A Truly Courageous Business Decision

Today marks the 55th anniversary of IBM’s announcement of the System/360 line…which made obsolete virtually all of its then-existing products.  The 360 line established a common architecture for machines of widely-differing price and performance characteristics, with the individual product implementations of this architecture differing considerably.  The goal was compatibility, so that customers could upgrade without extensive rewriting of programs.  Consolidating IBM’s diverse computer systems into this single system architecture was a bold decision; truly, a bet-the-company decision: in the end, it paid off, with devastating consequences for the ‘Seven Dwarfs’ who were IBM’s competitors at the time…but the implementation was frighteningly stressful.  A good article on the project recently appeared in IEEE Spectrum.

Tom Watson Jr, who ran IBM during this time period, discusses the 360 project extensively in his superb memoir, Father, Son, and Co.  I reviewed it here–highly recommended.

21 thoughts on “A Truly Courageous Business Decision”

  1. It truly was a bet-the-company decision. And I remember reading in the 70s, when its successor the IBM 370 came out – Forbes ran a list of the 50 biggest computer companies – and this is how it sorted out: IBM, then the 49 others whose revenues combined equaled IBM. #2 was Digital Equipment, now long gone.

    Something else: I loved the series Mad Men for their research into the norms of the 60s decade – from 1960-1969 – things changed radically.

    I saw things in this series that I had long forgotten – how people acted – like people leaving a picnic in the park, shaking the blanket and leaving all the trash on the ground.

    I was hoping this scene from Mad Men would show the new 360 displayed with glass walls in the lobby – but they didn’t show this scene. But they were talking about it in this scene – – that they know – they are coming into a new age. Companies wanted to showcase their very expensive mainframe computers.


    Don’t know what memory sizes the 360 had – of course – up through the range I’ll bet from a couple of meg to 64 MB?


    This old desktop I am using on Win 7 has 8 GB. Gigabytes.

    Cost about $20 IIRC.

    I know one of the top-of-the-line 370s – 370/168 – bank of America had one in San Francisco running all their ATMs – had 128 MB memory.

    This was core memory, an invention by An Wang (who had a minicomputer company in the 70s – core memory was very expensive.

    (My father in the early 70s bought a Burroughs B-700 computer – disc storage in platters was 2MB – had 48KB – KiloBytes of memory – IIRC you could get an extra 8 kb module for $8,000?

    BTW it was a Burroughs invention of virtual memory – having the operating system store needed tasks onto the disk drive when the main memory was full. Even your PC or Laptop is using this.

    Well, I am running on now…


  2. This post and the linked book review sure gave me a trip down Memory Lane.
    When I entered college in fall 1970, I ended up in a class that included computer programming, which was at that time a relatively new and esoteric offering.
    We ran ALGOL & FORTRAN programs on a Burroughs 5500 mainframe, punched into the legendary cards.

    (Turn-around time was spent studying for other classes. Programmers today have no clue…)

    In the fall of 1971, I went to the first meeting of my new programming class.
    “Good morning,” said the Professor. “We don’t have the B5500 anymore; we acquired an IBM 360/370 over the summer. It will be programmed in a language called PL/1.”

    That’s Programming Language One for you young’uns.

    “I don’t know PL/1,” he continued. “Your labbies don’t know PL/1. Your first assignment is due Friday.”

    And we still used punched cards, up into the early 1980s when I was teaching programming in another state.

  3. When I was going to school in Virginia – 1970-1972, I became enthralled with an HP 3000 Series II Mini computer. So much so that on a Friday I would stay until 3-4 AM, playing 3 dimensional tic tac toe. No monitors, but like playing chess through the mail, you would key in your move by coordinates, and it would give you its move on a printout.


    Now there’s a programming language from the past. IBM development wasn’t it? Like RPG.

    IIRC it was a combination of FORTRAN and COBOL.

    Had a class in it, but never programmed in it professionally.

    Now assembler – for the IBM Mainframes – try as I might never really got the hang of it with those registers and all. But if you wanted speed and non-trans-portability…

  4. I started working for Douglas Aircraft in 1958 as a junior engineer. It was the wind tunnel facility in El Segundo, CA. I was mostly working with a Marchand Calculator The wind tunnel had a one square foot chamber and our offices were in several trailers. The air-conditioning did not work well and we usually had the doors open. This resulted in flies and, being engineers, several devised novel approaches to the fly problem. Some of the guys would catch a fly and glue a tiny thread with a tiny square of colored paper attached. They would then release the fly, which would slowly buzz around the trailer dragging its flag. It was a bit like a small airplane towing a banner along the beach in summer.

    Douglas completed the new four foot wind tunnel and offices in early 1959. We now had luxurious quarters and an IBM 650 main frame. This had its own room which was kept at about 65 degrees. I was mostly working with the printers by this time. These printers could be programmed using a board like a switch board with little plugs and wires. They programming involved formatting the output. They took the standard stack of punched cards that was generated by the 650 running a program. The programs had to be quite small as there were only 2kb of memory. The data was all on cards.

    By the summer of 1959, I was starting to think of medical school and in the fall, I left to enlist in the Air National Guard as I was 21 and 1A. I had been going to night school taking pre-med classes since January 1959.

  5. @Mike – that fly story cracks me up. Except traveling flea circus sounds better than traveling fly circus. Been meaning to ask you how do you easily create links in your posts? Sure looks neater than my just printing out their URL. Do you have different privileges since you’re also a contributor?

  6. I have a collection of engineer stories. The links posting is just text

    The preview feature is nice because any extra or missing letter will make the link dead. The quotes have too be there, too.

  7. One thing that was key to the success of the 360 project was the management of the politics within the company. The is credited to Vin Learson, who was in charge of future product development. Watson:

    “He did it by applying a management technique called ‘abrasive interaction.’ This means forcing people to swap sides: taking the top engineer from the small-computer division and making him boss of the best development team in the large-computer division. A lot of people thought this made about as much sense as electing Khrushchev president.”

    Frederick Brooks was a passionate supporter of the envisaged 8000 large-computer product, which was planned to be killed as part of the 360 project. What happened:

    “Bob Evans immediately asked Brooks to develop the plan for a compatible family of computers. Brooks was flabbergasted, but he accepted, and with that the two engineering communities stopped feuding and began collaborating. There were still opponents in the company, but no matter—the trajectory toward a common system had been set.”

  8. He did it by applying a management technique called ‘abrasive interaction.’

    That sounds familiar. In his book, Ben Horowitz wrote about using the ‘Freaky Friday’ method in his company when the customer support department was warring with the sales engineers. He swapped the department heads, supposedly after watching the 70s movie.

  9. “IIRC it was a combination of FORTRAN and COBOL.”

    I once had an employer ask me if I coded my PL/1 as if it were FORTRAN or as if it were COBOL. I replied that I coded as if it were PL/1 (and yes, I did use the other two languages qua themselves).

    Mike K & the Fly:
    One of my early jobs was developing a construction cost accounting system to be used in the field by a power utility. Our equipment was primitive floppy disc readers hooked to programmable calculators (my electronic kitchen timer probably has more memory these days).

    One day the system crashed, and we joked about which of us was going to have to debug it. The engineer did the honors: after opening up the case of the floppy drive, he found a dead cockroach inside.

  10. Cribbed from the webz:
    a left-angle bracket .

    Specify the target in the argument (a href=” “). Then add the text that should work as a link. Finally add an (/a) tag to indicate where the link ends.

    Being interpreted – ignore the bolding, it’s just to make the relevant parts stand out more:

    Figure out your sentence and decide where you want to put the LINK that you are going to disguise with some TEXT word or phrase.
    Type the first argument a href=” ” but delimited by angle brackets instead of parentheses.
    Put your URL inside the quote marks, replacing the space: this is the LINK.
    Put your TEXT after the > (ignore the period; it’s grammar).
    Put the final tag /a delimited by angle brackets after your text.

    You should have something that looks like this in clear, but with the () to substitution, and MYURL being a regulation http thingie:

    I took this example from
    (a href=”MYURL”)here(/a).

    I took this example from here.

  11. You should have something that looks like this in clear, but with the () to ANGLE BRACKETS substitution, and MYURL being a regulation http thingie:

    HTML obviously takes any set of paired angle brackets to mean something, even if there is nothing inside of them.

  12. Revision 1.3
    a left-angle bracket or a right-angle bracket.

    I made the mistake of actually typing the angle brackets, thus eliminating the words inside them, but I would have sworn that it showed properly in the first preview.

    Clearly, one should not ever type a set of matching angle brackets even if the text between them is NOT a legitimate HTML command.

    Debugging comments is as bad as debugging code!

  13. AesopFan, thanks.

    Bill, anyone may post an “href” link here, following the instructions on the page AesopFan links to.

  14. One day the system crashed, and we joked about which of us was going to have to debug it. The engineer did the honors: after opening up the case of the floppy drive, he found a dead cockroach inside.

    That is alleged to be the origin of the term “Bug.” There was an example at Douglas when I was there. The main plant had a new IBM 704, which was replaced about the time I left with an IBM 7040, the first version with transistors instead of vacuum tubes. Anyway, there was a big problem at one time over at the main plant which turned out to be a “BUG” in the entomological sense.

    The 7040 gave way rather quickly to the 360 system but I was in medical school by that time. I did get one inquiry from somebody at IBM about working on medical applications but the systems at that time were not at all useful in medical applications, as best I could figure out. Medicine is pattern recognition, not calculation.

    My experiences with programming gave me a good perspective on the problems Sears had at the time. I had worked at Sears part time while in college. Maybe I should do a post about that one day.

  15. We ran ALGOL & FORTRAN programs on a Burroughs 5500 mainframe, punched into the legendary cards.

    (Turn-around time was spent studying for other classes. Programmers today have no clue…)

    How about waiting a day for the output, coming in next morning to find that all was well except for some stupid error on one card?

  16. @Mike, I have a similar fly story from when I was stationed in the Philippines. I worked at a comms site, so the air conditioning worked exceedingly well. No thermostat, 2 large units just ran full blast all the time with another as backup (we supported SAC and SAM flights such AF1 and AF2). Which kept it chilly in daytime in the dry, but at night during the (cooler) monsoon season, field jackets with liners, hats, and gloves would come out. And blankets for the operators. Anyway, enough flies would come in as people went in and out on smoke breaks, etc. One day when everyone else in maintenance was at lunch except me and my buddy, a very large horsefly had gotten in and was buzzing around. We stunned and caught him, superglued a bit of filament to his belly with a strip of rice paper glued to it. On it in red marker, we wrote “Eat me” (from the Deathmobile in “Animal House”) and turned him loose in the NCOIC’s office. It was hard keeping a straight face as he went back in after lunch. He came out in a few minutes with a funny look on his face, announced that obviously we two chuckleheads had far too little work to do, so…

  17. “How about waiting a day for the output, coming in next morning to find that all was well except for some stupid error on one card?”

    Been there, done that.
    Dropped a deck one time….

Comments are closed.