Liste des Groupes | Revenir à theory |
On 10/16/2024 8:10 AM, joes wrote:Am Wed, 16 Oct 2024 07:52:00 -0500 schrieb olcott:On 10/16/2024 1:32 AM, joes wrote:But it should work for DDD.Am Tue, 15 Oct 2024 22:52:00 -0500 schrieb olcott:It is true that a termination analyzer is not required to workOn 10/15/2024 9:11 PM, Richard Damon wrote:It definitely does. An uncomputable analyser is useless.On 10/15/24 8:39 AM, olcott wrote:Not at all. A termination analyzer need not be a Turing computableOn 10/15/2024 4:58 AM, joes wrote:But it needs to be computationally equivalent to one to ask aboutAm Mon, 14 Oct 2024 20:12:37 -0500 schrieb olcott:Even people of low intelligence that are not trying to be asOn 10/14/2024 6:50 PM, Richard Damon wrote:>On 10/14/24 12:05 PM, olcott wrote:On 10/14/2024 6:21 AM, Richard Damon wrote:On 10/14/24 5:53 AM, olcott wrote:On 10/14/2024 3:21 AM, Mikko wrote:On 2024-10-13 12:49:01 +0000, Richard Damon said:On 10/12/24 8:11 PM, olcott wrote:Can you please give the date and time? Did you also explicitlyI quit claiming this many messages ago and you didn't bother toTrying to change to a different analytical framework than theBut, you claim to be working on that Halting Problem,
one that I am stipulating is the strawman deception.
*Essentially an intentional fallacy of equivocation error*
notice.
disclaim it or just silently leave it out?
disagreeable as possible would be able to notice that a specified
C function is not a Turing machine.
Termination.
function.
correctly for all inputs.
That there is one way that HHH can consistently catch theDDD does terminate. Otherwise it would contradict the HP.
non-terminating pattern of its input proves that this can be done.
Mike suggested some different ways that would seem to be TuringWhat are those ways?
computable yet too convoluted to be time consuming for me to implement
in practice.
If it sees no termination condition in itself (it's a recursion), beingIt need not dig into an infinite emulation chain. It merely needs to seeThe basic approach involves the idea that every state change of theNothing special. They aren't even running on the hardware proper.
emulations of emulations is data that belongs to the outermost
directly executed HHH.
It is too convoluted for me to provide a way for HHH to look insideIt cannot dig into an infinite simulation chain.
all of the emulations of emulations and pull out the data that it
needs, so knowing that this is possible is enough to know that it is
Turing computable.
that DDD calls HHH(DDD) twice in sequence having no termination
condition within DDD.
The emulated HHH is merely data to the executed termination analyzer.When HHH is an x86 emulation based termination analyzer then each DDDIf HHH doesn't return, DDD doesn't either, not even the directly
*correctly_emulated_by* any HHH that it calls never returns.
Each of the directly executed HHH emulator/analyzers that returns 0
correctly reports the above *non_terminating _behavior* of its input.
executed one.
--In the same way that each person in a marathon that are running at theEach inner HHH must abort if the outer does,Every nested HHH has seen one less execution trace than the next outerWhen HHH is an x86 emulation based termination analyzer then eachOnly because the nested HHH doesn't abort.
DDD *correctly_emulated_by* any HHH that it calls never returns.
one. The outermost one aborts its emulation as soon as it has seen
enough. Thus each inner HHH cannot possibly abort its own emulation.
same speed and ten feet behind the person in front of them must win the
race.
The first person does not win the marathon they merely tie with everyone
else.
The fact that they start out ahead and remain ahead in the whole
marathon has no effect on whether they reach the finish line first.
since they are the same program. Of course, the outer doesn't simulate
the inner abort, because it has already aborted. Therefore the outer
HHH doesn't need to abort, because the inner HHHs already halt by
themselves. Reverse the if(Root) check on line 500 or what in Halt7.c.
As long as at least one of the infinitely many HHHs aborts, the whole
chain terminates (but the nested HHHs don't get simulated completely).
But they are all the same program,
so all of them must have the abort.
Les messages affichés proviennent d'usenet.