Better solution to Random>>initializa
Juan Cires Martinez
jcm at mat.upm.es
Fri May 28 09:32:42 UTC 1999
----- Forwarded message from jcm -----
>From jcm Thu May 27 16:09:17 1999
Subject: Better solution to Random>>initializa
To: squeak at cs.uiuc.edu (Squeak mailing list)
Date: Thu, 27 May 1999 16:09:17 +0200 (METDST)
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 8bit
Content-Length: 765
The problem with the initialization of a Random is in the statement:
seed := Time millisecondClockValue * self hash.
which has a problem when self hash evaluates to zero. The solution I
propose is to modify that line:
initialize
" Set a reasonable Park-Miller starting seed "
seed := Time millisecondClockValue * (self hash + 1).
a := 16r000041A7 asFloat. " magic constant = 16807 "
m := 16r7FFFFFFF asFloat. " magic constant = 2147483647 "
q := (m quo: a) asFloat.
r := (m \\ a) asFloat.
Greetings, Juan.
PS: I have translated to Squeak the C code to generate random numbers
with a normal distribution published in "Numerical Recipes in C", second
edition, and implemented the class NormalRandom. Is it legal to publish
such code?
----- End of forwarded message from jcm -----
More information about the Squeak-dev
mailing list
|