Rediscuss direction of link between ResearchProduct and ResearchProductVersion

[apdavison]

I would like to propose removing the property "hasVersion" from ResearchProduct and adding "isVersionOf" (or some similar name) to ResearchProductVersion**.

The reason is that with the current system, to add a new RPV requires a user to have write permissions for both the RPV and the RP, whereas if we reverse the direction, the user only needs to have write permissions for the RPV. The same goes when deleting a RPV.

In an EBRAINS context, this situation often arises when adding a new RPV (in a collab space) to an already-released RP (in a curator-only space), particularly for Models and LivePapers. The logic needed to manage all the possible situations, juggling a user KG client and a service client, gets very complicated. It would be greatly simplified if the link direction were reversed.

**(we might want to consider renaming "isNewVersionOf" and "isAlternativeVersionOf" at the same time, to avoid confusion, e.g. "isNewerThan" and "isAlternativeTo")

[lzehl]

@apdavison turning around the linkage between RPs and RPVs makes sense to me. Not sure why we did not set it up that way in the first place. We should re-discuss in this context again how to set up DOIs for the RPs and the inheritance rules between RPs and RPVs.

However, I think "isNewerThan" and "isAlternativeTo" are more confusing for relations between RPVs than "isNewVersionOf" and "isAlternativeVersionOf". For the latter it is always required to have editing writes on all alternative versions (I don't think there is a way around this).

[lzehl]

Let's have a look DataCite again (but I'm okay with what @apdavison suggested after the explanation)
https://datacite-metadata-schema.readthedocs.io/en/4.6/appendices/appendix-1/relationType/

[lzehl]

General discuss if version relations should be located in their own schema.

TODO 1:

• RPV/isVersionOf -> RP (instead of RP -> RPVs)

TODO 2:

• isNewerThan isSuccessorOf (isNewVersionOf) -> RPV (1)
• isVariantOf (isAlternativeVersionOf) -> RPV (N) [closer to datacite isVariantFormOf]
  (to be further discussed)