAD (JA

for~j~::J;'~.

IRDUF should

baving c.. size I.R+l, and
P't.'ll:pOSC e

17~y

reG.d operation in the

LR should be

colur:J1'lll aa are requ-ired

progl~

cV.n be

,n)

Transf.ers the data from InBUF to IA, converted to EBCDIC,
requil cd for
1t

c

sto~"ed.

Q.

cor,m

&l.'ray.

0.6

Ii, 'lhieh should be even, 19 the

DULlber of chara.cters transferred.

If N 16 oUil., U+l

chal.~cte:r8

are trans:f'er-,t1Ied$ the re:na1n1ns ch9.ractcr bei1lg fUled with a

blank.

After tranofer of clll'\l'acters, a

DCM

ea.rd is then read

into IlIDUF in overlapped m.cxle, rcadjr' for the next dn.ta transfer.

If XJ.1E.ID is called before E¢PF~ or m¢LD, then the

eom:p'.1ter "m:tts 'With AAOl displayed in the accumulator.
CALL rJl~LD (IHEm" ,LIt)

Identical to R~PEi:l, but does ))ot~ read a. new card.

Used to

c

Puw--w-n

19.

CAI,L mIpLD (cont' d)
re-init1aJ.1z(; the SUbl""outine on en-tcr1ng a. nC\'T link of n program

involving linked overluys.
These oubroutincr; uoe the library progrcEll

o

o

CJ~HDO.

20.

o
CALL QTYPE (IA, IT)

To accept data fro:l the 1131 Console kcybool"(l and to
this

a~ta

IA

in A2 for.cat to the

ar~

Pl"CSCD't

specified.

A one word integer array which is to receive the
data, frc:-1 the console kt'Yboard in J,2. forrJ.n,t.
ma.x1.l11Ul1l

The

pcmiosiblc length 1s 80 positi.ODS.

This ia en integer varia.ble or conote.nt 'Which defines
the number of characters to be

console ke.-yboard.
to be

anticip~tad

ente~ed

through the

This figure 10 (\ rnm.ximum f1(;Ure
for this

ke,yboa~d

operation.

Dctnilcd Description
This subroutine accepts infortlB.tion from the console keyboard,

typing each chal·acter as it is entered, accepting the entire
sequence of characters until. either the cherscter count as defined
by length 1s reached or an EOF character 1s entered.

This routine

provides for a carrier returne.nd line teed prior to each entry.
A

maxh...~

ot 80 characters is permitted. The output is presented

in A2 fOX"l&i,t.

All EBCDIC cba.racters are accepted.

uses IBM LIBF TYPEO subroutine and therefore

This routine

supports all

its

fea.tures with rC3pect to backspa.cing aud field ca.nce.llation.

The

c
;:.~

:7-."

21.

o
Cl\ I,TJ (l..::lPZ

( cont ' d)

entir"c 1nt(~i1nl 80 chnracter buttCl" is clcul"ed to blanko '

prior to each entry of keyboard data.
CAr,L

Q~~YHR

(lA, n) .

Fu.rtct~ion

To Pl':tl')t on the 113lConaole' Printer the eOlltc:nts of the A2
D"rrey contc1uillB the message.

Thla routine muat be used if

the IJrogl'dm alf~O contains QT'YPE and console typqlll·itel:' .output
ie . de.sired.

o

!1"'::.~~c~~ Descr:t~ion

IA

The nOJ.'ile of a one vord integer, arrey wb19h con~
ta1nsthe mesDage to be pl·1n~cd 011 the cOl1s()le
. type~iter ill A2 data. ,for1."'lD.t.

N

The integer variable or con:::,ta.nt which specifies
nULlber of cha.ractCl·s to· be transmitted.

be even :

'If

N should

it 1s odd then N·+ 1 ChtU.'Bcters will

be tranemitted.
Detv.i~~ cd D~scr~p.:!2:ton

T1::i1s routine USCG the IB!;1 LIDF TYPEO subroutine.

data. fron a. one ..Tord :J.D'tq~er

o

fori~::...~to

f'.tT'J.y

It acceptn

which data must be in A2.

1'hia me~llS packed tvo chnmc'ters per word.

A maxinl1.l!i1

--.•.

~.~.-

.. --,-.. -.-~----

. . '.,~
( C /",14"

t,i)
"

o
of 128 chnracters or 61~ "lords is a re~trict,ion to this routine.

Carrier return and res'l;orc are prov'ided before each IHcssage
is typed on the console typc\,1riter.
CALL QHRTY (IA,N)
F\U1ct.ion

To accept data from a one word integer array "Which is in A2
format and display tr~is information on the 1131 Con~ole
Typewriter.
Parameter

Descl'~"ption

Serne as QTYHR.

Detai1ed~cription

'l'his routine is similar to QTYVm except that. the IBM LIBF \olH'1.'YO

rout inc is als.o· used.

It must not be used if QTYPE is also

used in the same program for data input on the Console Keyboard.
Errors
The appendix B of the IE"! 1130 subroutine library listing

applies to the

Qjffi1~

error conditions.

c

o
ThiG cul)l"(Jutinc in d.esigned for :i.l'lcluo:J.on ill a cf'.rd inp'U.t,

printer output pro01'0'11 to proviclc Cl":cor Hc;sfic:.:e cnpab:tl:tty

on the 1131 CO!l301c P.:.:'1ntcl".

In t.hiG COl1:rieUl'Dt~_(;ll, it. usc::o

the lc:tst o.r:ount of core stOl'agC ot any of the typ,.:mritcr

routinoa.

Funct:ton

---'.~-

To pcrnit oJ.l :tnput/outPllt operationc to

o

cC':.1C

to

n!1

e~'lU

bcfo:... . c c, puusc or etcp atn.tc.1ci:ri; is cntcl'lcd.
.Pgrm·
e.,,·
---_·l.(;l;
__
·_,,_JD(;.nc:(~:t.nt:ton
__·.....
..
_·~_L

.....w.• , _ _

Hone
Dct~).11 c~l.Dc:::
crl nt.:t on
.....
_ _ ....... "..
......

...

~-'-

_~

~..:..l7~

Tbj.o routine

10

re.:j,uircd.

bCCo.U3C

the input n.od output snbroutines.

of. the intc;rrurr~ l1~:;Lul'c of
lThell 1t 1s a.cDl.red to

displr.y infol"Llation in t,he accumulr:J..to1" on tho console 1t is
necessary

th~t

all pending input or OU.tDut opel"ations be

bl·OUCht to an er!d 1'.H:::lOre the displfl,yeo. illf()r1:~.·).tj.Ol1 '\7ill

rcr:.c..in

in the accu':~·v~e.tor.' It is therefore ncccnr;o.ry J~o use the call

(IDI7'Ju:" Cl.l::1 CSP II subprogro,m)

o

o
p

r.:

PreviOilL!ly itcf'incJ hiCh l)rccinioll

Q

iii

Variable to be cO:Jpo.rc.d..

I

II:

Varic.blc to contain retul"ned

Sets to 1 if P i3

Sets to

grc~tc.:r

C6;:l:p~..'rc

VC.l\lO.

thn.n Q.

0 if' P cqunlo Q

0

Sets to ...1 if P is lens th!!n

To

intc~.;cr.

Q,.

t·,TO o11'):J.(,:::::1cric arrays IA cnd IB.

IA

I:

Prevlounl~r

J

-=

Colu1:m number lTithin IA at "h:i.r;h cc,nr>:3.r:tson

clefincd. CO:JBl' nrruy.

(le~t

to start

hnud end)

lB~

Prcviounly defined cor·!i];!' n"nY'.Y.

L

Colu!Jn

a::

numbCl~

within 11\

to start (le:rt haj:ld
}'in.i~1mum

I

-=

tluruber

Varj.ablo to
begins nt

or

COlrha,ins

thc-tll in

ftt

uhich

col'~1.rl

ehaructc~:j:'S

to be

~o:'ll't:.rcd.

vu.lno.

Cc::.:p::-·.y-iOOil

of 11:

C01U1'\.1.l3.

is found in the

a:rr~,y

e. ch:tracter \lh.:tch 10 htgl:l.C:l" 111 the collating
thr..~ cOl"rc~p(.)n~'11n~:;

a colu.mn 1£.1

fO\Ula

colmnn in B.

in A uh:ich hflO

10 .

J of ll..6.nd L of In, and

ln~'y':jrJ.uro.

colun~l'\

co~p:\rj.E~on

011(1)

con1~o.:i.n rctt~\rne!'l

continues for a
I is set equal. to 1 if n

i.~l

A llhich
Dc'.!u~':!nco

I 10 Get equal to -1 if

t\ ch:lro..ctc:l~

'Thich is 101<1el'

.".•..
IIIhiiWI."'I

'"."~'1IIIIIIiiIii~

_ ,#,"'1"1,,"%""
_ '''wrltH#'
_

""¢j":'" I' ":'2'1'$"

n

f thl ... I .. *j "db

thtbtttW"' #ebb "rib" riifn

.'hi ##t#±rtb¢*±_

nn

~---..~~~~~~~~~--W'

t $'1

,

751i""'"%('

'" ·"t"+±ditfhStiWfhii'ti

____

2: •

o

CCY'11X).1'e
S·tl.1r;)J~·o,r::.\":-:
;'.:.'._
_
.....'. _ _ _ _ _
_ _ .r.I: _ _ _ _ •. _ .... ,............

in B.

(cor.l"c' d)

I if! Cc-i:. equal to 0 :t:t" the

CXt'Y"i'.:,Yo

A nnd B f\~l'e found.
;l

to be ic1cnt1ccJ. in

t,h(~

colvnno cX!:Pllnccl.

•

The colletlr13
."1-$

! iI ,. I.~"

Y;

.'

sequC:Dce :tn v.sce:1(.linc order is cn follolTs:
blo.ph:,

.,

( ~ +,

e~,

$, ", ), -,
)",

/, , , •,

G, H, I, J, 1:, L, M, II, 0, P, Q, Rs; S,
0, 1,
I ~,.

2~

3,

l!~,

5, 6,

r( ,

A, B, C, D, E, ]1.,

Co': ,

~,

U, V, F, , X, Y, Z,

8, 9

IQBLK (lA, J ,1,1 )
~'est.f, the COI'/lET array IA for alphabetic blanks.

J

-

Colul11l1 numb(:r within IA at vlhich checking is
to start

M

=

NumbeI' of (!olurnTls to be checked

I

=

Variable to contain returned value

Set to

-1 if array is lO\>ler in collating sequence
tb.an alphabetic blt~nks (onl~r in t,egeJ:' numbers
can cause t.his condition)

Set to

o

Set to

1 if array is higher in collating sequence
than alphabetic blanks

if array is blank

r
.


-,
c-:n.

~'ritCe
S',~h-Pj~O·'f~'r.'.~J
--".
577M1.,.....;.,,~ \;......,........

o

CALL PLIST (p)

c·

r·

I:

milT

II

T3

:

2?.

o
CALL QPIIC H ( lA, t1 )
CO~v1E'r

To punch a

arrny

IA

=

Pr(;"i0l.lsly defined CO?-rds to regulc,.te t,'be re[ld ing !lnd 'Y1Titing of
records, uncl also is used for ussembJ.i.ng thC:' bJock of recard~,
for trHn~.:;fcr to nnc1 from the disk.
'lIte fo110\01i1'1B is the aI>r)cf"n~once

of the file buffer.
}i'~i.l(;

IBUF(l)
IET}FU~)

IBUF(3 )

()

Number

LOGical recOj.~cl lenr;Lh
Number of lor,ic8.l rccoTds in blocl~
File type
1 when bloc1\ has or.ly been lwed for
GE'l' operations (i. e. is idcnticrtl
to eq.ui valent rcc )rds on dis}\.)
= 2 \rtlen block h8S been used for
PUT operations
NmEbcr of fi.rst locical record in block.
Set to
zero if no reeol'd h[,s been transferred to buffer.
Blocked records

InrJF()~ )

IBUF(5)
IBUF( 6)
etc.

IBUF must HPIH:uT in the calling program \-lith dimension

LOGICAL RF;COrm LE.NGl'H

~:.

BLOCKING FACTOH + 5

Subroutine Cal2.Ll;; Sequences

IBUr-'
N

_.

,,1

=
=

LL

-

Call D¢PEl>T

o

Name of file buffel"
File number
Logical Hcco:t~d Lenc th
]~cngth

of file buffer

(ItUF,N,J ,L)

rfa open a file buffer.
Sets tnitie,l v[!lues to the fi1e control
words. No transfer of informaL ion to or frOl: di.sk tal,e s plRc e.

7.

.:.>

:)""e

Disk File
-.----.-.. -

-_

Hecord
DJ ocl\in{~
~~t~j')l'out ines
'--- ......... _--....._--... ------..--- ... - ---

Call DV~:r.)gN

=

=
=
=

(lEU1', N, ,1 , II)

CO!lt

'd

o

cont 'd

Name of f"1.l.c bllffcr

File nu.;:ibe!'
Loeic8.1 record length
Blocking factOl'

n¢r-'N and n¢pEH are alternative forms.

~rheir

effect is identicf\l.

Call DGET ( I BtJIi' , K, IA )

To transfer logical record K of ·the file to the arre.y IA.
If
the required. record is already resident in the buffer, it is
immediately t.rnncferred.
If it is not., the block of rcco:cds in
the buffer is stored if necessary, and the correct. block of
records obtained from disk.
Transfer then takcG ple.cc.
IBU£i'
K
IA

=
=
=

Nnmc of file buffer
Required logical record
Arra.y to conta.in·record obtnincd

Note that IA is integer.
If real values arc required., they can
be obtained. by UBe of a suitDble EQUIVAlENCE statement, in 'rlhich
real variables arc e.ssienec1 to EVEU locat.ions in t.he en'ray IA.
The real variables tht:n occupy the dl2signatcd loce.tion, end the
next lower locution in the array. e.g.

c

DIM8NBION IA(4o)
EQUIVAIJ!:NCE(B, IA, (16»)

B occupies IA(16) and IA(15)
Note:
The uSe of an EQUIVALENCE statement in this way is n.ot stl'ictly speakinz

permit.t.ed.

However, it works sat.isfact.orily, providing nothing is done
To

to force the addresses of real variables onto uneven \trord. numbers.
prevent this, the follo . . Ting rules should be Obscl'ved:
1)

Equivalf.:llCe only to even locations in the integer al'ray IA.

2)

The inteGer array IA sh<.)uld be dir(ICnsioned to an even number.
of words.

3)

If L,\ is in COlvG. '10N, then a.ny previous variables in COlvlf:!ON
together occupy an even nlJ!:'lber of words.

e.g.

CO!,~MON IX, 1Y ( 2 ) , IZ , IA ( 20 )
EQUIVALENCE (IA(2) ,n)

is va.lid

CQf,1i,10N IX, IlIl ( 20 )
EQUIVALENC}i~ (IA(2) ,n)

is not valid

c

tI"Irn t t t tt .... i,t

1M••• #1#&t" "

- lthtttffitf

- i#bfttW

*

r!ll

It

o
To tl'unsfcr the arl'etJ' TJ\. to logical record K of tIle file.
Op:::rn.tion is similar t.o DGET.
The content.s of IA are trPllsfel'Y'r:l
only eH far ns the file buffer, and not written on disk.
Gall DPTE; (lEU?, K, J.A)
To transfer t.he arro.y If.. t,;) logical record K of tbr file.
The opere,t.ion is similar to DPUr) except tr..3t it i s exp~C'ted
that the file will he written sequentially.
Hence a bloc1\:
of records i8 not read fralrl the 'disk befoJ'c transferr in b
data from IA to IOOF, in anticipation of later moclj fyinc;
all loeical records in the block.

CHI 1 DF:Um (JBUP, K )

To I)(lsition t.he (1...s1-: in the C0rrect position to p:coccss
loCical record K of the f5.le.
No transfer of inf'Jrrr.ation
ta)~es plo.ce.

Cell DCI¢S (IEUF)
To close the file.
If' a block or records requires transfer
to disk, the trunsfer is mede.

Call DGB'I'l(Il:iUF ,K,M, IX)

To transfer one vord of a logical record K to the location IX.
IBUF

--

Name of file buffer

K

=

Hequired logical record

1-1

=

IX

=

Hord nlun'ber W.l thin logical record K
data is to be obtained.
Variable to c:.;ntain returned valu(~.

frOitl

"lhich

Apart from the fact that only one word is transferred, t he
operation of this subroutine is identical t.o DGEr.r.\

To transfer one '\-l;)rd from

0

IBlJlc'
K

=

1-1

-

IX

o-

tht~

location IX to

th~

logical record K.

Nawe of fil e huffer
lJOgical r(~cord t.o be ITlodi.fied
Hord numuc;l" \7i thin. logival recol"d K into 'r!hich

data is to be stored.
\}ord to be stored

34.
Di~k

-- Subroutln~$.

File Record Blocking

~~,....~

..

",

.

o

eontfd

"'.

CallDPUTlcont fd

Aparttr9ln the :f'actthiltOnly 'one \1ord is transferred, the
operation of the suhroutinc'lsideutical to DPUT.

Call DUSE (IBUF,K,M,KEY,KL)
To generate a po1nter'toworclMot logical recordK.
If
the block contain ins therequlredlogical record is already
in the buffer, the pointer is gel1eratedimmediatcly.
If it
is not, tbeblock in the buffer . is stored if necessary, and
the required block obtained from the disk, after which the
generation of the pointer takes place.
IBUF -

Name of file butfer

K

Required logical l'ecord
Word ,reqv.ircd
Switch to indicate PU1--pose ,for which logical

M

KEY

KL

=
=
=

==

record is required.
= -1 norecol·d isbeil').g read or written.
Write tile bufter to disk if any record
has been modj,fied.
== 0 recordtare to be written sequentially.
=:
1 record is to be read
= 2recol'd 1s to be \lrittenrandom e.ccess
Pointer to -pas ition in IBUF of required '\-!ol'd

Normally PUS:E is not called by the user.
DPtJT, DGE'll,' DPTS , DCL¢,$, .. DP\J'l'l and .DGET1.
,

c

It isca.lled by
,

DPUT, DGET,DPTSandDcI46Sare secondarY'cntry pOints into
the subprogI"aJJl DUSE,WhichiS."T1ttcnin assempler language,

D~PEN and D~PIf are separate~sseJn;bler.lane;uage Subprograms
DFIND, DPUTlarid DGETl arewrltten in Fortran.

c

o
These ar-e required in core to sa.tisfy cnll£ from the nubprOgr3l~ls

listed e.bove.

PSFAC

l~veo

double "rord :from ACC and EXT to

FAC
QEDIT

(COME!' subroutine)

Edits alphabetic f:J.clcls

QER¢R

(COlvrn:r subroutine)

En'or l·outlne for COLlET.

Required to satisfy cull in QEDIT) bu'\i u(:ver

used, as the cond:i.t1cl1s li'hich cause

C.ll

e;rror in QEDrl' are teoted in PEDI':r before
QIIDIT is collcd.

QGHAD

()

(CO:,m,r r,u.broutine)
from a CO:,fET

EU~"[l,y.

Extracts sin61e C;lnr3.ctcr
C£:llnot be called eing test.d. involves the
operation by' cOlIlpiter ot a stor.._ .Y8$.eJIl, •,tor cOJalJercia1 vehicle. between
the paint and trim lines in one of our West Coast plants. A system. to
control a hiah bay warehouse for Itorage 01' truck parts i.-al8·0 under
development, The latter two 87,tema mq require replacement of the IBM
1130 by an IBM 1800 coaputer.
Se1ectionot hardW'a.recoflts.

c

EXHIBIT A

DATA PROCESSING ORPER

DUQUESNE LIGHT COMPANY

o

o,".r No.
Not",. of Work R.".,.,t.", _ _ _
M...a...t...e...r_i....a...l....s_a...,n_d......S....u_p...p....l_i..e..8______
O.port....nt R."".,tlnl Work:

Data Processing Department

Acco"nt To I. Chor..... __A..,c....,c...o._u._n._t__l_6_3....e...O_l_____________

Oat. Work Comr'.tlon R.qul,.'"

Continuing

Oat. Co,'.t.":

1402

p,..,.. , ... No. _ _ _ _ __
[]J
C.... lnul... D.'. 0,...,
Sp.clflc D.P. 0,.
NIA

o

D••crlptlon of Work'

Charge to this DP Order the time. material. rent and other
expenses incurred by the Operations Section in processing the complete
materials and supplies distribution application. Include such items
as preparing distribution check lists. material distribution reports.
stock ledgers and pricing out-files for Duquesne Light Company and
subsidiary companies.

o

S"",ltt'" "

0

Appr.,,"",

Manager.
General Accounting Section
Tltl.

1-~-6Z
0 ...

Woril Authorb'" I,

MaMler
Tltl.

D...

T'tl.

December 16. 1966
D...

EXHIBIT B

DATA PROCESSING ORDER NUMBER CODES

o

DATA PROCESSING NUMBER
Dept. or
Application
Division
lunction

01
02
03

Operations Function
Operations Division
G-1S operationa
Coal billing summaries
Radiation exposure reports

01
02

Development Function
Operations Division
Service interruption program
EEl Load Diversity Report

1

1

2

1

13

Operations Function
Sales Division
Billing data - cards and tape
Appliance saturation study

01

Development Function
Sales Division
Program conversion - billing data

01

Operations Function
Engineering & Construction Division
Tower Design - Phillips - North

01

Development Function
Engineering & Construction Division
Transmission tower design program

1

2
01
2

2
1

3
2

3
1
4

01
02
03
04
2

4
01
02
03
04

C:

Operations Function
Fiscal Divis~on
payroll and labor distribution
MAterials and supplies
Continuing property recorda
Accounts payable
Development Function
Fiscal Division
payroll and labor distribution
Materials and supplies
Continuing property recorda
Accounts payable

C

- fiT

EXHIBIT B
(Continued)

0
DATA PROCESSING NUMBER
Dept. or
Function
Division
Application
1

01
02

Operationa Function
System Planning Department
G-1S operations
Load flow studies

01

Development Function
System Planning Department
Finane ia 1 ana lysis program

01

Operations Function
Data Processing Department
Supervision - Manager and Directors

01
02
03

Development Function
Data Processing Department
Supervision - Manager and Dtrectors
General system development costa
Development Section Office Rent

01

Operations Function
. Personnel Department
Personnel data files

5

2

5

1
6

2
6

0

1
7
2

7

01
1
9

01
02
2
9

01

o

Development Function
Peraonnel Department
Opera tiona Function
Miscellaneoua
Labe 1 printing - Secretary Department
Forma control work
Development Function
Miscellaneous

EXHIBIT C
OPERATING EXPENSE AND CLEARING ACCOUNTS DETAil LEDGERS
--r'----,------,-----.,.----------

c ... Jlem or question arose ....•. due to our accounting strUcture,
there may be 100 different charge numbers open at anyone time and these
may vary from week to week. ,So if costs were charged, it would almost
have to be on a weekly basis ,and this. involves considerable more effort
than a monthly charging.
We also encountered the problem of budgeting.

Our manpower budgeting

is done for a six-month period and is done in man-months tO,the closest
tenth of a man";'month.

With our type of jobs it becomes almost impossible

to do any budgeting within these boundaries.
projects could not
times required.

c~~e

The people in charge of

close to estimating the programming and computer

This, by the way, was tried for one six-month period

and it was disastrous.

This signaled the demise of trying to do any

budgeting of programming time.
These factors all contributed to cause of an overhead shop.

A number

of years ago we were accepted as an overhead function by the cogizant
government accounting agency and this pretty well stifiled all more
recent attempts to do any direct charging.
I should point outtkat there is some direct charging within the

- 2 .;.

c

o

computer facility.

On any requested overtime, the labor charges are

charged to the requesting project; no machine time is charged, however.
I might add that this is not a great amount and only amounts to a few
dollars per month.

We use it mainly to make sure the work has be to

done during the overtime period.
These, then, are some of the reasons that we do charge most of our
charges to overhead.

They have worked well for us for the past twelve

years and certainly have made my job

a lot

easier.

I hope it continues

this way, but there are questions being raised on the feasibility of
it at the present time.

I do not foresee going to a complete turnabout,

but that maybe 25 or 30% of the charges will be made directly.

The

arguments given within this presentation are still valid and will
continue to cover most of our work.

o
R. H. Magee
6 September 1968

o

- 3 -

o

o

0'
SESSION Tl
SEPTEMBER 10, 1968

A SELECTIVE DISSEMINATION
OF INFORMATION SYSTEM
FOR MEDICAL LITERATURE

JAMES L. GRISELL, Ph. D.
ROGER GUDOBBA

o

APPLICATIONS/ BIOMEDICAL PROJECT
14 PAGES

COMMON MEETING
PHILADELPHIA,PA.

o

o

o

A SELECTIVE DISSEMINATION OF INFORMATION SYSTEM FOR MEDICAL LITERATURE
by
James L. Grisell, Ph.D. and Roger Gudobba
The Lafayette Clinic Medical SDI system has been designed to keep scientific
investigators routinely informed of the world's literature in any selected
area of medicine.

The system can provide either a current awareness of new

articles or a bibliographic search of an entire file of articles for references
relating to any topic.

At the Lafayette Clinic we use the system for the

literature on the mental illness of schizophrenia.
The current awareness portion of the system is desi.gned to work on an automatic
basis.

o

Each·investigator serviced by the system has on file in the Computing

Laboratory a list of key terms, designated as a profile, which define that
segment of the schizophrenia literature of interest to him.

The key te.rms use.d

in profiles are obtained from a dictionary of all terms occurri.ng in the entire
file.

Once a month, all profiles are compared against all new arti,cles entered

into the

sy~tem

during the preceding four weeks.

The investigator receives a

complete list of all articles which matched his profile.
IJ an investigator wants a bibliography, he prepares a profile which defines

his area of research interest.

When the key terms in the profile match the

key terms in an article, this article will be included in the bibliography.
This system has been designed to provide a maximum of flexibility in writing
profiles so that each investigator can define his area of interest with optimum

o

1

2

precision.

Continuing feedback is also provided regarding profile key terms

which are matched with article key terms.

The ultimate goal of the system

is to bring to the attention of each investigator only those articles which
are of interest to him.
Description of Entries
The ultimate success or failure of any current awareness or bibliographic
search system is the adequacy of the bibliographic reference data which is
used.

This data must meet two important criteria:

(1) it must cover the

world's literature of the area of interest as thoroughly as possible and
(2) it must contain an adequate, but concise, description of the contents of
each entry.
A source of bibliographic information which adequately meets both of these
criteria is the MEDLARS system of the National Li.brary of Medicine.

The

National Library is currently indexing approximately 25,000

containing

250,000 medical articles.

jo~rnals

rf~

\~

As each journal is received at the National ,Library

it is checked for articles of medical interest.
articles are routinely indexed.

For medical journals, all

For a journal such as Science, only those

articles pertaining to medical topics are included.
Each article is read by an indexer.

The indexer then assigns a series of

tags, or index words, which define the subject content of the article.
tags are words which appear in Medical Subject Headings (MESH).

These

The tags or

index words are also the subject headings under which articles are listed in
the Index Medicus.

To keep the size "of Index Medicus within reason, indexing

is done on two levels.

The first level consists of tags which are most repre-

sentative of the article's contents.

These are used for cross indexing in

C·,

3
Index

o

~ledicus

and are preceded by an

*

in a MEDLARS listing.

The other tags

are not used for Index Medicus but are available for searches of the MEDLARS'
master file.

The entries in the MEDLARS system are available from the

National Library on the basis of either a demand search, or a recurring search.
A demand search requires a list of tags and the dates within which the search
is to be performed.

A recurring search is routinely made on a monthly basis.

The Format of an Article
Each article in the MEDLARS system contains certain basic information descriptive
of the articles.

The following items are included:

(1) The author's name. If the article has multiple authorship all
are included in the listing, with the senior author listed first.
If there is no author (1.5%) it is listed as anonymous.

o

(2) The title. If the paper is in English the title wi.ll be listed
exactly as it appears in the article. If the paper is in a
foreign language, the title will be in English translation. The
language in which the article was written is indicated by a
standard abbreviation.
(3) The source of publication. The journals are given in standard
abbreviated form. The volume number, month and year o'f publica-·
tion and pagination are also given.

(4) Index terms or tags.

The index terms from MESH which describe
the subject contents of the article are also included.

Preparation ofMEDLARS entries
All articles received to date, and
for entry into our SDr system.

th~y

One of the guiding principles. of our system

is that it be as automatic as possible.
pre-editing of article listings.

now number 5025 ~ have been pre.pared

Consequently, we do a minimum of

Each entry is assigned a document number.

The order in which the entry parts are listed is consistent.

Each line of an

entry is given a sequence number and is labeled (A) author, (T) title,
(S) source, or (K) index terms (designated key terms).

o

keypunching and verifying.

(See Figure 1)

They are now ready for

· ..............-.•...•-.......-.".".-..

"~---~.-".-

.... ...
~

-,~--

..-

. . -..
-,

''''''"."

....... _,....."".,...... ... ..',
'

''

... ''''".,,,

...

... ,,,, .. ,, ,....

,

"''''''' ..,,,,''

...-..,,....

-.--.~~----

..

_-------_._...._--_._--4

Master Tape File
The articles are loaded on a magnetic tape in the follm'l1ing manner.
the article is stored in card image format.
maximum of 20 cards/article).

o

First

(There is provision for a

Secondly, the individual terms from the title,

author, source and key term cards are pulled out and stored in an array.
(There is provision for a maximum of 50 tenms/article.)

All duplications

and trivial words (such as and, or, the, etc.) are eliminated.
the number of characters for each term is computed.

In addition,

As you will see later

this technique of breaking down the article and counting the characters for
each term will save a great deal of computer time later on.

Thi.s tape can now

be used for the current awareness run and then added to th_e master file.
using MAGOP we can store 5000 articles on one reel of tape.

(See Figure..

By

21

Preparation of the Dictionaries
One of the most important aspects of our SDr system is the dictionaries.
Three separate dictionaries are maintained:

(1) authors (2) sources (3)

~ey

terms (from the title and key term cards).
Dictionary of Authors - Each author is included.

4266 authors in first 3800

articles.
Dictionary of Sources - Each source in its abbreviated form is included.

520

sources in first 3800 articles.
Dictionary of Key Terms - Since the goal of the SDI system was to make it as
automatic as possible, only single tenus are extracted from article titles.
To get mUltiple word terms would require the pre-editing of the titles and
designating which consecutive word groups should be treated as a s1.ngle term.
This is done in some SDr sys tems •

For example, one may wish to des.ignate

c

itHllltl!

r! [!

g:-

5

o

"social factors" as a two-word term.

However, this will appear in the dictionary

as two terms; "social" and "factors".

While this may be regarded as a short-

coming of the system, provision can be made in the construction of profiles
to treat these two words as a unit.
When dealing with the key term cards, mUltiple word terms are treated as one.
The purpose of this was to retain all of the terms as they appear in MESH.
6067 key terms in first 3800 articles.
The cwnulative total is kept for each term in the dictionary.
appearing for the first time are flagged.

Also, any terms

It is now a simple task to give the

investigators a listing of the new dictionary entries for the month.

If any

terms are of interest to the investigator he can add them to his profile before
the current awareness run for that month.

4C)

Since the articles on tape already have the individual terma extrapolated the
job of generating the dictionaries is simplified.

lS"ee Figure 31

Profiles
The ultimate success or failure of an SDI system is a function of the ease
and accuracy with which an investigator can define his area of interest on
the basis of the terms in the articles being searched.

This is done by con-

structing a profile of terms to be compared with the terms in an article.

If

the profile is a good one it will maximize the number of articles of interest
it finds and minimize the number of articles designated as interesting, but
which are not.

If the profile is not a good on"e, the reverse will then be true.

The profiles used in this system are similar to the profiles used in the IBM
" SDI system, but with modifications.

4C)

In the Lafayette Clinic Medical

system a profile has a hit level, which may range from -9 to +99.

snr

Each term

6

in the profile is compared with each term in the article and every time a match
is found the weight of the term in the profile is summed.

o

After the last

comparison is made, the sum is compared to the hit level of the profile.

If

the sum is equal to or greater than the hit level, then the article is designated
as being of interest.

If the sum is less the article is ignored.

In addition, tl;vO types of terms may be used:

complete terms and root terms.

A complete term must appear in the article exactly as it is in the profile
to result in an equal compare.

A root term, on the oth.er hand, will result

in comparing only as many letters in the article word as are in the root term
in the profile.

For example if "child" is designated as a root term in the

profile then it will result in an equal compare wi.th children, childhood and,
of course, child.

The use of root terms is a convenient way of encompassing

all variants of a term which may have one of several different endings..
One additional refinement in a profile is the use of modifiers:
not.

must and

A must modifier simply indicates that any time a must term is found the

investigator will get that article regardless of the hit level.
will do the opposite.

A not term

When a must term and a not term are both found in the

same article, the must term overrides the not term.
term can have either of the modifiers.

Either a root or a complete

(See Figure 4)

Searches
Since this is the most frequently used program in the system, a number of
techniques have been employed to decrease the time required to search the file.
First, the program handles one profile at a time and makes a pass through the
entire Master File of Articles.

This enables us to keep the output for each

profile separate without having to do a sort.
in core has the following advantage:

Secondly, keeping the profile

when the profile is originally read in

c

7
the program calculates the number of characters in each term.

o

Since the terms

in the article also include character counts this is used as follows:

For

complete terms the character counts are compared and if they are not equal
the alpha compare is not necessary.

For root terms the character counts

are also compared and the alpha compare is done only if the term iri the article
is equal to or longer than the term in the profile.
our own alpha compare routine.

Finally, we have written

Since the articles are stored with A2 format

to conserve tape and core requirements the use of an alpha compare routine such
as NCOMP (1130 Commercial Subroutine Package) would necessitate unpacking the
article terms first.

Our alpha compare routine will handle Al or A2 format.

(See Figure 5)
If the sum of the weights of the compare terms equals or exceeds the hit level,
or if a MUST term was found, the entire article is listed followed by the
profile term matches for that article.

o

term from the article is listed.

For root term matches the complete

Inclusion of the matching terms helps the

investigator make value judgments regarding the interest level of ,the material
a term brings to him and to modify his profile accordingly.

(See Figure 6)

Reprint library
If the system is to be effective 'in disseminating the world's literature on
schizophrenia to a group of investig,ators it is imperative that a complete
reprint library be maintained.

With approximately 40% of the articles i.n a

foreign language, appearing in some 25,000 journals, it would be of little use
if the investigator has to obtain the reprint by himself.
Consequently, we are also in the process of establishing a complete reprint file.
This has been accomplished to date by (1) sending reprint reque.stcards directly

o

to the author, e2l obtaining Xerox copies from Wayne State University Medical

8

Library, and (3) obtaining Xerox copies from the National Library of Medicine
for any articles that cannot be obtained by Steps 1 or 2.
to be both fruitful and worthwhile.

This has turned out

o

Of the first 3700 articles entered into

the system we have copies of all but two.
Conclusion
Although this system was designed for a specific medical research. area, namely
schizophrenia, it could easily be adapted to any area of any discipline.

Except

for the alpha compare routine, which is coded in Assembler, all of the programs
have been coded in FORTRAN.

o

''l'¥'R'''''\' ,

.. 1M

.tt

t

to.tt•• t".

itt#htiffti#s#±b""fffi""\

-

J

"TTl·

.... " Ii "Itwm

P·"·f

'W,-,.

9

FIGURE 1

o

Article Entries from MEDLARS after pre-editing
1
1
1
1
1

A
T
T
S
K

1
2
3
4
5

AALLL
(SYMPOSIUM ON SCHIZOPHRENIA-LIKE PSYCHOSES AND ETIOLOGY OF
SCHIZOPHRENIA. EXPERIENCES REGARDING THE TOPIC IN TANGANYIKA)
SCHWEIZ ARCH NEUROL PSYCHIAT 93,377-9, 1964
*SCHIZOPHRENIA, TANGANYIKA (1)

2
2
2

A 1
T 2
S 3

2

K

4

2

K

5

2

K

6

AARONSON BS
AGING, PERSONALITY CHANGE, AND PSYCHIATRIC DIAGNOSIS.
J GERONT 19,144-8, APR 64
ADOLESCENCE, ADOLESCENT PSYCHOLOGY, *AGING, DIAGNOSIS, *MENTAL
DISORDERS, *MMPI, *PERSONALITY, SCHIZOPHRENIC PSYCHOLOGY,
SOCIOPATHIC PERSONALITY

3
3
3
3
3
3

A 1
T 2
T 3
S 4
K 5
K 6

4

A 1

4
4

T
S

(GER)

ABELY P, LAUZIER B
(THE FATE OF THE CONCEPTS OF PERIODICITY, ATYPISM AND INCURABILITY
IN PRACTICAL PSYCHIATRY) (FR)
ANN MEDICOPSYCHOL (PARIS) 122,729-46, MAY 64
CLASSIFICATION, *MENTAL DISORDERS, *NOMENCLATURE, PERIODICITY,
PROGNOSIS, *PSYCHIATRY, PSYCHOTHERAPY, SCHIZOPHRENIC PSYCHOLOGY

ABRAHAM G
(THE PROBLEM OF MIXED PSYCHOSES) (FR)
ANN MEDICOPSYCHOL (PARIS) 122,481-90, NOV 64
4 K 4 CLASSIFICATION, *DEPRESSION, *EPILEPSY, *NEUROSES,
4 K 5 *PSYCHOSES, MANIC-DEPRESSIVE, *SCHIZOPHRENIA
2
3

~PSYCHOSES,

5
5
5
5
5

A
T
T
S
K

1
2
3
4
5

ABRAMS S
A VALIDATION OF PIOTROWSKI'S ALPHA FORMULA WITH SCHIZOPHRENICS
VARYING IN DURATION OF ILLNESS.
AMER J PSYCHIAT 121,45.... 7, JUL 64
DIAGNOSIS, DIFFERENTIAL, *RORSCHACH TEST, *SCHIZOPHRENIA

6
6
6
6
6
6
6
6

A
T
T
S
K
K
K
K

1
2
3
4
5
6
7
8

ABRAMSON HA
ANTISEROTONIN ACTION OF LS~25 AND OTHER LYSERGIC ACID DERIVATIVES,
FACT AND FICTION.
J ASTHMA RES 1,207-11, MAR 64
*ALLERGY, ASTHMA, AUTISM, CHILD, *HALLUCINOGENS, *LYSERGIC
ACID DIETHYLAMIDE, METHYSERGIDE (3), MIGRAINE, PHARMACOLOGY,
SCHIZOPHRENIA, CHILDHOOD, *SEROTONIN INHIBITORS, TOXICOLOGIC
REPORT (4)

ACHILLES M
(ATTEMPT AT A STATISTICAL DIAGNOSIS OF THE DRIVE STRUCTURE IN
7 T 3 PROBLEM STUDENTS) (GER)
7 S 4 PRAX KINDERPSYCHOL 13,177-81, JUL 64
7 K 5 ADOLESCENCE, AGGRESSION, AUTISM, CHILD, *CHILD BEHAVIOR
7 K 6 DISORDERS, EDUCATION OF MENTALLY DEFECTIVE, MOTIVATION,
7 K 7 PERSONALITY, PijBERTY, SEX, STATISTICS, THEMATIC APPERCEPTION
7 K 8 TEST
7

A 1

7 T 2

o

~,

;

'\.J
.

I-

."t..f

10

FIGURE 2
Article entries as they are used in the SDI System
1

A
T
T
S
K

2

A
T
S
K
K
K

3

A
T
T
S
K
K

AALL L
(SYMPOSIUM ON SCHIZOPHRENIA-LIKE PSYCHOSES AND ETIOLOGY OF
SCHIZOPHRENIA. EXPERIENCES REGARDING THE TOPIC IN TANGANYIKA)
SCHWEIZ ARCH NEUROL PSYCHIAT 93,377-9, 1964
*SCHIZOPHRENIA, TANGANYIKA (1)
28
SCHWEIZ ARCH NEUROL PSYCHIAT
6
AALL L
9
SYMPOSIUM
9
PSYCHOSES
8
ETIOLOGY
11
EXPERIENCES
9
REGARDING
5
TOPIC
10
TANGANYIKA
3
GER

o
(GER)

AARONSON BS
AGING, PERSONALITY CHANGE, AND PSYCHIATRIC DIAGNOSIS.
J GERONT 19,144-8, APR 64
ADOLESCENCE, ADOLESCENT PSYCHOLOGY, *AGING,DIAGNOSIS, *MENTAL
DISORDERS" *MMPI, *PERSONALITY, SCHIZOPHRENIC PSYCHOLOGY,
SOCIOPATHIC PERSONALITY
8
J GERONT
11
AARONSON BS
5
AGING
11
PERSONALITY
6
CHANGE
11
PSYCHIATRIC
9
DIAGNOSIS
11
ADOLESCENCE
21
ADOLESCENT PSYCHOLOGY
16
MENTAL DISORDERS
4
MMPI
23
SOCIOPATHIC PERSONALITY
ABELY P, LAUZIER B
(THE FATE OF THE CONCEPTS OF PERIODICITY, ATYPISM AND INCURABILITY
IN PRACTICAL PSYCHIATRY) (FR)
ANN MEDICOPSYCHOL (PARIS) 122,729-46, MAY 64
CLASSIFICATION, *MENTAL DISORDERS, *NOMENCLATURE, PERIODICIrY,
PROGNOSIS, *PSYCHIATRY, PSYCHOTHERAPY, SCHIZOPHRENIC PSYCHOLOGY
17
ANN MEDICOPSYCHOL
7
ABELY P
9
LAUZIER B
4
FATE
8
CONCEPTS
11
PERIODICITY
7
ATYPISM
12
INCURABILITY
9
PRACTICAL
10
PSYCHIATRY
2
FR
14
CLASSIFICATION
16
MENTAL DISORDERS
12
NOMENCLATURE
9
PROGNOSIS
13
PSYCHOTHERAPY

c

PHilP'

11

o

FIGURE 3
Samples from the dictionaries

AUTHORS
1
1
3
1

*

3

6
1
10
2
2

AARONSON BS
ABD EL NABY S
ABELY P
ABENSON MH
ABRAHAM G
ABRAMS S
ACHILLES M
ACHTE KA .
ACKER CW
ACKER M
SOURCES

14
2
4

o

1
2
6
1
1
117
7

*

*

BEHAV RES THER
REHAV SCI
BIOCHEM PHARMACOL
BOLL MAL ORECCH
BOLL SOC ITAL BIOL S~ER
BRAIN NERVE
BRAIN
BRATISL LEK LISTY
BRIT J PSYCHIAT
BRIT J SOC CLIN PSYCHOL
KEY TERMS

2
2
36
2
2
1
1
17
287
34

o

*
*

CHLORALOSE
CHLORAMBUCIL
CHLORDIAZEPOXIDE
CHLORIDES
CHLORMEZANONE
CHLOROQUINE
CHLOROTRIANISENE
CHLORPROMAZINE TOXICOLOGY
CHLORPROMAZINE
CHLQRPROTHIXENE

_ ..........." -_ _ .... _

... A ... "" ........ , ............... "~ .... ·J .......

12

o

FIGURE 4
Sample Profile
LAFAYETTE CLINIC MEDICAL SDI SYSTEM
BIBLIOGRAPHIC SEARCH OF SDI SCHIZ MADE ON 9/ 4/68
TAPE FILE CREATED ON 6/22/68 FOR ARTICLES
1 TO 5025
PROFILE DESCRIPTION
PROFILE 10000
MODIFIER
MUST
NOT

NOT

NAME SAMPLE PROFILE
WORD TYPE
COMP
COMP
COMP
COMP
COMP
ROOT

~ CO~ON

WEIGHT
1
1
1
2
1
1

MEETING

LOCATION

LC

HIT LEVEL

2

KEY TERM

C)

. ABRAMS S
CHILD BEHAVIOR DISORDERS
CLASSIFICATION,
DIAGNOSIS
NOMENCLATURE
RORSCHACH

c
0-

.'

J"

13

o

FIGURE 5
Sample run times on a 2 psec machine
Number of
terms in profile

o

4
4
5
7
7
7

16
50

Search made on 4600 articles

o

Number of
articles found

Time in
minutes

10
14

10

2

14
11

63
421
178
446

lJ
4.P
41
72

41

ib

14

o

FIGURE 6
Articles from Figure 1 which match the Profile in Figure 4
2

A
T
S
K

AARONSON BS
AGING, PERSONALITY CHANGE, AND PSYCHIATRIC DIAGNOSIS.
J GERON! 19,144-8, APR 64
ADOLESCENCE. ADOLESCENT PSYCHOLOGY, *AGING, DIAGNOSIS, *MENTAL
K DISORDERS, *MMPI,*PERSONALITY, SCHIZOPHRENIC PSYCHOLOGY,
K SOCIOPATHIC PERSONALITY

MODIFIER

WORD TYPE
COMP

3

2

KEY TERM
DIAGNOSIS

A
T
T
S
K

ABELY P, LAUZIER B
(THE FATE OF THE CONCEPTS OF PERIODICITY, ATYPISM AND INCURABILITY
IN PRACTICAL PSYCHIATRY) (FR)
ANN MEDICOPSYCHOL (PARIS) 122,729-46, MAY 64
CLASSIFICATION, *MENTAL DISORDERS, *NOMENCLATURE, PERIODICITY,
K PROGNOSIS, *PSYCHIATRY, PSYCHOnlERAPY, SCHIZOPHRENIC PSYCHOLOQY

MODIFIER

WORD TYPE
COMP

COMP
5

WEIGHT

A

T
T
S
K

WEIGHT
1
1

KEY TERM

CLASSIFICATION
NOMENCLATURE

c

ABRAMS S
A VALIDATION OF PIOTROWSKI'S ALPHA FORMULA WITH SCHIZOPHRENICS
VARYING IN DURATION OF ILLNESS.
AMER J PSYCHIAT 121,45-7, JUL 64
DIAGNOSIS, DIFFERENTIAL, *RORSCHACH TEST, *SCHIZOPHRENIA

MODIFIER

WORD TYPE

MUST

COMP
COMP

NOT

ROOT

WEIGHT
1
2
1

KEY TERM
URAMS S

DIAGNOSIS
. RORSCHACH TEST

o

o

o

o

Imll

o

1) .....".-&!
l) "'.r 'D
.,r~•
..

~.\.

i

~

J •

~
ft
'.
lJ .', • U"'IQVlsen

®

o

Il

Cfi-e ~ T 4 Ct:) L

'PJ~T:-l

g'

o

-

.....

".-'

..'y. "~

•F .
,,' t."

•.:

~.

...... :,." !r;J'- '...
~"""Ii' •• ,,~ . . . . . . . . . . . . . . .,.•• ~~~. . . . ~~ ....... _

•","'''

\

""'

,

•. ,!'

.,'

. . . \.' • .....;~t.

If....

.l.~ }
~'.'

\~

.••.•

o

1'\

.f

. . . . .

<';- -t'·
,~i'.~.'''·.rttf#f$tb Ct{ t

rlMrtH w" titrlettHttbtWt.f T r r

in

o

NAHE OF PRIME COW1ITTE:
SUR.TECT:

Data Reduction Techniques in a ClinicCll Laboratory

SPEAKER' S NA~1E:
C011PAr~\TY

Kiomedical

R. I. 0' nesky, ~'1arion Ball, and lL R. Stewart

SPEAKER REPRESENTS:

Universi ty of Kentuckv,

MAII. INC ADDRESS AND PHONE NIP·mER:

DAY,

o

o

TI~E A..~D

SESSION NUMBER:

NUMBER OF PAGES OF TEXT:

~edica1

University of Kentucky
I\fedica1 Center
Lexington, Kentuckv 40505
606-233-5000

Tuesday, 10-11 a.m., T2C

10 pages

School

o

c'

o

Data Reduction Techniques in a Clinical Laboratory

by R.I. Q'Oesky, Marion Ball and

\~.B.

Stewart

Within any clinical laboratory there are many problems caused by
automation.

At the University of Kentucky Clinical Pathology Laboratory,

Or. W.B. Ste\"art, Chairman of the Department of Pathology, is using a computer to assist him in more efficiently running his laboratory.

The

equipment Dr. Stewart choose to use in his lab is the ISM lCOO Data
Acquisition and Control System Computer.

The purpose of this paper is

to discuss the most efficient us·e of the facilities. of this system.
T~e

the 1800.

process of data acquisition is the

~ost

important function of

Since this acquisition occurs on a cycle steal basis of a

data channel, the 1800 Central

Proce~sing

Unit docs not use significant

amounts of time servicing this analog point recognition •. Ther~fore,
within the system, the data acquisi·tion is relegated highest priority.
Within any data acquisition computer there are three sources of
storage available.

These media are magnetic tape, disk, and corc.

Since magnetic tape is bulk storage media primarily used for
extensive·master files it will not be considered as critical to the
laboratory data acquisition system.

Its relativ.ely

510111

speed and se-

quentially organized files are not appropoe for t:he volatile acquisition
system.
The fastest storage medium available is core.

o

The problem arises

\'/hen it is realized that core is limited to 32~168 16 bit words of
1
storage.
Since it is desirable to take advantage of the IBM supplied

ISX 2 programming system, a prime consideration is the maximization of
variable core.

This means that the user would like to make the most

efficient use of the core storage which the Programming System requires.
This then allows the user to have the maximum amount of core storage
available for his programs which run as a foreground job in the timeshared environment.
The other storage medium is the replaceable disk cartridge.

This

means that limitless storage facilities are avai lable on these interchangeable cartridges.

't

In order to maximize efficiency in the system,

is desirable to keep the disk files as economical as possible.

This

is because the disk access time is much slower than core storage cycle
time.

An economical file reduces the number of disk seeks and reads,

iesulting in shorter access times for the pertinent information stored
there.

By keeping the files on a single .disk cartridge no operator

intervention is required.

This means the system is never waiting for

a.lIman response in order to carry out its function.

Also. economical

files allow more working storage to be available to the foreground programs running under time sharing.
At this point the speed, accuracy, and efficiency of storage media
for the system are to be considered.
The ideal situation would be a point by point representation of the
autoanalyzer output. 3 See Figure 1.

This sequence of points would then be

retained by one of the storage media.

Since core is limited, this implies using

disk storage to save all our point values.
scan basis, we would become disk
in a matter of minutes.

boun~

Operating on a continuous

and would exhaust our disk storage

Therefore, physical limitiltions prohibit the

o

page three

o

use of this method.
Two alternatives are now available to faci litate the user's handling
of the large qua~tities of data generated by the autoanalyzer scan.

These

alternatives are, first, a hardware data reduction technique and, second,
a software

4

technique.

Two distinct hardware approaches \'Iere considered.

The first hardware method takes advantage of the comparator
feature of the 1800.

5 See Figure 2.

Since the range of the analog

input signal is known, it iS,convenient to have the computer calculate
2% of this range.

6

let this 2~ of the range value be called DElY.

The lower limit of the comparator word is set to the base line value which is
determined by an analog read preceeding the initialization of actual
testing.

Let this value be called B.

parator word is set to B + DElV.

Then the upper limit of the com-

As we get interrupts we save the

B + DELY value and update the comparator word by replacing B
with B + DELV,7 At the position on the curve where the comparatqr

Interrupt~

by going below the lower limit, we save the 6 value and replace

B by B - OELY and ~:+ OELY by B.

In this manner we have defined

the curve by a much smaller number of points.
This is a practical sol~tion to defining the curve, but a large
section of. the disk is wasted by saving non-significant points.
an inher'e~t error of 2% of the range appears in the answer.

Also,

Thus, it is

desirable to consider another hardware solution.
.
.
8
T~e second hardware ~oJution uses the interval timers which are
available in the 1800. See Figure 3.

The timer will trigger an interrupt at

predetermined intervals., Servicing the interrupt consists of reading

0.

the analog signal, saving this value in either

a core or disk

table, and

page four

returning the computer to the status it had prior to the timer interrupt.

o

If the autoanalyzer tests are being run at a rate of 60/hour, this
means that one peak a minute will result:
second will result

i~

A scan rate of once per

a table of 60 points per test.

Again with this method,

a large number of superfluous points are being saved.

Thus. the

~'ost

efficient

,

use of storage media is not being affected.
Thus far only methods of defining autoanalyzer

outpu~

as a

sequence of points have been considered. Now the problem of picking
9
the peak value of the curve must be considered.
The most straight-forward method· of picking peaks is to scan the
ARRAY of polnts \

A spike or a shoulder

10

is not detectable with this

approach.
In viewing the entire system, the most desirable condition would be
the investigation of every point on the curve," as fn

t~e

first scan

n~thod

mentioned, and the saving of only the significant point for each

test.

Before formulating an approach to combining these two. methods, let

us recall the definition of a.derivative from elementary calculus.

c

page five

o·

Consider a function f, defined for values of the variable x in
the interval (atb).

Let x be any fixed point of the interval and

consider the ratio

f(x) - f(x o ) where x ; Xo and x is a variable

x-x
. 0
point of the interval.
. .I on: II
O
. e f ·I nit

The ratlo is called a

differenc~

qUOtic1t.

If the difference quotient approaches a limit as x

approaches xo' the limit Is called the derivative of f at x=xo and Is
denoted by f'(x o ).

Thus by definition
f' (x ) -= 1i m

f (x) - f (x )

o

0

provided the limit' exists.
, If x=x·o+A .x the defi nit ion becomes
f' (x) •
o

1110

o

~~-=-,O

-

Ax

This then becomes theequatJon which is to be considered.

In the chain

scanned system mentio'ned as the Ideal situation the analog values
corresponding to f(xo+Ax>' and f(x o) wou'ld be saved in core.

These

t,,,o values are the only data points \'Ihich have to be saved.

Thus our

core requ,rement is drastIcally trimmed by comparison ,to the ideal sItuation.
The Ax is very small
con$idered as an

12

and as a result

appr~xl~atl,on

th~

difference

quotlen~

can be

to the derivative.

Knowing that a validapproxhnatton to the derivative

Is

easily calculated

within the 1800, it is now advantageous to consider the classical applic-

o

ations of the derivative concept. See Figure 4.

In thrs sitvation it \'1ould

· page six
be convenient to be able to predict the value of f{x o+2L)x) from the given

two data points f(xo ) and f{xo+L.x).

lHXo+2 .iX)/~
spikes.

f' (xo+2 DX) +

e,

o

This is easi'ly dO!lc by sayi'ng that

and this Is the cri terion for el imlnating

Since f'{x} is calculated for the previous purpose, further usc

of this value would increase the efficiency of time spent for doing

The second use would be to determine when the curve

the calculation.
approaches a peak.

The phys ical peak occurs at

Xo

\"hen f'

(xJ : :

O.

Thus

only points which have to be saved are in an area of the curve where
f • ( Xo ) -L cC

f or c-very
~
11 13
sma.

Th'IS, t h\en, minimizes
•••
· number 0 f
tile

values which are to be saved on the disk.

The criterion for the

completion of saving points is that f'(x ) < O. Physically this means
o
that the apex of the curve has been passed. The table of saved points
14
is then scanned to find the maximum value.
This max value serves as
Input to the inte~polation routine which, on the basis of the standards,
yields the final reus Its

of

the laboratory test under consideration.

This final result is then collated with the patient record to result in
the final report.

Summary:

This paper is the result of an attempt to maximize the

efficiency of usage of both core and disk storage for an IBM 1800 OACS
used in a clinical laboratory.
chronologi~al1y

and by

and

calculattons~

culmina~es

The development ·of techniques is treated
with the

'.

deri~ative

method, which, experimentally

appears to yield the most efficient and economical

result.

c

I.

0\

The IUti 1800 DACS cornputerco,lc.J in core sizes of 8,192, 16,384, or
32,976 16 bi t "'lords.

2.

Time-Sharing Executive System monitor

3. The autoanalyzer is the instrument in the Cfinical laboratory
which is the source of the analog signal which the 1800 recognizes.

4.

Software is the user's program. to direct the computer as to what to do.

5.

The comparator performs sele'ctive checking on the digi tal values converted by the ADC.

A range type check is made to confirm that the
,,'

converted va 1ues are wi th i n spec if i ed 1 imi ts.

The 1 i mi ts are ob ta i ned

from the Multiplexer Address Table (one P-C cycle delay allows both
limits to be acquired) whenever a check is required.

The P-C is

informed of anout-of-limits condition by interrupt.

Def. taken from

IBM

0-

6.

JaDO

Functional Characteristics manual A26-5918-5 page

2% is an arbitrary value which can be as large or small as the
program~er

\

77.

desires.

It is also the maximum error value for any peak

readi n9.

7. This only considers the case for a increasing curve, but an analagous
situation

exi~ts

for the case of a decreasing curve.

8. An interval timer is an 1800 hardware feature which acts as a clock
to keep the computer informed of time status and conditions.

9.

The peak

~alue

is the maximum valid value on-the curve and represents

the test result which is the desired output of the entire testing

l'I

sys tern.

Il-~P_'
'V'" j f