search

SEMICeu / DCAT-AP

This is the issue tracker for the maintenance of DCAT-AP
https://joinup.ec.europa.eu/solution/dcat-application-profile-data-portals-europe
72 stars 24 forks source link

Checksum in DCAT 3.0 #244

Closed bertvannuffelen closed 2 weeks ago

bertvannuffelen commented 1 year ago

Checksum has been adopted from DCAT-AP.

So far there are no differences detected. The proposal is to adopt the proposed definitions by DCAT 3.0. This should lead to only minor adaptations to DCAT-AP.

init-dcat-ap-de commented 1 year ago

We agree and we will start the shift to use the official SPDX vocabulary of algorithms in the coming DCAT-AP.de versions.

(Note: not all supported algorithms are correctly shown: https://github.com/spdx/spdx-spec/issues/815 )

init-dcat-ap-de commented 11 months ago

Bug: The link in the Controlled vocabularies to be used is https://spdx.org/rdf/terms/#d4e1968 instead of https://spdx.org/rdf/terms/#d4e2129 (My guess is that they recreate the urls when they build the documentation, even though cool URIs don't change...)

Feature Request: The Usage note of the attribute checksum should direct people to the Controlled vocabularies to be used. The supportive entity Checksum Algorithm seems unnecessary, even confusing, considering that you should use a NAL. grafik

(This is true for other attributes using a NAL as well.)

bertvannuffelen commented 10 months ago

@init-dcat-ap-de, when investigating this case I find there are inconsistencies on the SPDX side.

The property spdx:algorithm is stated in the documentation with IRI http://spdx.org/rdf/terms#algorithm while the html local reference is https://spdx.org/rdf/terms/#d4e46. Thus it is unclear which one it to be used. The explicit named one or the opaque one. Given that the html documentation states an explicit IRI I propose to take these ones as normative and use http://spdx.org/rdf/terms#algorithm.

image

Secondly the range of this propertu is defined by SPDX. That has been taken over in the DCAT-AP specification. In SPDX they define a codelist as instances of the spdx:ChecksumAlgorithm class.

image

This is the reason why the link to the codelist is pointing to this location.

In essence from the moment you use the property checksum for a Distribution, the rules of the spdx ontology apply. Similar as my response to the use of NALs for spatial (#266) , this is another case where codelists are encoded in a different way as commonly recommended to rely on SKOS. However this instance list still matches most of the requirements for controlled vocabularies.

Note that for this algorithm codelist there is no "official name" given by SPDX. The instances are just managed as part of SPDX in general. So to refer to the list we use reference to a html achor in the SPDX specification, which happens to look like a persistent URI.

bertvannuffelen commented 5 months ago

Ensured that the reference to the instances is still correct. Further no change has happened.