Re: The proof that I was referring to is on the website

newstome_at_comcast.net
Date: 08/12/04 

Date: Thu, 12 Aug 2004 02:29:43 GMT

In comp.theory Peter Olcott <olcott@worldnet.att.net> wrote:
>
> <newstome@comcast.net> wrote in message news:mgrSc.243329$IQ4.70802@attbi_s02...
>> In comp.theory Peter Olcott <olcott@worldnet.att.net> wrote:
>> >
>> > <newstome@comcast.net> wrote in message news:T0iSc.238825$%_6.48565@attbi_s01...
>> >> In comp.theory Peter Olcott <olcott@worldnet.att.net> wrote:
>> >> >
>> >> > <newstome@comcast.net> wrote in message news:gK5Sc.262164$JR4.167687@attbi_s54...
>> >> >> In comp.theory Peter Olcott <olcott@worldnet.att.net> wrote:
>> >> >> > <newstome@comcast.net> wrote in message news:oUWRc.229589$a24.164871@attbi_s03...
>> >> >> >> In comp.theory Peter Olcott <olcott@worldnet.att.net> wrote:
>> >
>> >> > Mine always provides the correct results every time.
>> >> > If you do anything to change this, then it would necessarily
>> >> > have a different execution trace.
>> >>
>> >> FINALLY, you actually are on the right track! It will not have a
>> >> different execution trace AT ALL, so therefore will have to be correct
>> >> every time if and only if yours is correct every time!
>> >> Congratulations, now you see it!
>> >
>> > Line 01) void LoopIfHalts(string SourceCode, string InputData)
>> > Line 02) {
>> > Line 03) if (WillHalt(SourceCode, InputData) == TRUE)
>> > Line 04) while (TRUE) // loop forever
>> > Line 05) ;
>> > Line 06) else
>> > Line 07) return; // FALSE or UNKNOWN
>> > Line 08) }
>> >
>> > IFF LoopIfHalts has identical behavior whether or not WillHalt returns
>> > TRUE or any other value, then you are correct that the execution trace
>> > does not vary. IFF LoopIfHalts has different behavior for TRUE than
>> > it does for any other value, then you are incorrect, and the execution
>> > trace does vary.
>>
>> I'm not talking about the execution trace for LoopIfHalts. I'm
>> talking about the execution trace of WillHalt. LoopIfHalts is
>> completely irrelevant to this. In fact, it doesn't even exist as far
>> as WillHalt is concerned.
>
> LoopIfHalts is crucially relevant to WillHalt, because the change
> in behavior (based on the returned TRUE value as opposed to
> any other value) of LoopIfHalts, also changes the behavior of
> WillHalt.

No, LoopIfHalts will NOT change the behavior of WillHalt. WillHalt
has absolutely no idea if it is called from LoopIfHalts, run by
itself, or called from any other function.

> WillHalt could otherwise state that LoopIfHalts halts.

When WillHalt analyzes LoopIfHalts, it's just looking at the program.
It has no idea if that is the program that called it or not.

> In fact the only thing that is preventing WillHalt from determining
> that LoopIfHalts halts, is the fact that WillHalt is reuting the result
> to LoopIfHalts.

No, of course that's not the reason it can't tell if LoopIfHalts halts.

> Thus not returning a result even effects what the
> result itself is. If the result is NOT returned to LoopIfHalts, then
> the result is TRUE. If the result IS returned to LoopIfHalts,
> then TRUE can not be returned because it would be a wrong
> answer.

And if WillHalt could test this, then that might make some sense. But
since it can't determine that, this is completely wrong. 

-- 
That's News To Me!
newstome@comcast.net

Relevant Pages