Confusing Example for Class "Identifier", Property "schemaAgency"

[leightonlc]

This information will help the Core Team resolve issues with the draft DCAT3-US application profile. So we can track all issues, please submit different issues in separate issue tickets. The Core Team will review and adjudicate the submitted issue and provide comments on this ticket indicating steps taken to address the issue as submitted.

Name: Leighton Christiansen

Affiliation: Centers for Disease Control and Prevention (CDC)

Type of issue: Schema: Class

Issue: Confusing Example for Class "Identifier", Property "schemaAgency"

• https://doi-do.github.io/dcat-us/#properties-for-identifier
• https://doi-do.github.io/dcat-us/#identifier-schemaAgency
• https://doi-do.github.io/dcat-us/#other-identifer-example
• https://github.com/DOI-DO/dcat-us/blob/main/docs/examples/identifier.jsonld
• _I believe that in the examples for "Identifier" on the Candidate Snapshot https://doi-do.github.io/dcat-us/#other-identifer-example and in GitHub https://github.com/DOI-DO/dcat-us/blob/main/docs/examples/identifier.jsonld, the value in the Property "schemaAgency" is incorrect. The Property: schema agency https://doi-do.github.io/dcat-us/#identifier-schemaAgency does not have a definition. However, I am sure that what is meant here is the organization that is responsible for the creation and maintenance of the Persistent Identifier specification or standard. This "schemaAgency" may or may not be the same as the Identifier: Creator.
• In the examples posted https://doi-do.github.io/dcat-us/#other-identifer-example we have the following fields:
• a adms:Identifier ; adms:schemaAgency "Department of Interior" ; skos:notation "10.12345/ex.doi.67890"; dcat:version "1.0" ; dcterms:issued "2023"^^xsd:gYear. The value in the "skos:notation" field is a Digital Object Identifier (DOI). The organization that maintains the DOI schema is called the "DOI Foundation," see https://www.doi.org/. The Department of the Interior (DOI) does NOT maintain the Digital Object Identifier schema. However, Interior probably is the "creator" of the identifier that they cause to be associated with a specific dataset. Now, if the skos:notation value was referring to an "in-house" research project or output identifier assigned within the Department of the Interior research office, then yes, Interior could be both the "Creator" and the "schemaAgency." However when talking about Digital Object Identifies, Interior would not be the "schemaAgency" in the way that I have defined._ -This same misuse is present in the GitHub example at https://github.com/DOI-DO/dcat-us/blob/main/docs/examples/identifier.jsonld.

Recommended change(s): In the Property: schema agency https://doi-do.github.io/dcat-us/#identifier-schemaAgency please add the definition "the organization that is responsible for the creation and maintenance of the Persistent Identifier specification or standard." In the Property: Creator https://doi-do.github.io/dcat-us/#identifier-creator please add the definition "The entity which caused the identifier to be assigned to the dataset" or something similar. To the Identifier example https://doi-do.github.io/dcat-us/#other-identifer-example please make the DOI Foundation the "schemaAgency" and make Department of the Interior the "Creator". Please update all relevant examples in GitHub. -It would probably also be a good idea to add a Property called "identifierType" to help with confusion between maintained standards and agency-created internal IDs. I will create a separate ticket for that. See https://github.com/DOI-DO/dcat-us/issues/205

[dafeder]

Agree with this clarification. Also there is no apparent way in the example to actually provide an IRI for the identifier scheme itself. A better (IMO) example is provided at:

https://semiceu.github.io/Core-Business-Vocabulary/releases/2.2.0/#examples

Standardizing on a conformsTo field would provide a much more straightforward way to provide DOIs as Identifier objects. I'm not sure what the best IRI to provide would be but anything that actually resolves would be an improvement?