W3C home > Mailing lists > Public > public-prov-wg@w3.org > March 2012

Re: PROV-ISSUE-277 (TLebo): Supporting property chains [Ontology]

From: Graham Klyne <graham.klyne@zoo.ox.ac.uk>
Date: Sun, 04 Mar 2012 07:57:10 +0000
Message-ID: <4F532056.5010906@zoo.ox.ac.uk>
To: public-prov-wg@w3.org
I would hope this is a non-issue.  E.g. property paths in SPARQL include 
provision for including inverse properties that are not explicitly defined: 
http://www.w3.org/TR/sparql11-query/#propertypaths

(I guess this is just a reminder, but the "direction" of RDF properties places 
no technical constraint on accessibility - one can, in principle (and in 
practice with most triple stores) traverse a property backwards as easily as 
forwards.  Any need for explicit inverse properties is almost entirely for human 
consumption (and authoring), and their absence shouldn't constrain applications 
in any way.  Indeed, defining inverse properties is more likely to create 
problems of incompatibility by introducing different ways to express the same 
assertion.)

#g
--

On 03/03/2012 16:28, Provenance Working Group Issue Tracker wrote:
> PROV-ISSUE-277 (TLebo): Supporting property chains [Ontology]
>
> http://www.w3.org/2011/prov/track/issues/277
>
> Raised by: Stephen Cresswell
> On product: Ontology
>
> During our group telecon, someone (Stephen Cresswell?) mentioned a concern that the directionality of some properties in prov-o would inhibit the use of property chains.
>
> Although "directionality" can be handled with owl:inverses, we are not including many inverses in prov-o for brevity (however, we are maintaining a component at [1]). Although "anyone" can define their own inverse of a prov-o property to achieve their property chains, this will inhibit interoperability.
>
>
> [1] http://dvcs.w3.org/hg/prov/file/tip/ontology/components/inverses.ttl
>
>
>
>
Received on Sunday, 4 March 2012 08:51:50 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 13:06:58 GMT