RE: layering (5.3, 5.10): Sardinia compromise?

From: "Massimo Marchiori" <massimo@w3.org>
Subject: RE: layering (5.3, 5.10): Sardinia compromise?
Date: Thu, 20 Jun 2002 19:46:53 +0200

[...]

> Precisely so. When studying years ago the possible kind of semantical
> extensions (in the field of RDF-logic), I've met a variety of options,
> and explored the most liberal ones too. And the most liberal one
> (well, one of the most liberals) is just to use different domain
> of interpretations, and establish some very weak semantic link.

This is a rather unusual meaning for ``domain of interpretations''.
Basically what you appear to be proposing is that the semantics of OWL has
no necessary relationship to the semantics of RDF(S).

> The reason is, essentially, if X is an extension of RDF then:
> X-aware-processors will treat X-extensions as they like, and
> rdf-aware processors will treat X-extensions just as plain RDF.

This assumes that the syntax of X is the same as the syntax of RDF.  Most
extensions do not work this way - think of propositional logic to modal
logic or first-order logic to higher-order logic.

> This is the bottom-line. Then, the more semantic links you put
> between the domain of interpretation of X and the RDF one
> (or seen differently, between their two logics), the better,
> but that's just a plus.

I guess that by ``semantic links'' you mean any relationship between the
two semantics, from similarity of semantic structures to entailment
similarity.  

> That's why I've never digested the "extending RDF is impossible"
> leit-motive, because it just starts from very tight assumptions.

Well, I'm not the source of these assumptions.  I've argued against them at
length.  Feel free to chime in at any time.  :-)

> Now, quickly back on
> http://lists.w3.org/Archives/Public/www-webont-wg/2002Jun/0101.html
> Writing
> "RDF defines a large set of RDF graphs; for OWL we syntactically define a
> subset W (or alternatively its complement). OWL then only applies to graphs
> in W."
> is just a particular case of saying you're considering two different
> domains of interpretation. 

I don't see how this follows.  The stuff you include just says that not all
RDF graphs form OWL KBs.  It does not say anything about the relationship
between RDF meaning and OWL meaning.

> Like said above, an OWL-processor is free to
> treat its extensions however it likes.

But the definition of OWL KBs above has the consequence that all OWL KBs
are RDF KBs, and thus there are no OWL extensions.  Of course, if OWL
meaning does not have to abide by RDF meaning, the OWL is free to give
whatever meaning it wants to OWL KBs.

[...]

> So, in a nutshell: if in this OWL extension chain, we have a problem,
> it might just be RDFS and not RDF. Therefore, it would make sense
> to reshape RDFS, rather than touching RDF with dark triples.

Well, most of the problems do indeed have to do with RDF.  In particular,
RDF has the feature that all KBs are collections of triples of the form 
< URIref, URIref, URIref or literal >.  The RDF MT has the feature that all 
URIrefs denote objects in the domain of discourse.  The only added feature
from RDFS that could cause problems is the non-well-founded type hierarchy
that it imposes.

[...]

> I'll come back on the specific 101's approach at a later stage, but anyway
> I see it much better than dark-triples, as it embraces the different-domains
> idea
> rather than the dark-triples one. Just, beware of the "OWL extends RDFS"
> foundation,
> that might be the real source of all the trouble...

I don't think that RDFS is the problem, as detailed above.

> -M (likely really dense in this email, after re-reading, sorry).

peter

Received on Thursday, 20 June 2002 14:19:17 UTC