Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!uwm.edu!zaphod.mps.ohio-state.edu!samsung!munnari.oz.au!goanna!ok
From: ok@goanna.cs.rmit.oz.au (Richard A. O'Keefe)
Newsgroups: comp.lang.prolog
Subject: Re: Prolog PROG TECH. II
Message-ID: <3647@goanna.cs.rmit.oz.au>
Date: 31 Aug 90 02:50:12 GMT
References: <ganNChe00WB2QF5VM0@andrew.cmu.edu> <3581@goanna.cs.rmit.oz.au> <126688@kean.ucs.mun.ca>
Organization: Comp Sci, RMIT, Melbourne, Australia
Lines: 41

In article <126688@kean.ucs.mun.ca>, jgarland@kean.ucs.mun.ca writes:
> In article <3581@goanna.cs.rmit.oz.au>, ok@goanna.cs.rmit.oz.au 
[I wrote]
> > 	run(done) :- !.
> > 	run(Event) :- /* Event ~= done */
> > 	    exec_event(Event, NextEvent),
> > 	    run(NextEvent).
> > 		
> > If exec_event/2 is determinate, this is determinate.
> > If exec_event/2 is not determinate, this is not determinate.
> > 
> I say again, that I agree PDC (Turbo) Prolog cannot directly do some of the 
> interesting things other Prologs do.  However, Tim's question revolved around 
> memory conservation, not metaprogramming.

Yes, but _I_ wasn't talking about meta-programming _either_.  I too was
talking about memory management.  I was saying that this two-clause
approach is determiniate and tail recursive (provided only that exec_event/2
is determinate) so that >>>it runs in bounded stack<<<.

> Nothing he has proposed in any of
> his inquiries remotely requires that level of sophistication.

What on earth is sophisticated about a two-clause tail-recursive loop?
This is one of the cases where thinking in Pascal _does_ pay off; it is
the direct transliteration of

	while Event <> done do
	    Event := exec_event(Event);

If this is sophisticated, then heaven help us all.

>  So, if the O'Keefe structure is of use to you, have no fear of using it.  If 
>  it isn't, remember that backtracking points are the thing to focus on...there
>  mustn't be any before the recursive call if you want to keep them from being 
>  multiplied. */

But that is _precisely_ what the comment about wanting exec_event/2 to be
determinate was _about_.
-- 
You can lie with statistics ... but not to a statistician.