Meeting notes - August 24, 2022

[InKyungChoi]

Participants: Flavio Rizzolo, Francine Kalonji, Andrea Petres, Jenny Linnerud, Zoltan Vereczkei, Juan Munoz, Ayman Hathoot, Giorgia Simeoni, Wai Kit (Jackie) Si Tou and InKyung Choi

Issue #17 Registers in statistical production: their roles and types, and how to model them in GSIM

• This is a niche and underexplored area. We don't need to model the register maintenance, having a good clear definition that captures what we want in the model would be a priority
• To create a sub-group (Flavio, Zoltan, Dan, Jenny and Giorgia) to discuss and come back to GSIM group for reporting

Issue #6 Missing GSIM class – quality indicators

• Referential Metadata (RM) Subject should be generic (e.g., Statistical Program Cycle instead of "labour force survey" as it is mapped now). We need to leave open what RM Subject can be. If RM Structure changes depending on survey cycle, then it would be Statistical Program Cycle, if an entire domain shares the same structure, it would be domain.
• The current cardinality (1-1) between RM Subject and RM Structure is not right, the same structure can be re-used by multiple cycles, programs, or even domains (e.g., ESQRS can be used not only for labour force survey, but social survey, and so on). We can either change the cardinality from 1 to n, or introduce hierarchy in subjects (e.g., one RM Subject can be composed of other RM Subjects)
• The relationship and cardinality between RM Structure and RM Attribute seems to be wrong too (we would want RM Attributes to exist without RM Structure and re-use them for another RM Structures)
• Isn't RM Subject redundant? If RM Subject is supposed to exist to provide linkage mechanism, we can just link RM Structure directly to Identifiable Artefact, why RM Subject is needed? If we have additional information we want to store and organise in RM Subject, we might need to keep it. But currently, there is no additional attributes unique to RM Subject class anyway. We can remove RM Subject and have an attribute "subject" under RM Structure and link RM Structure with Identifiable Artefact.
• It is weird that RM Subject Item has composition relationship with RM Content Item, shouldn't RM Content Item be directly linked with RM Set (not with RM Subject Item), just like RM Attribute is linked to RM Structure? If we remove this relationship, we can also remove RM Subject Item as it is no longer in the middle of relationships.
• It would be helpful to have more specification on what can be RM Attribute or RM Content Item (e.g., value from codelist, number, freetext, table), indicating that RM Attribute can have different formats
• Name of class "RM Content Item" is confusing. If it is populated value for RM Attribute, why not RM Attribute Item (like RM Subject - RM Subject Item)?