Re: Daryl, Dave, Barb and the Georges' source of Confusion

From: Kenneth Doyle (nobody_at_notmail.com)
Date: 07/06/04 

Date: Tue, 06 Jul 2004 01:13:55 GMT

daryl@atc-nycorp.com (Daryl McCullough) wrote in
news:ccbf0i02uu0@drn.newsguy.com:

> This is one of those times where you are being misled by the
> informal treatment. For Turing Machines, there is no such thing
> as a "function call" and there is no interprocess communication,
> and there are no inputs (other than the parameters fed to the
> program initially) so the counterexample program doesn't
> "wait for inputs". Instead, the counterexample program is
> constructed using the *source code* of the candidate program
> (that is, the program that is purported to solve the halting
> problem).

That was my point when I wrote the program that calls his program. I first
wrote it to use an input file (mytext.txt) with an "N" in it (which is
"cheating"). Then I call his program with the text file as input
(hisprogram.exe < mytext.txt). But then, I call hisprogram.exe <
hisprogram.cpp (his program with the source of his program as input). I
cheated to show that the mechanism works (if you see what I mean), then I
called his program with its own source to demonstrate that his program is
constructed in such a way that it misrepresents the original situation by
calling its own source an "incorrect answer" (which seems to mean
"incorrect input", since we don't see "incorrect answer" when the answer
actually is incorrect). (I'm saying this more for Peter's benefit than for
your information.) So not only does his program fail to do what he wanted
it to do but it also fails to represent the original proof. Perhaps both
failures are really only one? 

-- 
CodeCutter - good, fast and cheap; pick two.

Relevant Pages