Re: Can an abort decider be defined that cannot be fooled by any pathological input?

On 3/25/24 9:12 PM, olcott wrote:

On 3/25/2024 7:47 PM, Richard Damon wrote:

On 3/25/24 8:37 PM, olcott wrote:

On 3/25/2024 7:22 PM, Richard Damon wrote:

On 3/25/24 8:11 PM, olcott wrote:

On 3/25/2024 6:47 PM, Richard Damon wrote:

On 3/25/24 1:33 PM, olcott wrote:

On 3/25/2024 11:16 AM, Fred. Zwarts wrote:

Op 25.mrt.2024 om 17:04 schreef olcott:

On 3/25/2024 10:48 AM, Fred. Zwarts wrote:

Op 25.mrt.2024 om 16:17 schreef olcott:

On 3/24/2024 3:51 AM, Fred. Zwarts wrote:

Op 24.mrt.2024 om 05:55 schreef olcott:

Can an abort decider be defined that cannot be fooled by any pathological input?
>
01 int D(ptr x)  // ptr is pointer to int function
02 {
03   int Halt_Status = H(x, x);
04   if (Halt_Status)
05     HERE: goto HERE;
06   return Halt_Status;
07 }
08
09 void main()
10 {
11   H(D,D);
12 }
>
Of all of the elements of the set of H(D,D) where H simulates its
input there are matched pairs of otherwise identical elements that
only differ by whether they abort their simulation or not.
>
The half of these that don't abort are incorrect because all deciders
must halt. This makes the other half correct about the abort/no abort
decision.

>
No. The other, aborting, half is just wrong, because it aborts when it is not needed. So, the half that aborts is wrong and it may be argued that it is better to not abort something that halts on its own and that

>
At least two software engineers with masters degrees in computer science
disagree.

>
Two is not many, considering that with Google for any invalid idea it is easy to find a several people with a master degree supporting it.
>

Exactly what are you software engineering skills?

>
I have been professionally programming since 1986 in several languages. (Non professionally I started programming in 1975). Since about 1990 I programmed in C and since about 2000 in C++.
>
>

I have been a professional C++ software engineer since Y2K.

>
I'm sorry to hear that olcott has been so smart, but now he does not even sees what even a beginner sees.

>
Can D correctly simulated by H terminate normally?
01 int D(ptr x)  // ptr is pointer to int function
02 {
03   int Halt_Status = H(x, x);
04   if (Halt_Status)
05     HERE: goto HERE;
06   return Halt_Status;
07 }
08
09 void main()
10 {
11   H(D,D);
12 }
>
*Execution Trace*
Line 11: main() invokes H(D,D);
>
*keeps repeating* (unless aborted)
Line 03: simulated D(D) invokes simulated H(D,D) that simulates D(D)

>
Even a beginner sees that, if the H that aborts is chosen, simulated H(D,D) aborts and returns false (unless aborted). So simulated D halts (unless aborted).
>

>
I am estimating that you must be fibbing about your programming skill.
The D simulated by any implementation of H (that aborts or does not
abort its simulation) shown above cannot possibly reach its own line 04
also shown above.
>

>
But that isn't the question.
>

>
*That <is> the abort decision question*

>
But you agreed that a correct abort decider oly NEEDS to abort its simulation if the correct simulation by a pure correct simulator of the input given to H (which doesn't change, so for this case, still calls that original H) will never reach a final state.
>

>

The question is does that machine described by the input Halt when run, or, alternatively, does its correct simulation (not just by H) run forever (and thus needs to be aborted)?
>

>
Since you know that H(D,D) must abort its simulation to prevent its
own infinite execution I don't understand why you would lie about it.

>
But an H that doesn't abort and an H that does abort are looking at different inputs "D", since you agree that the behavior of D changes based on the H that it is using.
>

>
Not at all. Of the infinite set of every possible implementation of
H where H(D,D) simulates its input everyone that chose to abort is
necessarily correct.
>
I don't understand why you persist in lying about this.
>

>
I really want to get on to the next step and see if any input can
fool an abort decider into making the wrong abort decision.

>
But you need to get this step right first.
>

>
Perhaps you already know that you are not up to this challenge?

>
No, it seems that YOU are not up to it, as you can't seem to understand the error that you are making.
>
You keep on lying to yourself about what your requirements are.
>

>
I am not the one that keeps using the strawman deception to change
the subject away from H(D,D) an abort decider for the above D.

>
Neither am I.
>
YOU agreed that the criteria for an abort decider is only CORRECT if a CORRECT simulation of the exact input given to H(D,D) (i.e UTM(D,D) ) does not halt, where D still calls that H(D,D)
>

 I never agreed to that.

Yes you did:
On 3/17/24 6:11 AM, olcott wrote:
 > On 3/17/2024 12:22 AM, Richard Damon wrote:
 >> To me, for H to NEED to abort its simulation, that means that when giving the input to a correct simulator, that simulator will not halt.
 >>
 > Yes that is correct.
 >
Since the ONLY definition of a "Correct Simulator" is a simulator that never aborts, you did..

 

>

>

>

Correct simulation refering to a simulation of the EXACT input given to H, that calls the H that you claim to be giving the correct answer (and not the "correct simulator")

>

>

>

>