Lack of Identifiers and Semantic Interoperability
Closed this issue · 2 comments
Feedback from our ALA colleagues: 'Within the general Darwin Core structure there is a tendency to allow an identifier to be associated with a more general textual term. For example, datasetName and datasetID from Darwin Core or measurementType and measurementTypeID from the extended measurement or fact extension. The measurementType would contain a string such as “area” and measurementTypeID would contain the URI “http://vocab.nerc.ac.uk/collection/P24/current/AREA/”. If a description can be unambiguously associated with a semantic URI, the inventory can be used by machines to aid search and discovery. A resolvable URI also aids user comprehension since it allows a user to locate additional contextual information about a term. As a general principle, any term that is associated with a formal vocabulary should have an associated ID-term. For example, identifierTypes and identifierTypeID Similarly, anything that might reference an external definition should also have an associated ID-term that can be used to locate a formal description of the definition. For example, samplingEffortProtocol and samplingEffortProtocolID.'
Response from Ming: 'Please see similar/related issues raised: Request to include iri terms in IPT - gbif/ipt#1947 → Why we deliberately chose not to mint the *ID term - #83'
Additional comment(s) from Task Group discussion: anything that can have a controlled vocab will have an IRI term so there is no need to create a new terms for IDs. Ideally we do not want to continue to add duplicative terms.