Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!rutgers!ames!ptsfa!ihnp4!homxb!houxm!hropus!jgy
From: jgy@hropus.UUCP (John Young)
Newsgroups: sci.math,sci.crypt
Subject: Re: Encryption of Hoffman coded text
Message-ID: <1227@hropus.UUCP>
Date: Mon, 10-Aug-87 10:06:31 EDT
Article-I.D.: hropus.1227
Posted: Mon Aug 10 10:06:31 1987
Date-Received: Tue, 11-Aug-87 04:28:12 EDT
References: <10489@linus.UUCP] <7876@mimsy.UUCP> <10604@linus.UUCP> <3515@cit-vax.Caltech.Edu>
Organization: Bell Labs, Holmdel, NJ
Lines: 19
Xref: mnetor sci.math:1825 sci.crypt:517

David Palmer writes:
> If you are going to encrypt a message by the standard breakable techniques,
> (Caesar cypher, Ultra, Purple, NBS DES, whatever technique the NSA is
> selling to our allies :-)  Is decryption more difficult if you use
> Hoffman coding (including Hoffman coding of common combinations such as 'th')
> on the plaintext before encypherment?  I am assuming that the people doing
> the attack know that you are doing Hoffman coding and know what the code is.
> .
> .
> .
> It seems to me that Hoffman coding makes the plaintext look more 'random'
> and, so, less amenable to statistical attack.  Is this so?
> 

My first reaction would be to say "yes it would be less amenable to
statistical attack", but then I wondered, if it's known your doing
Huffman coding (you say it is) then that algorithm has already done
the frequency analysis for the code breaker, and the partial-cypher
may now be in a "stricter" form.  Comments from experts??? (I'm not either)