Re: Can a regular Turing Machine provide Protected Memory?

From: Simon G Best (s.g.best_at_btopenworld.com)
Date: 08/29/04 

Date: Sun, 29 Aug 2004 21:42:53 +0000 (UTC)

Peter Olcott wrote:
> "Simon G Best" <s.g.best@btopenworld.com> wrote in message news:4130BFAD.7020501@btopenworld.com...
>
>>If Halts is a TM, then it can easily be used as part of another TM, and
>>Halts just cannot detect that this is happening. If, instead, it's a
>>program to be executed by a TM, then /that/ TM, with Halts as a program,
>>can easily be used as part of another TM. Again, there is /no way/ that
>>Halts, or the TM that Halts runs as a program on, can detect the fact
>>that they're being used within another TM.
>
> How does that defeat my process that shows that this is not true?
> Halt is run on another TM that provides UTM services to check
> whether or not there is any state transition out of the final state of
> Halt. If there are no transitions out of the final state of Halt, then
> Halt does know that it is not being called by another TM.

Halt is, indeed, the only thing that runs on your 'UTM'. It runs on
your 'UTM' as Halt, /not/ as part of anything else. Halt is left
completely unchanged. And your 'UTM' /is/ used as part of another TM.

When Halt checks with your 'UTM' to see whether or not there are any
transitions out of what is supposed to be Halt's final state, your
'UTM', which is /only/ running /your/ *unchanged* Halt confirms to Halt
that there are no such transitions. Halt runs *exactly* the same as it
does when running on your 'UTM' on it's own, /because it *is* running on
your 'UTM' on it's own/.

Your 'UTM' just happens to be being used as part of my TM, which may
well be Loopy Faults :-)

Simon

Relevant Pages