Re: Random Number Generator

[Douglas Handy <dhandy1@xxxxxxxxxxxxx>]

Rob,

>1)  What is z.digit(6)?  

Looks to me like he is just extracting the 6th, 4th, 5th, and 7th digit of the
ten digit result, ie the middle digits, in the above order.  So the actual
minimum and maximum values seem irrelevant as the pin is derived from
x,xx2,314,xxx where 1-4 are the location of the pin digits.

I have no idea of the distribution of the resulting PINs across all 86,400
possible timecode "seeds".  But this sounds like the type of application which
may not need statistically accurate randomness -- I think it depends on how many
PINs she is assigning per program invocation.

If Debbie is assigning PINs to accounts en masse in batch, then I'd want the
millisecond portion of the timestamp or RRN or something used in the algorithm
too so that accounts processed sequentially in the same second did not have the
same PIN.  And I'd care more about testing the real distribution of the PINs.

If it is for ad hoc interactive assignment, then it may well be "random" enough.
If you only did a few asignments per program invocation, how you choose the seed
seems to me more meaningful than the distribution when starting at a given seed.
And this algorithm seems at least as good as some methods of choosing a seed for
a "true" random number generator.

But lest you think Leif doesn't understand random numbers, see Chapter 22 of his
book.  The chapter title is  "How to Generate a Truly Random Number".

Doug

+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---