Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!mit-eddie!uw-beaver!uw-june!entropy!martin
From: martin@entropy.UUCP
Newsgroups: sci.math,sci.physics,sci.crypt,sci.math.stat
Subject: Re: Do you use RANDOM NUMBERS?
Message-ID: <560@entropy.ms.washington.edu>
Date: Mon, 6-Apr-87 01:02:45 EST
Article-I.D.: entropy.560
Posted: Mon Apr  6 01:02:45 1987
Date-Received: Wed, 8-Apr-87 04:52:45 EST
References: <5712@reed.UUCP> <464@batcomputer.tn.cornell.edu> <444@faline.UUCP>
Organization: UW MathStat, Seattle
Lines: 27
Keywords: Random numbers, leprechauns
Xref: utgpu sci.math:852 sci.physics:1050 sci.crypt:289 sci.math.stat:94
Summary: DES for random numbers

In article <444@faline.UUCP>, karn@faline.UUCP writes:
> My own feeling is that if you MUST use a deterministic algorithm to
> generate random numbers, then use the Data Encryption Standard. Pick a
> random number for use as a key (use a pair of dice as I previously
> described) and encrypt an incrementing 64-bit integer.  From everything
> I've read, most of the existing "simple" methods for generating random
> numbers have some horrible properties.
> 
> The excellent book Numerical Recipies also recommends DES as a random
> number generator, and gives listings in Fortran and Pascal. The main
> problem, of course, is that software DES is slow but  .........

This is an understatement. DES algorithms are VERY slow. The Num.
Recipies is a prime example. I expect that Phil has speeded it up
to the point where it might be marginally usable but that
does not mean that it is a good choice. I don't think that DES
has been tested very well as a random number generator. Nor did
Num. Recipies recomend it as a uniform random number generator,
it was proposed as a method of generating binary sequences. I
will agree that the distinction between binary sequences and
random binanary numbers is thin. For most applications, I would
recomend the shuffle algorithm in Numerical Recipies rather
than DES. If I wanted to check my application with another
generator I would add a shift regester ( modulo add or XOR )
to the shuffle.

Don Martin      martin@entropy