- From: Xiaoshu Wang <wangxiao@musc.edu>
- Date: Mon, 05 Feb 2007 23:35:43 -0500
- To: Harry Halpin <hhalpin@ibiblio.org>
- CC: "'Semantic Web'" <semantic-web@w3.org>, Brian Suda <brian.suda@gmail.com>, Norman Walsh <ndw@nwalsh.com>
Harry Halpin wrote: > In detail, these are the only parts of OWL used are cardinality > constraints [1], while in the original vCard/RDF[2] it was implied that > not having these sort of constraints was not a bug, but a feature: > > So, do we need these? Or do they complicate things? If we have > cardinality constraints they allow us "round-tripping" without loss of > data, but if don't have them them people can have much more flexible > names and organizations and the like, and keep vCard in RDF/S as opposed > to OWL. > Why not create two ontologies, with the second one (with cardinality constraints) further constraining the first one (those without)? So, when people use the vCard, they can use the rdfs:isDefinedBy to suggest to an RDF engine if additional constraints should be retrieved or not? I am a firm believer of ontology modularization. I think we should avoid build large monolithic ontology at all cost. We should build ontologies on small scales. Doing so will help users to pick and mix also it will help ontology's graceful evolution. Xiaoshu
Received on Tuesday, 6 February 2007 04:39:20 UTC