RE: DCAP in UML/OWL/XSD

Jodi,

 

The point on UML is a bit of a tangent: IMO, UML domain models are a
variant of OWL (80-20 rule). If seeing UML class diagrams help people
visualize an ontology, great. If not, ignore it.

 

In general, though, I think that Michael Panzer and I hope to encourage
the use of OWL as the preferred abstract model. Even though OWL
semantics specify  an open-world assumption, this only becomes relevant
at the point where reasoning is applied. Until then, the application
model is the same regardless. Michael's brief demo of the Pellet
Integrity Constraint Validator (ICV) at the Joint Meeting seems to show
promise as a bridge between OWL and closed-world views.

 

http://clarkparsia.com/pellet/icv/ 

 

Another way to impose a closed-world view on OWL is to translate it
(i.e. the application model expressed in OWL) into XSD, which is the
route I took. I would argue that the OWL and XSD (and UML class diagram
for that matter) are just variant Web document representations of the
exact same application model. In other words, any one of these three
could serve as the datum from which the others could be generated at
runtime (80-20 rule).

 

http://alcme.oclc.org/dcap/UML.png

http://alcme.oclc.org/dcap/model.owl

http://alcme.oclc.org/dcap/model.xsd

 

So in effect, this route could be viewed as an XSD "Integrity Constraint
Validator" that is analogous to what Pellet ICV is doing.

 

Jeff

 

From: Jodi Schneider [mailto:jodi.schneider@deri.org] 
Sent: Tuesday, October 26, 2010 6:40 AM
To: Young,Jeff (OR)
Cc: public-lld
Subject: Re: DCAP in UML/OWL/XSD

 

Thanks for sharing this, Jeff. Is the point just to show the translation
itself, or are there key facts that you're pointing out about what's
lost/changed in translation? (Just trying to make sure I'm not missing
something.)

 

-Jodi

 

On 25 Oct 2010, at 21:42, Young,Jeff (OR) wrote:





A zip file is attached with an example translation of an OWL ontology
into XSD to check closed-world assumptions. This was the essence of what
I talked about at the Joint Meeting:

 

http://www.w3.org/2001/sw/wiki/JointMeeting2010#Part_1_.2814:00-15:30.29
:_Review_of_DCMI_Abstract_Model.2C_brainstorming_on_requirements

 

The key documents are also browsable on the Web:

 

http://alcme.oclc.org/dcap/

 

I kept the XSD focused on a single individual of type Work, but in
principle XSDs could be built that systematically bundled more of the
surrounding individuals.

 

I think that Michael's demonstration of Pellet using SPARQL queries to
do validation is more interesting because it can be applied beyond RDF
document ingest.

 

Comments, questions and discussion are welcome.

 

Jeff

 

---

Jeffrey A. Young
Software Architect
OCLC Research, Mail Code 410
OCLC Online Computer Library Center, Inc.
6565 Kilgour Place
Dublin, OH 43017-3395
www.oclc.org

Voice: 614-764-4342
Voice: 800-848-5878, ext. 4342
Fax: 614-718-7477
Email: jyoung@oclc.org

 

<dcap.zip>