From owner-freebsd-hackers  Thu Oct  3 09:09:47 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id JAA05459
          for hackers-outgoing; Thu, 3 Oct 1996 09:09:47 -0700 (PDT)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id JAA05441
          for <freebsd-hackers@freebsd.org>; Thu, 3 Oct 1996 09:09:38 -0700 (PDT)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.7.6/8.6.9) id CAA20556; Fri, 4 Oct 1996 02:05:37 +1000
Date: Fri, 4 Oct 1996 02:05:37 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199610031605.CAA20556@godzilla.zeta.org.au>
To: ache@nagual.ru, joerg_wunsch@uriah.heep.sax.de
Subject: Re: rand() and random()
Cc: freebsd-hackers@freebsd.org, jmacd@CS.Berkeley.EDU
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

>random() isn't good enough too. However the problem is different
>with rand() case. The problem is that srandom() not accomodate
>startup value well, i.e. produce the same _few_ sequences for _many_
>different startup values. I remember bug report related to this problem
>(with suggested random() replacement) somewhere in freebsd mailing
>lists long time ago.

Another problem with srandom() is that it only gives 2^32 startup states.
I think you're supposed to use initstate() for serious work.  Don't know
if this gives better startup sequences.

Bruce