Path: utzoo!attcan!uunet!husc6!mailrus!ames!sri-spam!sri-unix!quintus!ok
From: ok@quintus.UUCP (Richard A. O'Keefe)
Newsgroups: comp.lang.prolog
Subject: Re: determinism, once etc, Trilogy
Message-ID: <982@cresswell.quintus.UUCP>
Date: 16 May 88 04:24:28 GMT
References: <2316@ubc-cs.UUCP> <6960007@hpfclp.SDE.HP.COM>
Organization: Quintus Computer Systems, Mountain View, CA
Lines: 19

In article <6960007@hpfclp.SDE.HP.COM>, fritz@hpfclp.SDE.HP.COM (Gary Fritz) writes:
> Obviously, if you know early in a predicate that a cut will be required,
> THAT is the place to insert the cut.  Such a predicate is not necessarily
> a good candidate for a deterministic declaration, if you use that as a
> substitute for the earlier cut.  If, however, you know that the predicate
> can only generate one solution, it may make sense to declare it deterministic
> _in addition to_ the earlier cut, so that OTHER predicates may call it
> more efficiently.
> 
> Gary

Agreed, it _may_ make sense.  It's highly implementation dependent.
On a WAM-based system like Quintus Prolog, SICStus Prolog, or SB-Prolog,
it would not be a good idea.  If the compiler were smart enough to
figure out that the additional cuts weren't necessary, then it would be
a good idea, but then you wouldn't _need_ the :- deterministic
declaration (though it would remain as a valuable declaration of intent).
(Functional dependencies such as Kamran Parsaye suggested years ago would
be even better.)