Class: Dataset Property: Other Identifier is not necessary

[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 & property

Issue: Class: Dataset Property: Other Identifier is not necessary

• https://doi-do.github.io/dcat-us/#properties-for-dataset
• https://doi-do.github.io/dcat-us/#dataset-otherIdentifier
• https://doi-do.github.io/dcat-us/#dataset-identifier
• Class: Dataset Property: Other Identifier is not necessary. The Class: Dataset Property: Identifier has a cardinality of 0..n, meaning it is repeatable. This makes "Other Identifier" redundant.

Recommended change(s): Please remove Class: Dataset Property: Other Identifier and update documentation.

[dafeder]

I think the better solution is to reduce the cardinality of Identifier to 1. Identifier is restricted to string literals which makes sense to simply present the catalog's canonical identifier for the record, while Other Identifier accepts full ADMS Identifier objects, which can contain contextual information such as the issuing agency and a scheme IRI/identifier. If all identifiers are presented in an array it will not be clear which is the primary identifier for the catalog.