>    Hi, Jinx!

Hi Mike!

>    You should prove that rejecting some values, you don't
> hurt randomness.
>    Why do you reject them at all?

Take the example of a table that has 1000 entries. I'd be looking
for a number between 000 and 3e7. Therefore the first random byte
(address high) is  0,1,2 or 3 and the second random number (address
low) is 00-FF if address high = 0,1, or 2 and 00-e7 if address high =
03. The third random number is the bit position, 00 to 0d

Rejection is done simply on the basis of the number being too big,
but every number accepted is random

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.