From owner-freebsd-current  Sat Mar 30 19:26:34 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id TAA09402
          for current-outgoing; Sat, 30 Mar 1996 19:26:34 -0800 (PST)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id TAA09384
          for <current@FreeBSD.org>; Sat, 30 Mar 1996 19:26:28 -0800 (PST)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id NAA08023; Sun, 31 Mar 1996 13:20:26 +1000
Date: Sun, 31 Mar 1996 13:20:26 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199603310320.NAA08023@godzilla.zeta.org.au>
To: ache@astral.msk.su, bde@zeta.org.au
Subject: Re: random .. not so ..
Cc: current@FreeBSD.org, davidg@Root.COM, imb@scgt.oz.au
Sender: owner-current@FreeBSD.org
X-Loop: FreeBSD.org
Precedence: bulk

>> >> >I plan to aply proposed fix, if nobody against.
>> 
>> It needs more thought.

>What thought exactly do you mean?

Think about rereading Knuth :-).

>Calling srandom(time() f.e.) is common case. Without this fix two
>programs calling srandom in _different_ times produces very predictable
>almost same sequences.

This must be a poor way to initialize random().  First, time() isn't
random.  Second, srandom() isn't claimed to give a state that varies
randomly with its arg.  In fact, it doesn't.  Third, and more
fundamentally, only 2**32 of random()'s (more than) 2**69 states are
reachable using srandom().

Bruce