Adding adoption declaration as a requirement to principle 8 (original suggestion: P3)

[matentzn]

While I do not see any land in sight on #2324, I think one thing that is pretty uncontroversial is that we should declare adoptions. I would like to propose to implement our previous decision to use rdfs:isDefinedBy to denote where a term belongs by making it mandatory to declare adoptions (if you don't know what I am talking about, read #2324). For example: BFO:0000050 is adopted by RO. I would like to suggest for practical reasons that we only make such declarations mandatory in cases od adoptions, not in general for all terms. We can do the latter in a separate vote, but it would effect too many ontologies in OBO Foundry that don't adopt or don't care about any of this.

Here is what I would concretely like to add to principle 3.

If a term previously defined in an identifier space belonging to an ontology A (e.g. http://purl.obolibrary.org/obo/A_123) is adopted by an ontology B (with a different identifier scheme, e.g. http://purl.obolibrary.org/obo/B_123) we should add the following annotation assertion:

http://purl.obolibrary.org/obo/A_123 rdfs:isDefinedBy http://purl.obolibrary.org/obo/b.owl

[nataled]

EWG discussion: 1) What happens when the ontologies have different licenses such that ontology A is CC0 and ontology B is CC-BY? 2) It is possible that the published releases are shifted in time such that it could be that the adopted term appears in both ontologies, or neither for a period of time. 3) How is this done for OBO format? Is this possible? Some ontologies are natively OBO format. 4) P3 refers to URIs at the ontology level, not at the level of specific terms. The closest principle appears to be P8 Documentation, the purpose of which reads, in part "Central to the issue of ontology documentation is ensuring transparency and traceability of artefact development." We could add a section specific to term metadata documentation. Would this work?

[matentzn]

What happens when the ontologies have different licenses such that ontology A is CC0 and ontology B is CC-BY?

Good question, in my view (with the little study of law I did during my undergrad), the identifier and all its metadata move into the license space of ontology B, effectively restricting access.

It is possible that the published releases are shifted in time such that it could be that the adopted term appears in both ontologies, or neither for a period of time.

Yes.

How is this done for OBO format? Is this possible? Some ontologies are natively OBO format.

It is the same; OBO format allows arbitrary annotation assertions, so no problem here.

P3 refers to URIs at the ontology level, not at the level of specific terms. The closest principle appears to be P8 Documentation, the purpose of which reads, in part "Central to the issue of ontology documentation is ensuring transparency and traceability of artefact development." We could add a section specific to term metadata documentation. Would this work?

I think this makes sense! :) Thank you!

[nataled]

@matentzn for our question regarding how to do it in OBO format, we meant literally what should appear in the stanza? http://purl.obolibrary.org/obo/A_123 rdfs:isDefinedBy http://purl.obolibrary.org/obo/b.owl is not valid OBO format.

[nataled]

@matentzn further indicates that a CC0 ontology cannot redistribute more restrictive content. In the principle, we can state that ceding a term to another ontology also explicitly means also that the term will henceforth be distributed under the license used by the adopting ontology.