search

w3c / dxwg

Data Catalog Vocabulary (DCAT)
https://w3c.github.io/dxwg/dcat/
Other
144 stars 46 forks source link

What to do with axioms in the DCAT 3 RDF serializations? #1402

Closed riccardoAlbertoni closed 2 years ago

riccardoAlbertoni commented 2 years ago

We discussed dropping some axioms from the DCAT 2 RDF serialization as they were not explicitly mentioned in the DCAT 2 REC. See #PR #1401 for the list of axioms.

How to deal with the same axioms in dcat 3. We might either a) delete the above axioms or b) add them in the DCAT 3 HTML specification and keep it in the RDF serialization.

Some of the above axioms relate to the discussion in separate issues, e.g., Axiom 2 in https://github.com/w3c/dxwg/issues/1242.

riccardoAlbertoni commented 2 years ago

Considering the discussions we had in the last DCAT meeting and the overall reluctance to add constraining in the DCAT RDF formalization. I would be prone to be provocatively proposing to delete in the DCAT 3 RDF the axioms 1, 2, and 3 as done for DCAT 2 RDFs (solution a) in PR #1401.

Any reactions to my provocation? If not, I will prepare a PR.

andrea-perego commented 2 years ago

+1 from me.

dr-shorthair commented 2 years ago

Yeah - too many axioms. That detail should be left to community profiles, probably formalized as shapes (SHACL and ShEX) rather than RDFS/OWL.

smrgeoinfo commented 2 years ago

Perhaps a 'recommended set' of axioms (shapes?) could be set in a separate module, to promote some level of consistency? A profile could choose to use (import) them or not.

riccardoAlbertoni commented 2 years ago

We removed the axioms in the DCAT 3 RDFs similarly to DCAT 2.

Perhaps a 'recommended set' of axioms (shapes?) could be set in a separate module, to promote some level of consistency? A profile could choose to use (import) them or not.

That could be an option, but at the same time, I think having further resources might be a source of confusion on what is the real expectation.
I think some users might have already difficulties in distinguishing normative from non-normative parts, and for that reason, I would be prone not to provide the axioms separately.

Can we close this issue?

andrea-perego commented 2 years ago

Can we close this issue?

+1 from me.