Re: Pseudo-random code in asm?

Perry E. Metzger (perry@piermont.com)
Mon, 24 Aug 1998 09:57:14 -0400

• Next message: Ben Laurie: "Re: Crypto-sendmail (was Crypto Coding Project)"
• Previous message: Andy Brown: "What benefit would there be in using triple RC4-40?"
• In reply to: Mail System Internal Data: "DON'T DELETE THIS MESSAGE -- FOLDER INTERNAL DATA"

jwashbur@whittman-hart.com writes:
> There is a great source. "The art of computer programming, Volume 2:
> semi-numerical methods", by Donald E. Knuth.
>
> If your Psuedo-Random Number Generator passes the 10 (and all 10) tests
> described in chapter 2 (Random numbers) you have a PRNG with no known
> flaws.

Not really.

Knuth's tests are way insufficient for finding cryptographically
strong PRNGs. They are only good enough for finding ones that are good
for certain kinds of non-cryptographic work.

Perry

• Next message: Ben Laurie: "Re: Crypto-sendmail (was Crypto Coding Project)"
• Previous message: Andy Brown: "What benefit would there be in using triple RC4-40?"
• In reply to: Mail System Internal Data: "DON'T DELETE THIS MESSAGE -- FOLDER INTERNAL DATA"