The Fallacy of the Almost-General-Purpose Computer

I was at a conference in Washington, DC on Friday and Saturday. Participants included some people who are reasonably plugged in to the Washington political process. I was stunned to hear one of these folks sum up the Washington conventional wisdom like this:

“The political dialog today is that the general purpose computer is a threat, not only to copyright but to our entire future.”

(It’s worth noting that he was repeating the views of others rather than offering his own opinion – and that he had a general-purpose computer open on the table in front of him as he said this!)

If I could take just one concept from computer science and magically implant it into the heads of everybody in Washington – I mean really implant it, so that they understood the idea and its importance in the same way that computer scientists do – it would be the role of the general-purpose computer. I would want them to understand, most of all, why there is no such thing as an almost-general-purpose computer.

If you’re designing a computer, you have two choices. Either you make a general-purpose computer that can do everything that every other computer can do; or you make a special-purpose device that can do only an infinitesimally small fraction of all the interesting computations one might want to do. There’s no in-between.

I can tell you that this is true. And I can assure you that every well-educated computer scientist knows why it is true. But what I don’t know how to do – at least not yet – is to give a simple, non-technical explanation for it. If anybody has a hint about how to do this, please, please let me know.