Re: [EXTERNAL] Follow-up on DCAT from D3

Dear Annette, thanks a lot for sharing this feedback - Dear Kale, many
thanks for contributing it!

Please find our replies inline.

> [snip]
>
> On Oct 28, 2020, at 9:10 AM, Aur, Katherine Anderson <kaaur@sandia.gov> wrote:
>
> Annette,
>
> We have a few comments on DCAT that might be of use to the committee:
>
>
> We added schema:version into the Resource, Dataset and Catalog Classes for versioning information. Might be better to have a DCAT term for this.

In the new versioning section
(https://w3c.github.io/dxwg/dcat/#version-info), we have not defined a
new property, but recommended the use of owl:versionInfo, as we think
that it addresses the necessary requirements.

Do you see any issue in using this property?

> We are using dct:relation, dct:replaces, prov:wasGeneratedBy to show how Resources are related and their provenance.
>
> The official DCATv2 documentation could be more explicit with how to handle these types of situations. We had a few confusing discussions around these terms.

Do you have any suggestions on how to improve the relevant examples
and guidance included in DCAT?

> We found the schema.org vocabulary to be more comprehensive for our purposes than the FOAF vocab for People/Organizations.

We still have an open issue on whether DCAT should elaborate more on
the description of agents (https://github.com/w3c/dxwg/issues/765).

It would be very useful for us to know which are your requirements on
this, and why FOAF was not considered sufficient.

> We used more schema.org vocabulary for publications (datePublished, author, publisher, numberOfPages, etc…). These seemed to fit our needs better than the recommended dcat and dct terms for this.

Besides numberOfPages, DCAT uses DCTERMS for datePublished
(dct:issued), author (dct:creator) and publisher (dct:publisher). As
they map pretty much to the relevant terms in Schema.org, could you
explain the issues you found in what DCAT recommends?

> Consider adding some DCATv2 terms for Software. We used a mix of our own terms and schema.org terms for this.

In its current version, DCAT is programmatically focussed on defining
terms concerning data resources, whereas requirements for other
resource types (or specific types of data resources) are meant to be
addressed by DCAT profiles.

We have however opened an issue on this topic:
https://github.com/w3c/dxwg/issues/1278

> I think in general the documentation for DCATv2 could use some more detailed implementation examples to help people better understand the use cases for the different vocab terms.

Thanks for raising this issue. May I ask which terms / aspects you see
as requiring further guidance and examples?

> We added Documentation and Software classes to describe new Resource that we needed that were not a Dataset or a DataService. We followed the advice on the DCAT pretty well and found this note to be helpful: ‘It is strongly recommended to use a more specific sub-class. When describing a resource which is not a dcat:Dataset or dcat:DataService, it is recommended to create a suitable sub-class of dcat:Resource, or use dcat:Resource with the dct:type property to indicate the specific type.’

Thanks for letting us know - it is also important to have clear
feedback on what in the specification is useful for adopters.

Best,

Andrea, on behalf of the DCAT editors

Received on Tuesday, 17 November 2020 19:48:52 UTC