From: "E. ALLEN SMITH" <EALLENSMITH@ocelot.Rutgers.EDU>
Date: Mon, 29 Jan 1996 06:34:18 +0800
To: pgut001@cs.auckland.ac.nz
Subject: Re: "Concryption" Prior Art
Message-ID: <01I0JULCV3Z4A0UMAT@mbcl.rutgers.edu>
MIME-Version: 1.0
Content-Type: text/plain


	It occurs to me that such a combined form would be useful for the
idea of having a message that can get multiple results depending on the
passphrase used. One example might work as follows:
	A section of data is placed at the beginning of the encrypted material.
When it is decrypted or encrypted (depending on how one wants to work things)
with a given passphrase, it turns out a series of bits, reiterated as needed.
Each x bits is used to say how far along in the encrypted material the next
piece of information making up one encrypted message (using the same
passphrase) is. If you put in a different passphrase, you get a different
series of bits, and thus use a different set of information for the encrypted
material. The 
	The major problem that I can see with this scheme is overlap between
messages. I would guess that one would need to keep coming up with different
data sections until one originated that wasn't a problem. How long this would
take would depend on the value of x and how long the data section was. However,
this should only need to be done once for a given set of passphrases and the
corresponding key (used for all of them). Any alternate suggestions?
	Not being a programmer, I have no real idea how to put this concept
into practice. (And, moreover, someone else came up with the idea of multiple
data sets from a given encrypted message; I am simply suggesting a potential
mechanism).
	-Allen