Consider distinguishing qualified relation and qualified attribution #80

New issue

Open

opened 2025-06-04 17:49:51 +00:00 by msz · 0 comments

msz commented

2025-06-04 17:49:51 +00:00

Owner

Trying to represent information about a dataset using the INM-7 schema, with catalog generation in mind, I noticed that:

DCAT distinguishes these two:
- qualified_relation ("Link to a description of a relationship with another resource"; range: dcat:Relationship)
- qualified_attribution ("Link to an Agent having some form of responsibility for the resource"; range: prov:Attribution).
DCAT's section 15. Qualified relations explains (in 15.1): "A general method for assigning an agent to a resource with a specified role is provided by using the qualified form prov:qualifiedAttribution from [PROV-O]."
The INM-7 schema only has qualified_relations; its range is a dlroles:Relationship, an object of which can be both Agent and Entity (thus covering both DCAT terms)
INM-7's qualified_relations (actually, dlroles:qualified_relations) declares an exact mapping to dcat:qualifiedRelation.

My concern: if coming from a catalog perspective (i.e. knowing DCAT but not INM-7 concepts) and trying to find Authors (Creators, Curators, etc.) I would look for something that maps to qualified_attribution, and not find it.

My question(s):

would it be beneficial to add a distinction for qualified relation and attribution in the INM-7 concepts (to improve the overlap with DCAT) or is it preferrable to keep the "capture both" meaning of the current qualified relation?
should dlroles:qualified_relations remain an exact mapping to dcat:qualifiedRelation, or should it be a close / related / narrow mapping to both dcat:qualified_relation and dcat:qualified_attribution instead?
or is this a non-issue / excessive concern about semantics?

Thanks in advance for a response -- this was a conceptual issue that kept me thinking. I have no practical experience with using mappings yet. And sorry if that has already been discussed (I remember that we had some discussion about using slots vs relations).

Initially described in inm7/annotate.inm7.de-data#13 (comment)

Trying to represent information about a dataset using the INM-7 schema, with catalog generation in mind, I noticed that: - DCAT distinguishes these two: - [qualified_relation](https://www.w3.org/TR/vocab-dcat-3/#Property:resource_qualified_relation) ("Link to a description of a relationship with another resource"; range: dcat:Relationship) - [qualified_attribution](https://www.w3.org/TR/vocab-dcat-3/#Property:resource_qualified_attribution) ("Link to an Agent having some form of responsibility for the resource"; range: prov:Attribution). - DCAT's section [15. Qualified relations](https://www.w3.org/TR/vocab-dcat-3/#qualified-forms) explains (in [15.1](https://www.w3.org/TR/vocab-dcat-3/#qualified-attribution)): "A general method for assigning an agent to a resource with a specified role is provided by using the qualified form prov:qualifiedAttribution from [PROV-O]." - The INM-7 schema only has [qualified_relations](https://concepts.inm7.de/s/base/unreleased/qualified_relations/); its range is a dlroles:Relationship, an object of which can be both Agent and Entity (thus covering both DCAT terms) - INM-7's qualified_relations (actually, dlroles:qualified_relations) declares an exact mapping to dcat:qualifiedRelation. My concern: if coming from a catalog perspective (i.e. knowing DCAT but not INM-7 concepts) and trying to find Authors (Creators, Curators, etc.) I would look for something that maps to qualified_attribution, and not find it. My question(s): - would it be beneficial to add a distinction for qualified relation and attribution in the INM-7 concepts (to improve the overlap with DCAT) or is it preferrable to keep the "capture both" meaning of the current qualified relation? - should dlroles:qualified_relations remain an exact mapping to dcat:qualifiedRelation, or should it be a close / related / narrow mapping to both dcat:qualified_relation and dcat:qualified_attribution instead? - or is this a non-issue / excessive concern about semantics? Thanks in advance for a response -- this was a conceptual issue that kept me thinking. I have no practical experience with using mappings yet. And sorry if that has already been discussed (I remember that we had some discussion about using slots vs relations). Initially described in https://hub.psychoinformatics.de/inm7/annotate.inm7.de-data/issues/13#issuecomment-4165