Proposal for adding property "identifier" (dct:identifier) to Class "Agent" (foaf:Agent)

[jimjyang]

dct:identifier (range rdfs:Literal) preferably as "recommended" (0..1)

[andrea-perego]

In case it could turn to be useful, identifiers for agents are supported in the DCAT-AP profile used in the JRC Data Catalogue (https://data.jrc.ec.europa.eu/) for multidisciplinary research data. BTW, the JRC profile of DCAT-AP uses for this purpose both dct:identifier and adms:identifier.

A discussion on the underlying requirements:

https://ec-jrc.github.io/dcat-ap-jrc/#req-identifiers

Identifiers for agents (organisations and individuals):

• dct:identifier:
  • https://ec-jrc.github.io/dcat-ap-jrc/#agent-organisation-identifier
  • https://ec-jrc.github.io/dcat-ap-jrc/#agent-individual-identifier
• adms:identifier:
  • https://ec-jrc.github.io/dcat-ap-jrc/#agent-organisation-other-identifier
  • https://ec-jrc.github.io/dcat-ap-jrc/#agent-individual-other-identifier

There are also a couple of examples showing how they are used:

https://ec-jrc.github.io/dcat-ap-jrc/#ex-agent-organisation

https://ec-jrc.github.io/dcat-ap-jrc/#ex-agent-individual

[bertvannuffelen]

what is the rationale for this request? In reference provided by @andrea-perego https://ec-jrc.github.io/dcat-ap-jrc/#req-identifiers you find a nice overview on the rationale applied by JRC. Covers that your request? As a good guideline or best practice explanation.

In addition I point you to an recommendation on joinup: https://joinup.ec.europa.eu/release/dcat-ap-how-use-identifiers-datasets-and-distributions

In order to make it recommended we should have the usecase clear. For me, identifiers are crucial to facilitate the data exchange, but I am not sure that recommendation of a property yields the desired result. Maybe a good best-practice document extending the joinup note and the JRC insights have more impact.

[jimjyang]

Sorry for my late response, @bertvannuffelen, and sorry that I didn't provide any rationale when I wrote this proposal.

Yes, the reference provided by @andrea-perego very well covers my request. Thanks, @andrea-perego!

[bertvannuffelen]

To the community:

This topic is also related with #187

• do we need to stress the usage/presence of identifiers of the key entities in DCAT-AP? A starting point for the usage might be https://w3c.github.io/dxwg/dcat/#dereferenceable-identifiers

In short, a proposal on this should decisions on the following:

• every entity is identified with an identifier
• the main/prime identifier is present in dct:identifier, other identifiers are present in adms:identifier (other identifiers)
• if no dct:identifier is present, the RDF technical identifier (rdf:about) is used a the identifier for identity processing
• by preference the RDF technical identifier and dct:identifier coincide as a dereferenceable identifier, if not coinciding then if dct:identifier contains a universal unique identifier in another scheme, this can be used as the prime identifier (e.g. a uuid), otherwise the RDF technical identifier is probably stronger and should be used. as the prime identifier for identity processing.

[bertvannuffelen]

decision from wg of 21 oct 2021: postpone the resolution of this issue to have a broader discussion on identifiers.

[bertvannuffelen]

We propose to close this issue as in release 3.0.0 in section on Identifiers a reference is made to the outcome of a community discussion on identifiers.

In case new use-cases or expectations or harmonisation proposals are needed we propose to open a new issue and start the discussion from that section.