Note that there is an OWL version created at Stanford that I would
consider to be "expert-created" -
http://protege.stanford.edu/plugins/owl/dc/protege-dc.owl
I think we should look at this to see if it seems to be semantically the
same as DCE. Note that all of the DCE properties are defined as
owl:annotationProperties, which is what I have been advised is the best
solution since no ranges are provided for DCE properties.
kc
On 10/9/14, 2:20 PM, Mariana Curado Malta wrote:
> On 09-10-2014 18:11, Kai Eckert wrote:
>> Hi Karen,
>>
>> thank you for raising this and for this comprehensive summary. I will
>> put this on the agenda for the Technical Board meeeting on Saturday (to
>> which everyone is happily invited who is interested in contributing to
>> the work subsumed under the TB, standards, communities, usage).
>>
>> Tom's summary is definitely a point to come back to, most of these
>> issues should be easy to fix or improve (notwithstanding pecularities of
>> the DCMI infrastructure).
>>
>> Regarding OWL support, I wonder why tools like Protege should not be
>> able to work with plain RDFS descriptions of vocabularies, which Dublin
>> Core provides. However, it wouldn't hurt to qualify our properties and
>> classes as owl:ObjectProperty and owl:DatatypeProperty. I wouldn't use
>> owl:AnnotationProperty as DC is not limited or particularly related to
>> describing (annotating) ontologies. Further OWL qualifications might be
>> suitable as well, like indicating inverse properties, equivalences,
>> transitivities.
>>
>> Any OWL aficionados out there who want to go ahead with a first draft?
> I can do that. I am an aficionada but not an expert. Anyway I would be
> glad to do it if there is no expert out there available.
>
> Mariana
>
>>
>> Cheers,
>>
>> Kai
>>
>> Am 30.09.2014 14:54, schrieb Karen Coyle:
>>> All,
>>>
>>> We had a discussion, without a real outcome, about issues of access to
>>> the DCterms vocabulary in RDF.[1][2] The problem there was (in my words)
>>> that the first de-reference point for a request for the vocabulary
>>> responds with the web page, not the vocabulary. This means that some
>>> "default" requests from within software return something that isn't
>>> usable, and this confuses people.
>>>
>>> Recently on the Protege support list, someone was trying to include the
>>> DC elements (DCE) vocabulary and that was failing, although for a
>>> different reason.[3] DC elements are not imported at all into Protege. I
>>> haven't tried with other software (if others can do so that would be
>>> very useful, thanks), but according to the developer responsible for the
>>> Manchester RDF2OWL software,[4] DCE does not provide enough information
>>> for *any* conversion to OWL using that software (which is very widely
>>> used). To remedy this, he suggested (in a private email) that there be
>>> an OWL version of DCE that explicitly designates the terms as
>>> owl:annotationProperty. [5] Stanford Protege provides such a file as a
>>> plugin. Of course, that is not an official DCMI version. [6]
>>>
>>> I do see a bit of tension at the moment between OWL vocabularies and
>>> RDF, so some may not see this as a problem. However, as OWL becomes more
>>> popular, in part because of the availability of tools, it is easy to see
>>> that it would be desirable for folks to be able to use DCE in
>>> combination with OWL vocabularies.
>>>
>>> I am far from expert in this area, so I leave this to others to discuss
>>> the technical details, but I do feel strongly that both DCterms and DCE
>>> should be available for use under what are becoming "normal
>>> circumstances" in linked data development, especially by those who are
>>> not expert enough to overcome the problems that these currently present.
>>>
>>> kc
>>>
>>>
>>> [1] Beginning of thread:
>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1405&L=DC-ARCHITECTURE&P=R2&1=DC-ARCHITECTURE&9=A&J=on&d=No+Match%3BMatch%3BMatches&z=4
>>>
>>>
>>> [2] Tom Baker's summary:
>>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=ind1405&L=DC-ARCHITECTURE&P=R1555&1=DC-ARCHITECTURE&9=A&J=on&d=No+Match%3BMatch%3BMatches&z=4
>>>
>>>
>>> [3]
>>> http://protege-project.136.n4.nabble.com/Import-DC-ontology-tp4661155p4661303.html
>>>
>>>
>>> [4] http://mowl-power.cs.man.ac.uk:8080/converter/
>>>
>>> [5] In fact, it would seem that owl:dataProperty would not be a bad
>>> solution, IMO, because most DCE values are strings, and
>>> annotationProperty's seem to be 3rd-class citizens in the OWL universe,
>>> ignored in some cases.
>>>
>>> [6] http://protege.stanford.edu/plugins/owl/dc/protege-dc.owl
>>>
>
--
Karen Coyle
[log in to unmask] http://kcoyle.net
m: 1-510-435-8234
skype: kcoylenet/+1-510-984-3600
|