Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!agate!eris.berkeley.edu!mwm
From: mwm@eris.berkeley.edu (Mike (I'll think of something yet) Meyer)
Newsgroups: comp.sys.amiga.tech
Subject: Re: ARexx psuedo-random number generator
Keywords: ARexx Random Number Generator
Message-ID: <1989Sep13.032352.10321@agate.berkeley.edu>
Date: 13 Sep 89 03:23:52 GMT
References: <19504@unix.cis.pitt.edu>
Sender: usenet@agate.berkeley.edu (USENET Administrator;;;;ZU44)
Reply-To: mwm@eris.berkeley.edu (Mike (I'll think of something yet) Meyer)
Organization: Missionaria Phonibalonica
Lines: 59

In article <19504@unix.cis.pitt.edu> ejkst@unix.cis.pitt.edu (Eric J. Kennedy) writes:
<I was having a little trouble with ARexx's random number generator, so I
<performed the following small test.  The 1 to 45 range is the range in my
<application, so I kept it for the test.  I'm using the time in seconds as the
<seed, which strikes me as being the most variable easily obtainable number.

Your use of the random number generator is incorrect. You should only
initialize it once. Typical output from a version that doesn't do that
(with performance tweaks, and output changed to self-format and be
runnable from inside of mg) is shown below.

As a general rule, serious applications should never use the random
number generator provided with a language. You should either use one
from a "known-good" library (i.e. - IMSL, if there's a random number
generator in it), or roll your own. Rolling your own can be done in an
afternoon, including chasing down and reading the references.  Knuth's
"The Art of Computer Programming" vol.  2 - "Seminumerical Algorithms"
- is the canonical reference, and has lots of theory plus various
algorithms for testing random-number generators. R. G. Dromey's "How
to Solve It By Computer" (part of the Prentice-Hall International
Series in CS, edited by C.A.R. Hoare) extracts the important part for
the do-it-yourselfer into a few pages that describe the best algorithm
from Knuth, what the parameters are, and how to choose them.

	<mike

hist.=0                         /* initialize array */
call random 1, 45, time('S')    /* initialize random number generator */

do index=1 to 10000
        r=random(1,45)
        hist.r = hist.r + 1
        end

out = ""
do index=1 to 45
        out = out || right(index, 2) right(hist.index, 5)
        if index // 4 = 0 then out = out || '0a'x
        else out = out || '09'x
        end
return out

 1   222         2   220         3   233         4   235
 5   207         6   193         7   207         8   231
 9   211        10   210        11   241        12   209
13   248        14   238        15   237        16   224
17   229        18   207        19   251        20   210
21   214        22   210        23   207        24   192
25   239        26   227        27   230        28   230
29   214        30   218        31   234        32   223
33   238        34   234        35   205        36   252
37   222        38   221        39   200        40   196
41   210        42   229        43   204        44   257
45   231
--
Il brilgue: les toves lubricilleux			Mike Meyer
Se gyrent en vrillant dans le guave,			mwm@berkeley.edu
Enmimes sont les gougebosqueux,				ucbvax!mwm
Et le momerade horsgrave.				mwm@ucbjade.BITNET