Agent-related properties and classes

[tombaker]

Four of the open issues are about creating terms in the dcterms: namespace that are related to agents: dcterms:Person, dcterms:Organization, dcterms:name, and dcterms:affiliation.

For three of these, BIBO itself already uses terms from FOAF -- foaf:Person, foaf:Organization, and foaf:name. FOAF has no property for affiliation.

Putting four new properties into dcterms: would send, to my way of thinking, a confusing signal about the intended scope of DC Terms. If properties and classes about agents are in scope, why would we in principle stop at just three properties and not mirror all (or most) of FOAF into DC Terms?

If on the other hand we were to create these terms in the BIBO namespace, would we not then want to create BIBO terms for all of the FOAF terms used in BIBO? (These are: foaf:Agent, foaf:based_near, foaf:depiction, foaf:Document, foaf:family_name, foaf:givenname, foaf:homepage, foaf:Image, foaf:name, foaf:Organization, and foaf:Person.) Perhaps the new BIBO terms could be mapped to their FOAF equivalents.

Alternatively, the SRAP WG could decide simply to use the FOAF terms, as BIBO does, but this would leave the awkward problem of affiliation, which does not exist in FOAF.

The path of least resistance, perhaps, would be for SRAP to use foaf:Person, foaf:Organization, foaf:name, and bibo:affiliation.