Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version VT2.1 1/30/84; site vortex.UUCP
Path: utzoo!watmath!clyde!floyd!harpo!decvax!vortex!lauren
From: lauren@vortex.UUCP (Lauren Weinstein)
Newsgroups: net.crypt
Subject: secure codes
Message-ID: <239@vortex.UUCP>
Date: Mon, 30-Jan-84 06:34:35 EST
Article-I.D.: vortex.239
Posted: Mon Jan 30 06:34:35 1984
Date-Received: Wed, 8-Feb-84 04:42:38 EST
Organization: Vortex Technology, Los Angeles
Lines: 28

In cases of very important military or diplomatic communications,
I believe that the coding technique of choice is still the time-honored
"one-time" code.  This technique involves pre-sending to your destination
(by trusted means, such as inside a diplomatic pouch) a long "sequence"
of data to be used for encryption and/or decryption purposes.  Essentially,
the concept is simple.  For each byte you wish to encode/decode, you perform
some function based on that byte and the NEXT byte in the coding "sequence".
You never use the same sequence data byte twice, and the coding sequence
itself (magnetic tape, ROM, or whatever) is also never used again.
The coding sequence itself would normally be as random as possible (for
example, based on the "noise" generated by the decay of a radioactive
particle).

In the old days, the coding sequence was sent on paper, thusly the
term "one-time pad."  Today, such data would usually be sent on magnetic
tape, on ROMs within calculator-like encoding/decoding devices, or by similar
means.  Of course, old-style pads are still a possibility in some
cases.

One-time codes are generally considered to be essentially impossible
to crack, since there is no "key" nor pattern to uncover.  They are
generally reserved for important communications, since they involve
considerable logistic problems, including maintaining the sync between
the encoding/decoding sides of the communication using the sequence
data, and the hassles of delivering the sequence data to the destination
party in the first place.

--Lauren--