Newsgroups: sci.math
From: haoyuep@aol.com (Dan Hoey)
Date: Dec 11, 1999 12:09 AM
Subject: Re: About James Harris

Clive Tooth <clive@pisquaredoversix.force9.co.uk> wrote:

>About James Harris (JHS)
>========================

>If you are new to sci.math and are considering responding to a
>post to or about James Harris you might wish to read this note first.

This is a very useful piece of information, and might be usefully
inserted every now and then--maybe every hundred posts or so--in
the JSH threads.

>James Harris thinks he has an elementary proof of Fermat's
>Last Theorem (FLT)....

Wait.  He _says_ he has an elementary proof.  It's questionable
whether he actually _thinks_ this nonsense is true, or whether he is
just lying.  It's hard to believe someone would lie so transparently
for so long, but it's also hard to believe someone could state these
falsehoods with belief.

>For several years Harris has been posting various
>versions of his proof to sci.math. His proofs are always
>wrong.  However they are virtually unintelligible so it is
>extremely tedious to find the mistakes. A lot of guesswork
>is required in order to make sense of what he is trying to
>say. He appears to be a simple attention seeker. Not caring if
>he is right or wrong, or even if he is intelligible or not.

It might be worthwhile to summarize JSHs's usual proof schema:

1) Assume a counterexample to FLT,
2) Manipulate the counterexample through changes of variables,
   factorizations, etc.
3) Arrive at a contradiction, and
4) Deduce that the counterexample (1) could not exist.

This method's effectiveness depends on being careless in step 2.
If you make enough random mistakes, sooner or later you always
arrive at a noticeable contradiction.  The errors also have
some resistance to easy falsification, since you can't just
put sample values in and watch for the manipulation to produce
an incorrect equation--the equations in step 2 depend on the
nonexistent counterexample of step 1.

Someone suggested that his method might arrive at a proof in
a few centuries.  I think they underestimate the power of random
carelessness.  It seems more likely that 10^100s of years could
be wasted on the errors that JSH is so fond of, than that he would
stumble on a correct proof in that time.  Since he is not interested
in learning anything, it's unlikely that the errors will become any
less frequent.

Dan Hoey <haoyuep@aol.com>