Re: Rand implementation question

[mpender]

Actually, it looks like randi may suffer from the same problem.  I found a
link for what appears to be the */randi/* code for Octave:
http://addi.googlecode.com/svn-history/r190/Addi/assets/m/general/randi.m

The line which computes the random integers is apparently:
ri = imin + floor ( (imax-imin+1)*rand (varargin{:}) );

Since it relies on */rand/* to generate the random values, then if */rand/*
occasionally takes on the value 1.0:
ri = imin + floor ((imax - imin + 1) * 1.0)
ri = imin + imax - imin + 1
ri = imax + 1

Of course, it is possible that the code has been changed since this was
posted.  The copyright date is from 2012 and it was not found on the Octave
Forge web site.



--
View this message in context: 
http://octave.1599824.n4.nabble.com/Re-Rand-implementation-question-tp4662472p4662705.html
Sent from the Octave - General mailing list archive at Nabble.com.