- From: Garret Wilson <garret@globalmentor.com>
- Date: Tue, 24 Jul 2007 12:56:19 -0700
- To: Bruce D'Arcus <bdarcus@gmail.com>
- CC: Harry Halpin <hhalpin@ibiblio.org>, Semantic Web <semantic-web@w3.org>
Bruce, (I hope you don't mind me moving this to the semantic-web list.) Bruce D'Arcus wrote: > I've not worked with them, but my understanding was if you define the > domain of some property as some class, and the rdf:type is not > present, it gets added. I hope not. If so, I'd prefer not to specify the domain of the properties. If someone can use an existing property with known semantics in a different context, I would encourage that. I'd hate for completely new properties to be created just because someone feared type inference. > Whether it's a bnode or not seems orthogonal. If inference occurs, then yes, explicit typing would be unnecessary. Of course we all know how well that default attribute values have worked with XML, since everybody ships their DTDs around with their XML documents... ;) > A more tricky issue is the VCard domain. If I have: > > <http://ex.net> a foaf:Organization ; > vcard:fn "ABC Books" . > > ... is that then not implicitly saying <http://ex.net> is a > vcard:VCard in an addition to being a foaf:Organization? Again, I hope not---I'd prefer not to specify the domain of that is the case. In fact, following a TBL comment I saw in one of the archived notes, the usefulness of a v:VCard class in the context of RDF is suspect, as it's really the *person* or the *organization* that's being described. Garret P.S. What would you say to my changing the "v:" prefix to "vcard:" in the next spec revision?
Received on Tuesday, 24 July 2007 19:56:37 UTC