[5168] in cryptography@c2.net mail archive
Re: A Massively Parallel Cryptosystem Based on Cellular
daemon@ATHENA.MIT.EDU (David Honig)
Wed Jul 21 10:21:15 1999
Date: Tue, 20 Jul 1999 18:47:26 -0700
To: Eugene Leitl <eugene.leitl@lrz.uni-muenchen.de>, <cryptography@c2.net>
From: David Honig <honig@sprynet.com>
In-Reply-To: <14228.46177.569389.978161@lrz.de>
At 10:43 AM 7/20/99 -0700, Eugene Leitl wrote:
>
>What is your oppinion on the security of this system. Any obvious
>flaws?
>
>http://www.santafe.edu/~hag/ca11/ca11.html
>
>
>A Massively Parallel Cryptosystem Based on Cellular Automata
>
>Howard Gutowitz ESPCI; Laboratoire d'Electronique 10 rue Vauquelin;
>75005 Paris, France gutowitz@amoco.saclay.cea.fr
If you're interested in security, then give us the C code, or some
other conventional language.
Its 'neat' that you can use a given architecture (CA for crypto;
neural nets for pattern recog; whatever) but one analyzes
an algorithm for competence, vs. machines for performance.
CA and other machine models (e.g., neural nets; systolic cpus; etc)
are perhaps fascinating and efficient ways of implementing
an algorithm, but the meat is the abstract algorithm, which could
be done by hand, or with an abacus.
There are a zillion ways to sort a bunch of numbers; and a zillion machines
that each implement these methods. But all that matters, at a certain
level of abstraction, is what it means to sort.
Cheers,
dh
