Re: Media types for the Semantic Web

From: Mark Baker <distobj@acm.org>
Subject: Re: Media types for the Semantic Web
Date: Wed, 31 Mar 2004 13:30:26 -0500

> On Wed, Mar 31, 2004 at 12:34:35PM -0500, Peter F. Patel-Schneider wrote:

[...]

> > > This seemed to be Dan's concern;
> > > 
> > > ""
> > > Now further up the spectrum, we might consider application/owl+xml.
> > > 
> > > I find that objectionable because it suggests that
> > > dublin core and adobe XMP and RSS and so on
> > > need their own media types, and it leaves me
> > > wondering what media type to use if for
> > > a document that mixes all these vocabularies
> > > together.
> > > ""
> > >   -- http://lists.w3.org/Archives/Public/www-webont-wg/2002Oct/0162.html
> > 
> > Well, what is wrong with this?  Well maybe it is an indictment of using
> > media types instead of some other mechanism, but why otherwise?
> 
> Deploying new media types is currently expensive; if you have the option
> between communicating some information using application/rdf+xml,
> instead of with application/prs.pps.foobar+rdf+xml, you should, in
> general, opt for the former (unless you know the recipient can handle
> it, of course, but obtaining that information doesn't scale).

Well, sure, if you have the option.  However, what if you don't?  

The example that I used before

 	ex:foo owl:sameAs ex:bar .	

has a different meaning when treated as an OWL ontology than when treated
as an RDF ontology.   Sending it as application/rdf+xml means, to me, that
you are sending as a triple with no special meaning attached to owl:sameAs,
whereas if it is sent as OWL (by whatever means you like) then the triple
means that ex:foo and ex:bar have the same denotation.

Actually I would much prefer to send it (or a close variation of it) as
the OWL ontology

	Ontology(
		Class(ex:foo complete ex:bar)
	)

which, to me, is much preferable in every way.

> I can forsee the day when we have dereferenceable media types so that an
> agent can, upon seeing an unrecognized one, dereference it to learn that
> it's just a specialization of some type it already knows (e.g. in terms
> of application/rdf+xml) with its own axiomatic triples.  But until
> that's common practice, I think it's best to avoid minting new ones.

Well, that assumes that you have some powerful logic backing up the
communication, which is certainly not the case with RDF.   Even if you do,
I would not advocate this approach, instead retaining an approach where the
formalism in which the communication is to be interpreted is sent by
reference and is not specified as a theory in this powerful logic.

> Mark.
> -- 
> Mark Baker.   Ottawa, Ontario, CANADA.        http://www.markbaker.ca

Peter F. Patel-Schneider
Bell Labs Research

Received on Wednesday, 31 March 2004 13:46:26 UTC