From owner-freebsd-hackers  Thu Oct  3 14:04:38 1996
Return-Path: owner-hackers
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.5/8.7.3) id OAA23422
          for hackers-outgoing; Thu, 3 Oct 1996 14:04:38 -0700 (PDT)
Received: from paris.CS.Berkeley.EDU (paris.CS.Berkeley.EDU [128.32.34.47])
          by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id OAA23410
          for <freebsd-hackers@freebsd.org>; Thu, 3 Oct 1996 14:04:36 -0700 (PDT)
Received: from paris.CS.Berkeley.EDU (localhost.Berkeley.EDU [127.0.0.1]) by paris.CS.Berkeley.EDU (8.6.11/8.6.9) with ESMTP id OAA07079; Thu, 3 Oct 1996 14:00:34 -0700
From: Josh MacDonald <jmacd@CS.Berkeley.EDU>
Message-Id: <199610032100.OAA07079@paris.CS.Berkeley.EDU>
To: Bruce Evans <bde@zeta.org.au>
Cc: freebsd-hackers@freebsd.org
Subject: Re: rand() and random() 
In-reply-to: Your message of "Fri, 04 Oct 1996 01:38:16 +1000."
             <199610031538.BAA19908@godzilla.zeta.org.au> 
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <7072.844376431.1@paris.CS.Berkeley.EDU>
Date: Thu, 03 Oct 1996 14:00:32 -0700
Sender: owner-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

> >> Because it is ANSI conrfmant (man rand).
> >
> >That doesn't mean it would be impossible to do
> >
> >int
> >rand(void)
> >{
> >	return (int)random();
> >}
> 
> It would have to use a copy of random() so that the functions don't
> mess up each other's state.
> 
> >The wrapping is needed since one function is declared to return int,
> >while the other returns long.  Other than this, both have the same
> >range for the result (RAND_MAX actually is (2**32)-1), so i fail to
> >see why random() should _not_ match the ANSI requirements.
> 
> The example implementation of rand() in ANSI is said to be bad.  The
> BSD implementation is much worse because it only throws away one high
> bit instead of 17.  We had a discussion a year or two ago about
> random() being bad too because srandom() uses essentially rand() to
> initialize things.
> 
> >I vote for killing the old rand().  Too many things (in particular,
> >games) behave very stupid with it.
> 
> Are they better with rand() replaced by (rand() % 0x7fff)?

The reason I originally sent this mail is that I had been showing the
Gimp authors something with their new version dated yesterday, I clicked
on the "about dialog", which has a randomly disolved-in image.  Well, the
image materializes not randomly, but like vertical blinds gradually closing,
because the rand() function is so terrible.  Also a few of their plug-ins
use rand() (because rand() is available everywhere, while random() isn't),
and display extremeley terrible effects, (for example, layer-dissolve, 
which is supposed to be completely random, makes a bunch of stripes through
an image).

Having a bad rand() function is killing the "portability" of these 
applications.  This gave them fuel in their ongoing battle to convince
me to start running Linux.  


-josh