From: Ian Farquhar <ifarqhar@laurel.ocs.mq.edu.au>
Date: Wed, 20 Jul 94 20:58:35 PDT
To: hal@mit.edu
Subject: Re: Clipper Chip retreat
In-Reply-To: <9407210232.AA08923@toad.com>
Message-ID: <199407210356.AA27526@laurel.ocs.mq.edu.au>
MIME-Version: 1.0
Content-Type: text/plain


>I'll leave it to someone else to post the entire article, but the gist
>is that Gore sent a letter to Maria Cantwell saying that the
>administration is willing to consider alternatives to Clipper that are
>based upon nonclassified algrithms, [...]

There are three of ways they could proceed from here.  The first is
to declassify SKIPJACK, which would (IMO anyway) be another welcome
boost for civilian crypto, in that it would be the chance to see
and analyse a NSA-designed cryptosystem.  For that very reason,
I doubt that they will do it.  (Small aside from another field: it will be 
_very_ interesting to compare the old SHA to the revised version.  One of
the reasons I suspect that the problem is quite significant is simply
because they have decided to revise it and run the risk of invoking
comparitive research which may disclose design techniques or methods
of attack.  If the later option is the lesser of two evils, I would
guess that it is a nasty hole indeed.)

The second would be to take an existing commercial cipher, and to
sanctify that as their recommended algorithm.  Again, assuming that
they selected a properly secure algorithm, the very features of the
algorithm they chose would be another point of interest.  This is
the least favorable option.

Their final option would be to release another cipher, but with a
reduced keysize or key entropy.  Not the best solution, but one
which I have a nasty suspicion will happen.

						Ian.