Re: PROV-O in LOV : URI, namespace and versions

Bernard,

On Jun 17, 2013, at 1:25 PM, Bernard Vatant <bernard.vatant@mondeca.com> wrote:

> Hi Tim and all
> 
> Thanks for the feedback. Not much time to parse all your answers right now, I will be back on this on Wednesday.
> 
> Quick answer for now. You wrote 
> 
>   30 <http://www.w3.org/ns/prov#>
>   31     a owl:Ontology .
> 
> 1237 <http://www.w3.org/ns/prov-o#>
> 1238     a owl:Ontology ;
> 
> HA! The LOV-Bot (and myself) did not parse down to the second owl:Ontology declaration, because it assumes there should be only one such  
>  ... but seems to me this is adding to the pile of issues rather than anything. Why the first declaration, then?

As Daniel described in his response, <http://www.w3.org/ns/prov#> is an owl:Ontology that is an aggregation of a handful of owl:Ontologies (prov-o is one of them, prov-aq, prov-dc are two others).
<http://www.w3.org/ns/prov#> is explicitly typed to an owl:Ontology because 1) it is one, and 2) some consuming tools required the explicit type assertion.


It seems to me that your bot should remember which ontology it requested, then look for that specific ontology's description in the representation that was returned.
So, if I want to know about owl:Ontology http://www.w3.org/ns/prov#, I request RDF from http://www.w3.org/ns/prov (since clients don't request fragIDs), and then look into that graph for descriptions of http://www.w3.org/ns/prov#. A bot shouldn't get distracted by other instances of owl:Ontology that may also be mentioned in the representation.

> 
> You wrote :
> "http://www.w3.org/ns/prov-o and http://www.w3.org/ns/prov are owl:Ontologies."

Apologies, I mis-spoke and was not being specific enough for the rest of the discussion. Rephrasing:

[[
yes, the namespace of the vocabulary is http://www.w3.org/ns/prov#
http://www.w3.org/ns/prov-o# and http://www.w3.org/ns/prov# are owl:Ontologies.
]]

(Though, I wouldn't object to anybody *also* typing http://www.w3.org/ns/prov-o and http://www.w3.org/ns/prov to be owl:Ontologies.)

You will note that one of the owl:Ontology URIs is also the namespace.

Representations for the owl:Ontologies http://www.w3.org/ns/prov-o# and http://www.w3.org/ns/prov# may be accessed from the URLs http://www.w3.org/ns/prov-o and http://www.w3.org/ns/prov, respectively, using appropriate content negotiation.


> and later on
> "Could you explain why you would expect http://www.w3.org/ns/prov-o to be a owl:Ontology?"

I'm not implying that http://www.w3.org/ns/prov-o is NOT an owl:Ontology, I'm simply trying to understand why you expected it to be.
As I mentioned above, it would *not* be wrong for one to state that http://www.w3.org/ns/prov-o is an owl:Ontology, but you'll never find that assertion from w3.org.
The WG used the convention that the "hash URI" instances are owl:Ontology.



> and finally
> "http://www.w3.org/ns/prov-o provides the resource representation of the owl:Ontology http://www.w3.org/ns/prov-o#"
> 
> Now I'm totally confused ... I have to munch over all this, but  all LOV architecture is based on discovering the ontology URI (here it seems to be http://www.w3.org/ns/prov-o# and not http://www.w3.org/ns/prov-o but I'm not sure) from de-referencing its namespace http://www.w3.org/ns/prov#


Perhaps the confusion is that there are a handful of ontologies.

http://www.w3.org/ns/prov# is an owl:Ontology that contains ANY term from ANY PROV-WG document, regardless of Rec or Note (see http://www.w3.org/ns/prov#)
http://www.w3.org/ns/prov-o# is an owl:Ontology that contains the Recommendation terms. (see http://www.w3.org/TR/prov-o/)
http://www.w3.org/ns/prov-aq# is an owl:Ontology that contains the Note terms for access and query. (see http://www.w3.org/TR/prov-aq/)
http://www.w3.org/ns/prov-dictionary# is an owl:Ontology that contains the Note terms for dictionaries (see http://www.w3.org/TR/prov-dictionary/#dictionary-ontological-definition)
http://www.w3.org/ns/prov-links# is an owl:Ontology that contains the Note terms for bundling linking (see http://www.w3.org/TR/prov-links/)
http://www.w3.org/ns/prov-dc# is an owl:Ontology that contains the Note terms for mapping to Dublin Core (see http://www.w3.org/TR/prov-dc/#bib-refinements)



> If you tell me the configuration you have enables to do that and that there is no possible ambiguity, please help us to understand how so that we can improve the LOV-Bot behavior accordingly …

I'm happy to point you to any configuration that we have. Which configuration do you need?

> 
> The bottom line is that after two years of work on this, hundreds of vocabulary URI and namespaces parsed, dozens of different conneg configurations found, and so many discussions and exchanges with so many SemWeb gurus, we're still enable to settle this issue properly. 
> 
> And yours is yet another configuration …


I'm familiar with LOV and I think it's a great project. 
If PROV isn't suiting LOV's bot, then I'm eager to reconcile it.

We tried to provide all of these relations explicitly in the RDF that we're publishing and its done in best-practice vocabularies, so if you think something is missing please let us know.

Regards,
Tim


> 
> More to come (have to pack for my night train)
> 
> Bernard
> 
> 
> 2013/6/17 Timothy Lebo <lebot@rpi.edu>
> Hi, Bernard,
> 
> I'm cc'ing the prov comments list to archive your comments.
> 
> @PROV-WG, I might need help refreshing on our provenance-of-provenance design to fulfill Bernard's needs)
> @Ivan, we have a 404 on http://www.w3.org/ns/prov-o-20120312 but it needs to be there like http://www.w3.org/ns/prov-o-20130430 is)
> @PROV-WG-chairs, are we still able to (or should we) use the tracker?
> 
> I respond within...
> 
> On Jun 17, 2013, at 9:32 AM, Bernard Vatant <bernard.vatant@mondeca.com> wrote:
> 
>> Hi Tim and Daniel
>> 
>> I updated today the LOV record for PROV-O [1]. Actually this should have been done well before, but the new version(s) had not been captured by the LOV-Bot, due mainly to a confusing (for me and the LOV-Bot at least) namespaces and URI policy.
>> I have several issues with this (important) vocabulary.
> 
> Thanks for reporting your challenges with consuming the PROV namespace. If PROV doesn't suit LOV, then it's a clear indicator that we aren't following common practice.
> I hope we can clear up the issues, either on your side or on ours.
> I've started a section on the semweb wiki to document your questions and their answers: http://www.w3.org/2001/sw/wiki/PROV-FAQ#The_PROV_URIs
> 
>> 
>> As I write in the vocabulary "review" : " Note that the namespace of the vocabulary is http://www.w3.org/ns/prov#, but its URI is http://www.w3.org/ns/prov-o. But the RDF file declares <http://www.w3.org/ns/prov#> a owl:Ontology, which seems confusing at least."
> 
> 
> yes, the namespace of the vocabulary is http://www.w3.org/ns/prov#
> 
> http://www.w3.org/ns/prov-o and http://www.w3.org/ns/prov are owl:Ontologies.
> 
> From the comment itself, I'm not sure what is confusing. Could you elaborate? (or, do the following three cover your concerns?)
> 
> 
>> 
>> ... In more details : 
>> 
>> 1. The namespace http://www.w3.org/ns/prov# does not dereference to the ontology, but to a general documentation page about various documents using the namespace, including the ontology itself.
> 
> Yes.
> 
>> How am I supposed to GET the RDF description of e.g., http://www.w3.org/ns/prov#Activity, from the namespace URI?
> 
> In short, use content negotiation.
> http://www.w3.org/2001/sw/wiki/PROV-FAQ#GET.27ing_an_RDF_description_of_a_PROV_term
> curl -H "Accept: application/rdf+xml" -L http://www.w3.org/ns/prov
> 
> 
>> 
>> 2. In the RDF at http://www.w3.org/ns/prov-o I read  <http://www.w3.org/ns/prov#> a owl:Ontology 
>> one should expect  <http://www.w3.org/ns/prov-o> a owl:Ontology seems to me …
> 
> There are two instances of owl:Ontology, at lines:
> 
>   30 <http://www.w3.org/ns/prov#>
>   31     a owl:Ontology .
> 
> 1237 <http://www.w3.org/ns/prov-o#>
> 1238     a owl:Ontology ;
> 
> Could you explain why you would expect http://www.w3.org/ns/prov-o to be a owl:Ontology?
> 
> As it stands, http://www.w3.org/ns/prov-o provides the resource representation of the owl:Ontology http://www.w3.org/ns/prov-o#
> 
> 
>> 
>> 3. The previous versions (Candidate Recommandation and Proposed Recommendation) are available as HTML documentation, but the respective RDF versions are not available.
> 
> 
> 
> 
> 
>> The link to the OWL version redirects to the current version.
> 
> Yes, to restate: http://www.w3.org/TR/2013/PR-prov-o-20130312/ points to http://www.w3.org/ns/prov-o in "The OWL encoding of the PROV Ontology is available here."?
> 
> The RDF representations for those versions are sitting around, so if you could describe in more detail how you'd like to be able to access them, perhaps we can update the provenance-of-provenance to 
> suit your use case.
> 
> From within http://www.w3.org/ns/prov-o,  the owl:versionIRI and prov:wasDerivedFrom are intended to provide access to the previous versions.
> But, I'm seeing a 404! :-)
> 
> 
> <http://www.w3.org/ns/prov-o#>
>     a owl:Ontology ;
>     rdfs:comment """This document is published by the Provenance Working Group (http://www.w3.org/2011/prov/wiki/Main_Page). 
> 
> If you wish to make comments regarding this document, please send them to public-prov-comments@w3.org (subscribe public-prov-comments-request@w3.org, archives http://lists.w3.org/Archives/Public/public-prov-c
> omments/). All feedback is welcome."""@en ;
>     rdfs:label "W3C PROVenance Interchange Ontology (PROV-O)"@en ;
>     rdfs:seeAlso <http://www.w3.org/TR/prov-o/>, <http://www.w3.org/ns/prov> ;
>     owl:versionIRI <http://www.w3.org/ns/prov-o-20130430> ;
>     owl:versionInfo "Recommendation version 2013-04-30"@en ;
>     :specializationOf <http://www.w3.org/ns/prov-o> ;
>     :wasRevisionOf <http://www.w3.org/ns/prov-o-20120312> .
> 
> 
> Thanks again for your comments. I hope that we can iron out some of the wrinkles that you ran into.
> 
> Regards,
> Tim Lebo
> 
> 
> 
>> 
>> Thanks for any clarification. I understand that since we deal now with a W3C Recommandation, if anything needs to be fixed, the process is likely to be long :)
>> 
>> Best regards
>> 
>> 
>> [1] http://lov.okfn.org/dataset/lov/details/vocabulary_prov-o.html
>> 
>> 
>> Bernard Vatant
>> Vocabularies & Data Engineering
>> Tel :  + 33 (0)9 71 48 84 59
>> Skype : bernard.vatant
>> Blog : the wheel and the hub
>> Linked Open Vocabularies : lov.okfn.org 
>> --------------------------------------------------------
>> Mondeca                             
>> 3 cité Nollez 75018 Paris, France
>> www.mondeca.com
>> Follow us on Twitter : @mondecanews
>> ----------------------------------------------------------
>> Meet us during the European Open Data Week in Marseille (June 25-28)
>> 
>> 
>> 
> 
>

Received on Monday, 17 June 2013 18:17:38 UTC