The Vision Thing

They wanted to understand the mind. To do that, Carver Mead and his associates, ensconced at Cal Tech, decided that they had to understand the brain. That search for knowledge led to madness and suicide, as one researcher delved too deep into pharmaceutically assisted research, and to unexpected discoveries. In the end, understanding the brain turned out to require understanding vision.

George Gilder's "The Silicon Eye" (Atlas Books, 318 pages, $22.95) traces the history of Foveon, the ground-breaking digital imaging company that grew out of the Mead team's efforts, and finds it to be nearly as tangled and interconnected as the axons and dendrites of the brain itself. The complexity accounts, in part, for the story arc of the book or, perhaps more accurately, for the lack of a story arc, for this is no straightforward tale of innovation. Mead's group started out in the 1980s with neural networks, took a detour into check-scanning machines, made a good deal of money with touch-pad technology and wound up, around the turn of the millennium, producing the Foveon imaging chip, which itself may wind up in a different class of devices than its inventors imagined.

Foveon is itself a story in search of an arc. The chip — which can image all three colors in a single pixel instead of relying on the Rube Goldberg array of filters and post-processing techniques employed by conventional digital cameras — is far more elegant than the technology it's slated to replace. In a conventional story, this superiority would translate into commercial success. But Foveon hasn't managed to find it yet, and the application of its technology to cameras has turned out to have a lot of rough patches. The first consumer-grade camera employing Foveon technology was recalled last month for poor image quality.

The muddled nature of Foveon's story, in fact, led me to wonder why Mr. Gilder chose to build a book around the company. I'm a digital photography fan (I own five digital still cameras, ranging from a superb but balky Nikon D70 to a rugged but mediocre Olympus 520, and two digital video cameras), and even to me Foveon seems an odd choice. The technology is great in theory but nowhere near as revolutionary as the book's subtitle proclaims: "How a Silicon Valley Company Aims to Make All Current Computers, Cameras, and Cell Phones Obsolete."