From: "A. Sinan Unur" <asu1 AT cornell DOT edu>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: One method for small random nuimbers
Date: Mon, 09 Jun 1997 07:54:57 -0400
Organization: Cornell University http://www.cornell.edu
Lines: 45
Sender: asu1 AT cornell DOT edu (Verified)
Message-ID: <339BEF11.BD5@cornell.edu>
References: <2 DOT 2 DOT 32 DOT 19970604162434 DOT 0069d50c AT gate> <3395B097 DOT 5077 AT cornell DOT edu>
	 <339830A6 DOT 672 AT blackmagic DOT tait DOT co DOT nz> <3397CBCF DOT 6536 AT pobox DOT oleane DOT com> <339C297F DOT 23A4 AT blackmagic DOT tait DOT co DOT nz>
Reply-To: asu1 AT cornell DOT edu
NNTP-Posting-Host: cu-dialup-0001.cit.cornell.edu
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Precedence: bulk

Bill Currie wrote:
> 
> Francois Charton wrote:
> >
> > Bill Currie wrote:
> > >
> > > Basicaly, I treat the number received from the main RNG as a
> > > bit stream and pull off the bits as needed.
> > >
> > I read about such a method being a bad thing in several books 
> > (Numerical Recipes warns a lot against this, and I remember reading 
> > something about it in Knuth, I'll dig it out tonight).
> 
> If you can send me a quote, that would be apreciated.
> 

i realize i am not the person to whom this request was directed, but in
light of the recent discussions, i think a few quotes can be useful.
(you can download chapters from 'merical recipes' at www.std.com/~nr/)

Numerical Recipes in C, 2nd ed. p.277:

(Linear congruential) generators have their low order (least
significant) bits much less random than their high order bits. If you
want to generatea random integer between 1 and 10, you should always do
it using high order bits as in:

  j = (int)(10.0*rand()/(RAND_MAX+1.0));

and never by anything resembling

  j = 1 + (rand() % 10);

more related to what bill currie was proposing ("I treat the number
received from the main RNG as a bit stream and pull off the bits as
needed.") is the remark on p.299, but their warning concerns the
opposite of what he wants to do:

A word of caution is: Don't use sequential bits from these routines as
the bits of a large, supposedly random, integer ...

pp.296-300 of their book is devoted to generation of random sequences of
bits.

  -- Sinan