Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.csd.uwm.edu!cs.utexas.edu!sun-barr!decwrl!decvax!ima!esegue!johnl
From: johnl@esegue.uucp (John R. Levine)
Newsgroups: comp.sys.ibm.pc
Subject: Re: Stuck on pc random number generation : help
Keywords: beta testing
Message-ID: <1989Aug31.151520.4079@esegue.uucp>
Date: 31 Aug 89 15:15:20 GMT
References: <7561@microsoft.UUCP> <1991@leah.Albany.Edu> <6474@columbia.edu>
Reply-To: johnl@esegue.UUCP (John Levine)
Distribution: usa
Organization: Segue Software, Cambridge MA
Lines: 19

In article <6474@columbia.edu> ari@sylvester.UUCP (Ari Gross) writes:
>Here is some Fortran code for generating random numbers.
>
>subroutine randu(ix, iy, yfl)
>...

The RANDU routine from the old IBM SSP package is notorious as the worst
alleged random number generator ever written.  See the October 1988 CACM for
an article describing the pitfalls of pseudo-random number generation and an
algorithm known to be robust.  If you want normally distributed
pseudo-randoms, see Knuth's Seminumerical Algorithms for some ways to do it,
since the SSP GAUSS routine is no better than RANDU.

Most of the SSP routines were pretty good, but they really blew it in the
random number department.
-- 
John R. Levine, Segue Software, POB 349, Cambridge MA 02238, +1 617 492 3869
{ima|lotus}!esegue!johnl, johnl@ima.isc.com, Levine@YALE.something
Massachusetts has 64 licensed drivers who are over 100 years old.  -The Globe