On 7/16/2021 5:24 PM, Alan Mackenzie wrote:
> Mr Flibble <flibble@reddwarf.jmc> wrote:
>> On Fri, 16 Jul 2021 19:46:28 -0000 (UTC)
>> Alan Mackenzie <acm@muc.de> wrote:
>
>>> Mr Flibble <flibble@reddwarf.jmc> wrote:
>>>> On Fri, 16 Jul 2021 19:28:12 +0100
>>>> Ben Bacarisse <ben.usenet@bsb.me.uk> wrote:
>
>>>>> Mr Flibble <flibble@reddwarf.jmc> writes:
>
>>>>>> A pathological program that executes a black box decider and
>>>>>> returns the opposite result can be detected by the black box
>>>>>> decider.
>
>>>>>> So there are THREE possible results the black box decider can
>>>>>> return:
>
>>>>>> 1) Program halts
>>>>>> 2) Program does not halt
>>>>>> 3) Program is pathological and can be discarded as invalid.
>
>>>>>> Halting problem solved.
>
>>>>> This one came up every year when one teaches this material. Every.
>>>>> Single. Year. In the end, I decided to bring it up myself and set
>>>>> explaining what's wrong the argument as an exercise. I wonder if
>>>>> any of the students here would like to have a go at that?
>
>>>>> (By student, I mean anyone reading this who has not read a
>>>>> textbook on this topic.)
>
>>>> I simply don't believe you. You appear to be a pathological liar;
>>>> feel free to provide evidence of what you are claiming.
>
>>> That's quite uncalled for. Reading (or browsing) these interminable
>>> threads over the last months, it's quite clear that Ben is an expert.
>>> You, by contrast, are merely exercising your freedom of expression.
>
>>> That you, in all apparent seriousness, put forward your 1), 2), 3)
>>> above indicates you are far from an expert. These things aren't a
>>> matter of opinion, they're a matter of settled fact.
>
>> Expert? I have a CompSci BSc (Hons) degree, dear.
>
> That surprises me greatly. Congratulations! As part of your course, did
> you study things like the halting problem?
>
>> Also, "far from an expert" is an attack on the person and not the
>> argument, a logical fallacy, dear.
>
> No, it's an attack on your 1), 2), 3). No expert could have seriously
> put 3) forward. Like Peter Olcott, you want people to fall into the trap
> of taking established falsehoods seriously and debating them point by
> point. When somebody such as yourself posts falsehoods, particularly
> ludicrous falsehoods, it is folly to go into an equal status argument
> with those falsehoods. Rubbish must be decried as rubbish, and the
> people advancing falsehoods must also be so decried. The truth matters,
> and in this matter, the truth has been settled for many decades.
>
> Every program is either halting or not halting. Assuming the black box
> returns the correct result, it can only return 1) or 2). There is no
> such thing as a "pathological program" in this sense.

That an input was intentionally defined to do the opposite of whatever a
corresponding TM decides <is> the exact same pathology as the liar paradox.

That an input was intentionally defined to do the opposite of whatever a
corresponding TM decides <is> the exact same pathology as the liar paradox.

That an input was intentionally defined to do the opposite of whatever a
corresponding TM decides <is> the exact same pathology as the liar paradox.

Flibble sees this. You as one of many mere sheep refuse to even look at
the counter-argument. That is despicable.

1,2,3 is the same as my work from 2004.

> How would it
> behave? Somehow neither halt, nor not halt? Get stuck in some sort of
> limbo where it's no longer running yet hasn't yet ceased to run? The
> notion is incoherent.
>
> Come on Mr Flibble, please assure us all that your 1), 2), 3) above was
> intended as some sort of joke, or parody, or something.
>
>>> When it comes to the halting problem, there is no dispute, except
>>> amongst cranks. It has a rock solid proof, much as do the trisection
>>> of the angle or the squaring of the circle.
>
>> Cranks? Ad hom logical fallacy (again), dear.
>
> Not at all. A statement of plain fact. The truth of the halting problem
> is settled, and only cranks dispute it.
>
>> Rock solid proof? Prove it, dear.
>
> No. I don't have the energy. You can find proofs far more eloquent than
> I could give in the text books you should still have from your student
> days. Failing that, ask Peter Olcott. I believe he keeps them on his
> website.
>
>> /Flibble
>

--
Copyright 2021 Pete Olcott

"Great spirits have always encountered violent opposition from mediocre
minds." Einstein

On 7/16/2021 5:54 PM, Alan Mackenzie wrote:
> [ Malicious cross-posting removed ]
>
> In comp.theory olcott <NoOne@nowhere.com> wrote:
>> On 7/16/2021 5:24 PM, Alan Mackenzie wrote:
>>> Mr Flibble <flibble@reddwarf.jmc> wrote:
>>>> On Fri, 16 Jul 2021 19:46:28 -0000 (UTC)
>>>> Alan Mackenzie <acm@muc.de> wrote:
>
>>>>> Mr Flibble <flibble@reddwarf.jmc> wrote:
>>>>>> On Fri, 16 Jul 2021 19:28:12 +0100
>>>>>> Ben Bacarisse <ben.usenet@bsb.me.uk> wrote:
>
>>>>>>> Mr Flibble <flibble@reddwarf.jmc> writes:
>
> [ .... ]
>
>>>>>>>> So there are THREE possible results the black box decider can
>>>>>>>> return:
>
>>>>>>>> 1) Program halts
>>>>>>>> 2) Program does not halt
>>>>>>>> 3) Program is pathological and can be discarded as invalid.
>
>>>>>>>> Halting problem solved.
>
> [ .... ]
>
>>>> Also, "far from an expert" is an attack on the person and not the
>>>> argument, a logical fallacy, dear.
>
>>> No, it's an attack on your 1), 2), 3). No expert could have seriously
>>> put 3) forward. Like Peter Olcott, you want people to fall into the
>>> trap of taking established falsehoods seriously and debating them
>>> point by point. When somebody such as yourself posts falsehoods,
>>> particularly ludicrous falsehoods, it is folly to go into an equal
>>> status argument with those falsehoods. Rubbish must be decried as
>>> rubbish, and the people advancing falsehoods must also be so decried.
>>> The truth matters, and in this matter, the truth has been settled for
>>> many decades.
>
>>> Every program is either halting or not halting. Assuming the black box
>>> returns the correct result, it can only return 1) or 2). There is no
>>> such thing as a "pathological program" in this sense.
>
>> That an input was intentionally defined to do the opposite of whatever a
>> corresponding TM decides <is> the exact same pathology as the liar paradox.
>
> What corresponding TM? I never mentioned one. Just that the program you
> would like to label as "pathalogical" either halts or it doesn't. There
> is no such thing as a "pathalogical" program in this sense. I note you
> decline to address my questions (below) as to its behaviour.
>

rec routine P
§L:if T[P] go to L
Return §

// Strachey(1965) "An impossible program" CPL translated to C
// https://doi.org/10.1093/comjnl/7.4.313
void P()
{ if (T((u32)P))
L: goto L;
}

Which Boolean value can the Strachey T return to the Strachey P that is
the correct halt status of P ?

>> Flibble sees this. You as one of many mere sheep refuse to even look at
>> the counter-argument. That is despicable.
>
> I have looked at it, analysed it in detail, and it is rubbish. A program
> either terminates or it does not. There is no magical third behaviour.
>
>> 1,2,3 is the same as my work from 2004.
>
>>> How would it behave? Somehow neither halt, nor not halt? Get stuck
>>> in some sort of limbo where it's no longer running yet hasn't yet
>>> ceased to run? The notion is incoherent.
>
>>> Come on Mr Flibble, please assure us all that your 1), 2), 3) above was
>>> intended as some sort of joke, or parody, or something.
>
>>>>> When it comes to the halting problem, there is no dispute, except
>>>>> amongst cranks. It has a rock solid proof, much as do the trisection
>>>>> of the angle or the squaring of the circle.
>
>>>> Cranks? Ad hom logical fallacy (again), dear.
>
>>> Not at all. A statement of plain fact. The truth of the halting problem
>>> is settled, and only cranks dispute it.
>
>>>> Rock solid proof? Prove it, dear.
>
>>> No. I don't have the energy. You can find proofs far more eloquent than
>>> I could give in the text books you should still have from your student
>>> days. Failing that, ask Peter Olcott. I believe he keeps them on his
>>> website.
>
>>>> /Flibble
>
>> --
>> Copyright 2021 Pete Olcott
>
>> "Great spirits have always encountered violent opposition from mediocre
>> minds." Einstein
>

--
Copyright 2021 Pete Olcott

"Great spirits have always encountered violent opposition from mediocre
minds." Einstein

On 7/17/2021 4:10 AM, Alan Mackenzie wrote:
> [ Malicious cross-posting removed ]
>
> In comp.theory olcott <NoOne@nowhere.com> wrote:
>> On 7/16/2021 5:54 PM, Alan Mackenzie wrote:
>
>>> In comp.theory olcott <NoOne@nowhere.com> wrote:
>>>> On 7/16/2021 5:24 PM, Alan Mackenzie wrote:
>>>>> Mr Flibble <flibble@reddwarf.jmc> wrote:
>>>>>> On Fri, 16 Jul 2021 19:46:28 -0000 (UTC)
>>>>>> Alan Mackenzie <acm@muc.de> wrote:
>
>>>>>>> Mr Flibble <flibble@reddwarf.jmc> wrote:
>>>>>>>> On Fri, 16 Jul 2021 19:28:12 +0100
>>>>>>>> Ben Bacarisse <ben.usenet@bsb.me.uk> wrote:
>
>>>>>>>>> Mr Flibble <flibble@reddwarf.jmc> writes:
>
>>> [ .... ]
>
>>>>>>>>>> So there are THREE possible results the black box decider can
>>>>>>>>>> return:
>
>>>>>>>>>> 1) Program halts
>>>>>>>>>> 2) Program does not halt
>>>>>>>>>> 3) Program is pathological and can be discarded as invalid.
>
>>>>>>>>>> Halting problem solved.
>
> [ .... ]
>
>>>>> Every program is either halting or not halting. Assuming the black
>>>>> box returns the correct result, it can only return 1) or 2). There
>>>>> is no such thing as a "pathological program" in this sense.
>
>>>> That an input was intentionally defined to do the opposite of
>>>> whatever a corresponding TM decides <is> the exact same pathology as
>>>> the liar paradox.
>
>>> What corresponding TM? I never mentioned one. Just that the program you
>>> would like to label as "pathalogical" either halts or it doesn't. There
>>> is no such thing as a "pathalogical" program in this sense. I note you
>>> decline to address my questions (below) as to its behaviour.
>
>
>> rec routine P
>> §L:if T[P] go to L
>> Return §
>
>
>> // Strachey(1965) "An impossible program" CPL translated to C
>> // https://doi.org/10.1093/comjnl/7.4.313
>> void P()
>> {
>> if (T((u32)P))
>> L: goto L;
>> }
>
>> Which Boolean value can the Strachey T return to the Strachey P that is
>> the correct halt status of P ?
>
> You're evading the question. This "pathalogical" program that neither
> halts nor fails to halt - what does its behaviour look like? You have no
> answer. There is no "pathalogical" behaviour, here. The program either
> halts or fails to halt. There is no third possibility.
>
> Regarding your question, which is something completely different, the
> answer is that the T cannot return a correct halt status of P. Why is
> that a problem? We've known for many decades that this is the case.
>

What has not been know for many decades is that undecidable decision
problems are merely errors and nothing more.

Mathematicians really hate to go to the philosophical foundation of the
notion of truth itself and instead continue the religion of what they
learned by rote.

>>>> Flibble sees this. You as one of many mere sheep refuse to even look at
>>>> the counter-argument. That is despicable.
>
>>> I have looked at it, analysed it in detail, and it is rubbish. A program
>>> either terminates or it does not. There is no magical third behaviour.
>
> [ .... ]
>
>> --
>> Copyright 2021 Pete Olcott
>
>> "Great spirits have always encountered violent opposition from mediocre
>> minds." Einstein
>

--
Copyright 2021 Pete Olcott

"Great spirits have always encountered violent opposition from mediocre
minds." Einstein

On 17/07/2021 00:32, olcott wrote:
> On 7/16/2021 5:24 PM, Alan Mackenzie wrote:
>>
>> Every program is either halting or not halting.  Assuming the black box
>> returns the correct result, it can only return 1) or 2).  There is no
>> such thing as a "pathological program" in this sense. 
>
> That an input was intentionally defined to do the opposite of whatever a
> corresponding TM decides <is> the exact same pathology as the liar paradox.
>
> That an input was intentionally defined to do the opposite of whatever a
> corresponding TM decides <is> the exact same pathology as the liar paradox.
>
> That an input was intentionally defined to do the opposite of whatever a
> corresponding TM decides <is> the exact same pathology as the liar paradox.
>

Sticking your fingers in your ears and shouting "Blah, blah, blah, I'm
not listening" does not constitute an argument.

> Flibble sees this.  You as one of many mere sheep refuse to even look at
> the counter-argument. That is despicable.
>

Generally speaking, when almost everyone else says you are wrong, it
does not mean you are the lone genius who sees the truth. It means you
are wrong.

>
> "Great spirits have always encountered violent opposition from mediocre
> minds." Einstein

And deluded minds encounter opposition from even more people.
Opposition does not imply that you are a "great spirit". You'd
appreciate that, if you understood anything about logic.

On 7/17/2021 11:24 AM, Ben Bacarisse wrote:
> olcott <NoOne@NoWhere.com> writes:
>
>> On 7/17/2021 4:10 AM, Alan Mackenzie wrote:
>
>>> Regarding your question, which is something completely different, the
>>> answer is that the T cannot return a correct halt status of P. Why is
>>> that a problem? We've known for many decades that this is the case.
>>
>> What has not been know for many decades is that undecidable decision
>> problems are merely errors and nothing more.
>
> Don't be silly. Asking if a program halts or not is not an error or a
> bad question. Asking if a context-free grammar is ambiguous or is not
> an error or a bad question. Asking if two groups are isomorphic or not
> is not an error of a bad question. You just say these things for the
> sake of having an opinion.
>

As always you diligently make sure to ignore the full context.

(1) When we ask does every element of the set of valid declarative
sentences have as associated Boolean value? The answer is yes.

(2) When we ask does every element of the set of sentences have as
associated Boolean value? The answer is no.

When you ignore the full context you ignore the distinction between (1)
and (2). In this case it is an honest mistake on your part because the
distinction can be difficult to fully appreciate.

> The set of subsets of N is uncountable. The set of Turing machines is
> countable. There must be an uncountable number of undecidable sets of
> numbers. That we've been able to find a few is handy, but it's not
> surprising.
>

--
Copyright 2021 Pete Olcott

"Great spirits have always encountered violent opposition from mediocre
minds." Einstein

On 7/17/2021 8:45 PM, Ben Bacarisse wrote:
> olcott <NoOne@NoWhere.com> writes:
>
>> On 7/17/2021 11:24 AM, Ben Bacarisse wrote:
>>> olcott <NoOne@NoWhere.com> writes:
>>>
>>>> On 7/17/2021 4:10 AM, Alan Mackenzie wrote:
>>>
>>>>> Regarding your question, which is something completely different, the
>>>>> answer is that the T cannot return a correct halt status of P. Why is
>>>>> that a problem? We've known for many decades that this is the case.
>>>>
>>>> What has not been know for many decades is that undecidable decision
>>>> problems are merely errors and nothing more.
>>> Don't be silly. Asking if a program halts or not is not an error or a
>>> bad question. Asking if a context-free grammar is ambiguous or is not
>>> an error or a bad question. Asking if two groups are isomorphic or not
>>> is not an error of a bad question. You just say these things for the
>>> sake of having an opinion.
>>>
>>
>> As always you diligently make sure to ignore the full context.
>
> You call is "extraneous complexity" but when you ignore it, it is
> usually crucial.
>
>> (1) When we ask does every element of the set of valid declarative
>> sentences have as associated Boolean value? The answer is yes.
>>
>> (2) When we ask does every element of the set of sentences have as
>> associated Boolean value? The answer is no.
>>
>> When you ignore the full context you ignore the distinction between
>> (1) and (2). In this case it is an honest mistake on your part because
>> the distinction can be difficult to fully appreciate.
>
> It's a trivial distinction. And an irrelevant one because every
> instance of the halting problem has a correct yes/no answer.
>
> For H to be a halt decider, H(P,I) must be true iff P(I) halts. The
> correct answer does not depend on anything but P and I. It does not
> depend on who or what we "ask". In fact, P(I) ether is or is not a
> finite computation even if we never ask anyone or anything about it.
>
> P(P) halts (according to you). H(P,P) == 0 (according to you). That is
> wrong (according to everyone but you).
>
> As far as I can see, there are no facts in dispute, only waffle.
>

void P(u32 x)
{ if (H(x, x))
HERE: goto HERE;
}

int main()
{ P((u32)P);
}

The fact is that the above computation never ever halts unless some H
aborts some P thus proving beyond all possible doubt that H[0] does
correctly decide that P[2] (zero based addressing) never halts.

--
Copyright 2021 Pete Olcott

"Great spirits have always encountered violent opposition from mediocre
minds." Einstein