Proceedings_COMMON_Meeting_Cincinnati_OH_196709 Proceedings COMMON Meeting Cincinnati OH 196709
Proceedings_COMMON_Meeting_Cincinnati_OH_196709 Proceedings_COMMON_Meeting_Cincinnati_OH_196709
User Manual: Proceedings_COMMON_Meeting_Cincinnati_OH_196709
Open the PDF directly: View PDF 
.
Page Count: 566
| Download | |
| Open PDF In Browser | View PDF |
PROCEEDINGS
C OM~10f\' ME E T I NG
C I f\lC INN AT I,
nHI 0
SEPTEMBER 6-7-8,
1967
TARLE OF CONTENTS
PAGE
o. () • ()
F''''
1.
FRS ~'1 FF T I ~.I(;
\/.1 • 1 • 1
GF f\1 F R. f\ L S F. S S I 01\1
It/ • 2 • 2
1. 1. 30 lJ S FRS ME E T I r--.J G
\:') • ? • Lt- [. \,1 • 3.4
36 () USE R S [VI E E T I ~.J G
r)~ 136() PROJECT tv1FFTI ~.!G
\;../.2.6
1 1 3 0 I. I SE R S ~~ F E TIN G
1,·J.3 .2
~YSTF.M ~FF~RENCF LIRRARY PAPER
v.l.3.6
n~/360 COMMITTEE MFETING
~··I • 3 • 7
t,'J. 4. 1
J. A ? 0 lJ S F. R S M F F T I "I r,
] 130 It SF R S ME F: T I f\JG
\../.4. ?
D.n.s. PROJECT MFFTING
1,·/.4. itF n I. I C /\ TIn 1\1 P R () ,j E C T ME E T I r\1 C;
1".4.6
f'
·1T II SYSTF ~11 2 40 0 A. T,l\ RET R I EVAL SYSTEM
\r. 4. 7
~.I
M F MR
2
15
16
17
18
19
29
30
31
32
33
34
f •
T• 1 • 1
T. J • 2
T • 1 • '-'.
T. 1.6
T. J.A
T .l.A
T. ?
1
T.2.3
T.2.4
T.2.5
0
T.2.5
T.2.5
LII[\IC~
T. 3. 1
T.3.2
T.3.3
T.3.4
T.3.4
T.3.6
T.4.1
T.4.2
o
n I S K n./\ Tf\ STOR AGE R0 IJ T I f\J E I [) (") SRI
nVFRL,t\PPFD 1/0 FOR. 1130 FORTRAI\I PRnGRAMS
RFPOPT OF 360 HARDWARE COMMITTEE MEETING
PF. R SO f\i N F= L S FLEe TIn [\1 A/\J 0 J 0 R nESC RIP T I 01\1 S
Tn-SMALL INSTALLATIONS
P R nGRAM r~ FRS E LEe TIn f\ 1
80
83
109
I 1\1 MEDIU M-
114
120
n R G /\ h.l T7. f', T I r) N A1\1 () S 1== L E CTIn N
L I I PRF S t:: ~.! TAT I n 1\.1 R Y I 8. M
1\\ rJ fII\ J ~ 1,6 T I f',l G C0 rv1 MIT TF E ME E TIN G
122
1 36
13 7
IPMt::PG' FAST SORT-MERGE FOR 1130
PRO FES S I fJ ,\, ALP R0 GRA~1 MER S A~H) ANA. L Y STS - PRO B LEMS
TN PERFORMANCE EVALUATION
PROGRAMMFR E\lALI.JATTON
PRO r-. R A tvl ~/1 F REV A L U A TIn N
ADDRFSS BY MR. G.W. WOERNER, I.R.M. VICE PRESIDENT
D.D.S. TIl
OS/360 MEETING REPORT
FORTR~N PROCESSOR FOR ORILLIMG TUBESHEETS ON N/C
MACHINE
CHEr--iT CAL ENG I NEE R. I 1\' GAP PL I CAT I 01\1 S 0 r; 'C S iV1 P ,
CUST0 ~'1 ER- I Rfv1 RELA. T ION SHIP HIT H RES PECT TO FE, S E ,
AND SALFS
'360' OPERATOR TRAINING
n P EI\I ROAR 0 ME ET I "IG
Pl.. A1\1 ~·lJ NG S ES S JON FOR E0 UCAT I C)f\1 PRn J ECT
139
P
202
207
20 8
224
231
240
241
254
258
278
299
3 27
SESSION
F.1.1
F. 1.2
F.l.4
F. 1.5
F.l.6
F.1.7
F.2.2
F.2.3
F.2.4
F.2.5
F.2.7
F.3.1
F.3.4
F.3.6
F.3.7
TOPIC
'360-30/40' RPG FOR FUN g PROFIT
REOUIREMENTS FOR THE PROGRAM LIBRARY
A COORDINATED STUDENT DATA SYSTEM
LABORATORY AUTOMATION BASED SYSTEMS
FUTURE MEETINGS PROJECT REPORT
'1620' PROGRAM FOR A.C. CIRCUIT ANALYSIS
OBTAINING MAXIMUM RESULTS USING A COMPUTER FOR
ENGINEERING DESIGN
SUBROUTINE 'DATAR'
AN EXPANDEO EDUCATIONAL COMPUTER SYSTEM-THE '1041'
IRM PROGRAM LIBRARY PROCEDURES
COMPATIBITY OPERATING SYSTEM MEETING REPORT
STORAGE & RETRIEVAL OF PERMANENT 1620-1311 FILES
360 PROJECT MEETING
COMPlJTER TIME STlJDY ANAL YSIS FOR WORK MEA SUR EMENT
05/360 COMMITTEE MEETING
1130
Af\.ID QUICKTRAN DEMONSTRATIONS
PAGE
32,S
360
363
384
401
403
C
430
436
448
459
464
466
474
476
488
489
LIST OF REGISTRANTS - ALPHABETICAL ORDER BY LAST
NAME
492
LIST OF REGISTRANTS - ALPHABETICAL ORDER BY CITY
525
c
0'"
I~
II
I"
c
SESSION NUMBER
0
SPEAKERS
JIM STANSBURY
LAURA AUSTIN
CHARLES MAUDLIN
DISCUSSION
WELCOME TO NEW MEMBERS. EXPLANATION OF KINOS OF SESSIONS AVAILABLE. SUGGESTIONS OF SPECIFIC SESSIONS OF INTEREST TO NEW MEMBERS.
DISCUSSION OF INFORMATION TO GIVE TO SECRETARY-TREASURER TO
OFFICIALLY BECOME MEMBER.
c
o
/,
SESSION NUMBER W.l.!
GENERAL SESSION
SPEAKERS
JAMES STANSBURY, PRESIDENT
c
C;'
',I,",
-
o
IB.A.N S C RI PTI ON
GENERAL ASSEMBLY
COMMON, CINCINNATI, 6 SEPTEMBER, 1967
c
o
Page 1
JAMES STANSBURY r CHAIRMAN At Boston l I took off on the members enthusiastically, wildly, and
with not too much result, but I can't preach to the people that attend
today. We Ive tried to set up a meeting that meets the objectives and
requirements that you people gave to us at Boston, and in conversation
afterwards. Welre still making mistakes, but I think we've done
far better than we did then. Hope you will enjoy it and benefit from
it.
lId like to say one thing about the interests of COMMON. live had
numerous inquiries from people who aren It certain whether COMMON
is what they want to join, or not. They would join GUIDE possibly;
if they had a large machine, they might join SHARE. They aren It
qualified for these on the basis of machines I so they say, fine I we'll
join COMMON. The special groups in this organization are the
ones you people want. If you people have a common interest then get
together with a birds of a feather session, organize a project, and,
if you convince the Executive Board that you are serious, you III get
to be a proj ect. The people here are the only ones to make those
projects run - make them beneficial. That's about all I intend to say.
I'll start off here by introducing members of the Executive Board, then
I intend to have each one of our Divisional Managers present a short
description of what his division is doing. Jim Tunney has some corrections on the agenda and IBM has requested some time to make a
short presentation, "it/hich they ,will think will be of interest to all
members - a new hardware announcement.
On my left over here is Dick Pratt, Executive Board Member next to
him is N orman Goldman, Pres ident of the Eas tern Region. Paul Bickford,
who was appointed to fill the vacancy created by Don Jardine's resignation. Bill Lane, Western Region President. On my right, Frank
Maskiell, Executive Board Member I and on the far right, Chuck Maudlin,
Secretary-Treasurer. Since Chuck is a rather ethical character 111m
going to apologize for him. There have been a great many problems
with communications with the Secretary-Treasurer - lack of installation
support. He's changed installationsi he assures me that he has his
backlog down now to a reasonable value, and that in the future we
can expect reasonably prompt responses from him. Is Laura Austin
I
o
Page 2
here?
Jim Taylor?
Weill start out with the Applications Division since we have
the Division Manager present.
FRANK MASKIELL I wasn It prepared to dissertate on the Applications Divisions. We
have six proj ects in the Applications Division. First, I might state
the objectives of the Applications Division. You all will have a
chance to read them more accurately in the COMMON Reference
Manual, which will be forthcoming shortly, I understand. In the
Applications Division/, we are particularly concerned with problems
of installations which are not particular machine problems. We Ire
concerned with what you Ire doing on tlle machine, on the computer, not
what computer you are using to accomplis h the res ults of thes e problems.
To this end we have these six proj ects, specifically I they are the
Techniques Project, which is concerned with mathematical packages,
some statistical work. We have the Electric Utilities Project, what
has operated in the pas t, particularly in the Eastern Region, as the
Electric Utilities Team. We have the Petro-Chemical Engineering
Project, hopefully serving the needs of installations in these industries.
The Civil-Mechanical Project r who at this meeting I believe, will
be meeting withsome of the members of Cepa. VIe have the Education
Project, which gathers together the large number of university people
who, in some sense of the word, have problems all of their own.
Meeting at Cincinnati, for the first time, are a fair number of individuals
interested in numerical control. This is a project which is just now
getting started. Vve hope to see it grow as others find information
available in this area, and can be served by this as a project.
I
CHAIR An now that Laura Austin is here the lady sitting on the far right, 11m
going to ask her to discuss the Administration Division next. She also
has some announcements to make.
I
IAURA AUSTIN Thank you Jim.
o
I would like to point out some of the Sessions to be held for the Administration Division first. The Administration Division is a working
division of COMMON. It is one from which you, as a individual, will
not realize a great deal of benefit for your installation, in terms
Page 3
programs that you have obtained through participation in a project.
It is a division in which we feel you can contribute a great deal to
COMMON, and thereby benefit your installation through the overa11
operation of COMMON, and through what we hope will be an efficient
operation of COMMON. The first session I would like to ca11 to
your attention is at 10: 30 on Thursday, which is the session on nominations. We will be I within the next year, drawing up a slate of
nominations for officers of COMMON. Since COMMON has grown
to be such a large organization, it is difficult for us to get around
to know each member. We would like to have people who are interested in serving on the Executive Board of COMMON or who
are interested in serving as Chairmen of Committees or as Managers
of Divisions, to make themselves known at this nominating sessions I
so that we can become better acquainted with you, and so that we
can have a slate of people to draw from for our nominations for the
next year. We feel that you will benefit a great deal from taking part
in COMMON from getting right into the working side of COMMON
whether it is as a Committee Chairman I or as a Proj ect Chairman, or
as an Executive Board Member.
c·
I
I
I
I
The next Session that I would like to call to your attention is at 8:30
on Friday. This is the Future Meetings Session. In this one we
will be discussing the locations of future meetings the cities and
hotels, and we also would like to solicit help from members of
COMMON for Program Chairman and Local Arrangements Chairman.
Again, you contribute a great deal to COMMON by offering your
services in this area. If you will be at all available for any of these
positions, we would like to have you come to this meeting on Friday.
We can give you more detail as to what is involved in carrying out
these duties at that time. Then, at 10:30 on Friday, is the Program
Library Project. We are hoping this will be one of the largest sessions
of the conference. In this particular ses sion, we will be discussing
the use of the PREP forms, what the PREP form is, what its purpose
is, and how it can be of help to your. We will be discussing the
Program Shipment Analysis form, in which we are trying to work with
IBM in finding out where the problems are in the shipment of programs,
and in the distribution of them. We will also be discussing the new
.360,1130,1800 Library, and I'm sure this is going to be of interest
to most members. We will be discussing the standards for submittal,
the procedures for ordering, and what catalogs will be available, etc.
I
I
Another part of the Administrative Division has been the Reference
Manual Committee. This has been largely made up of one person Mr. B. R. Russell, who has contributed a great deal of time and
effort in preparing for us a reference manual. I'm happy to say that
today, I have a preliminary copy of the Reference Manual for COMMON ·
o
o
Page 4
o
This will be mailed to every installation in about two weeks. You
should be receiving the complete binder I dividers and the material.
Now this first issue of the Reference Manual is not complete. There
are some sections that have not been submitted yet. We felt that
we wanted to get it out to the members as soon as possible. The
things that it does contain are something about the history of COMMON,
what are the advantages of membership I the obligations of membership,
what is the coming calendar for COMMON meetings I the organizational
structure. We hope here to have completed before too long a list of
all the projects that are currently active in COMMON - their scope and
objectives. The Reference Manual will not include current reports on
project progress. These will be covered in CAST but the Reference
Manual will give the scope and objective of each project; it will give
you a list of all of the Project Chairmen, so you will know who to contact when you want to have some correspondence regarding a project.
It also contains information for you about the information services of
COMMON. In other words I how do you submit things to CAST, what
can you expect to see in CAST and the Newsletter. How do you submit
information to the Newsletter, and what is the purpose served by the
Newsletter. And some information regarding the Program Information
Dept. I better known to you as the Program Library. This will talk
about the ordering procedures for material from the different machine
type libraries I will talk about submittal procedures for programs, the
PREP form, and the shipment analysis form. We hope the Reference
Manual will be of great benefit to you. There is one other thing it will
contain I a membership list which is indexed by company name and by
user group number. The company names are in alphabetical order I and
will give you the installation number. The full address of the installation
representative is given under the installation code listing. Later it will
also have a geographic list I listed according to region of the cbuntry.
I think that this describes what we have for you in the Reference Manual,
and rm sure you will be looking forward to receiving this within the
next two weeks.
I
I
:0
I
!
I
CHAIR The next division which we will consider is the Installation Management PAUL BICKFORD -
o
Thank you, Jim. I want to welcome all of you this morning to our meeting.
I would like to begin by outlining the objectives of our Division. We
primarily exist for helping to form guide lines to our management in the
areas of Personnel and Operation Management. We presently consist
of two projects. One in Personnel and Operations and, in a way ,one
1
Page 5
in Education. As a result of the Boston meeting, we outlined some
activities for this meeting, and are currently responsible for such
presentations as the Job Description and Personnel Selection Presentation. Also a presentation on Programmer Evaluation. We also
requested IBM to make a presentation on the Systems Reference Library
and one on 360 Operator Training. We would like to solicit at this
time any of you who are interested in participating in the activities
of our group. We will have a planning session this coming Friday.
We have about 30 people in our Projects. We would like to have
more people participate actively in our group. So, if you new members
or old members are interested in becoming involved in these activities,
please be at the Planning Meeting on Friday. Thank you"
C
CHAIR I would like to particularly suggest to all members I new and old , The Thursday afternoon session on the IBM Customer Engineer
relationship should cover a discussion of the APARS I and of something
called a Programming Systems Memorandum, which has just been
made available by subscription. It lists all the APARS for a given programming system, and comes out every two weeks. It1s the best
source I know of to find out what bugs are where and when. It I S extremely useful. Jim Taylor is not here; Dick, do you want to discuss
the Systems Division?
DICK PRATT Frank said he was unprepared, I'm extremely unprepared. About all I
can tell you about the Systems Division is that it consists of machineoriented sessions I and you will find these in your program. Some of
them have probably been changed. These will be announced. There
are sessions for 1620, 1130, 1800 & 360 and these will take most of
the day today, and part of the day tomorrow. I imagine you can find
the session you are interested in by just looking through your machine
type. There are some sessions scheduled at the same time you might
want to split yourself between - somebody has already complained to
me about that - unfortunately, there is only so much time. You just
sort of have to take a choice or go back and forthe. I don It really
have any information as to what is scheduled in any of these machineorient.ed sessions, except for the 360. The 360 will have some presentations by IBM on PL-I, which has just come out in DOS, and on
I
C
Page 6
o
DOS, Version 3, which has just been announced, and will be
available in April. And, while I am standing up, I would like to
say that lId like to meet very briefly after this session with the
Chairmen of the 360 committees, so we can sort of get things
straightened out. I assume the rest of the mcahine-oriented
ses sions are already straightened out. If they are not, I don't
know what to tell you to do about it.
CHAIR I should say that Dick is not the Systems Division Manager. He l s
the Chairman of the 360 Project, as well as a member of the Executive
Board.
There is one comment that I would like to make about that - particularly
in the 1620 area - there will be applications papers presented within
the Systems Division. Where we felt that they were sufficiently
machine-oriented that they would not be of benefit to anyone other
than a user of a particular machine type we put the applications papers
in the machine-oriented sessions.
I
Bill Lane has an announcement, and Jim Tunney has some modifications
on the agenda.
BILL LANE -
o
I notice that this is about the first and last General Session according
to the agenda. I notice also that, according to the agenda, Friday
afternoon until 3 P. M. there is a general time set out for planning
of the next meeting. The next meeting will be in San Francisco at
the Sheraton Palace and, getting word in for the people of the Golden
State, weld sure like to have you come out. Welre quite proud of
San Francisco and if you want something to see that is different go
out to Ashbury and see the Hippies. I went out to check about a week
ago I or two weeks ago I I gues it was, and they're fantastic. For
those of you who are worried, they are not around the Sheraton Palace.
But, anyway, I would ask that you harken to the call for papers I and
also the call for help, because the Executive Board can't put on the
meeting by themselves, and I sure as heck can It put it on by ftlyse1f.
Weld like you all to help. Weld like to have you consider today I tomorrow, and Friday morning I rather than just Friday afternoon I as to
what you would like in the next meeting I and what type of papers you
think would be appropriate I Please I when you do find these out I either
get information to the Division Chairman or I better yet I get information
both to the Division Chairman and to me.
Page 7
In CAST 7 there is a little questionnaire-type form that all you do
is check things off and send it back to me - it requires a five cent
stamp now I I guess. Weld sure appreciate it. 11m looking forward
to a good meeting today, tomorrow and Friday I as well as in San
Francisco. Come on out. res great in December - the weather's not
bad. In fact it might even be warmer than in the East.
lAURA AUSTIN In conjunction with Bill Lanels announcement, I might mention for
this Future Meetings Session that we have for Administration Division so that you can be considering this I maybe I should tell you where
the meetings are gOing to be so you would know whether you ld want
to volunteer to be Program Chairman or Local Arrangements Chairman.
December of '67 wi'll be San Francisco I as Bill mentioned April of '68 will be in Chicago - so anybody from the Chicago area we'd be interested in having volunteers to help on that program.
September of '68 will be Philadelphia.
December of [68 will be Houston, Texas.
April of J 69 will be Los Angeles.
That is as far ahead as I will go right now I but, if you are from any
of these local cities, we Id certainly like for you to consider volunteering
for help on those programs.
CHAIR Thank you, Laura.
Jim Tunney has s orne announcements and program changes -
JIM TUNNEY, PROGRAM CHAIRMAN Since there is going to be a series of program changes through out
this meeting, I am going to try each time to go down the lis t in the
same sequence these things appear in your program so that you can
make the changes as I go The fir~t change is on Page 14, I believe, the Session W2. 2, which
is 1130 - OK, page 12, in W2. 2, in which there is the 1130 project 1
there will be a presentation on 1130 Commercial Applications by Mr.
J. Elan.
c
!'~I
i'
I'
Page 8
0,·
Then on Page 16, W 4.4, the PL-I presentation has been moved
from there. That session will be chaired by Mr. Mc Ilvain, rather
than Richard Pratt, and will include comments on DOS. The PL~I
presentation has been moved to T2. l, which is on page 20 -- that's
the next item. T2.1, which is again Don Mc Ilvain's session, will
have the presentation by IBM on PL-I. T2.1 is PL-l under DOS.
The next change is on Page 22, which is T3. 1. T3-1, of course, has
the DOS Version 3, which is correct as stated. Down under T 3. 3, the
paper on Expanded AUTOS POT for 1130, which is B in T 3. 3, will be
given by Charles Newman, instead of D. Carlson. Page 23 - The
Education Project there - those presentations from IBM will be made
by Mr. H. Cod 0 w& G. Wolf. Their na me s somehow were omitted
here. That is in T3. 6.
Down at the bottom of that page, in T3. 8, R. Brennan is going to
give that paper.
lid like to talk to Jim Fisher after session here. His paper right now
is scheduled in F3.1, and there is some question as to whether thatls
when it will be presented.
Also, on page 27, in Fl. 7, the paper by Mr. Groft will be moved to
another session. I donlt know where yet, but the one that is scheduled
for 9: 15 on AC Circuit Analysis will really take place at 8: 30 inthat
session. I gues that s all I have.
I
CHAIR -
As usual there are always afterthoughts PAUL BICKFORD One little note - we are interested in forming a panel here at the
meeting of people, two or three people, interested in discussing 360
operator training. We would like for people interested to participate
in this panel. It will be Session F2. 2, on page 28. So if you are
interested, and would like to participate in a spontaneous discussion
here, please meet me up here after this session. Also, we are interested in getting together people who are interested in a CAl project Computer Assisted Instruction - Bill Lane is interested in meeting with
these people I so if you will see him here after the session I he will
o
Page 9
c
s peak with you then.
CHAIR We are trying to cut this session short today, so that there will be
time for registration for the people that didn It get there last night.
IBM has requested permission to make a presentation on some new
hardware Paul Manikowski of IBM will make the presentation.
CHAIR I heard the laugh when Paul read the statement of intent. How many
of you people have received CAST 7? Possibly half. There is a
letter in that, which is basically the same as the speech by Watson
to GUIDE, discussing IBM's policy regarding Program Announcements.
I don It know - we have not requested any specific coverage here, so
1111 give you the gist of it.
IBM has adopted a policy that program announcements will be deferred
until such time as they are reasonably certain that the program will
do what the intend for it to do. In some cases I it may even mean that
the program announcement will not be made until a program is actually
in Alpha or Beta Tests. Because I in many cases I this would be too
late to do the user any good they have indicated that they will make
a statement of intent I which indicates what they are trying to do I but
makes no committment on their part to do it. Their men are saying we are going to try to do this, it may not be exactly what we say
here.. We will do something in the area but we may even abandon
the proj ect. It I S not very good I but pos s ibly better than s lipping programs and having unsatisfactory programs issued. In addition to
that there was a statement made by Watts Humphrey, of IBM at
SHARE XXDC to the effect that there would be no extension to the
FORTRAN AND COBOL capabilities of OS. Any extension of capabilities would be done in PL-I, unless there was very serious
market pressure to implement an extension to FORTRAN OR COBOL.
I
I
I
I
I
With that - thatls about what we had for the General Meeting. I
know that there were a great many of you who didn't get registered
last night. This hopefully will give you time to do so. Also give
you time to catch up on the breakfast you missed.
Dave, or Jim Tunney - You know when the morning coffee break will
occur and where I Jim? There will be coffee at 10 olclock I outside
in the registration foyer area. I'll see you then.
Pages 13 and 14 were not made available when Proceedings were published.
I
I
I
~,
..
II,,,
~"
.
I
II',,~·'
"
c
o
/3
o
o
14
SESSION NUMBER
W.2.2
SPEAKERS
NO FORMAL PRESENTATION.
o
MEETING WAS CHAIRED BY LARRY ARMBRUSTER.
DISCUSSION
REQUESTS FOR INFORMATION AND ASSISTANCE COVERING THE FOLLOWING
TOPICS WERE MADE
MIXTURES OF FORTRAN & ASSEMBLER LANGUAGE
COMMERCIAL SUBROUTINES W/ OVERLAPPED I/O
ALLOWANCE FOR MECHANICAL FAILURE IN IDEAL FORTRAN
PLOTTER MALFUNCTIONS
BETTER ASSEMBLER LANGUAGE INSTRUCTION MATERIAL
SOURCE CODING OF THE OPERATING SYSTEM
DISK COPYING PROBLEMS ON THE 1800
EARLY MORNING START PROBLEM
USE OF OTHER PLOTTERS RATHER THAN 1627
GENE LESTER OF IBM WILL PRESENT A TALK ON PRIORTY, INTERRUPT
PHILOSOPHY AT SESSION W4.2.
A PROPOSAL TO SPLIT INTO SCIENTIFIC AND COMMERCIAL SUB-PROJECTS
WAS VOTED DOWN.
A SHOW OF HANDS INDICATES THAT ALMOST ALL PRESENT USE ASSEMBLER
LANGUAGE TO SOME EXTENT.
o
IS-
SESSION NUMBER
W.2.4. & W.3.4.
SPEAKERS
360 USERS THEMSELVES
MODERATED BY R.L. PRATT & D.R. MC ILVAIN
DISCUSSION
OPEN DISCUSSION ON PROBLEMS ARISING IN 360 INSTALLATIONS - MOST
ATTENDEES CONCERNED WITH DOS. LACK OF PROPER IBM ATTENDENCE
HINDERED RESPONSE TO MANY ITEMS. THE FOLLOWING ITEMS WERE
REVIEWED
DISCUSSION OF ERROR DIAGNOSTICS IN FORTRAN & LACK OF
INTERPRETATION.
BETTER REFERENCING AND INDEXING OF MANUALS IS NECESSARY
FOR THEIR EFFICIENT USE.
PSM NOW TO BE DISTRIBUTED BY IBM WILL GIVE USERS A BETTER
REFERENCE TO EXISTING APAR'S, ANTICIPATED CORRECTION TIME,
AND POSSIBLY IMMEDIATE TEMPORARY CORRECTION. THE RETAIN
SYSTEM FOR THE FE'S AND SECOM FOR SEwS IS IN USE BY IBM TO
AID THE DISTRICT OFFICES TO BETTER SUPPORT THE USERS.
FORTRAN DOES NOT AUTOMATICALLY OVERFLOW UPON SENSING
~
CHANNEL 12 PUNCH. A PATCH EXISTS FOR LEVEL 9, BUT WOULD NOT
FIX THE MOST RECENT ISSUE OF DOS.
029'S ARE SENSITIVE TO REPRODUCING HEX-PUNCHED CARDS AND
ARE LIKELY TO BREAK THE CODE PLATE, PRINTING OR NOT. IBM
MENTIONED THAT THE 029 HAS A NEW FEATURE AVAILABLE (CODE
INHIBIT FOR $3.00 A MONTH) TO LOCK OUT THE EXTENDED SET,
REDUCING THE KEYBOARD ENTRY ~OSSIBllITY FROM 64 TO 48 CHARACTERS. IT WAS REPORTED THAT AN 024 IS SATISFACTORY FOR
REPRODUCING THESE CARDS BUT SOME HAVE HAD POOR EXPERIENCE
HERE TOO.
REVIEW OF PL/I EXPERIENCE UNDER DOS. EXPERIENCE WAS
LIMITED BUT INDICATED GOOD ACCEPTANCE WITH THE COMMENT OF
POOR OBJECT TIME DIAGNOTICS.
IT WAS SUGGESTED THAT A FORM BE MADE AVAILABLE TO USERS
FOR USE IN SUBMITTING PROGRAMMING TIPS TO THE NEWSLETTER.
THIS WILL BE PURSUED ~Y THE DOS COMMITTEE.
JOB ACCOUNTING (AUTOMATICALLY) IS STILL A DESIRED FEATURE
IN THE SYSTEM SUPPORT. HOPEFULLY SHARE'S PRESSURE IN THIS
AREA WILL AID OUR REQUESTS.
THE oSR TYPE 3 PROGRAM IS AVAILABLE FROM THE LOCAL OFFICE
FOR INCLUSION AT SYSGEN TIME FOR LOGGING OF DIAGNOSTICS.
o
/h
I·
I
I
o
SESSION NUMBER
W.2.6
OS PROJECT
SPEAKERS
NO SCHEDULED SPEAKERS.
DISCUSSION
WE PLANNED AN AGENDA FOR THE FOLLOWING SESSIONS. THE THIRTEEN
ATTENDEES DISCUSSED PR08LEMS OF CONCERN TO THEM BRIEFLY. AFTER
ESTABLISHING AN AGENDA, IBM'S REPLY TO BOSTON RECOMMENDATIONS WAS
READ AND COMMENTED UPON. EVERYONE IN THE OS COMMITTEE AGREES THAT
COMMON SHOULD SUPPORT THE EFFORTS OF USASI X3.6 TO OBTAIN A
NATIONAL STANDARD FOR HAND CODED GRAPHICS.
c
i
I.
o
17
o
SESSION NUMBER
W.3.2
SPEAKER S
JIM ElAM OF IBM SPOKE ON 1130 COMMERCIAL APPLICATION PROGRAMMING.
DON GARDNER SPOKE ON SOCALS AND REMEDIES HE HAS FOUND.
DISCUSSION
DAVE DUNSMORE HAS BEEN ELECTED CO-CHAIRMAN OF THE 1130 PROJECT.
THE FOLLOWING SUBJECTS WERE OPENED TO THE FLOOR FOR DISCUSSIONIMPROPER DIMENSIONING
DISK READ ERRORS
o
I ..•
Ii
I'~
i"" "
c
SESSION NUMBER
W.3.6
SPEAKERS
MR. G.W. GOESCH, IBM CORP. ON THE
c
o
If
SYSTE~S
REFERENCE LIBRARY
o
COMMON
Cincinnati, Ohio
PROJECT:
Management Installation Division, Operation Project
SUBJECT:
The Systems Reference Library
SPEAKER:
Mr. G. W. Goesch, Manager, Product Publications
mM Corporation, San Jose, Calif.
Telephone (408) 227-7100
o
FOR
PRESENTATION: Wednesday, September 6, 3:30 PM, Session IV
8 Pages Text
0
1
..•
III
II
:i
0
The System Reference Library
1
I'm Gordon Goesch, Product Publications Manager, IBM, San Jose, California.
While I am not involved in the development of all types of IBM publications,
other areas of our organization have the same mission for similar publications
in various parts of the country.
When one looks at a manual, it is not impressive, there doesn't seem to be
much to it, but when you get into all tho ramifications in publications, it's a
little like an iceberg in the ocean - most of it is below the surface of the water.
We haven't solved all our problems and I'm not sure that we will completely but you may be assured that we (like yourselves) are certainly constantly trying
to improve our operation.
That is the reason I am always happy to talk to groups such as yours about our
publications. Because it gives us an opportunity to discuss with you our
Publication Library, its organization, its purpose, and revision service - as
well as to update you on the library's operation procedures - because even with
a good system - you must understand how to use it - if it is to be effective
for you.
I think that this type of a meeting can be a two-way street for information: We
develop the literature; you provide feedback. We do get feed back from you via
Reader's Comments forms - we want more of your comments and we certainly
appreciate them.
I don't know how knowledgeable you are about our publications; therefore, for
the benefit of the new members and also for the updating of the veteran members,
I'll run through a few slides which.!.. think will tell you the publication story.
Slide #1
BOL
To begin, we have what we call the mM Branch Office Library
(BOL). BOL contains much information. Not all, but the major
portion of BOL is made up of publications for users of our
equipment.
You have probably seen the publications display in the main
convention lobby. The display, I am sure, contains publications
of interest to you. Many of the publications on display have been
published since your last COMMON meeting. Feel free to
examine them in depth but please do not take them away, as
there is only one copy of each and we want many people to
benefit from the display.
o
Slide #2
FYI
In order to call to your attention what publications we have
For Your Information
1
~/
Slide #3
Swamped
and to help you avoid being swamped by ordering blindly as the
man shown in the slide
Slide #4
SRL
we have developed the Systems Reference Library (SRL). As
you probably know the SRL is a rather extensive library system.
Slide #5
Library
Shelves
However, before you start pulling arm fulls of manuals from
its library shelves
Slide #6
Mr. SRL
Helps Select
let Mr. SRL help you make the correct selection.
Slide #7
What,
Where,
How
He will
a.
b.
c.
Slide #8
System
Slide #9
System X
o
acquaint you with the SRL and tell you:
What is available
Where you can get the information
and, how you can go about getting it.
First, you must make some determinations:
a. What System or Systems interest you
b.
What size library do you want - everything for
that sy stem or just those parts that pertain to
your specialty?
(.~
. JiI
.
Considerable thought and effort was spent in the design of the
Systems Reference Library - for both the needs of the reader
and the type of publications necessary to support our products.
Slide #10
Each SRL is an encyclopedia for a particular system - with
separate publications for Ma.jor Subject areas. It consolidates
all the basic reference literature necessa.ry for you in:
Planning
Programming
Installing, and
Operating that system.
Slide #11
SRL Key
The key you need for opening any of the System Reference
Libraries
Slide #12
is the SRL Bibliography for the given system.
Currently there are 13 major System Reference Libraries,
ranging from the 1130 to 1he System 360 - and of course each
sy stem has its own separate Bibliography. You will find,
2
22
0
1
.,
o
however, that Bibliographies make cross references to
pertinent publications of other system s.
Each Bibliography is actually an Index of all the current publications about a specific system. In it, publications are listed
both by subject code and by machine number - and it contains
abstracts describing each available publication. (We will
discuss subject code a little later. )
You may have noted that now Programming LogiC Manuals
(PLMs) are listed in the Bibliographies. While they are subject to restricted distribution, they are available if a real need
is evidenced. The PLM details the internal logic of the program (like a large map of the listings).
Slide #13
Bibliography
& SRL
Newsletter
c
Now, how do we update the Bibliography?
Each Bibliography has its own Newsletter (its color is green)
and it is issued monthly (when there are changes). The
Newsletter updates the Bibliography, provides abstracts of new
publications, and lists Type I programs with their latest
modifications.
Actually, the Bibliography newsletter is a current "accumulative Index of Publications and Programs" available for a given
system.
From the publishing mechanic s point of view, Bibliographies
are periodically scheduled for revision - when that occurs, the
current information from the newsletter is merged into the
Bibliography.
Slide #14
SRL
Masthead
Each SRL publication, listed in a system bibliography, is
identified by a file number and a form number, located on the
upper right hand corner of the publication cover - as shown in
the slide.
The file number performs two functions: the first part, specifies the system (s) number; the last two digits the subject code
o
The Subject Code is made up of a group of two-digit numbers
(00-99) assigned to the various system components, e. g.
00 Includes, Bibliographies, System Summaries,
Configurator s
01
Machine System (CPU)
03
Input/Output Units
05
Magnetic Tape Units
20-50
Programming Systems
3
23
- ...;;.:..: .. -..... - .
A recent change now places Application Program manuals
under code 60.
o
The subject code 13 shown in this slide, indicates tbs.t this
publication is about Special and Custom features.
The form number is self explanatory; however, the form
number suffix indicates the editorial level of the publication.
Slide #15
SRL, TNL
Because of the dynamic nature of computer technical information, frequent changes occur.
When changes occur, technical newsletters (TNLs) are issued
to update the publications involved.
Consequently, not only is the Bibliography updated by its own
newsletter but each and every SRL manual can have its own
TNL.
In most cases TNL packages are made up of an identifying
cover page and replacement change pages for the parent
publication - .when you receive such a TNL you merge it into
its parent manual and throwaway the old pages.
inCidentally when a publication is ordered, you will automatically
receive the latest technical (suffix number) level copy as well as
all the outstanding TNLs available against that publication.
Slide #16
TNL
Mast Head
o
This slide shows the upper right hand corner of a TNL. It
indicates how the TNL identifies itself with its parent
publication.
It carries the file number and form number of the parent
publication it updates (and it is form number suffix sensitive)
Below that is the TNL's own number, publication date of the
TNL, and the form numbers of any previous TNLs outstanding
against the parent publication.
inCidentally, all page replacement TNL pages carry similar
identifying information.
Outstanding TNLs are incorporated into the parent publication
when it is being revised. TNLs may also be merged into the
parent publication when it is being reprinted.
I
0
I
I""
,1
4
24-
11
iG
Information regarding the technical level of a publication and
the TNLs that may have been merged into it may be found
inside the front cover of any SRL manual.
Slide #17
SRL, NL
Parent Pub.
Manual
This slide shows a manual, a corresponding TNL, and a green
SRL newsletter. With this combination on hand you have all the
publishing reference you need for the parent manual.
Keep in mind that the best SRL publishing information source is
the green SRL newsletter, because it not only updates its own
parent publication (the bibliography) but also lists all the
existing publications that are current for that system as well as
their outstanding TNLs
Slide #18
2 Biblios.
2 NLs
and of course that each major system has its own bibliography.
Slide #19
In-Out
Thus, the SRL system keeps you well posted on what is in
(current) and what is out (obsolete) for an effective library.
Slide #20
Wrapped Up
Basically, you have at your disposal a "living-doll" of a library
system.
o
Via the SRL, you can develop and maintain a library for one
system or more - and tailor your library to your own needs.
Slide #21
DPT
An additional source of information is the "Data Processing
Techniques" (DPT) Bibliography (Form F20-8172). It indexes
a series of publications of techniques for Study, Analysis,
Design, Implementation, Programming, Documentation,
Installation, Operation, Scientific, etc.
This Bibliography is also updated by its own Green Newsletter.
Slide #22
Series of
DPTs
This slide shows some of the DPT manuals.
Slide #23
KWIC
Another excellent Index for your use is the KWIC Index of
Marketing Publications (Form 320-1621).
Slide #24
KWIC &
TNL
It is published quarterly and updated by a monthly TNL.
The KWIC Index is based on an abbreviated 30-position publication title - listing and cross-referencing publications by the
important words in the title.
5
4l&QgMJM!la;;;;L",uxamui4J411iA¢ %J¢;¢,4 J ·k,,' ,;;;;,';;;
.,'t"
•
',.
S, .¥M??, _', .iII.""
The KWIC Index is listed in five ways:
1. Alphabetical
2. Machine or System Type
3. Form Number
4. Type I & II Programs in System Sequence
5. Type m & IV Programs in System Sequence
Slide #25
3 Way List
1-2-3
For example, this slide shows 3 separate word listings for a
single publication - "Programs for Petroleum Engineering"
Slide #26
2 More List
4-5
and here, the same publication listed by machine number and
by form number.
o
The latest KWIC Index was prepared from 11,546 publication
titles which generated 29, 122 listings,
Actually, the KWIC Index lists many publications other than
SRL publications, such as Executive Guides and Brochures,
tools and techniques manuals, applications manuals and briefs,
educational material,
Another way of putting it is that all SRL publications are
Marketing Publications but all marketing publications are not
SRL.
Hence the KWIC Index provides wide IBM marketing
publication coverage.
Therefore, to start a System Reference Library, it is necessary
that you contact your IBM repre sentative, and work through
him, using the three key publications we have talked about, the:
1. Bibliography
2. its SRL newsletter
3. KWIC Index of marketing publications·
You can select and build your own reference library to support
your system.
However, please do not order your publications by writing to
Product Publications (the address shown on the manual) or our
Distribution Center in Mechanicsburg, as it will only delay the
order. You must order through your local IBM representative.
Now that you have established the base for your library, let's
discuss its maintenance.
6
I.
o
a.
b.
Slide #27
SRL, TNL
Revision
Service
If you desire, you may continue ordering specific
publications through the mM Branch Office.
However, you may prefer to subscribe to the
"Publication Revision Service" that the System
Reference Library offers.
The Revision Service provides automatic shipping of revised
manuals and newsletters directly to you without having to go
through the IBM Branch office each time.
However, Subject codes 00-60 only are supplied by the
Revision Service. This excludes installation supplies,
education literature, and other supplementary information.
Here's how it works:
c
Slide #28
Subscrip.
Card
The IBM representative fills out a subscription card with you,
using the green SRL Newsletter to indicate those publications on
which you want the updating afforded by the subscription
service. After the card is approved, mM Distribution Center
takes over and a single copy of each TNL or revision involving
the indicated items will be mailed to you.
Slide #29
Mailman
w/Pkg.
Remember that only one copy per subscription can be mailed.
Additional copies must still be ordered through your IBM
representative, just as were your initial manuals. Any changes
in your Revision Service are made via a new subscription card.
Again, if multiple copies are ordered through the Revision
Service, it only complicates and delays matters.
It is important to set up a library and assign responsibility for
updating and maintenance to make the service effective.
Slide #30
SRL & Rev
Pkg
Now, you are all set, you know what is available, you know
how to select the publications, and you know where to get them.
Slide #31
The Key
Is Yours
The key is yours.
Slide #32
Key to
Knowledge
Use this key to open up a tremendous amount of timely
knowledge about your system through the IBM Systems
Reference Library.
o
7
27
Slide #33
SRL
I hope that the result of all this, will make you as happy as the
man in the next slide - careful planning of your library
may help!
o
Most of you have probably noticed the Reader's Comment Form
that is appearing on the back of many manuals these days.
Some of you may have filled one or more out. The response
to these has often been very gratifying and we appreciate it.
I want to encourage you to use them as it is one of the means,
along with meetings like this, by which we get the feedback from
our readers that is essential if we are to improve our publications and make them more useful to you.
This form is self-addressed and post-paid and will be directed
to the correct publications group. As a reminder, please
don't use the form to order manuals as we have to redirect such
orders back to the Branch Office with a consequent delay.
On behalf of the publications groups, I want to thank you for
your attention and for your comments and pledge to you our
whole-hearted effort in providing first rate publications support
for your sy stem s.
Thank you again for giving me this opportunity to talk to you.
8
I
I'
I:
c
SESSION NUMBER
W.3.7.
SPEAKERS
1. WADE NORTON
2. NORMAN GOLDMAN
3. NORMAN GOLDMAN
DISCUSSION
1. REREAD
2. GENERATING WITH A 2 DRIVE CUSTOMIZED
3. ACCOUNTING ROUTINES.
4. VARIETY OF TOPICS.
I'
I
'0
SYSTEM~
"1,1
I~
1
I
o
SESSION NUMBER
W.4.1
SPEAKERS
MR. HERBERT RUDERFER
DISCUSSION
MR. HERBERT RUDERFER PRESENTED BOTH THE FACILE & FACET
MR. WILLIAM SILER PRESENTED HIS PAPER IN ANOTHER SESSION.
30
PAPERS.
SESSION NUMBER
W.4.2
SPEAKERS
GENE LESTER OF IBM SPOKE ON PHILOSOPHY.
PAUL MANIKOWSKI OF IBM SPOKE ON EDUCATION INCLUDING COURSES,
PROGRAMMED INSTRUCTION, AND LITERATURE AVAILABLE.
DISCUSSION
TWO NEW PIECES OF LITERATURE TO AID IN ASSEMBLER LANGUAGE ARE
AVAILABlE~
1.
2.
PROGRAMMING THE 1130 AND 1800 BY R.K. LOUDEN - PRENTICE
HAll.
P.I. COURSE - CONTACT DON JOHNSON
HINSDALE CENTRAL HIGH SCHOOL
HINSDALE, ILLINOIS
c·
Ie
dIiAmlIlM\lA4l\TiiWllMWM;=4iJMCM. 00.4.\;; (·r1M.
o
SESSION NUMBER
W.4.4
SPEAKERS
MEETING MODERATED BY D.R. MCILVAIN.
DISCUSSION
THE PRESENT STATUS OF THE DOS COMMITTEES IN GUIDE & SHARE WAS
GIVEN. ESSENTIALLY NO COMMENTS HAD BEEN RECEIVED FROM THE GROUP
AT LARGE PRIOR TO THIS MEETING - PRE-SUBMITTAL IS NECESSARY FOR
EFFICIENT OPERATION OF THIS COMMITTEE. COMMENTS ON FORTRAN &
COBOL WERE PRESENTED BY MESSRS. GWILLIAM & CUNNINGHAM. THE
RESPONSE FROM IBM TO OUR QUESTIONS ON FORTRAN WAS RECEIVED BUT WAS
VERY POOR AND THE OUEST IONS WILL HAVE TO BE RESUBMITTED. GUIDE IS
MEETING IN ENDICOTT WITH THE IBM DOS IMPLEMENTATION GROUP LATE IN
SEPTEMBER, 1961 AND COMMONS DOS COMMITTEE WILL HAVE REPRESENTION
AT THIS MEETING AND WILL EXPLORE SOME OF THE ITEMS OF INTEREST
TO COMMON.
I
i"
~~~---.-.--~-
... --.------.--.--.-.---------~-.- •.--...-- ..----------
o
SESSION NUMBER
W.4.6
SPEAKERS
PANEL DISCUSSION I
DR. R. GA BR I EL
MR. M. GOLDBERG
MR. N. GOLDMAN
MR. P. KOEPSELL
PANEL DISCUSSION II
DR. R. GABRIEL
MR. H.B. KERR
MR. P. KOEPSELL
MR. D. LA PORTE
DISCUSSION
1.
ECONOMIC JUSTIFICATION OF THE UNIVERSITY COMPUTING INSTALLATION.
2. EQUIPMENT SELECTION FOR COLLEGES AND UNIVERSITIES.
ATTENDANCE
36
o
53
...
=. . =
= =...=..... =_
.....=
......"""'..
-~-----------,---~--"""""""-""'""""",.,-,~,,,,..=
:=="""'-
o
SESSInf\.1 "UJMRFR
H.4.7
SPEAKFRS
FREn
1/11.
MATFJCFK, cnMPtJTER CENTER ON NIU SYSTEM/240 A nATA
RFTRIFVAL SYSTEM AS APPLIED TO A LIBRARY INVENTORY
(""...-."'-.
~I..J' '
.
-I
j
j.i
1,1,1
1j
o
NIU System/240
A Data Retrieval System
As Applied To a Library
Inventory
Fred W. Matejcek
Computer Center
Northern Illinois University
o
... _ _ _
._._.~
.....,_ _
~~~~_~
_ _ _ _ _ _ _ _ _ _ _ _ _• _ _ _......._ _ _ _ _ _ _
•
~1.Jt:'1r'ri"·-~'··'··
···:~-"·'!"·~·"jj'I-t--··"~·-
....
-.~-"
..
-.~
............"-..... -.-.:w..:i,..l:o;.;.;....;,;,;,;.r...;.l.;W;;;,,;...;,,;;=
..
~~~~
_ _ _,
o
TABLE OF CONTENTS
Page
Acknowledgements • • • • • • • • • • ••
i
List of Illustrations. • • • • • • • • •
ii
Introduction • • • • • • • • • • • • • • iii-v
The Problem Defined
• • • • • •
1
Special Conditions • • • • • • • • • • •
2
The NIU Solution • • • • • • • • • • • •
4
• • • •
File Creation and Organization
Servicing Requests. • • • • • •
Maintaining the Inventory • • •
Maintaining Records • • • • • •
•• •
•••
• • •
• • •
·....·..
Appendix A . . . . . • • • • • • • • . .
Appendix B • • • • • • • • • • • · . . .
Appendix C • • • • • • • • • . . . • •
Appendix D • • • • • • • •
• • • • ..
Future Prospects • • • •
•
•
4
19
26
29
33
35
36
37
38
o
ACKNOWLEDGEMENTS
Much of the success of this system, for it has been
operating successfully more than a year now, must be attributed
to:
Dr. Robert Hunyard, Head of Northern Illinois University's
Audio-Visual Department for recognizing the problem and the
possibility of a computerized solution; Mr. Clyde Givens,
Director of the Computer Center, on whose analysis the system
was initiated; and Michael Roldan, Director of the Film Library
and his staff, who learned to use the various facets of the system
as they were designed to be used, without which the best design
never becomes operational.
Finally, the existence of the system
is in no small way due to the patience of my spouse, Elizabeth.
o
i
37
_~_~~""""",,",,",",,-
_________
""'lItu_'hIiWfW'~"
.. " .. .,. -.. . . ..'1lt'g.•~.--
.m.
o
LIST OF ILLUSTHATIONS
Page
Figure
1
Disk Layout • • • • • • • •
2
Disk TAble of Contents •
...• 6
....• • • 8
...
• .11
Film Number • • • . . . . . . . . • .13
Film Hecord Formats. . . .
• .15
3 Customer Files •
4
5
6 Invoking the System (Activity Chart) .18
7 IDNIC Index. • • • •
8
.....•
• • • 20
Processing Requests (Activity Chart) .22
9 Processing Billing (Activii.y Chart) •• 25
10
Request Growth Curve • • • • • .'. • .27
11 Updating The System (Activity Chart) .30
ii
()
3R
I
i
I\<
o
INTRODUCTION
NIU System/240; Northern Illinois University and the
system we developed.
It took on the project number "240"
identifying the application area with which we were dealing
within the University, Audio-Visual Aids.
This is a data
retrieval system developed on the 1620 to solve the inventory
problem faced by the film library at NIU in the AVA Department.
The manual solution to the problem was unable to handle the
growth both in terms of the student body (see next page) and
c'
something refered to as the "information explosionll •
This
is the first point we see under"justification" for the system,
the economic aspect.
Extra clerical staff to handle the
volume could not be hired for the same expenditure involved
in the operation of System/240.
The second point under justification is better and faster
service using System/240.
User requests that are received in
the morning are confirmed by the afternoon mail.
This is in
comparison to the normal two week response time by other
libraries in the area.
The third and last point is the additional predictive
information to be gotten from the system based on the records
kept by the system.
!o
So much for justification.
unique is the system?
iii
How new or
0
DU([J)/l?'l7'17.17ER/l¥l~
l'O/s-o
GrowTh
3,o/'10
I
NIII SYST"M/2'~-
IS-I<
I
IJAI J.IIIIE
I
NIII S-YST6M/~ ~D.
I
SrJII Ii T E.1J
f
I
1'2.0
IDk
t
"'~
~J
I
I
,
I
I
S-It
I
NIU
I
I
TI/lCNI~'$
COlI-lifE
STAT.
I
I
t9
~
1>
~
~
~
c
1>
U\~
~
r~
-
""jl
0'"
40
.---~~--.------
--:0
--~-.-.~-.-'~-~~--~-
o
A review of the literature shows that NIU System/240 was,
at the time it went on-line in April of 1966, the only
application of its kind in the country, and is to date the most
comprehensive of exisiting systems, irrespective of hardware.
How did we arrive at this solution?
To answer this question
we will look at the problems facing analysts when embarking
upon the creation of any data retrieval system to solve an
inventory problem and then look at how we faced these problems.
o
v
41
o
THE PROBLEM
To give structure to the following discussion we will
separate the problems and their respective solutions into
four classes.
This taxonomy covers the basic problems
faced in the design and implementation of a data retrieval
system.
Our first classification covers the creation of files;
something common to the early phases of most systems, because
in creating a new system or converting an old system from
manual or other means, the files must be put into ma.chine
usable form.
The second class of problems concerns the actual
servicing of requests by users for the item contained in our
inventory.
The third is the maintaining of-the inventory.
We must have a means of deciding how many of a given item to
have on hand.
The last basic class of problems is that of
maintaining records.
Our problems demand that we maintain
records on both our users or customers and our inventory.
Before elaborating on these four problem areas and providing
the corresponding solutions to the problems, let's look at
the special conditions which are placed on our example
application, the NIU film library.
o
o
A RECIRCULATING INVENTORY
A film library is a special sort of inventory.
It embodies
a recirculating scheme by which films are supplied to the user
and then returned to the shelves of our inventory.
Therefore
the normal input which maintains the "on hand" number of items
that we have decided upon, is from the user, not a supplier.
In other, or what we may call "one-wai' types of inventories,
the input which maintains the non hand" number in the inventory
is from an outside supplier.
The only time which the recirculating
inventory gets its input from an outside supplier is when we
c
wish to change the !ton-handlt level because of increased demands.
In designing the system, this recirculating scheme forces
special considerations in three areas.
All three fall within
the scope of the inventory file maintenance.
First, because we "reuse" items, we must keep track of
each and every item within a given type, rather than simply
being concerned with the total number of that item available.
We must know whether or not print three of a given film will
be back in the library in time to fill another request.
Seconq we need a method to tell us when to go to the
outside supplier and beef up our inventory.
This is more
complicated in the recirculating inventory than in the oneway inventory.
1
0
- 3 -
The third manifestation of our special conditions is
possibly a blessing in disguise.
c·
Because quite often orders
are placed more than a year in advance, our files must be large
and complex, but the potential blessing comes in the preview
of demands to come that we can glean from these advance orders.
But this is only potential.
We must take advantage of it.
Having now looked at the justification for the system,
the structure of the discussion to follow and "tne special
conditians a film library places on an inventory system, (and
hopefully having put you in mind of the special conditions of
your particular inventory problem) let's see what we did to
cOJae with the four general classes of problems embodied in
the design and implementation of a successful inventory system.
I·,
tl
Ci
THF~
NIU SOLUTION
First is the area of creating new files.
We Hill consider
our basic tools in terms of hardware and then software.
The
hardwarA c0r:c.::ists of an 113M 1620 Model II with 60K, which has
been undated to include a 1311 disk drive and 1443 printer.
The 1311 disk drivp is the medium that we use for our mass
stcra~e
of recQrds.
two parts; that
The software will also be considered in
whicr~
is sUPGJ.ied by IBM and that v-rhich is
sllpnlied by the Northern Illinois University Computer Center.
The t1oni.t0r I
CI:
i~
thA norrral suoport system supplied wi.th
the 1311 disk drive but as used, has been somewhat modified
frcm the original IbM version.
we do this
utili7.in~
Tho1Jgh we do compile FOFTRAN
another disk thus er.abling us to delete
FOPTRAN TT-D, its subroutines, and some unused utilities from
our nroduction disk and in turn make this room available to
System/240.
Sinee three of the four basic programs are in SPS
and 90 oercent of our running is execution rather than comniling
we do not really lose any flexibility.
The four programs developed as the full compliment in
NIU System/240 are named AV-LOD, AV-DLY, AV-RPT and QIKLOD.
AV-RPT is the only one written in Fortran and is used for
our quarterly reports.
AV-LOD and AV-DLY are the only ones
that are disk resident and used on a daily basis.
o
a key part of our back-up.
QIKLOD is
- 5Since we have only a one drive system, a strict disk to disk
o
back-up can not be provided, but with QIKLOD and the card output
that is produeed during our daily procedures we are able to
provide card back-up.
Figure I sho'W"s the status of our disk and
if you are familiar with the normal organization you can see
where things have been rearranged or deleted.
This includes
limiting the Monitor work area to the minimum 11 cylinders.
There are basically two files in the NIU/CC software.
The one is a file that contains all the information pertaining
to our users.
This file is broken again into two parts that may
be termed our "billing and shipping files".
We have run into
the necessity of creating these two files by the nature of our
users.
In dealing with schools or other large institutions we are
often asked to bill a central office or school district, but in
turn they want the items, in our case films, sent to individual
users within their system.
So we are faced with the problem of
billing to one address and shipping to several addresses within
that particular users juristiction.
We will look closer at
these files as we get into the structure of the actual records.
The other file that we keep on the disk is a complete file of
all the items, again in our case these items are films.
Our film library is divided up into 21 logical categories
by
subject matter.
We use these logical categories as actual
physical categories in our disk organi7.iation.
For each category
o
o
SEGTDQ
AD Dr-
~~
I
Ie'
i
I
01
.s
""
~
~
~8
A~
I
I
~
I
I
."
~
-If
V')
0
.,u
l
~
...c
>'8,
.s
d
~
~
"/9
- 9 -
would be exactly equal to the field of "B's" or the beginning
o
address of a given category, for the active level would be the
beginning address of the category.
The second field, or the
field of uN's" indicates the last film number within that
given category.
Each category starts with a film number one
and goes on to the la.st number within that categorYe
The last
number is the number that appears in this field so that in our
search we are able to check immediately the last number and see
whether or not the film specified is a legal film for that
category, i.e. if it falls within this range.
The last field
or the field of "L's" as it appears, is the limit to which we
may fill this category.
When a category is completely filled,
the converse of the situation we mentioned earlier occurs.
The
field of "Ats" and the field of "Lts" are equal because the
category has been filled to its limit.
Within that static table
there are two additional entries, one is the beginning address of
the customer, or the billing file; the other is the beginning
address of the shipping file.
Now going to the organization of our two customer files,
if we have an individual requesting a film and the request
indicates that he is to have the film sent to the address he is
to be billed at, this customer is entered in only one of our two
files - he is entered in our billing file.
If though, we have a
customer that wishes to be billed at an address separate from the
address to which he is having the film shipped, he is entered in
s-o
rtr·~.
~c';f
- 10 -
o
both files.
We gi.ve each customer a four digit number.
If he
is a customer with the same billing and shipping address this
is all the number that he gets.
If the customer does desire a
separate shipping address, one separate shipping address or
several, then on the initial cards that we use to create his
records on our disk file the two cards that contain his billing
address also contain a 1 in card column 7. 1
This indicates that
there will be shipping addresses to follow these billing
addresses and these will have the same four digit customer number
but will also have a two digit shipping number tacked onto the
end of this.
file.
c·
These records are then filed on our shipping address
Then we have a billing address number which is tied to
one or several shipping addresses and a sequence number which
makes each shipping address unique.
If we look at Figure 3, we
see the example of customer 101 in our billing file with a 1
coded following the 101 which indicates that in the shipping file
we will find at least one shipping address if not more for him.
When we make a request for the customer number 10102 this request
is serviced by going to the billing file, finding customer 101,
finding out whether or not the 02 is valid for this customer in other words whether or not the customer does have separate
shipping addresses by the fact that there is a one following his
number on the file and then going to the shipping file and finding
customer two or shipping address two of customer 101.
0·"·"
"
ICard Formats appear in Appendix A.
~/
Figure 3
0,
•
o
o
•
N
o
-o
o
I
I·
lI
•
()
I
--0
-0
2
~
- 12 -
Moving on to the organization of our film file, we'll look
at the key to this file or our film number.
find an example of our film number.
In Figure
4,
we
The first digit of this
film number is a length and color code.
The two digit category
which is the logical category mentioned above and the physical
category that we find on the disk.
We have then within the
category a four digit number which uniquely identifies a
particular title of a film.
Tacked on to the back of that we
have a two digit number which uniquely identifies a given
print for that given film.
conditions showing up.
Here again we see our special
In this recirculating inventory we have
to keep track of not only the total number of a given item but
c
of each and every item as a unique entity.
number
Cl
This is our print
We have multiples of a given item, i.e. a given title
of a. film.
Very popular films have several prints.
A footnote to this; the DAVY group which is to the educational
film industry as COMMON is to IBM, is trying to come up with a
coding system to be presented to the Federal Government as a
suggested national standard for the entire educational film
producing industry.
Though we did develope Northern's numbering
scheme independently, we find a great resemblence to the
DAVY number in the one created for Syatem/240 at Northern.
We ha.ve the two digit medium which corresponds to our category,
one digit storage area which corresponds directly to our length
code, a sequence number of four digits (this is our "title number")
0,
and the two digit print number which we just refered to.
In
'(",.,I!
addition they carry a two digit year of acquisition number.
S3
This
Figure
4
0
j
t
1
c:J.-~
~~
~~~
I
~~~
.~
1
I
;'d!
~
f
1
i
I
I
Q::
l4J
Il
~~
--J~
G:
.
y •
~~
~,,"~
~~
~~
.
.~&.
~
~/
~
<
C
I
1
H
>
C
-1~
o~
.-<6:~
•
.~~.)
J).
~~.
A
1
j
~6
4rro
.~,
~
. .,.~~
6 ..
~
·cC~
'-"
t'
~
?~
..,,-
C:
I"
- 14 number we carry internal to our record but do not put on our
key.
If we look then at Figure
of our film records.
5,
we have the two formats
The first is the general information on a
given film, information similar to that which we would expect
to find on any inventory system.
We have the film key which
we just mentioned above, the film title which corresponds to
the item description, a rental amount which would correspond
to a selling price, the cost of the film to us which would
correspond to a users purchase price and the number of times
this particular item was requested by a user plus the number
of times we were unable to fill a customer's request because
the film was not available at that time.
c
In addition we have
the alternate film number.
The alternate film number is a number that is used if the
customer indicates that the time period that he has requested is
very important to him but that the particular film is not
important and that a film covering comparable material would
be satisfactory to him if we can make a suggestion.
This
alternate film is automatically booked for him on the given
dates if he indicates that we can make this SUbstitution for him.
The second record is also an outgrowth of our special
conditions - added problems to the normal inventory system that
we had to cope with.
This is the record that we keep on each
individual print, each individual item within an item type.
gives us summary information, frequency of use, the contract
o
This
..-.-..
'-.""".""''''''''''
....
'-.-.-~'~~~"'--'-------------
Figure
5
o
~
~
(/)
~
~
~
t.a
•
~
~
0
0
II
0
U
til
~ ~3
~
()
~
~
/
:e
L»
~
1
1
I
~
~~
-0
2a.
::>
~j
Z
~b
I
05
~o.
5~
C,'
o
MAINTAINING THE INVENTORY
If we look at Figure 10, which is labeled REQUEST GROWTH
CURVE, we are trying to develop here a buyer's indicator by
keeping records on particular films.
This curve describes the
demand on a new item from the time of the user's awareness of
its availability to some time out beyond that.
for this film grows;
80S
As the demand
we go up on the axis labeled'trlumber
of requests", we can also mark it off in intervals indicating
number of prints of that film, or number of that item which
we have to keep in our inventory in order to satisfy these
C
1\
requests.
When we reach the doted line, we have a choice to
"
make, whether to purchase, lease, or by
so~e
other means obtain
the items in order to catch the peak of this growth and get
these requests, or at this time to stop our purchasing and to
realize that we are going to have to take some "not availables"
or find that we are unable to satisfy the customer's request,
but in so doing not have obtained films that will not be in
demand beyond time "y".
equation for this curve.
It is our job to come up with an
At the current time based on our
recrods, we can find a breaking point - point B - where the
curve starts to come back down or it starts to change direction
somewhat.
o
When this happens we can cut off or we can lease items
Figure 10
0
J.IJ
:>
I
~
I
CJ
I
I
~
I
I
S
<-!)
I
I
I
tl
JJJ
I
J
I
=»
~
Ld
--I~
I
C¥
~
I
\.
-,'
';../
I
J
~~
H
§
I
I
I
I
I
I
I.¥
0;"
'5'
- 28 -
for a short amount of time in order to fill in this area and
pick up these extra requests.
But our job is simply to supply
this buying information to management so that intelligent
decisions can be made on growth and trends within the inventoryo
This is the heart of the successful inventory and the conclusion
of the third portion of our solution.
o
o
MAINTAIl\fING RECOEDS
Having covered the creation of new files, the servicinE
of requests and the maintaining of our inventory, let's go now'
to the maj ntaining of O1Jr records.
We find that in the
maintaining of our records we are going back to the Same
system that provided us with the ability to create these
records.
We have if you refer back to
Figurf~
I, the disk
layout and Figure 2 the Disk Table of Contents provided for
our user department a system that very closely parallels IBM's
index sequential on the 360.
We have come up with an index
sequential type sytem for a 1620 disk system.
We have quite
the same capabilities here - the system can create files, add
to the files, delete from the files, process the
randomly or sequentially.
fi~s
All of these abilities are itemized
in Appendix D"
The updating is done on a daily basis, in other words, the
AV-LOD program which provides both the creation and updating
abl1ity is run daily just before the AV-DLY program which
1
services the requests.
Figure 11 shows the operation of AV-LOD
1
in the update mode, but it should be noted that in our system,
creation is a special case of updating.
1
I"
At this point we must
look closer at the back-up system that we mentioned earlier.
()
70
I,,·
1:.1
Figure 11
c
ta..
:::>
-z
~
~_ _..... c:J
o
7/
-~-----
--...-............ --....
.......
.,...
,~.~~~~-------------
- .31 -
If we h.ave to riu pI i . e!3. te our files we have to foo
steps.
Numt)er one
throuf'~h
three
c'
is to load tt, moni tor s.ystem on the disk,
numt'er two we hcl've to load our ba.sic files which amount to our
customers two part file, bi.lling a.nd shipping, ano our film
records for each category.
The
th~rd
step is then to reload
all the da.tes booked on thr3 fi lms we have ju st loaded.
The
loading of the customers and the films is done ae;ain by AV-LOD,
but the loading of the dates booked for a given film is done by
a unique
~hich
any
pror~ram
named QIKLOD
G
This uti 1 izes
01)r
a.cti vi ty cards
are kent up to da.te on the basis of our bilJing, i.e. for
filrnr~
that h2,ve t)een returned. and are no longer ovt of
li.LJrary the
t!'an9~wtion
th€~
or activity card is pulled i.'rom our ca.rd
fi le and storpd elsewhere.
So 1-Then loading up the peri.ods
booked for a given film we have
current bookings.
on~y
This
particular prop:ra.rn allo'Vls us to Toad someHhere in the nei.ghbor-
hood of :> ,000
bookin~8 .i.
n an hour.
The dai1y booki.nf, program
t.hat provi.des us with the conftrmations a.lloVJs us t.o book
something like a thousand in an hour.
from
We usually spend. anywhere
3 or L minutes to 15 minutes on the 1620.
time peri.ods thi.s smaIl the Ploat
compu ter time, but
set-u~'
8ivn~~ficant
Obviously in
time factor :i s not
tirr,e and genera.l operator intervention.
Having looked at the fOllr Darts of the problem and in turn
tte four parts of our solution, whi.ch involves creation of new
fi.les with program AV-LOD, servicing of requests using AV-DLY,
maintaining of an
:Lnv~:~y:t.cry
by Droducir:g
A.
bU'yer's indictor for
c
- 32 -
management and mainta.ining our records through the use of
QIKLOD and AV-LOD and having talked a little about the pleasantly
surprising performance we have gott.en from the system, lets
look at fut'Jre projections for the system and areas of growth.
o
o
FUTURE PROJECTIONS
Referring to Northern's enrollment growth curve in the
Introduction to our dlscussion, we can see the installation
of our Model 40 360 in August of 1967.
With the installation
of the machine under our belt, we are well on our way to
converting System/240 to 360 COBOL.
As we mentioned earlier,
a large part of the development of System/240 was devoted
to producing our own Index Sequential system for the 1620.
This having been done for us on the 360, the strict conversion
is going rather quickly.
Of course in a 360 we have additional
hard-v-Tare capabilities not found in the 1620.
Two of these
are the teleprocessing and multiprocessing capacities.
Using teleprocessing we can do several things, two of
which are; (1) give "irrunediate" response on film request and
(2) develop a cirriculum-building-tool in the University.
In
the first case, not only can an immediate "yes lt or "nott response
be given on a request, but if the response is "no" a calander
showing remaining available dates can be displayed on the CRT.
In the second case we can expand upon our existing quick
inde~
and on request, display available material pertinent to the
subject matter to be covered plus indicating proper sequences.
- 34 -
o
The multiprocessing capability gives us the technology
to be able to tie all the state libraries together so that a
request given to any of the libraries would exhaust the states
potential, i.e. over the phone, in a matter of seconds any
indi.vidual could have the entire states film resources at his
cornmand.
We have now a data retrieval system that is updated daily.
With the new hardware to fill the gap we will have a realtime system.
We feel at this point that the ability of the system to
grOl-T
rN'i th
increasing demands and greater the capabilities of
the hardware is li.mited only by the extent to which we wish
to Dllrsue it.
c
75
APPENDIX A
if.r007
P04884
1(eqV#.5i
IJrJlYIBItR
67
IOC: 00 b 7 00011° 1
R£TU~tJ
°
a 9 10 11 12 \3 14 15 16 11 11 I 20 21 22 23 24 15 " 21 1! 23 30 11 1213 34
Jj
l_
C
~AT~
IJ'STO/'lli=R
s""PPtlJfj
~IIM3ER.
°
l 0 10 ~ : 10 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 .: il R -; 0 0 0 0 0 ~ -:
0 0 0 0 0 0 0 0 0 0 0 0 0 0 n 0 U0 0 0 00
000000000
1 2 1 4 5 , 1
~o
SH IPf>;\J~
l>ATE
FILm
NilirT/Ji!fl
o
Cit
01 19 67
I
)075
I( ~nli66~Te31fl
SHIPP,,,,c:.
10 1"1 rlt
1
36 31 38 39 40.41 42 43 4~ 4546 41 4! 49 50 II 52 S3 5.4 55 5~.~1 5~_59 60,61. 2 63 S4 &'_ 66 &l 68 69/0 11 11 13 14 15 16 JJ 18 19 80
I
I
FIIoWI
~4I""B;:~
.,OOOOOO,O,OOO,OOOOOOOOOOOOOOOOOUOOOOOOOOOOOOOOOOOOOOOilnooooooOOOOOO~OOOAAOOOOOO
j
1 2 3 4 5 & 1 8 9 10 11 12 13 14 15 \6 11 18 19 20 /1 11 13 14 15 26 /1 18 29 30 31 32 33 34 .35 36 31 38 39 40 41 41 43 44 45 45 41 48 49 50 SI 52 53 54 55 56 51 58 59 60 61 62 63 S4
:
SlilElib071Cli6€lilj66
CVj~~J[alr
F'/L./tf
l)Are
P;;t't18{5R
])tJI£,,]
Sf::
~~o
(N
66 &1 6& 69 10 11 11 13 14 IS 75 17 18 19 80
4885 ,
~~tiI",I~K:
R.Ffiltl,Uu
(S-008)
·1
'
FRo.tII
6~
,'2
o • O. 0 0 0 0 0 0 • ... 0 0 • 0 • 0 0 0 • 0 0 0 0 0 0 0 0 fi
UU 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ,
I
2;
T;/~7RtXToR.
"be-p·r.
1"0
0 an,' 0 0 0 0 0 • 0 0 0 0 0 ~ 0 0 0 aDo 0 0 0 0 0
j
4884 .:
I
1 2 3 4 5 6 I 8 9 10 II 12 13 14 15 16 11 18 19 10 11 21 23 24 25 26 2J 28 29 30 31 32 33 31 35 36 31 38 39 ~o 41 4243 44 45 46 41 48 49 5051 52 53 54 55 56 51 58 59 60 61 62 63 64 6' 66 61 68 69 10 11 11 13 14 1J r JJ 13 J9 80
V 6PjlrE7r4fOO((I~2'670~3(rt:,~~1'Jl~~~.:5
eVtl!Ii!~...
,:',~,yr
.0
bAril'S
'R.;Q,Mt;O"
NI.INl8eR.
lArs
tRaM
o, 00 000000J
1 2 3 4 5 6 J
V-~
a
'I 'I
"'I.
1
'I
"
"
'1 '2
Ti
•,
Ii.. I!
'l 'l 'l 'l
OJ.
III
•
~
i-;i3I(,M.'f
_,
~
~ ;Jc;()~5j £0
Ai
'-1
()
-:
IJI.T2R{VtJTe"~ q
I?=~
J
r
4
- t-r
_ I 2 3 4 5 6 .1. t .9.10 Il.JZ 13. lUa .lULI8,\,9 20 2Ul 23 24. 15)6
to' ~TH
fL2L2~..3'L3.!.~ql
34. 35 36)1 38 39 4Q
~1
4143414546 H 48 49 50 51
~UHORA
SlREET
5~
.Joy
ell
I..
r
~
~ cl,
4
II
.
0 ADO 0 0 0 0 • 0 0 0 0 0 0 0 , 0 0 0 0 0 0 0 , U0 0 0 0 0 0 0 .oj 0 0 0 0 0 0 0 0 A 0 0 A 0 AnA.
/- 4p001 0 11 (79
T..
A A
0 0 0 A'
535455 56 51 56 59 60 61 61 63
A
7~ ~
/J1I"'AEjl.
1
.I
\1
~4
I
~.
1:
5' 66 6J 68 S9 ?O 11 12 13 H
1~
1.
I
.'
l
1118~1"{
--
(500~)
Il..
~H/~PINl.i
E
1
1LL 1NO'S 60504
ft.
a ~E
~
~
(5005')
0 0 0 0 ~ 0 0 0 0 0 0 0 0 (\,l1li
l'
,,·....It
~
~5~~lc060C')95~'(O::1~i~·
w··
'I. 't 'J. 't
F, '- h1
N.;meER
e.r.,.
(SDO~\
R£.
-~
E~il:;RYor'i I C liEVELOPNENT ·c~ J CK
""
••
r!
~,q;:
e.
0 A 0 0 0 0 0 • 0 0 J 0 0 0 • , ~ n •. U~~ ~~ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a ROO 0 0 0 0 0 0 0 0 0 0 ~ 0 .0 0 0 0 0 0 0 0 0 0
J'
F'U_m
A
\
c.
9 10 II 12 13 1 15 76 IJ 18 19 20 21 22 23 14 15 1 Zl 18 29 30 31 31 33 4 35 35 31 38 9 43 41 42 43 44 45 46 41 48 49 50 51 52 53 54 55 56 5J 58 59 60 61 61 63 64 6' 66 61 68 69 10 JI 12 13 1~ 15 16 JJ 18 19 gg
' 114°00;1°1
o000
TO'
!
~ 11
~I>PK.2SS
'tot
o &.;. 0 • 0 0 0 0 0 0 0 0 0 • 0 0 • • 0 0 0 • 0 0 0 0 Don 0 U0 0 0 0 ~ 0 0 0 0 0 0 0 0 0 0 0 0 • 0 0 • il • 0 0 0 0 0 0 0 0 0 0 0 0 ~ 0 0 0 0 0 0 0 0 0 0 0
1 7 1 4 5 6 1 ,8 9 10 II 12 131415 16 11 18 191021 22 13 242526 2l 28 2.9. 3a~I.. !~_}3 34J5~_6 JJ 38 3~_'O 4: 42 43 ~ .. ~! 4.6._4? 434950 51 51~3 .54 5556 51 ~~_59_60 61 615364 E' 66 6J 68691011 11 J3 H 15 16 1i J8 lJ 8J
r .;
x I u~
JUU 1 U 11 AURLIHA E
"~.,,
"
A V COORDlNATOR
SCHOUL
'2.'1
'I
'1.'1.
'1
') ') '2
o•
~ 0 • 0 0 0 A 0 0 0 0 0 0 0 0 0 0 0 0 ADO 0 0 0 0 0 G0 U 0 0 0 0 0 ~~ 0 0 0 0 0 0 0 0 0 ADO 0 0 0 a n 0 0 0 0 0 0 0 0 0 0 0 0 ~ 0 0 0 0 0 0 0 0 0 0 0
A
1 2 3 4 5 6 1 8 910 II 121314 15 16 11 \8 19202122132425261128293031323334353631383940414243 44454041 48 49 50 51,253545556515859 SO 616263646' 6& 6168691011 12
V"-·c~'oo 1
1. 'I 'I
1.,.
'I
o, ,
!l.
IF!> {)/Z#s"
j
0001
1ALiRORA F'-S'·
E'ti-f·s-r '-131· _.
Cu
II
"5 rc
'I
"
"'1..'1
e.
'1.2 ... , •
'1.'1
'J
~~
. .- ~OORD· it1StR-
3(4.~IAj()
Jj
~
"
Ii
r1ATRLS
'l.
'l.
L
L
!l;
LL
c.
•
.
I
000000000000
~ ~ 0000000000
35 ,J 1839 40 41 H4314 4i 16 4148 4950 5152535455565158 i9 60 61 62635.6' 66'1 caSl 1011 ni3 a J5 16 JJ 11 J9 80
!l.~; !lj
"'I "
14 J5 TS 11 78;980
(5 0 0=0
'I..
1
;1(
Sf:
~Jt
'1.
0 0 0 0 0 0 0 0 0 0 0 , 0 0 , , 0 0 0 ADO 0 0 DOG 0 U0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 A 0 0 A il
.t
I 1 3 4 5 6 1 3 910 "_12 131115_16 11'819102121231415.1621281930 lln1334
r]
1.
-a/'Ll/V.:)
,
n
HliROHA I LL I t10 'S 6050q
1417 i=" I to TH STF.:EET
....
76
.._....-.. . ......
(5"00/)
.
It
~
I
t~
APPENrJIX
NORT~I:RN
R
ILLINOIS UNIVI:RSITY
CONFIRMATI'ON COpy
EDUCATIONAL FILM LIBRARY
DE KALB, ILLINOIS 60115
REQUEST
NUMBER
I
IMPORTANT
SHIPPING
DATE
SHOW
DATE
DUE BACK
DATE
••••••••••••••••
...............
**
FILM
TITLE
FILM
NUMBER
RENTAL AND
• • • • • • • • • • • • • • •'
SERVICE CHARGE $ • • • • • • • • • • • • • • • ,
'
CUSTOMER
NUMBER
CUSTOMER
ORDER NO.
BILL
TO
SHIP
TO
• PLEASE NOTE ABOVE FILM IS A SUBSTITUTE BECAUSE FILM ORDERED IS NOT AVAILABLE.
•• 15¢ INSURANCE CHARGE ADDED FOR EACH FILM.
o
-------- NO~TH ;;N-~:L~ NO~-~~I~~;;I~Y -----------1
REQUEST
NUMBER
FILM
NUMBER
CUSTOMER
NUMBER
EDUCATIONAL FILM LIBRARY
DE KALB, ILLINOIS 60115
I
IMPORTANT
SHIPPING
DATE
SHOW
DATE
DUE BACK
DATE
FILM
TITLE
CUSTOMER
ORDER NO.
FROM:
AUDIO VISUAL CENTER
NORTHERN ILLINOIS UNIVERSITY
DE KALB, ILLINOIS 60115
TO:
BILL
TO
PACKING SLIP - NOT AN INVOICE
RETURN REQUESTED LIBRARY MATERIALS
16 MM MOTION PICTURE. NON FLAMMABLE
77
APPENDIX C
I-
I
NORT~ERN
ILLINOIS
UNIVERSITY
EDUCATIONAL FILM LIBRARY
DEPARTMENT OF INSTRUCTIONAL MATERIALS
DE KALB, ILLINOIS 60115
CUSTOMER NO.
INVOICE NO.
INVOICE DATE
PAGE NO.
**
15(t INSURANCE
CHARGE HAS
BEEN ADDED FOR
EACH FILM.
I
L-
~
REQUEST
NUMBER
SHOW DATE
FILM NUMBER
CUSTOMER
ORDER NUMBER
FILM TITLE
**
RENTAL AND
SERVICE CHARGE
t
PAY LAST AMOUNT IN THIS COLUMN _ _ _ _ _ _ _ _ _......
INSTRUCTIONAL MATERIAL
TRUST NO. 52428
SIGNATURE OF INDIVIDUAL PREPARING
MAKE CHECKS PAYABLE TO: NORTHERN ILLINOIS UNIVERSITY
MAIL TO:
NORTHERN ILLINOIS UNIVERSITY. BURSAR'S OFFICE. DE KALB. ILLINOIS 60115
78
INVOICE/STATEMENT
APPENDIX
D
AV-LOD
MANIPULATIONS
•••••••••••••••••••••••••••••••••••••
C'·"
"
•
FILMS
•
ADDRESSING
•
•••••••••••••••••••••••••••••••••••••
.TITLES. PRINTS. BILLING. SHIPPING.
• • • • • • • • • • • • • • • • • • • • • • • • • • • • •• • • • • • • • • • • • • • • • • • • •
•
•
•
•
•
•
• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •• • • • • •• •• •• • • • • •
.LOADING
1
1
1
•
•
•
•
• 1
• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • ••• •• • • • • • • • • • • •
4
.AUGMENTING. 2
3
5
•
•
•
•
• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •• • • • • • • • • • • •
-6.REPLACEMENT.
-7•
•
• • • • • • • • • • • •• • • • •• • • • •• •• • • • • • • • •• •••• • ••• ••• • • • •
BASIC REQUIREMENTS
AND
END RESULTS
••••••••••••••••••
1.
INFORMATION LOADED AS IT
APPEARS IN THE FILM AND CUSTOMER
FILE.
c
2. OPERATION FILLS GAPS IN TITLE
LIST OUTPUTED ON TYPEWRITER FROM
STEP ONE AS 'GAP' --INPUT FILM
CARD WITH NUMBER OF PRINTS TO BE
LOADED IN CC 9-10.
3. OPERAT10N ADDS PRINTS.
INPUT
FILM CARD WITH 9-10 BLANK AND THE
NUMBER OF PRINTS TO BE ADDED
APPEARING IN CC 19-8 U •
TYPEWRITER RESPONSE
• • • ••• • •••• ••• ••• ••
'GAP' CC NNNN FFFF
LEAVING A GAP IN FILM NUMBERS
C - CATAGORY
N - NEXT OPEN NUMBER
F - NEXT FILM NUMBER
PLUS APPROPRIATE ERROR MESSAGES
'ER' XX
XX - IDENTIFICATION NUMBER OF
ERROR
'AUGT' NNNNNNNPP LL
FILLING TITLE NUMBER GAPS
N - FILM NUMBER TO BE INSERTED
P - PRINT NUMBER FOR SAID FILM
L - NUMBER OF PRINTS STORED TO
DATE FOR SAID FILM
'AUGPt NNNNNNNPP LL
INSERTING PRINT
4. REPLACING POSITIONS HELD BY
DUMMY NUMBERS. AMOUNTS TO REPLACFM~NT
(7)
o
5. OPERATION ADDS SHIPPING
ADDRESSES FOR BILLING ADDRESSALREADY HAVING ONE OR MORE SHIPPING ADDRESS.
tINSS'
NNNNSS
(SHIPPING ADDRESS)
N - CUSTOMER NUMBER TO BE INSERTED
S - cORRESPONDING SHIPPING NUMBER
6. OPERATION REPLACES FILM W/{O
IN CC 80), OR W/O(BLANK IN CC 80)
CLEARING OF BOOKINGS ON ALL
PRINTS. REPLACES WITH SAME
NUMBER OF PRINTS. INPUT FILM
CAR~ WI CC 9-10 BLANK.
'FILM REPLACEMENT' NNNNNNNN - FILM NUMBER WITH STORED
BOOKINGS TO BE REPLACED
7. OPERATION REPLACES AND
INITIALIZES BILLING AND SHIPPING ADDRESS INDICATED.
'CUSTOMER REPLACEMENT'
(BILLING ADDRESS)
7r
o
SFSSlnf\1 f\JUtv1RFR
T.1.1
SPfAKFRS
MRS.
J nyc F F n n F R, F '" G T'" F F RIl'l G en MPUT I 1'.1 G L ARn RAT (l RY n (\1 n I S K
n ,6 T A. S Tn RAG FRO UTII" F f) 0 S R
/'~~\
~)
?O
I
DISK DATA STORAGE ROUTINE DDSR
c
Mrs. Joyce Fodor
Engineering Computing Laboratory
B554 Engineering Building
1415 W. Johnson Street
Madison, Wisconsin 53706
Description/Purpose
DDSR is a program,written in SPS II-D to facilitate permanent disk
storage of users' data from FORTRAN or SPS programs. The blocks of data
are given a name and table entries that are completely compatible with
the MONITOR I system. The user need not know what sectors on the disk are
available.
Machine Configuration Reguired
1.
2.
3.
4.
5.
1620 MOdel I or II
20 K memory
Card I/O
Indirect Addressing
1 Disk drive
General Program Description
Since rapid processing of the MONITOR system tables required approximately
10000 core positions, and it was undesirable to take this much core from the
user's program, the system uses the area from 02402 to 12000. In order to do
this, the routine must first store the contents of these locations on disk so
that it can restore them before returning to the main program. For this reason
the short form of the subroutines cannot be used when using these subroutines.
Both the disk write and the data recovery subroutines have routines which store
these core locations in cylinder fifteen of the work cylinders before calling
link to the main read and write routines. The read and write routines then
search the tables to find the particular entry called or the required storage,
and process the data.
All data blocks are given both DIM entries and Equivalence table entries.
When the routine is used to write data on disk permanently, the routine also
makes an entry in the sequential table.
Data blocks may be read from or written into di.k in the normal course of
any FORTRAN program, and may be used as often as the user desires, but the
maximum amount of core that may be stored at one time is 20000 digits. This
corresponds roughly to a singly dimensioned matrix of dimension 2000~ If
larger blocks of storage are required the arrays may be broken into parts and
stored that way.
Warnings
o
This program might not work with the short form of the FORTRAN subroutine
because core positions 2402 to 12000 are stored on disk and these core locations
are used as work area. If the call to these subroutines, and INDATA or OUTDATA
are located above this address the system should work.
B/
il
l'
All entries in an array should be defined prior to storing them on
disk o If the array is doubly subscripted this is very important. If
it is a singly subscripted array and all the elements from one to the
desired element is defined this is adequate.
o
1I
FORTRAN CALLING PROCEDURES
To Store Data
CALL INDATA (IND, LF, LK, NR)I ARRAY)
IND • 1
3
if data to be stored is a single fixed point variable or a
a fixed point array.
if data to be stored is a single floating point variable or
a floating point array.
LF
Is the floating point mantissa length for the program being used.
LK
Is the fixed point word length for the program being used.
NR
Is the number of elements to be stored, the dimension of the array
if it is single subscripted, or M N for a doubly subscripted
array.
A
Is the name of the array or number, fixed or floating point to be
stored.
*
To Call a Data Block
CALL OUTDAT (IND, LF, LK, NR, ARRAY)
where parameters are the same as above.
,~.
b~
Naming the Data Block
Whenever either of the routines is called, one data
from the card reader. This card should have name of the
read or written left justified in card columns 7 to 12.
to be sure that thi's card appears in the proper location
or error will result and the job will be terminated.
card will be read
block data to be
Care must be taken
in the data deck
DISK DATA STORAGE ROUTINE nnSR (1620 - 01.1.036) can be ordered from
the 1620 program library ~~rough your local representative or:
International Business Machines Corporation
DP Program Information Department
40 Saw Mill River Road
Hawthorne, N. Y. 10532
c~
,."
o
SESSION NUMBER
T.1.2
SPEAKERS
G. ROEMER OF IBM SPOKE ON THE CSMP FOR THE 1130. THE TOPICS
INCLUDED WERE THE DIFFERENCE BETWEEN DIGITAL AND ANALOG COMPUTERS,
BLOCK PROGRAMMING AN ANALOG COMPUTER, USAGE OF THE CSMP PACKAGE,
AND EXAMPLES.
P. WOODROW SPOKE ON BUFFERED AND OVERLAPPED 1/0. THIS PACKAGE
WILL BE RELEASED TO COMMON SHORTLY.
MEETING WAS CHAIRED BY LARRY WHELEN.
c
o
OVERLAPPED .1/0 FOR 1130 FORTRAN PROGRAMS
Peter J. Woodrow
Aeronautical Research Associates of Princeton, Inc.
Princeton, New Jersey
For most types of commercial computer applications, computation
is at a minimum and input/output operations are numerous. Thus it
is desirous, even at the cost of increasing core requirements somewhat, to use I/O routines that operate as efficiently timewise as
is possible. Unfortunately, the 1130 FORTRAN I/O routines were
designed to require minimum core. As a result, all FORTRAN I/O
routines (e.g. PRNTZ, CARDZ) use a common buffer that is filled
by SFIO. While presumably, ·on output at least, the I/O routines
could return immediately with SFIO waiting for completion before
starting a new line, this is not the case at present. All I/O
routines wait internally for the completion of the requested
operation. This means generally that I/O speeds from FORTRAN
programs are almost halved.
In order to speed up our cOITmlercial data processing, we decided to write some FORTRAN-callable routines, primarily for use
with COMET (since they all use A2 format in effect). iThese routines
were each written for a specific purpose but are presumably general
enough for different applications. We have never tried using these
routines with other FORTRAN I/O, primarily because we did not wish
to have SFIO loaded (approximately 800 words), but they should work
above modification level No. 5- if the interrupt levels used by
FORTRAN I/O are completely different from those used by these
routines. In particular, the card routine cannot be used by a
FORTRAN program that does FORTRAN I/O on paper tape or the
keyboard/printer.
o
o
2.
One point to note is that none of the routines is very complex
since all use standard assembler language I/O subroutines. There
would almost certainly be no great difficulty encountered in writing
a special version of PRNTZ to replace the supplied version and to
run at a speed twice as fast as the current FORTRAN version.
However, some core would have to be sacrificed. Replacing CARDZ
with a faster version is somewhat more difficult for reasons explained below in the brief write-ups that follow.
3.
o
SPCDR
This routine has three entry points; SPCDR, SRERD, and
SPCDP. The primary reason for writing this routine was not so
much to increase card reading speed as it was to perform some
special error checking and code conversion. If it is necessary
for computation to be done by the FORTRAN program in order to
determine whether to stacker select, this high card reading speed
is impossible. In addition, in most commercial applications,
each card is printed and hence(for the 1130 print speed is the
overriding consideration. For our application it was highly undesirable for the program to accept characters that would not
print on the 1132 printer (as does the FORTRAN card read routine).
In addition, since we have an 029 keypunch with the left-zeroinsertion feature, we felt it would be mandatory to accept -¢
(ll-¢ punch) which CARDZ translates into a blank. Following are
the various calls that one might use.
CALL SPCDR(NC, IFV, IFTV, IAREA, NERRC)
where
NC
IFV
IFTV
must contain the number of columns that are to be
read
a vector containing field end points. The routine
performs minor checking for specified fields on
the card. If, for example, the fields were 1 - 9,
l¢ - 19, .... etc., then IFV(l) = 9, IFV(2) = 19,
etc.
N.B. The last IFV element must be greater than or
equal to NC .
a vector containing on return to the calling program
the type of field corresponding to IFV. The
following codes are possible
1 - field contains only blanks
2 - field contains only numeric information
(no blanks)
96
4.
0
IAREA
NERRC
3 - field contains blanks and numeric information
4 - field contains only alphanumeric characters
5 - field contains blanks and alphanumeric
characters
6 - field contains alphanumeric and numeric, but
no blanks
7 field contains a mixture of all three; alpha~
numeric, numeric, and blanks
N.B. A (ll-numeric) punch in the last column of
a field is considered numeric (i.e. indicates
the field is to be considered negative).
a one-word integer vector dimensioned to contain
the input characters. IAREA must be dimensioned
at least NC/2 words (NC even) or (NC + 1)/2 words
(NC odd). The characters are packed two per word
as in the conventional A2 format.
returned as a zero if no errors were sensed; returned as a 1 if any character read is not
printable on the 1132 printer (the only exception
being -¢(ll - ¢ punch)). Characters in error
are replaced by blanks.
The above routine converts characters and checks field types as
the card is being read. After the card has been read, the routine
checks the error indicator provided by CARDl (indicating a 1442
sensed error) and retries the read and conversion if there was an
error.
In order to more fully understand the operation of the following
two calls, one must be aware of the fact that SPCDR contains an
internal buffer into which the card image is read. This is not
destroyed by conversion and is thus available for the following
two calls.
o
===;;;IIt1iML/JMwtttw: ijJ£(£WlJ 4,41#1 it ,.- ,,,.
..-.-""..,.,.,.--....
,
,~~~~--------
5.
CALL
SRERD(NC~
IFV~
o
IFTV, IAREA, NERRC)
This call is provided so that one may reread the same card with
different specified fields and thus first sense the type of card
to be processed and then actually check the appropriate fields.
No card reading takes place on this CALL. The explanation of the
parameters is the same as that for SPCDR above.
CALL SPCDP(NC, OAREA, NERRC)
where
NC
OAREA
NERRC
contains the number of columns to be punched
is a vector containing the characters to be
punched (see IAREA above)
returns the error code
o = no error
1 = at least one character to be punched not
printable on 1132 printer (except -¢)
2 - an already punched column (read on call to
SPCDR) is to be punched with a different,
non-blank character; i.e.,overpunching is
not allowed
If an error is sensed, the card will not be punched and an immediate
return to the calling program is instead executed.
Note that if SRERD is called after a call to SPCDP, it will in
effect be reading the Bunion" of the original card that was read
and the card to be punched, i.e. the final version of the card
as it would appear if it were read again after punching.
Note also that for proper operation all cards must be read by
SPCDR prior to being punched by SPCDP; otherwise, spurious errors
are likely to be generated.
0
1
6.
o
o
o
None of the above 'routines is particularly speedy. Their main
advantage is the thorough error checking that is done by them
rather than by the FORTRAN program. If maximum speed is desired
with no automatic error checking, then SFCDR which follows should
be used instead.
7.
o
SFCDR
This routine has four entry points; SFCDR, SFCDS, SFCDF,
and SFCDP. It was designed for a quite different environment.
Here we wished to read a considerable number of cards, operate
rather extensively on each card and maintain the results in core.
As a result, speed was the primary objective with no desire to
stacker select the cards. As a result, the routine was designed
with two buffers that alternate. Thus SFCDR immediately starts
reading the next card (unless otherwise instructed) and then
proceeds to convert the present, desired card. As a result of
this process conversion lion the fly" is not necessary and CARD¢
is used. However, the SPEED conversion routine is used and thus
all EBCDIC characters are converted properly and at maximum
speed. The FORTRAN CALLrs are as follows:
CALL SFCDR(NW, IAREA, NC)
c
where
NW
I AREA.
NC
is the number of words in IAREA. to be filled
(i.e. 1/2 the number of characters). The
number of characters must thus be even.
a one-word integer vector that is to contain
the converted card characters packed two per
word as in conventional A2 format. IAREA must
be dimensioned at least NW words long.
a code used to control the reading of the next
card while this one is being processed. If NC
is odd, no new card read is started (see SFCDS
below). If NC is even, then the next card read
is started unless the first column of the present
card happens to contain the card image code
equivalent of NC. Thus, if the last card of
one 1 s deck always had a 7 - 9 punch in column
one, then presumably one would not want to
start reading any card following a card with a
7 - 9 punch in column 1. The Subroutine Manual
90
()
..,
8.
o
gives the Hex IBM Card Code equivalent of a 7 - 9
punch as ¢¢5¢ which translates to a decimal
integer of 8¢. Hence, NC should equal 8¢ for this
example. If you wish to start reading the next
card in all circumstances, then NC should be set
to 2.
CALL SFCDS
This call is to allow the user's FORTRAN program to make a more
extensive test before starting a read operation on the next card.
Presumably, NC would be set to 1 in the call to SFCDR, a test
quickly made on the card, and then a call to SFCDS if the next
card is to be read. Note that SFCDS just starts a read operation;
it performs no actual reading and, in fact, has no effect if the
next card is already in the process of being read.
CALL SFCDF
c
This call causes a wait until current card operations are complete
and then initiates a card feed (with no read taking place at all)
on the next card. Note that if a card read operation were under
way at the time, then the next call to SFCDR will convert that
card and not the one after the card being fed through.
CALL SFCDP(NW, OAREA)
where
NW
OAREA
,0
contains the number of words to be punched (see
SFCDR)
contains the characters to be punched (see IAREA
above)
If a card read was under way at the time of this call, then the
read is completed, the punch data moved into the second buffer,
and the punch operation initiated. A call to SFCDR at this time
will cause proper conversion of the card that was read (and
which has already been punched by the call to SFCDP). If, however,
LS1&W1!GWlMJ J.C"J#.ifIi'Ifi"I/'A
'.~"
"".
9·
another call to SFCDP occurs prior to the call to SFCDR., then
the call to SFCDR will cause conversion of the data that was
punched into the second card via the second call to SFCDP.
c
Note that this routine always keeps track of whether the next
card read has been started or not; hence, it is unnecessary to
call SFCDS at any time unless the obvious resultant speed increase
is desired.
o
10.
c
SPRNT
This routine contains three entry points; SPRNT, SPRPT, and
SPRPC and is used to obtain high speed line-printer operation
from a FORTRAN program. Double buffering is not used because it
would result in only a very slight increase in speed at a cost
of an additional 60 words of storage. The calls are as follows:
CALL SPRNT(NW, OAREA)
where
NW
contains the number of words to be printed
(the number of-characters must be even and is
2
OAREA
*
NW)
is a one-word integer vector containing the
characters to be printed, packed two characters
per word as in the conventional A2 format
The print operation is started on this call and a wait for completion occurs either on the next call or on a call to either of
the following routines.
CALL SPRPT(NC)
where
NC
o
is returned as a 1 if a page eject occurred
after the last line was printed and is returned
as a ¢ otherwise.
Use of this routine will cause a definite degradation in the
printer speed. Generally, maximum speed will be obtained if.an
internal line counter is used to control page overflow. At any
rate, the call to SPRPT should be made immediately prior to the
printing of the next line (via a call to SPRNT). Note that the
page eje~t to the top of the next page (channel 1 punth in
carriage control tape) has already occurred automatically if NC
is returned as one.
93
Ii
I
11.
CALL SPRPC(NC)
, ,)
C
where
NC
contains a code for printer forms control. NC
is multiplied by 16 and transferred as is to
PRNTI (see Subroutine Library Manual) for printer
control. Thus NC = 16 will cause an immediate
skip to channel 1 and NC = 2¢8 will cause an
immediate space of 1
Note that the forms control command is merely initiated by this
call; it is completed on another call to any of the above entry
points.
I
j
I·
II
12.
FINAL NOTES
1)
All of the above sets of routines operate independently of
one another (except that SFCDR or SPCDR may be used, but not
both by the same program). In general, their operations overlap
one another with no problems and they have been used in a number
of applications with no difficulty.
2)
All sets of routines contain internal buffers into which the
characters are moved prior to a call to the proper I/O routine.
Hence the various character vectors may be operated on or changed
immediately after a call even though the operation may only have
been initiated.
I
C'
3)
Since two of the sets of routines may have I/O operations
under way concurrent with computation, PAUSE may not work
(Modification Level 5 should work, I hope). It it does not, then
use of IOND (in CSP Version 2) is suggested.
4)
I will be happy to send out binary decks, but may not get
around to submitting programs to COMMON for some time. I personally
do not feel that there is anything unusual or complicated about
these routines and make no comparisons with IDEAL or CSP, Version
2. They were written for our purposes and as a result have
features peculiar to our needs. It may easily be that these
features are wasted in your application.
" "" =~u;'n 1"t""" "!"''''. '"'~"'1·'~··'~'···""·"·!"
.,~ ..,.",. '·,~~,;~::,~;,:;:':.:,';:I·,;,;'-",.;;
,.
i
:1
II ASM
*L.IST
*PRINT SYMBOL TABLE
OOOR
225(3119
0000
22645644
nOBB
225(3117
OOO{) 1
0001 a
0002 0
0003 01
000:: 0
OOOe: 00
0008 01
ooaA
0
OOOR 1
oooe
OOOD
OOOE
0010
0
0
01
00
0012 01
0014 20
0015 0
0016 1
0017 1
0018 0
0019 0
OOIA 0
001B a
oOle 0
0010 0
OOIE 0
00lF 0
0020 0
OO?l 0
0022 0
0023 0
0024 01
OO?6 0
0027 0
O()28 0
0029 0
002A 0
0028 01
002D 01
OO?F
0031
0032
0033
0034
0036
01
0
0
0
01
01
0038 0
0039 0
003A 0
003R 01
0030 0
003E 01
0040 01
0042 0
0043 01
0045 0
0046 01
- - - - - - ..
~-.-----.---
-
0001
6973
6A74
65800000
6905
(5800000
D40000DF
700D
OOO(
6968
6A69
6580000B
(5800000
D40C)onDF
03059131
1000
OODF
OODA
(061
D061
(101
D03E
D027
(102
D042
(103
D05S
(104
D05A
(056
D480007D
(057
[')057
6100
C050
D055
6680007F
C60000DF
4(040c)2D
D04F
62(E
(04D
F6000162
4C18003E
7201
70F9
e043
D480007D
62CE
C6000194
4(1c)0048
C03C
95000045
E03(
EEOO0194
-_. .
ENT
SPCDR
ENT
SRERD
ENl
SRERD DC.
STX
STX
LOX
STX
LD
STO
i--1DX
speDR D(
STX
STX
LDX
LD
STO
LI8F
D(
SPCDP
.-
1 EXIT+l
2
11
1
II
EXIT+3
SRERD
SPCDR
L
IAREA
~DX
I\1DX
LD
STO
LDX
\1ATCH LD
sse
LD
IFADX S
AND
--.------~---
or~
-
REREAD CARD ItJ I TH NEW
FIELDS
C
0
RETRY
*
1 EXIT+l
2 EXIT+3
I 1 SPCDR
I1
a
L
IAREA
CAROl
(ALL SP(DR(!-.J(,IFV,IFTV,
IAREA,l\It::.RRC)
11000
IAREA
ERROR
ZERO
EF·· •
"
.
_o.:;,~....;;:...::..:-.w-;,;,.;;.;.
,,',""WI.... ,,'~ ..,".. ,I~,' ••• "'.',,'."IItb' .... ,I,',IIIt~~oIjnft".i.""'.,(10 ..,,,
..-......-
rJAGl:.
0171
0172
0173
0174
0175
0176
0177
0178
0179
O17A
017R
017C
O17n
n 1 7F'
()17r::
0
0
0
0
0
0
0
f)(
140C3
C
DC
DC
DC
D(
DC
IL+OC4
f)
140(5
140(6
140C7
140C8
140C9
IAODO
IAODl
IAODZ
IAOD3
IAOD4
IAODS
J 6
C.
01J0
F
c;
u 710
H
J"73u
f)(
0
AODO
ADDl
AOD2
AOD3
AOD4
AOD5
()
AClD6
f)C
()
AOD?
AOD8
DC
DC
GC
DC
DC
0
0
()
0
()
0180 0
OlRl
40(3
40C4
40C5
40C6
40C7
40C8
40C9
()
018;:> 0
()1 P·3 0
n 1 r~ 1+
r)
QIP5
0
o l'P6
()l P 7
0
AOC)9
40E2
40F.3
40E4
4UE5
IJC
DC
DC
f)C
DC
f)C
,)C
n1R9 ()
OlPA 0
nlRR. II
OlRe 0
OIRO 0
OIRt" 0
4CH:.9
C
DC
DC
DC
DC
2UFO
?()Fl
01 P,F 0
()190 0
()lql n
0192 ()
20F5
20Ff;
7UF7
?UFH
0193 0
20~9
0
OlAR 0
01Q4
'+Ol:.6
40E7
40E8
20F2
20F3
20F4
::>
I)
IAO'.J6
I A () [) 7
IAGDS
IAOD9
140E2
14G::'3
140EA
140t.5
14016
140't.7
06du
JILli
I
(-LEi~U
'j.J
)
J
K
J 7/-+'0
J7':J0
v 76":'
U7/U
u I C3 J
07')0
L
fVi
roJ
~)tiUv
0
P
U
vdlu
;~
0840
S
uo:JiJ
0dlU
0b .:JG
T
v~60
U
i...;b
V
U cH~ V
""X
0'jU.j
(v
vb')J
II+OER
y
140tSl
1
DC
12~)FO
J
J
DC
DC
120Fl
120 f:2
120F3
120F4
12:JF5
12DF6
12'JF7
120F8
IZOF9
1
\";'14lJ
'')'7:;-..)
f)(
DC
DC
r)c
)C
DC
DC
C
'J'71U
Uy~...;
J..
2
3
'-;lj\"!
v>7bv
Ii
5
0'-) -Iv
U:.1dv
(;
J'-J'7v
7
lvGv
f1
lu,lu
.i. ij 2. iJ
9
F. r\J I)
100
,~\
\~
~YMBOL
C
007F
0056
EXIT 0074
IFADX 0043
MSKCH 0083
PLP
0097
PSTO OOBF
SRCH OOD6
ZERO 007A
ceNT
CKEF
CHAR
EBCeL.
EXITP
LOOPA
MSKCT
PLPA
RETRY
SRERD
0081
0162
OOBA
0028
0084
OOAl
0018
0000
CHARLo
EROAD
IARAD
LOOPS
NOTS
PLPR
SLCH
rSTLC
TABLE
0130
007D
007C
0033
0048
009E
0085
006A
CHTYP 008U
ERRIC 0078
IAREA uODf
MATCH OO.3E
OFAD 0060
PMAT OOAB
SPCDP 008t3
TSTPF aace.
NO ERRORS IN ABOVE ASSEMBLY.
ie
/0/
eKEC
0066
ERROR OODA
IFAD 0059
MSKAB 0082
ONt:
007E
PREID ooce
SPCDI~ 0008
TWO
aUD~
PAGE
C:
:'1
C
0
0045
0046
0047
()O49
004A
0
0
01
20
DOF7
COOE
D4800048
03059130
1000
0
[)04B 1
004C
004C 01 74FF003E
004E 0 1000
004F 01 4(80003F
0000
0052
0052 1 OO~8
0053 1 OOEA
0054 0 0001
0055 0 0050
0056 0
0028
0057 1 0058
0058 0 (OES
0059 01 4410003F
0058 01 4(800057
0050 1 005E
C05E 0 6931
005F 01 658000;;D
0061 0 (101
0062 0 0018
0063 00 (5800000
0065 0
90FO
0066 a 4830
0067 0
1010
0068 a 80ED
0069 0 DODI
r06A ()
7102
006R 0 6926
006e 01 4(08008F
006E a 1001
006F 0 DOIB
0070 0 6100
0071 0 (8EO
0072 a 18DO
007'3 0 D8DE
0074 0 DOlg
0075 0 80DE
0076 0 D013
0077 0 80(3
0078 0 D01J
0079 0
D006
007A a (010
0078 01 D480008E
007D 01 (400a07F
007F 01 D5000081
0081 0 7101
0082 01 74FFOO7E
0084 01 74FFOO:3R
0086 a
70F6
0087 20 225(5144
00R8 0 0001
aOSA
0089 1
008A 1 008B
0088 1 008C
008C 20 03059130
008D 0 2000
RAD
STO
BUFA
LD
STO
LIBF
DC
K80
DC
MDX
NOP
BSC
BUFAD BSS
DC
RAD
CARDO
11000
...
L.
eCNT,-l
I
E
0
RCD
CBUFA
DC
CBUFB
ONE
DC
D(
1<80
K40
DC
SFCDS DC
1
80
40
*ceNT
L..D
aSl
as(
SFCDP DC
STX
LDX
L.D
STU
LD
S
BS(
SLA
A
STO
(\1DX
STX
SS(
PFAD
PTAD
SLA
STO
LDX
LDD
RTE
STD
STO
A
STO
A
STO
STO
LD
STO
LD
STO
*
I 1 SFCDP
1 1
PFAD+1
I1 0
K40
-z
16
K40
W(NT
1 2
1 PEXIT+l
L.
P'~ESI ,+
1
PCHCN
1 0
BUFAD
16
BUFAD
PAD
ONE
P(FAD+l
v'l(NT
PCFAD
PTAD+l
PCHCN
PAD
I
L
Ll
I\~DX
1
L
MDX
L
LIBF
DC
PCFAD DC
D(
PCH(N DC
LIBF
DC
SF(DS
1 PRESI+l
\1DX
rv1 DX
RCD.-
L.
I
*
*1
PFAD+ 1,-1
wCNT,-l
PFAD
SPEED
10001
*
*
*CARDa
12000
/03
2
,~-----,",.-----"",,,
...,",.,,"
~~-------
PAGt:.
OORE 1
008F 01
()O91 01
0093 1
0094 20
0095 0
009A 01
0098
OOEA
()1'1(
008F
65000091
4(O()OO93
009't
03059130
3000
4(800093
0052
0052
PAD
PRESI
PEXIT
SFCDF
DC
*
LDX Ll *
Bse L *
DC
*CARnQ
LIBf.
13000
DC
SFCDF
sse
CRUFA BSS
eBUFB BSS
END
3
,~)
~,
82
82
()
/o~
I
\'
I.:
SYMBOL TABl.E
0
~UFA
CFAD
FAD
0030
002A
002D
PCFAD 0089
PTAD 007F
SFCDP 005D
NO ERRORS
8UFAD 0052
CHeNT 002C
K40
0056
PCHCN 0088
RAD
0048
SFCDR 0000
CBUFA 0098
CODE 003'
K80
0055
PEXIT 009l
Reo
003F
SFCDS 0057
eBUFB OOEA
CONY 0028
ONE
00S4
PFAD 007D
RESIR 0037
TAD
002F
IN ABOVE ASSEMBL.Y.
c
o
lOS
ceNT
EXIT
PAD
003E
0039
008E
PRESl 008F
SFCDF 0093
weNT 0036
~.
I I ASi"1
*L.IST
*PRINT SY~BOL TABLE
0000
225D9563
0072
225D95E3
0080
225D95(3
0000 1 0001
0001 0 6929
0002 a 6A2A
0003 01 65800000
0005 a (101
0006 a D018
0007 20 176558F1
0008 0 0000
0009 0 70FC
OOOA 00 (5800000
OOO( 0 0028
OOOD 0
9023
OOOE 01 4C080012
0010 0 (020
0011 0 D023
0012. 0 COIF
0013 0
8021
0014 0
801E
0015 0 0008
0016 0
6200
0017 0
7102
0018 0
6916
(016
0019 0
001A 0 D019
001B 0 9019
001C 00 D4000001
001E 01 (6000020
0020 01 ·D5000022
72,FF
0022 0
1000
0023 0
0024 0
7101
0025 a 70F8
0026 20 176558F1
0027 0
2000
0028 1 0035
0029 1 008F
002A 01 6500002C
OO?C 01 6600002E
002E 01 4COOO030
0030 0 0000
0031 0 003C
0032 1 0035
0033 0
0001
0034
0001
0035
003D
0072 1 0073
0073 0
6987
0074 01 65800072
0076 20 176558Fl
0077 0
0000
0078 0 10FD
0079 0 COBA
007A 00 D5800000
007( 0
7101
007D 0
6981
007E 0
6AAE
ENT
ENT
ENT
~~~..;:u.--'-."-~
SPRNT
SPRPT
SPRP(
*
EXIT+l
EXIT+3
SPRNT
1
t<'lFOA+ 1
PRNT1
10000
TSTPL
0
OAREA
CaNST
*+2,+
CaNST
OAREA
OARAD
OAREA
ONE
i'v1TOA+ 1
0
SPRNT DC
STX
1
STX
2
LDX
I1
LD
1
STO
TSTPL LIBF
D(
t,1DX
LD
I1
STO
S
Bse L
LD
STO
LD
A
A
STO
LOX
2
MDX
1 2
STX
1 EXIT+5
LO
ZEI~O
STO
EPGFL
OAREA
S
STO L. 10001
MFOA LD
1.2 *
MTOA STO L1
*
MDX
2 -1
S.LA
0
MDX
1 1
MDX
MFOA
LIRF
Pf~NTl
DC
/2000
DC
OAREA
ERi~OR
DC
EXIT
LDX
L1
*
LOX L2 *
BSC L. *
ZERO DC
0
CONST DC
60
OARAD DC
OAREA
ONE
DC
1
EPGFl. BSS
1
OAREA BSS
61
SPRPT DC
*
STX
1 EXIT+1
LDX II SPRPT
TSTPF LIBF
PRNT1
DC
10000
MDX
TSTPF
LD
EPGFL
STO
I1 0
SPREX MDX
1 1
STX
1 EXIT+5
STX
2 EXt T+3
••• "-'-.".'-"--'-'
..
--:.:.~:.:.:.:.:,;;;~-~~:.;.;.;;.:.;.. :..:;,;~.:';':;,::,~:,:
-
.-
r~",
SPECIAL FOr~TRAN PRINT
FOR ACCOUNTING
SUBI~
V)
",
C
0
10'
I""
I.I~
II
I
-----------~----~-
pE'S
PAGE
0
Q07F
0080
0081
0082
0084
0086
0087
u
1
0
01
70AA
0081
69A9
65800080
00 (5800000
1004
0
0
EOOS
8005
0089 0 0001
008A 20 176558F1
008B 1 oose
008e 0 70EF
008D a OFFO
C08E a 3000
008F 1 0090
0090 0 COA2
0091 0 DOA2
0092 01 4C80008F
0094
0088 0
MDX
SPRPe DC
STX
L.DX
l.D
Sl..A
eXIT
*
1 EXIT+1
11 spa?c
11
f)
A-
MSK
AND
A
STO
CCWD
C(DWD
PRNT1
l..IBF
ceDWD DC
MDX
De
~SK
CCWD DC
ERROR DC
*SPREX
10FFO
13000
*ONE
LD
STO
sse
2-
1
EPGFL
ERROR
END
c
o
107
uamttklW, ,
,. \
, . 4M~44M : q
. - ,. ,
--- 4.. #. #44#.. ..
&W!W.!#k#¥iMf&hMAI1I!f.I:MQA.,· n_.
:1
:1
SYMBOL TABLE
CCDWD 008B
EXIT 002A
OAREA 0035
SPRPT 0072
CCWD
I'v1FOA
ONE
TSTPF
NO ERRORS IN ABOVt
008E
001 E
0033
CaNST 0031
MSK
008D
0076
TSTPL 0007
SPI~EX
007C
EPGFL 0034
0020
SPI~NT 0000
ZERO 0030
MTOA
ERROR OOaF
OARAD 0032
SPRPC 0080
o
ASSE~BLY.
c'
I
i
0,1".
10 f'
---~-------~----------------
.t
'W.:temWm'!!t1!!!!77!?I!7!!Z'""Zr
SESSION NUMBER
T.1.4.
SPEAKERS
J.L. TUNNEY, JR., CHAIRMAN S/360 HARDWARE COMMITTEE
f
I
DISCUSSION
THE MEETING CONSISTED OF
A SOUND-OFF ON HARDWARE PROBLEMS, AN
INFORMAL SURVEY OF MINIMUM CONFIGURATIONS THOUGHT TO BE REQUIRED
FOR VARIOUS SOFTWARE PACKAGES, AND A SURVEY OF USERS OF
NON-STANDARD EQUIPMENT.
f
\
I
l
Ie
I
109
.,14(1.'&;;
u.::a ,,,.,,
_._
,.#
H
REPORT ON MEETING OF SYSTEMS DIVISION, 360 PROJECT,
HARDWARE COMMITTEE - 9/7/67
Attendance:
Approx. 100
Chairman:
James L. Tunney
J. R. Ahart, Inc.
627 Salem Avenue
Dayton, Ohio 45406
Telephone 513 - 278-4754
The first meeting of this committee since its organization
was divided into three sections:
1. A discussion of users' hardware problems and how they
hav~ been remedied.
2. A discussion of software systems and what mlnlmum hardware requirements have proven to be in practice.
3. A survey among those attending of new or unusual equipment
that they might have on their "360".
SECTION I - HARDWARE PROBLEMS
~\
"1442" Card Read-Punch
V
The majority of the complaints about this unit involved an
apparent lack of checking circuits. Reports included machines
that would punch and read without cards, others instances where
the program would proceed normally without the punches actually
going down, and still other problems of punching without feeding
the cards. One user having both a "1442" and a "2501" found that
in order to "CATALR" from the "1442", both SYSIN and SYSRDR had
to be assigned to the "1442".
"1443" Printer
Most problems with the "1443" were evidently blamed on static
electricity by the FE's. One suggestion was to install a humidifier, another was to hang tinsel on metal frame over which the paper
moves, a third attendee mentioned that the box of paper must set
on the steel frame. Richard Pratt said that their printer could
throw the type-bar without leaving the ready state!
I
"2501" Card Reader
I
Four attendees reported that dust accumulated in one of the
read stations and caused frequent read checks. It was felt that
on heavily used equipment once-a-week PM was not sufficient to
relieve the problem and a design change is needed.
I/O
o
"2540" Card Read-Punch
Read-side: One person reported a machine that read 1200
cards per minute instead of 1000/minute when delivered. This
caused excessive reader checks and proved difficult to find.
Punch-side: Three installations reported excessive numbers
of partly punched holes without ,punch checks. New dies proved
to be the only effective remedy.
"1403" Printer
One user reported that on the QN print chain which is the FL/I
set with 45 preferred graphics, the single quote mark is a nonpreferred character. The net result is that any line having an
apostrophe prints at 330 lines/minute instead of 1100! Another
reported an oil leak that took 3 days to fix. Moral-put a pan
under it! Another installation reported that transfer of static
electricity from a belt to a wire resulted in character substitution.
"2311" Disk Drive
o
Six users reported oil leaks or hydraulic problems. Several
more said they frequently couldn't I.P.L. from certain drives. One
said the fix was to power the "2311" down, -slam the lid, and restart it!
Of more serious concern was the report that I.B.M. is currently
installing desensitizers on the heads of all "2311's". At least
two people reported that they had problems reading old files after
this engineering change. A theory was advanced that possibly old
"2311's!! were aLready below sensitivity specs!
"029" Keypunch
Duplicating ECBDIC cards with or without the printing mechanism
engaged is reported to damage the code plates. I.B.M. offers a
special "024 TT for this purpose but the users felt that the "026"
should be modified to do the job since it is the TT360" Keypunch!
SECTION II - MINIMUM HARDWARE REQUIREMENTS
FOR AVAILABLE SOFTWARE
o
T.O.S.
No comment other than no amount of core will speed it up!
D.O.S.
32K and 2 Disks should be minimum.
D.O.S.-FORTRAN IV
32K
D.O.S.-R.P.G.
minimum.
32K
minimum.
II I
D.O.S.-COBOL
D.O.S.-PL/I
32K marginal, 64K adequate.
same as FORTRAN - 32K minimum.
D.O.S.-Bill of Matl. Proc. with COBOL
M.I.T. Civil Eng. Pkg.
Proj. Mgmt. System
64K and several disks.
128K and 3 disks (comparable to what used
to fit on 40 or 60K TT1620 TT !
Rumor suggests 4 disks or 4 tapes!
D.O.S. Utilities and Sort-Merge
Good even In 16K!
SECTION III - NEW OR UNUSUAL EQUIPMENT
Record Overflow on T!2841 TT An excellent feature for $lO/month that
is rendered useless by lack of software support under D.O.S.
Storage Protect Out of 12 reported installations, 3 said it took
2 months or longer to get it to work!
"1012" Paper Tape Read-Punch R.P.Q.
AVlatlon, Denver, Colorado.
Leonard Sites, Sunstrand
r;\
"1231 fT Optical R~:;E. . Miami-Ddade qunbior cholle~e,. Miamlo, Florida d
reported that tillS lS a goo unlt ut t at It lS on l y supporte
~
under B.P.S. Lack of D.O.S. support results in an unrecoverable
unit check when D.O.S. is in core and someone readies the fT1221fT
or other unsupported device!
IT565IT CALCOMP Plotter on IT2701" Bethlehem Steel Company reports
that l t runs at !z "1620" speed under this hook-up.!
"470" CALCOMP Plotter off-line, 9 channel mag tape.
Don McIlvain
reports that 9 channel tape characters are very unusual and some
extra BAL programming is required for this arrangement.
IT2314" Disk Unit
Lear-Siegler Company, Grand Rapids, Michigan.
I.B.M. Visual Display Terminal
Data Cell
Also Lear-Siegler Company.
Data Corporation, Dayton, Ohio
R.C.A. Video-data Terminal
Also Data Corporation
/12
7:
::::n
o
SPF AK FP c:,
PI\f\IFL fl1\!
"r1 R
SF L != C T J n 1\1
nFSCRTPTTflf\IS
Tt\1
M Fin' IfVI
Af\\I! PFRsnf\.If\.IFL
T () SMA L L I t\1 S T l\ L L AT I fi ~,I S
PAl II. R Tr. K F n R n. [) F P A II hI I J 1\1 J \I FRS TTY
p ri P F R T C Cl R 1'.1 ELL, F F D. R F S. R Af\1 K, MI 1\1 f,1 F APn LIS
n P. L. 1-1. R ,l\ K!= R, PIn (\11: F R H I - R R F rJ C r) Rf\1 r. n •
n ~~.
P b, I I L H F R lr-' J T 7,
I. R • ~~ •
c
o
113
&&&&2&~.iMAU,g
ar 11(,..
."
J!II
q
o
Presentation Summary of a Member of a Panel
on
"Personnel Selection and Job Descriptions
in Medium-to-Small Computer Installations".
By
Paul A. Bickford
11'1
!!iw'ennrmw
C,
INTRODUCTION
,.
The comments below are centered around a small university (2,400
students) Computer Center which has an IBM 1620 20k, 1311 configuration
for education, a compliment of unit record equipment and an IBM 1401 Sk,
3-1311 system for University Administration Data Processing.
the main tasks of the installation are:
Some of
(1) Maintenance of a large
Alumni file (2) Student Admission files (3) Registration {4} Grade processing and reporting (5) Maintenance of Student Information files and
{6} Developments of Campaign Accounting files.
The Administrative
Data Processing group has always operated under a closed shop operation
philosophy whereas the Education group has always operated under an open
shop philosophy.
c)
All facilities are located
consists of eight people, one
secr~tary,
~n
one building.
The Staff
two Key Punch operators, one
File Clerk, one Programmer, one Machine Room Supervisor, Part-time
Delivery Boy and a Director.
PERSONNEL SELECTION
The procedures for hiring people are straight forward and derived
primarily out of necessity because of a very limited reservoir of prospective experienced people.
Attributes that the prospective em-
ployee's must have is a pleasant personality and who is reasonably
easy to get along with.
installation.
This is mandatory because of the size of the
One is constantly rubbing elbows and communicating with
another member of the staff and emotional friction here would be more
detrimental in disrupting output than an ailing
cpu.
o
/ 15"
&204=&.,:,
J
4
.j ., .
-
PA
.
&&&"'. >.4 ..
4..
..k&Wi&&HMl!IS.§#.L·· ..
- _ _ _ _ _ _ •__,._
. _.
~~
_ _.................:--_,_ _
~~
=.",.="
....=
......
_ _ _ _ _ _ _ _ _ _ _ _........_.,,,,,,,
... ,...,.
•.""",
.. ","""'"
••"""".H""""".,
.•.
=. . .'".'"."'.'""'. ;.;,".;,.,;.;. ....
~~_~
i
i:i
EDUCATION
A high school diploma is required.
for the positions of:
College experience is desired
Programmer and Machine Room Supervisor.
SEX
Women are employed for the positions of File Clerk and Key Punch
operators.
Men are employed for the positions of Machine Room Super-
visor, Programmer and Machine operators.
AGE
Women:
18 to 60
Men:
18 to ? depends upon persons experience and our needs.
Presently, the oldest man is 31 years.
JOB DESCRIPTIONS
Because the installation is small, there is an overlapping of
job responsibilities.
Each person must perform numerous and varied
tasks in order that continunity may be given to the flow of work thru
the Computer Center.
Sickness and vacations sometime present serious
difficulties because the overlapping of responsibilities never seems
to te adequately or properly defined.
The job titles are general ones
and the "other" tasks each person performs will be described.
KEY PUNCH OPERATOR(S)
Each must know how to operate the 548 Interpreter, 083 Sorter,
514 Reproducer, 085 Collator and the 407 Accounting Machine in addition
to the 026 Key Punch and 056 Verifier.
They must eventually become
familiar with (know by heart) the field formats of 20 plus different
1/6
L ... _
o
cards.
On some jobs, the Key Punch operator will receive information
directly in the mail, key punch and verify it, select the proper
Interpreter board (and sometimes wire one) process the cards thru the
548 and then file them in the proper file.
Many times the operator
must refer to a master code book in order that she may supply neede
source information codes to punch into cards.
Currently, one operator
is learning to program the 1401 in RPG language.
FILE CLERK
This individual primarily performs the task of hand filing updates
to the 75,000 card Alumni file which constantly changes.
Presently,
this procedure is followed even though the Alumni files reside on 1311
disks,. and ::shortly this hand task will be eliminated.
There are another
15 files that constantly change and the file clerk spends most of her
C'
time performing this function.
However, she must also be able to Key
Punch and operate the other Unit Record equipment when either of the
other two operators are ill or on vacation.
SECH.E'TARY
Besides performing the routine tasks of a secretary ie, typing
letters, distributing mail and filing etc. she must also be able to
Key Punch and operate most of the Unit Record equipment.
She also
updates Alumni address changes in an Alumni Directory which is 3 feet
thick.
Code sheets that are sent to the Computer Center are supplied
with additional codes (such as the student and alumni alphabetic
identification numbers) by the Secretary.
o
II 7
........... """ "..............................,......"'="'="., . . . .=
...=
. . . . '."""'.".=. . . =
. . . . . .=
. ." . =""=="="". =
. . . . .=
. . . .=.....="."..=.....=""....=. =...:;;
=.~
MACHINE ROOM SUPERVISOR
Much of the daily routine is closely watched and controlled by
the Machine Room Supervisor.
equipment.
He operates, wires and programs all
He accepts job requests (verbally or in writting) and
schedules them in a manner that will keep all equipment as busy as
possible.
With the many tasks that must "mesh" perfectly before a
job is completed he must be constantly aware of the state of completion that each job is at so that work can progress along.
This "art"
of scheduling seems to work reasonably well in our small shop environmente
The peak work periods seem to always present serious scheduling
problems and sometimes some jobs get completely neglected for awhile.
The Machine Room Supervisor carries the heaviest responsibility load
in seeing that jobs are finished in a reasonable length of time and
are well done.
He often communicates directly with the Faculty as
well as all Administrative offices.
One-fourth to one-half of his time,
approximately, is spent in programming the less complex "One-shot" jobs
that occur frequently.
PROGRAMMER
Is responsible for programming the more complex jobs such as
Development Fund Accounting, general Alumni and Student file maintenance.
He frequently takes requests directly, does what ever systems work is
necessary, writes and key punches his own program as well as assemble
and test the program.
On half of the jobs he is the only person who
knows precisely how his programs function.
Having only one person res-
ponsible for major jobs leaves a small installation in jeopardy.
The
c. ·:'
...,,1
II
e
~!
Programmer also writes the operating instructions for all of his
programs and is supposed to flow chart them also.
He is also responsible for knowing the more intimate details of
the IBM Sort Package, the Disk utility Programs and Disk File Organization routines.
DIRECTOR
Last, but I hope not least, the Director has the ultimate responsibility of the complete operation and he spends most of his time
communicating with the various departments of the University and
monitors the progress of new needed applications.
Sometimes he plays
the role of Programmer for a week or two, Machine Room Supervisor for
awhile besides carrying on the usual routine of answering complaints
(;.
(we have a few)
and carrying on with correspondence.
a course in Basic Computer Programming.
~
I~
111
He also teaches
PROGRAMMER SELECTION AT THE FEDERAL RESERVE BANK OF MINNEAPOLIS
~ -1,<,-1
Cortte.11
OUTLINE
sic StrJcture
...-
...
Programming staff of our company can be broken down into 5 basic units •
1. Programmer
2. Procedures Analyst
3. Systens A~alyst B
4. Systems Analyst
5. Senior Systems Analyst
Advancement along the scale follows naturally in that order, with the
"prog:c a.J.lli"Ue:c" being the bas ic unit, or s -carting po int.
v.
Even from the beginning, a programmer must be trained as a high-powered
individual in the many levels of work into which he may advance. His
duties include:
1. Present job analysis.
a. Discuss job to be programmed with supervisor or department head
in charge of the operation to be programmed.
b. Study and learn the "ins and outs" of the job by working with
the clerk; acquire a thorough understanding of all facets of the
job, including special operating methods, periodic, regular and
irregular error procedures.
2. Program design.
a. Determine program objectives through consultation with bank and
systems personnel.
b. Construct a flow chart of the work to be done by the computer.
3. Program writing/-cesting.
a. Writes the progra.J.~ - ideally in anyone of several programming
languages; selection of the one most suitable to the job.
b. Designs test procedure and sees to the preparation by keypunche~5
of a "test-data deck".
c. Run, test, debug program.
4. Installation/Implementation.
a. Assist in preparation of operator manuals.
o. Performs any necessary training of departmental and data
processing personnel concerned with the handling of the job.
~!el~ ')
0..
o.
~iring
4..
Normal procedure is to assign new programB'" to a "system" when
hired.
Most programmers will also be required, from time to time, to
help in the planning and programming of any special one-time
shots.
Techniques
We :~ire on three things:
3) ability to learn.
1) Pat score, 2) willingness to work; and,
C
"I
,..,iIi
/). 0
1
:tf
rrrrrF5'"
- 2 -
B.
o
Requirements
1. Education - High school diploma with good record> indication of
math interest. Any additional education, including college level
statistics, economics, accounting, or even IBM courses in functional
wiring or 'programming are considered; if absent, a willingness to
take co~rse work to fill in the gaps must be present.
2. Previous experience. We promote from within and train ourselves.
Expe~ience required varies with the job, and may be as low as none
for a programmer, to 5 years or more of bank experience for a
Senior Systems Analyst. We subscribe to the basic belief that
"Persons with an aptitude for this type of work (programming) can
usually gain job knowledge rapidly, but without the aptitude, no
amount of experience will produce a competent programmer.
3. Communications - In all levels of work, a fair amount of communication with other bank personnel will be required. The importance
of finding people'who are willing and able to communicate concretely
and calmly with others cannot be over emphasized. In his limited
supervisory level of work, the ability to communicate will be of
great value.
4. Personal charactc~istics.
a. Programming at any level requires initiative and ingenuity.
People must be found with the ability to absorb through observation, to suggest and implement improvement.
b. Programming is time consuming work, and requires a good amount
of initiative to complete a project ~ schedule.
c. Honesty important. As an employee moves into the level of
systems analysis, his principal goal will be to devise improved
methods, reduce costs and improve bank services. Among these
as well will be the responsibility to provide safeguards against
falsification and embezzlers, and against destruction - whether
inadvertant or careless - of valuable records.
o
/21
22&LC£ili&;:;MI#Ii';." IWS;,,,4Ji#U '·MMWW\4#Mi,
4¥;,¥i .#4
._ J
,~
o
Management Installation Division
ORGANIZA TION AND SELECTION
by
Dr. Paul S. Herwitz
IBM Corporation
Old Orchard Road
Armonk, New York 10504
(914) 765 - 4543
()
Thursday, 8:30 A.M.
Text, 7 pages
Graphics, 5 pages
Oi"
WUMttet:!!t!'fWt.' flT'W·· .,
c.
~··
·mw
ORGANIZATION AND SELECTION
,;,1
Although in the final analysis organization is people oriented, in IBM
we have a structure of jobs which seems to work very well in all of
our programming areas whether the group is large or small. I will
describe this structure to you in rather general terms without getting
into a completely detailed description of the job responsibilities and
requirements. As a matter of fact, this is an opportune time to talk
about the subj ect because I have been involved during the past nine
months in a study of the job descriptions that have existed for programmers for the last five years. We are just about at the end of
this now, and are making a number of revisions. In order to describe
the key positions correctly we conducted extensive interviews of
programmers throughout the company, and as a r-esult the new
descriptions give a pretty accurate description of the basic programming
jobs being performed today.
C\
I" ~
First of all (Figure 1) there are two ways to enter programming in IBMwith or without a college degree. Generally speaking the college graduate
can have any degree although we prefer a technical degree. (A little
later on I!ll give you a breakdown of the educational background of our
people.) When the college graduate is hired he enters in a classification
called pre -professional which is not exernpt from the Federal Wage and
Hours Laws. Depending upon the division in IBM in which he is hired,
our new pre -professional may have anywhere from six to twenty-six
weeks of progranlmer training. The twenty-six week training course is
populated by candidates for our programming systems activities. The
first twelve weeks of this course are formal class lecture with some handson experience. During the last/ fourteen weeks the trainee is a member of
a kind of prograrnming job shop, if you will, where he is managed by experienced programming managers and where :his assignment is to program
actual applications on a kind of sub -contract basis. This is a type of very
careful on -the -j 0 b training that is closely supervised by people who have
extensive management and training experience. At the end of the twentysix week period, those trainees who have successfully completed their
assignment (and not all of them do) are given a permanent assignment in
one of the regular programming groups.
Those new hires who do not go into the twenty-six week course, but who
have a shorter term - six or eight weeks, receive formal lectures and,
of course, some hands -on training. At the end of this formal training
period they go directly into a permanent assignment with one of the
programming groups.
-1-
/2.3
i&iIiLILM4&&MC;;;;.141 # (
I
(;
;;
; ,
(
g pg ,GP¥14W; . 4
.A4.•. 40
n.
Once the trainee has received his first permanent assignment, and
from then on, he is under the direct technical supervision of an
experienced pro-grammer. This experienced programmer will usually
be someone with at least two or more years of experience, and may be
the actual first line manager of the proj.ect.
From this point on, then, the pre-professional's progress depends upon
what he can demonstrate. On the average, after twelve to eighteen
months, the pre -professional is eligible for promotion to exempt status.
If he does not show that he is ready for promotion after two years, then
we feel he is not eligible to continue as a programmer. This means
either reassignment, or termination from the company.
For the non college graduate, we have several programming technician
positions which are designed to take care of two problems. The first
problem we are trying to solve is that of taking care of the nonprofessional activities that go with every programming job. When I
say non -professional I am really speaking in the legal sense as implied
by the Fair Labor Standards Act; and I am not trying to raise the
question of whether programming really is a profeSSion or not.
Remember that to be eligible for a profeSSional exemption one must
have as his primary duty IIwork requiring knowledge of an advanced
type in a field of science or learning customarily acquired by a prolonged course of specialized intellectual instruction or study, and this
work must require the consistent exercise of discretion and judgment. II
Alternatively, the work must be II original or creative in character in
a recognized field of artistic endeavor" and the result must depend
IT prirnarily on the invention, imagination, or talent of the employee. TT
Clearly there is much routine activity that goes along with all
programming jobs but does not meet the definition just stated for
profeSSional exemption. Some of these activities are for example,
setting up job decks for machine processing, key punching correction
cards, preparing in -put data for automated flowcharting, to sorlle
extent generating data for program testing, assembling program
documentation and writing procedures for the computer operator to
use, and up-dating operating systems by use of specifically designed
utility programs, and so on. USing these activities as a basis, we
have defined our first technician level job to consist primarily of $uch
activities. We think that one Programming Operations Aide can
perform these duties for a group of about a half dozen full time
programmers.
(~
~,~.y'
-2-
I"
tiM !feet r"MMnWi!! '7 r Kl!IY'Z7
The educational requirement for this position is a high school diploma.
This is a new position that has been in existence in IBM for less than
a year, and it is frankly experimental. We wonrt know for another
year or two just how well it will work out. If the candidate has college
credits but no college degree, after his initial training period he
becomes a Programming Technician. The job of the Programming
Technician includes some of the activities of the Programming
Operations Aide, but also includes coding of well defined sub -routines
from detailed flowcharts.
At the top of our technician path we have a Senior Programming
Technician. His duties are primarily to flowchart, code, and debug
complete computer programs from well defined specifications. These
programs are usually at the level of routine utility programs, and are
essentially always self-contained. The Senior Programming Technician
is of course also responsible for the documentation of his program, and
more or less for the installation of the program in an operational
enviroment.
The second problem we hope to solve by use of the technician career
path is, of course, the problem of scarcity of qualified programmers.
We hope to attack this problem in two ways. First of all, if indeed we
do succeed in defining the non -professional activities that accompany
the program task and in delegating these to our technicians, then we
have relieved the profeSSional programmer of some of his less productive activity; and he will devote, his time to more creative tasks, such
as systems specification and design. The net result then is to let the
more creative programmer spend more of his time in creative activity,
thus increaSing the total productivity of the entire programming group.
Secondly its clear that many of the non -professional activities I have
described don't require the attention of a college graduate. Thus, use
of the technician positions permits us to consider a group of prospects
that is much broader than we could consider without these positions.
Our first level technician requires only a high school or prep school
diploma. The top level technician requires two years of college or the
equivalent.
We think the technician career path is quite good -the first level
techllician, the Programming Operations Aide, can either learn
programming, or can go into machine operations. The top level
technician, the prograrnming specialist, is certainly sophisticated
enough to go into operations management of a reasonably largecomputing center. On the other hand, if indeed he truly demonstrates a keen
understanding of programming, and if his manager makes the business
-3I~S
judg-ment that he can be successful in programming, then he to can
be considered for promotion to the same exempt position to which
we promote the pre -professional. This is not a loophole; it is up
to management to present evidence that the man truly has learned
enough about the job that he can be expected to perform in all ways
as well as the pre -professional who was promoted. Our statistics
show that some twenty or twenty-one percent of our professional
programmers do not have a college degree.
c
Our first key position, then, is that held by our first level exempt
programmer. Traditionally he is known as an Associate Programmer.
The position is key because the Associate Programmer must truly
demonstrate that his career lies in programming.
The next key position, and probably the most important one we have,
is a position at what we call the staff level (Figure 2). There are
really three jobs at this level - all equally important to the company,
and all comparably rewarded. It is at this level that our first true
manager appears. This Project Programmer is in all respects a
manager; he is responsible for hiring, training, supervising, salary
administration, promotion, and firing. We expect him to be technically qualified to direct a project, and to be fully involved
technically in the programming activities of the project. Ideally he
manages a group of six to ten programmers. If the group is small,
he can spend roughly seventy percent of his time in technical matters,
the remaining time being given over to administrative and personnel
activities. If the project is larger, then he will typically have in his
group a programmer at his own level who will act as a project leader.
This project leader is called a Staff Programmer and is the second
of the three jobs I mentioned. The project leader has all of the
technical responsibilities that the proj ect manager has, but does not
have the administrative and personnel responsibilities.
Roughly half of our staff level programmers are managers, another
quarter perform the proj ect leader duties, and the remaining staff
level programmers -the third job-act in the classical staff capacity that is, they perform as an extension of management. Generally
speaking they act as trouble shooters, as gatherers and analysisers
of data, as technical consultants, and to some extent can make
decisions when the authority has been delegated to them. They also
double in brass as proj ect leaders when circumstances warrantusually for short periods of time.
-4-
I,
o
We are organized this way because we want our projects to be small,
and we think the person in the best position to perform the managerial
duties is really the leader of the proj ect. After all, he is the man on
the spot who is in the best position to know the capabilities of the
people who work for him and, therefore, given adequate training, he is
in the best position to make the managerial decisions that relate to his
people.
The use of the Staff Programmer as a project leader is really a
compromise. Since IBM's programming activities are expanding
almost as rapidly as .the programming field itself, it is always difficult
to find people who are well qualified to assume the combined technical
and administrative responsibilities of first line management as we
conceive it. Moreover, many of our technically qualified people are
just not interested in becoming managers. Thus there is always a
scarcity of first level managers and we are faced with the necessity
of using technical project leaders. The problem we have to guard
against is that if a proj ect grows too large then its first level
manager may tend to shift the administrative and personnel responsibilities onto the shoulders of the project leader. The proj ect leader
then becomes a manager without portfolio and without accountability.
~
!,'I
I
I
,
'\
I
c
From the technical standpoint, the project manager or project leader
works from objectives. He is responsible for a complex project, and
he is responsible for specification, negotiation of interfaces, implementation, documentation, the whole works. In general, he is not
responsible for the total programming system or large application,
but for a complex but recognizable sub-section. Typical examples
would be a FORTRAN compiler, a set of mathematical subroutines,
etc. The total systems or applications responsibility rests with the
next level or even the one beyond that. I won't go into detail about
these next two levels except to say that we have both managerial and
non managerial positions at these higher levels as you can see from
Figure 2. We have dual promotional opportunities in the managerial
and technical activities.
A little arithmetic (Figure 3) will show that if we restrict the project
manager to a group of six and if we restrict the second and third level
managers each to having six managers report to him, then if we
include the managers in our counts we can take care of a group of
fourty-three people with only two levels of management and this grows
-5-
12. 7
~~~~~----------.--.------,-~~~~---
to 259 people if we have one third level manager. Analogously, if
we allow each first level manager to manage ten people but still say
that a second or third level manager may only have six managers
reporting to him then the three levels of management include 403
people. In the first case if the ratio of people to first level managers
is six to one then the overall man-manager ratio comes out five to
one. If the first level ratio is ten to one, then the overall ratio
becomes slightly more than eight to one.
I have just one more comment on the responsibilities of the proj ect
manager. Clearly if the programs to be written are not so complex
as to require six programmers or more on the particular project,
there is no reason why a project'manager may not have the responsibility for implementing a number of small programs. In this case he
still manag es a small group in which one or more of his programmers
are responsible for a complete program.
We think the entire structure is quite flexible, and have found that it
works very well for us over a variety of technical activities.
As I promised earlier, I want to show you a breakdown of the
educational background of our programmers as of the end of 1966.
Figure 4 shows the educational level and Figure 5 the field of study of
our exempt programmers. Four percent of the records I consulted
did not give the educational level, and fifteen percent did not carry
the field of study. Otherwise, the figur~s should be self-explanatory.
Finally, I'd like to turn my attention now briefly to the question of
selection. I say briefly because I don't have too much to tell you. I
have talked about our educational requirements. Beyond this, all
inexperienced candidates are required to take the current version of
the PAT test. We are presently involved in a study which we hope
will help us better validate the PAT test and which we hope will also
give us a lead on other tests which might possibly be used in the
selection process. Presently our study has given us ample indication
that we don't understand all the ramifications of the PAT test. This
test was originally validated against performance in training programs.
It has never been satisfactorally validated to my knowledge in IBM
against actual programmer performance. We are attempting to do
this now, and will make the results available after the study is
completed.
1
(
~
.
I
lIII
l'
-6-
1·'1,1.:
II:
r
i
,
{~:::
,
,"
l,
j
~
nn'tf'tttwttlUt1!trr--:
c'
,rrl!
We also have indications that there will be one or two other tests that
might be very useful in predicting whether or not a programmer will
be successful. Unfortunately, at this point in time it is premature to
discuss our findings. Again, though, we will be happy to make our
information available when the study is complete.
Beyond this, I can only tell you that the judgments you must make
when you hire programmers are bound to be subjective. In my own
experience I have found that the interviewers who are right more
often than wrong in the subjective judgments are the interviewers
who are most people oriented. In lieu of any other more objective
indicators, I think you must select as interviewers those people on
your technical staff who appear to be most sensitive to the people
around them.
-7-
o
=. .
." .. ~,.."""""
..., .• "....
=
~ ....
"
"",,,=~,,,,,-~=".~,-,,.-=,,="'~"
". . . . . . . .- -.... ..
~.-- ,-~--
..,
.------------,
I ]) P OphS.
L.-y___ ,H~~t.1
___ J
,,~
\
\
,,/
\
r---~---,
I ~I)\(t~~
I
Op.,s. I
L __ jf ___ .J
t
( Tra\vaee. )
/
1l'
/
/
Go 'I~e
)eCflr~~
(-rra\~ee )
1l'
U-'i&cr.c..\t Dol
) I t> 10 "'i
~~
C
"''')
/30
--
--~-,-~-"---'
'_.
---,
"._,,------,------,---"
I,,,
o
EXEMPT PR06-RAMHlNG- 1>OSITloNS
o
/3/
\ ~t f,..>2.u-el
~
6
tao)
(10)
,,•
L
,
•
\
The~
!f
rV"tl .. t:.~n~~
S~i
Li\70\ ~~l\
( rJo""I'''\Vs
tv.-t I-'C,,&\ ""'\'"
- -"I
--
T)
NO\!\" ~«¢l vs
AU Mtts
( NOM· v>\fars
AU &\CtVS
- -saI
-...
f)
coo.,
/32
o
-PROGRAMMERS
\)\f fDf)cAtlo~
LfVE"L
!l!
Assoe,aie of Arls
2.
Co lI~te Gved.t, vao d.et~e~
11
q
No
-
c.ott~te. ~yedtt
1'"5
~e
BA O~ 1\S
5"1'
HA C~ MS
1>,,)
l'f
-I
l'
o
133
I
\
/r..",
~o7/
C
I ••
I,~
13J.1
:j{
r
o
PRO~RAMMfRS ~\f
FIEL'P C)f STt> DY
-010
Bu.~.~e\s
fcl,-"a:ho"
2-
E" t&'~..et.~~
J:,,,,~ &\\d Apphed Arts
IS-
t,.., beY'a I Ayts
i
CI
IO.t
MatlA
.sr
Ifl
')
Sc., ... ~c.~
!
t~
!O
/35
SESSION NUMBER
T.2.1.
SPEAKERS
DAN FULLAN (IBM) PRESENTED WM. GARRISON (IBM), MODERATED BY
D.R. MC ILVAIN
DISCUSSION
PRESENTATION ON PL/I BY IBM, INDICATING THE INTENT OF PL/It
USAGE & EXPERIENCE TO DATE, AND FUTURE PLANS (INCLUDING ADDITIONAL
PUBLICATIONS FOR USER EDUCATION).
FUTURE PLANS FOR PL/I UNDER DOS INCLUDE
1. ADDITIONAL DOCUMENTATION
2.
INCLUSION OF INITIAL ATTRIBUTE
3. SCIENTIFIC SUBROUTINES WILL BE AVAILABLE FOR OS PL/I AS
TYPE 3 SUPPORTC'
4. EXPANSION OF OBJECT TIME DIAGNOSTICS
5. CONSIDERATION OF CREATION OF IS FILE UNDER DOS PL/I.
6.
SIFTS FROM COBOL OR FORTRAN ARE BEING CONSIDERED.
7. PL/I OBJECT PROGRAMS WILL BE AVAILABLE FOR OBJECT TIME
EXECUTING IN FOREGROUND UNDER RELEASE OF DOS SCHEDULED
~\
FOR 4 1 6 8 . ' , . )
THE PHILOSOPHY BEHIND PL/I HAS BEEN THAT SUPPORT OF EXTENSIONS
IN 3RD GENERATION EQUIPMENT & TECHNOLOGY WOULD BE MADE
PREFERENTIALLY INTO PL/I OVER FORTRAN OR COBOL.
/3b
i
SESSION NUMBER
T.2.3.
SPEAKERS
LAURA AUSTIN
DISCUSSION
VOLUNTEERS WERE SOLICITED FOR SERVICE ON THE THE NOMINATING
COMMITTEE, EXECUTIVE BOARD, AND PROJECTS. THERE WERE 10 PEOPLE IN
ATTENDANCE. NAMES WILL BE REFERRED TO THE APPROPRIATE PERSONS.
C~
10
/37
____.,._ .4...
.eM ... L¥T4$!?
Wli,q..iili¥dJ-·
I
I
i
SESSION NUMBER
T.2.4.
SPEAKER S
BRIAN SWAIN OF SHAWINIGAN ENGINEERING SPOKE ON THE 1130 SING~E OISK
SORT PROGRAM. W.C. BLACKNEY OF DOW SPOKE ON RUNNING A MEMORYSCQPE
'THRU THE 1130, 1627 PLOTTER ATTACHMENT.
DISCUSSION
LARRY WHALEN CHAIRED THE MEETING. WE HAVE DECIDED TO ESTABLISH
A COMMITTEE TO INVESTIGATE FORTRAN AND MONITOR V2. THIS COMMITTEE
WILL BE FORMALLY ORGANIZED AT A LATER DATE.
/3(>
I'
I
()
PMERG
-
A FAST SOff!' -MERGE
SUBROUTINE
FOR
IBM
1130
by
B. J. SWAIN
THE
0'
SMW"INIGMT
ENGINEERTITG
SEFTEHBER
COMPANY
LIMITED
1967
/ 3 ~_
~~~~
_ _ _ _ _ _ _..._ _ _'_"'_""......"!_"'.='
..!,"='"'"'="''''=''"'''"","''K'''''''"''=''''"''''''';;;;'''=''"_"
~~~'"'
_ _ _ __
-18Appendix 2
PROGRAM S'l'RUC1I URE
The follcv.'ing prosra7l1s s.re included in the package:
PMERG
Principal subprogram, calls Sn:.rPH and MRGP!!
to perform sorting and merging operations
respectively_
SRI'PH
Extracts keys from records.
operation.
Perfor-ills sorting
Hakes up a file containing sort
keys and pointers to records included in the
sort.
}.n~GPI-I
Performs merging opet'B,tion.
Outputs the file
of pointers to records in the main file.
!FTSK
Function subprogre:=1,2
C 203 DO 209 IX=1,3
IP( IX)=IGHST(JG)
ro{
EXT
IX)=IG~ST(JG1)
JG~JG+l
209 JG1'=JG1+l
IF«~YTP-
214 CALL
2) 214,214,213
SD(~,P,Q)
IF(Ii\jT(~))
290,201,299
213 IF(P-Q) 290,201.299
C
CO~PARE ALDHAgETIC KEYS
20~
(L=~EYS{~,!K)
CAL L
,~
IF( IR)
207
C8 ,1,1 P \ r GH S T( J G ) , 1 , I G:~ S T ( J G1 ) , 1 , KL , I R )
299,287,290
JGI~(=(KL+1)/2
i\;(
JG=JG+J(;I
JGl=JG1"'JSI,\:C
c
<::YS ~~:: ID:::':TICt.,L
I F ( :CJ l - <~ 2 1 2 9 1 , 2 9 1 t 2 9 2
29:
I=:<~Y!<1
29:
IPTS{=l
--_._-,---
?92,291,291
172
o
"
-29-
o
PAGE 02
299 IF(KEYIK) 291.292,292
292 IPTSK=2
RETURN
END
FEATuRES SUPPORTED
ONE WORD INTEGERS
COQE
END OF
o
FOR IPTSK
VARIABLES
REQuIRE~ENTS
CO~~ON
0
22
PROGRAM
312
CO~PILATION
173
·30II FOR
*ONE WORD INTEGERS
*LIST SOURCE PROGRAM
SUBROUTINE DOPENCIBUF.N,J,L)
iHIS SUBROUTINE OPENS A FILE BUFFER.SETTING INITIAL VALUES
C
THE FILE CONTROL WORDS.
o
TO
C
C
C
C
C
DIMENSION IBUF(325)
ISUF = NAME OF FILE BUFFER
N : FILE NUMBER
J
= CORE RECORD LENGTH
L
= NUMBER OF RECORDS IN BLOCK
C
IBUF(l)
IBUF(2)
I BUF (3)
C
C
C
C
C
C
C
C
C
=N
J
=L
:II
IBUF(4) = FILE TYPE
=1 WHEN BLOCK HAS ONLY BEEN USED FOR TRANSFERRING A
RECORD FROM IBUF TO AN ARRAY(FOR GET OPERATIONS)
=2 WHEN BLOCK HAS BEEN USED TO TRANSFER A RECORD FROM
AN ARRAY TO IBUF(FOR PUT OPERATIONS)
SET = 1 INITIALLY
IBUF(4) = 1
IBUF(S)
C
= NUMBER
OF FIRST RECORD IN BLOCK
SET = 0 INITIALLY
C
C
IBUF(S) = 0
RETURN
END
FEATURES SUPPORTED
ONE WORD INTEGERS
CORE REQUIREMENTS FOR DOPEN
COMMON
0 VARIABLES
6
PROGRAM
50
END OF COMPILATION
17'1
----------------
THE
~
-31-
o
II FOR
*ONE WORD INTEGERS
*LIsT SOURCE PROGRAM
SUBROUTINE DGET(IBUF,K,lA)
C THIS SUBROUTINE TRANSfERS RECORD K OF A FILE TO ARRAY lA.
C IF THE REQUIRED RECORD IS ALREADY RESIDENT IN THE BUFFER, IT IS
C IMMEDIATELY TRANSFERRED
C IF IT IS NOT IN THE BUFFER, THE BLOCK OF RECORDS IN THE BUFFER
C IS STORED IF NECESSARY, AND THE CORRECT BLOCK OF RECORDS
C OBTAINED FROM THE DISK, AFTER WHICH TRANSFER OF THE RECORD TAKES PLACE
C
C
C
C
C
DIMENSION IBUF(325),IA(lOO)
IBUF = NAME OF FILE BUFFER
K
= RECORD TO CONTAIN ARRAY IA
IA
= REQUIRED ARRAY
CALL DUSE(IBUF,K,l,l,KL)
C HERE TO TRANSFER RECORD FROM IBUF
J= IBUF(2)
DO 5 JJ = 1,J
IA(JJ) = IBUF(KL)
5 KL = KL + 1
RETURN
END
FEATURES SUPPORTED
ONE WORD INTEGERS
CORE REQUIREMENTS FOR DGET
COMMON
0 VARIABLES
C~
();
6
PROGRAM
62
END OF COMPI LAT ION
17S
.______________________________ THE
-32II FOR
*ONE WORD INTEGERS
*LIST SOURCE PROGRAM
SUBRoUtINE DPUT(IBUF,K,IA)
C THIS SUBROUTINE TRANSFERS AN ARRAY IA TO RECORD K OF A FILE.
C IF THE REQUIRED alOCK IS ALREADY RESIDENT IN THE BUFFER, THE ARRAY
C IS IMMEDIATELY TRANSFERRED
C IF IT IS NOT IN THE BUFFER, THE BLOCK IK THE BUFFER IS STORED
C IF NECESSARY, AND THE REQUIRED BLOCK OBTAINED FROM THE DISK,
C AFTER WHICH TRANSFER OF THE ARRAY TAKES PLACE.
C THE CONTENTS OF IA ARE NOT WRITTEN ON THE DISK.
C
C
C
C
C
o
DIMENSION IBUF(325),IA(lOO)
IBUF = NAME OF FILE BUFFER
K
= RECORD TO CONTAIN ARRAY IA
IA
= REQUIRED ARRAY
CALL DUSE(IBUF,K,1,2.Kl)
C HERE TO TRANSFER RECORD FROM IA
J= IBUF(2)
DO 5 JJ = l,J
IBUF(KL) = IAeJJ)
5 KL = KL + 1
RETURN
END
FEATURES SUPPORTED
ONE WORD INTEGERS
CORE REQUIREMENTS FOR DPUT
COMMON
0 VARIABLES
6
PROGRAM
o
62
END OF COMPILATION
176
I
01"
i'Wt",--,-
-33II FOR
*ONE WORD INTEGERS
*LIST SOURCE PROGRAM
SUBROUTINE DCLOS(IBUF)
<:
C
THIS SUBROUTINE CLOSES THE FILE
C
IF A BLOCK OF RECORDS REQUIRES TRANSFER TO THE DISK. THE TRANSFER
C I S MADE.
DIMENSION IBUF(32S)
c
C
C
c
c
c
IBUF
= NAME
OF FILE BUFFER
KK = IBUF(4)
GO TO (3,4),KK
3 RETURN
4 N = IBUF(l)
Ll = (IBUF(Z)*IBUF(3») + S
LL
= NUMBER
OF LAST WORD IN BLOCK
= IBUF(S)
= MINO(320,(LL-S»
L
= FILE RECORD LENGTH
NREC = ( l l + L - 6)/L
NREC = NO OF FILE RECORDS
Kl
c
c
c
C
C
c;'
c
c
C
C
C
L
IN A BLOCK
K3 =«KI/IBUF(3»*NREC) + 1
K3
= RECORD
WRITECN'K3)
RETURN
END
NUMBER OF THE FIRST FILE RECORD IN THE BLOCK
CONTAINING RECORD K
(IBUF( 1),1 =6,LL)
FEATURES SUPPORTED
ONE WORD INTEGERS
CORE REQUIREMENTS FOR DCLOS
COMMON·
0 VARIABLES
16
PROGRAM
116
END OF COMPILATION
o
177
______________________
~
________________________________________________ THE
J&&£&d&EJ4#M4M.SJMhM""'U
t.
. ,., "¥%R4#4¢¥
42 A¥
--------------------------------------------------
--------~~.~
/ / FOR
*ONE WORD INTEGERS
*LIST SOURCE PROGRAM
SUBROUTINE DUSE(IBUF,K,M,KEY,KL)
DI~ENSION ISUF(325)
C
THIS SUBROUTINE GENERATES A POINTER TO ONE WORD OF A DISK
C
RECORD, SO THAT WORD CAN BE REFERENCED IN A SUBSEQUENT
C
OPERATION
C
IF THE BLOCK CONTAINING THE REQUIRED RECORD IS ALREADY IN
C
THE BUFFER, THE POINTER IS GENERATED IMMEDIATELY
(
IF IT IS NOT IN THE BUFFER, THE BLOCK IN THE BUFFER IS
(
STORED IF NECESSARY, AND THE REQUIRED BLOCK OBTAINED
C
FROM THE DISK, AFTER WHICH GENERATION OF THE POINTER TAKES
C
PLACE
C
IBUF =NA~E OF FILE BUFFER
C
K
=RECORD REQUIRED
C
M
=WORD REQUIRED
C
KEY =SwITCH TO INDICATE IF SUBSEQUENT USE OF POINTER WILL
C
CAUSE RECORD TO BE CHANGED. IF SO, KEY=2. IF NOT, KEY=l
C
KL
=POINTER TO WORD REQUIRED
C
FIRST 5 WORDS OF IBUF ARE CONTROL WORDS
C
IBUF(1l= FILE NU~RER (N)
C
IBUF(2l= LOGICAL RECORD LENGTH (J)
C
IBUF(3)= NO. OF LOGICAL RECORDS IN A BLOCK (L)
C
IBUF(4)= FILE TYPE DEFINITION =1 INITIALLY
=2 WHEN BUFFER HAS BEEN MODIFIED
C
C
BY PUT OPERATIONS, AND HENCE IS DIFFERENl FROM DISK
C
IBUF(5)= NO. OF FIRST RECORD IN BLOCK =0 IF BUFFER HAS NOT
C
BEEN FILLED
L=I8UF(3)
Kl= (( K-l) Ii..) *L+1
c
K1 =RECORD NO. OF FIRST LOGICAL ~ECORD IN THE BLOCK CONTAINC
ING RECORD K
J=IBUF(2)
K2=I8UF(5)
IF(KI-K2) 1,2,1
1 N=IBUF(l)
L2=J*L
c
L2 =NO. OF WORDS IN BLOCK
Ll=MINO(:20,L2)
C
Ll
=PHYSICAL RECORD LENGTH
NREC=(l2+Ll-l)/Ll
C
NREC=NO OF LOGICAL RECORDS IN PHYSICAL RECORD
C
-SAME AS BLOCKING FACTOR IF PHYSICAL RECORD LENGTH.LE.320
LL=L2+5
c
KK=IBUF(4)
C
C
C
C
C
INSERT THE FOLLOWING STATEMENTS IF A TRACE ON SW 15 IS REQUIRED
THESE CAUSE THE VALUES OF ALL PARA~ETERS TO BE PRINTED WHENEVER
A PHYSICAL RECORD IS TRANSFERRED TO OR FROM DISK
CALL DATSW(15,JJJ)
IF(~JJ-l)
99S,997J998
C 9 9 7 ~'J RI TE ( 3 , 9 9 9 ) ( I BUF ( I r I ) , I I I =1 ,5 ) ,K , M, KEY
C 999 FOR·".'IAT( t t ,lOIS}
C 998 CONTINUE
GO TO(3,4),KK
C
C
C
C
HERE TO STORE BLOCK ON DISK
4 K3= (K2/L)*NREC+l
K3 =RECORD NO. OF THE FIRST LOGICAL RECORD IN THE BLOCK
17i
----------------------------------------------------------------------------
o
I.
-35PAGE 0
0
c
CONTAINING RECORD K
WR I T E ( N I K3 ) ( I BU F ( I ) , I =6", L L )
c
C
HERE TO READ NEW DISK BLOCK
(K1/L)*NREC+l
READ (N' K3) (I 8UF ( I ) ,1=6 ,LL)
IBUF(S)=Kl
REQUIRED DISK BLOCK IS IN FUFFER
2 KL =(K-1(1)*J+M+5
IF(KEY-l) 6,6,';
HERE FOR PUT OPERATIONS
5 IBUF(4)=2
6 RETURN
END
3 K3=
C
C
FEATURES SUPPORTED
ONE WORD INTEGERS
CORE REQUIREMENTS FOR DUSE
COMMON
0 VARIABLES
END OF
o
16
PROGRAM
216
CO~PILATION
179
&&1;_
AiiIMAiiUkAUk.UU: hM",ijJUAiUm.u::; ale ,I er
"
,; 4#,M4##.w
"
-36II FOR
*ONE WORD INTEGERS
*LIST SOURCE PROGRAM
FUNCTION MINO(I,J)
c
c
c
c
()
FUNCTION SUBPROGRAM TO CHOOSE THE SMALLEST VALUE
OF TWO INTEGERS.
210
211
212
290
IF( 1)210,211,211
IF(J)212,290,290
IF(J)299.212,212
IF(I-J)290,29Q,299
MINO = I
RETURN
299 MINO
=J
RETURN
END
FEATURES SUPPORTED
ONE WORD INTEGERS
CORE REQUIREMENTS FOR MINO
COMMON
0 VARIABLES
2
PROGRAM
48
END OF COMPILATION
~O
I~O
__________________________________________________
~
________ THE
o
I
-31-
o
II JOB
SWAIN SORT-MERGE LISTING
II
COMET AND IDEAL SUBROUTINES USED BY SORT-MERGE PROGRAM
II ASM
*
*LIST SOURCE PROGRAM
0000
0000
0001
0002
0003
0005
0006
0007
0008
0009
0008
OOOC
OOOD
OOOF
0010
0011
0012
0014
0015
0016
0017
0019
001A
c
001C
0010
OOlF
0021
0022
0024
0026
0027
0029
180D6517
0001
QCOMP BSS
0
0
01
0
0
0
0
00
0
0
00
0
0
6A2B
2829
66800000
0
1001
96800003
00
0
0
0
00
0
30
0
00
30
0
01
01
0
01
THIS IS QCOMP
*
*
STX
STS
LOX
LD
STO
LD
SLA
S
C205
D024
C200
1001
96800001
BOOC
C202
8003
72FF
74FF0032
a 70EF
002A 01 D400002C
002C 01 6600002E
002E 0 2000
002F 01 4C000031
0031 1 0032
0032 1 0033
0033 1 0034
0018
0034
QeOMP
1
2 SAVE&l
SAVES
12 QCOMP
X2 5
STOR1&1
X2 0
X 1
12 1
TWO
SCRAD
A
0024
C6800004
0022
7206
6A19
D4000002
C017
181D9042
D016
C4000002
181D9042
9011
4C20002A
74FF0031
ENT
INST
LOOP
STO
LD
STO
LD
SLA
CHCNT
X2 2
X 1
S
12 3
A
MDx
STX
STO
LD
CALL
STO
LD
CALL
S
asc
MDX
TWO
X2 6
2 RETRN&l
L 2
SCRAD
QGRAB
ATEMP
L 10002
QGRAB
ATEt-1P
L sTOR1,Z
L SCRAD,-l
~~DX
MDX
MDX
STOR1 STO
SAVE LDX
SAVES LDS
RETRN SSC
SCRAD DC
eHCNT De
ATEMP DC
TWO
EQU
END
12 4
X2-1
L
L
L2
X
L
eHCNT.-l
LOOP
*
*
0
*
*
*
*INST&l
COMPOOIO
COMP0020
COMP0030
COMP0040
SAVE XR2
COtv1P0050
SAVE STATUS
COMP0060
LOAD XR2 WITH CALL&1
COMP0070
LOAD Ace WITH RET PARA ADD COMP0080
STORE IN STORE PARA INST COMP0090
LOAD se AREA ADD TO ACC
SHIFT LEFT ONE BIT
SUB COLUMN NO. SC. CHAR
ADO TWO
STORE INTO SCRAD
LOAD CHARACTER COUNT
STORE INTO CHCNT
LOAD SECOND AREA ADDRESS
SHIFT LEFT ONE BIT
SUB COL NO FROM ACC
ADD TItJO
ADD 5 TO XR2
STORE XR2 IN RET INSTR.
STORE Ace INTO XR2
LOAD SCRAD TO ACC
GET FIELD ONE CHARACTER
STORE CH INTO ATEMP
LOAD XR2 TO Ace
GET FIELD TWO CHAR
SUBTRACT ATEMP FROM ACC
BRANCH IF NOT ZERO
DECREMENT SCRAD
DECREMENT XR2
DECREMENT CHCNT
NOT ZERO, GO TO LOOP
STORE IN RETURN VARIABLE
RESTORE XR2
RESTORE STATUS
RETURN
CO~.-1P0100
COMP0110
COMP0120
CO~.-1PO 130
COMP0140
COMP0150
COMP0160
COMP0170
COMPOl80
CO~1P0190
COMP0200
CO~'P02~O
COMP0220
COMP0230
COMP0240
COMP0250
COMP0260
COMP0270
COMPO.280
COt-1P0290
COMP0300
COMP0310
COMP0320
COMP033Q
COMP0340
COMPQ350
COMP0360
COMP0370
CO~1P038 0
COMP0390
COMP0400
COMP04l0
COMP0420
COMP0430
NO ERRORS IN ABOVE ASSEMBLY.
o
181
__________________________________________ THE
...."""'.""".=
.....=..... =............._.=
...""".. . """"'
,. . ,,,"""'
. . . ."""".. .-=.....=......."""
......=
.......=
......""'"
.....=......
~-~~-~----------"",,,,,,,,",,,,,,,,,,,,~-.,,,,,,,
....
==~,;;.;;.,"';;
..;;.;;.;.··==~~~_
w.ii,;;;;.,~;"
·3a.
II ASM
*LIST SOURCE PROGRAM
*
0000
18109042
0000
0001
0001 0 1881
0002 0 0002
0003 0 1091
0004 01 (4000006
0006 0 4802
0007 0 1808
0008 0 E002
0009 01 4C800000
OOOB 0 OOFF
oooe
*
THIS IS QGRAB
ENT
QGRAB ass
SRT
STO
SLT
LOAD L.D
Bse
SRA
AND
MASK
Bse
DC
QGRAB
1
X 1
LOAD&l
X 17
L *
e
X
I
END
8
MASK
QGRAB
100FF
GRABOO10
G'RABOO20
GRABOO30
GRAB0040
DIVIDE BY 2
GRABOOSO
STORE WORD ADD INTO LOAD
GRAB0060
PUT REMAINDER INTO CARRY
GRAB0070
LOAD WORK TO ACC
GRABOO80
GO TO MASK INST IF CARRY 0 GRABOO90
RIGHT JUSTIFY CHARACTER
GRAB0100
MASK EXTRA BITS
GRABOllO
RETURN
GRAB0120
GRAB0130
GRAB0140
0
NO ERRORS IN ABOVE ASSEMBLY.
_ _ _ _ _ _ _ • _ _ • _ _ _ ww _ _ _ •• _ . _ _ • _ _ .,~_ _ _ • _ _ _•_ _ _ _ _ _ _ _• _ _ _•_ _ _ _ _ _. _
THE
I
~
1,1
IJ
-39-
0
II ASM
*LIST
0000
0000
0001
0002
0004
0006
0008
OOOA
OOOB
OOOC
OOOD
OOOE
OOOF
0010
0011
0012
0014.
0015
0016
0018
OOlA
0
01
00
01
01
0
0
0
0
0
0
0
0
00
0
0
00
00
0
0018 0
OOlC 0
OOlE
0
o
04262105
0001
69.15
65800000
CD800000
4C280011
4C08000C
COlO
7006
1090
4820
70FB
COOC
7001
COOS
D5·800001
7102
690!
65000000
4COOOOOO
FFFF
0001
0000
IDEAL961
IDEAL962
oISGN
ENT
IDEAL963
*CALL DISGN(OBL INT,ISIGN)
IDEAL964
*WHERE ISIGN IS RETURNED -1.0,=1 DEPENDING UPON THE IDEAL965
*DBL INT BEING RESPECTIVLY NEGATIVE,ZERO,POSITIVE
IDEAL966
DISGN ass
RETURN CALL &- 1 HERE
IDEAL961
1
1 INDX1&1
STX
IDEAL968
LDX I 1 DISGN
IDEAL969
GET DOUBLE INTEGER
IDEAL97Q
LDD I 1 0
BRANCH IF NEG ACCUM
IDEAL971
BSC L NEG.Z6BSC L ZACC,(,
BRANCH IF ZERO ACCUM
IDEAL972
LD
POS
PONE
LOAD ISIGN WITH PLUS ONE
IDEAL973
STORE
IDEAL974
MDX
IDEAL975
ZACC SLT
16
SHIFT EXT TO Ace
SKIP IF ZERO
IDEAL976
ase
z
IDEAL977
MDX
POS
ZERO
IDEAL978
LD
IDEAL979
MDX
STORE
NEG
IDEAL980
Lo
NONE
IDEAL981
STORE STO 11 1
IDEAL982
1 2
MDX
LOAD BRANCH BACK INSTRUCT IDEAL983
STX
1 BACK&l
INDXl LDX Ll *-*
IDEAL984
RESTORE INDEX ONE
BACK esc L *-*
BRANCH BACK TO MAIN PROG
IDEAL985
-1
IDEAL986
NONE DC
IDEAL987
PONE DC
+1
IDEAL988
ZERO DC
0
IDEAl989
END
NO ERRORS IN ABOVE ASSEMBLY.
1~3
THE S
-40/1 FOR
IDEAl993
*ONE-WORD INTEGERS
IDEAl994
*LIST ALL
IDEAl995
FUNCTION INT(DBLIN)
IDEAL996
C INT IS A FUNCTION WH1CH TESTS THE SIGN OF A DBl INT WITH AN IF STATE IDEAL997
C -1 IF MINUS,Q IF 0,=1 IF POSITIVE IS RETURNED.
IDEAL998
C FUNCTION·INT CALLS DISGN
IDEAL999
C SAMPLE
IF(INT(DIONE))NEG STATEMENT,ZERO STATEMENT,POS STATEMENT NUMBIDEAlOO&
CALL DISGN(DBLIN,ISIGN)
IDEALOOA
INT=ISIGN
IDEALOOB
RETURN
IDEALOOC
END
IDEALOOD
o
VARIABLE ALLOCATIONS
INT =0000 ISIGN=0002
CALLED SUBPROGRAMS
DISGN
SUBIN
CORE REQUIREMENTS FOR INT
COMMON
0 VARIABLES
4
PROGRAM
18
END OF COMPILATION
o
THE
mnSM1!'Purmn .
-41-
'0
II ASM
*LIST
0000
22100000
,-,
0000
0001
0002
0003
0004
0006
0008
OOOA
OOOS
OOOC
aOOE
OOOF
0010
0012
0013
0015
0017
0018
Cl
0
0
0
01
00
00
0
0
00
0
0
00
0
00
00
0
0
0019 0
OOlA 0
001B 0
OOlC 0
0010 0
OOlE 0
OOlF 0
0020 0
0022
0001
690F
280F
2000
65800000
CD800001
9D800002
4801
7009
DD800000
7103
6904
65000000
2000
4COOOOOO
74000032
70FD
(006
3000
(OE5
9004
3000
lOAO
70ED
DEAF
01C4
SD
ENl
SUBROUTINE NAME
*CALL SO
(A.B.C) WHERE A=B-C
*DOUBLE INTEGERS HAVE STD PREC REAL VARIABLE NAMES
*IOEAL 1130 FORTRAN ERROR CODE IS IDEAF IN ACCUM.
*TO DISPLAY STATEMENT ALLOCATION AODR IN ERROR, HIT
*START,ACCUM HAS FORTRAN STATEMENT ALLOCATION ADOR.
*HIT START TO CONTINUE.OUTPUT IS SET TO ZERO.
SD
BSS
SUBROUTINE ENTRY POINT
1
1 INDXl&1
STX
STS
STATS
LDS
INITIALIZE OVERFLOW
0
LOX 11 SO
CALL&l ADDR IN INDEX ONE
LDD 11 1
LOAD B OF A S-C
11 2
SUBTRACT C OF A B-C
SD
BSC
0
SKIP IF OVERFLOW IS OFF
TOBIG
MDX
GO TO IDEAL ERROR DISPLAY
OUT
STD 11 0
MOVE TO A OF A=B-C
MDX
1 3
STX
1 8ACK&1
INDXl LDX Ll *-*
STATS LDS
0
BACK BSC L *-*
TOBIG MDX L 50,0
MDX
TOBIG
INTERRUPT SERVICE LOOP
LD
HDEAF
IDEAL FORTRAN ERROR CODE
WAIT
DISPLAY ERROR CODE IN ACC
LOAD ENTRY ADDR,5UB ORG+2,
LD
SO
HOIC4
AND DISPLAY STATEMENT
S
WAIT
ALLOCATION ADDR IN ACCUM.
SLT
32
CLEAR TO OUTPUT ZERO VALUE
MDX
OUT
HDEAF DC
/DEAF
IDEAL FORTRAN ERROR CODE
HOIC4 DC
IOlC4
DISKZ ORGIN +2
END
IDEAL769
IDEAL770
IDEAL771
IDEAL772
IDEAL773
IDEAL774
IDEAL77S
IDEAL776
IDEAL777
IDEAL778
IDEAL779
IDEAL780
IDEAL781
IDEAL782
IDEAL783
IDEAL7d4
IDEAL785
IDEAL786
IDEAL787
IDEAL788
IDEAL789
IDEAL790
IDEAL791
IDEAL792
IDEAL793
IDEAL794
IDEAL795
IDEAL796
IDEAL797
IDEAL798
IDEAL799
IDEAL800
IDEAL801
IDEAL802
!DEAL803
IDEAL804
NO ERRORS IN ABOV E ASSEMB LY.
o
THE
_;&M4i'M.Na,Miu,&4J. ¥MiSMI,JA NfT%\f\%1##i&MQf¥4A#. "
p
-42FOR
*ONE WORD INTEGERS
*LIST SOURCE PROGRAM
FUNC1ION LARGECIREC)
c THIS IS A SAMPLE SUBPROGRAM TO SATISFY THE CALL TO DUMMY FUNCTION
c IUSE IN SUBROUTINE MRGPH
c THE PURPOSE OF THIS FUNCTION SUBPROGRAM IS TO DETERMINE IF A
c RECORD IS TO BE CONTAINED IN A SORT. ACCORDING TO RULES ESTABLISHED
C BY THE USER. IN THIS EXAMPLE, A RECORD WILL BE OMITTED IF WORD 10
c IS GREATER THAN 5
II
c
c
IREe IS ARRAY CONTAINING RECORD
C FUNCTION RETURNS 1 IF RECORD IS TO BE INCLUDED. 2 IF OMITTED
OIMENSION IREC(32)
IF(IREC(lO)-5
) 1.1,2
1 LARGE=l
RETURN
2 LARGE=2
RETURN
END
FEATURES SUPPORTEO
ONE WORD INTEGERS
CORE REQUIREMENTS FOR LARGE
COMMON
0 VARIABLES
2
PROGRAM
32
END OF COMPILATION
0,,,
....-.•.-.-......
--.-~-.-.--.---.--
-.-.--,---
..
- - -,------------.-------.---------.--- ...---- ..
--~
Tf-!
!
I
I,
-4~-
II * COMET AND IDEAL SUBROUTINES USED BY DEMONSTRATION PROGRAM
II ASM
()
*LIST
0020
14044240
0000
14062240
009E
14109040
0050
14889040
C
ENT
MADI
SUBROUTINE NAME
*CALL MAOI (INPUT ARRAY,FLO START.FLD END,DBL INT)
*CALL MAOI MOVES AND CONVERTS Al TO DBL WD INTEGE~
*AN 11 ZONE OVER RT MOST DIGIT NEGATIVE CONVERSION
*ALL ZONES EXCEpT RT MOST ARE STRIPPED OfF.
*DOUBL~ INTEGERS HAVE STD PREC REAL VARIABLE NAMES
*
ENT
MASI
SUBROUTINE NAME
*CALL MASI (INPUT ARRAY,FLO START,FLD END.INTEGER)
*CALL MASI MOVES AND CONVERTS Al TO SIG WD INTEGER
*AN 11 ZONE OVER RT MOST DIGIT NEGATIVE CONVERSION
*ALL ZONES EXCEPT RT MOST ARE STRIPPED OFF.
*
MDIA
ENT
SUBROUTINE NAME
*CALl MDIA (OUT ARRAY,START.END,DBL INT,EDIT CNTRl)
*CALL MDIA MOVES AND CONVERTS DBl WD INT TO A1
*EO I T CNTRL O,INSERTS 11 ZONE AT FLO END,NO ZERO SU
*EDIT CNTRL -N,- AT ENO+l,ZERO SUPP TO N RT DIGITS
*USE INT VARI FOR NEG CNTRL TO SAVE 5 WDS PER CALL
*EDIT CNTRL +N,-AT END+2,N DECI PL,ZERO SUPP TO DEC
*EDIT CNTRL N GREATER THAN FIELD WIDTH,N IS +0 OR-O
*DOUBLE INTEGERS HAVE STD PREC REAL VARIABLE NAMES
*
MS IA
ENT
SUBROUTINE NAME
*CALL MSIA (OUT ARRAY.START,END,INTEGER,EDIT CNTRL)
*CALL MSIA MOVES AND CONVERTS SIG WD INT TO A1
*EDI T CNTRL O.INSERTS 11 ZONE AT FLO END,NO ZERO SU
*EDIT CNTRL -N,- AT END+l.ZERO SUPP TO N RT DIGITS
*USE INT VARI FOR NEG CNTRL TO SAVE 5 WDS PER CALL
*EDIT CNTRL +N,-AT END+2,N DECI PL,ZERO SUPP TO DEC
*EDIT CNTRL N GREATER THAN FIELD WIDTH,N IS +0 OR-O
**IDEAL
1130 FORTRAN ERROR CODE IS IDEAF IN ACCUM.
*TO DISPLAY STATEMENT ALLOCATION ADDR IN ERROR, HIT
*START,ACCUM HAS FORTRAN STATEMENT ALLOCATION ADDR.
*HIT START TO CONTINUE.OUTPUT IS S.ET TO ZERO.
0000
0001
0002
0003
0004
0006
0007
0008
0009
OOOA
OOOC
0
0
0
0
0
01
0
0
0
0
01
01
0
0
01
01
0
eOOE
oeOF
0010
0012
0014
0015 0
0016 a
0017 01
0019 0
0000'
CO FE
0071
406A
74FF010E
C054
D04E
1010
2000
04000090
C48000E8
E04B
1808
D4000092
C4000090
1002
807A
1001
4C010117
8078
OOlA 01 4C290117
*MAS I
OC
LO
STO
851
ASI
MDX
LD
STO
SLA
LDS
STO
LD
AND
SRA
5TO
LD
SLA
A
SLA
sse
A
BSC
*-*
MASI
ARGMT&l
SAVE
L BACK&l,-l
HD480
OUT
16
0
L WORK
I
FLOST
HOFOO
8
L DIGIT
L VORK
2
WORK
1
L TOBIG,o
DIGIT
l TOSIG,O&Z
SUBROUTINE ENTRY POINT
CALL&l ADDRESS
SET UP FOR ARGMT ADDR LOAD
SAVE XRS & GET PARAMETERS
ADJ RETURN ADDR FOR 4 PARA
STC I SINGLE INT INST.
AT OUT INSTRUCTION
CLEAR Ace
INITIALIZE OVERFLOW
LOAD REMAINING INTEGER
FIELO CHARA TO Ace
Al DIGIT MASK
SHIFT T~ BINARY DIGIT SIGL
HOLD INTEGER DIGIT
LOAD PARTIAL INTEGER
WORK TIMES FOUR
PLUS ONE IS FIVE
DOUBLE EQUALs TEN TIMES
TOO BIG IF OVERFLOW
ADD D.l G I T
If 7
TOB I G IF .OVERFLOW OR NEG
lOEA~OOl
IOEALOO2
IDEAlOO!
IDEALOO4
IDEA L.OOS
IOEAlOO6
IOEALOO7
IDEAlOOS
ID£ALOO9
IDEAlO~O
I DEAL01I
IDEAlOl2
IDEAlOl3
IDEALOl4
IDEALOlS
[DEAL016
IDE.AL017
IDEALOl8
IDEALOl9
IDEAL020
tDEAL02l
tOEA l,.022
IDEAl023
IDEAL024
IDEAL02S
IDEAL026
IDEAL027
IDEAL028
IDEAL029
IDEAL030
IDEAL031
IDEAL032
IDEAL033
IDEAL034
IDEAL035
IDEAL036
IDEAL037
IDEAL038
IDEAL039
IDEAL040
IJ)EAL041
IOEAL042
IDEAL043
IDEAL044
IDEAL045
IDEAL046
IDEAL047
IDEAL048
IDEAL049
IDEALOSO
IDEALOS1
IDEALOS2
IDEAL053
IDEAL.054
IDEALOSS
IDEAL056
IDEALOS7
IDEAL058
IDEAL059
IDEALOI~P
_44",
PAGE
OOlC
OOIE
0020
0021
0022
0024
0025
0026
0028
01 74FFOOE8
01 74FFOOE9
70E9
0
0 D06E
01 C48000EA
0 E077
0 F074
01 4(200028
0 1010
0029 0 9066
002A 0
7028
002B 0 C064
002C 0 7029
002D
002E
002F
0030
0031
0033
0034
0035
0036
0037
0
0
0
0
01
0
0
0000
COFE
D044
003F
00 /+0
0
0038
003A
003B
003(
003D
OC3E
*MADI
004 L.. 01 4(290117
00L+6 01 74FFOOE8
0058
OOSC
01 74FFOOE9
70EC
D844
0
01 C48000EA
0 E04D
F04A
0
01 4(200055
0 lOAO
983C
0
7001
0
0 C83A
01 D48000EC
01 4COO010A
oFOO
0
D480
0
D(80
0
005D
0
004F
0050
0052
0053
0054
0055
0056
0058
005.~
COSE C
005F 0
0060 0
0061 0
0000
COFE
D014
4000
1090
L
OUT
\'JO RK
~1DX
OUT
DC
LD
STO
*-*
MADI
MDX
LD
STO
SLT
LDS
STD
LD
AND
SRT
STD
LDD
SLT
AD
SLT
SSC
AD
BSC
~~DX
t~DX
:viDX
STD
lD
AND
EOR
SSC
SLT
SD
MDX
LDAD! LDD
OUT
STO
esc
HOFoa DC
HD480 DC
HDC8'J DC
*
MSIA
FLDND
HFOOO
HDOOO
LDASI,Z
16
WORK
BSI
ADI
FLDST,-l
FlD\e
2
IDEAL061
IDEAlO62
IDEAL063
IDEALC64
IDEAL065
IDEAL066
IDEAL061
IDEAl068
IDEAL069
IDEAL07Q
IDEAL071
IDEAL072
IDEAL073
IDEAL074
IDEAL075
I DEAL076
IDEAL071
IDEAl018
IDEAL079
IDEAL080
IDEAL081
IDEAL082
IDEAL083
IDEAL084
IDEAL085
IDEAL086
IDEAL081
IDEAL088
IDEAlO89
IDEAL090
IDEAL091
IDEAL092
IDEAlO93
IDEAL094
IDEAL095
IDEAL096
IDEAL097
IDEAL098
IDEAl099
IDEAL100
IDEAL10l
IDEAL102
IDEAL1Q3
IDEAL104
IDEAL105
I DEAL1C6
IDEALI07
IDEALloa
IDEAL109
IDEAL110
IDEAL1ll
IDEALl12
IDEALl13
IDEALl14
IDEALll5
IDEALl16
I DEALll 7
I DEALl18
THE
0
it)
'C:
...
,
I
I':
j~
11
f
•
;
-42PAGE
C';
"I
,.",..,
0062
0064
0066
0067
0068
006.9
006A
01 C48000EC
01 44280083
0
D029
0 C028
0 1890
0 A82F
0 0025
006B 0
1098
006C 0 404E
006D 0 70F9
006E
006F
0071
0073
0075
0076
0078
007A
0078
007C
0070
007E
0080
0081
0082
0083
0084
0
01
01
00
0
0000
6DOO0108
2COO010C
65000000
C100
00 95800002
01 D40000F3
0 8010
0
IJ 085 0
0086
0087
0088
0089
008A
008e
0080
008E
0
0
0
0
0
0
0
0
0
0099 0
009A 0
0098 0
009C 0
n09D 0
0
009E
009F
OOAO
OOAl
OOA2
00A4
OOA6
0
*
SAVE
DC
STX
STS
ARGMT LDX
LD
S
5TO
A
STO
LD
A
S
STO
A
S
STO
STO
(103
!..!)
D065
STO
LD
STO
MDX
STx
SLA
LDS
0
0002
0000
0000
0000
0000
000.0
0000
0001
OOOA
DOOO
DFFF
FOOO
F040·
0000
0 CO FE
0 DOD3
0 40CC
01 CC8000EC
01 442800B3
DOEE
0
I
L
STO
SLT
BSI
MDX
C104
D064
0 7105
01 6000010E
1010
0
0 2000
01 4C80006E
0090
0092 0
0093
0094
0095
0096
0097
0098
D06E
C100
0 SOlA
00 95800001
0 0067
0 8016
0 9067
0 D065
0 D066
0
SIA
LD
SSI
STO
LD
SRT
asc
*
BSS
DIGIT DC
DC
EVEN DC
DC
000
DC
DC
ONE
DC
TEN
DC
HDOOO DC
HDFFF DC
HFOOO DC
HF040 DC
\~ORK
*
MDIA
DC
LD
STO
BSI
LDD
BSI
STO
II
L
ARG4A
NEGt&Z
WORK
yJORK
16
TEN
WORK
24
TEST
SIA
LOAD INTEGER
BRANCH IF NEGATIVE
SAVE INTEGER
LOAD WORK INTEGER
SET UP DIVIDEND IN A&Q
GET LOW DIGIT TO EXT
SAVE NEW INTEGER QUOTIENT
SHIFT DIGIT TO Ace
TEST FOR NEG & FIELD WIDTH
GET NEXT DIGIT
*-*
RETURN AD DR HERE
SAVE INDEX ONE
SAVE STATUS
ARGUMENT ONE ADDR TO INDXl
ARRAY ADOR TO ACCUM
SUBTRACT FIELD END COLUMN
SAVE SIGN ADDR
ADD ONE
FIELD END ADDR
ARRAY ADDR TO ACCUM
ADD ONE
SUBTRACT FIELD START COLMN
FIELD STARTING ADDR
ADD ONE
SUBTRACT FIELD END ADDRESS
STORE FIELD WIDTH
FIELD WIDTH TO ZERO SUPP
ARG~ ADDR To ACCUM
SAVE ARGUMENT 4 ADDRESS
ARGUMENT 5 ADDR TO ACCUM
SAVE ARGUMENT 5 ADDRESS
ADJUST RETURN ADDRESS
LOAD BRANCH BACK INST
CLEAR ACC
SET OVERFLOW & CARRY OFF
RETURN TO SUBROUTINE
INDX1&1
STATS
L1 *-*
1 0
11 2
L SIGN&1
ONE
FLDND
1 0
ONE
11 1
FLDST
ONE
FLDND
FLDwD
ZSCNT
1-
~
ARG4A
1
4
ARG5A
1 5
L1 BACK&l
16
0
I SAVE
E 2
0
E
0
0
0
0
0
1
10
IDOOO
/DFFF
/FOOO
IF040
*-*
I
MOIA
ARGMT&l
SAVE
ARG4A
l
NEG ,&Z
EVEN&1
CONVERSION WORK AREA
DOUBLE WORD DIGIT AREA
SECOND WORD OF DIGIT AREA
EVEN WORD ONE
EVEN WORD TWO
ODD WORD ONE
ODD WORD TWO
ONE
11 ZONE MASK
NEGATIVE ZONE MASK.
ZONE ISOLATION MASK
DIGIT MASK FOR Al FORMAT
SUBROUTINE ENTRY POINT
CALl&1 ADDRESS
SET UP FOR ARGMT ADDR LOAD
SAVE XRS & GET PARAMETERS
LOAD DOUBLE WORD INTEGER
BRANCH IF NEGATIVE
SToRE EVEN WORD OF DBl INT
/",
3
IDEALl19
IDEAL120
IDEAL121
IDEAL122"
IDEAL123
IDEAL124
IOEAL125
IOEAL126
IDEAL127
IDEAL128
IDEAL129
IDEAL130
IDEAL131
IDEAL132
IDEAL133
IDEAL134
IDEAL135
IDEAL136
IDEAL137
IDEAL138
IDEALf39
IDEAL140
IDEA Ll:41
I DEA-Li42
IDEALi43
IDEAL144
IDEAL145
IDEAL146
IDEAL147
IDEAL148
IDEALi49
IDEAL150
IDEAL151
IDEAL152
IDEAL1S3
IDEAL154
IDEAL15S
IDEAL156
IDEAL157
IDEAL158
IDEALlS9
IDEAL160
I.DEAL161
IDEAL162
IDEAL163
IDEAL164
IDEAL165
IDEAL166
IDEAl167
IDEAL168
IDEAL169
IDEAL110
IDEAL171
IDEALl72
IOEAL113
IDEAL174
IDEAL115
IDEAL17&
IDEALliA?
,g(h,-"~
-46PAGE
aOA7
aOAS
oo·A9
OOAA
OOAS
aOAC
OOAD
OOAE
OQAF
OOBO
OOBl
0082
0
0
0
0
0
0
0
0
0
0
0
0
1090
OOEE
C8EA
A8EE
00E9
1090
88ES
A.8EA
00E1
1098
4009
70F6
0083
0084
00B5
00B6
00B7
00B8
0089
0
OOBB
OOBC
OOBD
008F
00(0
00(2
00C4
0 0000
0 E8EO
01 D48000EA
0 (02E
01 4(0800(0
01 C48000ED
01 4C20010F
0000
0808
0 COE2
0 D037
0
10AO
0 9807
01 4(8000S3
0
00(6 01 C48000EA
00C8 0 EOD2
00(9 01 Dl+8000EA
00C8 0 1010
ooee 0 D021
OOCD 01 740100EA
OO(F 01 7 t+FFOOE9
OODl 0 .'( 00 1
00D2 0 7002
01 4(800088
00D3
OODS 01 (L~8000ED
0007 01 !+C18010A
00D9 01 4C1OOOEO
OODS 0
OC)DC
OODE
OODF
OOEO
oaEl
800F
01 4C2800FC
0 Dooe
701C
0
DOOS
0
0 (009
OOE2 0
9006
OOE3 01 !J.C1OOOEF
OOE5 0
OOE6 0
OOE7 0
OOES 0
OOE9 0
(082
D002
700A
OOEA 0
0000
0000
0000
aOES
0000
a
. - - -...... ...-....-".-..---.- ......-.-.......
,
SLT
STO
LDD
16
000&1
EVEN
TEN
EVEN&l
16
ODD
TEN
000&1
24
TEST
OIA
SHIFT ODD WORD TO Ace
STORE ODD WORD OF INTEGER
EXTRACT NEXT DIGIT
GET LOW DIGIT TO EXT
SAVE NEW EVEN WORD
SHIFT DIGIT TO Ace
ADD ODD WORD
GET LOW DIGIT TO ExT
S~VE NEW ODD WORD
SHIFT DIGIT TO Ace
TEST FOR NEG & FIELD WIDTH
GET NEXT DIGIT
4
IDEAL178
IDEAL119
DIA
IDEAL180
0
IDEAL181
STO
IDEAL182
SLT
IDEAL183
AD
IDEAL1S4
D
IDEAL185
STO
IDEAL186
SLT
IDEAL187
SSI
IDEAL188
MDX
IDEAL1S9
IOEAL190
*
NEG
DC
ENTRY FOR NEG ROUTINE
IDEAL191
*-*
STD
WO.RK
STORE DOUBLE INTEGER
IDEAL192
ONE
LD
SET NEGATIVE SWITCH ON
IDEAL193
STO
NEGSW
SET NEGATIVE SWITCH ON
IDEAL194
SLT
32
CLEAR Ace & EXT
IDEAL195
WORK
SO
COMPLIMENT VALUE IN WORK
IDEAL196
NEG
8se
GO TO CONVERSION ROUTINE
IDEAL191
IDEAL198
*
TEST DC
TEST RETURN
IDEAL199
*-*
OR
HF040
MASK DIGIT FOR Al FORMAT
IDEAL200
STO
FLDNO
OUTPUT RECOVERED DIGIT
IDEAL201
LD
NEGSW
LOAD SWITCH CONDITION
IDEAL202
SSC L POS,&
BRANCH TO pas IF sW OFF
IDEAL203
LD
I ARG5A
EDIT CONTROL PARAMETER
IDEAl204
sse L MINUS,Z
BRANCH IF NOT 11 ZONE (ODE IDEAL205
* INSERT 11 ZONE OVER FIELD END DIG1T
IDEAL206
LD
I FLDND
GET END DIGIT OF FIELD
IDEAL207
AND
HDFFF
INSERT 11 lONE WITH DIGIT IDEAL208
STO
FlDND
PUT DIGIT WITH 11 ZONE
IDEAL209
Sw
16
SLA
CLEAR Ace
IDEAL210
NEGSyJ
STO
SET NEG SWITCH TO ZERO
IDEAL211
POS
MDX L FLDND,&1
THIS IS NOW NExT COLUMN
IDEAL212
rv1DX
L FLD\tJD ,-1
THESE COLUMNS ARE YET TO G IDEAL213
tv1DX
AGAIN
SKIP THIS IF FLDWD IS ZERO IDEAL214
t~DX
EDIT
IDEAL215
,~.GAIN SSC
TEST
GET NEXT DIGIT
IDEAL216
IDEAL217
*EDIT LD I ARG5A
EDIT CONTROL PARAMETER
IDEAL218
sse L INDX1,&OMIT SUPP ZEROS IF ZERO
IDEAL219
Bse L DEC,NO DECI POINT
IDEAL220
A
lseNT
CAL-C LEFT ZERO MAX COUNT
IDEAL221
sse L SLZ,&Z
N GT FlDwD,CNTRL IS -0
IDEAL222
STO
ZSCNT
MAX LEFT ZERO SUPPRESS
IDEAL223
MDX
SLZ
BRANCH TO SUPPRESS ZEROS
IDEA L224
DEe
FLDWD
STO
STORE WIDTH OF DECI FIELD IDEAL225
LD
ZSCNT
ENTIRE FIELD WIDTH
IDEAL226
r LDVJD
SUBTRACT DECI WIDTH
IDEAL227
S
SSC L NOR,~, N GT FLDWD,CNTRL IS &0
IDEAL22B
LD
ONE
LOAD ONE
IDEAL229
STO
FLDviD
SET FLDWD TO ONE FOR SIGN IIDEAL230
MDX
SIGN
BRANCH TO SIGN
IDEAL231
FLDST DC
0
FIELD START ADOR
IDEAL232
FLDWD DC
FIELD WIDTH
0
IDEAL233
FLDND DC
0
FIELD END AD DR
IDEAL234
lSCNT DC
0
ZERO SUPPRESS MAX COUNT
!DEAL235
/90
THE 5
_._--.
.....
-
----~.--,--------- --.-~----
-~.--,-
...
·r,
V
~
. . rl
C)
I...•
-47PAGE
o
OOEC 0
OOED 0
OO::E 0
0000
0000
0000
OOEF
OOFO
OOF2
OOF4
oaFS
00F6
00F7
00F9
OOFA
DOFB
740100E9
65000000
C100
DIFF
7101
74FFOOE9
70FA
COlA
0
01
00
0
0
0
01
0
0
OOFS 0
DIFF
oaFC
OOFE
0100
0102
0104
0105
0107
0109
OlOA
010C
010D
01
01
01
01
0
01
01
0
00
0
00
74FFOOEA
C48000EA
E400005A
4C20010A
COOF
D/+8000EA
74FFOOEB
70F2
65000000
2000
4COOOOOO
uj,ur
u
I...UUO
0110 01 D48000F3
0112 01 4COOOO(8
0114 0 4040
0115 0 4B40
0116 0 6040
0117
0119
OllA
011B
OllC
OllE
OllF
0120
0121
0123
0124
0126
00
0
0
0
01
0
0
0
01
0
0
74000032
70FD
(009
3000
C4000074
9004
3000
lOAO
4(000056
01C4
DEAF
ARG4A DC
ARG5A DC
N EGSW DC
o
o
o
5
ARGUMENT 4 ADDRESS
ARGUMENT 5 VALUE
IDEAL236
I)EAL237
IDEAL238
IDEAL239
** SHIFT RIGHT TO FIELD END PLUS TWO FOR DEcr POINT IDEAL240
NORM STo
ZSCNT
MAX LEFT ZERO COUNT
IDEAL241
MDX L FLDWD,1
ADJUST FOR SIGN POSITION
IDEAL242
LOAD SIGN ADDRESS
SIGN LDx Ll *-*
IDEAL243
NEXT LD
1 0
LOAD CHARACTER
IDEAL244
STO
1-1
SHIFT RIGHT ONE COLUMN
IDEAL245
MDX
1 &1
INCREMENT INDEX FOR NEXT
IDEAL246
MDX L FLDWD,-1
DECREMENT DECI WIDTH CNTR IDEAL241
I"1DX
NEXT
SKIP IF LAST DEcr POSITION IDEAL248
lD
H4B40
LOAD DECIMAL POlNT
IDEAL249
STO
1-1
INSERT DECIMEL POINT
IDEAL250
* SUPPRESS LEADING ZEROS IN OUTPUT FIELD PER COuNT IDEAL251
SLZ
MDX L FLDND,-1
THIS IS NEXT COLUMN RIGHT IDEAL252
LD
I FLDND
LOAD LEFT CHARACTER
IDEAL253
AND L HOFOO
ISOLATE DIGIT
IDEAL254
Bse L INDX1,Z
BRANCH QUT IF NOT ZERO
IDEAL255
LOAD BLANK TO REPLACE ZERO IDEAL256
LD
H4040
STO I FLDND
OUTPUT BLANK TO FIELD
IDEAL2,7
MDX L lSCNT,-l
MAX ZERO suPP SKIP EXIT
IDEAL258
MDX
SLZ
SKIP THIS IF MAX CNT ZERO IDEAL259
RESTORE INDEX REG ONE
IDEAL260
INDXl LDX Ll *-*
RESTORE STATUS
IDEAL261
STATS LDS
*-*
BRANCH BACK TO FORTRAN PRO IDEAL262
BACK SSC L *-*
IDEAL263
* INSERT MINuS SIGN AT FIELD END PLuS ONE
Mr,~~u5 L.u
MOU,+U
MiNuS SiGN
i uEAL26it-STO I SIGN&l
AT FIELD END PLUS ONE
IDEAL265
sse L SW
I DEAL266
H4040 DC
/4040
BLANK
IDEAL261
H4B40 DC
/4840
PERIOD
IDEAL268
H6040 DC
IDEAL269
/60 /+0
DASH
IDEAL270
*TOSIG MDX L 50.0
IDEAL271
MDX
IDEAL272
TOBIG
INTERRUPT SERVICE LOOP
IDEAL273
LD
HDEAF
IDEAL FORTRAN ERROR CODE
WAIT
DISPLAY ERROR CODE IN ACCU IDEAL274
IDEAL275
LD
L ARGMT&l
DISPLAY ENTRY ADDRESS
IDEAL276
H01C4
AND DISPLAY STATEMENT
S
WAIT
ALLOCATION AD DR IN ACCUM. IDEAL277
SLT
32
CLEAR TO OUTPUT ZERO INTEG IDEAL278
IDEAL279
Bse L OUT
HOIC4 DC
/01C4
IDEAL280
DISKZ ORGIN +2
HDEAF DC
/DEAF
IDEAL281
IDEAL FORTRAN ERROR CODE
IDEAL282
END
NO ERRORS IN ABOVE ASSEMBLY.
o
,_ _.__-L-!...I1/~ _ _ _ _ __
,_ _ _ THE S
____
~~~~~
=
. . ·-"""""·
. - ·. "".
_ _ _ _ _ _ _ _ _ _ _ __.._.._=.-...,.""'' ' ' '"' '-,---'' ' ' ",.'' '"' ' '......
,.,.,'' ' ',=
....,,....
'-·m""""··'.."""'·
......·=,-·. =---...-·"""',
......
i.....
··-·'~-·,-~--=-i--·-::;;;;.
..---~--~~_.o....-..~_
_ _~_
.48.
II ASM
*LIST SOURCE PROGRAM
0000
0000
0001
0002
0004
0005
0006
0008
0009
OOOA
OOOC
0000
OOOE
OOOF
0011
0012
0013
0014
0016
0
01
0
0
00
0
0
00
0
0
0
00
0
0
0
185C18A2
0001
6A20
66800000
C200
1001
96800001
801E
D01B
C6800004
0019
C202
1001
96800003
801S
7205
6AIO
Df f-OOOOO2
COOE
0017
181D9042
0019
18888925
0018
72FF
OOlC
74FF002S
OOlE
74FF0026
0020 0 70F'5
0021 01 66000023
0023 01 '+COOO025
0025 1 0026
0026 1 0027
0027 0 0002
0028
00
0
30
30
0
01
01
NO ERRORS
*
*
THIS· IS QPASS
ENT
QPASS BSS
STX
LDX
LD
SLA
S
A
STO
LD
STO
LD
SLA
S
A
MDX
STx
STO
LOOP LD
CALL
CALL
MDX
MDX
MDX
MDX
SAVE LDX
RETRN Bse
SCRAD DC
CHeNT DC
T\vO
DC
END
QPASS
1
2 SAVE&l
12 QPASS
X2 0
X 1
12 1
T\'JO
SCRAD
12 4
CHCNT
X2 '".,
X 1
12 :3
TWO
X2 5
2 RETRN&l
L 10002
SCRAD
QGRAB
QSHUV
X2 -1
L SCRAD,-l
L CHCNT,-l
LOOP
L2
L *
*
*
*2
PASSOO1O
PASSOO20
PASS0030
PASS0040
SAVE XR2
PASS0050
LOAD XR2 WITH CALL&l
PASS0060
LOAD SOURCE AREA TO Ace
PASSOO70
SHIFT LEFT ONE BIT
PASS0080
SUB COLUMN NO OF SC FIELD PASS0090
ADD TWO
PASS0100
STORE ACC INTO SCRAD
PASSOllO
LOAD CHARACTER COUNT TO AC PASS0120
STORE IN CHCNT
PASS0130
LOAD DEST FIELD ADD TO ACC PASS0140
SHIFT LEFT ONE BIT
PASS0150
suB DEST eH NO FROM ACC
PASS0160
ADD TWO
PASS0170
ADD 5 TO XR2
PASS0180
STORE IN RETURN INSTR
PASS0190
STORE ACC IN XR2
PASS0200
LOAD SCRAD TO Ace
PASS0210
GET SOURCE CHARACTER
PASS0220
PUT IN DESTINATION FIELD
PASS0230
PASS0240
DECREMENT XR2
DECREMENT SCRAD
PASS02SQ
PASS0260
DECREMENT CHeNT
NOT ZERO, GO TO LOOP
PASS0270
PASS0280
RESTORE XR2
PASS0290
RETURN
PASS0300
PASS0310
PASS0320
PASS0330
IN ABOVE ASSEMBLY.
____~___~__~.__,. . . ______.-_____-,_~--~--"-.-191.--_.
---------------
-~-
0
C
o
_____________113_
· -...............""'""-'""""-·--'o...........""""'-............."""'"'-_ _ _ _ ~IJIIIr~.~.t_:·4'.,..,.. ,~.."••. . . - ' .... 'L'.. ','~""
:"50-.- ..
_._--------_._--------------
II JOB T
II FOR
*SEI0
42192
.1
SWAIN CALL ME
o
*IOCSCTYPEWRITER.l132 PRINTER,DISK)
*ONE WORD INTEGERS
*LIST SOURCE PROGRAM
C DEMONSTRATE PMERG BY SORTING 1000 RECORDS OF 32 WORDS
C USE 5 SORT KEYS. OCCUPYING SWORDS
C MAJOR KEY IS WORD 6, INTEGER
C SECOND KEY IS WORDS 14 AND 13. TREAT AS DOUBLE wORD INTEGER
C
(SIGN AND HIGH-ORDER BITS IN WORD 14, LOW ORDE~ BITS WORD 13)
C THIRD KEY IS wORD 21, TREAT AS 2 ALPHABETIC CHARACTERS
C FOURTH KEY STARTS IN WORD 16, REAL
C
(SIGN AND HIGH-ORDER BITS IN WORD 16
C
LOW ORDER BITS AND ExPONENT IN WORD 15)
C MINOR KEY IS WORDS 30 THRU 32 TREAT AS 3 ALPHABETIC CHARACTERS,
C
STARTING FROM RIGHT CHARACTER OF WORD 30
EXTERNAL LARGE
C . LARGE IS NAME OF PROGRAM TO DETERMINE IF RECORD IS TO BE
C INCLUDED IN SORT
DIMENSION IREC(32)
DIMENSION KEYS(4,5),rWORK(lS5Q),I8UF(325)
DIMENSION AREC(3)
DIMENSION NAME(25),IOUT(11)
EQUIVALENCE(AREC( 1), IREC( 16}), (DOUB, IREC( 14»)
DEFINE FILE 1(101.320.U.l1)
DEFINE FILE 2(30.320,U,I2)
DEFINE FILE 3C30,320,U,I3)
DEFINE FILE 4(4.320,U,I4)
C FILE 1 IS FILE iO BE SORTED.
C FILES 2 AND 3 ARE WORK FILES
C FILE 4 IS OUTPUT. CONTAINS POINTERS TO INDICATE PROCESSING ORDER
C AS DETERMINED BY SORT KEY INFORMATION
C
C GENERATE 1010 RECURDS OF RANDO~-.1
C PMERG WILL EXAMINE LAST 1000 OF
C INCLuDED IN SORT. AND THEN SORT
CALL DOPEN(IBUF.lt32,10)
K=31525
L=899
IX=1000
IK=S
IB=10
IL=IB+IX-l
C CREATE ALPHABETIC CHARACTERS IN
C ARE INCLUDED
NAME(1)=16459
NAME ( 2·) =19790
NAME(3)=20571
NAME(4)=2364S
NAME(S)=24673
NAME(6)=27517
NAME(7)=32449
NAME(S)=-1567"1
NAME(9)=-lS163
NAME (10) =-146 /+9
NAME(11)=-1413S
NAME(12)=-11822
NAME(13)=-11308
NAME(14)=-10794
Nl'MBERS
THESE TO SELECT THOSE TO BE
THOSE INCLUDED
NAME. ALL VALID PRINTER CHARACTERS
01,.
TH
-51PAGE 02
C"
"
l.'j
C
C
C
C
C
C
C
C
NAME(15)=-10280
NAME(16)=-9758
NAME(17)=-7196
NAME(18)=-6682
NAME(19)=-6168
NAME(20)=-5648
NAME(21)=-3598
NAME(22)=-3084
NAME(23)=-2S70
NAME(24)=-20S6
NAME(25)=-1792
STORE RANDOM INTEGER IN RANGE -9 TO +9 IN WORDS 1-14
DO 1 I=l,IL
DO 2 J=1,14
K=K*L
2 IREC(J)=K/3277
STORE RANDOM REAL NUMBER IN RANGE -10 TO +10 IN WORDS 15-20
DO 5 J=1,3
K=K*L
5 AREC(J)=FLOAT(K)/3277.
STORE RANDOM ALPHABETIC CHARACTER IN WORDS 21-32
DO 6 J=1,24
K=K*L
M=K/1311+25
6 CALL QPASS(NAME,M,IREC(21),J,1)
1 CALL DPUT(IBUF.I,IREC)
CALL DCLOS(IBUF)
CREATE KEYS TABLE, TO INDICATE LOCATION AND TYPE OF SORT KEYS
MAJOR KEY IS INTEGER. WORD 6
KEYS(1.1)=1
KEYS(2,1)=6
SECOND KEY IS HIGH PRECISION INTEGER. HIGH-ORDER pART IN WORD 14
SORT INTO DESCENDING ORDER BY THIS KEY
KEYS(1.2)=-2
KEYS(2.2)=14
THIRD KEY IS ALPHABETIC. FIRST CHARACTER WORD 21 COL 1, 2 CHARACTERS
KEYS(1,3)=3
KEYS(2.3)=21
KEYS(3,3)=1
KEYS(t+~3)=2
C
C
FOURTH KEY IS REAL, STARTS IN WORD 16
KEYS(1,4)=4
KEYS(2,4)=16
FIFTH KEY IS ALPHABETIC. FIRST CHtRACTER
KEYS(1,S)=3
KEYS(2,S)=30
KEYS(3,5)=2
KEYS(·4,S)=3
WOR~
30 COL 2. 3 CHARACTERS
C
o
C PERFORM SORT-MERGE
WRITE(1,102)
102 FORMAT(tSTART SORT-MERGE')
CALL PMERG(1,32,IB,IX ,KEYS,IK,IWORK,1850,IBUF,lO.2,3.4,
$LARGE,KOUNT)
WRITE(lrl03)KOUNT
103 FORMATC'END sORT-MERGE'.I6.' RECORDS INCLUDEDt)
c
C LIST FIRST 10 RECORDS IN SORTED FILE
---'-ff.____.____ ~ __._-
I~
-52PAGE 03
CALL
CALL
DO 4
CALL
DOPEN(IBUF,1,32,lO)
DOPEN(IWORK,4,1,320)
1=1,10
DGET(IWORK'!'IPT)
WRITE(:3,lOl)IPT
101 FORMAT(' RECORD NO,',IS)
CALL DGET(IBUF,IPT,IREC>
CALL QPASS(NAME,1,IOUT,21,1)
CALL MDIA(IOUT,l,lO,DOUB,-l)
4 WRITE(3,104)(IREC(J),J=1,12),IOUTtAREC,(IREC(J),J=21,32)
104 FORMAT(· ',1213,11Al,lX,3FS.4,lX,12A2)
STOP
END
FEATURES SUPPORTED
ONE WORD INTEGERS
IOCS
CORE REQUIREMENTS FOR
COMMON
VARIABLES
°
~ND
2324
PROGRAM
674
OF COMPILATION
,_ _ _ _ THE
-53II XEQ
L 1
*LOCAL,SRTPH,MRGPH
o
FILES ALLOCATION
1 0432
2 0497
3 0465
4 04D3
0065
OOlE
OOlE
0004
STORAGE ALLOCATION
R 47
0010 (HEX) WORDS AVAILABLE
CALL TRANSFER VECTOR
DISGN
QCOMP
INT
SD
QSHUV
QGRAB
MINO
IPTSK
DUSE
MDIA
DGET
LARGE
PMERG
DCLOS
DPUT
1C66
1C32
lBBA
1B94
1B7C
1B70
lA78
1940
1855
178C
16B1
168D
1638
15C6
1576
QD,A.S5
!'5~6
DOPEN
MRGPH
SRTPH
1498
1D25
1CF8
LOCAL
LOCAL
LIBF TRANSFER VECTOR
FSUB
SDRED
SDCOM
SDIX
SDWRT
FARC
NORM
EBCTB
GETAD
IFIX
PAUSE
SUBIN
STOP
SIOAF
SIOAI
SIOIX
SIOI
SCOMP
S\'-JRT
o
FSTOX
FDIV
1BCC
OD9E
ODC2
OD94
ODEC
1B4E
1B24
1B21
1AEO
1AB4
lAA8
1820
1814
OFEF
OFFC
1071
OFF7
OFDF
OFD6
145A
14F4
lif7
._ _ _
TH~
o
o
o
SUBSC
luE6
14C8
FSTO
FLO
147A
FLOA'
PRN-T t.
WRTYZ
145E
1398
1360
SFIO
lOAD
SOFIa
DISKZ
O·OF1
OOF4
SYSTEM ROUTINES
lF6F
ILS02
~
1C90
OB3C (HEX) IS THE EXECUTION ADOR.
FLIPR
t
~
RECORD NO. 764
4
-8 -2
2 -3
R~CORD NO.
~75
1 -1 -4 -8 4
RECORD NO. 861
-7 -5 9 -6 -8
RECORD NO. 349
-4
3 -2 -1 4
RECORD NO. 773
... e -7 6 -4 -8
RECORD NO. 860
4 o -5 7
0
RECORD NO. 632
-5 o -3 0 7
RECORD NO.
74
... 7 o -2 -5 -5
RECORD NO. 533
0-.
""
o -5 -1 -4 -6
589830
4.8669 -4.3433 -4.8663 XHLNP2LYVQ.8BAW'MGCPQMI)
0
9
0
589810
3.2807
9 -8
0
3
458760
3.5761 -4.8565 -6.3280 .D14+)K=02,J+&*U-PB*H.X8
-9
0
"9
1 -1
9
9.5688
2.8877 CODHXQ*(C=7'(SFTZ5(B&-.
-9 -6
9
-9 -7
4 -7
1
8
7
458753
1.701'
-9 -4
4
6
4
'3
2
458753
5.5974 -7.5419 -0.6350 7=,VNZ8K=+K60ZZC$R7ZIA.M
-9 -5 -8
o
-4
o -7
458750
8.7339 -7.7415 -0.0619
0
5
393219
-7.4418
9.3521
8.0967 K6&ILI1C*,4RM16141TYHZYW
-5 -5
393216
4.5422
3.7445
6.5855 N.+FF/MT.BAC-FL2LLE2UX(2
-1.5309
-9 -2
6
4 -1
.. 9
0
6
o
3
o "9 0 o -9 -6 -7
RECORD NO. 965
..4 -6
1 -9 5 6
0 9
8
1 -8
0
393207
3
3 -Q -6
327684
0
9.5115 -3.zu:;2 LW(N( (/H.=I$( 22R LZFbK*
3.5706 -9.7946
8,8004 -7.9325
v+w
,~57LVOFWD/E*6-62.11
'4V$G)UK56SCC3SZ/UY*+~+K
8.1925 Y'RJDQ1M7TYC"L U2$Z6CS2
o
START
ENn
SOPT-t~F.P.~F.
SORT-MFR~F
811 RECCRDS INCLUDED
/1'1
II JOB
II FOR
*5E10
42192
.2
SWAIN
*ONE WORO INTEGERS
*LIST SOURCE PROGRAM
FUNCTION INTAK(IREC)
C THIS FUNCTION CAUSES ALL RECORDS TO BE INCLUOEO IN A SORT
INTAK-l
RETURN
END
FEATURES SUPPORTED
ONE WORD INTEGERS
CORE REQUIREMENTS FOR INTAK
COMMON
0 VARIABLES
END OF COMPILATION
2
PROGRAM
14
o
o
SESSION NUMRER T.2.5
SPEAKERS
PANEL ON PROGRAMMER EVALUATION
A.S. GLOSTER, II, OAK RIDGE ASSOC. UNIV.
S.A. LYNCH, tl.S. REDUCTION CO.
DR. L.H. RAKER, PIONEER HI-BRED CORN en.
OR. PAUL HER""ITZ, I.B.M.
~Ol
o
PROFESSIONAL PROGRAMMERS AND ANALYSTS:
PROBLEMS IN PERFORMANCE EVALUATION *
By: Arthur S. Gloster II
Oak Ridge Associated Universities is a nonprofit corporation engaged in research and educational activities
located in Oak Ridge, Tennessee.
The primary function of the
corporation's data processing center is to apply electronic
data processing techniques, where feasible, to research and
administrative projects.
The center contains
of which 13 are analyst/programmers.
27
employees
The center is divided
into three groups: 1) scientific applications consisting
of 6 personnel and a group leader, 2) commercial applications
consisting of 5 personnel and a group leader, and 3) operations
section consisting of a Gupervisor and 10 other employees.
Oak Ridge Associated Universities has on site an IBM 1800
disk/tape system which is utilized by approximately 40% of
the programming personnel.
Approximately 60% of the personnel
use IBM 360-50, 360-75 and CDC equipment located in the area.
The analysis and programming function comprise a significant
portion of the operating costs of the ORAU data processing
center.
After discussion with personnel from other installations,
we found that there are no reliable standards by which costs
can be calculated in advance, schedules established, and the
performance of personnel evaluated,.
use~
to
establishschedul~s
Although the methods we
and costs are subjective and
arbitrary, they in no way approach the accuracy of the methods
c
-2-
o
developed in the hardware area.
For example, IBM has estab-
lished rates for EAM equipment and has even produced a slide
rule to use for estimating job times.
A means of evaluating the professional analyst/programmer's
job and a means of evaluating his effectiveness is highly
desirable but rarely accomplished.
Such means would be helpful
predictors in determining the staff needed for a particular
application or a data processing installation.
We have found
that records of intangibles, such as the time for the application analysis and problem definition, flow charting, coding,
debugging, checkout, and finally documentation would have to
be maintained continuously to have a base for predicting
analyst/programmer costs and for personnel evaluation.
In
predicting costs of computer programs and evaluation, we would
like to be able to have a magic number representing the proper
number of analysts or programmers that could be applied to a
given situation and for our center as a whole, but we found
this to be impractical because we were measuring intangibles
by SUbjective means.
We found that attempting to save expenses
by minimizing or restricting the availability of professional
personnel caused equipment to be used ineffectively.
The more
the programmer is annoyed with accounting for his time and the
more detailed the account for nonproductive time, the less
apt he will be in cooperating in a program that keeps up with
all of the various functions he performs.
o
~03
-3-
Programming personnel at ORAU are engaged in numerous
o
types of jobs; therefore, standards of work evaluation
could not effectively reflect the variety of tasks they
encounter.
One programmer may be responsible for coding X
number of instructions with relatively small amounts of logic
development, while another programmer may be responsible
for extensive logic development with relatively few instructions.
Thus, it becomes difficult to measure the amount
of work required on each program, and the total work effort
performed by a programmer cannot be assessed in standards
of comparison with another programmer.
At ORAU, we believe that the group leader of either the
scientific section or the commercial section, depending on
the particular area, should look at the problem in advance
and then meet with the data processing manager to establish
reasonable target dates for each of the previously-mentioned
phases on the basis of the nature of the problem and on their
past experience.
The manager or group leader must have a
detailed knowledge of the problem under study because it is
his responsibility to prepare the cost estimate-'and schedule.
He must keep up with the allocation of funds and judge progress
of the application.
After giving several methods of job
measurement trial, we have found that there is no substitute
for experience in the area of predicting costs, measuring
work, and evaluating programming personnel.
Programmers
:C'.
,-
-4-
o
respect a supervisor who gives them a job and can tell them
what performance is expected.
The supervisor is also respected
by his staff if he remembers that good supervision is the
least supervision needed to get the job done.
,.'\.
'-"
o
-5-
o
*From Oak Ridge Associated Universities, Oak Ridge, Tennessee,
under contract with the United States Atomic Energy Commission.
Programmer Evaluation in U. S. Reduction Co.
o
5. A.
LpC/A
u.
S. Reduction Co. is a producer of secondary aluminum alloys. Annual
sales are over $60, 000, 000. There are approximately 900 employees.
Corporate headquarters and one plant are located in East Chicago, Indiana.
There 'are four other plants in four other states.
Data-processing-services employees number six. In addition to a manager
and assistant manager, there are two programmers, one operator and one
key-puncher. In addition to these individuals, two other company employees
are closely related; one functions ae ~ senior systems analyst, the second
as a special project researcher who does his own prograrnrning. Finally,
use has been made of two outside progra:m.rn.ers on a contractor hourly basis.
U. S. Reduction Co. uses two IBM 1130 Systems and some time on System
360/Models 20 and 30. An expanded 1130 configuration has been ordered.
Methods of evaluation of programmer-effectiveness are based on subjective
criteria, augmented by certain measurable phenomena. These include:
a.
Demonstrated dedication to task and cooperation during its fulfillment.
b.
Speed of accomplishment of as signed tasks.
c.
Feedback from users serviced by applications programmed by the
programmer in question.
d.
Feedback from IBM personnel dealing with the prograrnrner on
technical matters.
e.
Extent and clarity of program documentation and general eas e of
implementation.
f.
Infrequency of undefined error halts, accuracy of output, and speed
of job execution.
g.
Personal evaluation by the programmer during the semi-annual
salary review interview.
h.
In all of the above full recourse must be made to comparisons with
the evaluator's personal experience as a programmer himself and
with other programmers he has known.
o
207
o
Management Installation Division
PROGRAMMER EVALUATION
by
Dr. Paul S. Herwitz
IBM Corporation
Old Orchard Road
Armonk, New York 10504
(914) 765 - 4543
' \
1
0'
".
Thursday, 10:30 A.M.
Text, 12 pages
c· , .
li~
Ii
o
c
PROGRAMMER EVALUATION
Whenever the subject of Programmer Evaluation comes up in
conversation, at meetings, and as the topic for panel discussion, I
have the feeling that the participants are looking for easy answers.
As an IBMer who has been in this business for a long time, I'll have
to disappoint you if you expect me to give you an algorithm for
judging the work that programmers do; because I don't believe one
ever will exist. It's easy to understand that in a profession that's only
20 years old, in which the technology is in its infancy, in which
40% or 50% of the people working in it probably have been working
in it for about two years, in which an operating system which was
a new concept in 1962 is old -hat in 1967, in which an experienced
professional is ancient if he's been in the business ten years, and
in which a manager is highly experienced if he has five years of
manag-ement -- it's easy to understand that most of us don't really
know what to expect of the practitioners of this peculiar craft
called programming-. I'm disturbed by the idea some people seem
to have that there may be a magic formula and some sort of
mechanical procedure that produces an evaluation automatically
without anyone's effort and judgment going into the process. You
know, programmers are a clever lot; and if someone comes up
with such a scheme, they'll find a way to beat it.
The evaluation of the work of professionals is just not an easy task.
It's especially difficult in a profession that's as young as ours. To
do the job properly requires a lot of experience, a lot of skill, -and especially it requires a lot of energy on the part of the manager.
It also requires an understanding of a company's philosophy -- a
knowledge of what the company expects to accomplish by evaluating
its people, and a knowledge of its standards. I think, also, that
there is an opportunity now to improve the case for professionalism
in programming. In an activity that's so new, -- which has been
entered by so many people so recently - - in which the demand for
practitioners far exceeds the supply -- it comes as no surprise that
advancement has been a matter of expediency more often than we
care to admit. This only works to the detriment of the profession.
We expect the number of programmers and analysts to double in the
next four or five years, and we expect it to triple in ten. It's time
to take stock. The opportunity won't last forever.
2.09
_ , ...
............
"'.~.,_
~~~~
_ _ _ _ _ _ _ _ _ _ _ _ _ _......,...,.................
' """',
.. t_"""_""".='."
..
._=
. . . ."_.,=
. . ........
. ·=--.....
-
""'""····=··,·~"""'··
"C-, .... ", ....., ... ---,.-
As I have said, the evaluation of the professional's performance on
the job is really very difficult. It's a continuing problem '- - one with
, no pat solution, only guidelines. Now one of the guidelines I'm going
, to speak about later on is knowing wha't the job is. And this applies to
the first line manager as well. This manager must know that
evaluation is his job and he, himself, will be measured on how well
it's done. This is something that's not well understood by new,
young managers, and I think the reason lies partly in today's
pressure environment and partly in the way we choose managers.
Many of us say that the progr,ammer lives in a state of constant
crisis. That may be true, but I don't think it's a characteristic that
applies only to the life a programmer' leads. I think if you will look
at any production-oriented business -- and I mean by that any
activity in which you produce a product -- I think you'll find the same
kind of perpetual crisis, or continuing pressure situation throughout
the business world and industry today. Requirements are always
changing, technology is constantly changing, and invention is made
according to schedule. This means management is generally facing
new technical problems every day. Combining this with the
explosive growth that has taken place in the programming field, it's
easy to understand the tendency to choose as a new manager the
person in the group who has been technically most competent.
Unfortunately, this doesn't mean that our new manager is the person
most likely to succeed as a manager of people. Technical
competence is, of course, a most necessary attribute of a manager,
but I believe that in the long run the person who is really well
oriented to the people who work for him will produce much more than
the purely technically-oriented manager. The Situation, of course,
tends to be self-perpetuating. The new manager probably learns
more from his manager than from anyone else. But the second line
manager, himself, is so busy with the problems of meeting
schedules and changing requirements that he is not likely to take the
time to tutor his new subordinate in the personnel aspects of the job.
Moreover, the second line manager probably lived in the same
environment when he first became a manager and learned very little
of this aspect of the job from his seco:p.d line manager. Now this is a
pretty vicious cycle, but it can be broken. Breaking it depends upon
management in general realizing the value of doing the job properly.
Just plain common sense tells us that there can be a real payoff.
In order to become a really top notch manager, a person must have
considerable insight which usually comes only after very extensive
experience -- job insight and insight into people -- what they're
-2-
210
o
o
,. II
capable of and what motivates them. In lieu of this extensive
experience, the problem is -- how can we do a job which in some
sense can be considered competent? Well, one thing we can do is
attempt to give our managers better education in the form of
instruction by his manager, and education in terms of formal
training. I think most companies probably do a pretty good job of
training new managers in various administrative aspects of their job,
such as a knowledge of company policies, what procedures to follow,
when it's necessary to requisition supplies, how to fill out the forms
necessary to give a man a raise, and so on. But although most
companies have forms to be used and to be filled out before and after
an appraisal and counseling interview is held, since the evaluation
process is so judgmental in nature, it's very difficult to tell anyone
how to make the decisions necessary so that he can fill in the forms.
My firm belief is that you learn this only by trying, by reviewing
your decisions with your own manager and by trying again. And, of
course, it's always extremely useful for a new first line manager to
talk it over with other first line managers who have had more
experience than he. At best, the only specific instructions that can be
given can only be in the form of guidelines; and we will discuss a
number of guidelines later.
Let's take a look, now, at the reasons for evaluating our people. In
IBM if we ask why we evaluate programmers, the answer is easy to
understand. And I think our reasons are pretty much the same as
yours. Evaluation really has a triple payoff - - for the company, for
the manager, and for the programmers.
0
~_)
If we look at the process from the company's viewpoint there are
three important reasons. First, we want to improve our products
and our services by finding an optimal deployment of our human
resources. Second, we want to be certain that the salaries we pay
are equable throughout all of our programming activities. And third,
something that isn't often considered, IBM's attitude toward its
programmers and toward the profession in general is bound to have a
major impact on all programmers. So we feel that learning to do the
job properly is a challenge and an opportunity to help the profeSSion.
If we look at evaluation from the manager's viewpoint, the reasons
are similar but more specific. First of all, the manager has to find
the people most capable of doing the job he needs done. This means
he must find the best way to utilize the people he has -- and then he
must measure them against his requirements. If he is really a good
-3-
I,:
---_._.. _ - - - - - - " , - - -
o
manager (and he, himself, will be evaluated on this attribute), he
will look beyond the immediate requirements of his own project and
will decide whether or not some of his people can contribute more
somewhere else. Secondly, since the responsibility for salary
recommendations lies with the first line manager, and since
increases are given for merit, the manager must decide who merits
a raise, what the amount should b,e, and when the increase should
be given. Finally, the manager must continually strive to improve
the performance of the people who work for him. This means he
must evaluate, he must instruct, and he must encourage his people.
This is his opportunity to contribute to the profession.
The third part of the evaluation payoff comes when the programmer is
told openly and honestly what the manager thinks of the programmer's
work. All of us want answers to a lot of questions -- our morale
depends on them. How do I stand?' How good was the job I did?
Where am I going? Will the jobs I do continue to be interesting -- to
be challenging? Even if the answers to the questions are not what the
programmer hopes for, his morale will be high if there is no
subterfuge on the manager's part. It's amazing how hard people will
work if criticism is constructive. And it's surprising how much they
will learn if they are given timely instruction.
c
:0
Now let's try to define some guidelines that will help us realize the
benefits of a good evaluation program. The guidelines I shall
propose are really a large number of questions that have to be asked
and answered on a continuing basis. They will take quite a bit of the
manager's time, skill, and energy to answer in any responsible
fashion. And this is one reason why I constantly preach that a
manager shouldn't have more than about eight programmers reporting
to him - - six would be better.
We are trying to evaluate programmers in terms of their present
performance and their potential for future assignments. We'll start
with present performance.
First and foremost, I think, it's necessary to know what the job is
that you expect the man to do. It's just as important that the man
know the job you expect him to do. Once there is agreement on what
the job is, I think half the battle is won. Moreover, one of the
factors that one must use in the evaluation is the level of detail
necessary in describing- the job; and this must take into account the
programmer's experience and pOSition level.
-4-
.----~-----..,...........-~"""--"-~~~~~--------.------------------...........- -.....'''-.-..."""'."""""".,.,...=
.....
Now it must be possible to measure how well the job is done. In
order to do this it follows that some quantifying" j ob parameters"
must be agreed upon. By job parameters I mean things such as the
length of time necessary to complete the program, some indication
of the size of the program, some indication of the performance
expected of the program when complete, and some statement of the
functional capability expected of the program. Remember, of course,
that it may not be possible to state such parameters when the program
is first conceived; and so the job assignment may very well be to
study the proposed program and arrive at such parameters.
Now depending upon the length of time needed to complete tre
program, it may be necessary to state some parameters that will act
as checkpoints. After all, it's easy enough to keep track of a
programmer who's working on a l-week assignment, but it may be
extremely difficult to keep track of the work a programmer does when
his assignment takes him 2 or 3 months or more. In the latter case,
it may be necessary to block out the job into sub-jobs, each of which
has some sort of a time checkpoint and appropriate functional
specifications, and so on. Checkpoints generally call for a review.
This, of course, can be a very iniormal thing or it can be quite a
formal undertaking if several programmers are involved in projects
which culminate in a common checkpoint. But these checkpoints are
necessary not only to see that the proj ect progresses properly but
also in terms of constant evaluation of the people doing the work.
Once the job is complete, of course, the other half of the battle
begins; and that is the assessment of whether or not the job was
completed as expected. It's not difficult to know whether a person
completed the job according to the required schedule or not; but if he
didn't, what are the circumstances surroundip.g this? Was he
permitted to do the job without interference? Were the specifications
changed during the middle of the job? Does this program interface
with other programs that changed and, therefore, required this
program to be changed?
Again, it's easy to measure the size of the finished program against
the agreed-upon size. But if the program is too large, we have to
ask whether the circumstances are extenuating or not. Was the
projected size based upon knowledge of similar routines which had
been done before for other machines or other circumstances?
Should there have been a direct carry-over from the previous
program to this? If this program is really a brand new program,
was the amount of work underestimated? Were additional functional
capabilities included in the program beyond the functional
-5-
o
o
specifications? Again, is the size of the finished program
considerably less than projected when the job was first described?
If so, is anything missing? Or was it just easier than we expected?
Or did we discover new programming techniques? Similarly, was
the job completed ahead of time? If so, was the estimate overconservative? And if so was that the manager's fault or the
programmer's fault?
How we measure the performance of the program is probably one of
the most difficult questions we could ask, particularly if the program
has any magnitude. But at least we can compare the performance to
other programs with similar functional specifications if such
programs exist. And the likelihood is that they do. Does this
program perform better or worse than the old program? In either
case, why?
"
~,I
Finally, in regard to functional specifications, does the completed
program really have all the functional capabilities specified? If not,
why is some capability missing? Was the feasibility of such a
capability not proved? Or was it something not understood by the
programmer in the first place? Again, if there is more functional
capability in the program than asked for, what did we pay for this
additional capability? Did we get it free in terms of schedule, size or
program, and performance? Probably not. If not, was the price
worth paying? One begins to see that these are not easy questions to
answer. And, therefore, it should be obvious that when I said
evaluation was a difficult and time-consuming job, I really meant it.
Once the initial assessment in terms of job parameters is made, we
have to undertake an additional evaluation which is probably even
more difficult. What is the quality of the work? We have already
addressed most of the external aspects of quality; that is, whether or
not the program met its size, performance, and functional
specifications. One other external aspect is the correctness of the
program. How well has this program been debugged? Of course, if
the program came in on time, this means that it was debugged on
time. Whether it came in on time or not, a good manager will ask
what are the nature of the errors that did turn up? What was the
debugging plan? Did the debugging plan take all reasonable
contingencies into consideration? And I guess, finally, would a
better debugging plan have produced the desired results in a shorter
period of time?
o
-6-
~/S
o
()
2.1'
o
Of course, to answer these questions, the manager must be pretty
familiar with the written code. As a matter of fact, I don't know any
way to judge the internal quality of a program without becoming
thoroughly familiar with the program itself. This, of course, implies,
too, that a manager can only make a quality judgment based on his
own experience. Another factor regarding the internal quality of a
program is the quality of the remarks that document the program. Are
these remarks succinct and to the point? Do they appear where they
are needed and not where they are redundant? Another factor one
must look at is does the program use proven techniques, or is it
innovative? I think that nine times out of ten the use of proven
techniques will get the desired result with much less heartache tmn
will innovation. This is not to say there is no place for innovation
because there is. Again, though, it's a judgmental factor as to when
one should be innovative. Finally, is the program elegant? Now how
do you answer that one? Well, I suppose you look for simplicity; you
look for coding techniques that save time, or produce more function,
and that are well documented; and you ask whether the job is
workmanlike. Are all the steps there, do they lead logically from
one to another to arrive at the desired result in the shortest possible
time? An elegant program to me is one which is characterized by
simplicity, lucidity, novelty, and good workmanship.
Before I talk about potential for future assignments, I would like to
pause a moment and look back over what I've said. At this point the
manager has evaluated a programmer's work with respect to a
particular job assignment; that is, he has determined whether the
job came in on time, that it met functional, performance, and size
specifications, and that it was of a certain quality. So now the
manager should be able to answer two questions. First of all -does the programmer meet the requirements for this job? If the
answer is yes, then the second question must be - - can he do a
better job and is there something more challenging for him to do?
If the answer to the first question was no, if the Job requirements
were not met, then the manager must ask -- why not? And here he
must make a real value judgment -- was the programmer over his
head, or was he not working up to his capabilities? If he was over
his head, can he be taught? Or should he go elsewhere? And
probably the most difficult question - - if he was not working up to
his capabilities, what are the chances that he will next time? If the
chances are poor, the manager bas a real problem. If there were
extenuating personal circumstances or job-related circumstances,
what can the manager do to help alleviate these? Finally, if there
-7-
2/7
are personal problems, the manager had best stay out. If there
are job-oriented problems, then the manager has a responsibility to
try to correct the situation.
Now let's look at the potential for the future. I'm going to assume
that the man being evaluated has performed satisfactorily in his
present assignment and that the assignment has been challenging.
Otherwise, he has either not met the job requirements or has not
shown any potential for promotion or added responsIbility. But if
his last performance has been acceptable, then we have to look to the
future. First, I guess, we would ask how broad is the man's
experience? And here we get into the old argument of whether a
person should be a generalist or a specialist. The argument, of
course, is that the specialist can produce a given program without
having to reinvent the wheel. My concern is that our people have
specialized without having been exposed to the broad basic
fundamentals of programming as a whole. This, of course, is due to
the nature of the rapid growth in programming, but it doesn't really
excuse the situation. There is no question that this is an age of
specialization, but specialization usually comes after extensive
schooling in the broad fundamentals of the discipline. Specialization
without a broad base to start from is really a pretty unhealthy
situation. Here again is one of the opportunities that we have to
improve the profeSSion as a whole. And I really don't think we will
be in good shape until the colleges and universities have developed
well-rounded curricula in the computing sciences. Whenever
possible, I think it's best to attempt to give our programmer some
breadth of experience.
Directly related with the question of how broad this man's
experience is is the question -- how up to date is his technical
knowledge? It's very easy for a man who has spent one or two years
in a high pressure proj ect to forget some of the experiences he had
in previous proj ects. This, again, I think is due to the lack of a
well-rounded initial education. If a man has a good base to begin
with, he is less likely to completely forget what he has learned in an
area he is not working in currently than if he has had only brief
experiences in several different areas. Several people have
suggested that at least part of the evaluation of a programmer be
based upon a well-defined skills inventory. I suspect this is a
rather difficult thing to construct, but it would be interesting to try.
Remember, however, that this is not a panacea. It is only one of the
tools that a manager must draw upon in evaluating one of his people.
-8-
o
o
f!
o
c
Another characteristic we must try to evaluate is the man's
teclmical foresight. How well does he foresee problems? Is he
able to avoid the problems, or bring them to his management's
attention early enough so that appropriate steps may be taken? Or
does he evade the problems? Another attribute to consider is the
man's problem-solving ability. This is always considered a key
characteristic in the search for new programming talent. But how
do we evaluate it? That's a very difficult thing to do without seeing
him in operation. I expect, though, that watching the man in action
gives a pretty good clue as to his ability to face problems by himself.
And I think that's what we really mean. Does he solve the problem
himself, or does he need to come to his management for help? A
corollary question, does he come to his management for help when
this is really necessary - - when the problem is outside of his range
of experience? This leads directly to another characteristic, his
judgment and decision-making ability. When the man is faced with
alternatives, how often does he make the right choice? Can he in
some fashion characterize the probabilities of success for the various
alternatives? Is his judgment good - - does he make the right choice
more often than not?
The next area that we have to evaluate is the man's self-knowledge
and his self-improvement activities. Does he understand his own
limitations and his own talents? What is he doing to remove his
limitations? Does he have his own plan, or must he be led?
Remember that one way to improve one's capabilities and one's
experiences is to constantly seek out new and broadening assignments.
Does he read the literature of the field? Does he go to seminars?
Does he go to meetings and bring back new ideas? Just how wide are
his horizons?
One of my pet peeves is the man who cannot communicate. Over and
over again I've run into people who cannot write clearly, cannot speak
clearly, and it's questionable as to how clearly they think. Although
many programmers tend to be reserved and introverted, this is not
true of all of them by any means. You might think if a programming
shop or a programming specialty is a one-man operation in a
particular installation, that it doesn't matter w·hether he can
communicate or not. Well, nothing could be farther from the truth.
If a programmer is to be successful in working in a team, he has got
to be able to communicate both orally and in writing. Particularly, if
a man is going to be an analyst of any sort, he should be able to
communicate clearly with the people he has to deal with in attempting
-9-
,
,.. .%# 4£ •au UOSAZ 4
ai,;: Wi TJAd ;.,,::,maSUM_a. .u
Mil_a:
=
to offer his services. So many times I've heard people who are not
computer-oriented say that they would love to use the facilities of a
computing center but they can't communicate with the analysts or the
programmers. I think it's up to us to see that we, as programmers,
are able to provide the right kind of information and to show the
uninitiated how our services might be used. And it does no good to
do this with a chip on the shoulder. So one of the major questions
that I would ask is how well does the programmer communicate with
his peers, his management, and with novices and the uninitiated? I
cant emphasize strongly enough that I think the man who can't
communicate) although he may be a fine programmer, is a man
whose usefulness is highly specialized.
Now let's consider what the programmer has done for the profession.
Has he had an opportunity to make a contribution? What has he done
within his company? Has he made any technical advances? Does he
make suggestions within his own area -- timely suggestions for new
programs? Does he see ways to simplify what's being done? Does
he attend seminars? Does he write papers and make presentations?
Does he go to computer meetings? What kind of contacts has he
made? Has he met people whom he can stimulate intellectually and
who can stinlulate him intellectually? If he makes presentations,
does he have something to say? How well is he respected within his
own area, within his company, and by others in the profession outside
of his company? I grant you that with so many people who are so new
in the profession, not a lot have been able to make contributions.
But I think again that it's our responsibility to stimulate our people
into contributing. And not on a haphazard basis either. They need
the guidaD ce of those of us who have had experi~nce.
Now I want to discuss one other item that needs consideration both
from the standpoint 0: present performance and potential for future
assignments. This isonewhich most of us don't like to talk about too
much, but it' s someth~ng- that's extremely important in this day and
age. That's the question of cost control. Primarily, we have to ask
how well our programmer has utilized machine time. This, of
course, tie.s in with his success in debugging. But it's of utmost
importance today because it's a major contributing factor to the cost
of production of programs. The main question is could the man have
used less machine time to debug his program? Three other cost
factors are the use of supplies, the use of travel, and the use of
overtime. Although the question of overtime is more likely
-10-
o
1
o
0
something that a manager will be judged on, unless a project has
been directed to go into overtime, a person's ability to do his work
during regular hours is one that should be looked at. I guess if
overtime is necessary in a particular instance, we have to go back
to the question of whether or not the job was completed on time. Of
course, there are always extenuating circumstances such as illness,
absence for other personal reasons, etc.
o
I I m certain that all of you can think of many other questions that one
might ask in the evaluation of a programmer. I don't claim to be all
inclusive - - II m just trying to suggest the general areas that I think
are important to evaluate. Of course, remember also that most of
the questions must be answered in light of the position a man holds.
We ask much less of the newcomer than we do of the oldtimer. We
ask considerably less of the junior programmer than we do of the
senior programmer. As a matter of fact, this implies that position
descriptions exist and that they're well written, that they're understood, and that they're generally applicable to the work being done.
Remember also that good position descriptions will address many of
the areas in which I have asked questions and hopefully will give
some guidance as to what is expected of the person in the particular
job. A good position description can be one of the manager's valuable
aids for evaluation. Good position descriptions also should address
promotability. I think there should be well defined criteria for
promotion with indications as to what criteria may be sidestepped and
under what circumstances. I'm certain most of you have seen cases
where a programmer was not promoted in one department, was then
transferred, and shortly thereafter promoted in another department.
At the very least this suggests that the two areas were not looking at
the same criteria for promotion. This is an extremely unfortunate
situation and one that's hard to handle.
I guess I can't tell you what kind of answers you need to these
questions to be able to say that a programmer is outstanding or merely
meets requirements or always exceeds requirements or does a very
poor job. Standards are going to vary from company to company. I
think most of us will find that we don't even have standards within
our own company. So we really have a job here in providing the
standards applicable to our own situation. There is a lot of work
involved; but, as I indicated earlier, we think it' s a challenge and an
opportunity.
-11-
C·\
"7
I.
Let's remember another thing, too. The manager who looks
constantly at all t.he attributes and all the characteristics of the
orogrammer he is evaluating and who instructs him and who
criticizes him constructively all the time is- a paragon that just
doesn't exist. If most of us can look at many of these questions
on a relatively constant basis and can do our best to instruct and
criticize constructively, then I think that we're dOing quite a
remarkable job. And I guess in closing I finally want to remind
you that when you try to answer many of these questions and when
you try to set up your own standards, don't forget the man who is
given an impossible task. We need a hero medal -- second class.
-12-
o
__
aWiM&
o
ADDRESS BY
MR. G.
W. WOERNER, JR.
VI CE-PR ES I DE "!T & REG I ONAl MANAGER
MIDWESTERN REGION
I.B.M. CORP.
c
c
COMMON Meeting
September 7, 1967
Cincinnati, Ohio
0.
1
Address by G .. W. Woerner, Jr. - Vice-President & Regional Manager - 1fIV.lR
Thanks for inviting me to your meeting and to the very fine luncheon this noon.
I must say it has been a real pleasure and I have been tremendously impressed
by the pace of your program and the am' :tious goals which you have set for
yourselves here in Cincinnati.
I am sure you must know how greatly IBM values the working relationship which
has evolved over the years between our people and your organization. There is
absolutely no substitute for the direct insight into YOlli'" individual and collective
problems which can be gained"through this relationship.
With the complexity of hardware, sofbNare and applications increasing at such
astounding rates, we need your advice, counsel and constructive criticism
badly. It provides us with a fundamental reference point from which to develop
our plans and focus our resources on areas that best meet your needs.
InCidentally, you may find it interesting that two particular subjects are coming
through loud and clear.
The first concerns the general area of our method of providing you with manuals
and "technical information.
The second question concerns the various types and levels of supporting education
required frorn IBM.
I can assure you that we are addressing these two particular points with
exceptional urgency as a direct result of the emphasiS" which this organization
has placed on them.
As I observe the many contributions your organization has made to its members,
I am espeCially impressed with the quality and quantity of the Type N Library.
Today it contains over 700 programs written by members of the COMMON
Organization. Your massive contribution to the 1620 Library has been of
significant value to your own members, to IBM and to the state of the art as a
whole. This experience should indicate the value of now turning your efforts
to the equally' productive Models 20, 1130, and 1800 to build a similar library
for your own use of these systems.
o
There are an awful lot of people these days sitting back and watching us in the
computing industry. In fact, it has become a popular pastime to comment
upon the effects of computing and automation on our society.
&$.1$,=,1#1\;4*1" $I .4iW;i4MmxwilAUZUac
&£bIh
COMMON }J1eeting Address
G. W. Woerner, Jr.
Page 2
We, as IBM'ers, are especially interested in these commentaries. And I, for
one, have been particularly concerned about one dimension of the industry
which has been getting a lot of attention lately and I would like to spend the
remaining time I have with you this noon talking about it.
It has to do with the brQad conclusion reached by a number of contemporary
writers that automation and, more specifically, the computer constitute a major
threat to the individual in our society. These writers point out specific instances
where automation has unfavorably impacted the individual and then conclude
categorically that these instances constitute a trend at. the end of which is the
comolete extinction of the individual in tomorrow1s society. And, that we are
destined to live in a faceless world with standardized values, regimented
routines of life and mechanical modes of thought. More often than not, the
computer is used as a symbol of that threat.
I guess the whole thought stream got started as early as the mid r30'swith
Aldous Huxley's "Brave New 'Norld". He pictured a technological SOCiety
living in doped-up bliss under the watchful eye of a powerful tyrant.
George Orwell depicted a similar grim scene of life under the ever-present
electronic eye and ear of Big Brother.
Today, writers such as Galbraith and Vance Packard are further soothsayers
of the impending doom. Packard, in his book liThe Naked Societytt, says,
"In the Western World today, swirling forces ·are causing whole populations
willy-nilly to change their attitudes, ideals and behavior patterns". One of
these forces he describes as "the electronic eyes, ears and memories".
Supplementing these more renown writers are self-appointed investigators
who use the public media to broadcast sensational exposes of how man today
is losing the battle with automation.
Most often they take some form of the theme of a man's personal identity being
replaced by a number of holes in a punched card.
They usually conclude that these mysterious forces are somehow thrusting men
pell mell toward a conformative life of grim, mushy blahdom overseen by the
sterile surveillance of the master machine.
Before we rush to take a position for or against these observations, let's pause
to define what it really is that we're afraid of.
-sts te-l-l us t--n'
. or~ us.'na~
...a.- ~UllltJl;;::J..Ll..Ll1.J
-. - -- --,,!--- .w::::;c;u
___ \,v v'.H.~","".L""U
",..~'"'"t....'; .... h h;C'
PSyChOiog
.
1
at eacn
........
personal worth, and that a key part of that need is identity as an individual.
..3
"',..
0..1
"'"rt71"'\
..., . . . . . .
o
COMMON Meeting Address
G. W. Woerner, Jr.
Page 3
I would define individuality as the desire of each' man in our society to direct his
own destiny. To set himself apart from the rest of society as a person who has
unique tastes, thoughts and goals. He needs to have the prerogative of living as
he wants and developing himself toward ends which he has defined. He doesnrt
want society or governments or employers or any other group (much less machines)
to plan and manage his life for him. He wants the freedom of choice which allows
him to select the kind of food, clothing, shelter, education, employment and
leisure which suits him and his family best. The uniqueness of this pattern which
he chooses identifies him as an individual and establishes his identity.
What we :re afraid of then is that our society is becoming so organized that the
above choices will not be ours, but will be provided according to some kind of
plan. Welre afraid that those in seats of power are developing some plan based
upon statistical models and nor ms describing what each of us should want and
need.
We're afraid of the Similarity of our housing, the food in the supermarkets, the
clo~hes on the racks, the cars on the lots, the standardization of our educations -and the impersonal treatment by service organizations.
In short, we Ire afraid because the patterns we see developing across our society
appear to be a challenge to our compelling' need for individuality and personal
identity.
I must join in these observations and agree that many of these trends are, in
fact, in motion. I canrt agree, however, with the conclusion that automation,
or computers, have been the cause of these trends nor do they constitute the
current threat.
Computers and their use are a result rather than a cause. A result of the very
same basic force which is responsible for the threat to our individuality. That
force stems from the historic insistence on the part of the "have nots" to join
the "haves II. Whether they be the poor or the emerg'ing nation reaching for food
and shelter, or the middle class stretching for higher living standards, they
represent a constantly growing demand for a higher standard of living for more
people.
Man's success in meeting these denlands, many times with the help of machines
but sometimes without, has resulted in a society geared to volume rather than to
individual needs.
o
4
~,
Now, whether we look at consumer products, human rights, education or personal
services, this demand and response goes through three stages.
:
T.:
,ItA4I••t£
4t2tA4US;ZJAgMWauuaa============w
A4A
I~
I
COMMON Meeting Address
G. W. Woerner, Jr.
Page 4
In the first stage -- the benefit is only available to a select portion of the
population. They are the "haves". They have an exclusive, if you will, on the
benefit -- and it uniquely distinguishes them from their fellows.
o
Stage two is a natural follow-on -- the "have nots" demand the benefit and
society's inqenuity to supply it is motivated. The result is that more and more
"have nots" become Ifhaves". But in dOing so, the benefit loses a big ingredient -its exclusiveness and thus, its .individuality to the possessor.
The third stage then is, or should be at least, an attempt to stylize or expand
the benefit into unique subs?ts so that even after mass production and distribution,
the benef~t regains what it lost in stage two --------- an appeal to the individual.
Let me illustrate with some examples -"
In the early 1900's every automobile was designed specially for the buyer and
then was hand made. This was the first stage. You can observe several things
about it --'
'
1.
Only a few could be produced.
2.
Only a very small percent of the population could afford them.
3.
Ownership of custom autornobiles was certainly a prestigious
and individualistic thing.
Within a few years, however, pressure developed from those who did not have
this luxury and we enter stage two -- as a response to the demand ...••. Henry
Ford's ingenuity created the technique of mass production and he cranked out
huge q1.;lantities or one modeL ••.. the Model itA" at a price that more people
could afford,and in the color they wanted .••.. as long as it was black.
At this point, Ford satisfied the demand of the masses but he lost something in
the process. He sacrificed the idea of a "custom" car designed and built
speciiically for an individual. The customer no longer had a car which
represented his individual tastes and distinguishe d him from his neighbors.
All had black Model itA IS".
I
,I
But the story doesn't end here -- rather it moves to a third stage and this is
the key point. Today we can select an automobile from an impressive array
of models, styles, color variations, and features which suit our individual
tastes. In fact, that myriad of options and combinations, is so great that every
Single car producea by Ford alone this year cuuiu hd.V~ 'beel-,L diff6i"6ri,t i:rGm ~ll
others without going outside the standard features.
o
------~~--------------~~~~~~~--~--
,I
COMMON Meeting Address
G. W. Woerner, Jr.
Page 5
What brought on stage three? What accomplished this return.to the custom car?
A demand for individuality responded toby more sophisticated techniques of
automation, aided in no small way, by the computer.
Let's take another example to illustrate the point .••.• education.
Stage one ....• in the good old days -- the right to education was enjoyed only by
the few, primarily royalty -- in fact, as late as 1900, only 94, 000 people in the
United States graduated from high school.
Today we consider education to be the birth right of every man, woman, and
child -- not just royalty.
The ~esult is the phenomenon realized in stage two - - a high school education is
a standard for the majority of our population and over a half million students
are receiving college degrees each year.
However, companion with this has been the depersonalized multiversity with
campuses swarming with 20 to 30, 000 students.
Overcrowded classrooms - where an instructor addresses a
hall with upwards of 2 to 300 students •....
and
Standardized curricula - where the pace of instruction and the
content is geared to the average rather than to the individual.
We readily see that in our attempt to satiate the demand, we've already sacrificed
the small classroom environment, the tailored pace of instruction, the personal
guidance and touch of the teacher, and to a great degr.ee, the ability of the
individual student to pursue courses of study which are to his particular interest.
Interestingly.enough, in this example one can't even associate the move to the
second stage with automation. For we accomplishe d both the response to the
demand of the many and the loss of individuality without it.
0
,
,
".
'
Stage three is still in its tnfancy. Significant developments are already on the
scene in the form of computer aided instruction. A system is now available
and in use which allows a student with his own visual display terminal to learn
with the computer in a conversational mode .••.. the process is personalized
to the extent that the material is presented and dealt with at the learning rate
of the student. The human touch is preserved by the teacher who, upon being
COMMON Meeting Address
G. W. Woerner, Jr.
Page 6
relieved of the mechanics of presenting volumes of material, has time to devote
herself to that student who is having problems or needs counseling o
These techniques show promise of moving education solidly into stage three. A
stage which satisfies the demand for higher education for more and more people,
and yet preserves the integrity of the individuals.
I myself have' explored many other examples and I know each of you could come
up with many' of your own. But these two suffice to illustrate my premise. Let
me review ....
0
First, there are today, trends in our society which are a continuing threat to
each of us as we pursue our personal identity and individuality.
Second, these trends do not stem from the use of automation or the compute r,
but rather are a by-product of man's organized response to the steadily increasing
demand of the many for the benefits enjoyed by the few.
And finally, the computer has demonstrated its capability to contribute to the
development of the last stage of these trends ...•. the pursuit of our individuality.
The advance of technology and automation cannot, and should not, be stopped -for it is not the adversary of the individual - - but the ally.
,';'"-".
)\.,..,"/
So I would say to you as members of this organization o • • • • as key participants
in the field of automatioI.L .•. _. we must be prepared to proceed with confidence -for there is no shame in progress -- our efforts are not a betrayal of ourselves,
but rather an expansion of our opportunity_
And this point must be made to our critics.
In fact, you might want to think about adding this one to the objectives of your
group.
I'd like to thank you again for the opportunity of joining with you today and wish
all of you a very successful conclusion to your meeting.
\,V
.~"
!230
o
SESSION NUMBER
T.3.1.
SPEAKERS
DAN FULLAN (IBM) - PRESENTATION ON DOS III •
J. ALVAREZ & W. SELSMEYER - QUESTION & ANSWER SESSION ON DOS.
MODERATED BY D.R. Me IL~AIN
DISCUSSION
DOS III (ACTUALLY A NEW RELEASE OF DOS II, NOT FORMALLY NAMED
DOS III) - SEE ATTACHED WRITE-UP. THIS IS SCHEDULED FOR RELEASE
IN S~ALL PART 11/67 & COMPLETE 4/68. QUESTION & ANSWER SESSION
FOLLOWED.
"' "
0
'.'
.'I
!f
~31
o
SYSTEMj360
DISK OPERATING SYSTEM
IMPROVEMENTS
...,./
/~
\\.
f(--"""
-.IY
'.,
~~.----.-----
.. -----.--------
-----_._._---
.~-
,... 1
Scheduled Improvements to DOS
I would like to spend the next few minutes or so discussing some significant
improvements to the Disk Ope~ating System. The.se improvements which
will be available with subsequent ,releases of DOS include Additional Device
Support, namely the 2314, Additional Features in the Supervisor to improve
performance, for example, seek separation, Increased Capabilities, such
as expanded muitiprogramming capabilities, and'a number of related features
to simplify operations, for example, label handling. Let's go over these
now in a little more detail.
.
2314 Support
The 2314 direc~ access storage facility has received wide acceptance in the
mtermediate systems marketplace. The programming support provided for
this device reflects this acceptance. Full system support will be provided
for the 2314. This includes system residence, that is, residence for all IBM
components currently supplied in DOS/360, except Autotest, systems input
and output on the 2314, full systems libraries support, full data management
support including QTAM, Sequential Access, Direct Access, Index Sequential
Access, and Device"Independent Access. All current DASD utility functions
will include support for the 2314. Additional support in the form of new
special purpose utility programs will be provided for both the 2314 and the 2311.
These utilities include a Volume Table Of Contents Display, an Initialize Disk
Program, an Assign Alternate Track Program, a program to copy Disk to Disk,
and programs to Copy and Restore Disk with Tape and Cards. Full language
support is provided for both compilation and object time support. This includes
the Assembler, COBOL, Basic FORTRAN and RPG. We intend to provide PL/I
and Sort/Merge support. Information will be provided at a later date. All support
will take advantage of the increased capacity of the 2314 as well as the speed of
the unit itself.
Simplified Label Handling
We have modified the labeling procedures in DOS to provide for a simpler
more efficient operating environment. These procedures will involve the use
of fewer cards and label cards that are reusable. 'rne formats have been improved and more room has been provided for standard labels. In addition,
standard labels can now be used for Index Sequential and Direct Access files.
Two type s of standard labels will be provided. Standard labels will be available
to programs operating in all three partitions while partition standard labels will
be available to programs in the specified partition only.\A total of six tracks for
user labels and partition standard labels are available. 'On the 2311 four
additional tracks and on the 2314 fourteen additional tracWs will be provided for
standard labels. We have then the possibility of ~abels appearing in one of three
parts of the label cylinder. It may appear as a user label associated with the
specified partition, as a standard label associated with that partition, or as
a standard label available to all partitions.
\
o
Ii
The increased capacity for standard labels and the ability to use standard
labels for direct access and index sequential files should greatly reduce
the number of label cards and handling required in an operating environment.
The number of cards required for label handling has been reduced. We
have a DSKL card which replaces the VOL and DLAB cards required
today. The parameters for this card may be expressed in a variable
format, some of the parameters have been made optional. Date need no
longer be expressed as absolute creation and expiration dates, a retention period may be substttuted. This will enable the DSKL card to be
reusable, that is, once label cards have been set up for a production
run, the same cards may be used each time the run is executed.
o
A new EXTNT card has been provided. In this card the upper and lower
limits of the extents do not have to be expressed in absolute form. You
may specify relative track, that is, the sequential number of tracks
relative to zero where the extent is to begin and the number of tracks
the extent is to contain. On input files the information that is supplied
by the extent cards can also be taken from the label itself so that in many
cases an extent card will not be required.
Data sets on tape are defined by means of a new TLAB control card which
replaces the current VOL and TPLAB cards. Here again a retention
period may be specified instead of an expiration date. When a volume
label is not found on a labeled output file, it will automatically be created
based upon a serial number supplied by the operator. A multi-volume
file may be opened at other than the first volume.
To summarize the label processing capabilities of DOS, the number of
cards required for disk labels has been reduced from three to two and in
some cases only a single card will be required. A single card will now
suffice for tape labels instead of the two cards previously required. In
addition, since retention periods can be specified these cards are reusable, that is, new label cards need not be punched on a daily basis.
The expansion of the standard label facilities both for the number of
standard labels that can be stored and the types of files that can use
standard labels provides for an even further reduction in the label cards
required in an operating environment. I am sure that those of you involved
in day to day operations will appreciate the reduced burden on operations
personnel made possible by these improvements.
Seek Separation
The seek separation feature was designed to improve the performance of
systems running under DOS. This feature, provided by a supervisor
option, enables the supervisor to separate a seek from its associated
read or write so that the seek can be separately scheduled. This means
that multiple seeks can be issued to devices on a channel and the reads
(;'
-3-
o
and writes scheduled as the seeks have been completed. As this is a supervisor function, it will automatically apply to programs written at any language
level and operating in all three partitions. The benefits derived from this
feature will increase the larger the number of devices on a channel. The
implementation of this feature is such that when a seek has been issued to a
device the arm cannot again be shifted until the I/O operation that initiated
the seek has been completed. In other words arm stealing has been prevente d.
Following the issuance of a seek the channel is available for scheduling other
I/O operations. In a multiprogramming environment this feature is particularly
important where the different partitions have a mix of input/output requests for
a single channel with multiple direct access devices.
The seek separation feature is specified at Systems Generation time and will
add approximately 200 byte$ to the size of the supervisor plus 4 bytes per
direct access device.
Full Track Add
c'
We h0-ve improved the efficiency of the index sequential access method when
adding records to the file by utilizing a full track add feature. This feature
enables the user to provide room in core for up to as many physical records
as can be contained on one full track. When a new record is added to an index
sequential file that record is inserted in sequence and records that have a higher
key are shifted to make room for it. The highest record on that track will then
be placed in me overflow area. Without me core data feature when a logical record
is added to a block each block on that track must be read, the records internally
shifted, and then written back to disk. When the core data feature is specified,
two or more physical records up to a full track will be read at one time. The
records shifted internally and men written out togemer. For example, when room
in core is provided for a full track, it will take 4 1/2 revolutions or less to add
a record to that track. While without me feature, 3 revolutions will be required
for each physical record on that track. Let us suppose a blocking factor of
4 physical records per track. When adding records to mat track using the core
data feature for a full track add, 4 1/2 revolutions are required compared to
12 revolutions without the feature or approximately a 300% improvement. To
utilize mis feature, an I/O size parameter must be specified in the DTFIS to
reserve room in core for the additional data area. Core data equals yes must
be specified in the index sequential module itself. The additional core requirements for this feature are 16 bytes per DTF and 1 75 bytes in the index sequential
module itself.
Cylinde r Index
o
We will also provide the capability to maintain all or part of the cylinder index
in core. This will have a significant impact in efficiency during random retrieval,
since the seek and read time of the index itself can be completely eliminated.
This facility is achieved by specifying the name and size of the area reserved for
the cylinder index in the DTF and specifying core index equals yes in the index
sequential module. Where there is not sufficient room in core for the entire index
it is possible to make only a portion of it core resident.
- 4The core requirements to use the cylinder index in core feature are minimal. An additional 24 bytes are required for the DTF plus 150 bytes in
the index sequential module. This of course is in addition to the core
required for the index itself.
Both the full track add and the cylinder index feature are available only
in the Assembler language.
Multiprogramming
In the area of multiprogramming we have made a number of improvements
to programs operating in foreground partitions. In the current system
foreground programs are initiated by the operator individually by a foreground initiator. We refer to this as Single Programming Initiation (SPI).
Included in the improvements for multiprogramming is another method of
initiating foreground partitions called Batch Job- Foreground (BJF). This
option will allow a batch job .::>tream to be executed in the fore'ground partitions. Other multiprogramming improvements include individual communication regions, checkpoint/restart and the use of systern logical units in
the foreground partition. These features allow for execution of all user
programs regardless of source language in the foreground partition
operating under BJF.
Let's examine foreground batch job capabilities. In addition to the background partition, one or both foreground partitions can be operated in the
batch job mode if sufficient Input/Output and CPU facilities are available.
The requirements for operating foreground programs in batch job mode
are a minimum 10K partition size, separate systems input and output
files for the partition and the specification at system generation time for
batch job program support. The batch job foreground option will add
approximately 350 bytes to the basic size of multiprogramming supervisor. This .includes the storage requirements for a communications
region for each of the foreground partitions. If disk system input and
output is desired for the foreground, an additional 250 bytes or a total
of approximately 600 bytes will be required in the supervisor. All
system class logical units except SYSLNK will be usable by foreground
partitions regardless of whether they are operated in the batch job or
single program initiator mode However, if the single program initiator
is used then SYSRDR, SYSIPT, SYSPCH, SYSLST must be aSSigned to
unit record devices. The IBM supplied utility programs will be
distributed to run in the background area but may be link edited by the
user to operate in the foreground with batch lob initiation. No other
types of IBM supplied programs are intended to be available to the foreground partitions.
o
Let's take a look at some of the environments that are possible with the
Disk Operating System. In a 16K system we can run a batch job stream,
where 6K is available to the supervisor and 10K available to background
programming One restriction in this environment is the inability to
compile COBOL programs since the COBOL compiler requires a minimum
of 14K.
0
-5-
O
In a 24K system with a supervisor that runs between 8 and 10K, we can have
a background program of 10K plus up to two foreground partitions operating
with a single program initiator.
t
,',
At the 32K level besides operating a 10K background program, we can have one
foreground program operating in the batch job mode and a second foreground
program operating with a single program initiator.
At the 64K level we can have three programs operating in a batch job environment,
background, foreground 1 and foreground 2. The supervisor requirements would
be at least 10K, with 10K minimum requirements for background and foreground
programs. The additional 24K that I have shown available for the background
could be distributed amongst all three partitions.
c
Let's summarize the improvements in the multiprogramming capabilities of DOS.
I have mentioned the features available for background programs, foreground
programs operating with the batch job initiation, and foreground programs
operating under the single program initiator. Program initiation itself is automatic when working from a batch job stream for both background and foreground
programs. Under the SPI it is still operator initiated. Individual communication
region will be provided for each foreground program operating under the batch
job mode. The CheckpOint/Restart facilities now available to background programs
will also be available to foreground programs operating under the batch job mode.
The systems service programs such as the librarian and linkage editor, the Language Processors and Sort/Merge can be executed only in the background partition.
Utility programs will be distributed for execution in the background partition but
can be link edited by the -qser for execution in the foreground where a batch job
mode has been specified. The multiprogram system utility macros as today will
be available in all partitio~lS. Minimum partitions sizes are for background program
10K, 14K if you want to compile COBOL programs, 10K for batch job initiation
in the foreground and 2K vlhen the single program initiator is used for foreground
programs.
Device Independence
There will be a new device independent access method for systems units. This
access method DTFDI and the device independent module will support sequential
processing of unblocked records for files on SYSLST, SYSIPT, SYSPCH, and
SYSRDR. Using this new access method the user can change device aSSignments
at object time without having to reassemble source program and without explicit
lmowledge of the functional characteristics of the assigned device. Besides
providing device independence at object time the use of the device independent
module will reduce core storage requirements since one module can be used
instead of several as are required today.
o
;).37
- 6Private Libraries
We are adding a new facility to DOS to provide for a librarian function
that will allow users to create and use private source and private relocatable libraries on packs other than the system resident pack. This
new facility will give the user greater flexibility in allocating library
space with the result that more room on the system pack itself will
be available for an expanded core image library. The number of private
libraries is not limited, as many private libraries as desired may be
created. However, only one of private source and/or relocatable library
can be operative on the system at anyone time. It is also possible to
use the relocatable and source libraries on the systems resident pack
in combination with private relocatable and source libraries. In addition,
it is possible to eliminate the source and relocatable libraries on the
systems residence pack. Library maintenance and service functions
will apply to these private libraries.
o
CSERV
We have added a new librarian service function called CSERV. This
function will enable the user to display, punch, or display and punch
a specified phase or complete programs from the core image library.
This facility will give the user the ability to transfer programs or
phases from the core Image library of one systems resident pack to
the core image library of another systems resident pack. Since
SYSPCH may be assigned to tape or disk an intermediate card step will
not be necessary. The display, the punch, and display punch functions
can be used for a phase, a program, or the entire library.
c
New Resident Utility Programs
We have provided in addition to the existing utility programs six additional
utility programs resident under DOS.
The VTOC display program will enable the user to display the labels
contained in the volume table of contents of a disk pack. This will enable
him to more easily keep track of his files and their extents. The output
of the VTOC program may be directed to a printer, a tape file or a disk
pack. Labels are identified by their locations within the VTOC and their
format types. There are no special utility modifier cards required for
this program. All that is necessary to execute the VTOC program will
be a job card, assign cards, and an execute card.
Another utility program is the initialize disk prograln. This program prepares the disk pack for use on the 2311 or 2314 disk drive. The initialize
disk program first checks the volume table of contents to verify that
there are no unexpired files on the pack. It then generates home
addresses for each track, does a surface analysis assigning alternate
tracks when required and pre-formats the volulne table of contents.
(;'
- 7-
o
The alternate track assignment program is used to assign alternate tracks
to replace defective tracks on the 2311 or 2314 at any tirDe other than when
initializing the pack. When an alternate track is aSSigned the records
contained on the defective track may transfer to the alternate track. Full
diagnostics are also provided.
Three separate programs are provided to copy and restore disk packs.
These are copy and restore disk with card, copy and restore disk with
tape, and copy disk to disk. These copy programs may be used to copy
the entire volume or just specified files. When the copy volume function
is used the entire contents of the volume including the IPL records,
volume labels, and the VTOC will be copied. The copy file function permits
the transferring of a data file from disk to cards, tape, or another disk
pack.
Availability
There is one more important subject to be covered. That is the availability of these improved facilities of the Disk Operating System. The
improvements in the tape label area will be available from the Program
Information Department November 17, 1967. The rest of the improvements
that I have discussed for the Disk Operating System will be available
April 5, 1968.
o
o
Sf-SSt ON NUMBER
T • 3. 2.
SPEAKERS
NO FORMAL SPEAKERS_ WADE NORTON, ACT 'G. CHAIRMAN, HAD BEEN
SCHEDULED TO PRESENT A PAPER ON RUST ENGINEERING CO.'S. 16K OS
WHICH RUNS ON A 65K MACHINE. THE OS PROJECT FELT THAT MOST OF
THEIR MEMBERS ALREADY KNEW THE ADVANTAGES, AND THAT THE
PR~SENTAT10N COULD BETTER BE MADE AT A LATER MEETING WHEN MORE
PUBLICITY COULD ATTRACT MORE DOS' ERS WI OR PLANNING 65K CORE.
DISCUSSION
WE THEN DISCUSSED SAN FRACISCO (OR LATER MEETINGS). THE NEXT
MEETING NEEDS TO COVER
A. LINK EDIT (I), SCHEDULER (I)
B. LINK EDIT (II), SCHEDULER (II)
C. OS FOR THE USER OF SMALL 360'S, PRESENTATION OF PAPER
ON RUST'S SYSTEM & PANEL OF OTHER SMALL OS'ES.
ON RUST'S SYSTEM & PANEL OF OTHER SMALL OS'ES.
O. PLANS & ORGANIZATION SESSION.
E. SQUAWK SESSION.
c
~I
'-'I
SESSION NUMBER
SPEAKERS
T.3.3
HARISH J. JAGTIANI, RICHMOND ENGINEERING CO., INC. nN FORTRAN
PROCESSOR FOR DRILLING TUBESHEETS nN NC MACHINE
('
~}
o
FOR TR AN PR OCESSOR
FOR
DRILLING TUBESHEETS ON NC MACHINE
o
by
Harish J. Jagtiani
Richmond Engineering Company, Inc.
Richmond, Virginia
presented at
COMMON, Cincinnati Meeting
September 6, 7, & 8, 1967
NUMERICAL CONTROL
Numerical Control (or NC) is a control systern usi.ng punched tape or other
automatic control devices to direct the operation of machines and 1.xlachine
systems. Numerical Control provides a highly accurate and efficient means
of positioning or controlling the path of a tool. Machine tools are classified
as being point-to-point or continuous i.n operation.
"Point-to-Point" implies that machining operations are only at certain
locations on a part and that the cutter is retracted [rorn the piece before it
moves to another location. Point-to-Point, particulary valuable for precision
drilling, boring, reaming, involves automati.c coordinate setting. Careful
layout, therefore, is less dependent upon the operator. "Continuous" (or
contouring) implies that there is no retraction of the cutter as it produces a
path in moving throughout a specifi.ed area. Poi-nt-to-Point machines operate
in two dimensions; continuous machines in two or three dimensions.
c
NUMERICAL CONTROL AT RECO
The application that we are particularly concerned with, at Richmond Engineering
is for NC Point-to-Point machine tools .. In the manufacture of process heat
transfer equipment, a good portion of-t-h-e--work is involved in the drilling of
holes in the tube sheets and baffle plates of heat exchangers and the bolt holes in
the flanges'. With a view to secure some of the benefits from automation,
Richmond Engineering decided to acquire a NC drilling machine. In 1961, we
purchased Pratt and Whitney's Tape-O-Matic numerical drilling machine. It
is a single machine with a table working surface of 30" x 20" and a transverse
table travel of 15". This machine is capable of taking care of the drilling for
R ECO Standard Products up to 16" diameter. The experienc e gained ·with the
Tape-O-Matic proved that NC machines could certainly provide tremendous
advantages. With an increase in business activity Richmond Engineering decided
to purchase a much larger drill press which would not only produce dollar
savings but replace some of the existing manual drills and still have enough of
excess capacity to handle increased business.
In late 1964, an order was placed for the American NC 819-32 Travelling
Openside Boring and Drilling machine. This has an actual work area up to 9Z
- 1 -
o
inches wide and length of 15 feet.
is currently in operation.
o
This machine has now been i.nstalled and
The job planning procedure for this machine was similiar in many ways to the
Tape-O-Matic. The machine has a "floating zero" or a "full zero shift" feature,
as shown in figure 1. The reference point is set at the center of the tubesheet,
which then allows the parts programmer to give all axi.s dimensions relative to
the workpiece. The programming done manually was hand-written on a process
data sheet.
DATA PROCESSING FOR NUMERICAL CONTROL MACHINES
In the preparation of "programs" or "instructions" to govern the movements
and processes of the numerically controlled devices, there was the problem
of cummulative tolerance errors being introduced through'manual computational
procedures.
An investigation was undertaken to determine what aid could be offered to alleviate
the problem of computational tedium and inherent accuracy difficulties associated
with manual preparation for NC processes. The thought of computations and
accuracy naturally suggested electronic computers as a possible means for
solution.
o
IBM has developed a simple language AUTOSPOT (AUTOmatic System for
Positioning Tools). AUTOSPOT is a general purpose computer program designed
to aid the parts programmer perpare instructions for NC Point-to-Point machine
tools. The parts programmer can describe the required operations in a familiar
language, without calculations, repetitions, and the tedium of output format
preparation.
The AUTOSPOT general program processor receives the input information, performs
various operations, and delivers outputs to the post processor. The post
processor converts the information into the data required to perform the
machining by a specific machine tool. Since there was not an existing post
processor for the American drill and writing up on would involve considerable
time, several months, investigation was directed to prepare an aid to our
specific pr,oblem that would specify the locations of the centers of equally spaced
holes on the tubesheet and also the locations of the centers of any number of
equally spaced holes on the bolt circle.
In this effort to relie've time consuming, laborious, and technician type work
from the shoulders of the part programmer, a generalized computer program
has been developed to perform position calculations for tubesheet hofes and bolt
circles configurations. These configurations are COlnInOn enough in our line of
products and for a firm using numerical control equipment to justify such an
approach.
o
- 2 -
--------------------
.........
.,,~,---,-~~-~~~~~~-~~~-~-----------
EXPLANATION OF TYPICAL LAYOUT
From a typical layout of a tubesheet you will notice the tube holes are equally
spaced and so are the rows of tube holes. Further the layout on one side of
the center line is generally a ITlirror iITlage of the other side.
There are tw·o configurations of tube holes that we are conc erned with.
tubes are either on a triangular pitch or a square pitch.
o
The
Lets consider the center of the tubesheet as the reference point with coordinates
of (O. 0, 0.0). Knowing the pitch of the holes, one can easily compute the
coordinates of each hole in each row.
Hence for the first row the coordinates of holes ITlarked A is (1. 0, 0.0), B
(2.0, 0.0) and so on. For the second row the coordinates maybe (1. 0, 0.5),
(2.0, 1. 5) and so on.
We can compute these coordinates manually and put them down on a process
sheet.
N ow if a row of holes wa s defined by the following statement,
SY
= 0.0,
SX
= 16.0,
EX
= -16.0,
NH
= 0.0
we can compute the coordinates of each hole in this row.
Pitch = (EX - SX) / (NH - 1)
= {16. 0 - ( -16.0)) / (33-1)
= 1.0
o
Hence
l\;
1
N
N
2
3
N
N
32
33
Y
Y
Y
0.0
0.0
0.0
X - 16. 0
X - 15. 0
X - 14.0
Y
Y
0.0
0.0
X -I- 15.0
X + 16. 0
This proc edure is the logic of the computer program. The flowchart further
outlines this routine to compute the coordinates of each hole, figure 2.
FILLING THE INPUT FORM
A typical input data form filled up is shown in figure 3.
The master card contains some general information followed by rCODE, the
reference point (corresponding to the Datum Reference to permit the
specifica~ions of point coordinates relative to the most convenient origin), total
of detail specification cards, total number of holes to be drilled, type of pitch,
o
pitch and bolt circle specifications if any.
The detail card specifies coordinates for eacH. row or part of each row in
some cases. It contains SX, SY, EY, NH, JT and EXCEPT.
Where
SX
SY
EY
NH
JT
EXCEPT
-
Starting X coordinate
Starting Y coordinate
Ending Y coordinate
Total number of holes
Total number of exception holes
Position of exception holes
EDITING THE INPUT
After the input form has been filled and the cards punched we run an edit of
the input data. The edit program checks the input data for completeness and
consistency.
The edit program checks the following:
1.
Code
2.
Bolt circle specifications
3.
Pitch of the holes
4.
Spacing of the rows of tubes
5.
Tot a 1 n urn be r of hal e s
6.
Total numbe r of detail cards
and so on.
The output listing from the edit may appear on the first run, somewhat as shown
by figure 4.
After the e'dit IS run the errors - if any - are corrected, and the input deck is
run through the processor. The processor punches the output ca-rd a'nd also lists
each block of the data for the NC machine on the printer. A listing of the output
is as shown in figure 5.
USING THE 3 - SPINDLE DRILL
o
On the American NC Drill we can also put a 3-spindle attachment. Pre-paring
a tape to run with the 3-spindle drill is naturally somewhat different from that
for the single drill.
The attachment we have and the usual pitch of the holes we maintain limits
drilling three consecutive holes. Instead we have to drill 3 alternate holes
at one setting.
For example if we had 6 holes equally spaced in a row, we first center the
drill over hole number 3 and drill holes 1, 3, and 5. Then we position drill
over hole number 4 and drill holes 2, 4, and 6. In other words we drill
groups of 6 holes in 2 operations.
()
The processor automatically takes care of this. Now we do have rows of holes
of less than 6. These holes we pick up on the single spindle drill. The
processor hence is designed to segregate groups of holes to be drilled with the
3 spi.ndle drill and those with the single spindle. These two data sets we
store on the disk. Each record in the first set is accessed and cards punched
out for the same. Then the same is done for the sec ond set of records.
The general flowchart of the processor is shown in figure 6.
LIMITATIONS OF THE PROCESSOR
Obviously the processor we are talking about has its limitations. It is not
as versatile or powerful as the AUTOSPOT processor. For instance, one
half of the tubesheet is generally symmetrical about the axis. If the parts
program were written in AUTOSPOT language, just one statement would
have sufficed to invert the pattern on one side to the other. But with our
present processor we h?-ve to define each row of the other half again .. Of
course this is not too difficult, because all the parts progralTIlTIer has to do
is to reverse the sign for the SX dimensions. However, this is not what you
may call very sophisticated. We also have the problelTI of not being able to
define patterns other than the bolt circle or a row.
Not long back we had a job which had a pattern of 3 concentric circles each with
36 holes and another pattern within the triangle forlTIed by the circles. There
were roughly 8000 holes to be drilled. Doing the programlTIing manually would
have taken ages. So we wrote a special program which consisted of routines
from the original processor. This was not difficult but certainly took som.e
effort and tim.e, which could easily be saved by using AUTOSPOT.
PUNCHED/TAPE PREPARATION
The data input to the machine tool is one inch perforated 8 channel tape ln
accordance with EIA standards.
Initia:ly the tape was manually punched on a Friden Flexowriter. This was
both time consuming and laborious task full with human errors and omissions, thus
offsetting some of the advantages gained by using the computer.
When we changed to card output from the cOITlputer we added a Friden Automatic
Card-to-Tape Convertor to the 2201 Flexowriter. Now the cards are fed in the
card reader and the tape is automatically punched out.
- 5 -
o
CONCLUSION
This processor as you can see is not very sophisticated or powerful as many
others. However it has proved extremely easy to use for the parts prograr,nmer.
The programmer does not need any knowledge of computers or computer
languages. It has considerably reduced the lead time in preparing tapes foythe NC machines -- from days to a couple of hours. And finally it has helped
a small manufacturing company with somewhat limited skills and resources
to inc rease its producti vity and product quality.
o
- 6 -
:J. '17
----
_._-- -----_._-_ ..
., .•
-
..
-
---_ .._-.'
,,_
..
-
.-
.---------.. ---------~---------... - ... -.- ..
------
#,
#2
BOLT
-x
___ ----'L-,.---
It
·v·
A•
./
AXIS
Establish At
C e n t e r 0 f W0 r k. P ieee after
Loading it on
the Table.
#4
X=-y=_
----TABLE------------
·x·
-AXIS
~--~~~~~--~A·
FIGURE 1
-_ .. _...... _---_ _.. - - - - - - - .1)£ TAI L C. /i 12))
.•....
0X,
.sY,
E
Y)
1'Jf1 )
NE, £,XCEPT(NE)
,
. . - - - - - - - -j~
. ----T
.j
/
.sy- EY
NH-i
_----l
'f
PI7 C I-I
PIlcH
.____. . . .1.
tV::=. N-f-i
SI
o
FIGURE 2
II
I
, , / C ,':)1,,1 PI) R. r=..
COrv;PUT E. j)
TO 0,IVe",;
py=-
l
I
Program. "NCRECO!'
I
Sheet
RICHMOND ENGINEERING COMPANY, INC.
2.
Of
NUMERICALLY CONTROLLED DRILL INPUT
;V b, c.~
Input Prepared By
70b7, 30
RECOJobNo.
r
Input Checked By
RECODrawingNo.J)-b72.02.-!
Custom.er Nam.e
--------------------
r"
job
nurnbe r
........
bv
r"
~
0
-oJ
E
'U
no.
of
reference
x-axi s
~
rj
rc
ending '+-lo (/)(J)
starting
starting
..--4
rj
;.. x-coo rd'lnaie y -<.:cordin ate y-cmrdinate o a jt
.-<
c...c:
,IX.,
."tv,v ,v, v
2-9
1
,L.
01/,
2
-
rt.
).l
t"I
2 ,
7.~
-I: 2,Q. ,
+1
3,9
3,t
.
r3.2
31
·
I
1
.,
1
1
,
..
·
·,
,
,
1
t
t
1
,
,
1
1
1
1
f
I
,
,
I
.
1
f
,
,
,
f
,
·
,
,
,
·
2.¢
,
.L
,
1
~
To
>
,
,
I
,
,
1
.
ALL COLUMNS SHOULD BE CODED
Frn F Ar:H LINE.
FILL WITH
ZEROES WHERE NECESSARY.
ICODE ;;; 1 for tube sheet only
= 2 for bolt circle only
= 3 for tube sheet &: bolt
circle
= 4 for tube sheet (3 - spindle
drill)
= 5 for tube sheet & bolt hole s
(3 - spindle drill)
JP = I for triangular pitch
= 2 for square pitch
LV
= 0 _ if
= I
·
·
I
,
,
.
·
,
3.4 ,
3,3
Q)
exc eption hole s
18- 25
- ;!. 3,7:;-
1,o,g7,S -tf!/16.S0
Xt.Xl-X ~X.xIXfX,~X.X X,Xi:'{ Xi-X
10-17
2
I
x."rx
;.:,v
tl.O
'"0
NOTES:
H
Y
bolt bolt cin:::le
dia- diameter
,
·
I
FIGU()3
,
,
I
I
1
JT
=
angle is given
for bolt circle hole s to be
straddled about cente r li!1.e
tota.l nUITlbe r of exc eption
holes
o
--------
Q
~----
~~
--
0
"
LISTING AND CHECKOUT OF INPUT DATA FOR AMERICAN NC DRILL
=======================================================
1 7057.30
l
0.0
2
DEC
8 21 67 3 40.0000 40.0000
23.7500
39
0
0
0
0
0
0
0
0
0
0
0
1.0825
24.3150
-24.3750
40
0
0
0
0
0
0
0
0
0
0
0
2
2.1650
-23.7500
23.7500
39
0
0
0
0
0
0
0
0
0
0
0
2
6.4950
-23.7500
23.7500
38
0
0
0
0
0
0
0
U
0
0
0
SPACE BETWEEN THIS ROW AND PREVIOUS - COMPUTED
COMPUTED PITCH # 1.2838
GIVEN PITCH # 1.2500
68 0.875 51.2500
0
4.3300
GIVEN
#
1.0825
DETAIL CARD NO •••
SPACE BETWEEN THIS ROW ANn PREVIOUS - COMPUTED #-2.1650
GIVEN
#
1.0825
2
5.4125
24.3750
-24.3750
40
0
0
0
0
0
0
0
0
0
0
0
2
"J.2475
24.3750
-24.3750
41
0
0
0
0
0
0
0
0
0
0
0
GIVEN
#
1.0825
GIVEN
"#
1.0825
2
4.3300
-23.1500
5
DETAIL CARD NO ••• 7
DETAIL CARD NO •• 7
2
7.5175
23.1250
DETAIL CARD NO •••
CIJ
-.....
2 1 1.2500
-23.7500
DETAIL CARD NO ••• 4
DETAIL CARD NO •• 4
'-'
35 1275 57.8750
a
23.1500
39
0
0
0
0
0
0
0
0
0
0
0
SPACE BETWEEN THIS ROW AND PREVIOUS - COM?UTED #-2.1650
COMPUT~~ PITCH # 1.2188
GIVEN PITCH # 1.2500
-23.1250
38
0
0
0
0
0
0
0
0
0
0
0
SPACE BETWEEN THIS ROW AND PREVIOUS - COMPUTED
2
8.6600
-22.5000
22.5000
37
0
0
u
0
()
0
0
0
0
0
0
2
9.7425
21.8750
-21.8750
36
0
0
0
0
0
0
0
0
0
0
0
2
10.8250
-21.250e
21.2500
35
0
0
0
0
0
0
0
0
0
0
a
2
11.9075
20.6250
-20.6250
34
0
0
0
0
0
0
0
0
0
0
0
2
12.9900
-20.0000
20.0000
33
0
0
0
0
0
a
0
0
0
a
0
14.0725
19.3750
NUMBER OF OETAILCARDS
-19.3150
32
0
0
0
0
0
0
0
0
0
0
0
2
NUMBER OF HOLES
#
ACTUAL#
ACTUAl# 1170
38
GIVEN#
GIVEN# 1215
35
#
4.3300
0
o 1
------~---------------_ _ _ _ _.,I"I.,
i'
________~J~O~B~N~O~.~.~.~~1~O~5~1~.~30~__~S~P~E~C~I~F~IE~O~R~Y~.~.~.~O~E~C____~D~A~T~E~.~.~.~*~*~*~*~**~lL9~6~1~____~
ICUOE ••• 3
I
RECO PROCESSOR FOR AMERICAN NC DRILLING MACHINE
TUBESHF.ET HULES COORDINATE DATA
TIJAESHEET SIZE
TOTAL NUMBER OF TUBES
_ _ _ _ _ _ _ _--:.-TO=...TAL NtJMHER OF RO~/S
NtH-1 BE R 0 F PAS SE S
PITCH - TRIANGULAR
57.875
1270
38
2
1.250
EQUALLY SPACED BOLT CIRCLE HOLES COORDINATE DATA
BOLT CIRCLE DIAMETER
56.2500
NUMBER OF EOUALLY SPACED HOLES 68
NOMINAL HOLE SIZE
1.0000
ZERO ANGLE CLOCKWISE FROM VERTICAL
DEGREES
2
MINUTES
38
SECONDS
49
REFERENCE POINT
N 1000
N 1001
N 1002
G 80
G 81
G 81
X-COORDINATE
V-COORDINATE
X 000.000
40.000
X
40.000
40.000
Y 000.000
Y 16.250
Y 17.500
M
R
R
0
1
1
~ ~gg~!~--~~~:~~----------------~~~~~~~:~6~~~g--~~~~~~--------~-----~N 1005
N 1006
N 1001
N 1008
N 1009
N 1010
N 1011
N 1012
G
G
G
G
G
G
G
G
81
81
81
81
81
81
81
81
Y 21.250
R 1
Y 22.500
M 1
Y 23.750
M 1
Y 25.000
R 1
Y 26.250
R 1
Y 27.500
R 1
V 2R.750
R
1
Y 30.000
R 1
N 101~3~__~G~8~1________________Y~~3~1~.~2~5~0__~R~~1~___________________
N 1014
G 81
Y 32.500
R 1
N 1015
G 81
Y 33.750
R 1
N 1016
G A1
V 35.000
R 1
N 1011
G 81
Y 36.250
M 1
N 1018
G 81
Y 37.500
R 1
N 1019
G 81
V 3A.150
R 1
N 102~O~--~G~8~1--------------~Y~~4~O~.O~O~O~--~R~~1--------------~----
_ _ _ _......_...pN,;..,..;1021
G
81
- ...
G Ml
V 41.250
R
1
N Ib~l~2~--¥G---8~1--------------~Y~~4~2-.~50~O~--~R--~1--------------------N 1023
G 81
V 43.150
R 1
----..:..:.-..:...::...:.:..::.--:=--::...=...---------..;..-..:..=...::;..;..::.,.';.--~-;::.----------
N 103h
V 60.000
R 1
__ __ ____
________~N~I~~~1____~G~8~1~__~__________~Y__~6~1.uZw5~Q~ ~R ~l
~__~__________
N 103A
G 81
Y 62.500
R 1
________~__~N~LQ3~9~__¥G~R~1~______~~----~y--K63~.7w'~Q~.__~8~Al__________~________
N 1040
G·81
X 41.G82
Y 64.315
R 1
~
~tM~_ _ _ _ _ _~N~10~4~1~~G~'~t~·~.d-t~--~--~X~6~3~.~t~2'~~R~~1~--__--~~__~i
N 1041
G 'II
Y 61.875
R 1
~
N 104:4
C81
V 60.625
R. 1
rtCuRE S
i
OR. JOHN MAN'OTES
COMPUTER TECHNOLOGY DEPT.
PUROUE UNlVERS'rv
CALUMET CAMPUS
HAMMOND, IN 46323
YGS
STOP
3- SPII'I.PL£
f2..EAj) ALL
:J)GTAIL &t£P-s
i:-
COIY'IPUTE.
N liTES
Row
COOf2j)I01= E~GH
S E ~I!.£e:, ATE.
01=' /loLeS
C1JM PC! Tt:
NO
J)12.i LL t;:...P
I
CJ
a>
Pol2 WINS
CCJMPlI/G. CoCilI
(:of!. .BoLT
CIi!.cL<='
HoLes
FIGURE 6
eY
SPzAlj)LE
PC/Nell
C/t-12.J).5
~R..
'1't'3£5 Hee7
~Il , f\A
S F S SIn f\'
RE R
T• 3 • 4
SPEAKERS
R • f).
R R F f\\ 1\' AN,
J R tv1 seT F~' T J FIe C E'" T ERn f\I CHF MI CAL f NG I '" F F R I f\! G
APPLI(ATIONS nF CSMP
'0
"{
",
DR. JOHN MANrOTES
COMPUTER TECHNOLOGY OEPt
PUROUE UNIVERSITY
CALUMET CAMPUS
HAMMOND. IN 46323
Petro-Chemical Engineering Project
Subject: CHEMICAL ENGINEERING APPLICATIONS OF CSMP
R. D. Brennan
IBM Scientific Center
2670 Hanove r Street
Palo Alto, California 94304
(415) 327-2300
T.3.4
•
.'
.~
'\
3 p.m.
Sept. 7, 1 967
.
kY
,(!HEMH!'At
ENGINEERING APPLICATIONS OF CSMP
The 1130 Continuous System Modeling Program, an adaptation of PACTOLUS,
is a new digital-analog simulator specially developed for online experimentation by
the design engineer. 1130 CSMP, in combination with the computing power of the
IBM 1130, has demonstrated itself to be an effective and economical tool for simulation studies across the breadth of engineering practice and the physica I and biological sciences. It provides a library of 25 standard elements plus five "special"
elements that the user himself may define. Interaction with the program is simple
and conven ient via the console keyboard and the console entry switches. From the
keyboard, or via punched cards, the user enters 1130 CSMP language statements
defining the configuration and associated parameters. During entry, automatica Ily
typed instructions and diagnostics guide the user through the procedures. These tell
him how to initiate data entry, how to select the variables for printer and plotter
output, and how to specify the integration interval, total run time, and output
intervals. He may interrupt a simulation run at will to modify or extend the simulation, and need not follow a rigid schedule.
The Continuous System S imu lator Approach
The S/360 Continuous System Modeling Program offers the best available
illustration of what can now be done in continuous system simulation. This IBM
program is an 'advanced version of DSL/90, the most powerful of previous efforts
of this kind. S/360 CSMP incorporates a number of novel features: it can, for
example, incorporate procedural coding within the definition of a macro element;
it can conveniently calculate initial and terminal conditions (thus placing the
entire simulation under programmed control). These new features provide the
engineer a significantly more powerful and sophisticated tool for simulation.
Working with a basic set of functiona I elements for model ing the components of
continuous systems, the S/360 CSMP user specifies the interconnection of these
functional elements by the (FORTRAN) equivalent of common mathematical notation for functional dependencies, namely, Y = F(X). S/360 CSMP thus represents
a rapprochement between analog block modeling and conventional digital programming. In addition to the block modeling capability shared with digital-analog
simulators, this kind of program provides the power and convenience of algebraic
and logical statements. Its FORTRAN-like language can be used as either a parallel
(non-procedural) or as a procedural programming language.
Irr'
t,\l./
The relative advantages of these two programs were compared within the
context of a chemical engineering example; this example was previously described
in connection with PACTOLUS in the March 1966 issue of Instruments and Control
Systems. (An expanded version of this simulation study is expected to appear in a
-(."."'".....
,)
... -)11
o
future issue of Genie Chemique.) Several other interesting applications in chemical
engineering are described in the Proceedings of the IBM Scientific Computing
Symposium on Digital Simulation -~rC~-!_~!.ir~~o~~ Sy~tems, IBM Form No. 32-0-1943.
Both programs are avai lable from I BM as Type II supported appl ication programs. The Application Descr iption Manuals for 1130 CSMP and 360 CSMP are
available under Form No. H20-0209-1 and H20-0240-1 respectively.
SESSION NUMBER T.3.4
SPEAKERS
CUSTOMER - I.B.M. RELATIONSHIP
ROBERT LUKEMAN, SALES SECTION
R.C. METEER, SYSTEMS ENGINEERING"
G.P. MONJEAU, FIELD ENGINEERING
o
hesanted
b7
J. R. Luke.man
IBM Corporation
COMMON MEETING
September 6 - 9, 1967
112 kat Post Road
lIl1t. PlaiDs, I. Y. 10601
(Panel Discussion o~ Customer - IBM Relationship with respect to FE,
SE, and Sales - Sales Section)
With the rapid changes we have been experiencing over the past decade
~e data processing industry with regard to hardware, programming
systems and applications, it is quite natural that there likewise be a
corresponding change in the role of the IBM account representative.
This change has not been an attempt to alter the primary responsibility
of the account representative, which has always been and will continue
to be ff service to the customer", but one of implementation of that responsibility.
in
()
The complexities of the industry today have brought about a need for a
type of salesman who is not only knowledgeable in all aspects of data
processing, but one who acts as an interface between the customer and
the IBM Company. No one individual is able to be an expert in all facets
of data processing and still possess the ability to comprehend every
conceivable application area the customer may have in his business. For
this reason there is a trend to cover our larger accounts with a self sufficient team made up of experienced employees with both marketing and
systems engine.ering backgrounds. Their job is twofold. One, (Slide 1)
they must represent all of the service, experience and technical expertise of the IBM Company to you the customer, and two, (Slide 2), they
must represent you, the customer, to IBM.
(Slide 3) - In carrying out this tw"ofold responSibility, a thorough knowledge of both organizations is a necessity. .M far as your business goes,
the IBM account representative must have detailed lmowledge of the
applications you presently are running as well as those that are in the
development stage. This includes not only the volume figures, input
sources and output requirements, but the interrelationship of these applications as well. In addition, he must know your organization, where
the data processing function. :fits into the corporate structure and what
the evolution of this organization has been. Another important aspect
the account representative must be keenly aware of, and one that our
company is putting a great deal of emphasis on, is knowledge of the particular industry, whether it be finance, education, insurance, manufacturing, and so on. Knowlege of not only what is occuring in the industry
today, but what are the trends of the industry. Finally, it is imperative
that the account representative have a good working·relationshipwith the
top executive management of his customer so he knows what the long range
qoals of the business are and can make long range recommendations
regarding data processing. As you are all aware, the implementation
of a data processing function, if done correctly, takes many weeks and
months of planning, testing, installing and converting. If this time
spent is not consistent with the long range objectives of the business it
can be disastrous to all concerned. To sum it up, the account representative must work hand in hand with every level of management (Slide 4)
aspect of your business and become as knowledgeable as you yourselves
are.
Once this knowledge is gained, and it is a continuing process to say the
least, the account representative must couple it with the knowledge,
experience and education he has received from the IBM Company. This
knowledge covers the full spectrum of IBM (Slide 5) products, services
and resources as well as our company policies and businej3§ practices.
It is the responsibility of the team manager to be sure that he has
developed the required talents within the team itself and to be cognizant
of these resources so .he may utilize them as the situation dictates. These
include such services as industry specialists, product speCialists,
application speCialists, operating system and programming specialists
and many more who are available from district, region and division
headquarters to assist the local people in fulfilling their primary role that of services to you the customer. The account representative must
have in-depth knowledge of the products capabilities, relationship to
other hardware, characteristics, potential and most importantly the
application value to solve a particular problem or set of problems.
Other information that the account representative must be intimately
familiar with in order to be of maximum service to the customer is, of
course, our own internal policies and business practices so as to be in
a position to advise and counsel in regard to contracts, deliveries, prices
and stay withing the general business rules that we must follow.
The account representatives can work hand in hand with you in such
matters as (Slide 6) providing professional guidance and advice on organization, both within the data processing group and the relationship of data
processing with the other departments of the business. Advice on budgetary matters very often can be offered by the account representative in
light of what other firms are doing and trends within the industry group.
Personnel matters is another area where the experience of the account
representative can be used to offer suggestions on position descriptions,
salary ranges, apptitudes, and labor markets. This advice can materially
assist you in your day to day operations as well as long range planning.
Planning sophisticated applications such as Management Information Systems
or Total Integrated Systems, Educational programs, Teleprocessing
systems and the long range data processing plan, can all be facilitated
within the assistance of the account representative.
c
~,
Jr
I
o
The things I've been mentioning all have been areas where you, as
customers, can benefit the most in developing a working relationship
with your IBM account representative. In doing so, the IBM representative is fulfilling one of his primary responsibilities, and that is
to make sure that every account in his territory is a satisfied customer.
A mutual respect for the other person's activities must exist between
the customer's personnel and the account representative.
This mutual respect refers to the relationship the account representative
must have for the supporting services available to him from the IBM
Company, as well. In acting as the interface between the customer and
the IBM Company, he must properly utilize these services in the best
interest of both. The group that he usually works closest with, of
course, are the Systems Engineers. We are fortunate to have with us
today an SE who's qualifications make him one of the finest examples
of service excellence in the IBM Company. I'd like to turn the conference
over to Russ Meteer.
c
o
",(4"'."
COMMON MEETING - Cincinnati, September 6 - 8
Operations Committee
Customer - IBM Relationship with Respect to FE, SE, and Sales - (SE
Section)
R. C. (Russ) Meteer
IBM Corporation
112 East Post Road
White Plains, New York
C'
10601
Area Code 914, WH9-1900
Extension 6153
Thursday, September 7, 1967
3: 30 p. m. Session VIn
3 pages of text
o
o
COMMON Meeting
September 6- 8, 1967
(Panel Discussion on Customer-IBM Relationship with Respect to FE,
SE, and Sales-SE Section)
As the complexities of the Data Processing industry have increased
over the years, it has become unrealistic to expect anyone person to
provide all of the knowledge and talents necessary to meet the varying
needs of a computer installation. I am sure if you, as a group, look
at your own operations, ~lOU will find Data Processing being used for
engineering and scientific computation, accounting and record keeping,
and management guidance and control. Each of these areas poses
unique requirements. The Systems Engineer brings specific ta lents
and sources of information to bear on these varied problems.
Generally, Systems Engineering will be most active during the installation planning phase. The installation of a Data Processing syste'Tl
consists of six major elements: (Slide 1) design, program.nl.iJ.lg, testing
and debugging, documentation, conversion, and implementation.
c.,'
,/
The most important factor which affects each of these six elernents is
your in-house capability to accomplish thenl. Education. is, and has
been for many years, one of the iTIl.portant initial pha ses of your experience with IBM, and is the beginning of the developrnent of this
in-house capability. The transition from forTIl.al classroom training to
the practical application of your newly gained knowledge to systeTIl.s
design, programming, and productive use of the system is one of the
key elements in the relationship between the SysteTIl.s Engineer and you
as a customer. The SysteTIl.s Engineer brings to you the technical
guidance necessary to TIl.ake this transition.
The services provided to you by the Systems Engineer will vary to a
great degree depending on the use you are making of your system.
The comTIl.on eleTIl.ent in any case is the iTIl.portance of TIl.aking the transition from theory to practice.
You might ask why IBM is so concerned with custOTIl.er self- sufficiency.
Only you can TIl.ake the TIl.ost effective use of the system in your organization. Many studies over the last few years have shown that the key to
a successful computer installation is user participation and TIl.anageTIl.ent
attention.
0'
Let's discuss the ways in which Systems Engineers will help you develop
your data processing skills during the installation planning. Design is
essentially defining what is to be done. If your applications are engineering oriented, this probably involves formulation and the trans lation of
-zthe forrnulas to FORTRAN. If your jobs are rnore accounting-oriented,
then design of the systerns flow and a definition of terrns is involved. In
either case, you as the user know best what your problerns are, and what
results are desired. The Systerns Engineer can often assist you by showing
you tried and proven approaches to obtain a solution. In addition, he can
discuss your own ideas with you and help you to evaluate their effectiveness in achieving the desired result.
Prograrnrning is one thing which you learned in sorne detail in your
education prograrn. There is a difference, however, between classroorn case studies and actual prograrns. Here the Systerns Engineer
provides guidance and helps you in the use of various prograrnrning
techniques until you have developed a proficiency of your own which
-allows you to work independently.
Testing and debugging usually provides the first opportunity that a custorner has to operate the systern. Those first unsure moments at the
console change very rapidly to an "old friend" relationship with the
proper guidance and counsel provided during the early testing phases.
Probably the most neglected and overlooked phase of installation planning
is that of docurnentation. And yet, for a smoothly operating installation,
docurnentation is ~ssential. The plan in a programmer's mind is not a
suitable substitute for readily available docurnentation. Providing that
you, the customer, have rna de normal progress in developing a programrning and testing proficiency, docurnentation can proceed without a
grea t deal of SE invol vernent. I would sa y tha t the rna in contribution in
this area is one of constant nagging to be sure that the job is actually
accornplished. We hum.ans seem. to have an aversion to writing things
down.
Conversion is largely a clerical function, but it requires close coordination
and rnonitoring to insure that all the records of your cornpany are correctly
transcribed into the form.at intended for processing by the programs which
have been developed.
And finally, irnplernenta tion. As the saying goes, "the proof of the pudding
is in the eating." I arn sure that all of you have had a few sleepless nights
during that period when you were irnplementing the programs on the systern
and m.aking thern work with real live data. At this time, the System.s
Engineer will be close at hand, but, hopefully, he will have been able to
help you develop to the point where you are able to do rnuch of the inlplementation yourself and the crises of irnplenlentation are nlininlized.
There is one thing I would like to nla:ke quite clear and that is tha t the
SE is not always all-knowing.
-
-
-----
He sOnletirnes runs down blind alleys and
--------------
o
- 3 -
o
finds problems which he cannot solve. However, he has available to
him a number of sources of information which can help him to help
you. (Slide 2) In each district, a Field Systems Center is available
to him a s a sour ce of te chni cal information and! guidance.
The Field
Systems Center encompasses the Education Center, Test and Datacenter,
and a Systems Design and Installation Center, better known as SD&I.
It is this last group, SD&I, which provides the necessary technical expertise
to assist the Systems Engineer in questions on programming languages,
operating systems and application-oriented programs, such as simulation,
netword design, etc. (Slide 3) The Field Systems Center in turn can go
to one of our programming centers, either in Poughkeepsie, or Endicott
in order to research problems and questions for which there are no
answers at a district level. The programming centers have direct
contact with the implementors and development groups to resolve any
questions which the programming centers cannot answer.
Another source of Systems Engineering assistance, which is available
in many of the metropolitan locations, is the Installation Center.
The
Installation Center concept is one in which the Systems Engineering
installation planning assistance is centralized in order to concentrate the
various resources of information together in one location so that they
might bear directly on the customer IS problem. In the Installation Center
the customer may do his programming and testing in a machine environment having constant guidance available from one of the Installation Center
personnel. C:urrently these centers support the 1130, S/360 Model 20,
and punched card equipment. Experience shows that the Installation Centers
accelerate the installation planning cycle and the customer I s learning
process.
In the past, the Systems Engineer also has been heavily involved in the
maintenance of Type I programs after installation. This function has
now been assumed by the Customer Engineer and will be discussed by
Mr. Monjeau from the Field Engineering Division.
We see then the role of the Systems Engineer as one of guidance, information
and counsel with the ultimate objective of developing customer capability.
Self-sufficiency on your part will provide you with the means to realize
the greatest return on your Data Processing dollar.
-.%4
ij
a.; ,t,,",;: ".lATZ, ,( P4fli/lUiA«UA.£Aj4AAAM1;;AOMMM14itJlW=_ _
Abstract of Presentation to be Made by G. P. Monjeau,
Field Engineering Division
Subject:
FE - Customer Relationship
Enclosed are copies of the Vl1 - graph foile which will be
shown with the presentation.
The presentation will open with a recap of the Customer
Engineer and a brief rundown on what he does. Next is a
description of the technical back-up for the CE for both
hardware and software followed by aids to the CEo
Following this is the discus sion of the customer - CE
relationship, what is required and what can be gained.
The subject of APAR's will next be discussed, touching on
the function of the CE and SDD's roll in APAR processing.
The last item to be covered is' problem determination.
Essentially, this covers who the customer sees when, he
has a problem.
o
---------~----------
-------------------------------
I
0
FIELD ENGINEERING
DIVISION
(CUSTOMER ENG INEERS)
o
WHAT DOES THE CE DO?
1. Physical Planning
· Environment
· Space
2. Installs
· Teams
· Shakedown
3. Production
(
,
"
· PM
· ECls
· Unscheduled Interruptions
· Program Support
o
CE Technical Support
o
I
SDD - SMD
I
Field Engine.ering
Techn ical Operations
Area Technical
Sup>port Staff
Designated Specialists
in Various Branch Offices
Branch Office
F. E. Special ists
o
AIDS TO CE
Education
Docu mentation
Hands on
Seminars
CAl
Self-Study
Classroom (formal)
Microfiche
Retain
Tools
Diagnostic Programs
Test Equ ipment
o
~?O
o
CUSTOMER - CE
RELATIONSHIP
o
:;. '71
CE MUST
KN OW OPERATI ON
KNOW WORKLOAD
KNOW CUSTOMER
o
HOW CAN YOU HELP THE CE DO HIS JOB?
1.
P. S. A. L
2. Trouble Log
3. P. M.
4. E. Co
c
o
273
APAR
0
A - Authorized
P - Program
A - Analys is
R - Report
CE
1. Identifies
2. Subm its
3. Bypasses - if possible
4.- Appl ies SOD suppl ied
temporary fixes (PTF's)
S00
1. Responds
2. Creates permanent fixes
3. Creates tempora ry fixes wh en
requ ired
4. Lends on-s ite assistance when
requ ired
C"
o
TYPE I PROGRAMS
INTERNAL
EXTERNAL
I,
I
Probl em Determ ination
What is it?
Problem Known
Hardware
Type I Program Internal
Contact
CE
Type I External
Other IBM Supplied Programs
User Program
contact
SE
t,
Unit Record Control Panel
Problem Unknown
SE is there
SE and CE are there
CE is there
Neither SE or CE is there
Contact
SE
Contact
CE
(0
~
SUPPORT R£SOURCES A'IAIl..ABl,E"
To youR CuSTOMER ~N~J~EER.
o.
PSM'$ ....
F,ELt)
4'"-------1
SDD
E N(;' NeE. tflRfS. ce.~ re:R.5
5UPPoRT'
CooRD,tll\T'oR
O.STRI6uTiON
t-----PRO,-RAM
SuPPOR'r
~~ouP
?r
f?EQusr
PTF'$
).'" ,",0 \J F(
Fe..
1=\
ss.tsm.....d:
SERi.CE
o
~77
I~
I
o
SESSION NUMBER
T.3.6
SPEAKERS
MR. G. WOLF, IBM EDUCATION DEVELOPMENT, SAN JOSE, CALIF.
MR. H. CADOW, IBM EDUCATION DEVELOPMENT, POUGHKEEPSIE, N.Y.
DISCUSSION
THE IBM EDUCATION EFFORT FOR THE 1130, 1800, AND 360 SYSTEMS
INCLUDING PROGRAMMED INSTRUCTION MANUALS, EDUCATION CENTER COURSES
AND THE CUSTOMER EXECUTIVE PROGRAM.
ATTENDANCE
39
-j
I
I
~
78
o
Presentation on 360 Operator Training
by Har17' Cadow
o
(1)
The purpose of this presentation is to describe the education program for
operations. In the context of this presentation, operations includes all
those non-management people whose responsibility it is to keep the wait
time of the 360 to a minimum.
o
This includes the traditional button
pusher through the lead operator or section chief.
(3)
Before getting into the discussion proper, let me state what the operations
education program does not solve, but what the user must certainly
consider.
These are the areas of Job Titles, salary ranges, career
paths, and personnel selection. Job titles as a consideration has always
been a frustrating topic in data processing.
As you'll see in a few
moments we've overcome the frustration by literally ignoring it.
Salary ranges is also a topic that generates lively discussion - "what
should you pay an operator?" is a tough question to answer because it
is entwined with the problem of job titles - just what is an operator?
Complicating the matter even further is the problem of career paths
for operations personnel. Some installations have regarded the position
of operator as a training ground for becoming a programmer if the
operator is "any good". It is somewhat difficult in these installations
to determine what ha,ppens to an operator if he doesn't make it to
programming. One SHARE member said that operators who don't
~
make to programming in his shop leave to go to another company for a
f\
considerable raise in pay.
A few installations have made considerable
effort to develop career paths within operations. One company has a
series of five positions, anyone of which an employee in computer
operations can aspire to.
The promotion from one level to another is
I.
(~
based upon several measurable items. 1) Classes attended and marks
received, 2) Length of time in the position, and 3) Management evaluation.
o
The last consideration concerns personnel selection o "How does one go about
choosing an applicant for a pOSition in data processing operations?" At
present there are a number of tests under development for assisting an
operation's manager to help choose a potentially successful operator.
Now back to why job titles are more or less ignored.
The main reason
is that as a guide line for IBM in developing operations courses, very
few users could agree on the description of "operator". Hence, we could
not use "job title" description as an aid in developing courses and texts.
(This problem existed not only with the job title of operator, but also with
such other data proceSSing areas as programming and systems analysis. )
Therefore, in the area of operations, we listed the tasks that must be
C.,
,)
performed.
,
(2)
Here we see five operation tasks.
The first one concerns dellce handling.
This includes such activities as obtaining and returning tape reels and
disk packs to the data bank or library, mounting and dismounting tapes
and disks, putting the correct cards in card readers and punches, and
putting paper and proper carriage control tape into the printer.
The second
task has to do with responding to operational messages printed on the
console typewriter and entering commands to the supervisor program
via the console typewriter. When I say "entering commands" I do not
envision the console operator deciding on the spur of the moment what
•
command to issue. Rather, I see the console operator referring to a
II
well documente d run sheet with a list of all the commands he may make and the
I
I
conditions under which he may make them.
The third task has to do with system updates.
This includes activities from the
o
addition or replacement of modules to systems libraries, to a complete system
generation.
The fourth task has to do with a very complex operation, thatGf identifying the
cause of a problem. Identifying the cause of a problem can mean doing a
complete diagnosis, recommending and effecting the fix, or, it can mean
calling either the systems programmer, IBM SE or IDM CEo
The task of
•
system restart is also an operation of varying implecation.
It could simply
mean re-IPL-ing and putting the job control cards of the errant job at the
beginning
ot the input stream,
or it could mean first running a series of
utility programs to restore the system to a checkpoint condition and then
re-IPL-ing. Running the utility programs is not t<:»difficult - it is choosing
the appropriate ones and preparing the proper control cards that requires
systems know-how.
(2a)
These five tasks lend themselves to categorization as to personnel
requirement - production people or analytical people. It's not that the
production people aren't analytically minded or that the analytical people
don't produce - it's more that the production people work from a recipe and
aren't required to make spur of the moment decisions, and that the
analytical personnel must deal with unusual and unpredicted events.
j
i
-1
3.
c
(2b)
The people requirements of production and analytical lend themselves
very nicely to functions of what has been called System Operation and
o
Operation Control. The tasks of device handling, responding to messages and
issuing commands, and updating the system come under the function of
System Operation.
The tasks of system restart and identifying the causes
of problems are the function of personnel in operation control. Also,
operation control is partially responsible for system update, viz., laying
out the systems residence pack and selecting the modules for a system
generation.
Now let's see what IBM has in the way of educational materials and courses
to assist you.
(17)
First of all, IBM has released within the past 4 weeks, four System Operation
Student Texts to be used for self study. The student text packages consist
of a text, a book of illustrations, and a hands-on exercise.
There is one
student text for DOS using the System/360 Model 40, one for DOS with
the Model 30, one for TOS with the Model 30, and one for any 360 model
greater than a 30 for
as.
Each of the four student texts assumes some basic knowledge.
This basic knowledge is primarily concerned with some very
simple data processing concepts. One of the ways the aspiring systems
operation student could have acquired these concepts is by having worked
in, or been close to, a 360 data processing installation. Some of basic
•
concepts and knowledge are:
1.
The relationship between input/processing/output
2.
Hexadecimal numbering system
3.
Decimal to Hex (and vica versa) conversion
4.
Function and description cf I/O devices and media
5.
Purpose and function of Channels
6.
Relationship between compiler, source program, and object
o
program
7.
(18)
Importance of documenting the jobs to be run
What if the systems operation trainee doesn't have these concepts and
knowledge in his repertOire? The answer is, "He enrolls in the programmed
instruction course Computing Systems Fundamentals or CSF for short. "
CSF is a major revision of the PI course called BCS which was released
for use in 1964. CSF is concerned only with system/3'60 topics whereas
BCS addressed tp.e entire IBM computer line manufactured in the early
nineteen sixties. BCS you may recall, required somewhere between 30 and
40 hours for the student to finish. On the other hand the entire CSF course
requires i5- 20 hours.
(19)
Whereas the entire CSF course consists of six units, the required prerequisites for a systems operation trainee are Units I, IV, V, and VI.
Unit I provides the student with a basic knowledge of problem solution,
numbering systems, and conversion, computing devices, and 110 media.
Flowcharts, decision tables, programming principles, and programming
languages are discussed in an introductory manner.
C,:
;'
Unit IV discusses the process of creating a program from previously prepared
decision tables and flowcharts.
Short sample problems are used to show the
development of programs in COBOL, FORTRAN, RPG, and PL/I.
Also
presented is a brief introouction to program compilation. Unit V discusses how
the computing system (CPU, storage, I/O devices) solves a data processing
problem. Information is traced through input devices, proc,'essing it, and
finally as it emerges from output units. Numbering systems, I/O devices,
storage devices, and processing units are illustrated and discussed.
Unit VI is devoted to the procedures for collecting and packaging the
documentation created during the programming and testing of a data processing
problem, and the documentation required for operations.
The amount of time required for a student to finish these four units varies
C'
(10)
from six to 14 hours.
You've probably been wondering about Units II and III: Unit II concerns
"D efining a Problem" and Unit III is about Analyzing a problem. Although
these two units are primarily targeted for a programmer or systems analyst
trainee and not absolutely essential for a systems operation student, we
recommend that if time permits, the student take these two units.
elaborate a little, Unit II concerns problem definiti. on in detail.
To
The student
is shown methods for determining input and output data requirements as well
as the calculations required to solve his data processing problem. Sample
I/O documents are discussed throughout the text.
Unit III is about the use
of decision tables and flowcharts.
Sample problems are analyzed by means
of decision tables and are flowcharted using the standard flowcharting symbolS.
We in IBM Education appreciate your concern with conservation of time.
That is why Units II and ITI are not mandatory - - - but we do strongly
recommend them. If the systems operation student does take these two
units, he will have a better understanding of the steps involved in producing
those computer programs for which he'11 be responsible in his role as an
operator.
(21)
Now let's take a look at the systems operation student texts.
The three
texts for the Models 30 and 40 DOS and Model 30 TOS are identical in
organization.
Section 1 is an introduction to the System/360 computer.
It can be classified as a review of CSF. Section 2 describes the role of
the operator. It also very appropriately tells what the role of an operator
,:-'f(' -'"
!,,,,"__ Jl
is not. One other major topic of Section 2 concerns a description of the Programs
to be run and the Run Book. Section 3, "Sample Programs" consists of
detailed step-by-step procedures for getting the components of a system
operational.
A description of sample programs on sample run sheets are
used to give cohesiveness to the step-by- step procedures.
Section IV is
devoted to Dos/TOS terminology and how the operator communicates with
the supervisor. Included in communications is a discussion on 110 device
assignments.
Also discussed in Section 4 are numerous messages that the
operator may receive and what actions might be taken on these messages.
~~,
'-J1
I
I
I
7,
After the student has finished the text, he should perform the appropriate
systems exercise.
o
This exercise will provide directed experience in
operating I/O devices, interpreting operator messages, and issuing
commands.
(22)
The exercise requires from 1/2 hour to 1 hour.
For Systems Operation Model 40 OS, the organization of the text is
somewhat different.
Section I is an introduction to the operating system -
specifically, the function of the job scheduler. When the student has
finished this section he is able to define such terms as Reader/Interpreter,
Initiator/TerJ?1inator, data sets, Volume Table of Contents,
DASDI, Job
\
Control Cards, data set disposition, and, IPL. Section II which is entitled
hardware familiarization is basically the same as "Sample Programs" for
DOS. Section III "System Exercise" describes the OS Inessage format,
lists the publications "Messages and Completion Codes" and "OS
""1\
'-.Y)
Operator's Guide" and how they are to be used, and, a description of the
exercise. One half to one hour of time is required for the student to run the
exercise.
(23)
Now let's turn our attention to the courses recommended for the personnel
who perform the operation control tasks.
The initial course the student takes may be one or more of three:
0,·,',
I,
1.
CS F (which has already been described)
2.
S/360 Introduction
3.
Fundamentals of Programming Languages
S/360 Introduction is a 5 day class in the education center.
The maj or
objective is to have the student able to describe the functional characteristics
and general principles of operation of the S/360.
Fundamentals of Programming Languages is also a 5 day course and is for
o
people with no data processing experience. It is primarily a CSF review
and a description of the various programming languages with analyses of
programs written in the different languages. f
If the installation is to transmit and receive data via terminals and common
carrier transmission facilities, it is recommended that the operation
control student attend the 1 day Data Communications Concepts course.
hoc..l
Here he will find out"data communications can effect the operation of the
data processing installation.
The Problem Oriented languages are COBOL, PL/I, FORTRAN, and RPG.
Some of
~ese
courses are currently available in both classroom form and
PI form, viz. FORTRAN and COBOL. In general, RPG is available only
c
in PI form. Although PL/I is currently being taught in education' centers
only, it will be available in PI form within the next 6 months or so.
The Assembler language coding course is also available in either form.
For those students who study ALP in PI form, there is a 3 day hands-on
wrap up class available at the education center.
The facilities class teaches the student how the operating system functions
and what facilities, in the form of pre-programmed routines, are available.
This includes the selection and use of utilities and the requirements for using
various levels of Access Methods. The DOS/TOS Facilities class lasts
for 3 days whereas the OS version is scheduled for 5 days.
-.J
!
:CI
.>-"'jIi
T'
The System Generation (or Sysgen) course covers the considerations for
generating a system when provided with the following items:
o
1.
Released tapes from PID
2.
Sysgen manuals
3.
The machine configuration and operating system options or
features to be included.
The last item is sort of misplaced. It belongs of course just after 8/360
introduction. This is where it will be placed in the futureo
c
o
ID.
2¢¥Q1lM. ti
: ."
4iji44iIij $
,JUUCU4IAlm,: a.ea;n4PM .tutU
.... _.................,...,., ...........:..."=
... ,.""""""""
..._ '
------~~~-
.. .. =..........-... -
.--...............--.~~~
ii"
1'1
I,~
II
o
en
~
z
0
0
-~~
~
p
~===:I
U
t~-n
a=
w
0
en
c::::::;::a
en
W
...J
~
.
0
U
~
(f)
;
'
I==-
CICI:ZII8
~~
(ITO
0
~
~
~
~
c§l
~
©:c:
00
X
t~
~
@-"w:~
rtf)
C
cmcr~
©1:
W
~
~
="'::;;?
~
d/
,.;::;:::0
>~
Q1
c=B
~
00
:~-;"".:
~~n~q
Be:
cg:
.~
U
c~
<0
"0/ '.
ef})
6d~
~_.fL~
~
c'
~90
e
o
o
OS OPERATION
---:::~.
TASK
REQUIREMENT
DEVICE HANDLING
PRODUCTION
MESSAGES
)..).
~
a
COMrv1ANDS
PRODUCTION
SYSTEM UPDATES
PRODUC1"-ION
Ar·,jALYTICAL
IDE~JTIFY
Af\J;l\LYTICAL
RFSTART SYSTEM
AN~.LYTICJ1~L
CAUSE
OF PROBLEM
'-
2
2~
FllNeTION
SYSTEM
OPERl\TIOr~
OPERATION CON1&ROL.
2b
__
L
__
SYSTEM OPERATION TRAINING MANUALS
TEXT
ILLUSTRATIO~~
E){ERCISE - ~~AN DS ON
MODEL 40 DOS
(light) BLUE COVER
RED COVER
..
-;,J
0 'DOS
.
l"/ rr·ODEL
/"'I
;\.iU\~
L~.·
u
~
MOD~
30 lOS
p/tJY 360
~v10DEL> 30
(dar~tl BLUE
as
COVER
GOLD COVER
1.7
~
C1
o
-
~~~--
dl
o
•
~
~
~
o
I
I'
["
I
.....
r,~,-"'-'::-'-:;;'
:::;:~~
~J
'j
c:.:::"",..;:;.)
W
rl
~
0
00
0
I""";
if)
~Jt...
DC
W
<:
0
b,.~~
~~
J
'i.
c::c:::-
()=~=-
(:?
~?
~
~
~-,
..:;-:-:.;.,;:a
-dl~
$~
~
2'-
=>
0
to;:
'-
~j
0_
-'
~
~;J
--p
(!)
(!)
0
>
===.J
D-
en
0::
DC
.~
ll~",;
~:~
~'..i_._ ......~.I'
:':"::~~~
("'';'''')
~
~
~
~
~
~
(j";;J
l'
c---;.
Z
W
r:~
\i~tj
n
~.:dJ
0
ce:(
~
U
~
~?
t,r.._J
--
~
0
Di:)
">:..,~.~",.
c::r~
p.
f!
-=-:Jl
. h___ ,
L~w
'j
P:=
tl
G;:.',~
~
~
-,
2
-=--
0
r~~
,~u
~
Z
-~
~>.
--P
6~
~=-
~>
c:.:-~
=>
~.J
0
C
r:,--!
,:~~
I:
--
~~u
p==-
II
6--""
-::::;p
~
:::)
o
o
.~
o
~
COMPUTING SYSTEMS FUNDAMENTpJ=S
<:.'
..
",...
.
",.·L~_
.-"",,::,:.:o=.=u
RECOw1MENDED IF TIME PERMITS
U~jIT
~
TI
UNIT TII
DEFIN~NG
A PROBLE0J]
p~NALYZING
A PROBLE~v~
~
\'l
20
SYSTEMS OPERATION
Gr.;•. ::'
MODELS 30 8 40 DOS
8!&&:5!"'L_.fr.:!El!!ll
SECTIO~
I
,"""""""',,,."'.·.c:;.C'DIIII_ _ _Bml_ _Gl3_ _ _ _ _ _!&:II_ _ _ _ _ _ _ __
INTR'O TO SYSTEM/360 COMPUTER
tl
I:
~.,. F:.~.c:?nO})~J
~bV
~ U 1!~J
fl.·'
e,
r
t[
ROLE OF THE OPERATOR
i:
I'
(~ r~(G/.i,C""?
': ,;,,:, \
1.1
U
~) l..:.= ",
IO'{J\j
I;:,'
~
c;]
"i\
V
SA~v1 PlE PROGRA~J1,S
"->
'"
~
_
0
(('"$ ~.="~.' C7-" ~j "
~Ll)~
ut::3Y
~
f'~
i '
S\fSTErv~
ll.
{.il
.~~
DIS~(
OPERATI NG SYSTEM
E)(ERC!SE
21
~
\
'\.\~
()
r'~
>/
~
..L
~
~
_ ..
_~
__ _
o
e
o
SYSTEMS OPERATION
~JJODEL
~
40 OS
SECTION I
INTRO TO OPERATING
SECTION ]I
HARD~4JARE FA~~iLIARIZATIO~J
SECTION JJI
SYSTEM E){ERCiSE
SYSTE~v]
11
.-~------
I--~ --~
--~--
@
-- --ij-
0-
c=-j
~n
h'
Ij
~rj
Oll
n
.
~~
, ::=:::l:)
U
i.
~r)
lOCI
n n
11
l.':;;;;';'; :1
H
,"';;1
L.:::._.~.:
... ,; j
,,~~\
ii
\..J J
C:;?
~, ..2
()
L:=a
~
~
~
~
,-1
1
ti'2
0
<:::3
c:.<~
0"
9
l_ .. _:...1
r",:z3I
0
(~~
n
~
d
c~~
-;':'/,
L==-
0
-)
0
~
0
~
~
rl'<'?
~
t»,.
r,
vii
0
• ,_;;;;a
, '.,,\
\,
..
;1
j
::uJ
(..
".,..
<0
@...
'-
l!...
~J
U
......
cp-'')
C\
. ""1
S ES S I 0 1\1
(\1 UMRF. R
T. 4 • 1
OPEN ROARD MEETING
SPEAKERS
JAr"1ES STAf\JSBIJRY, PRFSIDENT
C;
,"
TRANS CRIPTION
c
OPEN BOARD MEETING
COMMON
-
7 SEPTEMBER, 1967
The transcription below indicates that we have much to learn
about procedures for taped sessions.. I think that the idea is
good; it's the only way to get the verbal information at these
meetings into the proceedings
II
Many of the questions from the floor could not be understood
(on the tape); comments were generally clear I since most of
those used the floqr microphones. When I could understand
or remember the gist of the question I I have put it in; in some
cases I have reconstructed the question from the reply. In a
few places I I have been forced to delete passages; the context
was not clear.
I
In the future I we will have the floor mikes checked before we
start, to eliminate the feed back that was present at Cincinnati.
The rest is up tQ you. Name I rank I and serial number I please I
and
use the mikes.
The trans cription follows:
loO
Page 1
CHAIR This is not exactly an Executive Board Meeting I in the sense that
we don't vote on business I but we do have the Executive Board up
here as targets. We also have some things that we Id like your
opinions on. Some feed back from the members - and itls a
chance for you to bring up any gripes or complaints I pats on the
back or anything else that you think may be appropriate. lid like
to say before I start that there is some trouble with the mikes.
They must be pushed to talk There is a switch on each one of
them. Turn on the switch whenever you have anything to say.
Therels a report that I will present tomorrow at the Contributed
Program Library Session on the 360 Contributed Program Library.
This is a joint COMMON, SHARE GUIDE proposition. The minimum
standards that have been distributed by PID do not repre'sent
COMMONIS standards. They are minimal standards acceptable to
all user groups We Ire free to exceed them in anyway we want to.
Just not contradict them.
It
e"
Second item of business that I have - COMMON has been asked to
participate in a Contributed Program Library Catalogue sponsored by
the Joint User's Group and ACM. The Executive Board is in favor
of this. What it means is that the programs in the library of any
participating group would be part of a composite catalogue of all
User Group Program Libraries I regardless of the manufacturer and
machine affiliation. At present the ACM is talking about a subscription price of some $3 to $5 a year for a quarterly catalogue.
You don't have to subscribe if COMMON participates - that's up
to you. But I would like to know how you feel about our making
the COMMON Library Catalogue a part of this composite Contributed
Program Library catalogue First I is there anyone who wants any
further discussion, explanation of how it works I or anything of that
sort? Have you seen anything on it I heard anything about it anything else. Comments from the floor please.
It
I
QUESTION - What would be published?
Titles I authors -
ANSWER - Titles I authors 1 abstracts - any program entered in that
catalogue would be available to anyone regardless of
user group affiliation.
o
Page 2
QUESTION '- How available?
ANSWER - On order - the orders would be handled through your
User Group, so in practice they are restricting it to
members of organized us er groups I but it would be
available. There was a stipulation that a reasonable
reproduction cost could be charged.
The general idea frankly, is that COMMON IS Library, along with
those of several other user groups I is winding up in other catalogues,
anyway. Such things as COSMIC and some of the others. We feel
that we wouldn't be losing anything by making our library entries
available. We do feel that we might be gaining something in having
the information about the other User Group libraries. This would
not affect the COMMON Library catalogues in any sense. The only
additional thing that would be required is that someon~ who enters a
program under these circumstances would have to fill out two submission
blanks. One would go to PID, the other, without the program materials,
would go to the Contributed Program Library liaison group committee.
We are not promising to retro-fit existing libraries. If you don It
want a formal vote on it, we don It need it but I would like to have
some opinion as to what you think of the idea. In the back, please
and you may pave to use the mike I
l
QUESTION - How do SHARE and GUIDE feel about this?
ANSWER - At the Executive Board level, SHARE and GUIDE are both
in favor of this. I don It know if they have had a resolution, a vote
on it, because I frankly, we do not have a response from IBM yet
as to their willingness to participate. They promised such a response
by I hopefully I the first of October. That's the reason I am asking
for the feed back at this time.
FLOOR I think I could express the opinion of many of us here that this idea
could be considered twofold - one I we have a centralized way of
obtaining information and, second, with this type of catalogue I and
the whole combination of papers being presented in one place we
could all have these papers available when we enlarge, letls say,
to different systems. Welre a 1620 user now, and have a 360 on
order. 11m sure that many of you will express the same opinion.
I
Page 3
11m throughly in favor of it.
I would like to add one comment to that.
On the Contributed Program Library I COMMON I S Contributed Program Library I there will
be separate catalogues for the 360 1 the 1800 and 1130 I and the
1620. A user will receive the Library catalogue corresponding to
his machine type. The 1130 and 1800 catalogues are combined. T
The official rule is that the user will receive only his own machine
type catalogue I so you normally wouldn It know if you were a 1130
user what was in the 360 library without some such procedure as
this. Incidentally I you can get the catalogue for another machine
type by convincing your branch office that you need it ..
QUESTION - Wouldnlt the submitter have to sign a disclaimer?
ANSWER - They do have to sign a disclaimer when they send it in.
I mean now.
QUESTION - Just from a lega,""point of view - Would the individual
whose programs are now in the library have to be consulted on this I before we make any committment?
ANSWER - They had to sign a disclaimer when they sent it in. It
has always been required.
FLOOR The disclaimer gives any third party the right to reproduce the program
and give it to anyone they desire.
CHAIR AND I might add that IBM does exactly that. Apparently t they are
legally obligated to not use the Program Library as a means of sales
advantage. As a consequence I any requester can obtain a copy of
any program in our library I regardless of whether or not he1s even
using an IBM machine. These are the present rules •
•
Page 4
May I ask for a vote - a show of hands - as to those who would
agree that we should participate. Very good. Thank you.
There was to be some discussion of our proposed by-laws. How
many people have read them enough to comment? They are published
in CAST 7 which about half of the people received, and there were
additional copies made available here.
I
,I
PLOORNo time to read them.
I know it, but this is the last chance weill have for discussion.
Would you like a run-down from Laura as to the general changes
we Ire making?
IAURA AUSTIN I think generally you III find that there are relatively few changes except for wording. Jill just briefly go through them here, and see if
I can pick them out. I wasnlt prepared to do this I and I haven1t a
marked copy here.
We did change it to include the new machine types. Welve changed
the name throughout the py-Iaws from 1620 User's Group to COMMON.
Welve included the machine types that have since been made eligible
for members hip in COMMON. The purpos e and framework of the
organization remains the same as it was. In loss of membership,
there has been a change, because we Ire doing away with the regional
structure ~ Where we used to have two meetings in each region each
year I we voted to cut it down to three meetings a year. So we Ive
had to change our requirements for attendance at meetings. If an
installation is unrepresented at four consecutive meetings you will
be removed from membership. This should give you a period of
about 18 months to make a meeting and still maintain your membership.
We've added one other stipulation, so that we can keep our mailing
list up-to-date. Each year, the Secretary-Treasurer will send out
a registration form which registers your correct address I your machine
type l and so on. If this is unreturned after two consecutive mailings I
(we mail it I and if we don It get any answer, we try again) then you
(/0
{(~
I
Page 5
will be removed from the rolls. Now I on ~hat mailing I I shouldn It
say weill mail again if we don't get any answer. If the m.ailing is
unreturned, we will mail again to the installation, without an
individual addressee t because the installation maintains the membership, not the individuaL Also, any installation failing to respond
to two consecutive mail ballots will be removed from the membership.
We Ive provided for abstentions to be counted as voting. If you
wish to abstain, donlt just tl1row your ballot in the waste basket. Return it, because you will be counted as having returned a ballot I
eve~ though you are abstaining from the voting.
On the elective officers, there will no longer be regional Presidents I
but they will be Executive Board Members at large. The Executive
Vice-President will be elected from the Board by the Board members,
but the President I Secretary-Treasurer I and the Board Members will
be nominated from and by the membership at large. In the elections I
half of the Elective Board will be nominated every odd numbered year or every year half of the Board Members will be elected I so we will
not have a complete turn-over at anyone time. That means that we
will be having elections every year I instead of just every two years
as now. Half of the Board members will be elected every year. The
President and Secretary-Treasurer will be elected for a two-year term
of office I as will the Board Members. The election for the President
and Secretary-Treasurer will only be every other year I but there
will be Board Members elected every year.
There is some change in the wording on the Teller I so we wonlt have
more than one Teller from anyone installation. This was not restridted in the former by-laws.
On a q\lorum, welve had a difficult time in the past. You may remember
in New Orleans I for instance I where we had to declare the people
elected because there wasn It even a .,uorum voting for the people I
so we didn't in reality have an election. We are reducing what is required for a quorum. So welve said that, for formal business at a
meeting I at least one quarter of the member installations must have
voting representation at the meeting to constitute a quorum. Then,
for conducting business by mail, at least one-third installations must
reply to amke up a quorum l and as I mentioned, abstention shall be
counted. We hope it won't be so difficult to get a quorum anymore
•
Page 6
on our voting.
The decisions by committees have remained essentially as before I
about proposed standards on meetings - this will remain the same.
We have changed the by-laws to state three meetings a year that
they will be distributed geographically and temporally. This means
essentially we will have a meeting in the East, Mid-West, West,
Mid-West, East, Mid-West, W-est, so we will hit the Mid-West
twice as often as the others. We felt that this allows the East
to get to both theirs and the Mid-West, and the West to get to
theirs and the Mid-West.. So this the plan for coming meetings.
I
I
Welve also changed, under distribution of information, that any
material, except the membership list, which is distributed through
any COMMON channels, is assumed to be non-proprietary. We did
feel that the membership list should be considered proprietary to
COMMON,. It will not be sent out just on request, because we've
had difficulty with people putting COMMON members on mailing lists,
so we have restricted its use. Of course, in submitting a program
as welvealready mentioned, you submit a disclaimer at the time,
which does make it open for any kind of reproduction
I
II
You'll note that incorporated in the new by-laws is the proposal
that was submitted last year regarding financing for Executive Board
Members who must attend every meeting during the year. There will be
some subsidization made by COMMON I if necessary, to have a Board
Member in attendance.
We have also left it open tnat, if it is felt necessary in the future,
we might put in a subscription feed for CAST I if the cost becomes
prohibitive under our normal financing situation. As you know I the
only way we receive finances for COMMON right now is through
the registration at the meetings. We take the cost of the meeting,
and then some additional to cover the expenses of COMMON The
issues of CAST that we have sent out have run us in the neighborhood
of $1200 per issue of CAST. You can see that We are running into
considerable expenses in operating COMMON I but we have felt I
from the feed back we Ive gotten from the membership, that CAST is
well worth this ~ If we are to continue to issue CAST on a regular
basis, and to include all the information we have in the past I we are
q
c
Page 7
o
going to have to have some means of financing this. So we Ive
left it open in the by-laws that if deemed neces sary we might
put in a subscription fee. However I this would be submitted to
the membership for consideration before it was done. It's not
going to be anything that is just railroaded through ~ I think that
this basically covers the changes in the by-laws. Do any of the
other Board Members think of any I have skipped over?
I
CHAIR Thank you
I
I
Laura.
I would merely like to say one thing. I hope that the members see
fit to approve these by-laws I because I don It want to spend the rest
of my life running from the sheriff. Obviously r am approving
expenses of Executive Board Members to get them to these meetings;
11m not authorized to spend your money this way, but I am dOing so.
I
C. \'
There was some question at Boston about a statement of finances.
Chuck had to return to Dallas unexpectedly. Our present treasury
is something like $4750. We will make in the neighborhood of $2500
from this meeting. As far as the Executive Board is concerned I we Ive
already stipulated that the registration fee for San Francisco will be
$25 I and you will get only one luncheon out of it.
II',
The feeling at Boston was that higher registration feed would be acceptto the membership in preference to dues, and that a subscription
fee for CAS T would be acceptable in lieu of dues
They did not want
clues inlieu of a meeting attendance requirement. As a comparis on,
SHARE has considered imposing a subscription fee for SSD; but the
suggestion was voted down by the members. Every thing that they
get from a registration fee is net - There are no group activities that
are scheduled without charge. SCIDS registration is additional,
and on a usage basis. SKIDS is, a SHARE organized drinking session.
Six 'til two every night during the meeting. res very well patronized.
ab~e
it
In regard to the quorum, we felt that abstentions were legitimate votes
for COMMON members. There are many of you new members who I
for one reason or another I feel that you aren't qualified to make a
decision, but are willing to go along with the majority~ .. Now, we've
asked you to deliberately abstain. The vote on the by-law amendment
that we proposed earlier was very much in favor - I think something
like 80% - but there were not enough votes returned to establish a
"} Ot.
Page 8
quorum, so it was defeated. Welre trying to get around that,
but we didn't go as far as SHARE. In SHARE, if 20 ballots are
returned from a mail vote, the majority of the votes' determines
whether the thing passes, because every installation has an
opportunity to consider a mail ballot. I don't think that's quite
legitimate. We didn It go that far. Are there any comments from
the floor?
QUESTION - How many is 25% of the membership?
ANSWER - At the moment our membership is something like 1275.
Weare not a quorum here.
QUESTION - Is the omission of machine requirements deliberate?
ANSWER - It was deliberate. Wetll accept anything from a 20 to
a 91.
QUESTION - I would like to ask one question. Is there any provision
in the by-laws now for bonding of either the SecretaryTreasurer or the Local Events Chairman? I know that
Mr. Mauldin is undoubtedly a very honest person. Dave
Dunsmore is undoubtedly a very honest person, but to
protect them and COMMON I would hope there would be
saIne way of providing for a bonding of them ot of anyone
who must handle organization funds. I don It think there
would be an organization that could operate effectively
without some sort of a bonding procedure.
ANS\VER - I think you have a good point here on the Local Arrangements
Chairman, and I know that Dave himself, raised this question.
It does not state in the by-laws, and perhaps should, about
the Local Arrangements Chairman, but it does about the
Secretary-Treasurer. It says he should be the primary financial officer of the group I and, as such, will be required
to be bonded at the expense of COMMON.. I might add
that Chuck Mauldin has been requested by the Executive Board
to have himself bonded at the expense of COMMON for
his own protection. Further comments -.
l
Page 9
o
COMMENT - I think it has been rather the concensus of the
people at this meeting that the by-laws as amended
are going to be quite acceptable to the membership.
I realize that it will take a mail ballot to insure this I
but I think we could give you a s how of hands here to
indicate our approval
CHAIR Well , since the majority of you have not had a chance to read them,
or at least so indicated t I don't think it would be too appropriate at
this time.
IN THE BACK We don It need a show of hands,. What we mean is just five minutes
of people's time to take the time to mail the ballot.
CHAIR That is exactly right. rve forgotten exactly how many I but something
like 250 ballots were returned on that by-law amenpment. As I said,
it was about 80% in favor of it, but there were not enough under our
existing by-laws to establish a quorum.
QUESTION - Can you give some indication as to when the ballot will
be mailed and how long they will have to return it.
ANSWER - Yes" The amended by-laws I the proposed by-laws, were
published in CAST 7. About half of YOU about 40% of you, I
think, received them before you came to this meeting. The
rest of you should find them when you get back. Under the
by-laws there is a required discussion period - under the
existing by-laws and under the proposed by-laws. We have
to submit it to you, and we have to give you time to send
in written discussion, if you wish, prior to requesting a
vote. That written discussion and the ballot, together with
Executive Board comments I will be published in CAST 8,
which is scheduled for sometime around October 1 or October
15. You then have approximately a month to get your ballots
back. Comments should come in very promptly, please I
and we would appreciate favorable or unfavorable. Wetre
most interested in some idea of what you want.
I
o
Page 10
COMMENT - I HAVE read the by-laws. They look good to mel but l
from what you said a moment ago, I would suggest consideration of the question of the size of the quorum
neces s ary to conduct bus ines sat a meeting. If I in fact I
a group of this size is not competent enough to make
decisionsblnding upon the whole, I think we have a much
too conservative position expressed, and it might be
appropriate to lower the quorum when we do have a chance
to discuss back and forth, and each one adequately hear
the opinion of everyone else. Now I secondly I I take exception to your interpretation of what I believe to be the
argument presented at Boston. Ido not believe there was
a strong opposition to the payment of dues from this membership if it was rewarded by adequate, efficient administration.
o
Thirdly I I say that, maybe because of what we said,
planning and the conduct of this meeting has been vastly
superior I and I say thanks to whoever it was.
CHAIR To some extent I can take credit for that.
(LAUGHTER)
No I this is the reas on for my guilty conscience for subsidizing the
attendance of sufficient Executive Board members to establish a quorum
of the Executive Board at the planning meetings we had for this.. I
don't mean that I can take credit for the work. Work has been Dave
Dunsmore and Jim Tunney I Laura Austin, Division Chairman, Managers I
Eric Zielinski of IBM. They are the people who must really take credit,
but I think the Executive Board contributed a great deal to this, and I
would like to have my guilty conscience saf.ved. (Clapping).
We do hope and expect a better run meeting at San Francisco than we
had here. We are aware of the problem with the late agenda. They were
mailed on time. The next ones are going to have to go out first class,
which means more money. It is gOing to cost us about $1,000 for a
first mailing about the 1st of October, a preliminary agenda about the
1st of November. That's the reason for the deadlines that I mentioned
at lunch.
c
Page 11
o
IN THE BACK Do you know hovl many installations we have represented here?
How many people we have here?
ANSWER - The people are approximately 450. The number of installations,
however, is by no means anything like that. There are around 50
attendees from IBM, and I know of several installations that have
multiple attendees. I couldn It begin to guess I but somewhere in the
neighborhood of 200 or 250 installations.
FROM THE BACK Couldn't hear----------------CHAIR I would like to have that in writing, please. To CAST.
DICK PRATT rm not sure, maybe somebody can correct me about this if 11m wrong
but I believe the only way to get this provision in the proposed bylaws is to defeat them, and then propose a different set of by-laws
with this change in it. If this is true I and I think it is, I would like
to suggest to you that if this is the only objection you have, and
if your other obj ections are simi~arly minor you approve the bylaws as they stand, and then propose these things as amendments,
because we absolutely cannot operate legally under the present bylaws.
I
I
I
CHAIR Before we even comment on this - under our present by-laws we have
to have two-thirds to approve these amended by-laws. If we get
that done I then we can conduct business on a reasonable basis. That's
one advantage.
FROM THE FLOOR - SAM LYNCH I U. S. PRODUCTION CO. -
o
I agree with Dick I that if this is the only thing that prevents us from
getting these by-laws, for crying out loud, letts get these, and then
work on them.. But how can we find out right now whether Dick lS
assumption is correct? You know, if it is I then I wonlt even bother
writing CAST and making the suggestion, and 11m sure the gentlemen
back there won It either I but can we find out now what the legality is?
Page 12
CHAIR Chuck Mauldin knows I and hels not here. (After I got home. It1s
two thirds of the membership in favor I to amend the by-laws) •
, In any case I from our existing knowledge I I think we I d better do it.
DICK PRATT -
The point ls, I don It thlnk tbere ls anyway, either under the existing
by-laws or under the proposed by-laws I to amend a proposal to be
voted on by mail. You either accept it, or defeat it as it ls stated.
I think thaCs correct.
CHAIR -
I believe that one of the alternatives in mail discussion is to submit an
alternative proposal, but then that must be discussed, so it would
definitely delay it for two to three to four issues of CAST, which is
sometime next Spring.
FROM THE FLOOR I don't want to delay this, but, Dick, you imply that the restriction
is on the mail ballot. You know I this thing could be changed before
it is mailed to us •
CHAIR It has already been mailed.
FROM THE FLOOR -
Not in the form of a ballot - in the form simply of read this,
CHAIR This form has been mailed - must be discussed I and then the ballot
submitted.
FROM THE FLOOR -
O. K. In other words I the procedure has started I and we should not
abort them I but carry through. I would certainly go along. I withdraw
my comment for now.
o
Page 13
o
FROM THE FLOOR - WADE NORTON,. fro.m Southern Service, Birmingham.
I would like to ask the question if there is any way in which the
proposed by-laws are not as good as the existing by-laws?
Whether we have reached an ultimate is not necessarily the question.
Whether we have ever reached an ultimate js subject to some question
in my mind ISO would it not be better to go ahead and make progres s
in increments as we have been doing since we got our first computer.
(C lapping of Hands)
FROM THE FLOOR I have intermittenly received copies of CAST and various letters from
COMMON. If we are going to mail out a mail ballot, let's please
make sure that a1l the members receive a copy.. Myself for one would
like to receive a copy.
CHAIR -
c
All right, I can speak for Chuck on this. As you know, he has had
some trouble with his installation. Hets sti1l having trouble with his
installation, but not from lack of support. He informs 'me that he has
reduced his back log of mail from yea high, about four feed, to around
three inches and that, within roughly two weeks I there wi1l be a membership list complete with addresses I which includes every application
that he has. I will not guarantee that there is no incorrect information
in there, of course.. There should be no problem with people registering
or changing address I and not having it processed within a reasonable
length of time.
LAURA AUSTIN lid like to say that in the new members I meeting yesterday, some came
up and said they had not~received their mail. At that time I Chuck
checked the list which he has I which is current as of last Friday, and
all but one of the people's names were on there. We hope that things
are in m\1ch better shape right now than they were a week ago.
CHAIR My reason for giving the two weeks delay was that Chuck also said
that he has 75 new memberts names in Dallas I which are going in
immediately to Pat Lonergan for the final up-dating on their membership
list.
..,,'. ,t'
'"nNW
''t ......¥
' n C!!*''t "
··r:·'M¥I·'trwt/r,·¥!re±t"·r"2'··s··!;r"~M··~·"'"t···"····~····t"'~····'·'·~~·~···~'~· ... ····-···......
"··'ir~~~M"""'·-··--··=····-""""·~~~~............-
Page 14
FROM THE FLOOR -
I'm just wondering - I had problems here with Chuck, and it is
possible when you have something llkea ballot that it could be
flagged in some way so that you know it needs immediate attention,
or urgent attenti"on, and not be put in w;itheverything else in CAST.
(Sorry - everyone talked atonse here)
c
CHAIR -
Well, we assume that you would at least look over the table of contents. We do try to segregate such items, as a matter of fact, generally
not the front page, but we III try to make it the back page I normally
11
FROM THE FLOOR -
Different colored paper.
CHAIR-
Different colored paper is a. suggestion~ I don't know whether we
can do it or not, but we can certainly consider it.
FROM THE FLOOR -
If it were just typed in upper case in the table of contents - that would
tend to catch someone1s eye. Action is required#
CHAIR -
Well, I'll get hold of Chuck after I get back to New York I give him a
call, and suggest that he can possibly arrange to have an asterisk I or
some sort of indication, put beside the item in the table of contents
that requires immediate attention and your action.
FROM THE FLOOR-
If it's official action that is required of a member and this is pointed
out -CHAIR -
Something of that sort, but I think I will have to insist that you do at
least look over the table of contents. It is rarely more than two pages.
C\I.
I,
II
Page 15
o
FROM THE FLOOR -
I am wondering, as a suggestion, would it be possible to get regional
responsibility for contacting the members in the local area to find out
if contact has been made or just to remind them to mail this in.
CHAIR I might say that one of the reasons that we put that failure to respond
to two mail ballots in there I is that if you don't pay attention, you
aren't going to be a member very long. That won't affect this vote,
but if you pass these by-laws that provision will be in there.
FLOOR"-
In reply to that suggestion, it's my opinion that what we need is
responsible individuals who will take action without being led by
the hand, and not somebody that you have to constantly follow up
a nd follow uP.
FLOOR -
He might have been referring, tho, to those people who might not have
received it because they weren't on an address list.
CHAIR I might add in connection with this that quite a few of the local installations, Long Island Lighting for one, did not receive an agenda
till very late, because the third class mail service on Long Island
is lousy.. Jene Lewis called me, and I find that other members do
the same thing; they will tend to call the nearest Board Member, contact
like that, if they know that something is coming up_ . You certainly
should know because you would have to miss two copies, the one
that carried the item being discussed and the one with the ballot. I
think that to some extent it is back to a matter of installation responsibili ty . If you don I t get something you should, then find out why.
FLOOR -
o
May I make a suggestion - If you don't get CAST 7 write to Chuck .. This
way you will be sure of getting your ballot. When you get back to
your installation I if you haven It got CAST 7 I then write a letter to
Chuck right away so you can find out what they trouble is and be sure
you get your ballot.
Page 16
o
CHAIR-
Which takes care of only the people here.
FLOOR That was tried four weeks ago and I still haven't gotten an answer
back. I sent a telegram.
CHAIR To where? Terre Haute? Or Denton?
FLOOR Denton.
CHAIR He's not here to defend himself. I don't know the circumstances.
FLOOR We have a problem apparently to get this particular ballot approved
by a large majority. Would it not be worthwhile to make a special
firs t clas s mailing of this ballot only in an effort to get this back?
Otherwise~ you are going to have multiple mailings
anyway and you
are going to wind up with the same cost. It would be roughly five
cents apiece if you do it that way I rather than putting it in the next
CAST or some other method. This is a critical point. Once we get
past this our quantities are down. We don't have to have the same
reply.
I
I
CHAIR CAST goes out first class in any case.
FLOOR I know I but it 1s lost.
CHAIR A separate mailing is certainly something ••••••••••••
~
C'"
.,.)
Page 17
o
FLOOR -
I went back. I didn't know I had it. I went back and looked in CAST
and found it. I had gone right past it the first time and you're not
talking to people who are here. W,e'll all look for it. We1re not the
maj ority I by a long shot.
CHAIR -
Well worth considering. Thank you.
CHAIR -
The gentlemen in the back.
FLOOR -
You shouldn1t have to go through a special first class mailing just to
get you to read the appropriate place in CAST', or the appropriate ballot.
That1s what Cast is all about. This is what we are trying to do - save
a little bit of e,xtra money. For heaven sakes I please read it page by
page. Please read the letters from Joe Doakes to see what this is all
about.
CHAIR -
I have a gentleman down there.
FLOOR -
I just want to say in this case I as in so many cases I working with the
practical situation is going to require an awfully lot more than working
with the theoretical one. We can look out for this but what about the
one's who are not here?
I
CHAIR -
An there happens to be about 800 to 900 installations that are not repre s ented at this meeting.
FLOOR -
Hovv many must vote to establish a quorum for this?
CHAIR -
o
This has to be a mail ballot I which means two-thirds
I
something like
Page 18
800 or 900. However the present by-laws require the approval of
two-thirds of the members. They would all have to vote yes.
I
FLOOR -
You have 300 people here and there are roughly 500 .•••• •.•••••••
CHAIR We have 300 people but there are not 300 installations represented
here t I don't be Heve .
I
FLOOR -
Nevertheless t the first class mailing will call attention to it. This
has developed into a very difficult situation right now to make our
initial requirement for a quorum.
CHAIR Thank you.
FLOOR -
A very interesting point is that 1 if we can get this one ballot passed,
we can kick out all the dead beats who don't answer and then we'll
have a quorum again. But in the meantime l we have some number of
dead beats who never answer. O.K. so they don't read CAST so we'll
save some expenses by not sending them CAST or anything else after
this. We have to get this passed, because for now we can't kick them
out for not answering the ballot.
I
CHAIR I agree that almost any meas ures are warranted to get some res ponse
immediately.
FLOOR -
I recommend you that you put this first clas.s mailing to a vote.
CHAIR I hear your words I and, personally I I am inclined to agree with you,
but I will have to review our financial status. About 200 for postage,
plus some labor charges I of course. Folding and stuffing the envelopes
and so forth.
o
Page 19
o
FLOOR Suppose we don It get back enough votes, what happens then?
CHAIR Then we keep resubmitting it.
FLOOR I apologize immensely before I even start, but 11m going to say it.
We are a society of professionals in Data Processing. Welre supposed
to be experts in fine procedures and good systems. We are in charge
of installations spending lots of money for high powered·equipment.
Let's not be penny ante about running our own organization. I detect
this in all together too many res ponses of the Executive Board and
e lecte d officers I and I to my chagrin I I hear it from the floor. Let I s
spend $ 60. Let ISS pend $1,000. Let I s get an organization that works
for our benefit.
CHAIR At the moment, if we go wildly over budget I the registration for San
Francisco will not be $25 - it'll be $35. (Laughter and clapping)
FLOOR Wei 11 mail them out first class.
CHAIR They will go out first class.
CIAPPING
CHAIR There were two other hands raised. Same comment, or generally so?
Further comments?
FLOOR I just wanted to pose a question, Jim .. Is there any reason why or
why not our registration fee could not float between certain bounds
depending upon our particular situation and expenses ........... .
CHAIR -
o
It does now.
<.;.,.. "}
-~
Page 20
FLOOR -
Ididn't know if we were illegal in this respect.
CHAIR No, we aren't. We decide how much we need. We have to try to get
it out of meeting registration. That's the reason we left in our amended
by-laws the ability to impose a subscription feed on CAST. It is a
little unfair I because of this fluctuation in the registration feed. It
might always happen, we'll say, that one particular region wound up
having to pay the highest registration fee. We are considering other
methods of financing. Definitely. But at the moment we can't even
legally impose an subscription fee for CAST. Are there any comments?
FLOOR -
I just wanted to second again what the last gentleman said. We spend
many, many thousands of dollars on computers. Man-time cost thousands of dollars. Half an hour of computer time is not really thought
of as being expensive. Half an hour of computer time is $25 or $30.
Please don It think small. We've got bigger things to get out of this
instead of worrying about $10 more in registration fees.
CHAIR Part of our problem here is there are a lot of 1130 installations I and
1130 .installations are one and two man installations. And they do
have cost to worry about. They are in the same fix the 1620 group
was years ago" I I m talking about the maj ority now.
FLOOR -
Neither the typist nor I could get thas clearly enough for transcription.
Comments about the Boston meeting. res
CHAIR I know this.
It was a very poor meeting.
FLOOR -
I understand that people just sat and talked. I agree with this
gentleman - if it l s needed, we should go ahead and attempt to get
our finances straightened out now with a higher registration or with
some nominal charge for CAST. For instance fifty cents a year.
c
Page 21
o
CHAIR -
At the moment there is a subscription price for IBM employees who
want to get extra copies - you know - fOl' branch offices - things
like that. This is $15 a year and that is about a break even price.
FLOOR -
About a $1 an issue.
CHAIR -
Approximately that.
FLOOR -
Let' s charge it I then.
CHAIR -
At the moment, as I saYI under our existing bye-laws we cannot charge
for them. We can do it only if we are supply extra copies or something
of that sort. Copies other than the one that a member installation is
entitled to.
CHAIR -
(Section deleted; couldn It trans cribe) •
FLOOR -
It seems to me I Jim, that the message is clear that you ought to spend
the money to put on a quality program l and most of the people are going
to pay the money. The people who don't aren't going to recognize the
importance of operating their own installation I whether it l s a one-man
installation, or what it is. Let's spend the money and do a high class
job.
CHAIR -
We agree with that and we did it. And as I said I have a guilty conscience.
o
PageZ2
i~
I.
I
CHAIR -
o
O.K. let's wind this up. You people have indicated that you are
largely in favor and I am quite pleased, I might add , with your
response.. Is this pertaining to this?
FLOOR American Electronic Labs - I heard you talking about the COMMON
meeting in Boston about not being well run. That was the first COMMON
meeting I ever attended , and I don't agree with you at all. In fact,
I disagree violently. This meeting has been run very weel, but,
to me in Boston I whether it was planned well or not I it ran very well.
CHAIR I agree.
It wasn't planned very well t but it ran well.
FLOOR It got this meeting off the ground. A lot of things that happened here
resulted from that, so I see no reason for anybody to say to themselves you know - we have to look back at Boston, and say we didn't do a
good job. You did. I was very happy with it. In fact I I probably wouldn't
be here if not for that.
C
1
CHAIR And we didn't do a good job here in my sense of the word, and we
expect to do a better one in San Francis co.
Bill Lane, incidentally I needs a Local Arrangement Chairman for San
Francisco. He's handling both ends of the deal, but he's nearly 200
miles from San Francisco. He can't work as Program Chairman, and
as Local Arrangements Chairman. He shouldn't be working as either I
but we feel that we can It swap Program Chairmen at this late date ..
FLOOR I feel like the mandate is clear to those of us who are here I altho
we can It vote - we could vote I but you would tell us that we don't
have a quorum - so why don't we do that? (Comment concerned
registration fees for San Francisco).
CHAIR Never mind. I think the majority is quite clear. We hear that, and
,
(
,~'
I
{
Page 23
o
will consider that in discussing our San Francisco plans tomorrow.
There will be an Executive Board Meeting at lunch to discuss the
results that we are hearing now.. Bill Lane has some comments ..
He's the guilty party for San Francisco.
BILL LANE I hear all of these things .. We haven't discussed the raise in the fees,
but I can almost guarantee tney are going up. The first mailing for
the San Francisco meeting will contain approximately the following
material , and will come out approximately October 1. It will contain
three hotel registration cards. Please get them in. It will contain
a blurb about how great San Francisco is; I think it is the best city
in the world. I'm biased. It will have some information regarding
the wive's program. Weld like to see a lot of wives out there .. You
ladies bring your husbands along. Maybe we can send them out to
North Beach. It will have some general information on the meeting,
and will be about October 1. We expect, hope, and ask for returns
on this information. We would also ask that if you have some questions,
if you have some items that wish to be discussed, that you get to
your project leaders or to your Division Chairman, before you go
home tomorrow so that they will get on the agenda. I am not about
to put anybody's "non1l-return on ,the agenda. I'm not going to make
an agenda for you A s lot will be there, but in forma tion will be
printed only if it is submitted. And submitted means by deadline.
There will be a second mailing approximately (both of these are
going out first class; that's already been decided). November 1
which will contain, for those of you who did not get your hotel
card in, another hotel card, and the agenda. This is the tentative
agenda.. The final agenda Wil, be handed out at the meeting .. Please
get to your Division Chairman today. If you can It find him today,
corral him tomorrow. If you can't find him tomorrow, the names of
the various people are in the agenda. If you can It find anybody els e I
. get the information to me, and I will send it to your Division Chairman,
but let's get it out there. San Francisco can be a great meeting ..
In the past, the Western regions have had a slightly lighter attendance
than we ha-d· in New.Drle.ans.. I was not at the Bos ton meeting.. A
lighter attendance does not necessarily mean a poor meeting I and,
while we anticipate maybe a lighter attendance this time, start twisting
arms right now - you know the boss - tell him how you have to go
I
I
It
•
Page 24
to San Francisco. We will have a tour for you of IBM's campus
facilities in San 10se. IBM has a gentleman who has offered to
arrange bus service I and 11m going to take him up on that. There
will be available at the Data Center I and possibly other~laces I
computers that include the following: 1130 I 360-20 I 30 I 40 I 44,
50, and I think there is a 65 down there I isn It there? All you IBMlers.
You donlt know? What1s the matter with you guys? Well, anyway,
there's up through a 50 down there~ There are Data cells; you know I
thatls the little tweezer gadget. Anyway I the San Francisco Data
Center is well equipped. There are a number of pieces of equipment
there. You will be allowed to reserve time in the evenings to play
games I if you want. You will take care of this I WQn't you, Pat?
It has been suggested, I think-casually, that we also, for those
of you who are interested, hold a tour of the Napa Valley. I don't
know whether any of you know what the Napa Valley is, but that's
the greatest wine-growing country in the world, and there are lots
of wineries out there. This is by preference at a later date. Anyway,
San Francisco, the 11th through the 13th of December. Welre going
to see everyone of you there. It's the last month of this year.
Shortly after Thanksgiving and before Christmas. You can get all your
Christmas shopping done in San Francisco and things like that. Get
your papers in. There is something to be said about an unstructured
meeting an unplanned meeting but as an engineer it bugs the heck
out of me. I want to see some programs I and you are the people who
are making the programs. O. K:? Thank you.
I
I
CHAIR After that sales talk I give up I but I agree with Bill. There will be
a report by me on the 360, 1130 and 1800 CPL minimal standards
that have been distributed by PID for those machines. The 1620 would
not have received a copy of them, Again l these are minimal standards~
I'll tell you what we came up with, why w~ chose them, and what
SHARE'S z:-eaction to them was at the Program Library Meeting tomorrow.
I'll try to make that the last part of the meeting, in case there are
any conflicts with other meetings. If any of you are interested, or
have any ideas at all, about what you want to do with the Program
Library - fine. Otherwise, instructions are adequate. Any other
business!?
o
c
Page 25
o
FLOOR You have listed on your agenda discussions about IBM - customer
relations. For those of us who are first attending I Ild like to hear
what you have to say about the financial relations.
CHAIR Fine. IBM publishes our proceedings for us at the moment. They
have suggested that this would be a legitimate fiscal responsibility
for us, but quite agree that we are not in position to take it over
at the moment. We have to have a reserve in the treasury before
we can do that. They maintain our Program Library; there1s no question
about their continuing to do that. CAST and the Newsletter I which
preceeded CAST, and is now part of CAST have both historically been
published by COMMON. IBM provides certain help at these meetings
in the sense of IBM personnel - They give us an indirect subsidy in
the shape of IBM people attending the meeting and paying the registration fees.. There is no direct subsidy from - Oh I Yes I the coffee
breaks, audio visual equipment, special equipment of that nature I
signs - the administration facilities J for instance they have a 2400
Zerox in there for us, badges material of this nature - they do pay
for us. I should say that there is no cash subsidy. They help us
negotiate with hotels.
I
I
C"
;'
I
JUDI GREENE-
I would like to say thank you to the people of IBM. Those of us who
were at the 1130 meetings - we got a lot of very important information,
and a good deal of it was impromptu; we appreciate it very much those people who did speak to us.
CHAIR Incidentally I since there are new members present and Arnold Smith
was not introduced at the new members I meeting I or at the General
Assembly: stand up, Arnie, and take a bow - you and Pat, both.
On the transportation to the boat - there is a map illustrating the route I
something like five blocks. 11m told they are rather long blocks,
but we had planned that most of the people would walk down there.
Cab, that's fine.
o
Page 26
FLOOR -
Can we park?
DAVE DUNSMORE -
There is parking on the wharf down there. The boat is to be loaded
between 5:30 and 6:30. Cocktail hour until 7:30, and at 10:30 we
will be back. I VVill try to arrange for cabs to be present in that
general area I :because it is an up-hill walk back.
CHAIRIf there are no other comments I Ill! entertain a. motion that we adjourn.
Second? All in favor I move.
c' ,
7
\
o
SESSION NUMBER
T.4.2.
DISCUSSION
PLANNING SESSION FOR FUTURE MEETINGS OF THE EDUCATION PROJECT
OF THE APPLICATIONS DIVISION.
ATTENDANCE
20
o
317
C',
,.
SESSION NUMBER
F.1.1
SPEAKERS
W.K. THOMSON
J.R. AHART, INC., DAYTON, OHIO
DISCUSSION
USING 360/30-40 R.P.G. FOR FUN E PROFIT.
MR. THOMSON DISCUSSED HIS COMPANIES EXPERIENCE WITH R.P.G. HE
POINTED OUT IT'S POWERS, IT'S SHORTCOMINGS, AND WAYS TO OVERCOME
SOME OF THE SHORTCOMINGS. THE PRESENTATION WAS ABOUT AN HOUR
IN LENGTH WITH A QUESTION AND ANSWER PERIOD FOLLOWING.
APPROX. 50 PEOPLE ATTENDED.
c
o
USING 360/30-40 R.P.G. FOR FUN AND PROFIT
PRESENTED AT THE
COMMON
MEETING SEPTEMBER, 1967
CINCINNATI, OHIO
LIAM K. THOMSON
GRAMMER
o
J. R. AHART, INC.
DAYTON, OHIO
TABLE OF CONTENTS
INTRODUCTION
I. COMPILER EFFICIENCIES AND DEFICIENCIES
II. LANGUAGE PROBLEMS THAT CAN BE SOLVED AT SOURCE LEVEL
III. CODING TECHNIQUES WE HAVE FOUND TO PRODUCE MORE
EFFICIENT PROGRAMS
IV. EXECUTION PROBLEMS CAUSED BY D.O.S. DEFICIENCIES
V. CHAINING (INDEX SEQUENTIAL FILES)
VI. CHAINING (DIRECT ACCESS METHOD)
VII. TABLES
VIII. ASSEMBLY LANGUAGE SUPPORT BY USE OF THE EXIT INSTRUCTION
'0
.,
IX. IMPROVEMENTS WE WOULD LIKE TO SEE
330
INTRODUCTION
Using R.P.G. for FUN' and PROFIT?
Prior to the System "360"
Happy R.P.G. users were few and far between, but we feel it
has come of age and. if given a chance can find its place in any
installation.
The intent of this paper is to encourage the Use
of R.P.G. and. to pass on the good points and weak points we have
found through eight months of use.
There are six different coding sheets used in preparing a
program:
Fil:-e Description, File Extension, Line Counter, Input,
Calculation, and. Output.
(See Attachment 1 for examples)
The
R.P.G. manual has a section for each sheet which. covers all the
entries for it.
We have found no drawbacks in using coding forms
over free form which we had been used to.
On our first programs, we seemed to be bumping heads with the
lip, calculation, olp cycle that is the fixed logic of an R.P.G.
program.
(See Attachment 2)
We felt restricted because we
couldn't control the occurrence of the lip cycle.
We have found
that by making card.., disk, and. printer layouts and. then planning
our programs from these the logic fits many applications very well.
o
it
ijQ
,
'$I i
, ,,## ::;;m
, "MAM UtWtUl4,iTi4¥iillMiC"'
~
I,
1
. I. COMPILER EFFICIENCIES AND DEFICIENCIES
j
01
I-
To date we have found only one logic error in the compiler.
I
If the same name is used with different tag operations on the
calculation specifications, it is not diagnosed and the last
one encountered is used.
This is similar to duplicate statement
numbers in Fortran or duplicate labels in Assembly Language.
If the level break field of the first record of a file that
has level breaks specified is zero or blank, the level break
indicator is not on for the first detail cycle.
The compile time for
peripheral gear.
p~ograms
will depend mostly on your
With a (1443 Printer, 1442 Punch, 2501 Reader)
a (441) source statement program can be compiled and cataloged
in five minutes and seventeen seconds, (60) source statements or
less takes two minutes.
If you take object decks, the speed of
the punch will be the controlling element of compilations.
You
will get larger object decks from R.P.G. than you experience in
other languages because all fields are initialized.
To date, for
a 32K machine, the largest object deck we have generated is 500
cards and the smallest is 100 cards.
Because of the speed of the
compiler, we have found it economical on some jobs to use a compile and execute rather than maintain a source and object d.eck.
c~
I
o
II. LANGUAGE PROBLEMS THAT CAN BE SOLVED AT SOURCE LEVEL
Our early programs were frequently canceling due to reading
an undefined record type.
The HO indicator is used for canceling
a job for this reason and eleven others, so the programmer has
no way of deciding anything based on HO.
We have found that
leaving the record identification code blank on single record
type files and using a blank specification as the last one of
mUltiple record type file and then checking record type on the
calculation specs and taking action from resulting indicators
used on input gives the programmer good control.
(See Examples)
We found it was very difficult to obtain a skip to a new page
and put out headings after level breaks by trying to condition
headings on overflow alone.
By using the level break indicator
at detail time in a (or) situation with overflow - Group Printing
Feature - we have eliminated this problem.
(See Examples)
We found we were getting some extra pages in reports on which
we were doing an internal line count.
This is because if a print
file has been specified and an overflow indicator is not used as
one of the first three indicators of any olp line the compiler
generates an automatic skip to channel one on overflow.
We elim-
inated this by using a dummy line that is never executed with
overflow as one of its indicators.
Also there were some unexplainable lines at the beginning of
our reports.
These were the result of the.program looking for
first page output and allowing any detail conditions met to print.
o·
We find if you have to use a not condition for printing a detail
line., it must have NIP specified also to stop its printing on
first page o/p.
I!
We noticed. t:hat printer output was not at full print speed.
in programs that had little·calculation to slow them down.
We
found that this is a result of the program having to wait on a
line to finish printing in order to check overflow.
The use of
the line counter specification as an internal carriage tape
gives you maximum printer efficiency.
Unfortunately, it was not
working properly prior to release 11 of D.O.S.
o
----~-----~----~-----~.-------
III. CODING TECHNIQUES WE HAVE FOUND TO
PRODUCE MORE EFFICIENT PROGRAMS
o
The GO TO operation on the calculation specification was
said to be one of the improvements over 1400 R.P.G.
We have
found it to a powerful instruction for two reasons.
One is
probably obvious that is the ability to branch backwards in the
calculation coding.
But an additional use we find of great value
is that it gives you the ability to create your calculation in
modular blocks of routines to handle specific parts of a problem
rather than use a resulting indicator to allow or inhibit the
execution of coding.
We found that the latter approach not only
produced larger and less efficient programs, but was extremely
confusing to write and maintain.
(See Examples)
The total time block of coding is entered only once regardless
of the number of level breaks that have occurred.
All total time
calculations that can be done are executed at once, and then all
total output is done.
A second improvement over 1400 R.P.G. is the use of mUltiple
input files.
We have found several uses of this feature.
One is
specifying the same "file with two different names so you can process it twice in the same program.
An example would be where
totals are needed to calculate percentages of detail lines.
If you
specify the same file as a primary and a secondary with matching
record, R.P.G. will give you the detail records for a group twice.
You can add them as they are processed as a primary file and. then
calculate percentages and print them as they are processed as a
o
-A,;';'; H,
secondary file.
;, (,
,4,$" t. ,X, .QA#U$ Uk;aAJUi4hlUW&=,M&£Z
c
IV. EXECUTION PROBLEMS CAUSED BY D.O.S. DEFICIENCIES
When a disk file is specified as blocked and a full block is
not generated one time a record with a key the length of the
data is generated and a blank d.ata record written after it.
If a 'Utility created a disk file and it has not been sorted
by sort merge, an R.P.G. program will not recognize the end of
file written by the utility.
When a file has been specified as an 'Update file, the last
record to be updated is not returned to the file.
All of these problems have existing apar's to correct various
levels of D.O.S.
-----.-~.-----------.-----~--
..
----------
o
V. CHAINING (INDEX- SEQUENTIAL FILES)
If fi.le is unblocked, make the input record size ten bytes
larger than the record length. to handle records from an independent overflow area.
Be sure to include a LBLTYF card when link editing the program.
If key fields have been packed you can get a no record found
condition because the signs are different.
Chaining indicator is specified on chaining file only.
When creating the index sequential file, be sure the key
field is reproduced in the record for all files, blocked and unblocked.
c·
If at all possible do not pack key fields, becayse there is
no ability to chain from a card file if keys are packed.
If no record is found HO is turned on.
337
o
VI. CHAINING (DIRECT ACCESS METHOD)
The Addrout option of the sort merge program can be used to
produce a table of keys and associated track addresses.
This
table can be used to convert keys to a chaining address through
a table LOKUP operation.
This method will allow the user to add records with an R.P.G.
program.
By putting blank key fields in the original file, when
a no record found occurs in the table LOKUP fora key then a
LOKUP for a blank key will give a track address for new record,
and setting of a resulting indicator will let the
R.P~G.
program
know this is a new record.
c
c
o
VII. TABLES
If a HI/EQUAL or LOW/EQUAL LOKUP is to be performed., the
table must be specified as ascending on the file extension sheet
or an invalid. LOKUP will occur.
Also the use of the resulting
indicators is low gives next highest value, high gives lowest
value.
o
i.11
I~.'
I"
o
VIII. ASSEMBLY LANGUAGE SUPPORT BY
USE OF THE EXIT INSTRUCTION
We have found that the ,.ability to interface R.P.G. and the
assembler language through the RLABL, ULABL, and exit command.s
is very beneficial.
The linkage is not complicated as the
examples we have included will bear out.
The routines can be
in the relocatable library and will autolink.
The first routine we found to be of help was the ability to
block and unblock a segment of an input record by a displacement
and length.
The need for this arose when we wanted to loop on
the calculation specifications and move across an input record
picking up d.ifferent fields and returning an updated value.
(See
Block and Unblock Routines)
Another routine we have made extensive use of was a subroutine
that allows the programmer to cause the program to loop between
detail calculation and detail output until he resets it.
This
.
allows the formated output of tables at any time in the program.
(See R.P.G. LOOP)
A great aid in debugging is a routine that PDUMPS the status
of all the resultings indicators and the label fields at any time
specified.
(See PDUMP)
J?,s mentioned before the HO indicator can be turned on for 12
different reasons.
Logically, all of these cannot occur in one
program but several could at anyone time, and DO message is
issued to distinguish one from another.
~outine
Therefore we developed a
to analyze the reason it is on and print it on the console.
(See HALT)
----
-----~---------
Also we found there was a need to give the operator some
ability to align special forms prior to starting output.
Since
you cannot open or close an individual file in R.P.G. we developed a routine to allow some alignment.
(See ADJUS)
Finally we did not like the idea of having to read one card
to get the current date so we wrote a routine to get the date
from the communication region.
c
o
•
,\" .. U
, #I
; V¥UA«Alli4klUTtUM&;;;&C=M&LL22CL&Z2ERLL
(See DATER)
IX. IMPROVEMENTS WE WOULD LIKE TO
SEE~
o
1. Ability to create and add records to an index sequential file.
2.· Ability to control when chaining should or should not occur.
3. Make the date from the communication regions available in
the same manner as page number is.
4. Allow the programmer to specify if a table should be filled
by R.P.G. or just have space reserved and initialized.
This
could be done by leaving out the from file name on the extension sheet.
5. Allow a file to be closed and re-opened by the programmer.
6. Don't cancel a program because of a data exception either
set field to blanks or zeros or give control to a special
routine on calc specs for handling this.
7. Allow program to have a separate block of coding that is
executed. on first page only.
(Calculation and o/p).
8. Use a binary table LOKUP if an ascending or descending table
is used.
I
I
c
o
ATTACHMENT 2
Using D.O.S. 360130-40 R.P.G.
INITIALIZE
PAGE olp
19
Yes
PERFORM
TOTAL TIME
CALCULATIONS
No
PERFORM
TOTAL TIME
LOAD LABEL
FIELDS ON
lip SHEETS
FROM 1/0 AR
olp
PERFORM DETAIL
CALCULATIONS
PERFORM DErAIL
olp NOT COND.
ON OV.
SET
eV
No
Yes
SAl
PRINT olp
CONDITION ~--------------------~
ON OVERFLOW
o
~Ii
i
SUBROUTINE NAME:
LANGUAGE:
BLK and UNBLK
RPG (16 K.D.O.S.)
PURPOSE:
To allow the programmer to retrieve fields from an
I/O area by changing a length and displacement value
in the calculation coding.
CALLING SEQUENCE:
C
EXIT
6
28
C
EXIT
6
28
C
C
C
C
RLABL
RLABL
RLABL
RLABL
-28
6
BLK
(TO STORE IN I/O AREA)
33
UNBLK
(TO RETRIEVE FROM I/O AREA)
33
IPBLK
PIECE
LNGTH
DISP
Lj·3
(see Note 2)
(see Note 1)
30
30
31
NOTE 1:
PIECE may be established as an Alpha or numeric field. If
packed data is going to be retrieved PIECE must be numeric,
otherwise RPG will pack the field again if it is moved
from PIECE.
NOTE 2:
IPBLK must be an Alpha field and can have length established
on lip specs.
OPERATION:
BLK
UNBLK
Takes DISP and adds it to address of IPBLK and moves from
left to right out of PIECE the length of LNGTH into that
address.
Works the same as BLK except the move is from IPBLK to
PIECE.
.-~-----.------.---
c
~:~
*
-,.
RLK
WRITTEN BY R. c. nICE JR. IBM DAYTON
"-
o
BLOK
RLK
TITLE
START
EXTRN
USING
STM
LM
ZAP
CVB
ZAP
eVB
AR
BCTR
EX
L~1
'***
GENERALIZED
O/P
BLOCKING ROUTINE -- DOS RPG
IPBLK,DISP,PIECE,LNGTH
BlK,15
2,5,SAVE
2,5,REGS
DWRO,0{2,3)
GET 'DISP'
3,OI.-./RD
DHRD,OC2,5'
5,OWRO
2,3
5,0
5,MOVE
2,5,SAVE
GET 'Lr"-JGTH'
(2
BYTES)
BYTES)
ADO 'DISP' TO 'IPBLK'
SUBTRACT nNE FROM 'lNGTH' FOR 'EX'
MOVE 'PIECE' TO 'IPBLKCOISP+l)'
BR
14
SPACE
MOVE
MVC
0(0,2),0(4)
RF(;S
DC
ACIPBLK,DISP,PIECE,lNGTH)
SAVE
DS
4F
nl,\/RD
OS
o
END
END OF DATA
o
(2
***'
UNBlK
** WRITTEN
BY R. C. OICE JR. IBM
*
TITLE '*** GENERALIZED
UNBK
UNBLK
DAYTON
lIP
UNBLOCXING ROUTINE -- DOS RPG
**' .
START
EXTRN I Pt.l,L:-. ". i JP .. P ~ :_CE, Le-···'
USING tF \l",~_5
STM
2, ./"1
LM
ZAP
2,5,ttl such
statistics ..
a)
1!~r!4quency (iistri~jutions ,j,jased on Verual and i"iath
SA!l' scores, all(:i ra,nk in clas~~ carll i)(2 q(jn~.u~att;'!d for eaC(l
prc)qraro ()f (\~ac:) scllool by appoi{~\.; I accept~:k! 811d J.,)re-re9istered
applicclnt)i •
t»
Gco number,
address, ata. It may ~ \lSed to verify student, st.atus.
2.
Reminders.
Reminders on loanud books and overdue not.ices ma~i
00 produ(..~d for the library.. This flhould greatly facilitate ,-'*Ollection of loaned hooks.
1.
Room
Schedul~ls.
Roo.m schedules r~flect:ln9 its use can be made availabl.a to physical plant in order to facilitate service
12
o
schedules and allocation of rooms on special requests.
F.
Central Mailing
Central mailing may use the eSDS to produce mailing
labels for all general and selective mailings.
G.
Alumni Association.
A major report which may be produced from the Alumni
file is missing information lists. Periodically missing
information lists, designed to facilitate ti~e collection
of data to the Alumni file, can be produced. A typical
missing information list might be one containing all Alumni
who do no~ have a telephone number on file.
1.
Directories.
Al'~i directories can be alphabetical, geographic,
or class and school produced from the Alumni file.
o
2.
Fund Raising
Of all the areas in which the Alumni file is used
the Alumni Association, fund raising is undouotedly
the most important. For t..ltis reason, a mora detailed account
of this application will be given.
by
a)
Worker Assignment Lists.
Lists of all ~~umni with a better than average giving history can be produced. These lists may be used
to aid in recruiting workers that will participate in
the annual fund raising campaign.
Phonot..~.on Pledge Cards.
The phonothon pledge cards may consist of two parts.
One part may contain personal information, givint the
history of the alumni; and the other may contain his
na~e, address, and a place to write a pledge.
These
cards are used by the workers to contadt the alumnus
b)
by telephone and try to solicit a pledge.
One part of the pledge card is mailed to the Alumnus
reflecting his amount pledged; the other part is used
to enter the result of the telephone conversation in
the Almoni File.
o
11
o
e)
...lnden.
DariD9 the capalgn, tIOD~ly ..-iadera can be produced fna the Aluam1 fil.. !bea. reai.ndera are produc:ed for all al....1 Who be". not fulfilled their p1edge.
!'hey .bow the ano\1At: ple"pd, the MlOUD.t. 91ven. to date
(If any), ad a balance due. Rea1ndera are then mailed
to the Alwanu. 1n the .hope thatt • payment. for t.he .amO\lJlt.
pleated will be .ent.
ell
to
e)
. . .lpta.
Daily receiptJI of ...,..ntJI can be pJ:OC!uced and mailed
doDo" at:. the . . . . Uae that the file 1& updated.
~be
Block 8beeta.
Block sh••ts aa1' be 11st.lDg8 showing a detailed
l1at of 4oaon . .a ao\lftt. 91ven. by each. donor. They
also sbow tile tot.al dally latake.
f)
Worker Reports.
Worker reports can be produced periodically to report
to the individual workers the PX'09r889 of their group_
'fhe •• report. may contain the worker'. name, followed
by a list of proapeota, the amount pledged, and the
amount qiven by each proapect.
9)
Proqre88 Report••
Reporta can be produced to rafl.ct the amount of
proer of pledg•• and t.otal number
of donora can al80 be produced to ref.lace the overall
ProtX'es ••
h)
Year-end Analysi ••
Adota11ed year el\<1 analysis refleotin9 t.he fund
raising capaip can be produced. 'this analysis may
ran<]8 fZ't'ft a qeneral university-wide pertomance analY81. to •••lective qeograpblc and school-cluR ana-
lysis.
c
14
o
d.
l',)ursar
..
~_·~w
_~
'rh~
office of the
,~3ur9ar
can ex.t.ract. infortna\tion
from both t.he Admlssinn >1auter Filo and the Studl/!nt Cata
Fil(~ ~
",h:,.taltl anJe:(ception lists may
compris(,~
tho major
us(e vf the Admissicns ['~agter ,Pile by the nurS~tr.
Receipts
of tui tinn paym.ent. "may ,L'6 producecl by COMpnJter for t.he
T\ursar, but oftt.~n, t.h~ students c;l}ncf911ed check is reqarde6
S:J 9uffic.i~nt receipt. A l i .. t of allprc . . ·rcqistered Freshman. irH.,,:ludinq amount ,~f tuition prepa.1d, may be produced
?rlorto 1~qi8trat1on.
Th~~
1.
follo\'1intl arc
Tui lion and Fees
tt-IO
usee of the SDF hy the Bursar:
ra'Y"t~nt
l'ratHl{,lction.
Prior to registra,tion# a set of pre''''f,Hl.oched payn':ent.
cartiS l~',.ay be produced 'Ir/t'1ich conta.in tne type and amount
of any financial aid autilorized. These cards may then
,be used to record any payment made on tuition 'and fees.
and later used to update t.he SDF.
2.
Post-registration Totals.
After registration, a summary of totals can be provided for t.'l. Bursar. Totals, such as amount of tuition
and fees paid and balances due~ can he helpful in verification of account. balances; as well as enhancement of
the illumel control.
I.
Financial Aid
~........
J
•
r
T~1e Office of Financial Aid may utilize both the
all'.! t;:1C ~DF. The followln9'ma'y be output from t.he
.ru·r~'t
p.~~~p
ret-;ardin\J ineominq
1.
S\~ary
,A
1~1
Frt~sbman.
Il1formation Labels.
of summar} information can be provided for
all applicants who have lieen accepted for admission.
This infonnatlon call be used for cross reference on .11
who apply for financial ald.
2.
o
Financial Aid Applicants' List..
Prior t.o the Pinanelal Aida Award Committee meet.inq,
1$
c
a list. CWl be produced of t.hose applicants for which a
CSS card bas been reeeivod. Tbis list. may be ordered
by predicted OP1, adjusted need, or some other criteria.
It. sbould cont.ain the pertinent personal, academi.c, and.
financial information which .1s needed to make award deci-
sions.
3.
Pinancial. Ald Autherlaat.ioD '1'ransaction.
A .et. of markseniaf'f.t cards may be punched concurrent:ly
witb the financial aid applicants list for \188 by the
awards comtnittee. As awe,rd. are made, tbe amount alld
type may be marked on tbe card for later use 1n updating
'the NIP.
4.
Pinane!al Aid Avar(} Letters.
As the NW i8 updat.e.d with t.'e mark-senae authorization cards. a let.ter of award can ~ prepared for mailing' to ~ .tudent..
S.
Financial Aid Candidate List.
After the pre-registration de.tiline, a list. of tboae
st;u&!mt.s who haw accept.ed the awards CAll
~
'produced
as a follow-up pmcedure. TIli. listing may bet used as
• final office copy for report. preparations and inquiries.
, •
Suaaary Tot.als.
In con,unotion with producing the candidate list,
• ccmplete J!'891ater of summary totals can be periodically
print.ad. Such ltel1 and a'llllOunt accumulatJ.OD8 will be very
useful in maint.aining control of funds as well .s preparaUOIl of required Pederal and Stoats reports.
The SOP i.e used for 'procesalnq \U1dargraduates by
the Office of t:be l'iaanclal Aid in essentially s . . manner .a the AMP 1s . used. tor pzoce•• llV1 tnCC'Miaq freshman.
!'he slCJDlflcant cUffennce 1s tlle wae of Unlftnlt.y acadea1c
oat.a lnl1. . of Mgo .cnool 4ata ...ad boaz-tl &OOR5.
J.
<;:1ub.'~~.~!.&1 Ae8OC!~~~OB., B~! •
.,.. SOP may be usea by the varloue Unl".nlt.y club·.,
lIOOleU.a, aDd. otMz- oqaal ••Uoaa wboIIe tm4 lUly be aaa4ea1c f
athlet;lc, 800ial, etc. A ....l.ty of lafomat!on 1. 89&11.able fOlr aol1c1t1Ag pot.cmUal ...-ben, 01" OODt.a~lnt pnHDt - - . n . !'be . .'or ... thea '8 apec1alaail1Dg label.
or pze-add...... ea.velope••
377 -4
c
16
0 "',
"
K.
!,resldent
I.
Off~C?C!
The office of the President, or t.ll.Board of Trustees,
require the addit:ional aovantaqe of an information system over a record keeping system. That advantage is the
developement of informat.ive stat.istics from which deci-
sions on policy and future planning can be made. Prior
year to current year comparison totals &\0 ratios, wi"
the myriad of variables, could render the recor~~eepin9
• small portion of the system. For example, to know -.rl0
relative quantity of applicants requesting dormitory res! ....
Clence over lK'JIRmUurs could facilitat. the decision to
build a new dormiborJ.
o
17
C(h~CLUSIOtf
,
i
1 , .. lit
•
A Coordinat.ed St.'udel'lt Data System must encompass
all offices of the University that have a need for studant data. Sucb an a11-1nclusiva~ly8tem will enjoy the
followlnq advant.ages. Non-duplication of .ffort is avoided
throuqn sinqular 1ntroduction of source data. 'rbrough
the establish,;-nent of Off'i '~>39 of Primary Despon.tbil! t.y
(OPR), eas~ of file main~nance 115 aobMved.
Economic
availability of data for all University departments is
gained through cent.rally maintained student da,ta.
Con-
sistency of infomation 1. assured throuqh the inherent.:.
compatibility of common source data. Last., but not 1(t;8J:..t,
the Ut\lwrs1ty will realize an ,abundance o,fresearch dat.a
to l>e used in the analysis of University performanoo.
;'Jlth the addition of the necessary terminal equip:'1\;)nt; the eSDS 1$ potentially a :real-time system, t.hus
t!l'()vid.in~J the special advant.aqes of such a system.
C\
'·"1
-->'7(1'·-J i '
I
(l)
0
.....
DATA
. .. .
nPE
PRIMARY SOURCE
row
OP
PlUMAft RUPOtf •
17.
Spouse'.
18.
Reli9ion -
Name
C
C
Registration
roan
KP
ReglatratJ.oo
Student Hospital
Insurance
Activiti.s
20.
Studen~
21.
H1gh School Name
C
_9:latration
Pom
C
Student Affain
Pom
Kit
Regl.t.~...
Adtd . .ion
KP
A41I1••.toa.
Meqiftrar
AppllcaUoo
22.
High School AdcS.....
23.
High School ae
A
AdIIi ••loa.
Adrd ••101l
Applicat.ion
A
Adml. .lon
KP
Applloa~lon
24.
'l'ype of H1.9ill School
25.
Year of Graduation
MIIl••tea.
Applicat.ion
A
Uiqh 8011001 Average
27.
nigh School
28.
SAT Scone
Adm1 ••1011
AdlU ••101Ut
ItP
A
AdIIi ••iona
A
A
Adlai ••lons
H19ft School
KP/pc
H19ft School
KP/PC AdaJ. ••loaa
7ranacript
Ra"\k
AdIai ••lon.
Applicat..loD
(H1qh School)
26.
Re91auU'
Re91et.ru:
Pon
19.
orrICK
ttranaeript
Eft
PC
AdMl..1oaa
PC
AdIIi. .lone
QED Card
29.
Ach.......nt. Score.
A
ETS
CEED Card
30.
Transfer St.atu
A
Adad. ••loa
Applicat.ion
KP
Adlai••loae
31.
Year of Graduation
Present Program
A
AdIai •• lon
KP
Adai••lone
32.
School l::nrolled
A
Admi.sion
Application
1(P
Regist.rar
Application
381
()
(3)
0
••
Mft
........
33.
35.
....-
H.
Counu
37.
lIcbedul.
. f eboft
I ••
II.
0
o
h"09na Bazol1ed
'lIPS
A
....
Ra91.t...'t1oa
A
"'1.~ftS'
",iatE'eUea
...,1."U'
KP
"'1._~u
PiD_alal Aid
baa
II/S
Pia_aia1 Aid
.la_del Ald
N/8
P1DaDaial Aid
ft....c1a1 Aid
MIS
PlnMclal Aid
MIS
.1naac1al Ald
P
41.
Won
.
.uta.
.....vU'
GJ:'ade "po~
8abolanhJ.pe
.....17 PlnDOia1
at
A
40.
tl.
Reglsuu
Gude "POft
.
GS'_u 1D Aid
or
A
Lo. . Aathod.
42.
OPl'lCB
.PUMARY. USPOH • .Regletru
l'oEIl
It.
QIr_U
0
AdId. •• lon
1e918und toJr
........
AdIai •• ioll
10M
Appl1caUon
Maulo81Ued
Cl...
. .......
Appl'oa~lon
'leacbezo Cenif1oaUoa
eun.aUy
PRIMAn' 80URCE
rom
Foa
ria_lal Ald
..
ro_
Eft
ess Can
I'1DUCial Ald
o
S F S SIn f\J t\1I 1M R F R
F • 1 • ')
SPEAKFRS
R • A.
F f) IAI AR I) S,
I R Men R p n RAT Tn t\\ n t\! L ARn RAT n R Y AlJ Tn MAT InN BAS E n
SYSTFMS
c
o
o
Presentation
at
C01VIMON Meeting
Cincinnati, Ohio
LABORATORY AUTOMATION BASED SYSTEMS
I
!e
September 8, 1967
o
R. A. Edwards
IBM Corporation
112 East Post Road
White Plains, New York
LABORATORY AUTOMA_TION BASED SYSTEMS (LABS)
o
Definition
LAB Systems are defined as the application of largely sensor based
1800s and 1130s in the laboratory to perform functions such as:
A.
Data Acquisition
Recording of data at a rate dictated
by characteristics of the instrument
and the experiment.
Data Handling - Data Massaging
prior to analysis.
Filtering
Averaging
Scaling
Validity Checking
B.
Data Analysis
Selection of appropriate analysis
routine based on:
Accuracy Requirements
Speed of Response
Analysis for purposes of interpretation
of instrument output.
c.
Data Management
Identification- Comparison of interpreted
data to known standards.
Display of experimental data.
Cataloguing of experimental data.
D.
Experimental Management
Reset of experimental facility to
next condition
Real time interaction between the scientist
and his experiment
o
A laboratory is also defined as a facility in which various levels of
instrument (sensors) usage can take place. In addition,
at least
80% of the instruments found in a laboratory fall into the spectrometer
category.; t. e., they produce a graphic output as shown below which
is characteristic of a material/compound being analyzed.
Concentration
Energy
Etc·.
Wave length, mass #, energy distribution, etc.
Common Instruments
These instruments in a research laboratory can be involved in
activities such as:
Pilot Plant Studies
Quality or Production Control
Testing (Components or Systems)
M~terials Analysis (Analytical Chemistry)
Basic Research
It is therefore possible to find anywhere from one or multiples of one
to twelve different types of instruments in any given laboratory. For
instance, a petrochemical laboratory may house multiples of a doze.n
of the following types of instruments. The mix is dependent upon their
concentration on a particular classification of chemistry .
•
- 3Types of Instruments
c
Analytical
Amino Acid Analyzers
Mass Spectrometer
Nuclear Magnetic Resonance Spectrometer
Emission Spectrometer
X-Ray Diffractometer
Multichannel Analyzer
Infrared/Ultraviolet/Visable Micronove Spectrometers
Electron Spin Resonance Spectrometer
Spectro Photometer
Ph Analyzer
CHN Analyzer
02 Analyzer
Gas Chromatograph
Mechanical
Tensile Strength Analyzer
Dynamometer
Film Thickness & Hardness
Vibration Table
c
Medical/Clinical
Electrocardiogr aph
Electroencephlograph
Electromyograph .
Electrophoresismeter
Auto Analyzers
Coulter Cell Counters
Cell Scanner
Nuclear
Pulse Height Analyzers
Spark Chambers
Film Digitizers
Neutron Time of Flight Analyzers
X-Ray Flourescense
Electron Spectrometer
Beta, Gamma, Alpha Counting Systems
Other
C
•
Viscosimeters
Refractometers
Hardness
Flame Photometers
3f7
':I
J
1<111
I
- 4-
0 ',"
,"
Incentives
The incentives to automate the laboratory are both tangible and
intangible ... both benefits are clear advantages over any alternate
automation approach.
General
I
C'
o
a.
All the simple and obvious research has been completed.
Today scientists require elaborate experimental facilities
which may take years to plan and/or build. The ideas
behind the use of these new experimental devices must be
tested within a one or two year period after completion of
the facility. The experimentalist feels compelled to publish
his findings within that period or he may be pre-empted by
someone else taking a possibly more advanced approach.
b.
Because of the precise timing requirements of the experiments
conduct/control or the quantity of data generated, the computer
approach is the only reasonable way the experiment can be
accomplished ... some experiments generate thousands of events
which have to be individually analyzed to substantiate statistically
the existence of a new material.
c.
The computer can tell the experimentalist when things may be
going wrong ... is the experiment data valid? Also, the cost of
running experimental equipment may be significant. .. an on-line
computer may prevent malfunctions in its operation, thus
increasing its safety level as well as availability to the researcher.
In addition, the speed of the computer will allow optimum usage
of the experimental subj ect. .. as in the case of a fast changing
radioactive isotope which is decaying into other materials.
d.
The digital computer! s availability and versatility now allows it to
simulate experiments common to the physical sciences. Such work
often uses higher order languages which are written in terms understood by analog computer users or people used to solving complex
mathematical problems.
e.
The computer system allows the experimentalist to obtain improved
accuracy either directly by means of the A/D converter and the
!!absolute!! nature of the digital computers problem solving capabilities
or indirectly through the statistical treatment of meager or vast
amounts of experimental data to 0 btaiI1 correlations. It has been
stated that instruments have had their accuracies improved by one
to two decades through the use of the digital computer.
- 5-
o
Specific
a.
Assurance of greater precision for data acquisition, data
analysis and experiment control.
Data handled by a LAB System in digital
form (assuming a properly designed front
end) will provide significantly greater
accuracy than in conventional techniques.
Conventional Techniques
Method of Recording
Instrument Error
Readability
Strip Chart (inked)
1/4%
1 part in 150
Oscillograph (heat sensitive)
2%
1 part in 50
AM Magnetic Tape
3-5%
FM Magnetic Tape
1.75-3%
PDM/PLM
1-2%
o
Note: When the total data acquisition routines associated
with conventional techniques are conSidered, the
accumulative errors have been as high as 14%.
Significant additional errors are usually introduced
in the system through manual analysis of the data
and control of the experiment.
LAB Systems Techniques ('with a suitably designed front end)
operate with an instrument error in the order of . 01-. 1% and
has a readability of 1 part in 16, 000.
b.
Consistent (24 hr. /day) experimental observation and control.
The sampling of sensor based data can take place under
programmed computer control, on an interrupt basis at
varying time intervals: 1 sec., . 1 sec., .01 sec., .001 sec. ,
etc., or on a demand response basis. The important point
is that the computer can faithfully take data on a fixed or
variable basis presented by the experimental conditions during
the life of the experiment.
C
I,
-,
- 6-
o
c.
Decreased incidence of lost or worthless data due to instrumentation
or system malfunction.
A typical input from an instrument may contain valueless
data. In fact, a very large fraction of the data generally
falls outside of the areas of interest to the experimentalist.
The comparitor circ.uits or editing routines of a LAB Systerr.l,
can insure that only data within limits is digitized and stored.
In comparison when blind recording techniques (FM or analog
magnetic tape) are utilized, it is common to witness periods of
lost data in large blocks. Digital syst(;ms can be programmed
to self-checking.
d.
Release of scientists from routine tasks such as data taking,
scaling, calculating, experiment control.
A typical single crystal analysis may require 2, 000 different
adjustments to the goniometer/scintillation counter head; in
addition in excess of 100, 000 pieces or data must be extracted
from an analog record on a strip chart recorder. The
magnitude of this problem is obvious, considering that the
maximum manual digitizing rate is 8 points/minute excluding
the goniometer control problem. A LAB System can do
these tasks automatically.
e.
Increased control in the conduct of the experiment.
Research at times operates on the fringes of known operating
limits. It must push the state of the art. .. enter the unknown.
For examples, there are today complex analytical instruments
associated with small reaction chambers that operate at high
termperatures and pressures. The computer monitoring of
strain gauge placed on the exterior of the chamber will allow
the experiment to proceed under control at the computed
limits of the rated design.
f.
Conservation of expensive materials such as reactants, isotopes, etc.
Often times experiments require exact amounts of reactants
or isotopes which may be decaying into other isotopes at a
rate fast enough to produce erroneous results. Such experiments are enhanced through the use of an on-line computer
which can insert or inject samples on a closed loop basis in
reproductably exact quantities. In addition, if a known or
changeable amount of material must be introduced which
involves time dependency, the on-line computer can conduct
the experiment "flawlessly" within the given time frame and
thereby limit the effect of an isotope r s degradation upon the
experimental results. An ancillary benefit in such environments is that the computer can maintain an inventory of
isotopic materials under use and at hand. This is required
by AEC regulations.
I
o
I
;~
I:,.~
I
- 7-
I
g.
The fringe benefit of obtaining a greater understanding of the
complex research devices/instruments.
Instruments such as a mass spectrometer contaln pO~Ner
supplies, ionization gauges, sweep voltage controls, etc.
The mOnitoring of the criti~al points in a mass spectrometer's
components will allow the researcher to quickly determine if
the equipment is Ifin phase II, i. e., for a given situation (setting)
are all other subsystem outputs realistic. The data correlation
power of the computer will allow quicker debug time of the
instrument. In addition, a history of machine operation/malfunction
can be retained in disk memory to pinpoint wear in the components
of the instrument.
h.
All experimental data is in a retrievable form.
Data in a digital form, on any recording medium (except printed
on paper), i. e., magnetic tape, disk files, data cells, can be
quickly retrieved for analysis purposes at a later date. Also,
the data can be added to during the experiment. Such capabilities
are required to identify materials from a large data base such as
ASTM tables. These tables are used today in an off line manner
to identify compounds.
i.
A "Quick Look" at experimental results:
The results of an experiment's progress may apparent from
raw data. The on-line computer will allow the research to
analyze an envelope of data (a sampling of data over a given
time frame) and print such data on a typewriter or display it
on a CRT. This ability proves invaluable especially in complex
experiments where computed theoretical results can be compared
with the analyzed Ifquick look" results. Such information may
cause a complete redirection or reappraisal of the experiment.
j.
"Self-Checking" Capabilities
The on-line digital computer is a self-checking device ... the
only absolute device in the experimental system. It has the
capability, if properly programmed, to impose dummy voltages,
bit patterns, etc., in control circuits, and by putting the normal
control element in a hold or bypassed condition, check the correct
functioning of all components. The same program can even Ifexercise"
the control elements during an experiment Similar capabilities
allow the on-line system to standardize instruments automatically
on a timed or as required basis to eliminate base line drift problems.
k.
Operator Training
Routines ~an be written which will allow the new analyzer
operator to simulate the conduct of an experiment. As systems
become more complex, such capabilities will prove more useful.
Training can utilize the computer's spare time.
3qJ
o
c
o
I"
- 8 -
o
1.
Exact Time Correlation
The correlation of time with experimental procedures is an
often over looked problem. When recognized it becomes an
expensive problem to solve. Even if strip chart recorders
have time printed on them it is impossible to accurately relate
i~ to manual experimental procedures. On the other hand, the
digital clock in the on-line computer can allow all phases of
experimental conduct to be time correlated to the microsecond
level if necessary. Time can be in relative or absolute units.
m.
Batch and Time Shared Usage of the Central Processing Unit
The normal laboratory use of a computer today is in a batch
mode, i. e., the experimentalist takes his deck of cards to the
computer center for solution. NOVI programming systems exist
which allow a single computer to provide batch capability while
at the same time conducting on-line tasks such as data acquisition
and closed loop control of experiments.
n.
Efficient Facility Usage
The. ability of the on-line computer to compute results of an
experiment within minutes after its completion can greatly increase
the utilization of complex analytical equipment.
o.
New and Timely Experimentation
This incentive cannot be over looked. The technical world
changes very rapidly and what is research today will be routine
in the future. Instrument design and data rates are changing.
Experimental procedures are being enhanced by graphics, information retrieval and on-line interaction of the researcher 'vvith
his experiment. Only the computer can provide this flexibility
on a controlled and timely basis.
I
I
o
392
- 9-
o
Systems Approaches
The laboratories which are automating today are taking four systems
approaches.
1.
1 (one) instrument - very small CPU - 1 (one) researcher
4K-8K
DVM·I CPU
J.i.nstrument
Paper Tape
I/O
I
( Largely in basic
research environment)
Considerations
Programming in rnachine language
No growth system
Unsupported
Low initial investment
Dedicated system
2.
Multiple instruments of same type - small CPU - 1 supervisor
Card or Paper Tape
I/O
Instruments
~--I
------+
I
Ana~~Tg
I
MPX -lADe
I --
r---,-------- 'm' typewriters
r----'----
CPU *
I
*(1130 or 1800)
~
(Largely in QC
lab environment)
Considerations
Easily customized front end
lOW ClJu usage factor
fixed programs, low demand
difficult to add different instrumen.t
on CPU
fixed time slices in data acquisition mode
easily justified
easy to support under a TSX
0
- 10 -
o
3.
Multiple instruments of different types - small CPU - many different
users
Instruments
Analog
MPX
""""'"
ADC
Channel I
bigital
MPX
-
Card
I/O
I
~agnetic Tape~
CP~
1Disk)
Ix-Y Plotter I
I
CRT
Instruments
*(1130 or 1800)
Configuration for multi- purpose
laboratory facility
'0
Considerations
Capable of satisfying many
different users on a demand
response basis
Possible to dynamically
adapt to experiments
High initial cost
More sophisticated programming
required
Remote link to 360
Capable of more than
data acquisition tasks
o
314
- _·····u ._-.
·····"··~·~~·1····\-
·j-"· .. T......ltt'· ..•..•• ,
."..• ,...
,m
'1
1
- 11 -
4.
o
Total integrated laboratory information system
Time
Sharing
Terminals
Data
Cells
Insts
IDigital MPXI
I
I
Insts.
Analog
MPX
~
slow speed expo
up to 50K accrued
data rates
ADC
Small
Scale
CPU*
(
J
S/360 - 50 - 75
~I
*(1800
1130)
1
Disk
Files
CRT
Medium.
Scale
CPU
(S/360-44)
ADC
0'
Analog
MPX
( ith speed experiments
up to 500K words, accured
data rates)
For large laboratory /
corporate complex
CONSIDERATIONS
Total approach
Programming requirements are complex
Capable of widest
range of laboratory applications
Price is high unless remote S/360 is
time shared
Fast turn around time
Offers large CPU for total
problem solution
o
I
- 12 -
o
Modes of Systern Operation
Each different type of laboratory could conceivably write its OiJJIl unique
monitor system to fit the various levels of instrument usage which is
conducted there. The ~ive levels of activity found in typical laboratories
is shown in Table I.
Relevance of 1800 & 1130
The ability of the 1130 and 1800 to satisfy the demands of the LAB System
applications can be summarized as follows:
1130
Hardware:
The OEM and SAC channels on the 1130 allow this machine to accept an
analog or digital I/O (instrument based) front end. At least two manufacturers have announced A/D converters which can tie to the 1130.
o
Software:
No IBM TSX monitor exists for the 1130. However, it is known that at
least three users are writing I/O subroutines to support an ADC front
end. It is also known that two 1130!s are tied to x-ray and neutron analyzers
in a closed loop fashion under some monitor system.
1800
Hardware:
The I/O flexibility of the 1800 makes it an ideal growth system for laboratory
automation programs. The newly announced digital multiplexor R~ allows
data to stream directly into assigned core memory locations on a demand
reponse basis at data rates of 50K words/second. Thus each experimentalist
will think he has a computer always available to him during his asynchronous
method of operation.
Software:
Up to 50 chromatographs at one user!s laboratory are no\v operating under
the 1800' s TSX. The GCs are connected to the 1800 via the 1894 System
Engineering L...L~l.B special automatic gain change A.DC front end.
o
- 13 -
o
Laboratory Automation in IBM
IBM has now in progress at its own technical centers a number of
laboratory automation programs in support of its own research
requirements.
An 1800 at the Thomas J. Watson Research Center is tied
to an x-ray diffractometer in a closed loop fashion. In addition, a plate
film reader is also connected. At IBM's San Jose Research facility an
Electron Paramagnetic Resonance Spectrometer is tied to an 180:J also
in a closed loop mode. In both cases, the 1800's are controlling :he
experiment. Both facilities plan the attachment of additional instr G.Jnents
in the future.
These IBM research laboratory automation efforts have generated two
differing operating systems approaches to facilitate the interaction
between multiple different instruments and the 1800" The main features
of these operating systems is a so called It Lab Director Program lt w}iich
satisfies the demands of slow response analytical instrurnents in the basic
research environment.
These LAB directors are written to allov! the computer to respond to
the various tasks associated with each type of instrument in an efficient
manner; i. e., not under complete TSX control. So long as no interrupts
arise, the LAB director allows the 1800 to handle multiple different
instruments with relative ease. Modes of operation under the LAB directors
are as shown in the following diagram.
o
317
1
Summary
'c'. jjI
I
In summary, one finds the researcher or other lab personnel in the
I
process of modifying the existing TSX for the 1800 to adapt it to his
particular laboratory needs. Type III application programs for the
gas chromatograph, mass spectrometer, and Instron tensile strength
tester have been written by IBM and run under TSX. A plea is made
to COMMON members to put in PID their instrument oriented application
programs.
I
1
I
C!
I
I
1
1
Table I
PILOT PLANT
OPERATION
QU.l\LITY
CONTROL
Basic operational
philosophy of online computer
frequent changes in
a large math model
prograui - low
batch demands
infrequent changes frequent changes
to program, com- in small
puter dedicated ... programs
no background
infrequent
tfrequent changes in
changes in
~rograms associated
analysis routines, vli th wide variety of
frequent r.oll in & experiments
0
out of analytical
subroutines
~
Instrument
Characteristics
Slow response, wide
variety of II process II
instruments plus few
analyzers
slow response,
multiple instruments of one or
two types
wide response
range, mainly
slow. Generally
one instrument
type to satisfy
slow response
insts., are
gener al wide
variety of
analytical insts.
General philosophy
of system operation
to get the most out of
the pilot plants
performance
to get the most
out of the insts.
with as few
technicians as
possible
to recognize
to get the most
to get the most out
faults in products out of instruments out of researcher
final assembles
& components
as soon as possiblE
Turn Around
Requirements
Seconds
Minutes
Seconds to Hours
Computer Tasks
data acquisition
optimization
Monitor
Characteristics
Instrument Usage
Factor
COMPONENT
TESTING
MATERIALS
ANALYSIS
BASIC
tRESEARCH
wide range of inst.
response + wide
variety of instrumen1
Minutes
~illiseconds
~econds
data acquis.
data acquisition
performance
limi t checking
report generation summary report
generation plotting
data acquisition
data cleanup
table look up
report generation
data acquisition
klata analysis
data storage &
fretrieval experiment
control
none - process
demands will
use computer
scan instrument
at fixed time
interval - no
interrupts
scan instruments
at fixed time
intervals - rare
interrupts
scan instruments kiemand response
at fixed time
fruode necessary ...
interval when
no time for monitor
overhead
CPU is interrupted
100%
85-90%
100%
50-60%
10-15%
GENERAL OPERATIONAL CHARACTERISTICS OF LABORATORY O'PERATION
0
0
0
to
:1
SESSION NUMBER
F.l.6
SPEAKERS
LAURA AUSTIN
DISCUSSION
A CALL WAS ISSUED FOR PROGRAM CHAIRMAN AND LOCAL ARRONGEMENTS
CHAIRMAN. BOTH WERE LOCATED FOR THE PHILADELPHIA MEETING FOR
SEPTEMBER, 1968. WE STILL NEED BOTH FOR CHICAGO IN APRIL, 1968.
FUTURE SITES WERE LISTED AND DISCUSSION WAS HELD REGARDING CONTENT
AND FORMAT OF MEETINGS.
I
j1
C t
I
11
40/
II
wt'¥tt:!:r! tl'tt'rrm'MYrT :-nne-nnpnrr
o
SESSlnf\1 NUt-ARFR
F.l.7
SPEAKERS
DANIFL J. LAMPONE
en L Af\I n f\1 A 1 A 2 0 en MPIJ T E R PRn GRAM F n R A. c. c: IRe U I T
A[\1 ALY SIS I J TIL I Z I [\1 G T R Af\1 SIS Tn R Y PAR AI'" F T F REO lJ I \I ALE N T
CIRCUIT MnnFLlf\l(; TFCH[\IIOUES
E 0 M\1 f\1 n F.
o
W&22&;;;ua tt.MWi4;
I
¢! t ffiJ$, #4T¥4#¢
,44 -
.... 4$
10
August
28, 1967
o
A 1620 COMPUTER PROGRAM FOR A.C. CIRCUIT ANALYSIS
UTILIZmG TRANSISTOR ny" PARAMETER
EQUIVALENT CIRCUIT MODELDfG TECHlIQUES
by
Daniel J. Lampone
and
Edmund E. Colan
o
~ral ~~
Sylvania Electronic Systems,
Sylvania Electric Products Inc.
Williamsville, B.Y. 14221
I
o
I.
I
I
i.i:tti::i'HTtrtWWeruw::nr'
o
TITLE:
A 1620 Computer Program for A. C. :ircuit Analysis Utilizing
Transistor 'ty" Parameter Equivalent Circuit Modeling Techniques
Authors:
Daniel J. Lampont:::, Sr. Engineer
Edmund E. Colan, Research Engineer
Sylvania Electronic Systems-Centra.1
Wehrle Drive and Cayuga Road
Williamsville, New York 14221
I.
INTRODUCTIO~
This paper presents a simple but accurate method of obtaining circuit
A. C. analysis by use of the IBM 1620 computer.
gram utilizes the transiator in its
'~"
The unique feature in this pro-
parameter equivalent circuit and some
circuit transformation techniques to allow insertion of program input information in a normal topological manner.
The need for the use of "ytt parameter
representation became apparent when problems were encountered using a predecessor
program employing mesh current analysis and the transistor "h tl parameter model for
o
high frequency problems.
High frequency representation of the uh" parameter model
utilizing discrete components for the complex quantities created problems in circuits involving tuned amplifiers.
Details of this will be presented later.
Another
reason for resorting to the "Y" nodal analysis was that most vendor's infonnation
specify tty" parameters for high frequency transistors.
In the mesh current method,
inaccuracies resulted when converting "y" parameters to "h" parameters since conversion formulae were approximate.
Before getting into an actual circuit solution,
let us quickly illustrate the basic tty" model and sample nodal setup.
shows the basic four-terminal tty" parameter equivalent circuit.
Figure 1
The use of the "Y"
parameter equivalent circuit is essential in nodal circuit analySis since units of
current must be maintained,
by matrix algebra.
~th
the solution of nodal voltages being the outcome
This is illustrated by reference to the example passive circuit
and nodal equations shown in Figure 2.
o
-1-
404
U MiMI' "d ;, (I
Q.
o
Ii.,
+
-----
~
~
Lf~
V·IJ
(''1'/ I )
~trVO
Y1l)
10
+
~
'to
Vo
(Yzz)
~
0
FIC,URE
o
R,
E2
VV\,'\,..-.-_
+
e
,
C,
E3
I
CI
EQUATIONS:
EI (ZG, +
E, (- c;.)
t
uJ
Cl)
+
E2. (-
+
E2(bl+GL+i wC d
Gtd
+ E3
(0)
::: e c;,
+ E3
(-1 we,)
:: 0
(}
FI(;URE 2.
-2-
4o~
o
The Matrix solution is
[EO]
[Y]
=
[rJwhe:re the coefficient matrix is
made up of admittances rather than resistances in a mesh curr$nt solution.
The
right side of the t!quation is current sources associated with respective nodes.
II..
A. C. EQUIVALENT CIRCUIT SETUP
This disoussion will be centered around a sample RF amplifier circuit
shown in Figure 3.
ly. is.
The cireui t has been simplified for purposes of the
aaa~
Decoupling networks) d-c biasing components, &.no. AGC circul try -were re-
moved since they have no effect in the a-c analysis.
The circuit consists of
cascaded pairs of transistors and four tuned circuits (varactor tuned).
shows the a-c equivalent circuit with all circuit components and
t~
Figure 4
pa~meters
modi-
fied by "turns-ratio squared" factors to allow tor elimination of circuit ideal
transformers.
c
It would be best to interject at this time how to arrive at the a-c
equivalent circuit since the success of the solution hinges greatly on the correct
methods used at this point in the analysis.
It was necessary to do some hand cal-
culations because of the limited capacity of the 1620 computer or else we would
be into the category of a highly complicated a-c
anal~sis
used in the more pub-
licized and more automated programs allowed by the larger computers.
Certain
steps must be followed in setting up the a-c equivalent network in order to show
how to arrive at Figure 4.
First, starting with the original Circuit, substitute
the equivalent circuit of the transformers.
Most methods for these were taken from
"Communica.tion Networks", Volume 1, by Guillemin.
The obvious intentions were to
arrive at an equivalent tiT-network" and the following example and Figure 4A shows
the steps taken.
o
-3-
~1o~
Q42.~3904
r(.
•
!
•
0
1
RF
"
INPUT i
f\I'
I
fi
I:
FIGURE 3
.~
R.F AMPLIFIER.
~
•
.J:""
•
o
c
c
~ft
I
I
(I)
CI) .. 284
I
~ ,-
Z
®
.089
O-~--
'1·
QS's) z
I
- .03b4
r---T"~
~ so
eve
..=("
!
t~lT
L-
4 .Ot f
2\30
.0301
•
•
•
32.
Y22E
ff
_
----·(.313)Z
•
•
.. ,.
.0092.1
-.003"4
14.7 .0316
_
_
(.0154)l
.154
@ .0141
.00491
@ .0487
@l~
(j)e.
1-
,
~1---:368Z
Z3t
.0034
__ _
-.G02ab@)
P
I.
1
.0090S
3.2.
8.06
YrIB
t
.00334
1.785
YZ 1V7
a,
~.
I
I
a.t.'
I
~
':)
I
Q..,
a-Z.
:.
2.
Q2..
I
;1 a.: u..i ::
1
I
-3'':\= 3.2.
''';
I
\
.0154- :;
I
V1
I
Y,2. E
:; (C
'(ZIE
= (5'
-1 .\ Z.)
+i
,'- )
)( \ 0 -
VI~
~5
Eo:: .bO'
i
= (~4. 5' - 43) )( 10- 3 MHO S
'fIle::. (-:27 -1,51)
Y"
3 MHDS
-1 13 )
T'ZlE :; (.OlS"t 1,4)
8.0~
2N3904 AT 56 MHz.
2N 3933 AT 56 MHz.
'(II E :: (I. I +
I
a.3 - .'394' 1.42. '.2'2.'2. =
FIGURE 4
-_._---A.C. EQUIVALE.NT (,IRCUIT
e
I
YZ.I 6 :: (- 6,0
+i
I2.LB : (.2.7
+ 1,84)
49 )
~o
I
t+
8
J... . ....L • .L-
\
a2. - .394' \.42. :: 1.785
IB.4
·-1450
581
-rf
.OO;~3
..L .
.00286
I
,
I,
o
---II"-
bll
= Rl
+ jYlL l
b22
= R2
+ j wL 2
b12
= JwM
= b2 1
Defining the Mutual Inductance as Negative Numerically
(Arbitrary; for convenience only)
we can say:
bllll - b12 1 2 = El
-b2lI l + b 22 12 = E2
One type of network which can be made equivalent to the transformer
is shown below.
c
It consists of the T-network and an ideal transformer.
Various special forms result by giving the ratio "at! of the ideal
transformer particular values.
ZB
Gt.
o
V ~J~ ~
O~------~--------~
--
.----....
I l.
lDEAl
TRANSFORMER
Figure 4A
-6-
c
i
II
o
}t'or the T-structure alone:
ZcI2
(ZA + Zc)I 1 - -a.-
or:
(ZA +
Zc) 11
Zc
th~s
Zc
- -;- I2 ::: E 1
ZB + Zc
- ;- 11 +
If
= E1
a2
12 :: E2
network with the ideal transformer is to be identical in behavior to the
original transformer in Figure 4A, then,
Zc
a
=b12
ZA + Zc = bll
CI
which gives
= Rl + jwLl - j:waM
ZA = bll - ab12
ZB
= a 2b 22
::: a 2 (R2 + JWL2 ) - jwaM
- ab12
Zc ::: ab12
==
.1wa.M
Therefore we have the circuit in Figure 4B:
R,
L,
o------'V'Nv---~--
2-
-~M
a. Rz.
0--------.-----..".-.----..,,---"---.. . --.-.."----..- " - - . - - - ·
o
-CLM
L - -_ _ _ _~'VV11fV----J
Figure 4B
Q
IDEAL
XFORMER
.'{
-1)0
-
One must keep in mind that the transformer conversions need only be
d.erived once and retained for reference in anticipation of future problems Involving transformers.
o
Initial preparation of these equivalent networks of
transformers is painstaking but nevertheless valuable.
By assigning any value to the ratio a, we immediately have the lmpedances in the T-structure to be used in conjunction with an ideal transformer
having this ratio in order to replace the given transformer.
A special case: when we make a = I the ideal transformer vanishes
(this is the case for a 1:1 transformer).
Then we have:
ZA
:::
bl l
b12
ZB
::::
b22
b12
Zc
=b12
c
and this is:
ZA ::: RI + Jw(LI - M)
ZB
= R2
Zc
=
+ jw(L2 - M)
JwM
This result is illustrated in Figure 4c as follows:
(LZ- M)
Rz.
Ow:. I
o---------------~!-----------------~o
Figure 4c
-8-
411
o
Our original circuit also showed step-down and step-up auto-transformers.
For purposes of brevity, the solution for the equivalent circuit of the autotransformer follows without detailed discussion 1n order to approach the next
point.
R.,
L,
---... 1z.
0
Rz.
E2.
L,
bl l
=
Rl + R2 + jw(L1 + L2 + 2M)
jwL2
b 22 = R2 +
b12 = b21
=
R2 + jw(L2 + M)
II
--.-
1.. A l B
I,
----- a...
~------~Iv\~--'---~AAAr-----------
o
a...
-9-
o
per above:
= b
a
= b
12
22
= b
+ jwL
2
= R
21
2
= R
2
+ j w( L
2
+M)
Substituting:
Zc = a [R2 + jw(L2
+M~
ZA :: R1 + R2 + jw(L1 + L2 + 2M) - a.
~2 + jw(L2 +M~
~ = a 2 (F2
MIT
+ j'wL2 ) - a [R2 + ,jw( L2 +
Impedances are:
Zc = aR2 + a.L 2 + aM
ZA = R1
ZB =
(1-a)R 2 + L1 + (1-a)L2 + (2-a)M
2
(a 2 -a)R 2 + (a -a)L2 - aM
+
which yields the following equivalent circuit shown in Figure 4D:
RI
(I-a.)r<.z.
L,
(I-Cl)LZ
(z.-a..)tv1
2
.
(a.,2-a) Rz. (Cl -a.) LZ
0----1\IV'v---'WVV-
-a.W1
(a..)
r-
j
0
~: ~
Nz.
1DEAL
TRANSFORMER
Figure 4n
Step-Down Autotra.nsformer
-10-
4)3
0
o
As you can see, each equivalent circuit of the transformers includes
an ideal transformer.
ryrh~
next step is to eliminate the idea.l transfonners in
order to avoid discontinuities in the circuit solution.
ideal
tran8~ormers
In order to eliminate
all components succeedins an ideal transformer must have im-
pedances modified by the "turns-ratto-squared" factor
[=~]
2.
Now, looking back at Figure 4, the circuit was split up at four different points where the ideal transformers should have existed.
For this circuit
the following turns-ratios existed for the four transformers:
81 =
.394
82 =
1.42
a3
= .222
84 =
4.88
As shown on Figure 4, the first section impedences succeeding the first ideal
transformer get multiplied
and so on.
o
section.
by
.3942
= .155, the follOwing section by
~ 1.42)2( .155~
Notice, the multiplication factors become cumulative for section after
This is an absolute necessity since the source generator must see all
components in the proper perspective.
Conversely, it becomes necessary to modify the t'Yil parameters of the
transistors (admittances) by multiplying Yll and Y22 by 1/82 , Y2lv l and Y12V2 by
l/a.
This can be understood easier by observing the following calculations.
For the 2N3933 of the first cascoded pair:
1
:: -1 - = 3.2
.313
Now:
Y1lE = (l.1xlO- 3 + J 4.2xlO- 3)3.2 :: 3.52xlO- 3 + j l3.42xlO- 3
Y12E :: (OxlO- 3 - j.16xlO- 3)1.785 = 0 - j.285 x 10- 3
o
Y2lE
Y22E
= (5lxlO- 3 = (.015xlO- 3
J 13xlO- 3 )1.785
=
9lxlO- 3 -
j
23.2xlO- 3
+ j.4xlO- 3 ) 3.2 = .048xlO- 3 + j l.28xlO- 3•
-11-
I'
,I
o
The same multiplication factors apply to the common base 2N3904 of the first
ca.scoded pair.
For the second cascoded pair, the same process is carried out, only
now the multiplication fa.ctors obviously become:
::
1
2
a 1 a-2 2 a 32
=
1
.• 394
x
1 x l : : 8.06
1.42
.222
65
We now arrive at what looks like Figure 4, the a-c equivalent circuit.
The foregoing set-up seems a little messy (yet straightforward) but after experience with a couple of circuits, adeptness, efficiency and accuracy are very much
enhanced.
III.
PROBLEM AREAS
Minor problems existed and we will now.discuss them.
Originally, it was intended to solve this circuit by the loop (or mesh)
current method using the "hybrid" equivalent network for the transistor.
The
"hybrid" equivalent network is show in Figure 5 with a few rules governing its
insertion into the circuit,
This "h" equivalent was derived from information
contained in the text "Transistor Circuit Analysis and Design" by F.C. Fitchen.
The information in Figure 5 is only included as extraneous information
to serve as a guide in cases where choice is made to apply a mesh current solution,
for example, low frequency analog circuits.
As I previously mentioned, most vendor's information gives "yu parameters
for high frequency transistors, and these are given in complex quantities.
Con-
verting to "h" parameters is necessary to perform mesh current analysis and these
obviously become complex.
Utilizing discrete components for the complex quantities
0
-12-
1
CO LLE C. TO R.
EMITTE.R
Circuit Insertion Rules
1_
hrb-v cb is always same as,v
2.
Polarity of hfb .ib must always be related to vi as follows: either,
hob
both must be positive at collector and ba.se (terminals) respect1v.ely
i
in sign.
(as shown in sketch), or both nega.tive at these terminals_
Base current
(i b ) must·be consistent with vivcb is interpreted as voltage at collector relative to base. Polarity
hfb
of the voltage generator ( - • i b ) is indicated (as determined in 2).
hob
Voltage drop and polarity in resistor l+hfb is determined by direction
hob
of current flow assumed for i c •
4.
When inserting parameter data into equations, dontt forget that h fb
is negative.
5.
The generators will take on polarities with respect to loop currents
according to the following manner:
+
- -......-
...
o
Loop current
Loop current
Figure 5
-13-
created problems
cities.
~lere
tuned circuits could interact with the transistor capa-
Reference to the grounded base configuration in Figure 6 will illustrate
o
this.
EM\TTER.
COLLEC.TOR
L2.
Figure 6
The capacitor C1 and resistor Rl are the components resulting from the conversion.
During compnter solution, Cl' which is small, resonated with the combination of
C2 and L2e
C2 and L2 formed a high value of XL rather than a high resistive
value for the tuned circuit.
Consequently, while attempting to tune a circuit
for maximum output voltage, the solution proceeded to find the peak series resonant point rather than the antiresonant point which is desired.
The series re-
sonant point CRuses a much higher output 'Volta.ge, Eo' than the parallel peak.
(This can be demonstrated mathematically).
This is a false resonant condition,
even though Eo is higher, because of interaction 'oetween transistor para.meters
and actual external tank eircui tS e
rPhe series resonant peak is close to the
antiresonant peak and the skirts of the response curve blanket any identification
of the actual circuit parallel resonant condition.
Conversely, a solution for
antiresonance by searching for the minimum loop current does not succeed because
C1 eliminates the occurrence of the current "dip".
-14-
4)~
c
Attempts to utilize a current generator in the h-parameter transistor
model with paralleled
~2
components were not successful.
Introduction of the
current generator - or any equivalent alteratiQns -, could not be made co. .tible
with the loop analysis.
Therefore, without the success of using the "htt equiva.lent loop (or mesh)
current ana.1ysis, we proceeded to investigate the tty" equivalent nodal analysis.
Some nodal analysis problems also occurred but were "ironed" out after
some proper manipulation.
It so happened that with the circuit previously shown
containing two (2) pairs of cascoded transistors, it was found that the input
admittance of the second transistor in the pair (the grolmded base) loaded down
the groilllded emitter transistor output so tha.t a loss resulted rather than aga.in
across the pair.
This problem was circumvented by combining the transistor data
so that two cascaded transistors are combined into one single model to represent
the cascoded pair.
The combined admittance matrix takes on the following form:
Y
yl
=
Yl2B
12E.Y2iB
Yo
where Yo
= Y22B
-
Y2lB Y12B
Y22E +y llB
The new combined fly" model of' a. common emitter - common hase pair has its elements
varied by the a.bove formulae (the y lll~ and y 2ll~ remained unchanged ohviously).
Another small problem that occurred in the process of detennining the y-parameters
for a cascoded pair is where the grounded emitter transistor has a degenerating
o
emitter resistor.
shown
The emitter resistor is combined by matrix manipulation as
in Figure '7.
-15-
o
c
A
FIGURE
7
-16...., ..
"
l,1 ( . .
~.
.
.
~
o
A short computer program was written to accomplish what is shown in
Figure 7.
The y- parameter data initially used was the transistor data modified
by the "turns-ratio squared" factors to fit the A.C. equivalent circuit with ideal
transformers removed o
This modified data was then combined with the emitter re-
sistor value to obtain the resultant matrix for each cascoded pair.
however, yield.ed improper pa.rameter data.
This method,
Oddly enough, this particular problem
was rectified by reversing the order of solution.
By first combining the basic
transistor data with the emitter resistor and then modifying by the "turns-ratio
squared" factors, the proper data. finally resulted for the ca.scoded pair.
By combining the transistor data. for each cas coded pa.ir, the circuit
reduces to a 12 node problem.
Figure 8 shows the new revised. areas containing
the resultant combined "Y" parameters of each cascoded pair.
The emitter re-
sistors have now been combined into the new "yU parameters representing each
c
cascoded pair.
r.J •
COMPUTER PROGRAM
The final category that will be discussed is the actual computer input
data set-up.
With most people this should be elementary since matrix algebra
determines the final solution and a little knowledge of circuit analYSis can
aid the programmer in inputing the information.
The technical portion that I
have previously discussed would probably need the assistance of an engineer to
properly set-up the a-c equivalent circuit.
But again, practice with a few
circuits would lead to more familiarity and shorter turn-around time.
An explanation of presenting the input information follows.
Byre-
ferring to Figure 9 and Figure 4 with the combined cascode pair configuration of
Figure 8, the input information is easily understandable.
o
-17-
1 ~ PAIR
o
.02.0
,0521.
WAS®
WAS@
®
®
-.0174
a...-..,f\FV'ln--.I
WAS®
14.7
0)0
.031' .0117
-.00364
-........
(-.1'3+11.15)
10-
3
•
[( 88.2 -
i 3 2..3) 10-3 ]
Vs
2. ~ PAIR
WAS
0.)
(j)
WAS
WAS@
.0092.1
.0034
~
-,00701
(68.8
+i
l.7B) 10-
®
®
WAS
.,54
@,
@)
~----~----~----~.
-,002.03
- ,DOlE>
@
(-3.3 +
3
FIGURE
i 23.3) '0- 3
8
c
-18-
420
o
Read cards order:
1st card:
N and Frequency
where N = number of nodes
2nd set of cards: II, JJ, R, XL, C, rTYP (Impedance Branches On~y)
where: II
= matrix
row element (node number)
JJ = matrix column element (node number)
R, XL, C = resistance in ohms, inductance in henries,
capacitance in farads associated with II, JJ
impedance branch.
=0
ITYP
if symmetrical, 1 if asymmetrical
All impedance branches are symmetrical.
The YV (current generators) branches
are sometimes asymmetrical and will be assigned as such when we present the admittance branches later.
c
By looking at Figure 9 and Figure 4, three branches
are common to node 1 and must be listed separately in the 1-1 position.
The
program combines these in the final 1-1 position in the coefficient matrix.
A
branch need not be repeated when symmetrical and off-diagonal (for instance the
1-2 common branch) i. e., the program makes A (II, JJ) = A ( JJ, II), however,
the off-diagonal elements should be assigned a minus (-) sign.
Also, the program
converts all these impedance entries to admittances 1n complex notation to keep
the matrix solution in proper units.
Notice II cannot be larger than JJ.
3rd set of cards: Follow these cards with a blank card.
4th set of cards:
II, JJ, YR, YI, rrYP of Y parameters (admittances)
where: II = matrix row element
JJ = matrix column element
YR
o
= real
part of Y parameter
YI = imaginary part of Y parameter
ITYP =
a
if symmetrical, 1 if asymmetrical
INPUT
Tt~~O
-
I).
ZZJOB 5
ZZXEQSIMPAD1
12
1
.284
-..:::~ 28£1
.284
2
.189
3
3
- . 1 ~9
-.02q~
~-O6
.189
.0293
E-06
3
-.0034
.003A
-.0034
.123
.0521
-.0521
3
2
2
O.
--------3--· ..-3· "-"'-0.3
3
4
4
4
4
4
4
4
,,_.' '-.- '---'-'4"-'---5
o.
5
6
6
7
5
6
7
7
.0521
14. 7 316
-14.'316
1L'l.'316
7
7
.00905
---....- ... - --.-.- ---""'--7'---7"-' _.--- cY02- 2-"
~-
7
8
- .. 0022
8
8
• 0022
8
8
O.
9
9
9
10
10
10
10
11
11
10
11
11
11
11
12
lC
. -r'c" fT
f\)
;-
12
12
1212
o
I
?~
O.
--;048 7
-.C487
.0437
18.4
5
3.4
6
~.1.?
..
13.8
.206
o
o
153.
-1t::;3.
153.
2130.
85.0
-85.0
85.0
n
E-12
E-12
TUNE
~-12
E~12
'cf""
E-12
E-12
E-12
~} TUNE
o
o
O.
O.
··0"·--------·--..·---------------···----·..-·
d.
o.
o
o
O.
0.
a.
o.
'-::::-'--.--.--~----~.-.-~--.-----
E~06-
v.
.. ...
~
~-.-
o.
O.
1040.
E-12
O.
o.
o--~
.
1500.
o.
E-06
E-06
E-06
E-06
E-86
E-Q6
E-06
O.
o.
o.
O.
o.
o.
o.
O.
E-03
E-06
o
6
o.
o.
.0141
5
o.
O.
O.
O.
O.
.00205
-.Q0205
.00205
.00286
.01404
-.01404
.01404
E-12
O.
O.
.00123
E-06
-.00123 E-06
0.
.0141
-.0141
5'
E-06
E-06
E-06
E"-06
E-06
.00581
F-06
-.00581 E-06
.00581
E-06
.0646
E-06
.0026
E-06
..
-.0026
E-06
.0026
E-06
.00806
E-06
-.00806 E-06
.00806
E-06
.00697
E-06
-.bb 12'3
.154
-.154
- ---------Tb'-·fO ---- .. - .154
r
4.
2
2
o
o
O.
2
2
""
)'
'.J
o.
o.
.0589
-.0589
.0589
.0301
.0293
1
1----2"--
<
NODES
3
56.E+C6
50.
o.
I
RF Atv'lPLI FIER-
E-03
E::-06
E-12
o
o
a
..--;:;-_. -
o
o
o
o
o
..-. --..·--0" ....
0
.-
----·-~-.---_-
____I_".
TUNE..
TUNE:.
o
o
o
o
._- ". ". _..... -----6---.... _. -o
o
o
0
F I C-.UR. E.
PAC.....c I
~
.. ---".-
-,.-~-----
~
q
c
o
-0 0
Cl
-
-0
i •
Ole
n f'
0
()\
OJ
~
(10
(J 0
-,I _.
C
0
I
I
\OJ\()
\0 :\l: \f.j \() '4..i '0
(,!(
C () c' C
I ii' I I I
1J' :IJ. 11. LI.J lL ll.i
ole
("') i
('";
{\j 1:('")
{\:
,{\j
....
ill
.;t
t~~ ;(; II: \[) q' 0
L i (I. C U.i 0 ~'j
(, ~ : (
(' J (I!
-:j
()
C
0 C
- 0 • (.) CJ C •
C I.....
k
(. ~.:;
• tJ • • • • • i • • • •
c. cole c e o
0
I
!
l
I
I
0
t c: ;. (; ~
(~i
lfi
iii t ! I
Li I 11.1 ILl ILl I
...
0
.
If,
0:.
I
I\() \(.) :\() \()
HL! I,; III III
co co
-
I
I
I
i
o
0:0 0 ( ' ) 0 0
I
clo
(1:0 C
0
I
-I • • - - CICOOCO
I
~
Q
c
I
1
W
0
I
(\J
.... '..........
I I I I Ii. I
LiJ W hi llJ tLl
ltl,
")<
I.
C
_; • • • •
,
\:)
~
'\(
I
:t
I
:<.
<:I' ,
!LJ
..;;;t
•
.;j
O.
"-l
0'
!
:\(' ., a (I.! Q'
- {\ • (I. , tre.I.:c
:IJ (
q
• (, I i · I
<\1
•• 'I
U
..... , - .-' (\
cI
(l
("",
c·l/ ......
:x:
f'.1 f'.'
f'. f'.1
..
(\
~.(\!
,.
I
I
....
~I
1-' ("", :t- r)p {\,: ~j~ HI(\!
•
t-• .....
(Y;
c :> C c>
(',
-Ie
c' •
• r) {\i
l'l ! • .... it ..... • I ....
C' c\.) I
.....
Q: ,e'
...... M \(...
r--
I
•
<::It- . ' . - . I- C•
I I I e <:l" ~,; 1"-
.... ' ,.;;:I
-
.... <
~
... ........., ~. ..........
(\; (\
!"""
(\)
......
"
I
c:
_.
(Y'; (\j
;.-.
(\j
(\~
•
(\!
(
-
,
... - ,.... .... -:-
c' 0 c c c
..... ......
- Il)
(\ '" r--.
I ;'
1
<......
l.l
I -
(\,1 , -
I:
11,1 I
• :1'" ,..'
([',e(\
(,(
_ ..... 'c
fa
Ie
- Ill':'
i
-'1.1'- F'"
..... • <:l I'r-- cr, t \{' - F"
-
r. 1"-!I
'e
!
!
•
U,:Ul IJ.:
1
1
It
(1';'
!
I
(',
!
i
1
(Y'l !('
("') r'l (') \(") f') :M ('~ ('" (") i(') f"; iM
C' :e 0 io 0 Ie 0 0 c, C
I
i
; lC (\ " I ! : l f
C'c, a' e e (\ U (\ Il .;;:I Q
c
c,
1
(') i(1'i r, '("'~
t I (Y (") (y,! (') (Y'~ ri I('~
<. ( ( ( c' 0 C
c'
I 1 I; I • I ., 1 I III I I I J
I I
I I 1 1 I I I I f t
l1i IIi llJ lLl 1.1' ti' lJ IJ lJ I.JJ lJ! lJ.i 1.1' lJ.,' II.!, tL ,lll I)! il1.i IJ Ill' Il.' ill. Iii 'lli lI.1LL~ III :ll,' ll.' l.J IJ.'
C'''
C
C
1''"
.-t j......
t-4
-(\!
r,
M f"',
•
o
SFSSIOf'.1 NUMRER F.2.2
SPEAKERS
HARRY CADOW ON 360 OPERATOR TRAINING.
GARTH GROFT nN OPEN SHOP OPERATION IN AN FNGINEERING
DEPARTMENT.
C~I
o
i
I
I
j
O·
o
OBTAINING MAXIMUM RESULTS USING A COMPUTER
FOR ENGINEERING DESIGN
by
Garth E. Groft
York Division, Borg-Warner Corporation
ABSTRACT
For more than five years a major manufacturer of refr i gerat i on and
air-conditioning equipment has made comprehensive use of an electronic
computer as an engineering tool. Its use is directed toward increasing
the product i v i ty of the des i gn and test eng i neers, there by reduc i ng
the cost of engineering.
CI
It is well known that a substantial part of engineering is involved
with tedious and repetitive calculations.
Much of this work can be
done more quickly and more accurately by a computer. Both the quanti ty and qua Ii ty of product des i gn and .deve lopment are enhanced as
faster, more accurate, solutions al low more time to explore a wider
range of des i gn poss i b iii ties and to conduct more extens i ve i-est
programs.
The faci I ity is operated on an "open shop" basis whereby the engineers
have direct 24 hour/day access to the computer and are encouraged to
use it as freely as they do a sl ide rule. Courses are frequently given
to train the engineers so that they themselves can program and operate
the computer.
To date, 94 engineers have completed the courses. This
arrangement minimizes the number of steps in the communication chain,
and has resu Ited ina most effecti ve use of the computer for des i gn
and development work.
This paper discusse~ in detal I the manner in
which an "open shop" operation is achieving outstanding results.
Page I of 6
o
OBTAINING MAxIMUM RESULTS USING A COMPUTER
FOR ENGINEERING DESIGN
by
Garth E. Groft
~ork
Division, Borg-Warner Corporation
For more than five years a major manufacturer of refrigeration and
air conditioning equipment has made comprehensive use of an electronic
computer as an engineering tool. Its use is directed toward increasing the productivity of the design and test engineers, thereby
reducing the cost of engineering.
PURPOSE OF THE ENGINEERING COMPUTER
It is wei I known that a substantial part of engineering is involved
with tedious and repetitive calculations. Much of this work can be
done more quickly and more accurately by a computer. By using the
computer, the engineer is no longer limited to approximations, oversimplified formulas, and short-cut methods that have been in wide
use for many years. The computer can be programmed to include
equations in their entirety and to consider factors that previously
had to be ignored.
The computer provides more reliable, more consistent results.
The
engineer is assured that a computer program wi I I repeatedly handle
data in precisely the same manner - that is to say, al I calculations
are performed the same way each time the computer is used. Hence
al I data has a common basis of analysis. He is further assured that
the computer is not subject to errors caused by distractions, nervous
tension or fatigue.
Both the quality and quantity of product design and development are
enhanced, as faster, more accurate, more re I i ab Ie so Iut ions a I low
more time to explore a wider range of design possibi lities and to
conduct more extensive test programs. The computer has contributed
significantly toward reducing the total design and development time
cycle. The effectiveness of the computer in saving time is emphasized by the fact that the work being done by this installation is
equivalent to an additional staff of more than 40 engineers using
sli,de rules and desk calculators.
Page 2 of 6
c
o
GROWTH OF THE FACILITY
The first engineering computer faci I ity was an IBM 1620 Model I Paper
Tape System, instal led in December 1961. Acceptance of the computer
as an integral part of engineering was exceptionally favorable. This
was due largely to the extensive personnel conditioning, training,
programming and general preparations which took place before the
equipment arrived. The demand for the computer quickly exceeded expectations as wei I as its capacity, necessitating upgrading just a
year later, to a faster, more efficient card system. The core storage was also increased from 20,000 to 40,000 6-bit digits.
This
provided sufficient thruput unti I early 1966. Additional capabi I ities
were then provided by converting to a 1620 Model II having twice the
speed of the Model I. This system provided more effective operation
and permitted a higher level of programming capabi lity. A later appraisal of future engineering computing needs and proposed improvement
in test data collection equipment and techniques dictated the further
change to an IBM 1800 Data Acquisition and Control System. This was
instal led in Apri I, 1967.
The 1800 is ideally suited to engineering use. It can execute both
process contra I . and non-process contro I programs concurrent Iy . Batch
processing and time sharing at the computer console or from remote
terminals is provided by the system. The system software and most of
our program library reside on two disk packs of 512,000 words ~ach.
A core storage of 16384 16-bit words was found to be equivalent to
some 50-60,000 digits.
We have also determined that for our work the
1800 is 2-1/2 to 3 times faster than the equipment it replaced, or
about 15 times faster than the original tape system.
COMPUTER TRAINING COURSE
Frequent courses are given to train engineers so that they themselves
can program and operate the computer. These 50 hour courses are
divided into two-hour sessions, so as to minimize interruption of the
eng i neer 's regu·1 ar work. The first part of the course presents
1800 FORTRAN.
Because FORTRAN is machine independent, the engineer can be problemoriented, i.e., he can direct his attention toward the method of
solving the problem itself. The FORTRAN language is quickly learned
and a program can be written with relative ease. Examples of FORTRAN
statements from actual programs are used in these sessions to i Ilustrate the flexibi I ity and practical ity of the FORTRAN language. In
the interest of saving training time and avoiding confusion, the
design and internal workings of the computer are pointedly avoided.
o
Page 3 of 6
431
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _....'.....".....
',,'' ' '""'' ' ' ,.'_.'•..'""""...""'","","""'"."."""',...,'_".."""'
.... """"',""==",..''''"'''w=n''''''''
....
'j"''=''i'=""
.. '.""-""",,..........""........,.....
The second phase of the course concerns the three major areas of difficulty which tend to discourage our engineers from writing computer
programs:
,,'.,=
.....'""""._"f..
~~_
o
I). Methods for representing tabular or graphical data in a
computer program,
2)
Techniques for solving trial and error problems, and
3)
Means for referencing properties of refrigerants ann
other fluids.
Curve fitting and surface fitting techniques are presented as a likely
method for representing graphs and tabular data which form smooth
curves. The applications of table look-up procedures are reviewed.
Norma I Iy th i s approach is not recommended as the equat ions resu It ing
from curve, or surface fitting are easier to program, require less
storage and generally provide faster, more accurate results.
Techniques for solving trial and error problems are treated in detai I.
The advantages and disadvantages of various methods are explored in
detai I with actual programming examples.
Sub-programs for the properties of commonly used refrigerants have been
written from equations developed by a major manufacturer of refrigerants.
These same equations generated the tabular refrigerant data frequently
used in manual calculations. These properties can be referenced freely
in any program and are as easy to use as sine, cosine, and square root
determinations.
Examples:
= PfT (T1 )
Pressure of saturated I i qu id P1
Specific volume of saturated vapor Vl
=
VPTCP1,T1)
Enthalpy of vapor Hl - HPT(P1,Tl)
Enthalpy of vaporization HLAT1
Where
T1
=
Temperature in of
P1
=
Pressure in psia
=
HFGT(T1)
The refrigerant properties include acoustic velocity,heat capacity at
constant pressure and at constant volume; enthalpy and entropy of vaporization; specific volume, pressure, temperature, enthalpy, and entropy
of the saturated liquid, and of the superheated and saturated vapor;
liquid and vapor viscosity; and X and Y compressibi lity functiQns.
Other properties wi I I ba addod to this set as they are needed. SubPage 4 of 6
43112.
c
programs for the properties of water, air, carbon dioxide, methane,
ammonia, lithium bromide, chlorine, and other such fluids are also
avai lable.
During the third and final phase of the course the engineer programs
several typical applications in class. This training is to prepare
him for the ultimate objective of the course - to write a useful program himself and get actual results on the computer. Experience has
shown that only by actually writing and running a computer program
The
does the engineer acquire good recal I of the course content.
course throughout avoids the commonly encountered textbook approach.
Instead, material, examples and illustrations gathered from everyday
experience are used.
The engineer continues to receive training on an individual ba~is each
time he uses the computer faci I ity. He is given an Engineering Computer Guidebook, an in-house publ ication which provides detai led programming and operating information specifically designed to promote a
greater degree of self-sufficiency and improve the overal I productivity
of the engineer using the 1800 computer.
To date, 94 engineers have completed the FORTRAN course and most of
these are now making effective use of the computer. Our own Program
Library has grown to 175 programs covering a broad range of applications. The major areas where the computer is uti lized are:
I.
Design
2.
Data reduction/analysis
3.
Performance ratings
4.
Physical science
5.
Mathematics, statistics and rei iabi I ity
6.
Information processing and retrieval
7.
Management science
8.
Numerical paper tape control of machine tools
OPEN SHOP OPERATION
Before the initial computer instal la~ion, it was recognized that it is
easier to train engineers to write their own programs than to fami I iarize
a machine language oriented programmer with the broad aspects and fine
variations of engineering design and development.
o
For this reason, ,the computer has always been used on a strictly "open
shop" basis, whereby the engineers operate the computer and do most of
their own programming. Direct 24 hour/day access to the computer is
provided and they are encouraged to use it as freely as they do a sl ide
rule. Computer oriented help is avai lable, however, for consultation
and to assist in the more involved problems.
Page 5 of 6
A magnetic type schedule board is used to reserve computer time. Permanent reservations are provided for certain test activities which go on
around the clock. A I I other reservat i.ons are made as needed. Con f Ii cts
in scheduling or other such difficulties are resolved by computer personne I.
The keypunching of new programs and input data cards is done by a keypunch operator. The engineers, however, are trained to punch their own
cards so that they are able to use the computer after hours without
assistance.
The "open shop" operation minimizes the" number of steps in the communications chain. The engineer does not have to explain his problem to a
programmer who, no matter how intelligent, resourceful, and eager to
help, rarely has as much knowledge and experience as an engineer in his
own bai liwick. The problem of retrieving and relating al I pertinent
information and ensuring that nothing is lost in the translation is
eliminated when the engineer does his own programming. Whi Ie programs
written by engineers may not always be as efficient as those created
by programming specialists, the qualitative benefits of "open shop"
programmi ng far outwe i gh the disadvantages of any s light I y 51 ower runn i ng
times.
Like any commercial product, a program is subject to occasional or frequent revision a~dictated by technological innovations and advances.
The enginee.rwho does not write his own program frequently experiences
difficulty and delays in keeping the program up to date. Furthermore,
as the engineer operates the computer, he acquires valuable feedback
regarding the effectiveness of his own program and can detect unsatisfactory program performance, implement necessary improvements, and
prevent needless delays caused by incorrect input data.
Moreover, hands-on experience contributes greatly toward bui Iding up the
confidence of the engineer in uti lizing the computer faci I ities.
It
removes the mystery, awe, and fear often associated with computers. The
engineer who uses the computer himself no longer feels that it is a
gigantic brain with which he is i~ fierce competition.
This attitude, although diminishing with each passing year, presented a
real problem with early instal lations and is sti I I of consequence. Under
"open shop" conditions, the engineer acquires a genuine appreciation for
the computer and gradually realizes that, even though inanimate, it is a
valuable partner in a closely knit engineering team.
The "open shop" phi losophy coupled with "in-house" training has proven a
most effective approach to computer usage in design and development work.
It is undoubtedly the basic reason that this installation is achieving
such outstanding results.
G.E.G.
7/19/67
Page 6 of 6
o
1800 FORTRAN LANGUAGE
REFERENCE iv1ANUAL
MC CRACKEN
SESSION #
Or------1
INTRODUCTION
TO FORTRAN
CHAP. 1 , PAR. 1. 1
2
CONSTANTS,
VARIABLES &
EXPRE S S I 0 f\l S
( I f\! CLU0 I I\! G SUBSCRIPTED VARIABLES)
1.5,
CHAP. 1 , PAR. 1.3
PROBe 1
7
CHAP. 5, PAR. 5.1, 5.2,
S.5
PP. 1 - 5 , 18
( 0 I MEN S I 0 f\.1 )
ARI THr·1 ETIC
CHAP. 2, PROBe 1 - .3
CHAP. 5, PROBe 1 , 2, 3,
8, 9, 10
P. 6 (COL. 1 )
PP. 22 - 23 (FORTRAN
.sUPPLIED SUBPROGRAHS)
...
-.~
& 1.2
-
-
- - ---_..-
STATE~~ENTS
QUIZ #1 ON
SESSIONS 1
(1/2 HOUR)
-
19
-
...
3
-
3
- - - ..._--_.
4
CONTROL
STAT EI'·, Ef\l TS
CHAP. 3, PAR. 3.4 & 3.5
CHAP. 4, PROBe 1 - 13
P. 6 (COL. 2 ) - P. 7
(DO STAT Ef"1 Ef\! T )
P. 9 (PAUSE, STOP, E~·J D)
5
DO STAT E;·1 ENTS
CHAP. 5, PAR. 5.3,
PROBe 5, 6, 7, 11
CHAP. 6, PROBe 1 - 9
P. 7 (DO STAT E,. , Ef\lT )
P. 9 ( CO[\lT I j\.JlJ E )
-
P. 7 (DO ST1"4 TEi,'l Ei\! T )
P. 9 ( CO i'lT I j\J UE )
-
--'- - - - - - - -
<'
I DO STAT E~.'1 Ej\j TS
I
CHt\P. 6, PROBe 10, 13
-
17
----r-.
7
REVIEH & QUIZ #2
Oi'l SESS IOi'IS 1 - 6
( 1 HOUR)
8
DIS CUS S I,0 r,J OF
QUIZ #2
I f\1 PUT I (J UTPUT
STATE:'~
MC
1800 FURTR/.,·! L/I;~GtJ:\(;f.
REFERF.JJCF.-: i·~r\,\'IJ(\L
CRACKE~!
---------------------0
il:-3 ON
SESSIONS 8
~UIZ
14
-
12
INPUT/OUTPUT,FOR~AT,
S TAT EiiI Ef\.1 T F Uf\j CT ION S ,
SUB PROGRAf/,S
(1-1/2 HOURS)
----- _._15
DISCUSSIOi\J OF
11~ 3
CASE STUDY 1
CASE STUDY 8
FLO\--! 01 AGRAi·j
QUIZ
CHAP. 9,
tHAP. 9,
----- ----
PP. 62
PP. 73
-
-
-----+-------.-------...
-
16
j·1 ETHODS
17
18
OF
IT ERAT I
63
77
-.~.-
.
O~I
_._..._--_ .. ...._-19 GE~·.)ERAL 01 S CUS S IOI\J
-
OF PROJECT
o EVE LOP r··l Ef\1 T TO DATE
20
--y-_. . _ -
COi:;PUTER o Ei'-'iO GEf\!ER/\ L S E S S IOi\1
> - - - _ .•- .._ - - - - - - - - _ . _ - - _ . _ .
CASE STUDY 2
HOTOR E F F I C I E!\\ C Y
CAL CU L A T I Oi\1 S
22
cnrAPUTER DEHO ':i I TH CASE STUDY 2
23
CHAP.
CASE STUDY 4
TRAi\lSFER
HEAT
PROB. i
_
......
----r-_·-·---_-.---------.. -.-.
---.-----.-~.----
CHAP. 9,
21
P.
64
~
c
---.------
~--
------
Cni:: PUT E: t<. DEi/tO ~,I I TH C/\SE: STUDY 4
- - - ----_.
f
9,
PP.
66
-
68
..•. - ..
---.~.-
24
25
--
TEST & EXECUTE
PROJECT P RCJ GRA!": S
---- ---_.
v:"
o
S F S S J n ~,I 1\1\ I fV1 R F R
SPFAKFRS
c
C"·:
,.'}
F •
2•3
o
Subroutine DATAR
Summary:
I am going to speak to you today about the subroutine
DATAR and some of the history behind its conception.
DATAR is a format-free input subroutine designed to read
numerical input data with or without a decimal point. A
blank space is used as the delimiter (separator) and as many
pieces of data (or as few) may be punched on a data card as
desired. The call statement is CALL DATAR(K,X) which initiates
the reading of a data card and the beginning of the search
for K pieces of data which will be returned as floating point
(real) variables in the vector X.
Background:
We at General Foods have enjoyed the usage of "format-:free"
input capability as long ago as 1961 when the first Fortran
compiler was made available for the IBM 1620. The only problem with the input capability of that compiler was that each
time the READ statement was encountered, a new data card was
not read. That is, the reading of new data carried on in the
input area of the compiler where the reading of old data left
off. This caused problems when one had more data on a data
card than one wanted to read. This problem was so l ve d" when
the next compiler from IBM and all subsequent ones for the
1620 (and the present compiler for the 1130) incorporated the
format concept of input data that we are all familiar with.
Other compilers for the 1620 (AFIT and PDQ-written by users)
attempted to give format ... free input capability, but each had
its limitations.
II
The few programs having the input capability we wanted
were written in SPS (the assembler language for the 1620)
and utilized a general purpose format-free input routine
also written in SPS.
There were many other capabilities lacking from the
earlier compilers for small scientific computers such as
the 1620 which drove others to write a group of input/output
subroutines (also written in SPS) for use with commercial
applications utilizing programs \vritten in Fortran. These
subroutines were called FORCOM, an acronym for Fortran
Commercial.
-2-
At the Denver COMMON meeting in July, 1966 Jim Kokie of IBM
was explaining the FORCOM subroutines written for the 1130
(subsequently changed from a Type III program to a Type II
one entitled 111130 Commercial Subroutine Package
He made
what appeared to be an astounding statement when he said,
liThe interesting aspect of seven of these eight subroutines
is that they are written in Fortran! Astounding, indeed, for
he had just said in effect that with 1130 Fortran one now
has assembler language maneuverability at a higher programming level.
ll
).
II
Method:
The method employed in DATAR and the other routines I
have mentioned is a relative simple one.
In 1130 Fortran
(and also the 360 Fortrans) one has the capability of reading
information from a data card in Al format from 1 to 80 columns.
When. the variable used for reading is an integer variable,
the resultant value after the read (under AI) is a numerical
one which can be used for comparing and other purposes with
no problems (Data read under an A2 format results in similar
numerical-type values). Table 1 attached shows the numerical
equivalent values for all valid 1130 characters which have
been read under an Al format.
(This is p. 150 from the 1130
Commercial Subroutine Package, Version 2, Program Reference
Manual, H20-0241-2).
This means, of course, that one can easily identify any
valid 1130 character which is to be read in on cards by storing
the appropriate decimal equivalents in your program and comparing
these stored values with the ones read-in under the Al format.
If we wanted to recognize a slash (/) mark, for instance, one
might put the statement K=24896 in his program, read the appropriate variable in Al format, and compare the two.
One can
do many nice programming tricks this way.
It is worthy to note in passing the relationship among
the decimal equivalents of the ten digits. Each is 256 larger
than the preceding one which allows for one to translate the
decimal equivalent (K) of a digit read in Al format into its
digit equivalent (I) with a single Fortran statement:
o
I=(K+4032)/256
437
-3-
The letters can be related to the numbers 1-26 in a similar
way, but one must program around the two discontinuities noted
between the letters I-J and R-S.
Remarks:
DATAR was possible only because of the capability of reading
characters on a data card in an Al format.
The attached abstract,
flow chart, and Fortran listing of the subroutine clearly shows
what is taking place. However, a few remarks are in order concerning
the philosophy of this format-free input subroutine.
1.
Every time the call statement CALL DATAR(K,X) is encountered in a program, a new data card will be read
and scanning started in column 1. As many data cards
as necessary will be read until the K pieces of data
(separated by at least one blank column) have been
encountered. All numbers with or without decimal
points will be converted to floating point and returned in the vector X.
2.
Only numeric data can be read by DATAR. Any character other than a decimal point (.), minus sign (-),
plus sign (+), or the digits (0-9) will cause an
error message to print on the printer.
If an error
is encountered, the routine will halt on a PAUSE and
will return to the first statement in the subroutine
after PROGRAM START is pushed.
(Thus, one can fix
the error and continue at the point of the calling
statement) .
Note: Both an 026 plus sign (12 punch) and an 029 plus
sign (12-6-8) can be recognized.
3.
A decimal point alone, a minus or plus sign alone,
or a minus or plus sign with a decimal point alone
will result in the variable being set to zero.
4.
The subroutine at present scans columns 1-72 for data.
If any other limit besides 72 is desired, a one statement change is all that is necessary.
o
-4-
o
)ther Applications:
I have already mentioned the Commercial Subroutine Package
)ut it is worth mentioning again, particularly Version 2 which
:ontains some twenty-three subroutines most of which are
rritten in Fortran.
Clearly, one could use this method very effectively when
ne wants to generate alphabetic information from within one1s
rogram. Of course, the "H" specification would also work
nless the information is to be printed in variable places
n the page (like the word "TOTAr,.:: as a heading of a column
f figures whose placement on the page is variable depending
pon problem size).
I have used this idea many times in many programs, some of
hich are shown below:
C:
1.
In a "scatterplot ll program each X-Y value is plotted
as an "Alf.
If more than one data point falls within
the same grid, it would show as a IIBII etc.
2.
In another program I convert the digits 1-6 to the
letters A-F which is quite easily done (could be
done, of course, for all the letters).
3.
The following table shows five separate row labels,
anyone of which could be encountered in a program.
The index i is used to generate the proper row label.
Condition
--L
I
2
3
4
5
4.
A
B
C
In another program (familiar to some of you) the row
labels were generated from the i index according to
the following correspondence.
i
1
2
0
A
A
B
A
AB
B C
C D
D E
3
4
5
6
7
Row Label
A
B
A
B
C
A
B
A
C
C
B
C
43(J
-5-
5.
In the same program as (4) other row labels were
generated in this way:
i
1
2
3
4
5
6
7
8
Row
1
1
1
1
2
2
2
2
o
Label
1
1
1
2
2
1
2
2
1
1
1
2
2
1
2
2
As before, the index i generated the numbers in Al
format so that blanks could be printed if the problem size were smaller than three.
6.
In a one column distribution program one could separately identify all digits, letters, and special characters.
In closing, I shall only say that the use of the Fortran
language in this way is only limited by one's imagination.
DSG/mh
CI
0
C
EBCDIC CHARACTERS AND DECIl\1:AL EQUIVALENTS
A
-16064
S
-7616
blank
16448
B
-15808
T
-7360
• (period)
19264
C
-15552
U
-7104
< (less than)
19520
D
-15296
V
-6848
(
19776
E
-15040
W
-6592
+
20032
F
-14784
X
-6336
&
20544
G
-14528
Y
-6080
$
23360
H
-14272
Z
-5824
*
23616
I
-14016
0
-4032
J
-11968
1
-3776
- (minus)
24640
K
-11712
2
-3520
/
24896
L
-11456
3
-3264
M
-11200
4
-3008
%
27712
N
-10944
5
-2752
If
31552
0
-10688
6
-2496
@
31808
P
-10432
7
-2240
' (apostrophe)
32064
Q
-10176
8
-1984
=
32320
R
-9920
9
-1728
23872
27456
o
-150-
441
Subroutine DATAR
Purpose:
To read data in "free" format from cards, each piece
separated from each other by at least one blank column (between cols 1-72)
o
Usage:
Call DATAR(KTOT,A)
Description of Parameters:
KTOT - number of pieces of data to be read
A - the output vector containing the KTOT pieces
of data
input
output
Remarks:
An error condition will result if any character other than
the digits 0-9, a plus sign (+), a minus sign (-), or the decimal
point (.) is encountered.
The "end of card" specification is set at column 72.
If another specification is wanted (say, column 80 for a complete
card scan); change the first card of the subroutine (DATAR 13)
accordingly (i.e., KK = 80).
C
I
~··
.'
Subroutines Reguired:
None
Input - Output Devices Used:
1442 Card Reader, 1132 Printer
Method:
Each data card is read with the format 80 A 1. Each
character is scanned and the proper numbers formed arithmetically.
o
442
I
I-
J,;
't
~
rt~
~
,..
~
u..
"
~
"-
It.
~
\L.
~
&
""'
-.D
~
~.
..
:)
R'
.0
-;:r..,
k
(5)
443
/ /
~!-
/ /
,~
DUD
o
"';;::-U::oTC"
=(jn
'! \! ~
I".f
r. ~ i)
I i\i T t G '- r, .:..
C; U p q 0 UTI f\i ~ i) A T A C. ( K TOT
.
~ f\ )
,""
'-
,..
F ~ E E F Q R ~.f. t\ TIN PUT
'-
C
C
,..
\
R (1 UTI r--! ~
KT0T=Nlfv!8ER OF
,lI
! JT P :.1 T V '- C T r, p
=('
I C (\1 t. ( 1 )
r:- ,\, S I 0 (1.1 (I.' ( 8 I)
f) T r-,,'
1) T'vI
c
c
ELE!\/;tNTS TO BE READ AND PLACED I N
()
F V ,A PI'. ? L ;:- ,S
1\1 5
C"
)
SET *END OF CARD* AT COLUMN 72
-
'-
Ki(:::72
!( :::
1
,\
] 03
R F.: A f)
(
=
2 ~ 4 ()") ( N ( I ) , T 1 ,K K )
40 FOR'I.AT(SOA1)
='
'. "3 7
'!'
c::. T (: r-,! :::
1•
NDEC=O
1 C) 5 I;: ( j\!
(
I ) - 1 6 4 UK)
1 3 2 , 1 I'"' 1 , 1 '3 2
,..'
101 IF( I-KK) 104,ln3,1n1
104 I=I+1
GO Tf) 10.'5
CH ;:C !(
132
138
I
=( ~!
I F ( j\:
(
(
c
133
D L USC,
I ) -? n n '3 /)
I ) - 2 0 :, 4 4)
( H ECK
10?
1 !) 6
r ::; P
~ (' Q
:A
I~(N(I)-?4~40)
c:, T r-. \1
=- 1. •
IF( I-KK)
1("\7
T=I+1.
1{)8 Nc:,TDT=!
I G~\I
(1 2 - 6 - 8 C.A R D P U 1\] C H - 0 2 9 )
1 3 11 , 1 '3 '3 , 1 ~ H
1 0 2 , 1 3 '3 , 1 (\ 2
1 N! J C,
c:, I r, ~,!
( - )
10?,lnA,1nQ
-lU\"l~
DATAR
DATAR
DA T A,R
DAT AI~
DATAR
DATl\9
DATlI.9
DATAR
OATAP
DATAR
DA T A,R
OA Tl\R
DATAR
nATt\p
nATA?
riA,T t-,p
n,ll,TAQ
DATAR
DA, T AR
DAL,\R
nA T [:.,r;:;
rJATAR
DATA?
l"'\fI T l,?
nATAP
DATAR
DAT A.R
DATAR
DA T ,AR
DA T p,R
f)A,TliQ
f)ATAR
1
2
3
4
5
6
7
8
Q
In
II
12
13
]4
1s
lA
17
18
19
20
7.1
22
23
/4
?C)
26
27
28
2G
C'
.,
30
?-1
32
f'lt\Tt.r:;>
~~
DATAR
DAT,AR
DATA.R
3 L.
35
36
":37
DATAR 38
f)AT.M~
i)ATAq
j) A T llR
f")A T L\R
D,ATAR
DATAR
DATA,R
OATAR
') A'j AR
r)ATAD
-:1,,",,
4n
41
42
43
44
4~
."J
46
47
fJATAR 4A
[)ATtq 4')
1()7,1?8,128
f)ATAR
')1"')
DATA~
51
S2
nAT t.R
DA T AI~
nATAp
nATt.;:-
5~
nllT~r")
::;r.,'
5"3
~t...
C'i"
"I
"il
CHECK FOR
r
c'"'
c
c
100 1FeN(1)-19264)
109 NOEC=1
112.1()9,11~
CHECK FOR fNG OF
110 tFeJ-KI()
111 1=1+1
DATAR 57
DA T~.R 58
DATAR 59
POINTe.)
D~CIMAL
C~RD
(COL *KK*)
111,11;',11'i
r
CHECK FOR
(
c
112
PLANK
1~(NeI)-16448)
130,113,130
r
'-
ACC~PT
r\ .
130
~TGIT~
rF(N(I)+40~2)
0-9,
QCJ~CT
ALL
OTH~R
(HARACT~pc
131,122,115
135 IF(N(I)+1728) 122,122,131
131 W~IT~(3,14n) N(I ),1
140 FORMAT(1Hl,?lhNON-NU~SRI(
CHARACTER~1X,Al,lX,9riIN
PAUc,=:
G0 TO 1
c
c
~IO
C
r
'-
114
"J!)!= ('
1,
~I
t;
r::
I~:
()~TAR
=I
= T-1
f),A.T,AP
DAT t.R
DATA?
DATAR
DATAR
DATAR
DATAQ
Go') T0 117
r
~
y!:"S
(:
r
'-
116
~FIN=I-2
r
'.
(
r
IF (.),
(-),
(+),
(-.),
OR (+.)
SFT A(K)=n.
.--
128t..(!()=O.
(-0.
1
c
120
T('.
1 R L =!\: F I 1\' -
L =0
.1\1 1'"\ t:" ('
+,
I"1i
~ Lj
6·1
S2
6~
64
65
66
67
6~
69
7r:
71
77
73
74
75
76
77
78
70
8~
81
d"''''l
L.
83
84
8e:.
- -'
86
87
8~
8)
on
91
92
':;3
f)AT~R
94
0AT~Q
q~
OATAr:? q6
DATA9 °7
DATAR 93
DA1AR 9;
DATAR1:)O
DATARl!")1
!lATA?l"?
rAT~,~1f)4
I t:" T
OATAR10S
DATAR1!J6
DATtJ)107
DA.T AR 108
DATAQ109
LL=IARS(L)
S~I~T=lO.~!-~:LL
1 ~"~
60
f)ATAo1r:\~
L GREATt:"R THAN 0
cYIFT Lt:"FT
L L~SS THAN n
cHtFT RIGHT
TO
COLUMN,11)
DA.Tf·R
DATAR
DAT AR.
IlATf!.D
DATAP.
DATAR
OAT A.R
I)ATAR
DATAP
DATAR
DATAR
DATAR
f)ATi\R
DATAR
DA-T AR
DATAR
DATAR
DATAR
DATAR
r.JATt.R
r>A,TAR
f)l\Tt.r:?
OAT t,R
DATAR
DAT !:..R
DATAR
;) 0
=r
t:"1}
DATA~11(')
T /\\ (., ? r:: p (' ( !:" c)
r:: 0
('\ ~.'
/\11, J ~/~
a r:: 0
!)ATflJ?111
~flTf\q11?
=
o
K:( I( j\! S r R T
J0 162 J J J
=I( K ~~ , :\' F I /\1
IF(N(JJJ)+4n??) 161,162,161
16 2 ~! S TRT= NS Tq T+ 1
DATAR113
DA,T A.R 114
DATAR115
DATf.,P116
GnT{) 129
,..
\-
...
r
-
r
161 FFF=O.
DO 160 JJJ=NST~T,,\lFrN
160 FFF=FFF*10.+FLOAT( (N(JJJ)+4032)/256)
IF(L) 150,]52,1'51
150 FFF=FFF*SHIFT
(-f")
15?
Tf')
151
C!=F=i:;!=!=/~~TFT
152
,~'( I() =FFC'Y.'~ T (;1\1
CHECK FOR LAST VARIABLE
\...
r
'-
r:
129 IF(K-KTOT)
120 i(=1(+1
c
IFCT-KK)
c
120 119,]}0
CHECK F0R EN0 OF CAqD (COL
r
*~K*)
1?1,1()"l03
121 1=1+1
GO TO 137
HAS DFCIMAL POINT BCFN
~NCOUNTF~!=D
c
YES - SHIFT
C
DECI~AL
C
c
124 N(I-l)=N(I)
GO TO 110
c
DIGITS ONF SPACE TO LEFT
DATAR129
DATAR130
DATAR131
DATA,R132
DA.TAR133
DATA,Rl'34
DATAR135
Dl\TA,Rl':\6
DATAR137
DATAR138
DATAR139
DATAR140
DA,T AR 141
I)ATAR142
DATAR,14-:l.
DA TA,R 144
DATAR145
DA.'T A~ 146
DATAR147
DATAR148
DATA P 149
f')ATAP1'1 rJ
r'lt,Tf\01t:.1
12?
TCCT-KIt)
'?~,1?f..",~
125 I=I+1
GO TO 100
126 NF I ~1= I
~lr'I~C=
I+1
GO Tn 117
110
OCTIIO".1
r'I,ATl\Plr;?
r)ATA.Q153
DATAR154
DATAR155
DATARl5A
DATARl'37
DA,T,A o l::,R
nAT~o,t:.1")
I I
r'lIF'
* STOP':
I,-!S
IJA
o
flAT~R1?~
C
r
DAT/\Rl17
DATt-RI1S
I)ATl\.f?llq
rJATf)01?0
DATAR121
DATAR122
DATA,R123
DATAR124
DATAR125
r)AT,A,t:?1?6
f)ATAP127
I)~T,AP
c
S F SSt n~,I
1\1' J "A R F R
F•
2•4
SPFb,KFP S
n A\! Tn,). MAR T I 1\1, C HAT T A1\1 n GAS TAT F T F CH. I 1\1 ST., n 1\1
F n I. I rAT
I n (\1 ALe n MPI) T FRS Y S T F Iv1
o
447
II, I\!
F XP Af\1 f) F n
The
o
1041
AN EXPANDED EDUCATIONAL COMPUTER SYSTEM
A paper
Submitted
by
c
David J. Martin
Associate Professor
Data Processing
Chattanooga State Technical Institute
COMMON Meeting
Cincinnati, Ohio
September 7, 1967
o
o
In the fall of 1965, the State of Tennessee entered upon a new adventure in the
field of higher education, when Chattanooga State Technical Institute opened its
doors for the first time o The purpose of this school was to provide, during a
two-year period of study, a junior college education geared, not so much
to the development of theories, as to the practical application of these theories.
Qne of the fields of study included in its program is a data processing curriculum which is divided into three distinct areas o The first of these is, of
course, the various programming courses o In the 14 courses in this category,
the student is introduced to variety of subjects.
From an introduction
beginning with unit-record equipment and its applications, he progresses into
machine-level language programming languageso
From higher level languages such
as COBOL, FORTRAN, AND PL/l, he then goes into Systems Design and Analysiso
o
In keeping with our philosophy that the data processor needs to have a wellrounded educational backgroung, the student is also required to take a group
of courses called related studieso
Among these courses are English, Public
Speaking, Mathematics, and Social Scienceo
Additionally, the student chooses
to complete his course of study with courses either in Accounting or in Mathematics and Science.
Our students' courses in computer programming initially included FORTRAN and
1620 SPS during their first yearo
In order to expand their programming back-
ground to include work with other systems, we planned to provide training and
experience on the 1401 during their second year.
Since we believe it is necessary
for a student to compile and execute programs before he can begin really to
o
understand a computer-programming system, we faced a very definite problem.
could we teach the 1401 when our school only has a 1620 computer?
The ideal
How
Ii
I"
\
solution was to rent a 1401 computer; therefore, we called up our "fr j (md J 'j
o
IBM. representative" and asked the cost for renting an appropriately equipped
system o When we heard the price, we realized that our ideal solution would
have to be compromised somewhato
Then one of our faculty members suggested
that some pub1ic-spiritec company might be happy to allow us to use their
1401 system free of charge or at a reasonable costo
But when we realized the
amount of machine time and/or cost involved, this idea also faded into the
background
0
And so, you guessed it, I was appointed the task of making our
1620 meet this new challenge.
Up until this point, I had always believed that the 1620 users-group library
was a worthwhile undertaking.
Now it had a chance to prove itse1fo
At the
first of September a year ago I ordered a group of programs known as the 141
It included two 1620 programs--a simulator and
Educational Computer Systemo
an SPS assemb1ero
The simulator portion of the 141 System provided for the simulated execution
of 14 1401-type instructions
0
Among these were the basic input-output
instructions--read, write, and punch--the arithmetic instructions of addition and
subtraction, the data manipulation instructions of move and load, the various
forms of the branch instruction and some few miscellaneous instructions o
The
indicators that can be tested are those that result from a compare operation;
high, low, equal, and unequa10
The only provisions made for carriage control
on the printer was the testing for channel 12 and skipping to channel 1, this
function being automatically performed by the simulator, rather than being contro11ab1e by the program.
Also provided were subroutines for multiplication,
division, zero suppression, and editing of fields in the form of a dollar and
cents amounto
-~--.----.---~~------
We used this system through the entire fall quarter and were very pleasantly
surprised with the results obtainedo
However, it had several dificiencies that
we wanted to overcome, so in December of last year I proposed to begin work on
a project that led to the development of the 1041 simulatoro
One of the first problems encountered in the 141 System concerned the amount of
core storage available.
Of the thousand positions available the first three
hundred were reserved for use as input-output areaso
In addition, the four
subroutines mentioned above required approximately one hundred positions
apiece; therefore the average student problem was limited to less than five
hundred positions, which meant that the complexity and size of the various
problems ahd to be limited.
As a result, the first change was to increase the
amount of core to two thousand positionso
This change provided greater variety
in the programs that could be solved by the studentso
It also afforded the
student experience in using a three position address to denote an address
greater than 9990
A second need was for the student to be exposed to the power of the 14010
This
need necessitated making available for his use the complete instruction set for
a 1401 card system rather than simply the 14 instructions then available.
In-
cluded in this group of additional instructions were some to provide greater
ease in handling problems which were commercial in natureo
For example the move
characters and edit instruction, probably the most powerful instruction on a
1401, was implemented in full powero
which he wanted information displayedo
Now the user could specify the format in
In the edit-word he could denote the
location of the decimal point, any commas, the desire to have leading zeros
o
suppressed or asterisks inserted in their place (a feature known as check
protection), the insertion of a dollar sign, and the use of a CR symbol to denote
\~
I~
I'
a negative valueo
The zero-suppression routine, likewise, was replaced by
a single instructiono
The other instructions added were zero and add, zero
o
and subtract, the multiple-output instructions, and store A and B
address registers to facilitate subroutine linkages
0
In order to allow printer operations, a control-carriage instruction was
includedo
Finally, for compatability with the 1401, the select stacker
instruction was included.
Simulation of this instruction is of course
physically impossible with a 1620, therefore it is treated as a no-op instruction unless there is an address included with it, in which case the
appropriate branch is taken.
We expanded the usefulness of the Branch on Indicator by allowing it to check
for the Last Card, Channel 9, and Channel 12 situationso
c
In order to teach the 1401 indirect addressing scheme, it was necessary to
implement the feature known as index registers.
By storing appropriate
values in the index registers, effective addresses could be generated without
modification of the actual addresses in coreo
and more efficient use of storageo
The results are faster execution
An additional feature of the 1401 allows
for the CPU to utilize the contents of the A or B address registers, rather than
obtaining new addresses from coreo
This feature, like the index registers,
results in better use of core and faster program execution.
As a test for this new simulator, one of the programs chosen was the Autocoder
Pre-List programo
the systemo
It worked, and so did some 500 student problems run against
o
'1
The final goa1--that of increased Through-put--was achieved in many ways, some
o
of which we have already pointed outo
In addition to these, I completely
rewrote twelve of the fourteen instruction simulation routineso
For instance,
new algorithms had to be developed for addition and subtraction since the highorder position of the fields did not necessarily have to be numerico
We also
used,wherever possible, the indirect addressing feature of the 16200
The net
result is considerably greater Through-put over the 141 simu1atoro
This
was especially gratifying when you remember that, in addition to simulating the
1401 instruction, we were also maintaining the current status of the A and B
registerso
In addition to the ability to execute programs, both simulators have four
routines providing the following features:
1)
A formatted dump of the 1401 Storage area
2)
The ability to alter the contents of the 1401 storage area
3)
The ability to designate at execution time the address from which
execution is to commence
4)
The ability to initialize the 1401 storage area to b1ankso
The
formatted core dump includes displaying the contents of the various
registers and the op-code that was last being executedo
All of these functions are under the control of the operator and are initiated
through the use of the 1620 console typewriter.
As all of you are very much aware, the development of a computer is composed of
two phaseso
The first of these, which we have
discu~sed
up to this point, is the
development of the hardware, but, in order to utilize effectively the elements
o
of the hardware, it is necessary to develop appropriate software
look at what we have provided in the way of softwareo
0
Now let's
IIi"
I!
I
As you remember, there was provided in the original 141 system an SPS Assemblero
It of course provided mnemonic op-codes corresponding to the 14 instructions
of the 141 simulator plus the pseudo-op-codes necessary in a symbolic
language 0 The highest address it could assemble was 999.
for index register specification.
It made no provision
One of the interesting features of it was
that it allowed one-pass assembly only for programs containing fewer than 100
source cards o Finally the label table had a capacity of ninety entries.
Next we see a 141 Autocoder assembler that was developed by two of my secondyear studentso
Though written independently of the 141 SPS Assembler, their
program provided many similar featureso
Several minor and two major modifications allowed the 141 assembler to be
expanded into the 1041 SPS Assemblero
One major change allowed the address
generation algorithm to specify four-position addresses in three positions
and to include any index register associated with that addresso
c
A second
major change made better use of the available 1620 core storage in order to
process much larger programs in
900 on a 40K machine.
on~
pass--250 to 300 in 20K or up to 700 to
This was accomplished by storing only the amount of
information necessary to produce the source card image on outputo
of the 141 assembler stored columns one through 550
The writers
I had discovered after
analyzing many source decks that storing ten to fifteen columns was more than
adequate 0 At long last the 1620 user has for his use) for all practical
purposes) a 1401 computer system.
I hope many of you are asking yourselves what is necessary for you to use
this systemo
Needed is a 1620 with indirect addressing and with the additional
0
',
"j
o
instruction seto
executiono
These two features could be eliminated at the cost of slower
I will be glad to talk with any of you about other modifications
that you might desireo
We plan to submit the 1041 system shortly to the 1620
Users' Group Libraryo
Alas, as I began to breathe a sigh of relief, we recieved conformation that
our school's new 360 System is in the process of being installed today.
In
anticipation of this installation several weeks ago, the question was asked as
to when I 'p1an to have a 1401 simulator available for our 3600
I categorically
denied even the idea of such a project, but I must confess to you that the
idea of having a 4K simulator with actual stacker select capability has me
already mentally drawing the flow chartso
o
o
INSTRUCTIONS
1401
141
G'
1041
BRANCH ON INDICATOR
CHANNEL 9
X
X
12
X
X
lAST CARD
X
X
UNEQUAL
X
X
X
EQUAL
X
X
X
LOW
0
X
X
HIGH
0
X
X
BRANCH WORD MARK
OR ZONE
X
X
COMPARE
X
X
X
NO OPERATION
X
X
X
READ
X
X
X
WRITE
X
X
X
WRITE & READ
X
PUNCH
X
READ & PUNCH
X
X
WRITE & PUNCH
X
X
WRITE & READ & PUNCH
X
X
SELECT STACKER
X
X1<'
CONTROL CARRIAGE
X
X
X
X
STORAGE
X
1000
2000
CHAIN OF ADDRESSES
X
X
INDEX REGISTERS
0
X
LEGEND~
X=STANDARD
O=OPTIONAL
S=SUBROUTINE
0
0,
~I
~--.---.--
..- --- --
-
-
...
---~~-
Attachment
o
SUMMARY
OF
Number
1
FEATURES
1041
mSTRUCTIONS
C':I
ADD
X
X
X
SUBTRACT
X
X
X
MULTIPLY
O/S
S
S
DIVIDE
O/S
S
S
ZERO & ADD
X
X
ZERO & SUBTRACT
X
X
MOVE CHARACTER TO
A OR B WORD MARK
X
MOVE CHARACTER TO ..
0
MOVE AND SUPPRESS ZEROS
X
MOVE NUMERIC
X
X
MOVE ZONE
X
X
MOVE CHARACTER & EDIT
X
S
X
X
X
X
SET WORD MARK
X
X
X
ClEAR WORD MARK
X
X
X
CLEAR STORAGE
X
X
X
STORE A-ADDRESS REGISTER
0
X
STORE B-ADDRESS REGISTER
0
X
BRANCH
X
X
X
X
iii
S
X
LOAD TO A-FIELD
WORD MARK
X
0
4~7
X
Ji't~'."''''''~~I~''''iw'~'_''''''*~
o
SESSION NUMBER
F.2.5
SPEAKERS
DAVE DYE (PID)
JOHN KEITH
JIM STANSBURY, CHAIRMAN
GAYE BABER
DISCUSSION
MR. DYE DESCRIBED THE OPERATION OF PIO, POINTING OUT THE
PROCEDURE FOR HANDLING ORDERS AND TELLING HOW THE USER CAN HELP
IMPROVE THE SERVICES. JOHN KEITH DISCUSSED THE SHIPMENT ANALYSIS
QUESTIONNAIRE, PURPOSE AND USE OF THE QUESTIONNAIRE, AND A SAMPLE
COpy IS INCLUDED IN THE REFEREt\!CE MANUAL. JIM STANSBURY DISCUSSED
THE STANDARDS FOR THE 1130/1800/360 CONTRIBlJTEO PROGRAM LIBRARY.
THESE STANDARDS WILL BE ISSUED ON THE FIRST UPDATE OF THE REFERENCE
M.A"'UA L.
(;1
o
o
Cincinnati - COMMON
Sponsor:
Administrative Division, Program Library Project
Subject:
IBM Program Information Department -- Program Library Procedures
Speaker:
D. R. Dye
Company Represented:
IBM
Speaker's Address:
IBM Corporation
Program Information Department
40 Saw Mill River Road
Hawthorne, New York 10532
914 -- 592-5790
Number of Pages:
4
o
A presentation concerning the IBM Program Information Department
(PID) was made at Cincinnati COMMON. Slides were used in conjunction
with the verbal presentation Which was educational in nature. An open
discussion period was included.
o
Although the presentation was formatted somewhat differently than the
following narrative, it answered these same questions about PID that
you (as a user) might ask.
WHAT IS THE IBM PROGRAM INFORMATION DEPARTMENT (PID)?
PID is the central control point for availability and distribution of IBM
Programming Systems, IBM A.pplication Programs and Contributed
Programs. PID serves IBM customers in the United States from an
extensive facility in Hawthorne, New York. Customers of the IBM
World Trade Corporation are serviced from similar facilities in Paris,
Rio de Janeiro, Toronto, Tokyo and Sydney.
PID's growth from its inception in the late 1950's has been dramatic.
From an elementary exchange service for a few customer and IBM
authored programs, PID has become a carefully organized operation
which distributes hundreds of thousands of programs to users each
year.
WHAT DOES PID JV1EAN TO lVIE AS A USER?
The availability of almost 3000 programs for IBM equipment means
that users can often save considerable programming, procedural
and systems design effort that would be required if they started from
scratch. Often the economic payoff on an installation can be improved
and the effective utilization can be increased and/or realized sooner.
PID can take a load off your back by s erving as the distribution agency
for your programs through the Contributed Program Library.
PID also means that, as an IBM customer, you have a dependable,
accurate and timely source of IBM authored programs to support your
system installation. When an improvement is incorporated or a bug
is fixed, in an IBM program, users registered with PID automatically
receive the modification or a notice of availability of the release.
o
-1-
----------------------------------------j
o
HOW DO I KNOW WHAT PROGRAMS ARE AVAILABLE FROM PID?
Abstracts of all programs available from PID are published in the
"IBM Catalogs of Programs II'. Copies are available through your
local IBM office. In addition, IBM sales and systems engineering
representatives receive automatically, timely fact sheets called
Programming Announcements (P-Letters) which give information
on the availability and features of IBM Programming Systems and
IBM Application Programs. In addition, a Memo to Users is sent
to each customer on the PID file announcing the availability of a new
release.
Once you have ordered an IBM program and registered with PID as
a user, you will automatically receive direct announcements of new
releases of that program.
HOW DO I ORDEg, A PROGRAM FROM PID?
Normal program orders should be submitted using Program Order
forms available from local IBM offices or special prepunched order
cards supplied to you by PID. By a normal order we mean one that
will be processed under PID's standard in-house processing cycle of
up to ten working days. Transit time to and from PID is in addition
to the ten working days in-house.
Orders that require faster handling should be submitted through your
IBM representative after he has contacted PID by phone to make
arrangements for special handling.
It is a good practice to consult your IBM representative before submitting an order. Certain information must be provided on the form
and sometimes magnetic volumes must be submitted with the order
form. Without all needed information and magnetic volumes, Pill
cannot complete the processing of your order. The IBM representative
can help you in determining if the order submittal requirements have
been met.
HOWlS THE QUA,LITY OF Pill, DISTRIBUTIONS ASSURED?
Extensive use is made of special hardware designed and built for Pill.
Also, sampling, chain copying, programmed comparisons and assigned
personal accountability techniques are used.
o
-2-
4~1
,I
I'
]~
I
!
The result is that the quality level of program materials distributed
by PID is very high.
o
One particularly interesting piece of special purpose gear used for
this quality control is the IBM 7299 Tape Copy and Compare System.
This system automatically performs a bit for bit compare, in core,
of written output against input and checks record counts and control
labels. This assures the readability and completeness of users r
magnetic tapes onto which programs have been copied.
If you receive a magnetic tape bearing an external label stating that
the tape was "verified readable at PID" but you experience difficulty
with it as input on your own system, we suggest that you try to read
it on another drive and even another system, if possible. If a problem is still encountered, your CE should be notified so that he may
determine the precise nature of the reading problem.
WHAT CAN I DO TO USE PID MORE EFFECTIVELY?
A few simple checks when submitting an order will insure that you get
the program you need when you need it.
1.
Be sure that the order form is complete. These items are needed
by PID to control the accuracy of the order process:
o
a. IBM customer number and branch office number (your local IBM
representative will fill in this information for you if requested).
b. Type of material required.
c. Track mode and recording density.
d. Full return address including ZIP code and "Attention of".
2.
Be sure to submit a magnetic volume to PID if it is required for the
program being ordered. The program abstract in the "IBM Catalogs
of Programs If always indicates if a volume must be submitted. A.lI
magnetic volumes should carry external labels identifying the sender
and the program being ordered.
3.
Send the order form and the magnetic volume to PID as one package.
-3-
o
o
Plan ahead -- allow adequate time for shipping to PID, up to ten workdays at PID for processing and up to four days for return shipment. If
faster service is needed, have your local IBM systems engineer or sales
representative call PID.
Report packaging and distribution problems to your local IBM representative immediately if a replacement is needed. The Program Distribution
Questionnaire, included with every shipment from PID, is a convenient
means of reporting on the condition of program materials. Please use it.
-4-
o
o
SESSION NUMBER
SP E A.KER S
NONE -
F.2.7
GENERAL DISCUSSION
DISCUSSION
C.O.S. - COMPATIBILITY OPERATING SYSTEM.
SOME 20 PERSON PARTICIPATED I'" A GENERAL DISCUSSION OF C.O.S.
VARIOUS LEVELS OF C.O.S. EXPERIEf\.ICE LED TO A GOOD OISCUSSIO~'.
L E~, GTHO F THE MF E T I 1\1 G t..J AS ABO UTI H0 URI 0 ~1I Nil T E 5 •
o
o
SESSln"" "HIMRFR
F.3.1
SPEAKER S
JAMFS N. FISHER, THE RAnGER co., INC. nN STORaGE & RFTRIFVAL OF
PFRtv'1A~!FT FILFS FnR THE 1620/1311 MO".IITnR SYSTFM
o
o
Storage & Retrieval of Permanent Files
for the 1620/1311 Monitor System
o
by
James N. Fisher
The Badger Co., Inc.
Cambridge, Mass.
o!
THE BADGER COMPANY, Inc.
o
A method of perll1anently loading of inforll1ation (either in the forll1
of nUll1eric data or Hollerith-type Forll1at headings) has been found which
fills a gap inherent in the IBM 1620/1311 Monitor Systell1.
consists of three steps in which:
This ll1ethod
(1) a Fortran prograll1 is written to
read the data and tell1porarily store it in Working Storage (this is done
through the usage of the RECC}5RD statell1ent), (2) a set of three cards
designed to write a lOO-digit indentification tag preceding the data files
is then read into the cOll1puter, and finally (3) a*DLC}5AD operation is
perforll1ed which is used to relocate the information froll1 Working Storage
into a perll1anent disk storage address.
The data ll1ay be used at any till1e by copying it into Working Storage
and then calling for it whenever necessary.
steps:
This is accoll1plished in two
(l) a *DC¢PY operation used to copy the perll1anent data files into
a working storage area and (.2) a FETCH statell1ent used to call in the data
by a ll1ain prograll1.
The prograll1 is set up for a Monitor systell1 with Working Storage
beginning at 00219.
If the Working Storage area has been redefined this
value must be altered accordingly.
STORING DATA FILES ON DISK
I.
Loading Inforll1ation into Working Storage
A Fortran program is written by the programmer to read the
desired inforll1ation into the Working Storage Area.
A prograll1
o
THE BADGER COMPANY, Inc.
-2to read in an alphameric set of headings is shown as follows:
i-i-J¢B
o
5
i-fF¢RX
DEFINE DISK (nl, n3)
DIMENSI¢N YH (n 1'. n2)
D¢
2 YH
2
(N)
N=l, n 4
=
0.0
READ 1001, Nl, N2
READ 1002, (( YH (J, K), J = 1, N 1), K= 1, N2)
N=2
REC¢RD (N)
1001
(( YH (J,K), J=l, Nl), K=l, N2)
F¢RMAT (2I4)
c
1002 F¢RMAT (n l A4:)
END
Where:
n l = number of individual items to be written per data record=Nl
n2 = number of data records to be stored = N2
+1
n3
~
n2
n4
~
nl . n2
This program, since it is to be used for alphameric headings, uses
an A-Format in Statement number 1002 to read in data.
For nu-
meric information the programmer can use either E-, F-, or IFormat instead.
Note that in the third instruction after StateInent 2,
o
THE BADGER COMPANY, Inc.
I~
-3-
o
N is set equal to 2.
This is done in order to leave the first sector
of working storage available for the 100 -digit identification tag of
Section II.
Also:
The maximum nUlnber of sectors/ data record
= 2.
This is
determined by the value of f and k (See Service Manual, File
No. 1620-36, Form C26-5739-3).
II.
Writing Identification Tag
Next, the following three cards are loaded into the computer, with
the Working Storage area defined as beginning at 00219 in Card 1.
Cd. 1.
3600100005003600180005003400050007013800050007024810021900100100
Cd. 2.
9878971000100101001200008040010659998
Cd. 3.
Blank Card
Note:
Some of the inputs in Cd. 2 are not necessary.
However, it
is strongly suggested that they be used exactly as presented
in this paper.
IlL
Permanently Loading of Data
Finally, the data is loaded into a designated disk address.
This is
accomplished by means of the following set of cards:
ffJ¢B
ffDUP
*DL¢'AD
5
(COl.
.Col. 21
name (1 002.191nSnSnSnSnSl n6 n6 n6 n 6 n 6
49
DI
o
THE BADGER COMPANY, Inc.
I~
-4Where name = name assigned to the data file by the programmer (6
o
characters or less).
ns
= 00219
n6
=available
+ n7
- 1
sector address - able to contain n7 data records.
This
may be determined from a DIP* Listing of the Monitor System.
n7 = N2 if there is one sector/data record; otherwise use 2(N2 ) for
two sectors/data record.
Note:
See note for Section I.
Both nS and n6 are five-digit numbers which must be rightjustified.
UTILIZING DATA FILES
1.
Copying Data Into Working Storage
When the data files are desired, they must first be made available to
a program by copying them into Working Storage.
Copying of infor-
mation from a permanent disk address into Working Storage is accomplished in the following manne r:
Co121
f;t:DUP
*DC¢PY
name
Where name
*
(ln6 n 6n 6n 6 n 6 InSnSnSnSnSl0021 9
= same
as that in previous section.
a DIP Listing consists of a map of all programs, their corresponding
disk locations, plus all available sector locations.
obtained through the USER'S Group.
476
This may be
The IBM Program Library
THE BADGER COMPANY, Inc.
o
-5Nwnber is 1. 6. 141 ..
o
ns = n6
Note:
+ n7 - 1 and must also be a five-digit right-justified number.
It is strongly suggested that the programmer carefully
check the *DC¢FY and *DL1>AD cards.
Errors in these
cards have been known to render all disk-loaded programs
completely useless.
II.
Operations with Data File
The program reads in the data files by means of the FETCH statemente
The record number is set equal to 2, as was indicated in
Step I of the Storing Data Files on Disk Section. For a detailed explanation as to the proper usage of the FETCH and REC¢RD statements, see page 106 of the IBM 1620 Monitor II System Reference
Manual - File No. 1620-36, Form C26 -5739-3.
In order to fetch and
utilize this stored information, the program must contain a DEFINE
DISK statement, in which the value of n 1 must remain unchanged,
while all n
2
and n3 terms may now be less than or equal to those
corresponding to Step I of the first section.
A sample segment of a program used to fetch the alphanuITIeric inforn1ation of Step I, Section I and complement it with cOITIputed values
is pre sented below:
N =2
FETCH (N) «YH (J,K), J=l, NI), K=l, N2)
..../.;' /
THE BADGER COMPANY, Inc.
-6-
· ...................................... .
·...................................... .
· ...................................... .
o
PUNCH 1003, «(YH (J,K)., J;;:l, Nl), (Z¢ (N,K), N=l, 3), K=l, N2)
1003 F¢>RMAT (nl A4, FIO.O, 2F13.4)
Where:
Z¢ (N, K) = the computed values used for output by the program.
The following output was obtained from such a sample segment:
DIAMETER,
FEET
1.0000
2.0000
3.0000
TRAY RING,
INCHES
4.0000
5.0000
6.0000
DOWNCOMER WIDTH,
INCHES
7.0000
8.0000
9.0000
PATH LENGTH,
INCHES
10.0000
11. 0000
12.0000
DC-TRAY CLEARANCE,
INCHES
13.0000
14.0000
15.0000
INLET WEIR HEIGHT,
INCHES
16.0000
17.0000
18. 0000
OUTLET WEIR HEIGHT,
INCHES
19.0000
20.0000
21. 0000
HOLE DIAMETER,
INCHES
22.0000
23.0000
24.0000
HOLE PITCH, TRIANGULAR,
INCHES
25.0000
26.0000
2,"{.0000
C
At the moment, permanently-loaded information may only be used by
a disk-loaded program.
This may be improved upon in the future.
The advantage of permanently storing data files on the disk is that a
programmer now may be able to limit the number of F¢>RMAT statements appearing in his program.
Further, independent programs
may now be allowed to use the same records in places, thus further
reducing the amount of storage required for Bome programs.
Finally,
I
OJ·
;
,
THE BADGER' COMPANY, Inc.
•
5 ••
-7long chains of alphameric headings complemented by output data
o
may now l::e permanently stored as data files.
Caution, however,
should be used in experimenting with the Storage-System, for as
mentioned in Step IJ Section IIJ some errors may result in the
destruction of disk-loaded programs.
r1'
N. Fisher
o
4'-1?)
THE BADGER COMPANY, Inc.
======AtU'k::JNi4i@$;aW;:;"C=U4MM aM,
U;;4i4 \ U* ,4$.
z:
4J ,
,PI,
T
U4i¥#&#M#%M 44
• • au
SESSION NUMBER
- - - - - - - : - - ... ....
'"
-.~--
•.---'"-"
F.3.4
SPEAKERS
NO SCHEDULED SPEAKERS.
DISCUSSION
PLANS WERE MADE FOR THE 360 AGENDA AT THE SAN FRANC1SCO MEETING.
THE GOOD AND BAD RESULTS OF THE SESSIONS AT THIS MEETING WERE
EVALUATED, AND WE DECIDED WHAT PLANNED PRESENTATIONS WERE DESIRED
AT THE NEXT MEETING.
c
o
S F S SIn [\!
1\11 I M R F P
~
•A
R F Y 1\.1 L r.-)
s.
F•
SP~!\KFR~
(~ F ClP ~ F
".
n
G F 1'1 r:: R A L
T J fJl F S T I J r) Y fJ 1'.1 to L Y S T S F n R
C
!
I~
o
Ivl (l T fi R S
vI C) R!<
T ~! S T TTl .fT E
j\A F A S l )P
F rv' F ~.1 T
n [\1
C n,., 1 P l , T F= R
o
COMPUTER TIME STUDY ANALYSIS FOR WORK MEASUREMENT
George J. Reynolds
Assistant Professor of Industrial Engineering
General Motors Institute
Flint, Michigan
ABSTRACT
c
This article describes the format for the observation,
recording and calculations of work standards. In
addition, a computer program for the IBM 1620 Model II
40K is included which provides a more efficient calculation time for each time study. This in turn
reduces cost per work standard.
o
COMPUTER TIME STUDY ANALYSIS FOR WORK MEASUREMENT
George J. Reynolds
Ronald W. Cox
o
The determination of a fair day's work has progressed from haphazard methods to rather scientific methods.
One scientific approach to
this is called "Time Study." Time study can be explained as a technique
for the measurement of the time factor in the utilization of men, materials,
tools, and equipment.
In most industrial organizations, time study is one
of the common techniques used for setting standards which are used as a
basis for determining a fair day's work.
In many organizations, the
department doing time study work is called Industrial Engineering.
Taking a time study consists of breaking down the operation into
parts and recording the time taken for each of these parts by means of a
stop watch.
It also consists of recording job data, and setting the
This standard is expressed in time per unit and
standard for the job.
is considered representative of a fair day's work.
A time study primarily
records the cycle time of the operation.
In properly recording a time study there are certain data which
must be clearly set down so that anyone referring to the study at a later
date may know all the facts surrounding it.
Some such items are obvious
in their purposes - as for example, the part name.
Others are more ;n the
nature of manufacturing data.
The following is a list of those items which are most frequently
found on time studies:
o
Date of Study.
This establishes the day, month and year in which
the study was made.
Under certain conditions it might be necessary to
477
1
I:,
1
2
I
indicate the hour or the shift.
Effective Date.
This establishes the day, month and year on which
the study became effective.
It is highly important that this date tie
o
into ·the routing date.
Material.
This item should adequately describe the material used.
The kind, the specification number, the dimensions and any other facts
required for identification must be set down.
This is quite important
since the study is limited to the specified material, and any change in
the material may automatically call for a new study.
Part Name.
Carry here the full part name as it is carried on the
bl ueprint.
Part Number.
Accurately record the part number as it appears on
the print.
Operator Number and Name.
Aside from identifying the man upon
whom the study was madethe very act of learning the \'lOrkman ' s name -
o
if done in a tactful and friendly way - can help the time study analyst
in selling himself as well as letting him know the man.
This is impor-
tant in establishing a better attitude toward the study on the part of
the operator.
Department Number.
To show where the study was made.
Machine Number and Name.
This item should indicate by name, model,
and inventory number the specific machine studied to identify it since
no t\,/O machines, even though they be of the same make and model, perform
identically.
Cutting Compound.
Show the kinds of cutting compound used such as
oil, soap, etc., and wherever available refer to the specification
number.
The compound frequently has a very direct effect upon the
cutting time and the correct recording of this item is an essential.
o
"11
IJr
3
o
Operation Number and Name.
t~~
A code to identify the operation and to
it up with the routing.
Speeds and Feeds.
A machine time study that does no: rarry complete
and accurate speed and feed data has little or no value since it is such
data that tell s us
hO't1
effect; ve 1y the machi.ne was performi ng duri ng the
study.
Tools and Tool Number.
Here identify whatever die, tool or fixtures
are used in the study by number.
Where necessary a brief description may
be used as for example: Carboloy tips, special diamond dresser, abrasive
wheel
specif;catif~,
etc.
Like material data, speeds and feeds, and
cutting compound, this identification of tools and tool number definitely
limits and clarifies the study and a careful and accurate recording is
important since any change would qualify the operation for a recheck.
o
Production Per Hour.
When you have completed the study through the
development of the standard. calculate the number of pieces by dividing
1.0 hour by the standard time per unit.
Male or Female.
Indicate here the sex of the operator.
This in-
formation may be called for in labor discussions relative to whether a
job is correctly classified as male or female.
Allowances.
This space provides for entering certain allowances
for items not covered in the elemental detail of the study.
Total Tirre of the Study.
Here is Y"ecorded t;.,:: starting and stopping
time of the study, or the total overall time of the study.
This data
may be used as a check against the accuracy of the study.
Name of the Observer.
The time study analyst's name should be
signed here.
o
47q
4
Reason for Study.
This space is set aside for such general remarks
as the time study analyst may wish to make, for example, "Rechecked
o
Standard ...
Approvals.
Here is recorded the foreman's signature and all the
necessary approvals required by management.
Work Place.
A sketch to work place, piece part showing operation
or area worked on, etc.
The more completely and accurately we record on our studies all
the data relative to the job, the more valuable our studies become, both
to us and to those functional areas which will call on us every day for
information.
As time passes, we very quickly forget the details surround-
ing the taking of a study and when it becomes necessary to discuss it, a
month or year or two years later, we very often cannot recall the whole
picture to mind.
Therefore, it is for this reason that the true value of
full and correct job data is necessary.
The actual mechanics of determining the standard time are not
complex, but must be thoroughly understood and correctly done in order
to arrive at a basis for getting an accurate work standard.
The generally
accepted method is:
1.
Obtain time data from observation by gathering stop watch
readings, element frequency, foreign elements, performance,
and allowances.
2.
Calculate leveled average time for each element.
3.
Determine allor/ed time per piece.
4.
Calculate standard time.
c
~I
5
o
After completing the recording of stop watch readings, earmarking
irregular elemental times, and recording elemental frequencies, then
individual representative time values must be determined for each element.
These times for each element come from the recorded snapback or continuous
stop watch readings.
This procedure can be explained by means of a sample
continuous recorded study (Figure 1, page 11).
The first recorded time
from the starting of the watch until the end of the first element is .07
minutes; so this figure would be recorded opposite (I) in element #1.
Then the time for element #2 can be obtained by subtracting the continuous
reading of element #1 from that of element #2.
Thus, .38 - .07 = .31 and
this figure is recorded opposite (I) in element #2 and so on, until all
times are obtained.
These times, then, will reflect the standardized
method and conditions under which the operation is to be performed and
C,
c
l'
I
should be averaged in order to determine the leveled average time fur each
element.
This leveled average should be calculated by dividing the total
of all time intervals, which are not earmarked,by the number of such
intervals.
For example, the completed study of the first element in
Figure 1, page 11, covers a total number of 10 observations for the
cycles. With one of these recorded times earmarked, the leveled average
time should be determined by dividing by 9 the total time consumed by
the remaining 9 time intervals.
In this manner, the time for this regular
element will reflect the basic established method for the operator in the
performance of this particular part of the operation.
Earmarked irregularities are called foreign elements.
o
Where
the irregularity is timed and recorded with a regular element reading,
1-RI
6
the foreign element can be calculated by subtracting the leveled average
time from the recorded watch reading.
o
After this time value has been calculated, the prorating can
be done the same as if the irregularity had been timed and recorded
separately. These calculations with respective frequencies should be
plainly shown in the provided space on the time study form.
In addition
to foreign elements, the details of delays for allowances should be
filled in.
These will tend to vary depending on company policy.
If company policy involves performance rating, the next step
is to determine the operator performance by comparing the operator to
the average normal operator concept.
The objective of this rating is
to establish an allowed time which is representative of the average
normal worker.
This rating should be expressed as a percentage above
or below 100% which represents the average normal operator.
c
Next is to determi ne the allowed ti me per piece for each element
of the study.
This is done by multiplying the leveled average time by
the performance and the elemental frequency for each element.
Finally, thi.s method of calculation involves the summation of
all allowed times and all detail of delay allowances.
After this is
completed, subtract the total time for allowances on a per shift basis
from the total shift time to determine the minutes available for work •
..
Next, the minutes available for work time is divided
by
the summation of
the. allowed times plus delays on a per piece basis. This yields standard
pieces per shift •. Oivide this number by eight hours which gives standard
pieces per hour.
Divide the standard pieces per hour into one hour and
the final work standard is .expressed in hours per piece.
o
7
1
All of the calculations for leveled average time, allowed time
per piece, foreign elements, detail of delays and the setting of the
final work standard can be accomplished by tnp IBM 1620 Model II, 40k
computer.
In )rder for the computer to accompli sh thi s, the time study
I·i
analyst must prepare his studies in a language the computer understands.
Therefore, lt is required that all data be in decimals whereby the key
punch operator can take this data directly from the time study observation
sheet (Figure I).
The IBM cards should be prepared in the following
manner:
FIRST READ STATEr1ENT (.5F7.0)
a.
Part Number
b.
Operation Number
c.
Type of Study Code
C
= A" Continuous
12 = IIAII Snap Back
13 = IIBII Continuous
14 = IIB" Snap Back
15 = "C" Continuous
15 = "C" Snap Back
11
It
d.
Number of elements in the study.
e.
Number of columns of element readings.
SECOND READ STATENENT (lOF8.2)
Recording of Element Reading
a.
Keypunch individual stop watch time values in vertical
order regardless of the type of study.
o
b.
For any missing time values or those disregarded because
of some assignable cause code as
II
4f3
Ua:00::"4#::&*
(·M*"t#,,==.,",
_
zero • II
8
c.
In the case of a continuous type of study where a value
has been coded as "zero" add 1000 to next acceptable
o
reading.
d.
If any of the readings are earmarked as a foreign element
code as follows:
Type "A" add 900 to the value
.. "B"
.. "e"
.. "0"
..
II
Ell
800
..
II
.. 700 ..
..
.. 600
II
II
500
II
II
II
II
..
II
.
II
Note: Add zero if necessary for a total of fi vei terns.
THIRD READ STATEMENT (lOF8.2)
Element Performance Evaluation and Element Frequency
The performance and element frequency must be key punched
C
consecutively for each element respectively.
FOURTH READ STATEMENT (15F5.2)
a.
Foreign element frequency
b.
Detail of delay allowance
1.
Minutes/shift items
2.
Minutes/piece items
The frequency and detail of delay allowance items are key punched consecutively, i.e., frequency, minutes/shift item, minutes/piece item as line 1
horizontally, etc.
If any categories do not have five items in the
particular study, key punch zero as the missing items until a total of
o
five items each have been listed.
#4
1
9
o
FIFTH READ STATEMENT (7A4)
I.
Name card.
b. Continuous run - if more than one tfmestudy is run
at the same time, a 1 is key punched in column 30 to
separate the studies.
For example, with the data from the time study shown in Figure 1, page 11
as input to the IBM 1620, the computer would give the following output
for all the necessary calculations involved in stop watch time study
analysis.
c
o
,4
,iiiueam,.awu==iil4&&MA&
1
1.,II
111
10
TIME STUDY OBSERVATION OUTPUT
NAME
RON COX
PART NU~1BER
88128
OPER NUMBER
12352.
TYPE OF STUDY
11.
LEVELED AVERAGE
MINIMUM PIECE ALLOWED TIME
LEVELED AVERAGE
MINIMUM PIECE ALLOWED TIME
LEVELED AVERAGE
MINIMUM PIECE ALLOWED TIME
LEVELED AVERAGE
MINIMUM PIECE ALLOWED TIME
LEVELED AVERAGE
MINIMUM PIECE ALLOWED TIME
LEVELED AVERAGE
MINIMUM PIECE ALLOWED TIME
LEVELED AVERAGE
MINIMUM PIECE ALLOWED TIME
LEVELED AVERAGE
MINIMUM PIECE ALLOWED TIME
ELEMENT 1
.0733
.0366
ELEt~ENT
2
.3320
.1411
ELEMENT 3
.7270
.3453
ELEMENT 4
.1300
.1300
ELEMENT 5
.5114
.4347
ELEMENT 6
.0430
.0204
ELEMENT 7
.2566
.1219
ELEMENT 8
1.5260
.2746
DELAY MIN/PIECE
A=
.0106
o=
0.0000
B=
C=
.0258
.0282
ALLOWED CYCLE TIME/PIECE
ATTAINABLE HOURLY PRODUCTION
TOTAL DELAY (MIN/SHIFT)
TOTAL DELAY (MIN/PIECE)
MIN AVAILABLE FOR WORK
STND PIECES/SHIFT
STND PIECES/HOUR
STND HOUR/PIECES
1.5048
29.871
34.
.064
446.
284.155
35.519
.0281
o
I
I
,
'--"'-'~---
~~--~~-
ElEMENT DESCRIPTION
NO I
•
=1
!
UfT SIDE
~1 ement
I
RIGHT SIDE
._.
.
rA
5
6
7
8
I
1
9' flO
AVER
I I
.
~fL~
LEVEL-
PER-
ED
AVER.
FORM-I'
1
19511/2
\
\100 I 1/1
.07 3.57
I
I
I
I
I
f
I
I
/100
AL~?:eED
f MINS/PC
FREQ.
/ 85 I 1/2
I
I 1/2
I
I
c 1.1014.6218.26112.30\15.88119.82123.44127.09130.75134.74 1
1
El ernen t
4 I
I
I
t
J
~
3
I
t-ct--.-38+1-3.-92--11-7-.5-2+-1"-1.-57-+1-15-.1-31-\19-.0-7+\2-2-.72-+:\2-6-.3-81l-30-.-02-1--~134-.-02-11
c
Element
2
I
I
I
I
HeI
1
l
7.17 ~_1.2~14.82118.75122.38126.03129.68131691
11;1
Element
I
ELEMENT TIME
ELE.
NO.
t
I
J
-':4
Zl
I-t--.--t-4-.7-6-+--8-.4-11--.-4'+-16-.0-0+-,-g.-97-+-23-.58--+-2-7
c 1 22
12
>< ><:
,I
Flp-m~nt ia
J
C
I
I:.lement
B
1.71 5.29 8.88 12.93 16.88 20.47 24.10 27.75 31.77 35.54
I
Ip
I
I
'.74 5.34 8.9112.96 16.93 20.53 24.16 27.80 31.81 35.57
C
I
e
2.01 5.58 9.52113.22 17.19 20.79 24.4128.0532.07 35.83
I
I
e
1.
I
H----+--t---t---t---+---4---+--+--+---i
Elemen t #17
Element -f,fff
T
+
:
I
II
I
Ie
DETAIL OF DELAYS:
MIN/SHIFT
MIN/PC
Fumble
PERSONAL
P & cu
Stuck Part
Drop Part
24
10
ATTAINABLE
HOURLY PRODUCTION
60 MINUTES
CYCLE TIME)
= (ALLOWED
1/2
95
1/2
90
1/5
Dj
480 MIN. -
I
TOTAL
OR
------------
__
ALLOWED CYCLE TIMf/PC
~ = ________________
_______________________ ,
MIN )
AVAILABLE
( FOR WORK
~
(ALLOWED
CYCLE) ( STD PCS
TIME + DELAYS
IN MIN/PC
/SHIFT
(STD PCS/HR)
+ _______________________ = ___________ + 8 = _____ .. __ .__ .
STANDARD HOURS/PI ECE _____________________ ._. _____ _
TOTAL
1£-7
0
=
FINAL CALCULATION OF STANDARD:
ALLOWED MIN
( DELAY SHIFT
I
95
I~
FOREIGN ELEMENTS
-
1/'
t'-
~
...
~
c:
e 3.51 7.09 11.07 14.76 18.68 22.31 25.94 29.60 33.61 37.36
1
85
o
o
~
o
SESSION NUMBER
F.3.7
SPEAKERS
NONE - SQUAWK SESSION
DISCUSSION
PREPARED A LIST OF RECOMMENDATIONS FOR IBM.
IBM, WAS WITH US.
OICK PHILLIPS, OF
c
o
4ff
o
SPECIAL ADDED ATTRACTIONS
IBM 1130 SYSTEM
I RM QUI CKTRAN 0 F. M0 NSTRAT I 0 ,,!
o
o
4ft]
o
IBM QUIKTRAN DEMONSTRATION
Parlor J, 4th Floor, Netherland Hilton Hotel
QUIKTRAN is a time-sharing data processing system that brings
the power of a modem large-scale computer directly to the desk
of the engineer and scientist, the mathematician, the business
and financial planner - anyone who solves problems with numbers.
The QUIKTRAN system uses the FORTRAN language and has available a library of widely used programs.
To show you the QUIKTRAN system in
the following DEMONSTRATIONS:
o~
ration, we have arranged
c
1) Common library programs including:
a) ROOT, a program which finds the nth root of m
by Newton's iteration method.
b) QUAD, which finds the real and imaginery roots
of the equation ax 2+bx+c.
2) Programs written especially for the COMMON meeting
designed to demonstrate the conversationality of QUII)TRAN •
Demonstrations will be given:
Tuesday, September 5
Wednesday, September 6
Thursday, September 1
Friday t September 8
6:00
9:00
9:00
9:00
p,m.
a.m.
a.m.
a,m.
-
9:00
9:00
6:00
5:00
p.m.
p.m.
p.m.
p.m.
C
i
'" ",,;
410
---~-.--------------~------------
o
IBM
1130
SYSTElVI
An IBM 1130 has been provided for the COMMON Meeting. It will
be available from Wednesday (Sept. 6) through Friday (Sept. 8), to
all members for testing or demonstration purposes.
The 1130 consists of a 1442 Card Read/Punch, an 1132 Printer, and
8K disk. Two key punches, disk packs, and a System Reference
Library are also available for your use. The following programs have
been loaded on disk packs:
lVlonitor (Mod. level 4)
Commercial Subroutine Package
Scientific Subroutine Package
Statistical System
Vvork Mea surement Aids
Structural Engineering System Solver (STRESS)
lVlechanical DeSign System
Numerical Surface Techniques
Critical Path Method/PERT
Other Demo Packages & Utilities
The equipment is located in Parlor I of the Netherland Hilton. You
may sign up to reserve time for testing or demonstration. Please
feel free to stop by any time; there will always be someone there to
assist you.
o
4ql
LIST OF REGISTRANTS
SF:PTE~'1RF.R
C0/'"f'<10f'I
r~F.F.TI"'G
SEPTEMBER 6-7-8, 1967
o
CINCINNATI, OHIO
ARf'-tFR, ,I R
5904 SEHFLL RD
PEf\.ISACOLA,
ALLRR
FLA
32504
TTTON, E ,J
&3 BOX 191
C L AR0 S V ILL F,
t·10
63336
ALLEN, J E
1206 MULRERRY ST
DES tv10 I ~\I E S,
I 0 I~J A
ALVA.REZ, J J
IBM GLENDALE LAR DEPT
E1\1 I) I COT T, f,! • Y•
IV,tDER SON,
50309
265
13 7 60
B K
20 SOUTH ROAD
S 0 lJ T H I I'l GT(1 r", C n (\.1 N
ARt"1RRUSTER,
c
L F
25136
ARf\,lTSON, W M
1700 W. THIRD AVE.
FLINT, MICH
L~8
50 2
A J
ARTHUR,
3270 CARRILLO AVE
SANTA, CALIF
95051
ATER, G T
35 PAGE PO
CHILLICOTHE, OHIO
45601
AUSTIN, L B
1700 W. THIRD AVE
FLINT, MICH
RABER,
G
48502
~1
3000 SPOUT RUN PARKvJAY
ARLINGTON, VA
22201
BAILIE, 0 L
455 C STREET
v.JASHOUGAL, ''''ASH
98671
BAILEY, D C
MONTEREY AND COTTLE ROAD
SAN JOSE, CALIF
o
LIST OF REGISTRANTS
SEPTEMBER COMMON MEETING
o
SEPTEMBER 6-7-8, 1967
CINCINNATI, OHIO
BAKER, L H
1206 MULBERRY ST
DES MOINES, IOHA
r~
BALL,
50308
J
159 IDLE HOUR DRIVE
LEXINGTON, KY
"-0502
BALLENTINE, J D
547 PAIGIE STREET
SCHENECTADY, N.Y.
12307
BA.Rr40RE, D R
4780 SNO~' DRIVF
SAN JOSE, CALIF
95111
BARf\IFY, P L
740 S. ALABAMA
INDIANAPOLJS, IND
BAR R,
46206
S
222 BROAOIr-.IFL
NEH YORK, N.Y.
RE"IFOICT, 0 E
2000 FORRER RL\lf)
DAYTON, OHIO
45401
BERGER, 0 E
2605 REYNOLDS CR.
HUNTSVILLE, ALA
35810
R J
,)00 I
DR
HAI'-1~'1nND,
LA
B ER [\1 A R 0,
606
BERItI I C1<, J
B I C1< FOR 0,
70401
~:1
f'·IA II\!
H A1t! K E S RU RY,
rJ f\1 T
P /J,
GREENCASTLE,
o
10038
1;'·'0
46135
BLAC KI\! E Y, \~! C
T5&0, ARC, 2020 BLDG no I·.! (Htvl
r1 I 0 L A"I D, "·1 J C H
1
RLA.TCHLfY, C G
300 BENT RD
PA
t-.Jy~,jCOTF:,
19095
,'
',,t~1
I
LIST OF REGISTRANTS
SEPTEMBER COMMON MEETING
SEPTEMBER 6-7-8, 1967
o
CINCINNATI, OHIO
BLISS, R J
5830 WESTHENRIETTA RD
HENRIETTA, N.Y.
14606
BOBAY, J P
1000 5TH ST
COLUMBUS, IND
47201
BO"'IO, W F
656 STORM AVE
BROOKHAVEN, MISS
39601
BOSCHAN, C
144 EAST 24 ST
NEW YORK, N.Y.
10010
BRADY, J ..J
4620 FOREST AVE
NORWOOD, OHIO
45212
BRANAGAN, R E
2442 TRENTON AVENUE
MONTREAL,
BRASKAMP, B
1111 CONNECTICUT AVE
WASHINGTON, D.C.
QUEBEC
N.t-J.
20036
BRECHBILL, 0 0
625 CLEVELAND
COLUMBUS, OHIO
43215
BRENNER, R L
P.O. BOX 561
BURL I NGTON, IO\NA
52601
BRENNAN, R D
2670 HANOVER
ALTO, CALIF
BRIGGS, 0 R
430~6 BEDFORD
MIDLAND, TEX
79701
BROWN, L W
P.O. BOX 2328
MOBILE, ALA
36601
BUCKLEU, R E
1414 BEECH ST. S.E.
DECATUR, ALA
35601
o
LIST OF REGISTRANTS
SEPTEMBER COMMON MEETING
o
SEPTEMBER 6-7-8, 1967
CINCI.NNATI, OHIO
BUESKING, C W
4617 TAYLOR AVE
EVANSVILLE, IND
47715
BUFE, 0 E
2300 CHESTER AVE
CLEVELAND, OHIO
44114
BURGESON, J W
220 EAST UNION ST
WHEATON, ILL
60187
BlJRGGRABE, W F
1400 SOUTH THIRD ST
ST. LOUIS, MO
63166
BURNS, R A
ALGOMA STEEL CORP LTD
SAULT, ONT RIE
CANADA
BURRO~JS,
H
A
PITTSBURGH, PA
15225
BYTHER, T E
31 OAK ST
OLD TOWN, ME
04468
CAMPBELL, M
2240 SOUTH LONE PINE
SPRINGFIELD, MO
65804
CAPLAN, F L
215 ROSE HILL AVE
NEW ROCHELLE, N.Y.
10804
CARLSON, 0 R
5707 LINDENWOOD LANE
FAIRFIELD, OHIO
CARLSON, D M
327 S. FOURTH AVE
ANN ARBOR, MICH
CASTELLAN, N J
DEPT OF PSYCHOLOGY I U
BLOOMINGTON, IND
47401
o
CEELY, F F
6206 BREN MAR DRIVE
ALEXANDRIA, VA
;"
1,$1
k.. ' .;
48108
,$,
St$., ·,UL "
¥
I,
W4S
tiQJ4WA;;;Itc=*=====
22312
\l
1
I
LIST OF REGISTRANTS
SEPTEMBER COMMON MEETING
SEPTEMAER 6-7-8, 1967
o
CINCINNATI, OHIO
CHAIKIN, A ,J
9300 GEORGE PALMER HWY
LANHAM, MD
20034
CIPRIETTI, B J
34 IPSHICH PLACE
HAMILTON, ONT
CANADA
CLARKE, J R
449
vJ.
5 T H ST
CHILLICOTHE, OHIO
CLARK, LA
3520 t.-.I. ROI.J~!TR FE
SPRINGFIELD, MG
CLARK,
1t.J
i
45601
65804
H
EDGE BROOK ROAD
BINGHAMTON, N.Y.
13903
CLARK, A G
132 DAVIS STREET
PAINTED, N.Y.
14870
CLEGG, J B
346 COMPTON RD
CINCINNATI, OHIO
45215
4
c
CLOSMAN, S
112 EAST POST ROAD
WHITE, N.Y.
COLE, C T
106 KATAHDIN DR
POLAND, OHIO
44514
CONROD, R L
P.O. BOX 208
BEDFORD, MASS
01730
COOPER, T
USDAARSBSSARCBLDG226
BELTSVILLE, MD
20705
CORDING, B L
2705 DELLWOOD OR
ORLANDO, FLA
32806
CORNELL, R L
41 37 BE AR0 AVEt-.HJ E SOU TH
MINNEAPOLIS, MINN
55410
o
LIST OF REGISTRANT5
SEPTEMBER COMMON MEETING
o
SEPTEMBER 6-7-8, 1967
CINCINNATI, OHIO
COTTON, B W
P.O. DRAWER G
GRAPEVINE, TEX
cox ,
76051
R C
1429 KEED AVENUE
BATON, LA
CRAFT, J
70806
C
2109 NORRIS RD N.W.
HUNTSVILLE, ALA
35810
CRUMB, H F
33 LIBERTY ST
NE~'
YORK, N.Y.
10045
CUM·M I NGS, L R
FEDERALSBIJRG, MD
C·'
21632
CUNNINGHAM, A J
25 MAIN ST
ANSON I A, CO"IN
06401
DAGENFIELD, R L
1214 JAEGER ST
COLUMBUS, OHIO
43206
DANZEISEN, L A
WOODVILLE ROAD
TOLEDO, OHIO
43601
DEAKIN,
G R
302 MOUNTAIN DRIVE
PEARISBURG, VA
24134
DECK, J C
661 CHESTNUT STREET
VALPARAISO, IND.
46383
DEGENNARO,
M G
175 OLD COUNTRY RO
HICKSVILLE, N.Y.
DEL n~lG,
11801
H
SIBLEY STREET
HAMMOND, IND
629
o
463~O
DFr-...1JNG, J hI
3451 MCHENRY AVEf\!lJE
CINCINNATI, OHIO
1~5
225
LIST OF REGISTRANTS
SEPTEMBER COMMON MEETING
SEPTEMBER 6-7-8, 1967
o
CINCINNATI. OHIO
DEUTSCH, E
COMPUTER CENTRE
GENESEO, N.V.
DEWEY,
14454
G C
11444 LACKLAND ROAD
ST. LOUIS, MO
63141
DICOSTANZO, J A
P.O. BOX 390
POUGHKEEPSIE, N.Y.
12602
DONALDSON, D L
1570 WOODMAN DR 15
DAYTON, OHIO
45432
DONEGAN, A W
107 SMITHFIELD DRIVE
ENDICOTT, N.Y.
13760
DONNELLY, M J
906 BELGIAN AVENUE
BALTIMORE, MD
21218
DOUCETTE,
J
E
109 CHANDLER STREET
BOSTON, MASS
DOWD,
02116
C K
LOCHEARN DRIVE
BALTIMORE, MD
~711
21207
(1)RAY, R D
47 PAGE ROAD
CHILLIcotHE, OHIO
45601
DUNSMORE, D It.
414 WALNUT STREET
CINCINNATI, OHIO
45202
DUQUETTE, D J
THl~D AVE & FORDHAM ROAD
BRONZ, N.V.
10458
DWYER,
J
R
9321 E. 84TH TERR
RAYTOWN, MO
DYE • 0 R
40 SAWMILL RIVER ROAD
HAWTHORNE, N.Y.
64138
o
LIST OF
o
ReGISTRANT~
SEPTEMBER COMMON ME'ETTNG
SEPTEMBER 6-1-8, 1967
CINCINNATI, OHIO
EATON, T J
400 WASHINGTON AVENUE
ST. LOUIS, MO
63102
EDWARDS, R A
112 E. POST ROAD
WHITE, N.Y.
10601
ELWELL, W G
7030 STARR
LINCOLN, NEBR
68505
ENYEDY, G
AUBURN ROAD
PAINESVILLE, OHIO
44077
FAERBER, R B
1266 AVALON DRIVE
SAN JOSE, CAL
95125
FALCONELLO, P
THIRD AVE & FORDHAM ROAD
BRONZ, N.Y.
10458
FANUELE, V L
12 WILDWOOD DRIVE
WAPPINGERS, N.Y.
12590
FELICE, L
1341 BALCOM AVENUE
~EW YORK, N.Y.
10461
FELLER, G G
1402 10TH AVENUE S.E.
ROCHESTER, MINN
55901
FINCH, 0 G
3316 CROSS COUNTRY DRIVE
WILMINGTON, DEL
19803
FISHER, J N
363 THIRD STREET
CAMBRIDGE, MASS
o
02148
FITZPATRICK, E 0
ILLINOIS ,STATE UNIVERSITY
NORMAL, ILL
61761
FLEMING, C 0
P.O. BOX 34380
DALLAS, TEX
75234
LIST OF REGISTRANTS
SEPTEMBER COMMON MEETING
SEPTEMBER 6-7-8, 1967
o
CINCINNATI, OHIO
FO()OR, J E
ENGINEERING BUILDING
MADISON, WIS
FOERSTER, C S
112 VALLECITOS WAY
lOS CATOS, CAL
95030
FOLLAND, G H
MILFORD, MICH
48042
FOLTZ, T V
4818 AVONDALE DRIVE
FT. WAYNE, IND
46806
FORSTROM, R W
19401 DORAL COURT
YORBA, CAL
0
FORTUNE, F C
19394 GULFSTREAM DRIVE
TEQUESTA, FLA
FOWLER,
J
G
3730 MEADOWBROOK DRIVE
ZANESVILLE, OHIO
43701
~,
FRASER, W C
36 DOBIE
MONTREAL, QUE
FRASER, W I
21 ETON COURT
CAMlACHIE, ONT
CANADA
FULLAN, 0 J
112 E. POST ROAD
WHITE, N.Y.
GABBERT, D A
2117 INDIANA STREET
PARKERSBURG, W. VA
GABRIEL, R F
SOUTH ORANGE AVENUE
SOUTH, N.J.
07079
GANATRA, J t
Source Exif Data:
File Type : PDF
File Type Extension : pdf
MIME Type : application/pdf
PDF Version : 1.3
Linearized : No
XMP Toolkit : Adobe XMP Core 4.2.1-c043 52.372728, 2009/01/18-15:56:37
Create Date : 2014:03:05 14:58:16-08:00
Modify Date : 2014:03:05 15:07:55-08:00
Metadata Date : 2014:03:05 15:07:55-08:00
Producer : Adobe Acrobat 9.55 Paper Capture Plug-in
Format : application/pdf
Document ID : uuid:549c76c9-f47a-0249-912b-3380d7e6eb4e
Instance ID : uuid:94c8485c-7f84-3d4f-80ff-3930c4d2ff73
Page Layout : SinglePage
Page Mode : UseNone
Page Count : 566
EXIF Metadata provided by EXIF.tools