Game plan: What should happen on 01.01.2024 when passing Dashboard becomes mandatory? about obofoundry.github.io

[nlharris]

An intermediate step could be to add an icon (a medal? 🏅) to the ontologies that are passing the dashboard on 2024-01-01, and allow a grace period before moving to the split dashboard.

I assume all the ontology developers will be contacted well before that date to alert them about the upcoming change and give guidance on how to pass the dashboard?

[matentzn]

Personally, I think the 6 year grace period with huge red X's on the dashboard, and all ontologies already have a OBO Dashboard badge on their site: https://obofoundry.org/ontology/cob.html. But yeah, if there is a majority that wants this a bit more gentle I guess we can wait a bit longer. Just for the record, we have given 3-4 presentations about the OBO Dashboard and told people about the deadline (Town Hall, Ontology Summit, ICBO Workshops); we will add this as a major item to the newsletter as well.

[pfabry]

I think the passing/not passing dichotomy is too coarse. Some red crosses are redder than others, and splitting the dashboard along those lines would put in the same category an ontology that can't be reasoned on with one that has multiple labels for a term (Granted, multiple labels is an easy fix and shouldn't be a problem).
I would lower the threshold on what is passing, but I would be more strict when an ontology fails it (cf. issue #2348 )
Ultimately the question is what is the purpose of the OBO Foundry? Is it a biomedical ontologies repository or a biomedical ontologies standard organization ? It can't be both in my opinion, and I think there is a bit of "schizophrenia" right now between the standards we enforce for the new ontologies (which is a good thing) and the ones we adopt for the ontologies already accepted.

[matentzn]

I think the passing/not passing dichotomy is too coarse.

We do need some kind of threshold to develop "carrots and sticks".

Some red crosses are redder than others, and splitting the dashboard along those lines would put in the same category an ontology that can't be reasoned on with one that has multiple labels for a term (Granted, multiple labels is an easy fix and shouldn't be a problem).

It's true that some crosses are redder than others. But we have with year long deliberation set a very very small bar for stuff that is fail. I think we can debate certain determinations on a case-by-case basis. Maybe we need to relegate some tests that are currently "errors" to "warning" - and I 100% think this should be done.

I would lower the threshold on what is passing, but I would be more strict when an ontology fails it (cf. issue #2348 )

Yes, but on a test by test basis. Right now, what is considered an error has been determined by the OBO operations committee in the last year as minimal. If you see specific red things you do not want to be considered as errors, you should report them on individual issues.

Ultimately the question is what is the purpose of the OBO Foundry? Is it a biomedical ontologies repository or a biomedical ontologies standard organization ? It can't be both in my opinion, and I think there is a bit of "schizophrenia" right now between the standards we enforce for the new ontologies (which is a good thing) and the ones we adopt for the ontologies already accepted.

I think we are and should be the latter - a standards organisation. Our goal should be to promote standard practices and make ontologies more interoperable. Some of us also are ontology providers and build repositories, but that is separate from the OBO mandate. I believe we should consider ourselves a bottom up standard body.