Re: basic decisions underlying DAML-ONT (defined classes)

Frank van Harmelen wrote:
> 
> "Peter F. Patel-Schneider" wrote:
> > > >
> > > > Why not create a new kind of object, perhaps called a description, which
> > > > would ``contain'' restrictions and qualifications.   The meaning of a
> > > > description would be the intersection of these restrictions and
> > > > qualifications.  A class could then be defined as ``equivalentTo'' to a
> > > > description.
> 
> From: Dan Connolly <connolly@w3.org>
> > > Yes, that's the basic idea. It's called reification
> > > in the RDF specs; it's called quoting elsewhere.
> 
> "Peter F. Patel-Schneider" wrote:
> > I don't think that I am asking for reification or quoting, just something like:
> > [... Details of Peter's proposal deleted, FvH]
> 
> Maybe I'm missing something here, but can this not already be done in current DAML?
> The crucial point is in the following qoute from an earlier off-line email exchange with Dan:
> (excuse the now obsolete list-syntax):
> 
> > In DAML, equivalent isn't a relation between two
> > class *names*, but between two *classes*. So of course we can say:
> >
> >  <Class ID="human">
> >    <equivalentTo>
> >      <Class>
> >        <UnionOf parseType="daml:collection">
> >          <Class about="#Man"/>
> >          <Class about="#Woman"/>
> >        </UnionOf>
> >      </Class>
> >    </equivalentTo>
> >  </Class>
> 
> [a] This does give you defined classes, doesn't it?
>     (it gives nec. & suff. conditions for the class "human")

Well... it completely defines that particular class.
I'm not sure everything that can be expressed as
a defined class in OIL can be expressed this way.
If so, then I guess we don't need any quoting/reification
stuff.

By the way... the semantics (that I understand)
for daml:equivalentTo mean that
the above is the same as writing:

 <Class ID="human">
        <UnionOf parseType="daml:collection">
          <Class about="#Man"/>
          <Class about="#Woman"/>
       </UnionOf>
      </Class>
 </Class>

> [b] This also shows that DAML-ONT's is not similar to SHOE's <DEF-RENAME> element,
> as Jeff Heflin suggested, since <DEF-RENAME> is a syntactic operation concerning class names, not a relation between two classes?

I don't think so. First, it's a little odd to say
(as I did) that equivalentTo is a relation between *two* classes,
since it asserts that the classes are identical; i.e.
there's just one class.

Second, in formal systems, identity comes down to the
sort of syntactic manipulations that Jeff H. is talking
about, no? i.e. the semantics of
	(= X Y)
is nothing more and nothing less than saying that if
we see
	(P X)
we can write
	(P Y)
and vice versa. (This is, of course, an informal
paraphrase of the substitution-of-equals-for-equals
inference rule that may, in particular formal
systems, take pages and pages to state precisely.)


-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/

Received on Saturday, 14 October 2000 21:53:20 UTC