>>Is anyone aware of reliability checks of FoxPro's RAND() function to determine whether it adequately
>>simulates random numbers on the interval (0,1)? Thanks.
>
>I could be wrong but from my college days I believe that there was no way to generate truly random numbers with just software. I remember reading a couple of years ago that Intel was going to provide some functionality in their processors to read the tempature variations of the processor or something like that to generate truly random numbers. Basically you need some physical process to generate truly random numbers. There is acutally a book of just random numbers that were generated by reading atomic decay of some material or something like that.
>

If you want to roll yer own, there's a lot of discussion of pseudorandom numbers in Knuth's "SemiNumerical Algorithms", but it is not Fox code (in fact, it's all built around a theoretical assembly language called MIX). I don't know of any Fox-specific books, but general references on cryptology and number theory are likely sources of algorithms that can be used to generate pseudorandom sequences.