Document policy on identifiers

[cmungall]

Should NMDC mint new identifiers, or reuse? Does this depend on the source? Do we also maintain mappings, e.g. between gold and ncbi?

[cmungall]

I will link ESS-DIVE/ISGN presentation here later

[cmungall]

When we use gold IDs etc we should always use CURIEs/prefixed IDs, and have these registered

[wdduncan]

Do we want to identify a namespace curie for the URIs?
e.g., nmdc: -> http:purl.nmdc.org

We use append custom obo-style ids.
e.g., nmdc:GOLD_0001, nmdc:EMSL_00001, nmdc:ESS-DIVE_00001 etc.

[cmungall]

Yes, we should use jsonld contexts to map prefixes

but we shouldn't mint new IDs for GOLD. We can register GOLD etc on n2t/identifiers.org

[cmungall]

Update to the above:

but we shouldn't mint new IDs for GOLD. We can register GOLD etc on n2t/identifiers.org

Reddy did this (https://registry.identifiers.org/registry/gold), so identifiers will now resolve

E.g. the CURIE GOLD:Gp0119849

Can be resolved via

http://identifiers.org/GOLD:Gp0108335

This doesn't work on nt2.net though, I will follow up with them

[aclum]

@cmungall anything left to do here?