Open matentzn opened 2 years ago
You may want to consider if you want to distinguish between the person who asserted the mapping and the person who represented it in SSSOM.
PAV does this with
author
the agent that asserted the mappingcreator
the agent that put it into its published form PAV is a good resource for clarifying the roles, thank you @AlasdairGray !
Another resource to consider is Contributor Role Ontology (CRO): https://data2health.github.io/contributor-role-ontology/ cc @mbrush
For the first version, we will probably go with @AlasdairGray's suggestion. We map the author to pav, but because of historical reasons continue to map the creator to dc:creator. We also add a new category of reviewer:
creator_id:
description: Identifies the persons or groups responsible for the creation of
the mapping. The creator is the agent that put the mapping in its published form,
which may be different from the author, which is a person that was actively involved
in the assertion of the mapping.
Recommended to be a (pipe-separated) list of ORCIDs or otherwise
identifying URLs, but any identifying string (such as name and affiliation)
is permissible.
slot_uri: dc:creator
range: uriorcurie
multivalued: true
creator_label:
description: A string identifying the creator of this mapping. In the spirit of
provenance, consider to use creator_id instead.
range: string
multivalued: true
author_id:
description: Identifies the persons or groups responsible for asserting the mappings.
Recommended to be a (pipe-separated) list of ORCIDs or otherwise
identifying URLs, but any identifying string (such as name and affiliation)
is permissible.
slot_uri: pav:authoredBy
range: uriorcurie
multivalued: true
author_label:
description: A string identifying the author of this mapping. In the spirit of
provenance, consider to use author_id instead.
range: string
multivalued: true
reviewer_id:
description: Identifies the persons or groups that reviewed and confirmed the mapping.
Recommended to be a (pipe-separated) list of ORCIDs or otherwise
identifying URLs, but any identifying string (such as name and affiliation)
is permissible.
range: uriorcurie
multivalued: true
reviewer_label:
description: A string identifying the reviewer of this mapping. In the spirit of
provenance, consider to use author_id instead.
range: string
multivalued: true
Should we suggest reviewer as a new property in PAV?
After discussions with @NuriaQueralt, its clear we need to distinction a few more roles than just "creator".
Does this make sense?