- From: Seaborne, Andy <andy.seaborne@hp.com>
- Date: Wed, 9 Jun 2004 11:44:35 +0100
- To: Rob Shearer <Rob.Shearer@networkinference.com>, Dan Connolly <connolly@w3.org>, Kendall Clark <kendall@monkeyfist.com>
- Cc: Jim Hendler <hendler@cs.umd.edu>, RDF Data Access Working Group <public-rdf-dawg@w3.org>
-------- Original Message -------- > From: Rob Shearer <> > Date: 8 June 2004 23:42 > > > "The protocol should allow construction of notional RDF > > graphs inferred (e.g. using standardized semantics such > > as RDFS, OWL or emerging technologies such as SWRL or N3 > > rules) so that queries may be posed against the inferred > > knowledge base." > > I have a problem with that because it seems to imply that all knowlege > *about* an RDF graph can be encoded *within* an RDF graph, and that's > clearly not the case. (The charter similar slants things this way as > well, which I think is a mistake.) It's perfectly sensible to know, for > example, that an RDF graph must contain at least one of two possible > triples, but not which one. Are you allowed to use that information to > help answer queries about the graph? I think it would be a major mistake > to make the use of such information a violation of the spec. I don't read Dan's wording like this. In partcular, you talk about "all knowledge about an RDF graph". Assuming by "about" you mean "contained" ("about" suggest stuff like who authored it, when it was last changed etc - external assertions aboiut the graph as well), then I don't see why the wording implies "all knowledge", rather then some additional knowledge. Some extra knowledge, but not all possible knowledge, would meet the wording. Such "virtual graphs" don't have to be everything an inference system can derive - they are access to some of the extra information. Andy
Received on Wednesday, 9 June 2004 06:45:44 UTC