Hi, Paul.

PM> PE> RS>      Ok , i admit - you can't have perfect random numbers

PM> FM> For truly random numbers, try using radioactive decay.

PM>On the topic of random numbers, I seem to remenber that there is no way to
PM>tell if a finite series of numbers is *really* random or not - although you
PM>can calculate a probability that it is truly random. Chi sqared or something

True, and I can't remember the stats test to use either! OTOH, for
*most* applications it doesn't matter - the pseudo-random numbers
produced by a decent generator in most languages/libraries is quite
satisfactory, with the additional advantage that for debugging (and
in encryption applications) it's repeatable!

PM>Basically, unless you are using an infinite numer of random numbers,
PM>srand((int)time(NULL)) will give you random numbers more-or-less
PM>indistinguishable both theoretically and for all practical purposes from
PM>a sequence of truly random (probabilistic) nubers.

Oh, I just said that :-)

PM>PS - instead of radioactive decay, perhaps you could rig up a noisy
PM>transistor.

Yep.

PM>PPS - yes I know that you didn't make the original posting, I hope you don't
PM>mind this being adressed to you.

No worries. I think it was Paul Edwards, I'm sure he'll see this.

Regards, FIM.

 * SLMR 2.1a * I've got 256K of RAM, so why can't I run Windows 3.0?

--- Maximus 2.01wb