Early distributed computing video, 1959, prefigures the net

Rogue archivist Rick Prelinger sez, "Simon Ramo (the 'R' in TRW Corp.) commissioned this film in 1959 to describe his concept of 'polymorphic computing,' which means distributing computer power over many generic machines. With friendly moving wooden blocks, charcoal drawings, acoustic guitar and a German-accented narrator, this is the best unknown film-prefiguring-the-Net I've ever seen. Please download, show and propagate this gift from Prelinger Archives."

Simon Ramo's 1959 film on "polymorphic computing" anticipates the Net (Thanks, Footage, via Submitterator)

29

  1. Based on his bio, I’m willing to give the double-Caltech-PhD-at-23 Simon Ramo all the credit for being what is clearly the genius behind this video.

    Seriously, this is so far ahead of its time that it feels like a modern-day ad, done in a retro fashion by what would still be considered futurists.

    So much so that I’m suddenly wondering what credit this guy Ramo has gotten? It often occurs that science re-invents something someone thought of (and died believing, alone)–but history corrects itself and the poor brilliant guy gets credit.

    When I saw this I thought, “That’s a Greek von Neumann,” thinking of the every-man approach I (unfairly?) associate with Greek mathematicians. Evidently Ramo knew von Neumann, but was wholly American.

    I am very quickly interested in knowing more about this guy.

  2. 1959 is only a couple of years before ARPA started funding the networking research that built the ancestor of the Internet. Distributed computing was definitely one of ARPA’s explicit goals. So I wouldn’t be at all surprised if Ramo’s advocacy was one of the factors that inspired the ARPAnet.

  3. Music by whom? (Rather poor eyesight up in here, complicated by Mad Men related Canadian Club consumption…)

  4. It’s not surprising to me that he’d come up with this idea, since the computers of the day were as big and inflexible as he describes. It’s interesting that it took a couple decades for it to come to pass.

    I think you’d have to credit companies such as BBN and DEC and Sun for implementing these ideas in the 1970s and 80s. Of course, by then it was obvious that computers would exist in networks, but that was not at all obvious in 1959.

  5. A little searching has turned up that Ramo was interested (obsessed?) by tennis strategy–I count three books on the subject. Given that his textbooks involve control theory and chaos, I’d speculate that Tennis by Machiavelli is a pretty damn good read.

    Anyone read it?

    Given that there’s no obvious biography and that all of the above are out of print, I’d further speculate that the odd science/history student could make a great thesis on even these dying embers.

  6. Not only is there obvious genius in the ideas behind this film, it has managed a feat I have long thought impossible.

    It has made someone with a German accent sound friendly.

    1. Not familiar with the disney character Ludwig Von Drake, are we?

      I think Cory linked to a video where said character was used to introduce color television to the US people.

  7. This system is, IMO, somewhat different from the-internet-as-we-know-it. It’s describing a -modular- system where specialized (and, presumably, standardized) sub-systems are connected via a switchboard. In the internet, you have standardized -connection protocols- among autonomous, independent processors– leading to a self-organizing and self-repairing web. Not the same thing.

      1. No, MattF got it right. what they are describing is a single computer system where you can “plug in” additional functional units. What they are describing is closer to a bus-based system where you can insert things like memory boards or additional multiply units or …

        What is quite clear from the video is that the cubes are not independent and fully functional computers, but rather are component modules for a complete computer. Given the incredible expense involved in (say) building a arithmetic unit, the option of custom tailoring a mainframe computer to your particular needs had a lot of appeal.

        The idea of multiple independent functional units was widely adopted during the sixties and seventies, but the idea of pluggable units was not. There was a lot of innovation during this period, and it is hard to say who influenced who.

        This was a good, or perhaps a great, idea, but it really isn’t all that analogous to either distributed computing (except that brief mention of tying multiple systems together) nor of the Internet.

        By the early seventies, there were people who foresaw those things a lot more clearly – Ted Nelson being one notable example…

      2. And sometimes i find myself wondering what could have been during the BBS era, if a single phone line could have carried multiple calls so one could access multiple boards at the same time.

  8. That sounds like a Hungarian accent to me!

    What he’s describing sounds a lot like a modern compute cluster..

  9. Note at the end, he says they’ve already built one. I’m going to think of a problem and try to get some time on this polymorphic computer to solve it.

  10. Looks like the Intel 432 architecture almost verbatim. I sat through internal training which showed a matrix of processors on the left and memory on the bottom connected by a switching chip. The switching chip may never have been produced.

  11. It all becomes clear when he notes, in passing, that it can be called “The Matrix”. I wonder when in 1959 they plugged us all in for power?

  12. not so much the internet, but this is exactly what goes on in a CGI render farm.

    So much so i may actually send this video to clients who ask what a render farm is.

  13. Directed by Bill Orr!

    Intersting. Orr was integral to setting up the First Motion Picture Unit during WWII. This group of Hollywood-filmmakers turned-“soldiers” included Frank Captra, among others, and produced thousands of training films during WWII. It’s only mildly surprising that he would have remained involved with defense contracting companies and help them out after that time.

    He became the first head of Warner Brothers Television.

  14. I’m guessing this is also the only vision-of-future-computing video to ever use Old Joe Clark as background music (around 2:45). Frailing banjo FTW!

  15. He barely touches on “distributed computing”. Most of what he’s discussing in the first 2/3 of the segment are system architecture developments that have allowed for automatic time-sharing of the resources available to a single CPU (i.e. DMA (direct memory access), where a controller outside the CPU takes charge of multiple memory load/store ops; thus freeing the CPU of this comparatively mundane task, in favor of more arithmetic work, messaging, or whatever else it can accomplish while waiting for the memory work to finish)

  16. This is an ad. Where did it play? In movie theaters or on some of the 3 dozen 6″ TV sets that existed? Maybe on a very early version of YouTube that the guys waiting for computing time could watch?

  17. Rufus — if you listen to the end, it sounds like the film was made for a small number of potential customers who might want to buy a polymorphic computer.

    Unlike many other advertising films of the era, it wasn’t made for theaters — that would be a far too general audience. Nor was it made for TV, though many other sponsored films were shown to tens of millions of people in the late 1950s (by 1960, there were 50 million TV sets installed in the US, and many of them were 25″ color receivers).

  18. This is a nearly exact description of the AN/UYS-2 Enhanced Modular Signal Processor (EMSP) developed at NRL and Bell Labs in the early 1980’s for submarine signal processing (SONAR, etc.) It consisted of a number of different elements connected by a very high-speed switch that could route data streams from the output of any one component to the input of any other. The elements came in several different flavors:

    o general-purpose computers, originally Motorola 68000’s, programmed in Ada.

    o microprogrammable array processors, 4×4 SIMD multiprocessors. A fast Fourier transform would be a typical instruction/operation.

    o memory stores – big RAMs with some logic to do memory management.

    o I/O processors that read sensors and drove displays. A typical sensor would be a large array of microphones being towed behind the submarine or ship.

    The whole thing was programmed in data flow diagrams that were run interpretively by one or more of the general-purpose computers. I take pride in the fact that that was my idea originally, when I was at NRL, and that I led the team that developed ECOS — the EMSP Common Operating System. The architecture is still in use, after several generations of hardware upgrades.

    1. God I would love to work on a project like that. My education feels so limited, but I was taught (and learned well) how to implement most of the applications you listed for that sub, with CPUs, EPROMS and discrete components.
      Now if I ever want to get hired in something fun like that, I’ll have to go back to school – a trade school – so that I’ll have experience implementing such projects either with wasteful PLC’s, or thousands on thousands of lines of volatile, less dependable HLL code.

  19. Ok, i am officially freaked out by the spread of people that finds their way into the boingboing comments. Wish i could buy you a beer, Bob.

  20. Regarding the narrator’s accent. Somebody said Austrian, but this is not Austrian. This is a Hungarian accent. No doubt about it.
    Cheers.

Comments are closed.