- From: Robert Sanderson <azaroth42@gmail.com>
- Date: Thu, 9 Apr 2015 09:19:14 -0700
- Cc: W3C Public Annotation List <public-annotation@w3.org>
- Message-ID: <CABevsUES+oGmcr1k6-CPzOk_qNXSK5cvbXqUyDOvxrat_hkYtA@mail.gmail.com>
To also try to summarize so I can make sure that I understand as well ... Due to an inference rule in PROV, the current mapping that serializedBy is a subPropertyOf prov:wasAttributedTo, made for complications. If it's important to distinguish the document at all, then it needs a different mapping. The suggestion is to define a oa:serializedInto predicate, and implicitly a AnnotationDocument class to be the range of that predicate. Then serializedBy / serializedAt are shortcuts for the property chains through to the properties of that resource. So I agree with Randall, that the mapping appendix needs to be updated, the ontology needs to define that predicate and class, and the rest of the spec stays the same. The question asked by Ray is whether the value of the serializedAt/By information warrants the inclusion at all, or whether we can simply drop the information completely. I guess the question is whether implementers intend to populate these fields in their systems? Rob On Mon, Mar 30, 2015 at 12:12 PM, Randall Leeds <randall@bleeds.info> wrote: > Comparing the original diagram of the mapping and the revised proposal it > seems that there are two differences to my undestanding: > > - invert the derivation: anno1 hadDerivation annoDocument1 > - add the generation provenance for the serialization: annoDocument1 > wasAttributedTo agent2 > > I fully admit that I don't understand completely what the implications > are, but it seems to me that if the goal is these two points then all > that's needed is to update the mapping diagram. Any concern about > complication can be removed by not adding an oa:serializedInto property. > Those who would want to do so can model as in the (revised) mapping > diagram. Those who don't need this information don't need to be aware of > any of this complication. > > On Mon, Mar 30, 2015 at 6:02 AM Ivan Herman <ivan@w3.org> wrote: > >> Hi Luc, >> >> First of all, I hope you ate the cake you baked:-) >> >> More seriously: just to understand the issue in somewhat more laymen's >> terms. The way I explain myself what you say is that the current annotation >> approach does not really get into the details of the 'activities', ie, does >> not really go into the detail on how the serialization really happened in >> terms of timing and in terms of process. On the other hand, PROV considers >> the details of this workflow as essential. Is that correct? If so, then >> what you propose at the end of your blog is to ensure that the >> serialization process is clearly separated from the act of annotation by >> introducing a separate oa:serializedTo property whose object will, >> conceptually, separate the serialized document from the annotation itself. >> >> Model-wise or, if you like, OWL-wise I think I can follow this and it is >> o.k. What I do not know is how this will translate into practice. Is this >> something that remains on a modeling level, hidden in the guts, so to say, >> of the annotation model, or is this something that will appear, as extra >> triples, on the serialized annotation that is traveling over the wire? In >> other words, can we simply use the oa:serializedBy axiom using >> propertyChainAxiom as part of the core vocabulary specification of OA, and >> let those processors who are really interested by these guts use those, >> while being ignored by lighter weight Web Applications? >> >> My concern is that we get into the type of modeling details that many in >> the group and the community will be scared by and will consider mostly >> unnecessary and more of a nuisance. We should find a good balance here. >> >> Thanks! >> >> Cheers >> >> Ivan >> >> > On 30 Mar 2015, at 12:56 , Luc Moreau <L.Moreau@ecs.soton.ac.uk> wrote: >> > >> > Hi Rob, all >> > >> > Here is my proposal for the mapping to PROV. >> > As I thought this was a useful technique I wrote a standalone blog post >> about it. >> > https://lucmoreau.wordpress.com/2015/03/30/provenance- >> recipe-mapping-shortcuts-to-prov/ >> > >> > Feedback welcome! >> > >> > Luc >> > >> > >> > On 18/02/15 22:00, Rob Sanderson via GitHub wrote: >> >> @lucmoreau Could you propose a solution? >> >> >> >> The options that occur to me: >> >> * Simply drop the subPropertyOf relationship to prov for >> >> serializedBy/At. This doesn't feel clean. >> >> * Split the annotation node to always refer to the concept and then >> >> not require a node in the graph for the document. So if you want to >> >> give serializedBy/At you have a new node that otherwise isn't present. >> >> We could then just use the prov terms directly. This feels cleaner >> >> .... but a lot more expensive when the information should be captured. >> >> >> >> Also discussed on the call today was: >> >> * Drop serializedBy/At completely, but this was not viewed >> >> favorably. >> >> >> >> Thanks! >> >> >> > >> > -- >> > Professor Luc Moreau >> > Head of the Web and Internet Science Group >> > Electronics and Computer Science tel: +44 23 8059 4487 >> > University of Southampton twitter: @lucmoreau >> > Southampton SO17 1BJ, UK http://www.ecs.soton.ac.uk/~lavm >> > >> > >> > >> >> >> ---- >> Ivan Herman, W3C >> Digital Publishing Activity Lead >> Home: http://www.w3.org/People/Ivan/ >> mobile: +31-641044153 >> ORCID ID: http://orcid.org/0000-0003-0782-2704 >> >> >> >> >> -- Rob Sanderson Information Standards Advocate Digital Library Systems and Services Stanford, CA 94305
Received on Thursday, 9 April 2015 16:19:49 UTC