review on the draft model of the Data Usage Vocabulary

Dear Bernadette and all,

I have been reviewing the draft model of the Data Usage Vocabulary, and I
am very glad we now have some concrete artifact to work on.

I have the following remarks:

1) In my opinion duv:Publisher should not be a notion defined in DUV
because of a matter of conceptual scope. The notion of "publisher" is used
in DCAT borrowed from DublinCore Terms. See
http://dublincore.org/documents/2012/06/14/dcmi-terms/?v=terms#terms-publisher


DCAT does not specialize dct:publisher. It only adopts it informally
(suggesting that it should range over foaf:Agent with domain dcat:Dataset.
(I think that it would have been better to include publisher formally in
DCAT as a subPropertyOf dct:publisher, but this was not done in DCAT...) A
minor problem that we have to address is that dct:publisher's range is
dct:Agent and not foaf:Agent, which I understood we intend to use in DUV.

2)  With respect to duv:Consumer, this is certainly in scope, but I believe
this should be represented as some relation between an agent and a Dataset.
Either because it produces an application that refers do some dataset or
because the agent cites some data. So, if we are consistent in our
representation style, and considering (1), there should be no class for
Consumer, but instead, properties relating an Agent to the dataset it
cites, or an Application to the dataset it uses. Another option should be
to make duv:Consumer the range of properties that represent this.

3) What is the semantics of DatasetUsage and why it "provides" a Dataset
while an Application (a special kind of DatasetUsage) "consumes" a Dataset?

4) Why is Citation not a kind of DatasetUsage?

5) I agree with others that "Blog" should not be a subclass of Feedback.

best regards,
João Paulo