Closed apdavison closed 1 year ago
Please double check in how far this issue is not already covered by the core schema for Identifiers.org (for Pubmed cf. https://registry.identifiers.org/registry/pubmed)
identifiers.org is limited in scope to life sciences. As an example, it doesn't include https://hal.science identifiers (as far as I can see) or other national science repositories.
@apdavison should we then not define "type" / "name" in the schema but just identifier and the emitter? I think we should push people to register official identifiers through typed schemas (because they can actually validate the format of the identifier).
I agree that the typing all of them is though a bit cumbersome. The clean alternative for this would be the following:
schema for identifierType: (with controlled instance library in openMINDS)
schema for digitalIdentifier:
Research products would then only point to digitalIdentifier schema. Unfortunately this approach would cause a lot of change to the running system. It might be worth it though...
That would be a cleaner way to do it. I suggest we think about making that change for the next-but-one release. This schema is then a stop-gap, to provide the flexibility needed in https://github.com/HumanBrainProject/openMINDS/issues/51 but without defining a whole bunch of explicit identifier schemas.
The schema needs "type" or "kind", since a given Organization might emit more than one type of identifier.
@apdavison I agree with the plan to push this change to the next development round. (also because Oli would kill me if we do this now :see_no_evil: )
cf https://github.com/HumanBrainProject/openMINDS/issues/51