Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: notesfiles - hp 1.2 08/01/83; site hp-pcd.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!unc!mcnc!decvax!ittvax!dcdwest!sdcsvax!sdcrdcf!hplabs!hp-pcd!joseph
From: joseph@hp-pcd.UUCP
Newsgroups: net.math
Subject: Re: Re: 4-color CONJECTURE
Message-ID: <13900004@hp-pcd.UUCP>
Date: Tue, 14-Aug-84 15:53:00 EDT
Article-I.D.: hp-pcd.13900004
Posted: Tue Aug 14 15:53:00 1984
Date-Received: Sat, 1-Sep-84 09:21:50 EDT
References: <14798@arizona.UUCP>
Organization: Hewlett-Packard - Corvallis, OR
Lines: 16
Nf-ID: #R:arizona:-1479800:hpcvlo:13900004:000:899
Nf-From: hpcvlo!joseph    Aug 28 14:53:00 1984


Concerning the 4-color conjecture and its demonstration by computer,
I would like to add the following to what already appeared above:

Even assuming that the algorithm and its coding is provably correct,
(absurd as this assumption may or may not be) calling its execution 
a proof, simply baffles me. Even the mean time to failure of the hardware
is less than the running time of this program. Re-running of the program
to check for the same results has not and probably will not be performed.
(computer time is too valuable)  The non-zero probability that there 
was a hardware error implies that there is a non-zero probability
that the "proof" is wrong. (a bit could have flipped yielding incorrect
behavior) However unlikely this event may be, anyone who thinks that proofs 
can be probabilistic, should be referred to the nearest logic textbook.

                             --Joseph Albert