W3C Technical Plenary Week | Agenda
I'm a member of the Web Services Description Working Group that works on WSDL 2. Together with Bijan Parsia, I'm working on the WSDL RDF mapping where we define a simple WSDL ontology along with the transformation from valid WSDL files into RDF.
Within this work, we are bumping into scope limits. For example, WSDL uses XML Schema, and we want to point to XML Schema element declarations and type definitions. However, XML Schema doesn't give us any RDF form for these things, and it's not in our scope to define an RDF mapping for schemas. We have similar issues with SOAP, too.
In this talk I suggest that the W3C should make RDF mapping a mandatory part of every new language work. We already got used to having XML schemas for all languages, and what I'm proposing is much easier than creating XML schemas.
Our experience from WSDL indicates that creating such RDF mappings need not be complex work.
In the XML schemas that we are already producing, we rightfully try to encode all the restrictions of the language, so that then we can do validation. A lot of ontology work is describing language constraints, but here that would be redundant, because I'm talking about RDF mappings, which turn valid documents into RDF. The resulting data will automatically conform to the constraints of the language.
An RDF mapping is a simple task with little room for introducing new issues to the working groups. It can, however, uncover existing issues, but that's a good thing.
With RDF mapping for every W3C language, we would have an RDF representation for all W3C-compliant data, and we could put it together, and we would be closer to the Semantic Web.
As a last remark, I hope the microformats people will also produce some GRDDL transformations. Unfortunately, I understand it's a relatively bigger piece of work in their context.