How do you intend to prove that?
>
>>
>> The fact that it can't give the right answer, just shows that the
>> problem is impossible to do, not that it is wrongly defined.
>
> If the simulation method is valid then it can give the right answer;
> for a [Strachey 1965] "Impossible Program" I believe the right answer
> is INVALID INPUT whilst Olcott believes it is NON-HALTING. If you
> continue to refuse to allow me the reasonable conceit of extending the
> definition of what constitutes a halt decider then I would have to
> concur with Olcott's answer of NON-HALTING
And what is "Invalid" about the machine?
Yes, if you want to CHANGE the definition of what you are looking at,
and thus are NOT computing the "Halting Function" but the
Flibble-Halting-like function, you can. It just has no impact on the
original problem or proofs claiming show that the original problem is
uncomputable.
Your method is functionally no different than just defining that any
machine that runs for more than, say, 100 steps, is non-halting. That is
just proposing an "alternate" definition for Halting and showing that
you CAN compute this alternate definition.
>
>>
>> Many problems are shown to be impossible to do, so that by itself is
>> not an error.
>
> We are not trying to SOLVE the Halting Problem, Olcott is trying to
> refute the Halting Problem proofs and I believe he has done so: his SHD
> highlights the category error present in those proofs.
>
> /Flibble
>
And saying that an input is non-halting when it Halts does not refute
the proof,
Neither does showing that you can perhaps build a Flibble-Halting-Like
decider, or at least one that appears to handle this particluar case.
Since you do this by just redefining the "correct" answer, it doesn't
show anything about the original question.