From: Adam Back <aba@dcs.ex.ac.uk>
Date: Mon, 15 Apr 1996 10:17:53 +0800
To: jeremey@forequest.com
Subject: Re: Is crypt(1) a prohibited export?
In-Reply-To: <Pine.BSI.3.91.960414014821.7867A-100000@newton.forequest.com>
Message-ID: <199604142234.XAA00563@adam.test.net>
MIME-Version: 1.0
Content-Type: text/plain



Jeremey Barrett <jeremey@forequest.com> writes on cpunks:
> [...] But since a one-way hash function (implemented as a one way
> function) can't really be used for encrypted communication, 

One way hashes can be used to produce an block encryption system
running in CFB mode, eg: Peter Gutmann's MDC.

(CFB mode is where you encrypt an IV (just a random salt) with the
key, and XOR the result with the data (you chaing the encrypting for
subsequent blocks) -- to decrypt you just repeat the process, use the
same IV and encrypt again.  As you never actually use the block
ciphers decrypt function, you can therefore (and this is what MDC
does) use a one way hash in the place of a block encryption algorithm
in CFB mode.)

> I don't think it is subject to export restriction.

They don't seem to be subject to export restrictions, but then perhaps
that is paradoxical, perhaps they should be.  (Well no crypto should
be export restricted in my view naturally, but for logical consistency
you understand...)  It would in my view be a good thing if one way
hash functions were declared to be non-exportable, as this would
clearly hinder commerce, as they are used in signatures and
authentication.

Also I seem to remember that Dan Berntstien's case was based on the
prior restraint in him not being free to discuss his method for
turning a one way hash into a symmetric key function.

Adam
--
Exported RSA today?  --> http://www.dcs.ex.ac.uk/~aba/rsa/

#!/bin/perl -sp0777i<X+d*lMLa^*lN%0]dsXx++lMlN/dsM0<j]dsj
$/=unpack('H*',$_);$_=`echo 16dio\U$k"SK$/SM$n\EsN0p[lN*1
lK[d2%Sa2/d0$^Ixp"|dc`;s/\W//g;$_=pack('H*',/((..)*)$/)