Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail
From: dbush <dbush.mobile@gmail.com>
Newsgroups: comp.lang.c
Subject: Re: Proving the: Simulating termination analyzer Principle
Date: Sat, 5 Apr 2025 18:27:42 -0400
Organization: A noiseless patient Spider
Lines: 61
Message-ID: <vssaot$389d8$8@dont-email.me>
References: <vss56v$375du$2@dont-email.me> <vss6ie$389d8$3@dont-email.me>
 <vss7av$375du$8@dont-email.me> <vss7em$389d8$5@dont-email.me>
 <vssa6p$3b2j0$2@dont-email.me>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Injection-Date: Sun, 06 Apr 2025 00:27:41 +0200 (CEST)
Injection-Info: dont-email.me; posting-host="bdf677863da665303917a0e1a85da663";
	logging-data="3417512"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX18RvnSrElDS8a5eRQFvsZmq"
User-Agent: Mozilla Thunderbird
Cancel-Lock: sha1:ahGeNPTEULnB09L/s/uvn5SDyVE=
Content-Language: en-US
In-Reply-To: <vssa6p$3b2j0$2@dont-email.me>
Bytes: 3187

On 4/5/2025 6:18 PM, Richard Heathfield wrote:
> On 05/04/2025 22:31, dbush wrote:
>> On 4/5/2025 5:29 PM, olcott wrote:
>>> On 4/5/2025 4:15 PM, dbush wrote:
>>>> On 4/5/2025 4:52 PM, olcott wrote:
>>>>> *Simulating termination analyzer Principle*
>>>>> It is always correct for any simulating termination
>>>>> analyzer to stop simulating and reject any input that
>>>>> would otherwise prevent its own termination.
>>>>>
>>>>> void DDD()
>>>>> {
>>>>>     HHH(DDD);
>>>>>     return;
>>>>> }
>>>>
>>>> Except when doing so would change the input, as is the case with HHH 
>>>> and DDD.
>>>>
>>>> Changing the input is not allowed.
>>>
>>> You may disagree that the above definition
>>> of simulating termination analyzer is correct.
>>>
>>> It is self-evident that HHH must stop simulating
>>> DDD to prevent its own non-termination.
>>>
>>
>> Changing the input is not allowed.
> 
> You're right, but it doesn't matter very much as long as terminates() 
> *always* gets the answer right for any arbitrary program tape and any 
> data tape. Mr Olcott's fails to do that.
> 

Of course you're correct. His criteria is basically what happens if you 
replace the code of X with a pure simulator and run X(Y) for some Y.

This is essentially the same as the halting criteria in cases where Y 
does not call X at some point, but not in cases where it does.

He's explicitly agreed that this characterization is correct:


On 2/22/2025 1:02 PM, olcott wrote:
 > On 2/22/2025 11:10 AM, dbush wrote:
 >> On 2/22/2025 11:43 AM, olcott wrote:
 >>> The first point is DD correctly simulated by HHH cannot
 >>> possibly terminate normally by reaching its own "return"
 >>> instruction.
 >>
 >> In other words, if the code of HHH is replaced with an unconditional 
simulator then it can be shown that DD is non-halting and therefore 
HHH(DD)==0 is correct.
 >>
 >
 > Wow finally someone that totally gets it.


But has since distanced himself from it because it makes his error too 
clear.