Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!ut-sally!pyramid!decwrl!decvax!minow
From: minow@decvax.UUCP (Martin Minow)
Newsgroups: net.crypt
Subject: Re: randomly adding bits/bytes (compressing text before encryption)
Message-ID: <251@decvax.UUCP>
Date: Sun, 17-Aug-86 18:47:31 EDT
Article-I.D.: decvax.251
Posted: Sun Aug 17 18:47:31 1986
Date-Received: Tue, 19-Aug-86 02:01:30 EDT
References: <8608042018.AA04376@ucbjade.Berkeley.Edu> <7024@utzoo.UUCP> <806@petsd.UUCP>
Reply-To: minow@decvax.UUCP (Martin minow)
Organization: DEC - ULTRIX Engineering Group
Lines: 16

In discussing ways to make data encryption more robust, several
people suggested running the plaintext through LZ-compress first.
utzoo!henry noted that compress prepends a fixed length (and content)
header to the text, giving the codebreaker a hook into the encription.

The LZ-compress header exists only to communicate information about
the data file (number of bits in the compression and algorithm) that
aren't interesting in this context (they can be sent by some other
secure channel or hard-wired into the encryption/decription code).

One interesting thing about LZ -- one bad bit in the file may garble
the entire remainder of the file.  This, in itself, should make the
codebreaker's work harder.

Martin Minow
decvax!minow