random generate small numbers using Qc3GenPRNs

[<jfranz@xxxxxxxxxxxx>]

I'm using IBM's api Qc3GenPRNs to generate 3 digit random numbers (repeats are 
ok). 
Using IBM's example it functions well, and you can pass it the "max" number so 
can be used for any (realistic) # digits. 
However, I find in small #digits like 3 (max 999) it consistently returns higher 
numbers (over 500) than lower numbers. This causes a problem - regulatory 
requirement to randomize a group of records & select all that are <= xxx (varies 
from state to state). The last statement of the pgm takes the huge # generated 
(** 64) and divides by the max (999) - i see that as the problem, but my math 
brain is limited. I've resolved the issue (i think) by using the last xxx digits 
(truncated) of the huge#, and not dividing. does this make sense? I need to be 
able to say this is "standard" solution. 
Jim Franz