🚀 go-pugleaf

In article <1jk77nINNeh3@darkstar.UCSC.EDU> speth@cats.ucsc.edu (James Gustave) writes:
>
>  I need a source for truly random numbers for an encryption program I'm
>working on.  I was thinking of using PGP's approach and time keyboard input
>with a fast timer, then use the low-order 8-bits as a random number.  I'm
>working on a Mac, so I'd like to use key and mouse events.  How can I tell
>if the numbers I'm getting are good enough to be considered cryptographically
>random?


One way is to do the same sequence of operations over and over again,
trying to be as repeatable as possible, timing it each time.  Then take the
several series of numbers and reduce them statistically to a set of basis
functions.  You should learn from this how many bits of each timed event
are noise and how many are part of the timed sequence itself.  The noise
ones are the only ones you want to use.

It might be, by the way, that there is a regular source of real noise.  You
might therefore take a number of sequences of different actions but the
same length and try to reduce them.  Any common basis found might be noise
from your computer or OS.

Of course, you have to look at the results and interpret them before getting
the answer to your question.  I hope you publish the results here, when you
do get the answer.



-- 
-- <<Disclaimer: All opinions expressed are my own, of course.>>
-- Carl Ellison						cme@sw.stratus.com
-- Stratus Computer Inc.	M3-2-BKW		TEL: (508)460-2783
-- 55 Fairbanks Boulevard ; Marlborough MA 01752-1298	FAX: (508)624-7488