On 11/12/2022 9:03 AM, Mr Flibble wrote:
> On Sat, 12 Nov 2022 09:52:12 -0500
> Richard Damon <Ric...@Damon-Family.org> wrote:
>
>> On 11/12/22 9:41 AM, Mr Flibble wrote:
>>> On Sat, 12 Nov 2022 09:32:23 -0500
>>> Richard Damon <Ric...@Damon-Family.org> wrote:
>>>
>>>> On 11/12/22 9:09 AM, Mr Flibble wrote:
>>>>> On Fri, 11 Nov 2022 19:24:53 -0500
>>>>> Richard Damon <Ric...@Damon-Family.org> wrote:
>>>>>
>>>>>> On 11/11/22 6:55 PM, Mr Flibble wrote:
>>>>>>> On Fri, 11 Nov 2022 18:25:58 -0500
>>>>>>> Richard Damon <Ric...@Damon-Family.org> wrote:
>>>>>>>
>>>>>>>> On 11/11/22 4:54 PM, Mr Flibble wrote:
>>>>>>>>>>> void E(void (*x)())
>>>>>>>>>>> {
>>>>>>>>>>> H(x, x);
>>>>>>>>>>> }
>>>>>>>>>>>
>>>>>>>>>>> Any H that does abort its simulation to prevent the infinite
>>>>>>>>>>> execution of E is correct to report non-halting. No shell
>>>>>>>>>>> game can correctly deny this.
>>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Any H that does aborts its simulation is INCORRECT because
>>>>>>>>>> the CORRECT simulation, as will the diret exectuion, will
>>>>>>>>>> halt. Note, such an H doesn't do a correct simulation, so any
>>>>>>>>>> arguement based on it doing so it just WRONG.
>>>>>>>>>
>>>>>>>>> The need to abort the simulation is due to the self reference
>>>>>>>>> category error present in the proof; what Olcott is getting
>>>>>>>>> wrong is the mapping of the need to abort the simulation to a
>>>>>>>>> halt decision of non-halting; it needs to instead be mapped to
>>>>>>>>> INVALID INPUT.
>>>>>>>>>
>>>>>>>>> /Flibble
>>>>>>>>>
>>>>>>>>
>>>>>>>> Nope, no self reference.
>>>>>>>>
>>>>>>>> E just has a copy of the H that claim to be deciding it, not a
>>>>>>>> "reference" to it. Turing Machines do not have the power to
>>>>>>>> directly express a reference.
>>>>>>>
>>>>>>> Nope, if it isn't a self reference then it is infinite copies
>>>>>>> all the way down so is the same category error manifesting in a
>>>>>>> different way.
>>>>>>>
>>>>>>> /Flibble
>>>>>>>
>>>>>>
>>>>>> Only if the "decider" makes that happen, in which case it isn't
>>>>>> actually a decider.
>>>>>>
>>>>>> If we assume a prospective decider exists, then the "Impossible"
>>>>>> program is simple to make from it, and is given one copy of the
>>>>>> description of itself, which is also simple to make.
>>>>>>
>>>>>> When run it makes a second copy of its description, and then
>>>>>> calls the decider.
>>>>>>
>>>>>> After that, it is the deciders job to make the decision in finite
>>>>>> time, by whatever method it wants. If it gets stuck in your
>>>>>> infinite loop, the decider is just wrong.
>>>>>>
>>>>>> The proof shows that what ever answer the decider does give (if
>>>>>> it gives one) will be wrong, and thus the decider doesn't meet
>>>>>> the requirements.
>>>>>>
>>>>>> No "Self Reference" in sight there only a program being given a
>>>>>> copy of something that just happens to be its own description.
>>>>>>
>>>>>> The only place we get any form of "Reference", is when we try to
>>>>>> ANALYSE or DESIGN the H to try to meet the challenge. There the
>>>>>> effect of the Self-Reference just lets us see that the task turns
>>>>>> out be be impossible, so no such program exists.
>>>>>
>>>>> You are fractally wrong on all fronts: in the traditional halting
>>>>> problem proofs based on [Strachey 1965] the program is impossible
>>>>> not due to self reference or infinite copies but because the input
>>>>> tries to do the opposite of what the decider decides; the category
>>>>> error that I have identified is different: it is an error of self
>>>>> reference and/or infinite copies; it is an error related to the
>>>>> fact that the input references a decider rather than being related
>>>>> to what the input does with the decision result of a decider.
>>>>>
>>>>> /Flibble
>>>>>
>>>>
>>>> But the infinite copies is a error in the Decider, not in
>>>> Strachey's program. The decider is SUPPOSED to be able to handle
>>>> ANY input and answer in finite time, If an input causes it to make
>>>> infinite copies, then the decided just doesn't meet its
>>>> requirements.
>>>>
>>>> Turing Machine can ALWAYS be legally built based on another Turing
>>>> Machine as a base. The only reason it wouldn't be allowed is if H
>>>> isn't actually a Turing Machine, so it CAN'T be a category error
>>>> if H is actualy a Turing Machine.
>>>>
>>>> All your declaration of a "Category Error" here is doing is
>>>> admitting that your H can't actually be a Turing Machine, but must
>>>> be of a HIGHER order logic system, which means H fails the
>>>> requirement to be the needed decider.
>>>
>>> In which case we get infinite turning machines all the way down: yet
>>> another manifestation of the category error I have identified.
>>>
>>> /Flibble
>>>
>>
>> Where are infinite machines? There is ONE machine being run, either H
>> or D, and it SIMULATING others, and if we get an infinite sequence of
>> simulations we have just shown that H was defective because it failed
>> to answer in finite time.
>>
>> This isn't a category error, but a design error in H.
>>
>> Note, when we start H, there is exactly two machines present in
>> representation on the tape, and two is much smaller than infinity.
>
> Nope, if,
>
> a) H is a copy, and
> b) H is a Turing Machine, and
> c) D is an input into H, and
> d) D references H, and
> e) H references D,
>
> then (d) and (e) repeat ad infinitum so we get infinite Turing Machines
> all the way down: a manifestation of the category error I have
> identified.
>
> /Flibble
>
void E(void (*x)())
{
H(x, x);
}
The point is
that D correctly simulated by H would never reach its own last
instruction and terminate normally after 1 to ∞ steps of correct simulation.
When H returns 0 to main() it is indicating
that D correctly simulated by H would never reach its own last
instruction and terminate normally after 1 to ∞ steps of correct simulation.