NBO Styles and Conventions

[DitchingIt]

Most of the original structure for the NBO was created over a decade ago. I have been looking at where things are currently and some of the contents may need a bit of a dusting. In particular, I am interested in four contributions to developing principles:

1. OBO Foundry naming conventions
2. OBO Foundry textual definitions
3. Guidelines for writing definitions in ontologies
4. Chris Mungall's comments on the last resource

I would like to have a go at reviewing a sub-branch of the process branch and see what comments people have. If that seems worthwhile, I will put it as a new issue and keep this issue for a general discussion about labels and definitions in NBO.

[DitchingIt]

Just a note to add a resource:
5. Chris Mungall's comments on ontology mappings

[DitchingIt]

Having reviewed the stress related behavior sub-branch #123, I can see that this thread has the potential to generate questions about the NBO styles and conventions. I have copied a note from my review as a contribution to the general thinking:

• (S1) Historical definitions throughout NBO chop and change styles, and there does not seem to be control over NBO label vocabulary. Having, "a standard style... can make it easier for others to write definitions when creating new classes." Rather than rewrite the whole label system (with its own complications related to familiarity and inter-ontology consistency), I have opted to use hierarchical terms in the definitions. Although this introduces a greater risk of circularity through synonyms (G7), it offers
  • conformity: "Harmonize the definitions in the ontology in order that they all conform to a single set of typographical conventions and editorial guidelines" (G1.1),
  • controlled vocabulary,
  • consistency,
  • conventionality, e.g. "Do not replicate the name of the class in the definition,"
  • an opportunity to increase the currently sparse scattering of OWL definitions (S11), although I am not familiar enough with OWL to have attempted this.

[DitchingIt]

Attached is my proposed Convention sheet for editing old and adding new classes. It should be referenced in the contributing.md but doesn't change anything there (except the emphasis which needs to be a little stronger if we agree to use this as a touchstone).
Conventions 230128.pdf

• Should we implement the OBO Foundry policies in full?
• Should our next line of reference be Seppälä?
• Do we agree with my suggested exceptions?
• Do we agree with my suggested additions?
• Are there other issues to add to this or the contributing.md?

[DitchingIt]

See #16 regarding vague definitions.

[DitchingIt]

I have drafted an updated Definition Conventions Reference Sheet. In particular it

• adds reference to OBO Foundry Principle 13 (Notification of changes)
• gives advice on multi-parent classes.

Conventions 230731.pdf

[DitchingIt]

It was pointed out to me recently that whereas it's easy to disjoint a term in another ontology, doing so may actually affect how the other ontology works. Unless discussion here changes things, I propose:

• that NBO convention be NOT to disjoint terms in other ontologies.

[matentzn]

Nice initiative!! And, I am sure, quite lonely. Sorry I only see this now, if you need help on anything ping me!

[DitchingIt]

The current Ethology module editing project principally focuses on editing what exists within the NBO and only adding new terms which help with its interpretation.

A future project needs to look at adding new terms not just for the sake of taxonomic and habitat balance (if not completeness) but to tackle what Chris Mungall calls the single-child anti-pattern and the single-child-by-axis pattern.

[DitchingIt]

Whereas I am withdrawing from working on the NBO (see #124 ), I strongly recommend that the conventions sheet developed as part of this issue be adopted by any future editor, unless it be amended by consensus.