From: deeb@meceng.coe.neu.edu (Stephen Humble)
Date: Wed, 4 May 94 10:40:20 PDT
To: ecarp@netcom.com
Subject: Re: Lobbying/Politics/etc.
In-Reply-To: <Pine.3.89.9405031055.A14425-0100000@netcom10>
Message-ID: <9405041548.AA04593@meceng.coe.neu.edu>
MIME-Version: 1.0
Content-Type: text/plain


Ed Carp <ecarp@netcom.com> sez:
> Consider a successful terrorist attack against a significant 
> group of innocents (the larger the number killed, the greater the horror 
> and shock value).  The terrorists were using PGP-encrypted email to plan 
> out the thing.
> 
> Now, how long do you think it would take before ALL crypto was outlawed?  
> Who would benefit from such a thing?  Consider that it's child's play to 
> finance, arm, and train a group of people to conduct a terrorist attack 
> and (conveniently) they all get killed in their attack.  No one's going 
> to complain too loudly - after all, they *are* terrorists, right?

I suspect significant problems implementing a law that criminalizes
crypto.  The government currently spends $billions per year trying to
eliminate illegal drugs, to very little effect.  Drugs should be
easier to eliminate than crypto since phys-obs can't be copied ad
infinitum as bits can.

There's also the matter of recognizing crypto in use.  A program that
transforms its input so that the output can be converted back to the
input but has maximum entropy is a good compression program and might
also be an encryption program.  If a TLA taps my phone and finds a
mysterious bit sequence, how can they distinguish reliably and cheaply
between an encrypted conversation and a download of
emacs-19.22.tar.gz?

I don't claim *they* can't try to outlaw crypto, and I certainly don't
claim they can't kill millions in the effort, but I *do* claim that
eliminating crypto is a very hard problem.

Inspired by my recently-arrived "Cypherpunk Criminal" t-shirt,

Stephen