Re: decisions at recent RDF Core WG F2F

From: Brian McBride <bwm@hplb.hpl.hp.com>
Subject: Re: decisions at recent RDF Core WG F2F
Date: Tue, 16 Jul 2002 15:46:22 +0100

> At 18:36 15/07/2002 -0400, Peter F. Patel-Schneider wrote:

[...]

> >It appears that the RDF Core WG is edging towards a two-semantics
> >approach.  This approach has caused considerable problems for DAML+OIL.
> >Perhaps it would be a good idea to solicit advice here.
> 
> I take it you are referring to describing the semantics of RDF(S) in 
> LBase.  The current position is that the semantics of RDF(S) will be 
> defined by the model theory.  The semantics document will contain a 
> non-normative appendix describing the semantics in terms of LBase on the 
> grounds that this will be easier for those less familiar with model theory 
> to follow.
> 
> LBase is a proposal for specifying the semantics of several semantic web 
> languages.  I have argued that the suitability of this approach must be 
> discussed in a wider forum than RDFCore and I hope such a forum will be 
> found/created soon.  I have argued this in part so that you will be able to 
> influence that decision.
> 
> Regarding advice, I hope Peter, that you don't feel you need of an 
> invitation to warn us when you think we may be making a misstep.  If you 
> wish to offer some observations, or suggest someone we should approach for 
> input, please feel free at all times to do so.
> 
> And in this particular case, what is it that you feel we ought to know?

[...]

> Brian.

Two semantics leads to the definite possibility of conflicts.  This has
happened in DAML+OIL, with negative consequences.

Having one semantics be normative and the other non-normative is better,
but even better would be to not include the non-normative semantics in any
working group documents.  If LBase is to become part of the semantic web,
this should be discussed in a different forum than the RDF Core WG.

By the way, there are a number of other issues concerning the Semantic Web,
including layering, that deserve to be discussed in detail within W3C but
that do not fit within any of the W3C working groups.  Some of these issues
are currently adversely affecting work in the WebOnt working group.

Peter F. Patel-Schneider
Bell Labs Research

Received on Tuesday, 16 July 2002 11:02:09 UTC