Path: utzoo!mnetor!uunet!seismo!sundc!pitstop!sun!decwrl!decvax!ucbvax!pasteur!ames!nrl-cmf!cmcl2!brl-adm!brl-smoke!gwyn
From: gwyn@brl-smoke.ARPA (Doug Gwyn )
Newsgroups: comp.lang.c
Subject: Re: Random Numbers ...
Message-ID: <7363@brl-smoke.ARPA>
Date: 28 Feb 88 00:55:19 GMT
References: <7097@sol.ARPA> <225800009@uxe.cso.uiuc.edu>
Reply-To: gwyn@brl.arpa (Doug Gwyn (VLD/VMB) <gwyn>)
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 16

In article <225800009@uxe.cso.uiuc.edu> mcdonald@uxe.cso.uiuc.edu writes:
>Has anyone out there considered how to write a random number generator in
>which the output is truly random?

Yes, you need a truly random mechanism, such as a radiation counter or
resistive thermal noise (as you say you once did).  Deterministic
algorithms will of course always have intrinsic regularities.

>In my tests it seems that the low-order bits have a very short repetition
>period.  I've considered using two ordinary generators giving 32-bit periods
>and patching together the high order words. Is this better?

It depends on how you're using the pseudo-random sequence.  I suggest
you read Chapter 3 of Donald Knuth's "The Art of Computer Programming"
(in Volume 2, "Seminumerical Algorithms", 2nd ed., Addison-Wesley 1981,
ISBN 0-201-03822-6).