Since when has DC had complex, context-specific machine processable
constraints to represent?
If and when we run into this, RDFS was designed to be incrementally
extensible: Schema can quite properly include additional
constraint information, expressed in extension languages such as DAML+OIL
or W3C's planned Web Ontology language.
Dan
On Thu, 18 Oct 2001, Carl Lagoze wrote:
> I'm uncomfortable with the notion of RDFS as the "canonical schema".
> Usually the notion of canonical implies full information - e.g., there
> are other forms of expression that may be lossy but the canonical one is
> most expressive. RDFS, however, is not a fully expressive schema
> language, limited to types of relationships and not fully expressing
> data constraints. We need to consider whether this is relevant before
> canonicalizing the RDFS.
>
> In sum, I agree with Aaron (and Andy, Simon, etc.)
>
> Carl
>
> > -----Original Message-----
> > From: Thomas Baker [mailto:[log in to unmask]]
> > Sent: Thursday, October 18, 2001 6:02 AM
> > To: [log in to unmask]
> > Subject: Re: DC-ARCHITECTURE Digest - 16 Oct 2001 to 17 Oct 2001
> > (#2001-55 )
> >
> >
> > On Thu, 18 Oct 2001, Rachel Heery wrote:
> > > What I am curious to explore is how the various schemas
> > would be kept 'in
> > > synch' in this scenario?? will they all have a separate DC
> > Usage Board
> > > 'approval and amendment' track? or will they all be derived from one
> > > canonical schema??
> > >
> > > I would suggest the latter, that we need a canonical schema and this
> > > would best be expressed in RDFS at present.
> >
> > I agree with this, especially from a Usage Board point of view.
> >
> > Tom
> >
> > ______________________________________________________________
> > _________________
> > Dr. Thomas Baker
> > [log in to unmask]
> > Birlinghoven Library, Fraunhofer-Gesellschaft
> > Schloss Birlinghoven
> > +49-2241-14-2352
> > 53754 Sankt Augustin, Germany
> > fax +49-2241-14-2619
> >
>
|