- From: Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>
- Date: Thu, 29 Sep 2011 10:46:57 +0100
- To: Provenance Working Group WG <public-prov-wg@w3.org>
On Thu, Sep 29, 2011 at 10:26, Luc Moreau <L.Moreau@ecs.soton.ac.uk> wrote: > It's a good point, inline with mine below (not used to make a representation > of something in the world), > and also echoed by James in his response to Satya. Paolo (I think) mentioned yesterday that although you can describe an apple as an prov:Entity - we don't make :Apple a subclass of prov:Entity. Similarly, although we might in some cases be talking about a provenance container as a prov:Entity, we don't make prov:ProvenanceContainer a subclass of prov:Entity - we've not invented a new owl:Thing which every class will subclass. So although it might sound from the prov:Entity description that "almost anything" can be a characterized thing in the world, it does not mean that you have to do it. You would only do it where it would make sense for the purpose of tracing provenance, in which case a particular instance becomes a prov:Entity. (I assume however that it should be fine to create subclasses of prov:Entity which are always used to describe provenance entities, for instance wf:Value in my workflow example). It might be worth modifying the conceptual model to highlight what is the purpose of describing an Entity. Initially we had the formulation along the lines of "Something which we want to assert the provenance of" - which is somewhat the intention, but might seem to exclude entities which are only prov:used or controlling a PE. -- Stian Soiland-Reyes, myGrid team School of Computer Science The University of Manchester
Received on Thursday, 29 September 2011 09:47:48 UTC