Path: news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: Mikko <mikko.levanto@iki.fi>
Newsgroups: comp.theory
Subject: Re: ChatGPT agrees that I have refuted the conventional Halting Problem proof technique --- Full 38 page analysis
Date: Thu, 26 Jun 2025 09:36:36 +0300
Organization: -
Lines: 74
Message-ID: <103ippk$3alk7$1@dont-email.me>
References: <103acoo$vp7v$1@dont-email.me> <728b9512cbf8dbf79931bfd3d5dbed265447d765@i2pn2.org> <103ag9k$10fmp$1@dont-email.me> <7b01bff1fe560095410422094a05ccac24c9fa7a@i2pn2.org> <103bodf$1a3c8$1@dont-email.me> <1b5b8f6a6c809724740bc68be167c5d535031e06@i2pn2.org> <103c3ir$1cme6$6@dont-email.me> <2d41ccbe7effb0813b82db178812d695d8884c90@i2pn2.org> <103cgtn$1gvif$2@dont-email.me> <587683d619382eee02219afcfb95b5247c984bb8@i2pn2.org> <103ebi6$22250$3@dont-email.me> <103g8lp$2ks4d$1@dont-email.me> <103h555$2rinm$3@dont-email.me>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Injection-Date: Thu, 26 Jun 2025 08:36:37 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="5969cac5bec68bb40ce1ba5008bf02cb";
	logging-data="3495559"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX19AK9s/9S4S6zIyjKpZt3BI"
User-Agent: Unison/2.2
Cancel-Lock: sha1:0q6WyoognBFEqCBM9bdMG7xRqxo=

On 2025-06-25 15:38:13 +0000, olcott said:

> On 6/25/2025 2:32 AM, Mikko wrote:
>> On 2025-06-24 14:09:10 +0000, olcott said:
>> 
>>> On 6/24/2025 4:27 AM, joes wrote:
>>>> Am Mon, 23 Jun 2025 16:28:23 -0500 schrieb olcott:
>>>>> On 6/23/2025 2:58 PM, joes wrote:
>>>>>> Am Mon, 23 Jun 2025 12:40:43 -0500 schrieb olcott:
>>>>>>> On 6/23/2025 10:34 AM, joes wrote:
>>>>>>>> Am Mon, 23 Jun 2025 09:30:07 -0500 schrieb olcott:
>>>> 
>>>>>>>>> If you read the 38 pages you will see how this is incorrect. ChatGPT
>>>>>>>>> "understands" that any program that must be aborted at some point to
>>>>>>>>> prevent its infinite execution is not a halting program.
>>>>>>>> Such as HHH, making it not a decider (when simulated).
>>>>>>>> 
>>>>>>> My claim is that
>>>>>> [blah blah non sequitur]
>>>>>> Well MY claim is that HHH simulated HHH (itself) doesn't halt.
>>>>>> 
>>>>>>> obvious
>>>>>> You know what, it actually IS obvious that HHH can't simulate past the
>>>>>> call to HHH. Thanks for coming to my Ted talk.
>>>>>> 
>>>>> Thus when HHH is simulating DDD and DDD calls HHH(DDD) the outer HHH
>>>>> does simulate itself simulating DDD.
>>>> Sure, it simulates *into* the call, but it never returns, which is
>>>> precisely why you abort it.
>>>> 
>>>> [more irrelevant stuff]
>>> 
>>> 
>>> void DDD()
>>> {
>>>    HHH(DDD);
>>>    return;
>>> }
>>> 
>>> *This is the question that HHH(DDD) correctly answers*
>>> Can DDD correctly simulated by any termination analyzer
>>> HHH that can possibly exist reach its own "return" statement
>>> final halt state?
>> 
>> Answering that question prevents HHH(DDD) from answering any
>> other question because it can only answer one question.
>> A termination analyzer is required to answer a different
>> question, which HHH(DDD) does not. Therefore HHH is not a
>> termination analyzer.
> 
> It turns out that the question a halt decider must answer
> has always been a bogus question because no TM can ever
> takes a directly executing TM as its input.

It is not a bogus question. It is a question that must be asked
and answered about every program that is intended for routine
use. A partial termination analyzer may be useful for the
purpose. If it fails something else may be used or possibly
the program can be chaged so that the analyzer does not fail.
In any case, if the answer cannot be determined then the
program is not acceptable for general use.

> Deciders must always
> compute the mapping *from* inputs thus are not allowed to
> report on non-inputs. Partial Halt Deciders are actually
> required to report on the behavior that their input specifies.

In particular, a partial decider is not allowed to report on its own
behaviour except that it may report that it is unable to report on
the behaviour specified by its input.

-- 
Mikko