November 20, 2004

Jacquard's Web, James Essinger

Bruce Sterling divides technology into string and glue, doesn't he? String is older. Information technology was first a child of string (some children were lost).

Most people who care already know that the Jacquard loom used punch cards to store the amazingly complex patterns of fashionable brocade and damask cloth; Essinger can only stretch that out into a few chapters, mostly related to the economic importance of cloth, because not a lot is known about Jacquard's life and what is mostly comes before his inventing. The rest of the book is about the currently-more-interesting descent of computing, of which Jacquard's looms were parents twice: once to part of Babbage's machines, d.s.p., again to telegraphic paper tape, most fruitfully to Hollerith's punchcards. Hollerith was related to a weaver/industrialist who used Jacquard looms.

Quite a lot of this book is devoted to the history of I.B.M., descended mostly from Hollerith's company but also some others, including one that made cheese-slicers... Remington Rand is a more suitable example, since it was a powerful maker of typewriters partly from its experience in making sewing-machines, which required speed, precision and enough reliability to run without a dedicated maintenance team. It's not actually very surprising that clothing should have several times impelled a technological leap. It's a tempting use of capital, since so much money is spent on it at all times, and the standards are high; really good handwork is still better at some things than what machines can do. If the next industrial revolution is 'mass customization', it will probably start in clothes again. I hope so, because it's an offense to aesthetics that so many people have so many clothes that don't really fit.

I wonder if Hollerith's cards were the first time data lived naturally in a database. It's an error-inviting pain to fill out even a two-dimensional table by hand, let alone a deeper one.

It's a pity that information never went from cloth back into the machines; I imagine it as long-term storage: punch the cards until the design is right, weave a reference sample, and when the cards wear out unweave the sample through a machine that generates cards from cloth: then save the first weaving from those cards as the new reference. This wouldn't be a good idea, since punching more cards from cards is easy.

I also wonder where Jacquard got the idea for punched cards. (Or where one Falcon, who built the first but worse punchcard loom got it, if Jacquard got the idea from Falcon.) I have a Theory, actually; bobbin-lace patterns. Bobbin-lace was as expensive and slow to make as brocade, and the patterns changed with fashion much faster than one person could make up a suit of lace. Complicated patterns require pinholes punched into stiff card, which give a skilled lacemaker enough direction to make up the pattern. Middlemen made up many many cards corresponding to small pieces of a fashionable pattern and handed them out to lacemakers as they picked up the finished pieces from the last pattern. It would have been important that the patterns lined up well to be invisibly sewn together, although the threads did not weave from one piece into the next. I think there must have been a lot of these cards around, especially in a town as devoted to luxury clothing as Lyons was. It's still a big intellectual jump to switch from a human feeling with a pin to know where thread-crossing should go, to a machine that always crosses in the same places feeling a card to decide whether a crossing should happen; but it would explain why sheaves of punched cards 'looked like' information storage.

Online glossaries give 'lace cards' as a synonym for punchcards, but they also sometimes suggest that that only refers to a card with all possible holes punched out, giving it a resemblance to simple lace. On the other hand, that resemblance would provide an easy false etymology.

I can't find an online picture of how the early automatic lacemaking machines work, although Nottingham has a promising history of mostly-Nottingham lace machine inventions; the Jacquard idea came in after decades of improving knitting-frames to approximate the action of lace-twisting.

ISBN: 0192805770

So wrote clew in Clothing. , History (18th c.). , History (19th c.). , History (20th c.). , Technology. | TrackBack
And thus wrote others:
TrackBacks turned off...