chaoplexity and institutions

John K. Taber jktaber at onramp.net
Thu Aug 13 16:20:38 PDT 1998


Tom Lehman wrote:
>
> Dear Barkley,
>
> Do you recall assertions in the late 1970's that there was no such thing
> as a computer generated random number. Does anyone still hold this
> view?
>
> Sincerely,
> Tom L.
>
[snip]

Yep, but the quote is a little off. There is no algorithm for it, meaning a deterministic way, and in fact, genuinely random numbers are not wanted because any given sequence could not be repeated.

Von Neumann once said something to the effect that anybody thinking of generating random numbers by computer algorithms is in a state of mathematical sin.

Cryptography is one field where genuinely random numbers are useful. There is no deterministic way of getting them that I have heard of, but PGP measures latency between keystrokes. I don't think anybody can really prove that is random, but it seems to be good enough.

Practical implementations to get random numbers are not easy to accomplish. Yes, you can use white noise, but there is always something to interfere with the results, like house power, or plenty of other things to insidiously prevent randomness.



More information about the lbo-talk mailing list