Best way to get started with XML/RDF/OWL

[Ray]

I have a Java application at work that essentially stores Enterprise
Architecture data in XML. In looking at the future we will want to
export this data into the Federal Enterprise Architecture reference
models that requires the use of OWL so that the relationships between
the elements and the elements themselves, etc are described. I have
been reading Lee Lacy's book on OWL, but am having trouble
understanding. Is there some sort of tutorial (e.g. like a cookbook)
out there that I could follow to go about generating an OWL file. Or is
there a better book someone can recommend out there? Thanks Ray

 

[Joe Kesselman]

Have you tried looking at the W3C's OWL hub at
http://www.w3.org/2004/OWL/
? That includes pointers to some tutorials and articles, as well as the
official documents.

(I don't know much about OWL, so I can't offer much help beyond that
right now. Personally I'm still skeptical about whether the Semantic Web
concepts will be scalable enough to be genuinely useful, but I'd be
delighted to be proven wrong.)

 

[Andy Dingley]

Is there some sort of tutorial (e.g. like a cookbook)
out there that I could follow to go about generating an OWL file.

Download a copy of Protege from http://protege.stanford.edu (best
ontology editor around), switch it to OWL-DL mode and have a hack
around with that. Lots of examples around, mailing lists etc

There's a useful tutorial (the pizza ontology) at
http://www.co-ode.org/resources/tutorials/ProtegeOWLTutorial.pdf
and the infamous "Ontology Development 101" paper (and its wine
ontology) is on the Protege site.

Don't lock yourself into OWL (and Description Logics) thinking though.
The Frames approach still has a lot to offer, particularly for learning
ontostuff. Protege is nice because it supports both.

You'll probably want to be coding in Java (practically all the good
tools are in Java) and a copy of Jena wouldn't hurt either. Something
the world really needs though are good docs on Jena examples. It's
certainly worth reading (and re-reading) the 2004+ RDF documentation
pack (W3 site) because you have to grasp the concepts of RDF pretty
well before you make any headway with the tools.

 

[Ray]

Thanks for the advice. I am a little familiar with Protege having
followed it's development. I was really pleased with the addition of
OWL to it and the tutorial looks like it will help quite a bit.

 

[Andy Dingley]

Thanks for the advice. I am a little familiar with Protege having
followed it's development.

In that case you should have no problem. The current Protege is a much
nicer interface too. Spend some time looking at their examples from
other users - there's some really impressive stuff happening out there.

I was really pleased with the addition of
OWL to it and the tutorial looks like it will help quite a bit.

OWL itself is easy, it's the mindset for description logics that will
hurt!

Right this minute I'm trying to use Protege to develop an ontology for
software release management and Jena to back-end my build process /
source control / bug tracking system. My current project has some very
awkward customer / version / customisation problems and making reliable
builds is a nightmare. The usual Bugzilla/Subversion/Ant combo is well
out of its depth.