Re: Can you find anything wrong with this solution to the Halting Problem?

From: Peter Olcott (olcott_at_att.net)
Date: 07/14/04 

Date: Wed, 14 Jul 2004 00:18:30 GMT

> Look. Whenever someone produces a groundbreaking proof like
> Turing's, lots of mathematicians look it over to see if they
> can find mistakes. For something like Wiles's proof of
> Fermat's last theorem, it takes a long time because the
> mathematics involved is really difficult. For proofs of the
> halting theorem the maths is relatively easy, so checking it
> isn't terribly hard. In fact it's much simpler than the
> stuff on your web page, because the model of computation is
> completely specified, whereas yours involves talk about
> "functions" and "screen data", and I don't know what you
> mean by that. Note that the description to which you link
> from your page is only a sketch of the proof, not a proper
> proof -- he doesn't show how to convert the functions to
> Turing machines, for example.

A very important point. I will correct this immediately.
Here is what I added at the top of the new wep-page:

Only the conclusion from the original proof is retained
everything else has been explictly discarded. All that I
need to do to disprove the original proof, is to
disprove its conclusion, nothing else is required.

> Of course it is! If a well understood theorem says P, and
> someone comes along and says ¬P, the theorem is already a
> refutation of ¬P, so there's no work to do.

And of course every theorem is created only by infallible
people that never make any mistakes, thus (as you are
directly implying) no theorem has ever once been found
to be incorrect in the entire history of mankind.

Sorry for the sarcastic rant but I have seen your form
of reply far too many times.

> Other people have already pointed out flaws in your
> argument. The easiest one to grasp should be that it doesn't
> matter if WillHalt is in protected memory, because if
> WillHalt exists, it has a representation that can be fed
> back into LoopIfHalts whether or not we know which value it
> is, because it's supposed to work on /all/ values. You are
> playing a game that's like this one:

That original methodology has been disabled by my methods.
Go back and read it again. Study it as if it was true, then
(especially to section on possible bases for refutation)

you might be able to see that it is true. Glancing at a few
words prior to spouting off one refutation or another
is not going to ever possibly have success, because there
are always far too many holes in these sorts of refutations.

>
> Player A thinks of a positive whole number but keeps it secret.
> Player B has to name that number.
>
> Player B: 1, 2, 3, 4, ...
>
> and you, as player A, are claiming that because player B
> can't read your mind (and you are right in that: he can't!),
> he'll never name your number. He might never know that he
> has, but that's only because you don't have the grace to
> admit defeat.

Do you ever admit defeat when you are totally 100% correct?

Relevant Pages