Hi All,
I wanted to give more "legs" to this discussion point that came up at
the end of today's HCLSIG TCon, as I think it's extremely important
for us users of RDF & OWL.
Sorry - I don't mean to be a "Chicken Little" on this issue - but as
someone who needs to make practical technical recommendations to a
large bioinformatics infrastructure project, this issue has me
concerned.
Here is the link to the OWL ED 2006 manuscripts:
http://owl-workshop.man.ac.uk/accepted06.shtml
I'd particularly recommend absorbing the details given in the
following, as I believe these do an excellent job summarizing the
justification being given by the developers of the OWL 1.1 spec for
this move away from RDF:
"Next Steps for OWL" (http://owl-workshop.man.ac.uk/acceptedLong/
submission_11.pdf)
"Problems with OWL Syntax" (http://owl-workshop.man.ac.uk/
acceptedLong/submission_13.pdf)
"Supporting Early Adoption of OWL 1.1 with Protege-OWL and FaCT+
+" (http://owl-workshop.man.ac.uk/acceptedLong/submission_15.pdf) -
describes the new RDF-free version of ProtegeOWL with v1.1 support
"The Manchester OWL Syntax" (http://owl-workshop.man.ac.uk/
acceptedLong/submission_9.pdf)
"Putting OWL in Order: Patterns for Sequences in OWL" (http://owl-
workshop.man.ac.uk/acceptedLong/submission_12.pdf)
These each present different aspects of the argument for moving ahead
with OWL 1.1 independent of the requirement for the entirety of the
normative spec to be expressible in RDF.
Just as a footnote, the DIG 2.0 API will include support for OWL 1.1
operators:
"Towards a Flexible Interface for Description Logic
Reasoners" (http://owl-workshop.man.ac.uk/acceptedLong/submission_3.pdf)
As Eric pointed out, there has been considerable discussion of this
move by members of the SemWeb community - e.g.:
http://lists.mindswap.org/pipermail/owl/2005-December/000097.html
http://www.mindswap.org/blog/2006/11/13/11-alternative/
There has also been some attempt by OWL developers to ease the
transition:
http://owl1_1.cs.manchester.ac.uk/rdf_mapping.html
As Eric said, some of the limitations, as opposed to be limits to the
expressivity of RDF per se, are based on constraints imposed by RDFS
and/or an RDF/XML representation - which as we all know is only ONE
WAY of expressing RDF triplets.
However, I'd point out, in the last document, where they describe a
mapping of OWL 1.1 to RDF, they make the following caveat:
Not every OWL 1.1 ontology can be serialized in RDF. In particular,
ontologies using the following features of OWL 1.1 cannot be serialized:
1. punning and
2. annotations on axioms.
My understanding is some of these constraints (and other listed in
the papers from the OWL ED meeting last week) are limited to RDFS,
but I don't believe that is true of all of the constraints that - as
they say - make some OWL 1.1 constructs inexpressible in RDF. I'm
far from an expert in the formal under-pinnings, however, so this is
a question for the experts to answer.
I would add, the latter is one of the most valuable additions to the
OWL 1.1 spec that runs right up against all the RDF/RDFS constraints
imposed on annotations in OWL. Those of us building ontologies in
OWL & RDF recognize the need in the scientific domain to track the
"evidence" for any ontological assertion. Many projects - SWAN & Bio-
Zen are good examples - are very much ABOUT formally tracking data
provenance and distilling it into evidence-supported ontological
assertions. I know from trying to use AnnotationProperties in the
BIRNLex ontology, OWL 1.0 has very serious problems, when it comes to
supporting the required links to "evidence". The GO folks know this
well, too, as "evidence codes" have always been at the heart of GO
based annotations. In fact, Chris Mungall and others have been
working toward developing a more generic means of providing this
capability to those both developing and using OBO Foundry-based
ontologies.
Anyway - I think this is an important topic for us to consider - one
for which - on this list - we have the ears of the experts - both in
the design and use of OWL & RDF - who can help us all get a clearer
sense of what the practical impact will be on our use of OWL & RDF.
I think Alan Ruttenberg and Jonathan Rees in particular - and many
others I'm certain - have some insight to offer based on the work
they've been doing.
Cheers,
Bill
Bill Bug
Senior Research Analyst/Ontological Engineer
Laboratory for Bioimaging & Anatomical Informatics
www.neuroterrain.org
Department of Neurobiology & Anatomy
Drexel University College of Medicine
2900 Queen Lane
Philadelphia, PA 19129
215 991 8430 (ph)
610 457 0443 (mobile)
215 843 9367 (fax)
Please Note: I now have a new email - [EMAIL PROTECTED]