`bibo:contributorAffiliation`

[tombaker]

bibo:contributorAffiliation
  definition  [The agent's institution or group in the context of the resource.]
  domain      bibo:Contribution
  range       dct:Agent

See also #100 (comment) for a comparison of proposals.

[tombaker]

On the call, we discussed:

Tom>   The affiliation of the agent who made a contribution to the resource.
Karen> The agent's institution or group in the context of the resource.

I still think we need to use a contrib* word in the definition, like this:

The agent's institution or group in the context of their contribution to the resource.

[annakasprzik]

I think we should use the indefinite article "an agent's [...]".

[kcoyle]

We did define affiliation in SRAP as having a range of IRI or BNODE and a value shape of organization. The organization shape, though, allows a literal institution name. When people are cataloging with the item in hand the affiliation will be a text. I don't have a strong enough use case that would justify searching for an identifier for the institution.

I think that our use of IRI or BNODE in the valueNodeType is wrong, and that cell should be blank when there is a valueShape value. @philbarker ?

[philbarker]

@kcoyle you could allow IRI, BNode or Literal values for affiliation in the TAP. However most implementations would need to be either Literal or non-literal (e.g. Object or String in JSON-LD). Using a BNode with no property other than name with a Literal valueType need not look any different to an end user than using a Literal directly. I would stick with what you have (affiliation as BNode or IRI). That would then allow data providers who do have an IRI for the Organization to use it and those who do not to just use the name, and implementers not to have to worry about trying to accommodate both Literal and Non-Literal values.

[kcoyle]

Thanks, @philbarker . What I think we were thinking is that the valueShape would be either an IRI or BNODE. Leaving it blank (which we've done in examples in the past) makes the DCTAP less oriented to RDF and more open to, for example, XML solutions. We do seem to be defining SRAP as an RDF profile, so perhaps our usage is appropriate.

[philbarker]

@kcoyle that's maybe a slightly different to what I was thinking of, but I think my answer is the same. I remember we didn't define defaults for TAP, but I think most people would interpret a blank cell for valueShape as "no constraint, anything goes". In RDF, I think IRI or BNode covers either a link (with or without a name) or a name with no link (in a BNode) in a way that causes fewer problems that allowing literal values.
I guess the follow-on question is, if you allow IRI and BNode do you want to make a name mandatory even when you have a an IRI? Having the name is not necessary if you have the IRI, but it would be easier to show in the TAP and would help data consumers.

[tombaker]

PROPOSED

bibo:contributorAffiliation
  definition  An institutional or group affiliation of the agent.
  domain      bibo:ContributionContext
  range       dct:Agent

[annakasprzik]

The institution or group the agent was affiliated with at the time of the creation of the resource.

[tombaker]

An institution or group with which the agent was affiliated when the resource was created.
An institutional or group affiliation of the agent.

[annakasprzik]

DECISION (5 upvotes on June 23rd, 2026)

bibo:contributorAffiliation
  definition  An institutional or group affiliation of the agent.
  domain      bibo:ContributionContext
  range       dct:Agent