`bibo:contribution`

[tombaker]

bibo:contribution
  definition  An agent and the context of their contribution to the resource.
  domain      dct:BibliographicResource
  range       bibo:Contribution

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

[tombaker]

Regarding the name -- dcterms:contribution vs dcterms:qualifiedContribution or some other name with a "meta" character -- it was pointed out on yesterday's call that BIBO already has at least one property-class pair with names that differ only by case (bibo:issue/bibo:Issue; Argie: and, as Osma pointed out, bibo:chapter/bibo:Chapter). The name dcterms:contribution happens to align with Bibframe. Also, I have been convinced that qualifiedContribution is jargony, which would require extra explanation, while a has name -- hasContribution -- is not generally distinct in meaning from properties without has -- eg, contribution-- and is therefore a pattern we should avoid.

A few alternatives were discussed:

• "A contribution of an agent to the resource." (Tom two weeks ago)
• "The agent that has a function in relation to the creation of the resource." (Karen two weeks ago)
• "An agent and their context in regard to the resource."
• "An agent in the context of their contribution to the resource."
• "The context of an agent's contribution to the resource."

EDIT: I was fine with "An agent and the context of their contribution to the resource" but have changed my mind.

[annakasprzik]

I still think, since the range of the property is not an agent, that "an agent" should not be the subject of the definition. Once again, I am putting up my suggestion for a vote:

• "The context of an agent's contribution to the resource." or even
• "The details/specifics of an agent's contribution to the resource."

[tombaker]

@annakasprzik I agree that "an agent" should not be the subject of the definition, so

The context of an agent's contribution to the resource.

is better.

[kcoyle]

The April model is more than the context -- it does include the agent: (from #97)

</project/2023/112/output/2025/1> a bibo:AcademicArticle ;
  bibo:hasContribution [ a bibo:Contribution ;
      bibo:hasAgent </person/Somebody> ;
      bibo:hasRole lcrel:aut ;
      bibo:hasAffiliation </org/SomeUniversity> 
] .

That the hasAgent value is a shape that contains the agent details doesn't change this. Agent statement is within contribution.

[tombaker]

@kcoyle I'm unsure if you are disagreeing with the definition "the context of an agent's contribution".

The binary relation between scholarly resource A and agent B (A dcterms:contributor B) needs additional context, so the contribution node X provides role and affiliation. The node is pointed to from A and points to B. (It has this directionality because SRAP takes the POV of describing scholarly resources.)

The April Model property bibo:hasAgent is what we are now calling bibo:contributingAgent. In defining "the context of an agent's contribution", the Contribution node points to the agent in question with bibo:contributingAgent (and in that sense, it "includes" the agent, and the "Agent statement is within contribution").

I'm not seeing a problem here - other than the obvious problem of explaining the pattern itself and its rationale.

[tombaker]

Would it help if we called the class bibo:contributionContext? Forgive me if this is something we have already discussed and rejected.

Or if we do not put "context" into the name of the class, might we refer to a Contribution-typed node as a "context node"? To me, the broader picture has to do with turning flat descriptions (ie, centered around one node) into somewhat more nested descriptions (with additional nodes) for the sake -- in this case -- of differentiating between essential relations (scholarly resource A to agent B) and more contextually bound information (role and affiliation).

[philbarker]

I think calling the class bibo:ContributionContext would help people understand what it is (I think it better represents the current draft definition) and avoid the case of the initial c being the only difference between the class and property. Even the best of us sometimes sometimes make mistakes with that.

[kcoyle]

@tombaker I think that calling it "context" makes it sound like it's just role and affiliation, which to me are the context relating to the agent. But the key thing is that the contribution node includes the agent, so it is agent + context. We began this because we didn't want the agent info together with the context info, which is what we originally had in SRAP. See lines 55-60 of the dctap.

The agent will always be in the contribution node, but the context information might not be recorded. Having the context be its own node ... makes some sense.

    flowchart LR
    bibo:contribution-->bibo:contributionContext
    bibo:contribution-->bibo:contributingAgent
    bibo:contributionContext-->bibo:contributionAffiliation
    bibo:contributionContext-->bibo:contributionRole

Loading

With that model, the contributing agent can be a shape with all of the agent information. The reason this makes sense to me is that I see the agent as being the primary/required information in the contribution node, and the context as less important and optional. However, this design suggestion may just be hair-splitting.

All that said, even with the April design I think we need a definition that makes clear that the agent is included in the contribution.

[annakasprzik]

Whereas I see it more abstract, in the passive voice: something got contributed to a resource, and the context then specifies who the agent of the contribution event was, what their role in the creation of the resource was, and what their affiliation was at that time.

[kcoyle]

I did think about modeling the contribution as an event, which could how your "passive voice" might be modeled. I'm not sure that the domain of BibliographicResource works with an event model, though. BibResource is a thing, not an event nor a statement of an action. I can't think of a good domain at the moment, but it would need to be in the sense of bibliographic resource creation.

[annakasprzik]

I am not sure why that would not work -- there is a resource and associated with it (expressed by the property) are potentially multiple contribution events that have led to the existence of said resource. That is exactly what domain and range of this property seem to say to me.