HP Forum Archive 19

[ Return to Index | Top of Index ]

IBM 360 is 45 years old! Message #1 Posted by Don Shepherd on 6 Feb 2009, 1:45 p.m. An article in my local newspaper from 2004, on the 40th anniversary of the IBM 360. In 1964, there were no PCs....

Re: IBM 360 is 45 years old! Message #2 Posted by Garth Wilson on 6 Feb 2009, 2:48 p.m., in response to message #1 by Don Shepherd So interesting: http://en.wikipedia.org/wiki/IBM_System/360 The 360 was the first computer I had real contact with in school. We had to write our Fortran programs out by hand on paper, then go to the punch-card machines and put one line of the program on each card, rubber-band the set together with our account number and so on, put it in a cubbyhole, and come back a couple of hours later to get a printout of all the reasons the program wouldn't run, then try to fix it, and try again. The people who could actually touch the computer seemed next to God. Running our simple programs got so much easier when I got myself a programmable calculator. Edited: 6 Feb 2009, 2:49 p.m.

Re: IBM 360 is 45 years old! Message #3 Posted by Michael de Estrada on 6 Feb 2009, 3:32 p.m., in response to message #2 by Garth Wilson Which version of FORTRAN did you use? I used FORTRAN IV and it was at my second job doing engineering problems. I've since used FORTRAN 77 for DOS PC's and Lahey FORTRAN 90 which is Windows compatible, except not Vista. I remember putting statements in my programs directing the computer techs to mount tapes, and they would actually physically mount a tape on a tape drive. My first mainframe computer was an IBM 1130 using paper tape input and programming in pseudo assembly code as a graduate assistant at MIT. Output was binary LED on a console.

Re: IBM 360 is 45 years old! Message #4 Posted by Namir on 7 Feb 2009, 1:08 a.m., in response to message #3 by Michael de Estrada The University of Baghdad school of engineering had an IBM 360. I learned programming in FORTRAN IV on that baby!!! I used to use it to do design projects. My classmates hated me for that!! Namir

Re: IBM 360 is 45 years old! Message #5 Posted by Jon S Canale on 6 Feb 2009, 3:33 p.m., in response to message #2 by Garth Wilson Fortran IV with WATFOR and WATFIV! Boy, we're getting old. Still have that dang book! Can't throw anything out!

Re: IBM 360 is 45 years old! Message #6 Posted by Forrest Switzer on 6 Feb 2009, 6:03 p.m., in response to message #5 by Jon S Canale Fortran II, in 1966 on a GE 225, I believe with probably 8K of memory and about 8 huge tape drives. Problem - just dump memory to paper and find it. Three years later - a different world. Probably I am a couple years older than you. Let's see in 1966 I was 23. :| Forrest

Re: IBM 360 is 45 years old! Message #7 Posted by Palmer O. Hanson, Jr. on 7 Feb 2009, 10:49 p.m., in response to message #5 by Jon S Canale Quote: Fortran IV with WATFOR and WATFIV! Boy, we're getting old. Still have that dang book! Can't throw anything out! I, too, can't throw anything out. I still have my copy of WATFIV that I used when I was introduced to FORTRAN on a Sigma 5 in the 1970's. I also have the FORTRAN AUTOTESTER, the pocket manual before there were pocket calculators. My first digital computer was a RemRand (later UNIVAC) 1103 at the University of Minnesota in 1960. I still have my class notes including a "Repertoire of Instructions". My notes say the 1103 had a 1024 word fast memory (CRT storage) and a 16,384 word slow memory (magnetic drum). Input was paper tape, magnetic tape or IBM cards. Outtput was paper tape, IBM cards, or magnetic tapes. Other computers mentioned in the class were the IBM 701-704-709 which had 2048 words of electrostatic storage and 8192 words on a magnetic drum. My second digital computer came in 1960 as an M-252 airborne machine manufactured by Hughes which was part of an inertial system by Honeywell installed in a drone by Fairchild. My third digital computer came in 1963 as an airborne machine manufactured by Honeywell which was part of an inertial system by Honeywell which was installed in the black birds. My fourth digital computer came in 1966 wih an airborne machine manufactured by Honeywell which was part of the first airborne inertial system using electrically suspended gyros. Programming on all of these machines was in ones and zeroes. We did our own scaling. I still have (I really can't throw anything out!) a set of "Numeric Conversion Tables from Octal to Decimal and Decimal to Octal" published by American Bosch Arma in 1959. My fifth machine came in 1968 and was actually the Honeywell computer network based on a DDC 516 in Minneapolis. That was where I learned BASIC. Then came the Sigma 5 and FORTRAN. A little story about my experience with the Sigma 5. We used IBM card input. For our first programs the instructor took my program cards, combined them with the program cards from other class members and submitted the batch to the machine. Later we were allowed to submit our own programs, but the insructor failed to mention the use of a couple of additional cards which would terminate our program, say in case we had an infinite loop in our program. One of my programs had a infinite loop and as luck would have it, it was the last program in the queue for the night which was run unattended. The machine ran all night and the accumulated charges would have been astronomical except that we hadn't been told about the additional cards which were needed. Palmer Edited: 8 Feb 2009, 9:46 p.m.

Re: IBM 360 is 45 years old! Message #8 Posted by Marcus von Cube, Germany on 8 Feb 2009, 5:27 a.m., in response to message #5 by Jon S Canale Memories arise! :) I onced managed to stop our university /360 with a WATFIV program. WATFIV (at least for us students) was a manged environment with very strict rules what could be done and what could not. I tried to exercise a sort of some kind on a virtual file defined with a DD statement. There was something wrong with the statement but the error handler was buggy and so the system stalled. This could be seen easily on the outside of the machine room which was separated from the data entry room with the card reader and punchers by a glass wall. There was a red/green light on the outside that turned read. And there were some people looking stunned at the machine's console. The problem was solved later by changing the DD card to use physical temporary storage and I could continue learning how to do a data sort in FORTRAN IV. I've never learned the successors of FORTRAN IV, I did some SIMULA programming for the Gesellschaft für Mathematik und Datenverarbeitung (GMD, now Fraunhofer Gesellschaft) in St. Augustin. After I left into the industry in 1985, I've never touched a mainframe terminal myself again. Marcus, just four years older than the /360... Edited: 8 Feb 2009, 5:28 a.m.

Re: IBM 360 is 45 years old! Message #9 Posted by Ed Look on 8 Feb 2009, 8:58 p.m., in response to message #5 by Jon S Canale Quote: Fortran IV with WATFOR and WATFIV! Boy, we're getting old. Still have that dang book! Can't throw anything out! Hey!! That's what I learned it with. But, as late as just a few years ago, I taught FORTRAN 77 to p-chem students. (Yeah, the nice ones grinned and the bad ones challenged me as to why I was using archaic things. But they all finally, at least in my presence, and at least verbally said my line that it gave them the basic mental framework for learning ANY programming language from that point on.)

Re: IBM 360 is 45 years old! Message #10 Posted by Martin Pinckney on 6 Feb 2009, 3:50 p.m., in response to message #2 by Garth Wilson Quote: The 360 was the first computer I had real contact with in school. We had to write our Fortran programs out by hand on paper, then go to the punch-card machines and put one line of the program on each card, rubber-band the set together with our account number and so on, put it in a cubbyhole, and come back a couple of hours later to get a printout of all the reasons the program wouldn't run, then try to fix it, and try again. The people who could actually touch the computer seemed next to God. Running our simple programs got so much easier when I got myself a programmable calculator. Sounds so familiar I could have written this myself! So it seems that when I entered engineering school at USC (the real USC, not Southern Cal!) in 1967, the 360 system there was practically brand new. Used it for Computer Science and Advanced Structural Analysis. Eleven years later in graduate school (this time at Louisiana Tech), they still had the same 360 (370?) set-up, except you could actually put your stack of cards in the card reader, and get your results right off the master printer yourself. They also had "dumb" terminals in the Civil Engineering department. However, having been away from the engineering profession and computers for years, I went with what I already knew, and made many trips to the computing center with my stack of punch cards. Because it was so busy during the day, I would go down there in the wee hours, when I could run my stack through the reader, and then in a few minutes (sometimes seconds) get my results right off the printer. A story to tell on myself: one graduate project was to write a simplified user-manual for a complex FORTRAN program to do pipe network analysis. The professor gave me the source code on paper. I spent many hours coding and debugging until I finally got it working. Only then (after proudly informing my prof) did I learn that the program was already on disk, I only needed to write the manual!

Re: IBM 360 is 45 years old! Message #11 Posted by PeterP on 6 Feb 2009, 6:12 p.m., in response to message #10 by Martin Pinckney :-) makes me think of a similar goof I did during one of the first weeks of my -psst - MBA... We had to analyze how bottlenecks in serial processes can hold up total through-put time and how to improve on through-put, cycle time etc by parallizing etc. They assignment questions said something about using a simulation program and I naturally assumed that we had to write a program to simulate things. This sounded quite the logical thing to do for me. So I wrote a - rudimentary but nevertheless quite functional - simulation program and answered all the questions. A class-mate had some trouble with the assignment so he asked me for help, which I did, showing him my little program and we went to class. He was a bit surprised that it needed a program and mentioned something about 'who the hell here can write a program like this'. This did not make a whole lot of sense to me, it was not that difficult to write in the end, so I did not think much of it. Boy was I wrong... When we went to class and started working through the assignement questions, it became apparent that our distinguished institution had purchased a simulation program for X k USD for the students to avoid overheating of grey matter. Literally with pressing of a few buttons it would answer the assignment questions. No need for writing it yourself (or thinking for that matter) Boy did they laugh about me (my friend could not help but tell 'only his best friends in the section' that I had written a program instead of reading the assignment carefully and downloading it... Cheers Peter

Re: IBM 360 is 45 years old! Message #12 Posted by Sancerre (Phoenix) on 6 Feb 2009, 6:52 p.m., in response to message #11 by PeterP That is classic. I'll remember your little story forever. Being a software developer at heart, I am pursuing my MBA and will finish in a few months. Your story is totally plausible, as hard as that might seem to be for the non-techies in your MBA. Naturally, I have no need to justify your response to the vast majority of the HP forum members. By the way, I have to tell a confession --> one of the best parts of getting an MBA is utilizing my modest arsenal of HP calculators. --Tod Edited: 6 Feb 2009, 6:52 p.m.

Re: IBM 360 is 45 years old! Message #13 Posted by Garth Wilson on 6 Feb 2009, 10:15 p.m., in response to message #11 by PeterP Quote: that I had written a program instead of reading the assignment carefully and downloading it Maybe the professor's assignment-writing was as bad as modern manuals. And for Michael de Estrada above, the answer was Fortran IV. Edited: 6 Feb 2009, 10:17 p.m.

Re: IBM 360 is 45 years old! Message #14 Posted by Mike Morrow on 6 Feb 2009, 11:09 p.m., in response to message #1 by Don Shepherd At Georgia Tech in the late 1960s and early 1970s, the IBM System/360 was in an admin building and was restricted to university administrative (financial and academic records) services. The scientific and computer services were handeled by Univac 1108 and Burroughs B5500 systems in another building. I spent many an enjoyable evening with the U1108, especially after I got ASR33 terminal priviledges. The B5500 stack machine, with Algol as its principal programming tool, remains to me the most intriguing system. I always considered the HP RPN stack machine calculator and the Burroughs RPN stack architechture to have much in common. Following university graduation and naval service, I attended for short while (in 1980) the U. of Arkansas. There I got my first actual experience with the System/360 in the form of the later System/370. Does anyone here recall IBM's cryptic Job Control Language (JCL), and acronyms such as HASP (Houston Automatic Spooler Program)? There was a famous (but likely ficticious) bumper sticker of that era that read "Honk if you love JCL!" I will say one thing...I wrote IBM Corporation as an EE grad student for technical documentation for that System/370 and received without charge several hardware manuals totaling several thousand pages. I wonder if there are any System/360 or /370 units still in operation in the world today. For that matter, are there any Univac 1108 or Burroughs B5500/6500 type systems still in operation? These are all great milestones in the history of computing.

Re: IBM 360 is 45 years old! Message #15 Posted by Don Shepherd on 6 Feb 2009, 11:31 p.m., in response to message #14 by Mike Morrow Yes, I too learned JCL for the 360. It was maddening! Like you, I later worked on Univac 1108's at the Census Bureau beginning in 1974, using EXEC-8 and FORTRAN-V, as I recall. One thing I remember from those days at Census was that they did not use the standard FORTRAN I/O; they claimed it was much too slow, and the Bureau's processing was much more I/O oriented than CPU crunching. So the Bureau systems guys wrote a custom I/O package that all of us programmers used. The systems guys also wrote a custom security sign-on system for the "demand" terminals, when they replaced the keypunchs and ASR-33 teletypes. I doubt that many of these old systems exist anymore, outside of a few museums. The first commercial installation of a digital computer--Univac 1--was right here in my hometown at General Electric Appliance Park, around 1953 or 1954. I wrote them a few years ago to see if they still had that historic system in some warehouse somewhere, but it had vanished over the years, much like my Beatles record albums, unfortunately.

Re: IBM 360 is 45 years old! Message #16 Posted by Herbert Crepaz (UK) on 7 Feb 2009, 2:25 p.m., in response to message #15 by Don Shepherd For all of you who enjoy to timewarp back to those days of cryptic command line sessions download the excellent Hercules IBM mainframe emulator (http://www.hercules-390.org/ ) and try OS/360 or MVS on your laptop for some big iron nostalgia.

Hercules and more Message #17 Posted by Mike (Stgt) on 9 Feb 2009, 7:33 a.m., in response to message #16 by Herbert Crepaz (UK) Last month Hercules was updated to release 3.06, see here: http://www.hercules-390.org/hercnew.html With it I get about 10 MIPS on an Atom-CPU (MSI Wind-sized toy), not overclocked. For those interested in WATFIV, look for Sim390. A must-have for mainframe-dinos (like me). Ciao.....Mike

Re: IBM 360 is 45 years old! Message #18 Posted by Keith Gobeski on 9 Feb 2009, 11:54 a.m., in response to message #14 by Mike Morrow The Burroughs B5000/B5500 descendants and the Sperry 1100 descendants are the current Unisys Libra and Dorado product lines. The high-end machines in the product lines still use proprietary processors, but the rest of the two product lines use Intel processors executing the same machine code as the high-end machines but in virtual environments.

Re: IBM 360 is 45 years old! Message #19 Posted by Karl Schneider on 8 Feb 2009, 3:38 a.m., in response to message #1 by Don Shepherd Hmm, lots of quotes in this thread to which I can relate... Quote: Does anyone here recall IBM's cryptic Job Control Language (JCL), and acronyms such as HASP (Houston Automatic Spooler Program... Yes, I too learned JCL for the 360. It was maddening! Like you, I later worked on Univac 1108's at the Census Bureau beginning in 1974, using EXEC-8 and FORTRAN-V. One thing I remember from those days at Census was that they did not use the standard FORTRAN I/O; they claimed it was much too slow, and the Bureau's processing was much more I/O oriented than CPU crunching. So the Bureau systems guys wrote a custom I/O package that all of us programmers used. In the latter 1980's, I programmed on Sperry/Univac 1170/1180/1190's. These had replaced IBM System/360 mainframes, I was told. Sperry's Executive Control Language and unique terminology could also be arcane. Its command for printing was @SYM, for "symbiont". Executable files were called "absolutes", and compiled object-code files were called "relocatables". Our organization also wrote custom I/O programs based on Sperry's I/O functions, because they were much faster than standard Fortran I/O. Later, a directive came down to make our software as ANSI-standard as possible. Quote: I used FORTRAN IV and it was at my second job doing engineering problems. I've since used FORTRAN 77... Fortran 77 was the standard in the 1980's, but some of the earlier software had been written in Fortran IV. Quote: We had to write our Fortran programs out by hand on paper, then go to the punch-card machines and put one line of the program on each card, rubber-band the set together... ...put your stack of cards in the card reader, and get your results right off the master printer yourself. They also had "dumb" terminals in the Civil Engineering department. The Sperry systems included an industrial-grade card-reader in case the "dumb-terminal"-based I/O failed. Quote: Problem - just dump memory to paper and find it. Sperry's core-dump files were encoded in octal, to match the 36-bit words in RAM. I wrote a program on my HP-15C specifically to decode the floating-point numbers printed on a stack of fan-fold paper. Had I known about the HP-16C, I might have bought one and written a more-refined program. However, the HP-15C did the job. Sperry merged with Burroughs to form Unisys in 1988. -- KS Edited: 8 Feb 2009, 3:42 a.m.

Re: IBM 360 is 45 years old! Message #20 Posted by Forrest Switzer on 8 Feb 2009, 4:37 p.m., in response to message #19 by Karl Schneider Yep, Octal dumps on fanfold paper, until memory got too large, then we had to set up a way of printing variables at different points through the program to make sure that everything was progressing properly. We wrote card images on pads of paper - 25 lines (cards) per sheet and 132 columns per line. We then gave these sheets to card punch operators where each letter or number on each line, in each column, was typed and each line became a card. Then the cards were given to another card punch operator where each card was again verified. I thought at the time that this must be the worst job in the world. Eight hours per day, 5 days per week making sure you typed every character in the correct column. I figured I had it easy - writing flow charts and the Fortran commands on the sheets. In 1966 I could walk in and talk to the computer operator while he ran my code. In 1969, the computer facility was in a building with wire mesh in the walls to prevent the possibility of monitoring the computer from the outside. There were two doors you had to be buzzed through to get into the computer. The punch card room and the computer room were classified secret, with special times where that was even more restricted. When I left in 1970, we were running 24-7 and sending stuff to Befesda(sp?) Maryland, Slidell Louisana, Boulder Colorado, and Port Hueneme California. I still can't believe that I was involved in such an incredible growth period during those years. My last day at that job, I drove out, stopped at the computer building, looked at what was run the night before, made some suggestions on the printout, and drove 1800 miles home for Christmas. Too many years ago. Forrest

The Early Days!! Message #21 Posted by Stuart Goldberg on 8 Feb 2009, 9:17 p.m., in response to message #20 by Forrest Switzer I started as a programmer trainee on a Univac 490 Real Time system at Bethlehem Steel in Bethlehem, PA, back in 1965. We wrote our programs in an assembly language called "SPURT" (Symbolic Programming for Univac Real Time) computers. We were just starting to play with COBOL. To create a COBOL executable, we ran our source code through a pre-compiler which created a SPURT program which then had to be assembled. I still remember proofreading a listing of a program punched card deck as opposed to "wasting" computer time compiling it right back from keypunch. On one of my program tests of an assembly program, my instruction sheet to the operator requested a memory (core) dump of the program area after the program had ended as I had left several "breadcrumbs" in memory rather than printing them out. The operator forgot to perform the manual memory dump request and I therefor had to re-run the test. When I called the operator to complain the next morning, he innocently (and in all seriousness) asked "Where was your program loaded? I can print that memory area for you now!" Today, almost 45 years later, I can still recite the major U-490 machine instructions and their octal op codes. I have several sheets of core memory from that old machine on my desk today and enjoy explaining to the young computer kids today how the destructive read-out worked. They believe me when I tell them it was invented by my cousin, "Rube Goldberg". :-) Several years later, I was providing support to the U.S. Navy's Univac 494 systems at the Aviation Supply Office in Northeast Philadelphia. I remember helping programmers debug their COBOL programs by reading thru the octal memory dumps and reverse compiling the instructions by hand back into assembly language which I could logically link back to the original COBOL source to follow the internal flow of the program. Some of the programmers were astounded that the core dumps did not contain the literal text of the COBOL instructions (e.g. multiply ordered_quantity by unit_price giving ordered_cost.) Stu

[ Return to Index | Top of Index ]

Go back to the main exhibit hall