Newsgroups: sci.crypt
Path: msuinfo!caen!sdd.hp.com!cs.utexas.edu!uunet!boulder!mumble.cs.Colorado.EDU!barrett
From: barrett@mumble.colorado.edu (Dave Barrett)
Subject: DES Preserving Filesizes
Message-ID: <1992Jan23.063858.23100@colorado.edu>
Originator: barrett@mumble.cs.Colorado.EDU
Keywords:  DES CFB RSA
Sender: news@colorado.edu (The Daily Planet)
Nntp-Posting-Host: mumble.cs.colorado.edu
Organization: University of Colorado, Boulder
Date: Thu, 23 Jan 1992 06:38:58 GMT
Lines: 18

Actually, it is possible to use the DES in CFB mode to encipher 8 bytes at
a time and use the same method to encode the last n (where n < 8) bytes of
the text.   Thus, there is no penalty for using CFB mode over ECB mode in
software implementations.  Of course, the last n (n<8) bytes are subject to an
active wiretapper substituting another message there.  In addition to 
preserving the filesize, it appears to be tolorent of flipped bits, but
not bit-slip (If I'm not mistaken).

Is CFB mode less secure than CBC?  If so, in what way?

I would LOVE it if someone could publish a secure public key algorithm that
wasn't patented.  Or alternatively, if RSA would be more accomodating about
licensing their patent to individuals wishing to use it in programs they write
for personal use.  I'm patiently waiting for the RSA patent to expire in a
(very) few years.

Dave Barrett
barrett@asgard.cs.colorado.edu
