Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Re: A UTM halt decider <is> a pure function of its inputs (halt deciding)
22 views
Skip to first unread message
olcott
Mar 22, 2021, 5:41:03 PM3/22/21
to
On 3/22/2021 2:50 PM, Malcolm McLean wrote:
> On Monday, 22 March 2021 at 02:06:18 UTC, olcott wrote:
>>
>> OK so your whole basis in anchored in the notion that a Turing
>> equivalent machine cannot possibly derive a protocol for a simulated
>> computation to invoke any functionality of its simulator.
>>
> Exactly. Linz's proof refers to halt deciders which are Turing machines.
> If you relax this requirement, there are often ways of meeting the challenge
> of producing an H / H_Hat pair which behave consistently.
>
> But that should have been accepted long ago. The question is whether,
> by relaxing the requirement that H and (by extension) H_Hat be a Turing
> machine, you've changed the question in any interesting way, or
> if you've just passed a global flag to H allowing it to flip its answer based
> on its call context. If fundamentally you've just passed the global
> flag , but disguised it cleverly, then that would also be interesting in its own
> right, but as a trick rather than as a contribution to computer theory.
>
A global UTM / x86 emulator derives an execution trace incrementally one
simulated instruction at a time that is a pure function of its inputs.
No machine descriptions are simulated outside of the scope of this
global UTM / x86 emulator.
After the simulation of each instruction UTM / x86 emulator invokes a
halt decider that derives a return value to the simulator EXECUTING /
HALTED / WILL_NOT_HALT as a pure function of its input execution trace.
The global UTM / x86 emulator continues simulating its input while this
return value from the halt decider is EXECUTING.
http://www.liarparadox.org/Halting_problem_undecidability_and_infinite_recursion.pdf
--
Copyright 2021 Pete Olcott
"Great spirits have always encountered violent opposition from mediocre
minds." Einstein
> On Monday, 22 March 2021 at 02:06:18 UTC, olcott wrote:
>>
>> OK so your whole basis in anchored in the notion that a Turing
>> equivalent machine cannot possibly derive a protocol for a simulated
>> computation to invoke any functionality of its simulator.
>>
> Exactly. Linz's proof refers to halt deciders which are Turing machines.
> If you relax this requirement, there are often ways of meeting the challenge
> of producing an H / H_Hat pair which behave consistently.
>
> But that should have been accepted long ago. The question is whether,
> by relaxing the requirement that H and (by extension) H_Hat be a Turing
> machine, you've changed the question in any interesting way, or
> if you've just passed a global flag to H allowing it to flip its answer based
> on its call context. If fundamentally you've just passed the global
> flag , but disguised it cleverly, then that would also be interesting in its own
> right, but as a trick rather than as a contribution to computer theory.
>
A global UTM / x86 emulator derives an execution trace incrementally one
simulated instruction at a time that is a pure function of its inputs.
No machine descriptions are simulated outside of the scope of this
global UTM / x86 emulator.
After the simulation of each instruction UTM / x86 emulator invokes a
halt decider that derives a return value to the simulator EXECUTING /
HALTED / WILL_NOT_HALT as a pure function of its input execution trace.
The global UTM / x86 emulator continues simulating its input while this
return value from the halt decider is EXECUTING.
http://www.liarparadox.org/Halting_problem_undecidability_and_infinite_recursion.pdf
--
Copyright 2021 Pete Olcott
"Great spirits have always encountered violent opposition from mediocre
minds." Einstein
Bonita Montero
Mar 26, 2021, 9:01:27 AM3/26/21
to
STOP POSTING IN comp.lang.c/c++, IDIOT !!!
Bonita Montero
Mar 31, 2021, 2:19:16 AM3/31/21
to
STOP POSTING in comp.lang.c/c++ !!!!!!!!!!!!!!!!!!
0 new messages