mixed case identifiers

[micheldumontier]

I worry about the case sensitivity of curies such as
NCBITaxon:9606

i would recommend that identifiers adopt a single case (upper or lower), as it would likely reduce errors. In Bio2RDF, we adopt lower case.

[cmungall]

I'd like to adopt this for all OBO ontologies moving forward, apart from the ones grandfathered in (e.g. many of the anatomies).

For bioinformatics databases we've historically been using mixed case, especially for the MODs: FlyBase, WormBase, PomBase, dictyBase. This is the metadata that GO, UniProtKB/GOA and the MODs have been using https://github.com/geneontology/go-site/blob/master/metadata/db-xrefs.yaml - it goes back over a decade so it would be hard to reconfigure. But we have to think about it for synchronization with MIRIAM.