Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!munnari.oz.au!metro!macuni!mqccsunb!lhamey
From: lhamey@mqccsunb.mqcc.mq.oz.au (Len Hamey)
Newsgroups: comp.ai
Subject: Re: TM's (was Searle and Radical Translation)
Message-ID: <383@macuni.mqcc.mq.oz>
Date: 13 Aug 90 03:09:11 GMT
References: <614@ntpdvp1.UUCP> <5644@puggsly.cme.nist.gov> <620@ntpdvp1.UUCP>
Sender: news@macuni.mqcc.mq.oz
Organization: Macquarie University, Sydney
Lines: 22

This discussion of real computers vs TMs is becoming a little convoluted.
Am I missing something obvious or is Ken Preston?  What is it about
Turing Machines that says that they cannot be run for a long period
of time performing many computations just like real computers?  Clearly,
the "input tape" of the Turing Machine is an abstraction of various
components of a real computer.  Thus, the output terminal (or the output
port) may be viewed as a portion of the tape - writing to that portion
of the tape causes the written characters to appear on the screen.
SImilarly, the keyboard input corresponds to a portion of the tape
(if time is not of importance, then the entire input stream can be
collected and stored on tape in advance, else the input stream must be
considered to be 'creating' the end of the tape as the computation
progresses).
Note that all real computers have physically limited storage, hence
a mathematical limit on their states, so the actual compute engine
is a finite state automaton -- the only thing that may be of unlimited
size is the input stream and the output stream, and it is easy to see
how these two streams can be stored on the TM tape in an interleaved
form so that both streams may be infinite.

Len Hamey
len@mqcomp.mqcs.mq.oz.au