[m-dev.] from [mercury-users] Re: Question regardi^[[Dng determinism

[Thomas Charles CONWAY]

Mark Anthony BROWN, you write:
> That piece of code should be described as a temporary workaround
> only.  I wouldn't relish having to explain to somebody why the
> empty quantification is necessary.

I don't disagree with you at all, but it is easiest to describe the
empty quantifier as a hint to the compiler that the goal doesn't have
any output bindings. Of course, this isn't quite true, since the goal
could have output bindings, and the quantifier would still be legal, but
writing an empty quantifier is nicer than writing:

	call(((pred) is semide :- list__member(o(Obj, _), Model))).

Thomas
-- 
Thomas Conway <conway at cs.mu.oz.au>
Nail here [] for new monitor.  )O+