OBOFoundry
Document synchronization process with identifiers.org and n2t.net
Currently the process is not clear
Are OBO submitters expected to register with these resources prior to or after submitting to OBO? Or do these resources pull from OBO? Do these resources mutually sync? If so, how regularly?
How are conflicts to be resolved? We are currently considering a process that prospectively avoids conflicts (see #1519) but there should be some documented SOP for various scenarios, including grandfathered prefixes such as geo, or cases in the future
I think Identifiers.org has an automated API for submissions, but it's a black box so I wouldn't trust anything that doesn't have an associated GitHub issue or email thread with the maintainer (iirc, Henning).
I'm working on making a system for automating submissions to the Bioregistry that directly creates a pull request where review can be done in the open. However, the Bioregistry pulls from OBO Foundry on a nightly basis so there isn't actually any action necessary from the OBO Foundry side to keep the Bioregistry up to date with its content.
I would really like to get the governance/coordination of the prefixes part worked out first ;-). Is that what this ticket is about? Discussion about this on #1519 but got wandery. Where will the prefixes be curated as a central registry? originally this was going to be prefix commons, do we need to merge Bioregistry with prefix commons in some way? What happens to prefix commons? Agreed its gotten a bit ignored, but was a significant earlier effort with much good work done and coordination realized. I think the most important thing we need to decide is where and how we will be curating and governing prefix creation. Once that is done, synch should be straightforward.
Having not been part of earlier discussions, it wasn't apparent that there was any coordination between these services except the explicit mention that N2T and Identifiers.org are synced. I also got the feeling that many of these are quite opaque with respect to where the underlying data is stored, who's responsible for it, how it ends up listed on the site, and how the data is used to power the resolver. Maybe that is all documented somewhere, but not in a way that it's findable for a newbie like me
It is important to separate concerns
- Prefix Registration
- Prefix Resolution
- Prefix Search
- Prefix Context documentation
- Prefix Metadata
- Resource Metadata
On the issue of registration, I don't think it is wise to have another service in the mix (This list circa 2017 may be even larger now). It could be a last resort, if you really need a prefix that none of them will refuse to mint. However I would expect this to be a very rare exception indeed.
