From: khijol!erc@apple.com (Ed Carp)
Date: Sat, 21 Aug 93 11:47:34 PDT
To: gnu@toad.com (John Gilmore)
Subject: Re: Cracking & auditing crypto protocols
In-Reply-To: <9308211555.AA14075@toad.com>
Message-ID: <m0oU0Ky-00022PC@khijol>
MIME-Version: 1.0
Content-Type: text


> There is no way that the selected group of people could crack a 
> half-reasonable cryptosystem in a few weeks.  Real Cryptanalysts spend
> months and years working on cracking cryptosystems, and none of the
> panelists was a Real Cryptanalyst.  We had all the details of DES,
> and it took 15 years to make a dent in it.

That's one of the strongest points in favor of crypto in wartime, for example.
The usefulness of a cryptosystem is not just a function of its resistance to
attack, it is also a function of how long it *has* to resist attack.  For
example, if the Nazis had broken a message regarding D-Day, encrypted with
a cipher in such a way that if the message were compromised it wouldn't
compromise the system itself, it wouldn't matter when they broke it, as long
as it was after June 4, 1944.  After that time, it's useless, and many
messages during tactical operations have an effective lifetime of days, if
not hours.
-- 
Ed Carp, N7EKG			erc@apple.com			510/659-9560
                            anon-0001@khijol.uucp
If you want magic, let go of your armor.  Magic is so much stronger than
steel!        -- Richard Bach, "The Bridge Across Forever"