[topicmapmail] Topic Maps, Description Logics and Semantic Networks

[Lars Marius Garshol]

* Murray Altheim
>
> I'm assuming you mean that FOL is a subset of DL? If you look at it
> that way there should be no confusion. By "subset" I simply meant
> that FOL is a component in the construction of a Description Logic.

Well, DL and FOL are different, but closely related, and it's easy to  
translate DL into FOL. However, DL is less expressive than FOL, so  
you in DL you can only say a subset of what you can say in FOL (hence  
my contracted "DL is a subset of FOL"). It's partly because of this  
that DL is interesting at all, since unlike FOL DL is decidable.

> One might also note that DL is not a single logic but a family of
> knowledge representation languages. It just happens that one flavour
> of DL ended up being talked about to the exclusion of others by the
> "Semantic Web" people (not to be confused with the "TV people"). To
> my knowledge there isn't *one* DL that is better than any other;
> they're all just tools to be used for specific purposes.

Yes.

> Or to put it another way, FOL is a fundamental and simple foundation,
> one of the tools upon which a DL is built. There are a lot of things
> in DL (i.e., a lot of things that can be expressed in DL) that go
> beyond what can be expressed FOL, [...]

It's the other way around. You can say more in FOL than in DL.

> I believe (and someone can correct me if I'm wrong) that anything  
> expressed in DL can potentially be deconstructed to an FOL expression,
> [...]

Translated, yes. (You seem to be contradicting yourself here, BTW.  
How could this be possible if DL were richer than FOL?)

--Lars M.