[2892] in cryptography@c2.net mail archive
Re: Search for "better" F box for Skipjack
daemon@ATHENA.MIT.EDU (Marcus Leech)
Tue Jun 30 11:35:13 1998
Date: Tue, 30 Jun 1998 06:25:11 +0200
From: "Marcus Leech" <Marcus.Leech.mleech@nt.com>
To: Matt Thomlinson <mattt@microsoft.com>
CC: cryptography@c2.net
Matt Thomlinson wrote:
>
> marcus,
>
> if you've got source I've got a fast dual-alpha that I could throw at this..
>
I've got source, but it's unlikely I can give it to you,
unfortunately. Getting even "recreational" source code
out of this place is a nightmare...
Carlisle Adams developed a really nice S-box generator
system for CAST that may be better at this task than
mine. I don't know if he'd be able to share it with
the world or not. I'll ask him privately.
My generator is very simple minded: it generates a random
S-box of the appropriate size, then measures the boxes
properties. I used it to generate the S-boxes in my
own algorithm, CRISP. It took about a week to generate
all the S-boxes, but produced some good S-boxes when
it was done (maximum pairs-XOR 30/2048 and minimum
hamming distance of 0.4512).
