- From: Sandro Hawke <sandro@w3.org>
- Date: Sun, 21 Apr 2002 13:06:18 -0400
- To: Drew McDermott <drew.mcdermott@yale.edu>
- cc: drager@bbn.com, www-rdf-logic@w3.org
> Is it supposed to be the case that just mentioning an ontology's > namespace means that the file uses that ontology? The namespace declaration is just syntactic sugar (it doesn't appear in the N-Triples generated, etc), but if we phrase your question differently, the point remains important: Does using a term defined by some ontology O in some RDF graph G mean that G is using O? I think a "yes" answer would be convenient but wrong in the long run. I think this is a special case of broader RDF question: Does using a term T in RDF graph G mean that the author of G believes and and all other RDF graphs which use T? (obviously not....) ... that the author of G believes some graph which "defines" T? (maybe, if we could figure out what "defining" meant; is that what you get when you try to GET from the namespace address?) My working answer is this: when you're looking at a graph G and you come across a term T... 1. Other graphs which contain T may have relevant information, but you can't assume the author of G knew about them or believes in them. If you have your own reasons to trust both them and the author of G, you should use the merged graph. 2. The author of G should be able to state a belief of some other graph, of which the ontology O is a special case. This is something like daml:includes, but more general. Something like "<http://address.of.some.RDF.content> rdf:type rx:TruePage." 3. EricP (and SOAP) keep saying they want "must-understand" messages, which is relevant here, because you might want to say "you must subscribe to O before looking at my message; if you don't then you'll understand my message incorrectly". I don't have a use case for this, but I do have solution, which is that you provide your graph in the form "if O then G", so the reader must understand both your "if...then" vocabulary and O before it even sees G. I know we're working in some different areas, but I hope this still made some sense. -- sandro
Received on Sunday, 21 April 2002 13:11:11 UTC