Bill -
None of these programs are truly random. They can't be because they
are defined reproducible processes and no defined reproducible process can
be random. For this reason they are referred to in the literature as psuedo
random. What SEED did (and still does) was look at the system clock and
take the rightmost n bits and use them as the starting number for the
"random" numbers generated. n depended on the computer make, system clock,
and whatever programming language SEED was written in. The number returned
by SEED was then passed to the random number generator, which was usually
written in FORTRAN, along with several other arguments which specified the
range and statistical distribution desired.
A great deal of work went into making such generators as "random" as
possible and I did a lot of statistical process modeling using them in the
early '70s.
Bernie
___________________________
At 12:39 PM 4/23/2001 -0500, you wrote:
>On Mon, 23 Apr 2001 13:16:24 -0400, Jim Limburg wrote:
>
>>Ben is right. Haven't tested it lately but when I first started learning
>>VB the random function it used did the same thing..
>
>When I first started learning BASIC in the late 60s -- way before Bill G co-
>opted it, made it Visual, and thumbed his nose at the rest of the
>members of the ANSI standards committee on BASIC -- I remember John
>G. Kemeny teaching us that RANDOM always generated the same
>sequence of random numbers, which made testing a lot more
>manageable. Then there was another command, maybe SEED, that
>turned on true randomization. That BASIC was written in FORTRAN.
>
>Bill
>
>
>
>
>
