Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!henry
From: henry@utzoo.UUCP (Henry Spencer)
Newsgroups: net.crypt
Subject: Re: randomly adding bits/bytes
Message-ID: <7031@utzoo.UUCP>
Date: Fri, 8-Aug-86 16:34:29 EDT
Article-I.D.: utzoo.7031
Posted: Fri Aug  8 16:34:29 1986
Date-Received: Fri, 8-Aug-86 16:34:29 EDT
References: <8608042018.AA04376@ucbjade.Berkeley.Edu>, <1026@g.cs.cmu.edu>
Organization: U of Toronto Zoology
Lines: 17

> ... choose as a function of the key a mapping... such that...
> the number of ciphertext blocks
> that map onto each plaintext block approximates the frequency of that
> plaintext block... [to encrypt] *randomly* chose a representative ...
> Immune to frequency analysis, no special location for uncompress keys.

This technique has routinely been used in both ciphers and codes for
centuries.  It doesn't make them unbreakable, but it does make it harder.
Computerizing eliminates some of the weaknesses of manual methods, notably
the tendency to pick ciphertext versions systematically.  (In particular,
virtually all codes had multiple ciphertext equivalents for "STOP", but
it was all too common for code clerks to memorize one or two of them to
speed things up.  Cryptanalysts could sometimes tell when a new code clerk
arrived because of a sudden surge of new groups for "STOP".)
-- 
				Henry Spencer @ U of Toronto Zoology
				{allegra,ihnp4,decvax,pyramid}!utzoo!henry