: New: Book Report: The Silicon Eye

You think that you understand something, but then you figure out that you don't understand it after all. You've built up this model in your head, then you see something that doesn't fit the model. Today I saw a line of people waiting outside an Apple store in the rain. And I realized that my mental map of Apple fanboys was off—I'd forgotten the level of devotion. I'd forgotten that these people would wait in the rain. My model was off.

My point? I'm getting to that.

This book is mostly about the development of the Foveon X3 sensor. There's a neat idea behind it: different wavelengths of light can penetrate silicon to different depths. So use a silicon chip as a a camera's "eye" with detector circuitry embedded at different depths in the chip, exposing the chip's "top" to whatever light you want your camera to sense.

I guess that's how it works. I don't trust this book's accuracy. Partway through, it talks about the history of the Z80.

Faggin led the design of the Z-80, a new microprocessor, which was decisively superior to the Intel processor and would enjoy a longer life span embedded in low-rent electronics products. However, the central role of the microprocessor in a computer architecture is to define its "instruction set," the portfolio of things that it can directly do, such as add, subtract, or fetch from memory. [so far, so good]

Directly addressed by every computer's operating system (OS)—such as Windows or Linux [uhm, what? the instructions are used by all programs, not just the OS]—these instructions vary subtly between different microprocessors, even very similar ones like the Intel 8088 and the Z-80. The subtle variations among instruction sets pass on as a kind of DNA into the software operating systems that run the basic operations of a computer [wtf?], and then from the OS into all the applications such as Word and PhotoShop that use the OS.

[The Z80's designer] Faggin set the Z-80 to execute Design Research DOS instructions, thus obviating the need to pay Microsoft fees.

Oookay, so the book's author confused the role of the microcontroller's instruction set with some of the other "instruction sets" running around inside a computer. And if you already know something about microcontrollers, you can guess at what he's trying to say. But suppose you didn't know anything about microcontrollers. You might trust this book and think that's how a microcontroller really worked.

So... for the one part of this book where I already knew the facts, this book got the facts wrong. That's a small sample space, but it's all I have to work with. And it tells me not to trust this book.

The Foveon X3 process sounds neat, and I bet there's an interesting story behind it. But I don't trust this book to tell me that story. And I should remember not to buy any more books by George Gilder, lest my brain picks up another model that doesn't fit facts.

Tags: book unfinished

blog comments powered by Disqus