[m-dev.] Tabling [1/3]

[Fergus Henderson]

On 12-Mar-1998, Oliver Hutchison <ohutch at students.cs.mu.oz.au> wrote:
> > I think there should be some way of saying that you want
> > memoized execution without changing the declaration semantics.
> > I think `pragma memo' should be that way.
> 
> Yes I think you are right. Perhaps I should change the name of the current
> nondet memoing to minimal_model? and reimplemient nondet memoing with
> respect to the completion semantics. 

Yep.

> > If there is some other useful semantics, e.g. minimal
> > model or well-founded model semantics,
> > then there should be different pragmas to ask for them.
> > For example, that's what `pragma minimal_model' is for.
> > 
> > If I recall correctly (a big if! ;-), SLG resolution implements
> > the well-founded model semantics, which (again if I recall correctly)
> > is an extension of minimal model semantics to programs with negation.
> > I thought we decided to use `pragma minimal_model' simply because
> > that was a slightly less frightening term than `pragma well_founded_model'.
> 
> Yes. But SLG-d is a subset of SLG supporting only definite programs.

Oh, I see.  So what you've implemented is exactly the minimal_model
semantics.  That is probably why we called it `pragma minimal_model' ;-)

Can you remind me what the difficulty in implementing full SLG
resolution is?  i.e. what would be required to implement
`pragma well_founded_model'?

> I think perhaps I have described what happens in the minimal model just
> called it the memo model :-(

Yep, well you certainly had me confused ;-)

-- 
Fergus Henderson <fjh at cs.mu.oz.au>  |  "I have always known that the pursuit
WWW: <http://www.cs.mu.oz.au/~fjh>  |  of excellence is a lethal habit"
PGP: finger fjh at 128.250.37.3        |     -- the last words of T. S. Garp.