Simon miles

kaiec edited this page Jul 5, 2012 · 5 revisions
 (kai) I leave this paragraph as it is for now, as I am not sure how to formulate it less technically. I think we could address your comment by adding a paragraph to the introduction where the fundamental concepts of PROV and DC are explained first. I keep this as an open issue.
 The documents seem to assume that we are mapping DC RDF to PROV-O (rather than DC to PROV more generally), but I didn't see this explained anywhere.
 Many points (such as the questions at the end of "What is ex:doc1") would benefit from a running example to make the ideas concrete.
 (kai) Right. However, the first sentence is: "The mapping is based on the definitions of the DCMI terms and the latest version of the PROV ontology." I added: ", so we focus on the RDF representations of both models."
 I think it would help clarity to come up with more descriptive names than "Stage 1" and "Stage 2".
 (kai) I removed them and reformulated the text.
 As it's a draft, there are clearly some explanations missing, and so I note some points that need to be clarified:
 - In "What is ex:doc1", option 1: why are the mappings potentially bloated? Needs higher level explanation.
 (kai) See above.
 - In option 2, why are the PROV semantics unclear? I wasn't clear what you were trying to say.
 (kai) We already changed the text lightly. I keep it as an issue that should be reviewed before the final publications.
 - What is the connection of the dc:publisher figure (which needs explanation itself) to the text?
 (kai) Added caption and reference in the text.
 - Are "PROV Specializations" the same as "specializationOf" in PROV? I didn't see the connection, and if there is no connection we should not use the term.
 I have to think through all the proposed mappings. Some direct mappings seem maybe not intuitive. I believe that PROV aims to cover a smaller area than DC (i.e. only provenance) but more generally (i.e. any kind of past occurrence).  Therefore, I would not expect PROV terms to usually be subclasses of DC terms.
 For example, I don't think wasRevisionOf is intended to be more specific than isVersionOf, even if under certain readings of the words "revision" and "version" this might be intuitive (also, I remember debating whether isVersionOf actually links different versions of a resource, or links a version of a resource to the general document).
 (kai) yes, they are the different. I changed the PROV specializations in the sense of newly introduced subclasses to refinements. I am not sure if the term refinement fits best, but at least it is now distinguished from the entity specializations and we can easily change the term if we have something better. I keep your point about the actual mapping as issue, that will be discussed separately.