Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!ucbvax!pasteur!ames!haven!purdue!decwrl!sun!pitstop!sundc!seismo!uunet!munnari!murtoa.cs.mu.oz.au!djk@murtoa
From: djk@murtoa (David Keegel)
Newsgroups: comp.lang.prolog
Subject: Re: NU-Prolog times/3 (was: IC-Prolog)
Message-ID: <1208@murtoa.cs.mu.oz.au>
Date: 6 Feb 89 09:34:15 GMT
References: <187@aucsv.UUCP>
Sender: news@cs.mu.oz.au
Reply-To: djk@murtoa
Distribution: comp
Lines: 17

From article <187@aucsv.UUCP>, by ok@aucsv.UUCP (Richard Okeefe):
] ... Note that
] 	times(X, 0, 0)
] has *infinitely* many solutions.  Such a thing in NU-Prolog would have to
] wait until someone else comes up with a binding for X.  Even for
] 	times(X, Y, 720)
] --720 being a rather small number-- the number of solutions is large.

The current implementation of NU-Prolog gives a division by zero error
and fails for ``times(X, 0, 0).''

Also, ``times(X, Y, 720)'' will delay until one of X or Y is instatiated.
(Rough translation: "Figure out the value of one of those variables, _then_
we'll talk.")  This is presumably for ease of implementation.

                        David Keegel    (djk@munnari.oz)
   "Flattery will get you nowhere, unless someone else does it to you"