Path: utzoo!utgpu!news-server.csri.toronto.edu!clyde.concordia.ca!uunet!samsung!zaphod.mps.ohio-state.edu!uwm.edu!bionet!arisia!quintus!dave
From: dave@quintus.UUCP (David Bowen)
Newsgroups: comp.lang.prolog
Subject: Re: taking a predicate name as an argument
Message-ID: <1353@quintus.UUCP>
Date: 15 Apr 90 01:05:01 GMT
References: <35047@brunix.UUCP> <2624D004.805@orion.oac.uci.edu> <28908@cup.portal.com>
Reply-To: dave@quintus.UUCP (David Bowen)
Organization: Quintus Computer Systems, Inc.
Lines: 27

In article <28908@cup.portal.com> pgl@cup.portal.com (Peter G Ludemann) writes:
>- Functor names can't be variables.  It would seem that this
>  could be overcome by using =../2:
>
>        question(Pred, X, Y, Answer) :-
>                F =.. [Pred, X, Y],
>                Answer is F .
>
>- But general terms aren't allowed with "is" (at least, this is
>  the case with Quintus 2.x; I assume it's true of others); if	
>  you try this, you will get a run-time error for is/2. 
>
>So, the answer would seem to be "it can't be done" in many
>popular Prologs; you'll just have to list out all the arithmetic
>operators yourself (there aren't too many, fortunately).

It can be done, like this:

        question(Pred, X, Y, Answer) :-
                 F =.. [Pred, X, Y],
                 call(Answer is F).

In future releases of Quintus Prolog it won't be necessary to put in the
call/1 wrapper.

If efficiency is of any concern, though, it would be better to have a 
clause for each arithmetic operator as Peter suggested.