- From: Paul Gearon <gearon@ieee.org>
- Date: Tue, 29 Sep 2009 10:03:59 -0400
- To: SPARQL Working Group <public-rdf-dawg@w3.org>
On Tue, Sep 29, 2009 at 6:20 AM, Alexandre Passant <alexandre.passant@deri.org> wrote: > HI, > > On 15 Sep 2009, at 13:38, Seaborne, Andy wrote: <snip/> >> Having access to the length of the path needs to say what happens for >> loops and for multiple paths between the same two graph nodes. > > While imo relatively useful, I also agree that using path length can be > tricky in many cases, esp. combined with entailment. Only if the predicate is considered transitive (the issue that you demonstrated in your email). But I don't believe that path length makes sense for a transitive predicate, since the path length is always "1", by definition. Instead it becomes useful for other predicates, particularly non-transitive subproperties of transitive predicates. For instance, family:hasAncestor is transitive, so path lengths don't make sense, but the sub-property family:hasFather is non-transitive and the path length makes a lot of sense. foaf:knows is another example, this time without a transitive super-property. So if a predicate is being treated as transitive, then the path must be 1 or 0, but otherwise it could be useful. Paul
Received on Tuesday, 29 September 2009 14:04:42 UTC