On 09/11/03 07:49 AM, [EMAIL PROTECTED] wrote: > Is this s Perl random function? If so, I would complain to those that > ported it. > > In any case, you could write your own randomizer as: > > S' = A * S mod P > R = S' / N > > Where: > S = seed, > A = const 16807 > P = const 2**31 - 1, > N = scale, random number R returned between 0 : N-1 > > > I have been using this for decades in various forms on a wide variety > of platforms. Results are the same for any given platform... > > At 11:04 +0200 09/11/2003, Louis Pouzin wrote: >> Hi, >> >> In a script intended for mac, unix and windows platforms, I use the >> "rand" function. It is essential that it produces the same sequence >> on all three platforms. This is the case with mac and unix. On >> windows, it's a completely different set.
Likely this is neither a Perl issue nor a M$oft issue. The same problem arises in C implementations. A common reason for getting different results has to do with whether or not the microprocessor is BIG-endian or little-endian. >> Is there another random library routine designed to be independent >> from Micros--t shenanigans ? What you need is a *portable* RNG. See Perl code in my other email. --njg ___________________________________________________________________________ PERFORMANCE DYNAMICS COMPANY(sm) http://www.perfdynamics.com/ ___________________________________________________________________________ Consulting Services Educational Services 4061 East Castro Valley Blvd. P.O. Box 1238, Magalia Suite 110, Castro Valley California 95954, USA California 94552, USA Bookings and registrations FON: +1-510-537-5758 FON: +1-530-873-0575 FAX: Dial FON (Handshakes automagically) FAX: +1-530-873-6697 NET: [EMAIL PROTECTED] NET: [EMAIL PROTECTED] ___________________________________________________________________________
