Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.theory > #12255 > unrolled thread

When will the establishment see the light?

Started byNewberry <newberryxy@gmail.com>
First post2017-10-16 21:55 -0700
Last post2017-10-17 21:04 -0500
Articles 3 — 3 participants

Back to article view | Back to comp.theory


Contents

  When will the establishment see the light? Newberry <newberryxy@gmail.com> - 2017-10-16 21:55 -0700
    Re: When will the establishment see the light? Ben Bacarisse <ben.usenet@bsb.me.uk> - 2017-10-17 16:40 +0100
    Re: When will the establishment see the light? "Shobe, Martin" <martin.shobe@yahoo.com> - 2017-10-17 21:04 -0500

#12255 — When will the establishment see the light?

FromNewberry <newberryxy@gmail.com>
Date2017-10-16 21:55 -0700
SubjectWhen will the establishment see the light?
Message-ID<os42gh$drf$1@dont-email.me>
When will the establishment see the light? I claim the existence of a 
sound and consistent partial halting decider that returns “yes” or “no” 
[or nothing], and is in a certain sense able to identify the cases it 
cannot decide. What is this “certain sense”? The Halting Theorem is 
usually proven by constructing a self-referential “counterexample”, a 
program that attempts to say that it itself does not halt. I claim the 
existence of a program that is able to determine that its own attempt to 
determine (that the counterexample does not halt) does not halt.

I do not know if anyone can actually deny the existence such a program 
because I have a simple example running. 
http://xnewberry.tripod.com/ProvableNotProvableNotHalt.c The question is 
if the program contradicts itself. I am suggesting that it does not.

https://www.researchgate.net/publication/316562253_Getting_around_the_Halting_Problem

[toc] | [next] | [standalone]


#12258

FromBen Bacarisse <ben.usenet@bsb.me.uk>
Date2017-10-17 16:40 +0100
Message-ID<87r2u13hr2.fsf@bsb.me.uk>
In reply to#12255
Newberry <newberryxy@gmail.com> writes:

> When will the establishment see the light?

When you publish a paper in journal read by "the establishment"
(provided the paper is, in fact, enlightening).  Whoever "the
establishment" are, they almost certainly don't read this crank-infested
corner of Usenet.

> I claim the existence of a sound and consistent partial halting
> decider that returns “yes” or “no” [or nothing], and is in a certain
> sense able to identify the cases it cannot decide.  What is this
> “certain sense”? The Halting Theorem is usually proven by constructing
> a self-referential “counterexample”, a program that attempts to say
> that it itself does not halt. I claim the existence of a program that
> is able to determine that its own attempt to determine (that the
> counterexample does not halt) does not halt.

See Rice's theorem.  If you have a counter-example, that will get
published like a shot, and you will be famous.

> I do not know if anyone can actually deny the existence such a program
> because I have a simple example
> running. http://xnewberry.tripod.com/ProvableNotProvableNotHalt.c The
> question is if the program contradicts itself. I am suggesting that it
> does not.

The existence of partial halting deciders is trivial.  The existence of
a program that can determine any non-trivial property of the language of
a TM (such as the one you describe as determining "a certain sense")
would be headline news.

-- 
Ben.

[toc] | [prev] | [next] | [standalone]


#12260

From"Shobe, Martin" <martin.shobe@yahoo.com>
Date2017-10-17 21:04 -0500
Message-ID<os6cs6$hqg$1@dont-email.me>
In reply to#12255
On 10/16/2017 11:55 PM, Newberry wrote:
> When will the establishment see the light? 

When you shine some light where the establishment can see it.

> I claim the existence of a 
> sound and consistent partial halting decider that returns “yes” or “no” 
> [or nothing], and is in a certain sense able to identify the cases it 
> cannot decide. What is this “certain sense”? The Halting Theorem is 
> usually proven by constructing a self-referential “counterexample”, a 
> program that attempts to say that it itself does not halt. I claim the 
> existence of a program that is able to determine that its own attempt to 
> determine (that the counterexample does not halt) does not halt.
> 
> I do not know if anyone can actually deny the existence such a program 
> because I have a simple example running. 
> http://xnewberry.tripod.com/ProvableNotProvableNotHalt.c The question is 
> if the program contradicts itself. I am suggesting that it does not.
> 
> https://www.researchgate.net/publication/316562253_Getting_around_the_Halting_Problem 
> 

Oh the errors.

1.1
	1)	Neither "sound" nor "consistent" are properties of programs.
	2)	Given the clarification, the two criteria for A(n,m) are the same.
	3)	You never get around to the A(n,m) claim.

1.2
	1)	Program A only sometimes enables you to determine when a computation 
does not halt. You, yourself admit that it doesn't later.

1.3
	1)	C_k(s) halts if and only if C_s() halts. That's how C_s() is defined.
	2)	There was never any requirement or expectation that C_k(s) would say 
something about C_s(). The only requirement is that they say the same thing.
	3)	C_s() halting implying that C_k(s) halts is not a contradiction. 
It's required by the definition of C_s().
	4)	That imagined contradiction is not a restatement of the Halting 
Theorem proof.
	5)	H expresses the premise that "C_s() halts" by halting and returning 
the value that indicates that C_s() will halt. That is, after all, the 
purpose of H.
	6)	C_s() halting does not help H express that "C_s() halts".
	7)	C_s() halting would not contradict itself.
	8)	The only way H() is allowed to indicate halting is by halting and 
returning the value that indicates halting.
	9)	The second "proof" is not a more verbose form of the first proof. 
The fourth line is wrong since it follows from C_k(s) not halting that 
C_s() also doesn't halt.
	10) 	The argument that eliminated option 1 was flawed.
	11)	The halting theorem does not say that A() cannot determine that 
C_s() does not halt.

1.4
	1)	The best way to think about H is that H is a partial decider.
	2)	Once again, the fate of R() is known. If T(r) computes nothing, then 
R() computes nothing.
	3)	Once again, C_s() halts -> C_k(s) halts is not a contradiction. 
C_s() is defined in such a way that C_s() halts <-> C_k(s) halts and you 
never establish otherwise.
	4)	You don't know whether or not it will halt. However, that doesn't 
change the fact that it either will halt at some point or not.
	5)	One possibility is to prove that it doesn't halt. In some cases, 
this is easily done.

1.5
	1)	This is not a restatement, you don't have as many requirements on H 
as you do in the body of the paper. As defined here, H is a partial 
decider. In the body of the paper, H is a partial decider
that always gets the C_n(m) halts cases correct.

1.6
	1)	The halting problem does not make a case for non-bivalent logic.

Martin Shobe

[toc] | [prev] | [standalone]


Back to top | Article view | comp.theory


csiph-web