Re: [LC Response #2] to Jan Wielemaker, Re: Triples and OWL2

Dear Ivan,

Thanks for the motivation. In the formal sence I'm `satisfied', though I
still do not like it.

If we proceed in this direction, we will end up with a growing number of
XML applications to cover the semantic web. To me, that is much worse
than poor/impossible handling of RDF/XML by generic XML tools.

I would be much happier if RDF/XML was, if necessary, extended (e.g, by
introducing additional parseType declarations) to support OLW-2 and the
XML tool issue was fixed at the level of RDF/XML (although reading the
TriX report gives the impression this might not be feasible).

Ideally, an OWL-2/XML syntax should be a sub-language of RDF/XML, i.e.,
force the use of certain RDF/XML constructs to describe OWL-2 classes
such that they become more canonical and possibly allow for processing
using generic XML tools.

 Regards --- Jan

On Wednesday 18 March 2009 07:28:27 pm Ivan Herman wrote:
> Dear Jan,
>
> This is an answer to your message
>  http://lists.w3.org/Archives/Public/public-owl-comments/2009Mar/0000.html
>
> giving further comments on the OWL 2 Web Ontology Language last call
> drafts.
>
> You are correct that a completely XML "friendly" encoding of RDF could
> indeed be used to encode OWL 2 ontologies and could, therefore, be used
> as part of a more complete XML workflow. There are, however, several
> issues that must be considered.
>
> First of all, as you yourself noted, developing such an XML encoding is
> not in the charter of the OWL 2 Working Group. Taking into account the
> fact that there is no group at the moment whose charter could reasonably
> include such a development, that means that, in the meantime, the needs
> of a distinct community would not be fulfilled for a long time.
>
> There is, however, a further issue to consider. Let us suppose that a
> regular XML encoding, closely reflecting RDF triples, was used
> (something like TriX[1], for example). That would mean that OWL
> construct would have to be encoded in, essentially, an XML
> transliteration of N-triples. Though this would be well defined, it
> would still be complicated to manage the resulting XML content through,
> say, XPath, and almost impossible to define an XML schema that could be
> used by a schema aware editor. This is simply due to the fact that the
> triple representation of OWL constructs are, by their very nature,
> fairly complex (think of the representation of class intersections using
> RDF lists). One could of course imagine a slightly more complex XML
> encoding of RDF, but it is unclear at the moment what that would be. In
> other words, relying on a generic XML format for RDF may not satisfiy
> the requirements end users have for such a serialization of OWL due to
> its inherent complexity.
>
> Note that having specialized formats for 'sub'-languages on the Semantic
> Web is not specific to OWL. A typical example might be the XML encoding
> of Resource Descriptions in POWDER[2], which provides an XML syntax for
> end users but also defines a formal transformation of that XML encoding
> into OWL. As long as these languages clearly map on a common and
> required exchange format (which is the case for OWL 2), they can be
> valuable in serving various specialized communities without damaging
> interoperability.
>
> Please acknowledge receipt of this email to
> <mailto:public-owl-comments@w3.org> (replying to this email should
> suffice). In your acknowledgment please let us know whether or not you
> are satisfied with the working group's response to your comment.
>
> Regards,
>
> Ivan Herman
> on behalf of the W3C OWL Working Group
>
>
> [1] http://www.hpl.hp.com/techreports/2004/HPL-2004-56.html
> [2] http://www.w3.org/TR/powder-primer/

Received on Thursday, 19 March 2009 14:46:54 UTC