- From: Jacco van Ossenbruggen <Jacco.van.Ossenbruggen@cwi.nl>
- Date: Thu, 04 May 2006 16:17:51 +0200
- To: "Jeff Z. Pan" <jpan@csd.abdn.ac.uk>
- CC: swbp <public-swbp-wg@w3.org>, Mark van Assem <mark@cs.vu.nl>
Jeff, Mark, I've revised Jeff's example in [1]. I think the previous example suffered from the typical thing that many of the points the original example wanted to make did not survive the (too) many rounds of simplification by (too) many different people. It was also suggesting that the DC-based approach is different from the ontology based approach, which is a bit unfair because DC actually recommends using values from controlled vocabulary and is organizing their own properties more and more in a nice hierarchy. I also did not like that the first example was an instance level annotation and the other a schema level class definition. I also doubt that many readers will understand and/or appreciate the subtle differences between the example OWL class definition and a similar definition in UML or Java (with the risk of them thinking: oh now, yet another language for OO modeling). So instead, I put in two lines of RDF based on the Wordnet vocabulary, to stress the difference between annotating using string values versus RDF -defined concepts such as WordNet's. Jeff, could you check if you still agree with the whole section. Mark, I've used dc:subject to refer to two wordnet concepts: <dc:subject rdf:resource="http://www.w3.org/2006/03/wn/wn20/instances/synset-Indian_elephant-noun-1"/> <dc:subject rdf:resource="http://www.w3.org/2006/03/wn/wn20/instances/synset-Ganesh-noun-1"/> Have I used the right URIs? Other comments? Thanks, Jacco [1] http://www.w3.org/2001/sw/BestPractices/MM/image_annotation.html#semweb_intro
Received on Thursday, 4 May 2006 14:18:11 UTC