W3C home > Mailing lists > Public > public-prov-wg@w3.org > November 2012

Re: future-proofing prov-o.owl (and "namespace concatenation")

From: Ivan Herman <ivan@w3.org>
Date: Tue, 20 Nov 2012 13:04:19 -0500
Cc: "public-prov-wg@w3.org WG" <public-prov-wg@w3.org>
Message-Id: <B585DF74-8C8D-4C58-B2F3-8A721F847E88@w3.org>
To: Timothy Lebo <lebot@rpi.edu>

On Nov 20, 2012, at 11:38 , Timothy Lebo wrote:

> Ivan,
> 
> On Nov 20, 2012, at 11:26 AM, Ivan Herman <ivan@w3.org> wrote:
> 
>> Some questions
>> 
>> - (just to make it sure) you staged these under CR-prov-***/ns but it is not the intention, is it, to put this into the /TR/2012/.../ns directory. The target for all these files is www.w3.org/ns 
> 
> Correct. The directory organization that I chose is to tell you "this is what to publish into /ns, as of CR". I intend that NONE of this appears in /TR.

Ok. Just checking:-)

> 
>> 
>> - The .owl extension goes to RDF/XML files, I presume.
> 
> Yes. Are you suggesting to add .rdf in addition, or to replace .owl with .rdf?
> 
> 

Actually, as I said below, I was suggesting to use extension-less URI-s, then populate /ns with .rdf and .owl (with identical content) and .ttl for a turtle version. I think that is what we have already.


>> I wonder whether we should not use extension-less URI-s everywhere, eg, http://www.w3.org/ns/prov-o-20120724, and let conneg work to choose among RDF/XML or turtle versions.
> 
> 
> That is what I would prefer. But I think we're bending certain ways to help the bad tooling that is out there.
> The tradeoff that is in place is to point to the extension less URIs using specializationOf.

You mean that is what you will do (at the moment, all the URI-s have an extension)

> 
> 
>> You should check whether an average OWL processor handles such conneg properly…
> 
> I'm not sure what an "average OWL processor" looks like…

:-)

> I can try it on Protege. We could find someone to try it on TopBraid Composer. What others?
> 

Pellet, at the minimum, but also browsers like TimBl's tabulator (although I suspect tabulator is fine, it will do conneg)


>> 
>> - What will I get if I dereference  http://www.w3.org/ns/prov ?
> 
> More than prov-o is returned. The broad design of "namespace concatenation" was agreed to by the WG this summer, but it hasn't been implemented yet (Paul asked me to finish it by CR publish time).

ok

Ivan

> 
> Poor quality and out of date notes: http://www.w3.org/2011/prov/wiki/ProvNamespaceManagement#Solution_2.2_Use_owl:import_and_return_full_merge_of_PROV-O_and_all_Notes
> 
> -Tim
> 
>> 
>> Ivan
>> 
>> 
>> On Nov 20, 2012, at 10:25 , Timothy Lebo wrote:
>> 
>>> Ivan, prov-wg, OWL-in-practice gurus, and anyone concerned with the "prov ns concatenation" issue,
>>> 
>>> I've staged the prov-o ontology files to go into /ns:
>>> 
>>> http://dvcs.w3.org/hg/prov/file/tip/ontology/releases/CR-prov-o-20121211/ns
>>> 
>>> The directory contains the files for {prov, inverses} \cross {versioned, unversioned} \cross {rdf/xml, turtle}
>>> 
>>> Each of the two ontologies (provo, inverses) has a owl:versionIRI, prov:wasRevisionOf, and prov:specializationOf, i.e.,
>>> 
>>> from provo http://dvcs.w3.org/hg/prov/file/de8cda493917/ontology/releases/CR-prov-o-20121211/ns/prov-o.ttl#l30 :
>>> 
>>>    30 <http://www.w3.org/ns/prov#
>>>> 
>>> 
>>>    31     a owl:Ontology ;
>>>    37     owl:versionIRI <http://www.w3.org/ns/prov-o-20121211.owl
>>>> ;
>>> 
>>>    38     :specializationOf <http://www.w3.org/ns/prov-o.owl
>>>> ;
>>> 
>>>    39     :wasRevisionOf <http://www.w3.org/ns/prov-o-20120724.owl> .
>>> 
>>> and from inverses http://dvcs.w3.org/hg/prov/file/de8cda493917/ontology/releases/CR-prov-o-20121211/ns/prov-o-inverses.ttl#l7 :
>>> 
>>>     7 <> a owl:Ontology;
>>>    11    owl:versionIRI        <http://www.w3.org/ns/prov-o-inverses-20121211.owl
>>>> ;
>>> 
>>>    12    prov:wasRevisionOf    <http://www.w3.org/ns/prov-o-inverses-20120724.owl
>>>> ;
>>> 
>>>    13    prov:specializationOf <http://www.w3.org/ns/prov-o-inverses.owl> .
>>> 
>>> 
>>> The PROV HTML points to the "unversioned" OWL files, specifically:
>>> 
>>> [[
>>> The OWL encoding of the PROV Ontology is available here (http://www.w3.org/ns/prov-o.owl)
>>> ]]
>>> 
>>> and 
>>> 
>>> [[
>>> For convenience, this file (http://www.w3.org/ns/prov-inverses.owl) lists the resulting inverse properties.
>>> ]]
>>> 
>>> 
>>> I believe from this setup that consumers will be able to:
>>> * recognize when the OWL encoding was updated (via a new owl:versionIRI), 
>>> * find the previous version (via wasRevisionOf), and 
>>> * find the latest version at any point in the future from their current copy of the OWL encoding (specializationOf).
>>> 
>>> 1) Do you agree? Is this an acceptable arrangement?
>>> 
>>> 
>>> 2) Looking ahead to the "namespace concatenation" [1], I would expect that I should change the URIs for the "provo" and "inverses" ontologies, to something like:
>>> 
>>> http://www.w3.org/ns/prov-o#
>>> and
>>> http://www.w3.org/ns/prov-o-inverses#
>>> 
>>> Does that sound reasonable?
>>> 
>>> Thanks,
>>> Tim
>>> 
>>> 
>>> 
>>> [1] http://www.w3.org/2011/prov/wiki/ProvNamespaceManagement#Solution_2.2_Use_owl:import_and_return_full_merge_of_PROV-O_and_all_Notes
>>> 
>> 
>> 
>> ----
>> Ivan Herman, W3C Semantic Web Activity Lead
>> Home: http://www.w3.org/People/Ivan/
>> mobile: +31-641044153
>> FOAF: http://www.ivan-herman.net/foaf.rdf
>> 
>> 
>> 
>> 
>> 
>> 
>> 
> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
FOAF: http://www.ivan-herman.net/foaf.rdf
Received on Tuesday, 20 November 2012 18:04:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 20 November 2012 18:04:46 GMT