Re: Can a regular Turing Machine provide Protected Memory?

From: Keith Frayne (Unknown_at_nospam.thanks)
Date: 08/30/04 

Date: Mon, 30 Aug 2004 21:35:20 +0000 (UTC)

"Peter Olcott" <olcott@worldnet.att.net> wrote in message
news:EIlYc.270179$OB3.94947@bgtnsc05-news.ops.worldnet.att.net...
|
| "Nick Smallbone" <nick@nick8325.freeserve.co.uk> wrote in message
news:4131d798$0$22765$db0fefd9@news.zen.co.uk...
| >
| > "Peter Olcott" <olcott@worldnet.att.net> wrote in message
| > news:Oe2Yc.264727$OB3.21461@bgtnsc05-news.ops.worldnet.att.net...
|
| > But how does the Halt Analyzer know that its output is really going
to a CRT
| > screen? Let's say that instead of the normal transition
function/table we
| > have an extended one, which maps each (state, input) pair to a
(state,
| > output, direction of tape, character on screen) tuple instead of the
more
| > usual (state, output, direction of tape). What is there to stop
| > LOOP_IF_HALTS from 'simulating' the operation of your halt analyser
and
| > reading off the characters written to the screen from the transition
table?
| >
| >
| If this is framed as a condition under which my method would
| fail, then it would fail to meet the burden or proof of proving a
| negative. It would fail to meet this burden specifically because
| it fails to refute the case where the CRT is real.

I've been trying to make sense of this for about a week now.

What are you all talking about? Turing machines do not have CRTs, they
have tapes. Turing machines do not have protected memory, they have
tapes and states.

And what on earth is this 'proving a negative' business?

I'm sorry if these are stupid questions, but I can't make head nor tail
of what this is all about.

You started this thread Peter. What is it all about? 

--
Keith

Relevant Pages