- From: Rob Shearer <Rob.Shearer@networkinference.com>
- Date: Mon, 12 Apr 2004 10:51:52 -0700
- To: "Jos De_Roo" <jos.deroo@agfa.com>
- Cc: "Dan Connolly" <connolly@w3.org>, "RDF Data Access Working Group" <public-rdf-dawg@w3.org>
> Rob, I'm also afraid that I don't follow... > Could you maybe formulate your original use case > as a test case > > given following triples: > ... This I can do: :Al rdf:type Employee :Carl rdf:type Employee :Bob rdf:type Employee :Al :hasGroup :Management :Carl :hasGroup :Engineering :Bob :hasLocation :SanFrancisco I'll even give you some OWL used to supplement this data: Employee => Exists hasGroup (Management OR Support OR Engineering) SanFrancisco is a member of the "NoEngineering" class (Employee AND Exists hasLocation NoEngineering) => NOT (Exists hasGroup Engineering) > asking following triples: > ... I'm afraid not. A use case shouldn't include one arbitrary guess at its conclusion. There is no query language for RDF. I can tell you what the user wants to do: he wants the names of all employees who are in either management or support. That's all he knows. He's not interested in whether he's going to write owl:unionOf or '||' or "or" or whether he's going to use triples or SQL or whatever. He wants to get his answers in the easiest manner possible, and specifying any of this stuff about which he really doesn't care just makes the use case worse. Ignoring the OWL data, the user expects to get back Al but not Carl. He's not sure whether or not Bob will be returned. Using the OWL data, the user should get back Al and Bob, but not Carl.
Received on Monday, 12 April 2004 13:52:57 UTC