Responsible organisation role: why multiple occurence?

[PeterParslow]

Responsible organisation role has "multiple" occurrence, but it is an implementation of ISO 19115's CI_ResponsibleParty.role, which can only occur once.

This has been the case since GEMINI 2 in 2009. Perhaps it was prescient in that ISO 19115-1 sort of supports this, because many different "responsibilities" (point of contact, maintenance contact, distributor, constraints, releasability...), can point to one "party".

But given that GEMINI adheres to the earlier ISO 19115, perhaps this should always have been single? It's impossible to encode multiple roles, so noone could have implemented multiple roles even when they wanted to.

[nmtoken]

Good spot, it should be one in the current GEMINI standard.

[PeterParslow]

I think there used to be a concept of "GEMINI that isn't ISO 19115" - so "GEMINI" wanted multiple roles, but "GEMINI in ISO 19115" could only have a single one. But years ago, we locked GEMINI into ISO 19115 via INSPIRE ... so until that changes, I guess we're stuck with 'one'.

[nmtoken]

In ISO 19115/19139 XML you can encode multiple roles for an organization by having multiple CI_ResponsibleParty elements, (each with one role). The guidance perhaps confuses the GEMINI abstract data model, with the encoding?

[PeterParslow]

We (GEMINI WG) need to decide whether the "GEMINI abstract data model" is tied to the ISO 19115:2013 conceptual schema, or whether that is a step in the encoding of GEMINI.
(In ISO/TC 211, ISO 19139 provides the encoding of the 'schema' provided in ISO 19115. ISO 19115 isn't actually explicit about what level of abstraction it models - something to fix! - but in ISO 19101 / 19103 terms, I'm pretty sure it's a conceptual schema/abstract schema)