Re: [Yap-users] Strange tabling behavior with dif/2

[Eyal D. <eya...@gm...>]

Why do you need this diff/1 at all?

Why not?

:- table same/2.

same(1, 2).
same(1, 3).

same(4, 5).
same(4, 6).

same(X, Y) :- same(Y, X).
same(X, Z) :- same(X, Y), same(Y, Z).

you can exclude same(X, X) in a new predicate like:

same1(X, Y) :- same(X, Y), X \= Y.

Eyal




On Tue, Feb 16, 2016 at 2:05 PM, Edward Schwartz <esc...@ce...> wrote:

> Hi Eyal,
>
> Thank you -- this seems to explain the behavior I'm seeing.  Do you have
> any suggested alternative?  A version of dif/2 that doesn't use
> co-routining maybe?  Does such a thing exist?
>
> Thanks,
>
> Ed
>
> On 02/16/2016 01:29 PM, Eyal Dechter wrote:
> > Hi Ed,
> >
> > I think that this is because dif/2 is part of the co-routining extension
> > and the tabling extension (YAPTab) doesn't support the co-routining
> > extension. According to the manual (section 18) this should through an
> > error; that it doesn't is likely a bug.
> >
> > I think the reason it is not supported is because tabling in YAP depends
> > on being able to ask when one term is a variant of another. But in the
> > presence of constraints like dif/2, two terms that are syntactic
> > variants are no longer necessarily unifiable.
> >
> > Best,
> > Eyal
>