[mercury-users] Re: Question regarding determinism

[Peter Schachte]

[I mailed Tom privately about this, thinking it wouldn't be of general
interest.  Seeing his reply, I now think it is.  I've also Cc'd it to
Mercury Developers, because DJ moved the discussion there.]


On Wed, 19 Aug 1998, Thomas Charles CONWAY wrote:

> > > moddat__add_assumption(Obj, Model, [a(Obj)|Model]) :-
> > > 	some [] (
> > > 		list__member(o(Obj, _), Model)
> > > 	).
> > 
> > I would have expected this to be the right code: 
> > 
> > 	moddat__add_assumption(Obj, Model, [a(Obj)|Model]) :-
> > 		some [X] list__member(o(Obj, X), Model).
> 
> That is equivalent to my code except for the singleton variable
> warning for X. Arguably, X isn't singleton since it's mentioned
> in the quantifier and in the goal. Perhaps this is a bug.

Definitely.  Variables appearing in quantifier lists have to be considered
for singleton warnings, otherwise you'll get singleton warnings for my code,
which shouldn't give a warning, and not for code like

	some [X] foo(1,Y)

which should give a warning.

But I'm still confused.  Why isn't `some [] X' always the same as just `X'? 

Come to think of it, aren't variables supposed to be quantified in their
closest enclosing scope?  In this case, the closest enclosing scope for the
anonymous variable is the list__member goal, so I wouldn't think the
explicit quantification would be necessary.  Why is it?


-Peter Schachte               | 640K should be enough for anyone!
mailto:pets at cs.mu.OZ.AU       |     -- Bill Gates, 1981 
http://www.cs.mu.oz.au/~pets/ | 
PGP: finger pets at 128.250.37.3 |