Re: PROV-N to PROV-O mapping

Hi Héctor,

I'm not sure if anyone has responded to this, if not, sorry for the delay.  The history there is that the ProvRDF page was started when it started to become clear that it was difficult to keep the PROV-O and PROV-N notations (as described in the two large documents) synchronized.  To my knowledge, the ProvRDF page stopped being maintained once the existing specifications had stabilized to the point where it was possible to keep track of the individual changes.  I believe Tim Lebo and Stian Soiland-Reyes were the main maintainers.

Giving a clear mapping between the two is important, e.g. for really being sure about how to enforce constraints on PROV-O-encoded provenance, but doing so has also been more or less placed outside of the scope of the WG's remit, rather it is something that could be done by a future WG or member submission once the need for it is clear.  I believe this was for pragmatic reasons, i.e. lack of energy/volunteers. (Someone should correct me if I'm wrong, not speaking for the group here).  

Nevertheless, it would be great to bring the ProvRDF page up to date with the CR versions of everything, before the wiki is frozen, and perhaps link to it from the FAQ.  Something similar could be done with the XML schema, but there the mapping is a lot more obvious.  If you are already working on understanding the mapping, and have suggestions on how to bring the mapping page up to date, I'm happy to make the changes (if no one else is).  If there is enthusiasm for developing this to the point that it could be included in a note, e.g. PROV-SEM, so much the better.

--James


On Jan 23, 2013, at 9:17 PM, Héctor Pérez-Urbina <hector@clarkparsia.com> wrote:

> Hello,
> 
> We have been working on capturing the various PROV inferences in an ontology by extending the PROV ontology (PROV-O).
> 
> In the process we were wondering if you guys have defined a mapping between PROV-N and PROV-O. We have been resorting to analyzing the test cases in order to find these correspondences. This is obviously suboptimal and error-prone.
> 
> We came across the ProvRDF document (http://www.w3.org/2011/prov/wiki/ProvRDF), but it's incomplete and not up-to-date. Will the WG update this document?
> 
> -- 
> Best,
> Héctor