Liste des Groupes | Revenir à theory |
Mike Terry <news.dead.person.stones@darjeeling.plus.com> writes:"The computation in question only halts because it is halted by the decider on which it is built."
On 23/05/2025 19:37, Keith Thompson wrote:Maybe it makes less sense out of the context it was posted in. This wasBen Bacarisse <ben@bsb.me.uk> writes:>Mike Terry <news.dead.person.stones@darjeeling.plus.com> writes:[...]And the big picture is that this can be done because false is theHmm. I don't read that the way you do. Did I miss something?
correct halting decision for some halting computations. He has said
this explicitly (as I have posted before) but he has also explained it
in words:
>
| When-so-ever a halt decider correctly determines that its input would
| never halt unless forced to halt by this halt decider this halt
| decider has made a correct not-halting determination.
It assumes that the input is a non-halting computation ("its input
would never halt") and asserts that, in certain circumstances,
his mythical halt decider correctly determines that the input
is non-halting.
When his mythical halt decider correctly determines that its input
doesn't halt, it has made a correct non-halting determination.
It's just a tautology.
You're reading it the way most people would, and in the way I said Sipser
would be interpreting the oft-quoted "Sipser quote". I don't think you've
missed anything particularly.
when he was being less obtuse. The computation in question only halts
because it is halted by the decider on which it is built. It is a
halting computation, but according to PO it can reported as not halting
because of what would happen if it were not halted by the decider from
which it is derived.
Subsequent wordings have all been about hiding this. Just prior to thisYes, current posters are right on top of that, calling it out as "changing the input". I'm not sure PO realises he is changing the input, and if he does, whether he understands /why/ that is completely out of order. PO has recently started talking about what happens when we hypothetically "change the HHH/DDD input /pair/ " so that makes me think he does now explicitly realise that's what's going on, but not why it's a mistake.
wording was the even more explicit claim that non-halting is correct
because of what "would happen if line 15 were commented out". It's
always been about what would be the correct decision were the
computation not what it actually is.
Les messages affichés proviennent d'usenet.