Houghton,Andrew wrote:
> XML and XML schema are
> easily understood technologies compared with RDF/OWL.
Are they? This mail to XML-DEV summarising comments from the
XML Schema workshop would suggest otherwise:
http://lists.xml.org/archives/xml-dev/200507/msg00057.html
There perspective there is that XML Schema is difficult to
work with.
> RDF suffers from very obtuse syntax that allows you to specify
> the same thing by multiple means, thus adding to the confusion.
The syntax does get a lot of flack. Dan Brickley has written
some notes on why the syntax is the way it is [1]. Dave Beckett's
paper on the development of the revised RDF/XML syntax (the
specs were recently updated for those who last looked in several
years ago) is also worth a read [2].
> Why couldn't they just have
> provided three XML attributes that you could sprinkle on *any*
> XML schema or grammar in a similar fashion as the XLink
> specification did?
I'm not sure that XLink can support everything that RDF can.
Remember RDF is a data model, the syntax is secondary.
That side, you can harvest RDF from XLink [3]
> That would have been immensely simpler on
> the technology side allowing organizations to model their
> business problems in a way that made sense to them while
> providing the Semantic Web with a mechanism to extract semantic
> relationships by parsing *any* XML schema or grammar.
Schema annotations would be a good way to capture this mapping,
I believe there's work happening in that area.
> As it stands now the W3C wants you to recast your business
> problems in terms of RDF so they can extract value out of your
> metadata for the Semantic Web.
That's a little disingenuous. I don't believe the W3C are after
anyone to solve their problems for them. The issue is more: does
RDF (OWL) offer any benefits? Experiences seem to differ wildly.
It would be interesting to identify why, but I've rarely encountered
arguments against RDF that go beyond the syntax.
> BTW, I should point out that had the authors of the MARC-XML
> specification created an element based XML grammar, rather
> than a mixed element/attribute model, you could have dropped it
> directly into the RDF model with a simple RDFS schema. An
> example of a compatible MARC-XML like grammar, which I call
> MARC-RDF-XML, in its more verbose RDF form, e.g., without an
> RDFS schema is:
Yes. Here are some other guidelines to keep in mind when
making RDF friendly syntaxes [4].
Cheers,
L.
[1]. http://www.w3.org/2001/sw/Europe/200407/swintro/syntaxdesign.html
[2]. http://ilrt.org/people/cmdjb/2003/05/iswc/paper.html
[3]. http://www.w3.org/XML/2000/06/xlink2rdf.htm
[4]. http://www.xml.com/lpt/a/2002/10/30/rdf-friendly.html
|