Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: Mike Terry <news.dead.person.stones@darjeeling.plus.com>
Newsgroups: comp.theory
Subject: Re: Bad faith and dishonesty
Date: Fri, 30 May 2025 17:31:05 +0100
Organization: A noiseless patient Spider
Lines: 45
Message-ID: <101cmga$imoa$1@dont-email.me>
References: <m99YP.725664$B6tf.610565@fx02.ams4>
 <100uct4$184ak$1@dont-email.me> <100v9ta$1d5lg$7@dont-email.me>
 <1011eai$1urdm$1@dont-email.me> <10121bt$22da5$4@dont-email.me>
 <8bb5266e35845a4d8f2feb618c0c18629c04e4e7@i2pn2.org>
 <1012oj1$278f8$1@dont-email.me>
 <1196d9de2e2aebc1b6d1a85047192e8ea1aeb1f1@i2pn2.org>
 <10137lv$2djeu$1@dont-email.me> <ewIZP.135645$vK4b.131815@fx09.ams4>
 <1017l6l$3cerk$1@dont-email.me> <1017tr1$3drlu$5@dont-email.me>
 <1017ufm$3e54m$6@dont-email.me> <1019vm1$3u8nj$3@dont-email.me>
 <101a65n$3vsp7$1@dont-email.me> <101a86h$3vfam$6@dont-email.me>
 <101a9np$gl7$1@dont-email.me> <101bt7o$58on$1@dont-email.me>
 <101cis6$hv12$1@dont-email.me> <101cjjo$hqle$2@dont-email.me>
MIME-Version: 1.0
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 8bit
Injection-Date: Fri, 30 May 2025 18:31:06 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="544f6bd2ceaa26c9e82aa18c0abd7bf3";
	logging-data="613130"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX1/ApWyf6yyhkLVtsaPMcMzE42jE0bkzkmc="
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101
 Firefox/91.0 SeaMonkey/2.53.18.2
Cancel-Lock: sha1:BsYc0ljRNSqQpi5wduGSI4hdkOc=
In-Reply-To: <101cjjo$hqle$2@dont-email.me>
Bytes: 4089

On 30/05/2025 16:41, Richard Heathfield wrote:
> On 30/05/2025 16:29, Mike Terry wrote:
>> @Richard:  so you cannot make HHH decide non-halting simply by looping for a long long time, 
>> hoping HHH will get fed up!  That would just result in HHH simulating for a corresponding long 
>> long time. You need to feed it a program that halts, but matches one of his non-halting behaviour 
>> tests.  For example DDD.
> 
> What if I don't know whether it halts?
> 
> I followed up to vallor with a pseudocode sketch of such a program.

That was the Goldback Conjecture counter-example searcher?  If it halts it has found a 
counter-example, so GC is false.  If it never halts GC is true.

If GC were easy to prove/disprove it would have been settled a long time ago.  It is not going to be 
settled as a result of someone writing a partial halt decider that decides your program.  People 
have already tested GC up to around 4000000000000000000.  In practice, if you gave your program to 
HHH you would just see HHH running and running, which is not useful to anyone.

> 
> If HHH could deliver a reliably correct report for that program within a year or so, that would 
> probably be enough to earn Mr Olcott a place in the history books.

HHH obviously cannot do that.  Also, PO does not claim HHH is a (full) halt decider, so it does not 
affect his claims.

> 
> But at some point we have to place a ceiling on "long long time". A reporting program that keeps 
> saying "maybe next year" isn't much of a reporting program.
> 
We have two requirements:
a)  people want to actually /use/ real life halt analysis tools in their daily work.
     For such people, waiting a year for a result is no good, like you say.
     HHH is not a candidate for people wanting such a real-life tool.

b)  people want to understand the /theoretical/ limits of computation, hence the Halting Problem.
     The HP places no limits on how long a program can run, or how much storage it can
     consume.  PO's HHH is an attempt to invalidate one particular proof of HP.  It does
     not run for very long when running HHH/DDD, so we never have to face the
     "maybe next year" scenario.  [Related HHH/DDD scenarios that never halt are
     easily seen to never halt by simple code analysis.]


Mike.