versioning check is too strict

[balhoff]

The OBO Foundry versioning principle says:

Version identifiers MUST either be of the form “YYYY-MM-DD” (that is, a date) OR use a numbering system (such as semantic versioning, i.e, of the form “NN.n”). Each version MUST associate with a distinct official release. The date versioning system is preferred, as it meshes with the requirement that version IRIs be specified using dated PURLs (see below).

However in the 2023-11-21 dashboard report it flags the CHEBI version IRI as in error: "Version IRI has neither a semantic version nor a date". The current CHEBI version IRI is http://purl.obolibrary.org/obo/chebi/227/chebi.owl. This resolves properly to the right ontology release. This version IRI uses a numbering system as required. In my reading, precise adherence to semantic version is not required.

[anitacaron]

The specification for the semantic versioning is of the form NN.n. It's expected to use points between numbers. That's why CHEBI is not valid.

[nataled]

There is a misunderstanding here. The text of the principle states: "Version identifiers MUST either be of the form “YYYY-MM-DD” (that is, a date) OR use a numbering system (such as semantic versioning, i.e, of the form “NN.n”)."

I added the emphasis on 'such as'. In any case, it indicates that semantic versioning is an example of a numbering system, but it is the numbering system that is important, not the form it takes. So CHEBI is indeed valid.

[anitacaron]

Ok, I'll add the validation for the CHEBI numbering system.