Young Von Neumann Encounters Punched Cards

[beamjockey]

In addition to his work on pure mathematics, John von Neumann contributed fundamental advances to dozens of fields, from quantum mechanics to weather prediction. In particular, he was a pivotal figure in the development of electronic digital computers.

I've been reading George Dyson's terrific book Turing's Cathedral: The Origins of the Digital Universe. Within its pages, I found an interesting quote. Von Neumann's father Max was a Budapest banker who discussed his work with his children at the dinner table. Dyson quotes Nicholas A. Vonneuman* in John von Neumann As Seen by His Brother. Dyson writes:

Max believed in demonstrating practical examples of the industrial applications of finance. "If these activities involved financing of a newspaper enterprise, the discussion was about the printing press and he brought home and demonstrated samples of type,” says Nicholas. “Or if it was a textile enterprise, e.g., the 'Hungaria Jacquard Textile Weaving Factory,' the discussion centered around the Jacquard automatic loom. It probably does not take much imagination to trace this experience to John's later interest in punched cards!"

This offers a connection between Jacquard's loom and an individual deeply involved in developing the stored-program architecture now featured in virtually all computers. I should add this quote to my "Babbage's Favorite Picture" talk.**




* Different members of the family Anglicized their Hungarian names into different spellings.

**It took me fifteen years before I managed to find an image of the Jacquard Jacquard, but I now see that it pops up on Wikipedia for all to see in seconds.

Comments

[mmcirvin.livejournal.com]

I recall once going to a science or technology museum (it might have been the Smithsonian's National Museum of American History, which has doubled as a tech museum at various points in its career) where there was an operating model of a Jacquard loom all done in transparent Plexiglas, so you could see what was going on.

[neowolf2.livejournal.com]

I wonder how early anyone theorized about cells having a Jacquard-like programmable mechanism for making proteins. I think the idea that proteins have specific structures was actually fairly late (even though it should have been obvious after the first protein was crystallized.)

[mmcirvin.livejournal.com]

Erwin Schrodinger wrote a book in 1944 that had the basic idea, but I wouldn't be surprised if you could go earlier.

[mmcirvin.livejournal.com]

...according to Wikipedia, H. J. Muller had most of the theory in 1922.

[neowolf2.livejournal.com]

Yeah, I had looked up the wikipedia page on "What Is Life?" before I made that earlier post, but it didn't say anything about whether Schoedinger had that idea. He had the idea that genetic information was stored in molecular structures ("aperiodic crystals", had he not heard of polymers?), but I don't know if he took the next step.

[mmcirvin.livejournal.com]

...the Avery-MacLeod-McCarty experiment in 1944 identified DNA as the carrier of genetic information in bacteria... but I don't know if anyone had an inkling of just how that information was expressed in the phenotype:

http://en.wikipedia.org/wiki/Avery-MacLeod-McCarty_experiment

[mmcirvin.livejournal.com]

Von Neumann's work on self-replicators in cellular automata was sometime in the late 1940s, though I think much of it wasn't published until the Sixties.

http://en.wikipedia.org/wiki/Von_Neumann_universal_constructor

I assume that von Neumann was thinking of the thing as a rough analogue of a living replicator. He designed it as a universal computer using a long "tape" as its program, and it'd copy the tape and run the rest of the program to build another one of itself.

More recently, people have played around with actual implementations of von Neumann's replicator, and also with various kinds of constructor in Conway Life. One of the more interesting realizations in recent years is that, for at least the lion's share of the process, you maybe can get away with much much less than a Turing machine: a simple tape-driven mechanism more like a Jacquard loom or a player piano is enough. It's sort of come full circle. I discussed a not-quite-replicator in Life here (not quite a replicator because it doesn't copy the "DNA"):

http://mmcirvin.livejournal.com/458266.html

[del-c.livejournal.com]

I saw a stunning documentary last year on the Antikythera mechanism (probably "The Two-Thousand Year Old Clock") that, as well as impressing me with its sophistication as an astronomical computer, convinced me that it was not a one-off, but one of a population of such instruments, and that this population led, via a thin but real line of development through the Muslim world, to mediaeval European clocks.

Another good documentary from last year was Professor Simon Schaffer's "Mechanical Marvels: Clockwork Dreams". This started with mediaeval clocks, and led, via mechanical toys, to Jacquard looms. Put those two together with the Jacquard/Von Neumann link, and you have a much more interesting line of continuous technological inheritance than that silly thing about horse-drawn carts and the Space Shuttle.

[beamjockey.livejournal.com]

convinced me that it was not a one-off, but one of a population of such instruments,

So the experts say, which sounds convincing-- it's a mature design, but its ancestors (and other contemporary gadgets which may have existed) are lost. Best guess: old bronze things get melted down, especially if nobody understands them any more.

...and that this population led, via a thin but real line of development through the Muslim world, to mediaeval European clocks.

You're saying that the Antikythera Device had descendants? That its technology was not fully lost? First I've heard of this; would like to know more.

This started with mediaeval clocks, and led, via mechanical toys, to Jacquard looms. Put those two together with the Jacquard/Von Neumann link, and you have a much more interesting line of continuous technological inheritance than that silly thing about horse-drawn carts and the Space Shuttle.

I have a Girl Genius talk that covers 18th-century performing automata. This could plug into that, with a little work. After all, Vaucanson invented both the fabulous Mechanical Duck and a drum-driven programmable loom that inspired Jacquard.

[del-c.livejournal.com]

Here's a naughty copy of the BBC documentary: the speculation starts at 53:30, and at 54:30 someone is demonstrating what he says is the world's oldest geared device after the Antikythera mechanism, found in Lebanon.

[beamjockey.livejournal.com]

Found it! Mechanical Marvels: Clockwork Dreams.

[jarodrussell.livejournal.com]

I'm sorry I don't have anything really relevant to add as a comment to your topic, but thank you posting a link to George Dyson's book. I suspect I know what I'll be buying soon.

[maiac.livejournal.com]

Greenfield Village has a Jacquard loom in its weaving workshop. I don't remember ever seeing it in use, but they display textiles that have been made on the loom.

I had a conversation with one of the docents once about the relationship between Jacquard looms and computers.

[beamjockey.livejournal.com]

I have heard that they demonstrate it something like once a year.

I would be interested in seeing this, but would have to plan my life around Greenfield Village's schedule and then get myself to Dearborn on the appointed day... so I don't know whether it will ever happen.

[maiac.livejournal.com]

I'll need to find out on what day they'll be running it, and make sure I'm there. I'll take my camera. It's in an awkward position to get good photos, but I'll give it a try.