obophenotype / ncbitaxon

Build for NCBITaxon
BSD 3-Clause "New" or "Revised" License
24 stars 7 forks source link

Prefix #102

Open chrishmorris opened 3 months ago

chrishmorris commented 3 months ago

Thank you for your ontologies, especially ncbislim.obo.

It uses the prefix "NCBITaxon:". This does not conform to the standard for CURIEs, which should be lower case, and is not the one used by identifiers.org, which uses "taxonomy". (BioPortal, which is similar to you but also non-compliant.)

I would like to suggest "ncbi.taxon:" as a prefix that perhaps everyone can agree on.

There is some more discussion on this here: https://github.com/identifiers-org/identifiers-org.github.io/issues/171 and in some email correspondence which I will be happy to copy you in to.

Convergence on this would help make everyone's data FAIR.

chrishmorris commented 3 months ago

Correction. I see that this resolves: https://bioregistry.io/ncbitaxon:/9606. So "ncbitaxon" might be best.

matentzn commented 3 months ago

Hello @chrishmorris. Thanks for reaching out! This is a politically, and technically, a quite complicated issue unfortunately. There are a number of clashing assumptions here, and unfortunately there is no easy way out.

The ontology here, in essence a "database converted to an ontology", is a member of the OBO Foundry library of ontologies. As such, it is subject to its rules on prefixes and identifiers: https://obofoundry.org/id-policy.html.

This preference has also been enshrined into bioregistry (https://bioregistry.io/registry/ncbitaxon), but even if NCBI takes control of the preferred prefix, it wont be possible to easily change the prefix on the OBO side of things. Technically, we would have to change https://github.com/OBOFoundry/OBOFoundry.github.io/blob/master/ontology/ncbitaxon.md?plain=1#L21, which is pretty much impossible given how widely the obo-version of the ontology is used.

So, unfortunately, I dont think there is anything that can be done about this.. You could bring this issue to an OBO Operations Committee Meeting, but I doubt anyone would agree to a change like that due to the wide-spread of the NCBITaxon (which is widespread as it is part of the OBO IRI).