- From: Graham Klyne <GK@ninebynine.org>
- Date: Thu, 03 Nov 2011 13:35:07 +0000
- To: public-prov-wg@w3.org
Luc, There are possible choices here - probably not a single correct response. I would consider the option that the different entities can be associated with provenance graphs (using separate hasProvenance statements), but that such statements might refer to the *same* graph. Thus, a single provenance graph - maybe for a single overall workflow execution - could be constructed then associated with each of the artifacts generated by that execution. Or each entity could have a minimal graph dealing with just the details pertaining to the provenance of that entity. I'd like to see some implementation leeway here, as long as there's no possibility for mis-interpretation of the resulting provenance. (It also relates to the possible advantage I hinted at in relation to the duplication of the entity URI - the provenance graph is able to include additional information that may not be specifically about the original entity.) #g -- On 03/11/2011 12:56, Luc Moreau wrote: > Hi Graham, > > Thanks for exploring simplified ways of expressing provenance. > > How do you see your approach with "hasProvenance" [1] with named graphs > when you consider more than one documents? > > Concretely, here is an example in the ASN, with aDoc1 and aDoc2. > > entity(aDoc1, [type=ex:Document]) > entity(aDoc2, [type=ex:Document]) > agent(meritoriousMeerkat) > entity(meritoriousMeerkat, [foaf:name="Meritorious Meerkat"]) > wasGeneratedBy(aDoc1, pe1, qualifier()) > wasControlledBy(pe1,meritoriousMeerkat,qualifier=(role=dcterms:creator)) > wasUsed(pe1, aDoc2, qualifier()) > > Would both aDoc1 and aDoc2 have their how provenance graphs? > > Cheers, > Luc > > [1] > http://www.w3.org/2011/prov/wiki/Exploring_provenance_model_complexity#Use_named_graph_for_provenance > > > On 03/11/2011 12:39, Graham Klyne wrote: >> In a previous email >> (http://lists.w3.org/Archives/Public/public-prov-wg/2011Oct/0228.html) I >> indicated that I wanted to explore some simplified ways to express provenance >> information that could still be related to the provenance data model. >> >> I've now taken a first stab at such an exploration, which I've posted in the >> working group wiki at >> http://www.w3.org/2011/prov/wiki/Exploring_provenance_model_complexity >> >> Part of what I wish to explore is ways in which the provenance model can be >> expressed in RDF, without necessarily having a 1:1 correspondence between the >> data model terms and the RDF terms used. To my mind, one of the main >> advantages of having a separate abstract syntax is that it allows exploration >> opf alternative representations while retaining the essential underlying >> structure. >> >> #g >> -- >> >
Received on Thursday, 3 November 2011 14:14:28 UTC