Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site sbcs.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!mhuxn!mhuxr!ulysses!unc!mcnc!decvax!genrad!mit-eddie!think!harvard!seismo!cmcl2!philabs!sbcs!debray
From: debray@sbcs.UUCP (Saumya Debray)
Newsgroups: net.ai
Subject: Re: Re: AI and Turing's Thesis (asynch processes)
Message-ID: <324@sbcs.UUCP>
Date: Wed, 12-Jun-85 08:39:08 EDT
Article-I.D.: sbcs.324
Posted: Wed Jun 12 08:39:08 1985
Date-Received: Mon, 17-Jun-85 02:49:58 EDT
References: <113@nvuxf.UUCP> <388@linus.UUCP> <1647@psuvax1.UUCP> <185@yetti.UUCP>
Organization: Computer Science Dept, SUNY@Stony Brook
Lines: 22

> The only *easy* way I know of simulating an asynchronous process on a TM 
> requires the ASSUMPTION that all the events  in the asynchronous process
> can be clocked by a global clock.  (Imagine what a mess it would be if
> they couldn't be! How would we know how many states the process as a whole
> actually had?  How could we define temporal orderings between parts of the
> whole process which are not in direct communication? 

Individual processes in an asynchronous computation may have incomplete
knowledge about the states of other processes (e.g. their local clocks),
but this has nothing to do with a TM simulation of the computation: in any
particular asynchronous computation, events have to happen with respect to
a (possibly external) global clock, e.g. the one at Greenwich (*), and
there's no reason why this can't be handled by a NTM.

(*) Please, let's assume an inertial frame of reference.
-- 
Saumya Debray
SUNY at Stony Brook

	uucp: {allegra, hocsd, philabs, ogcvax} !sbcs!debray
	arpa: debray%suny-sb.csnet@csnet-relay.arpa
	CSNet: debray@sbcs.csnet