Hi Markus,
I agree we can add "subject" and "bibliographicalCitation" to the list of predefined variables. dct:modified is already present in "lastUpdated".
We never considered multiplicity of environment variables or relationships between them. In principle they should be just single independent values. My suggestion is to keep using this simple approach, unless we really need more complex functionality.
Anyway, adding more values to the official list of variables is the kind of change that we should always be able to do without bringing impact to the existing implementations. Perhaps we could be more strict by adding a prefix to all TAPIR predefined variables and then defining the corresponding type as the union between the list of prefixed values and any string that does not start with the prefix. This way at least we completely avoid name clashes with custom variables.
I'm using simple rules to get the content of technicalContactName and technicalContactEmail. I just get the first system administrator of the first technical host entity. In the case of contentContactName and contentContactEmail I get the first data administrator of the first data supplier entity. The specification should also be explicit about this.
Let me know if you have other ideas.
Best Regards, -- Renato
Renato, as you might have guessed all your proposed changes are fine with me.
Regarding the metadata variables I would make all existing metadata values available as TAPIR variables. Maybe we can leave out the indexing preferences. This would mean to add dc:subject, dct:bibliographicCitation and dct:modified
I am not sure yet how you will retrieve metadata values for RelatedEntities though. There might be multiple and do you have a simple rule in mind what to do in that case? That's the case for all contact data.
cheers
Markus