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

From: The Ghost In The Machine (ewill_at_aurigae.athghost7038suus.net)
Date: 07/10/04 

Date: Sat, 10 Jul 2004 20:01:09 GMT

In sci.logic, Peter Olcott
<olcott@worldnet.att.net>
 wrote
on Sat, 10 Jul 2004 15:40:48 GMT
<46UHc.79720$OB3.8783@bgtnsc05-news.ops.worldnet.att.net>:
> Only direct refutation or confirmation of this message will
> be replied to, anything else will be considered off-topic and
> ignored.
>
> http://www.netaxs.com/people/nerp/automata/halting2.html
> I am using this as the basis of my discussion so that the
> specific constituent parts of the Halting Problem can be
> explicitly referred to by their counter-part in this example.
> From what I understand this version of the Halting Problem
> is sufficiently analogous to the original.
>
> I am not sure that this "solution" is correct. There may be one or
> more errors or loopholes that I have failed to discover. Since the
> basic conclusion of the Halting Problem is that it is impossible
> to construct a WillHalt() function that will correctly determine
> whether or not any other program will halt or not, I have
> freely modified the other aspects of the problem definition
> to refute this single conclusion.
>
> (1) The LoopIfHalts() function can not see the screen.

ITYM tape. Turing machines don't have screens.

>
> (2) The only access that the LoopIfHalts() function has to the
> WillHalt() function is through a function call.

Turing machines don't have function calls.

>
> ***************************************************
> To solve the Halting Problem all that is needed is for the meaning
> of the return value from Willhalt() be kept from the LoopIfHalts()
> function. The WillHalt() function arranges a coded reply to the
> human user.

I could post a base64-encoded reply here of a certain swf file,
but I'll desist in deference to the other posters. :-P

As it is...how are you going to enforce keeping the return value
of the WillHalt() function from the LoopIfHalts() program?

Bear in mind that what LoopIfHalts() may be using is a *copy*
(or alternate implementation) of the WillHalt() function, as well.

>
> It could be as simple as one and zero. The meaning of true would
> be assigned to either one or zero, the meaning of false would be
> assigned to the other. Before the program takes the input of the
> LoopIfHalts() function it outputs either a one or a zero to the
> screen. Whichever (1 or 0) is output, holds the meaning of true.
> Whichever one it outputs is generated by a hardware noise based
> random number generator. (thus it is impossible to predict which
> of these two will be generated) Now the LoopIfHalts() function
> has no way to thwart the WillHalt() function, and the human user
> can understand the result.

It would be interesting to develop an alternate model of computing,
which uses neural networks. I'm not sure what the state of neural
computing is although it's already being used to filter out spam.

> ***************************************************
>
> 

-- 
#191, ewill3@earthlink.net -- insert random fuzzy logic here
It's still legal to go .sigless.

Relevant Pages
Loading