Explaining General Purpose vs. Special Purpose Computers
An item from Ed Felten asks how to give a "simple, non-technical explanation" for the truism:
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.
Here's my try at such an explanation, geared to Washington concepts:
Suppose you want telephone calls answered, for an office. You can either hire a human and have that person be a receptionist, or buy an automated telephone answering machine. The human receptionist who has the task of answering telephone calls will also be able to answer letters or do any other clerical task. The automated telephone answering machine will never be able to do anything other than answer telephone calls. There is no in-between, where there's a machine which will do all general clerical work, but nothing else.
Moreover, to continue the analogy, the human receptionist, as a consequence of general-purpose ability, will also be able to tell unauthorized people who has been telephoning the office. And perhaps even what the contents of the telephone calls contain (copying!). An automated telephone answering machine will never be able do that either (on its own).
This is simply two sides of the same coin of having general-purpose ability. Note this problem has been well-known since ancient times - where rulers would maim servants in various ways (e.g. cutting-out the tongues of slaves) in brutal attempts to prevent what might be called nowadays, unauthorized information transfer. Recent legislative proposals are perhaps the modern equivalent of those crippling practices.
By Seth Finkelstein |
posted in copyblight
, infothought
|
on October 14, 2002 06:16 PM
(Infothought permalink)
| Followups
