Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!gamma!epsilon!zeta!sabre!bellcore!decvax!decwrl!greipa!pesnta!hplabs!sri-unix!KATZ@USC-ISIF.ARPA
From: KATZ@USC-ISIF.ARPA
Newsgroups: net.physics
Subject: reversible computing
Message-ID: <329@sri-arpa.ARPA>
Date: Thu, 27-Jun-85 17:05:44 EDT
Article-I.D.: sri-arpa.329
Posted: Thu Jun 27 17:05:44 1985
Date-Received: Mon, 1-Jul-85 07:09:48 EDT
Lines: 32

From:  Alan R. Katz 

In response to the question of tunneling:

I have not read all of the Sci Amer article, so I can't answer about this
particular point.  However, Richard Feynman (and others) have constructed
Quantum Mechanical models of reverisble computers, so such machines
are realizable in the Quantum world.

The trouble is that these models consists of constructing a wierd
Hamiltonian for a theoretical system.  They do not really say
how an actual system could be contructed.  The one(ones) mentioned
in Sci Amer may have problems, but in principle, you could do it.

For those who have not read the article, it is about making computers
that generate essentially no heat, and use essentially no energy.  The
key is using reversible microscopic processes (such as billiard ball
collisions) as gates in the computer.  Because the computation is
reversible, any noise will just "undo" the computation a little, rather
than mess it up.  Since energy is conserved, it takes no (very little) energy
to do the actual computation.  You get strange things like the number of
bits being conserved (so when you generate an answer, you also must
generate its complement).

After discussing various ways to do these computations based on 
billiard ball type collisions, you must take quantum mechanics into
account.  Feynman and others have done this and it still all works,
at least theoretically.


			Alan
-------