BIBO Ontology 2.0 -- a proposal

[tombaker]

BIBO Ontology 2.0 – a proposal

The UB chairs propose that we create a new version of the BIBO Ontology that is scoped only to properties and classes directly defined in the BIBO namespace, http://purl.org/ontology/bibo/. This will be the object of future maintenance. BIBO Ontology 2.0 will point back to BIBO Ontology 1.3 as the previous version.

The BIBO Ontology 2.0 will include everything in the current BIBO namespace subject to several exceptions and fixes:

• All declared properties will be included and will be of type owl:DatatypeProperty or owl:ObjectProperty.
• All declared classes will be included and will be of type owl:Class.
• Properties and classes will be described, as they are now, with rdfs:comment, rdfs:label, rdfs:range, rdfs:subClassOf, rdfs:domain, rdfs:subPropertyOf, owl:equivalentclass, owl:inverseof, owl:deprecated, and with various SKOS note properties.
• Two instances of owl:NamedIndividual in BIBO will be included: bibo:bdarcus and bibo:fgiasson.
• Assertions using rdfs:isDefinedBy will be fixed as per current usage.
• Metadata about the BIBO Vocabulary will credit bibo:bdarcus and fgiasson as creators and add DCMI.

The BIBO Ontology 2.0 will not include the following things from the BIBO Ontology 1.3:

• Controlled vocabularies for "degrees" and "status", each of which has its own namespace URI, which will be moved to their own standalone documents and defined as SKOS Concept Schemes.
• All assertions about OWL restrictions and union classes will be dropped. This will have the effect of dropping all assertions that touch on blank nodes.
• All assertions about terms in other namespaces will be dropped.
• The use of vs:term_status will be dropped.

[tombaker]

This proposal will be discussed on an open Zoom call on 17 February, after which the Usage Board will make a formal decision.

We encourage everyone to post comments about this proposal in this issue.

[niklasl]

On a technical/maintenance level, this should hold:

1. The BIBO namespace IRI is and will remain to be: http://purl.org/ontology/bibo/
2. That IRI does, and will remain to, denote BIBO — the bibliographic ontology.
3. It does, and will remain to, support content-negotiation; resolving to either an HTML or an RDF description of BIBO.
4. We propose to clean up the RDF description (per the issue description above).
5. We may make some changes regarding location of serializations, discovery of previous versions, etc.

(See also #5 for more details.)

[aisaac]

This plan sounds sensible. I have two questions about the things to be removed:

• "All assertions about terms in other namespaces will be dropped." Is this about scope notes? If yes, then where will these scope notes be kept, if they had any value to the understanding of the BIBO patterns?

• "Controlled vocabularies for "degrees" and "status", each of which has its own namespace URI, which will be moved to their own standalone documents." I reckon that these are on 'sub-'namespaces and thus you can do as suggested. But when a concept scheme has fewer than 10 elements and is meant to be used (and maintained) in tight alignment with the core set of classes and properties, I'm not sure there's much to gain by asking everyone to work with 3 documents instead of 1.

[niklasl]

@aisaac Good questions!

For the first: yes it is about the scope notes (among other things, such as "unwarranted" annotation property assertions). Perhaps defining e.g. a SHACL shape to describe such intended usage would be of use? (The scope note assertions aren't "scoped" to BIBO anyway, unless you e.g. track the association using named graphs, which really isn't a conveyable practise.)

For the second: I agree that it warrants discussion. How about two documents, one for the ontology and one for the two "mini" schemes?

[aisaac]

@niklasl thanks for the answers.

For the first yes it's always not great practice to put statements onto resources that are defined elsewhere. So a shape may be formally the best way to proceed.

For the second, I think I would be in favour of having just one document for everything, especially if it's not so long. But I agree that having two documents is already much better than three :-)

[tombaker]

BIBO 2.0 Proposal - Zoom call agenda

This agenda: https://hackmd.io/@dcmi-ub/bibo-proposal

Tuesday, 2026-02-17 1500 UTC

Zoom: https://us02web.zoom.us/j/89590350199

BIBO Ontology 2.0 – a proposal

• BIBO Ontology 2.0 -- a proposal #8

We will discuss a proposal to create a slimmed-down BIBO Ontology 2.0 document that describes (only) the properties and classes (and two individuals) that are identified with BIBO namespace URIs.

The term descriptions will remain unchanged with the following exceptions:

• Assertions of rdfs:isDefinedBy will be corrected.
• Assertions of vs:term_status will be dropped.
• Assertions about OWL restrictions and union classes will be dropped.

Note:

• All assertions about terms in other namespaces will be dropped.
• Also dropped will be two controlled vocabularies -- one for "degrees" and one for "status" -- each of which has its own namespace URI.

Questions for discussion:

1. "BIBO Ontology 2.0"

• Is the planned document appropriately named and scoped?

2. "Everything else" from BIBO Ontology 1.3 (current version)

• Where should the two controlled vocabularies be documented?
• Should the assertions about terms from other namespace be captured anywhere, eg in SHACL?

Issues discussed on this call will be considered by the Usage Board before deciding on the proposal.

[bdarcus]

I support this plan.

[kcoyle]

A BIBO (2.0) vocabulary is absolutely logical.

I believe that today's BIBO 1.0 could be developed as an Application Profile, possibly in DCTAP once the "superBIBO 2.0" exists. I think the hardest part will be determining if the "superBIBO 2.0" makes up a coherent vocabulary, which means analyzing where BIBO makes use of outside vocabularies. For this, examples and diagrams might be more informative than alphabetical lists.

[kcoyle]

Will it be necessary to use owl:DatatypeProperty and owl:ObjectProperty?

Is there any thought to using dcam:rangeIncludes/:domainIncludes instead? That would allow the widest re-use of the properties.

Will there be relations between the 1.3 and the 2.0, à la DC 1.1 and DC Terms. (Yes, I realize this is basically the opposite case, since 2.0 is a kind of reduction.)

[tombaker]

Re: calling BIBO 2.0 an "ontology" or a "vocabulary"...

Data Catalog Vocabulary (DCAT), which is now in Version 3, has consistently been called a "vocabulary" since 2014.

The landing page for the DCAT 3 Vocabulary namespace document says: "This is a namespace for use with Data Catalog Vocabulary (DCAT) Version 3", and: "The vocabulary is published both in human- and machine-readable forms."

There is some confusion where it points back to earlier versions under the heading "Older Versions of Ontology files". It says: "The ontology for DCAT 2 can be seen in Turtle or RDF/XML or JSON-LD", even though the page for DCAT 2 says it is a vocabulary. It does not look to me like the latest version is anywhere actually referred to as the "DCAT ontology", though it does say clearly: "The DCAT vocabulary is an OWL2 ontology".

Note that the DCAT editors faced precisely the situation we face with BIBO. DCAT also provided "additional documentation or usage notes" about "externally defined terms". They solved this by putting the additional documentation to a separate Turtle document: "The RDF for externally defined terms where DCAT has provided additional documentation or usage notes can be found at dcat2-external.ttl."

In my opinion, the DCAT editors get the messaging around DCAT -- as vocabulary, ontology, and namespace -- exactly right. It provides a good model to follow for BIBO.

[kcoyle]

I prefer vocabulary, and to reserve "ontology" for those efforts that attempt to be comprehensive of a discipline or part of a discipline. I see BIBO as designed to be practical and useful but without a claim to be formally correct.

I still want to see whether the BIBO namespace elements feel "complete" on their own or if, without the terms from the other namespaces, BIBO has an air of being a bit random. I suspect the latter, which could be another argument for going with "vocabulary".