- From: Larry Hunter <Larry.Hunter@uchsc.edu>
- Date: Tue, 22 Aug 2006 10:28:03 -0600
- To: John Barkley <jbarkley@nist.gov>
- Cc: Ivan Herman <ivan@w3.org>, Matthias Samwald <samwald@gmx.at>, public-semweb-lifesci@w3.org
On Tue, 2006-08-22 at 10:05 -0400, John Barkley wrote: > I also think we should consider a recommended practice of using OWL Lite or > DL where possible (i.e., when the knowledge base can be expressed in no more > than DL). I don't want to get into a religious war here, but I think the advantages of OWL-DL are generally oversold. The crux of the above statement is what "can" be expressed in OWL-DL. In my experience, many complex knowledge modeling projects benefit from the use of metaclasses. For example, if the domain of a relationship is limited to several specific classes, it makes sense to model those classes as members of a particular metaclass (i.e., one that supports a particular slot type). There are gyrations one can do to avoid metaclasses (the GO/OBO folks have really been trying hard), but I think that these load the knowledge models with counter-intuitive structures for very little gain. The "non-computability" of OWL-Full is a worst-case problem that doesn't seem to me to have been a practical concern in many real-world cases. Larry
Received on Tuesday, 22 August 2006 16:28:39 UTC