Wednesday, 5 April 2017

digital foundry or prototype

In the context of debates on the privacy and over-reach of contemporary facial-recognition technologies, Tedium takes a look back to when teaching machines to read—that is recognise text that was also contoured for human eyes (as opposed to bar-codes)—was the hurdle to vaunt. Aptly, it was the offices of Reader’s Digest that first employed optical character recognition (OCR) in 1954—not for skimming the abridged versions for the busy reader—but rather to manager their subscribers’ data.
The prototype was invented two years earlier in the attic workshop of a tinkerer who called it “Gismo.” As the idea spread, a need for industry standards arose and type-face designers had to make a departure from the quiver of fonts that printers used, sourced from medieval scripts, and designed a character set that could be easily interpreted by all eyes and avoided ambiguity. As the visual acuity of machines has greatly increased (although there are still notable hang-ups) along with reading comprehension, we are not restricted to a certain font family and can be expressive with materials fit for consumption by electronic or biological minds, but for legacy hardware, such typefaces are kept around—for instance on the raised digits on a credit card. The inventor, David Hammond Shepard, also designed the latter numeric font called Farrington B and devised the automated telephone answering tree, interactive voice response technology.