- From: Jim Hendler <hendler@cs.rpi.edu>
- Date: Mon, 12 Nov 2007 22:12:19 -0500
- To: OWL Working Group WG <public-owl-wg@w3.org>
In separate discussions with Bijan (via chat) and Boris, I've raised the issue that from a Web language usage point of view, the proliferation of typing terms is a problem (the size of the OWL language will now be twice what it was, which is a way to discourage usage IMO). - A couple of ideas have come up. It appears to be that if we could either have a way to enumerate all the classes and properties in an ontology (both local and remote) or if we could do some language design to simplify things -- for example, to make this one feature in the RDF with some keys -- ie. some thing like having one Datayping property construct to which the many different types are key attiributes or something like that. Boris and I also agreed that we could not think of any reason why anyone would ever want a datatype and an objecttype property with the same name -- so if we added to the OWL namespace document that these two kinds of properties are disjoint, then we would be able to rule out the case of a consistent ontology ever having this done, which would therefore make parsing/processing easier, and we might not need so many property types. (apologies to Bijan and Boris for any misquoting I've done in explaining what I thought we said) I have created a wiki page for this at "PropertyProliferationControl -JH
Received on Tuesday, 13 November 2007 03:12:46 UTC