RE: ACTION 2001-11-02#02: Datatyping use-cases from CC/PP

At 03:29 PM 11/14/01 +0200, Patrick.Stickler@nokia.com wrote:
>No. But I'd hope that we *are* the right WG to define RDF Schema to
>say things like 'an instance of the person class must have two
>and only two values for the property x:parent and the values of
>those properties may not be identitical (share the same URI).

Not, as far as I am aware:  the DAML+OIL and Web ontology groups are 
building those structures on top of RDF.

>Right? Otherwise, we are left with *no* means whatsoever to
>define constraints about manditory knowledge for specific
>applications. XML Schema cannot define constraints for knowledge
>bases build from the syndication of multiple instances, so
>the constraint functionality offered by RDFS is critical.

Yes: no such means within core RDF(S).

>Am I really understanding you correctly as saying that RDFS
>no longer defines any constraints whatsoever?

That's my understanding.

>Where then does the responsibility get shifted to? DAML+OIL?
>The WebOnt WG deliverables?

That's my expectation.

..

You (and others) might argue that this means RDF(S) is only "half a 
solution".  For some purposes, that may be true, but I strongly believe 
that for many other applications it is enough.  The availability of a 
simple, consistent language for applications to express ground facts is a 
very powerful tool.  Its's simplicity (from a developer perspective) will 
be, in my view, a key factor in getting lots of raw metadata out there on 
the web for more advanced reasoning engines to chew upon and generally work 
their magic.  The logical consistency of this simple language is crucial to 
ensuring that more powerful generic layers (DAML_OIL, WebOnt and more) can 
be built on this foundation.

Here are some notes I used to present RDF to our development team:
http://public.research.mimesweeper.com/RDF/RDFMetadataForEndToEndContent.html

#g


------------------------------------------------------------
Graham Klyne                    MIMEsweeper Group
Strategic Research              <http://www.mimesweeper.com>
<Graham.Klyne@MIMEsweeper.com>
       __
      /\ \
     /  \ \
    / /\ \ \
   / / /\ \ \
  / / /__\_\ \
/ / /________\
\/___________/

Received on Wednesday, 14 November 2001 13:01:34 UTC