version tagging

[carueda]

(related entry in old system: mmisw/mmiorr#19)

The general version tagging needs some careful analysis. For example, we should probably "respect" any original version tagging --like in the set of SWEET 2.3 ontologies. What we currently call "version" in the ORR is basically just a submission timestamp. Users should have the option to explicitly indicate the version string, etc.

[carueda]

@graybeal any comments?

[graybeal]

I've engaged in 3 different ontology versioning discussions/threads in the last 24 hours, counting my update of https://marinemetadata.org/apguides/ontprovidersguide/ontguideconstructingiris, so I suppose I should pony up.

I agree users should have the option of setting the version attribute if they are submitting on ontology; it is a best practice we should not override. I think the demand is not strong for using the standard OWL(?) versioning in the vocabularies that MMI ORR stores; however, since there will likely be a few people who want to manually version their ontologies, there is no reason to be inconsistent with the way the standard feature is used in ontologies. However, we can confidently predict, based on plenty of evidence in BioPortal and in life, that the version string as set by users will be unusable for unambiguously defining the version of the ontology.

The mechanism that we originally called versioning, as we originally intended for application to both ontology IRIs and term IRIs, is essential to the ability to persistently reference a unique version of any ontology or term, as corresponds to its submission to the portal. If we cede to creators of original ontology the standard attribute, we should define our own attribute (say, submissionID) which can be used to define the deterministically unique version of the ontology and its terms.

[graybeal]

Note the information re versioning with OWL 2:

• https://www.w3.org/2007/OWL/wiki/Ontology_Versions
• https://www.w3.org/TR/owl2-syntax/#Ontology_IRI_and_Version_IRI
• https://www.w3.org/TR/owl2-syntax/#Versioning_of_OWL_2_Ontologies

[carueda]

Thanks for the pointers. In particular I'm happy to see very good alignment in ORR to what's said in https://www.w3.org/TR/owl2-syntax/#Versioning_of_OWL_2_Ontologies. But I still have to look more closely and see what might be appropriate to incorporate as key basic functionality in the ORR.

(Since your previous comment includes aspects of "best practices", just one general observation. In v3 of the ORR I have purposely attempted to make it even more flexible, that is, not enforcing any particular practices --especially regarding ontology submissions-- while of course providing, as best as we can, the key basic functionality to support a variety of likely practices.)

[carueda]

@lewismc fyi the referenced ticket briefly mentioned today.