Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!rutgers!nike!ucbcad!ucbvax!hplabs!sdcrdcf!randvax!jim
From: jim@randvax.UUCP (Jim Gillogly)
Newsgroups: sci.crypt
Subject: Re: Decryption Program
Message-ID: <622@randvax.UUCP>
Date: Mon, 20-Oct-86 11:59:28 EDT
Article-I.D.: randvax.622
Posted: Mon Oct 20 11:59:28 1986
Date-Received: Wed, 22-Oct-86 22:30:03 EDT
References: <2121@mtuxo.UUCP> <828@chalmers.UUCP>
Organization: Banzai Institute
Lines: 26

augustss@chalmers.UUCP (Lennart Augustsson) writes:

>Monoalphabetic substitution can be handled by an algorithm described in
>"Breaking Substitution Ciphers Using a Relaxation Algorithm" by Peleg and
>Rosenfeld, published in CACM Vol 22 pp. 598-605 (Nov 1979).

Have you (or anyone other than the authors) been able to duplicate the
results in this paper?  Independently I and three other members of the
American Cryptogram Association tried to program this up (one each PL/1
and Pascal, two C efforts) and we weren't able to reproduce the reported
success.  At least three of us corresponded with the authors, who were
forthcoming with both their data ... but we couldn't make it happen.
Each of our efforts died the same way: the substitution would degenerate
to translating the same ciphertext letter into one plaintext letter (e.g.
converting everything to a "T").

The examples they gave in the paper used very long texts by puzzle standards:
the Gettysburg Address, for one; and we couldn't get it to work with that
either.

If anyone has *successfully* implemented this algorithm, I'd like very
much to see your code and data.
-- 
	Jim Gillogly
	{hplabs, ihnp4}!sdcrdcf!randvax!jim
	jim@rand-unix.arpa