Case study: Author

[mslw]

This is about slots with two possible classes in their range - and about reusing schema.org definitions when we in fact rely on a modified subset of their properties.

Current state

Our Dataset class defines author slot as:

crc-schema-draft/src/sfb1451_schema.yaml

Lines 158 to 168 in 3a233f5

	author:
	slot_uri: schema:author
	required: true
	multivalued: true
	range: Author
	description: >-
	One or more entities (natural persons or organizations) that
	are considered authors of the dataset. These authors need
	not be identical to an author list of an associated
	publication. Any entity listed in this table will be
	credited on the catalog page of the dataset.

In https://schema.org/author (as in real life), an author is expected to be an Organization or a Person.

In SFB tabby, we expect an author to have name, email, orcid, affiliation.

We model that as an Author class:

crc-schema-draft/src/sfb1451_schema.yaml

Lines 238 to 272 in 3a233f5

	Author:
	class_uri: schema:Person
	description: >-
	Entity (natural person or organization) that is considered an
	author of the dataset. Dataset authors need not be identical to
	an author list of an associated publication. An entity listed as
	dataset author will be credited on the catalog page of the
	dataset.
	slots:
	- name
	- email
	slot_usage:
	name:
	required: true
	description: >-
	The full name of the author. For example, the name of a CRC
	member.
	email:
	required: false
	description: >-
	An email address with which the author can be contacted.
	attributes:
	orcid:
	slot_uri: obo:IAO_0000708
	required: false
	recommended: true
	description: >-
	ORCID of an author, to uniquely identify a researcher.
	affiliation:
	slot_uri: schema:affiliation
	required: false
	multivalued: true
	description: >-
	One or more names of organizations or institutions an author
	is affiliated with. Affiliations are free-form.

Similar case: DataController.

Issues

• Despite the description, we assign class_uri: schema:Person to our Author
  • I don't know how to express range: Organization or Person in linkml?
  • It would not be possible to distinguish organizations and persons from the sfb tabby input
• Despite class_uri: schema:Person, we only use a subset of https://schema.org/Person properties (name, email, affiliation) and define one additional field (orcid)
  • orcid could be a https://schema.org/identifier
  • we currently define the orcid field as https://purl.obolibrary.org/obo/IAO_0000708 ...
  • ... but it is a Class

Thoughts

Should our Author really have class_uri: schema:Person, or be our "own" definition of an Author which is related to a schema:Person? LinkML allows to define mappings, such as close_mapping which map to SKOS data model.

This would shows in the docs as:
URI: sfb1451_schema:Author
and then in the "Mappings section" as:

Mapping Type	Mapped Value
self	sfb1451_schema:Author
native	sfb1451_schema:Author
close	schema:Person

OTOH I am not sure what other practical meaning that would have. E.g. if what I care is feeding the data to the catalog, it's probably better to carry over the schema IRI?

[mslw]

A snapshot from a whiteboard discussion, to refresh our memories: