Decide license

[sbliven]

What license should we use for the OSC-EM schemas?

There are three components here to keep separate:

• The datasets themselves, including data and OSCEM-format metadata files. This can be customized per dataset during ingestion. PSI recommends to CC-BY-SA 4.0. EMPIAR and EMDB use CC0.
• The schema. This refers to format itself and whether other initiatives could create derivatives of OSCEM schemas. Ontologies linked by OSC-EM also have independent licenses (at least according to the OBO interpretation that using the PURL counts as attribution). OBO recommends either CC-BY or CC0 for ontologies.
• Software for dealing with the schema (eg the python API generated from the schema, converter tools, or CI/CD code)

References:

• OBO discussion on whether to recommend CC-BY or CC0 for schemas.
• CC BY and data: Not always a good fit.. Compares CC-BY and CC0 for datasets, concluding that CC0 is better since scientific ethics already requires attribution.

My understanding is that the three topics are independent, eg a CC-BY-SA dataset can be produced by a BSD licensed tool according to a CC0 schema. Please correct me if that's incorrect.

My personal recommendation would be CC0 for the schema and BSD3 for the software (like SciCat), with datasets up to the depositors.

[cossorzano]

I don't have any strong opinion about licenses. For the issue completeness, I am adding the links to CC and BSD

https://creativecommons.org/share-your-work/cclicenses/
https://en.wikipedia.org/wiki/BSD_licenses

Without any special preference, I agree that CC0 would probably be better for the schema, datasets depend on the depositors, and software, again I don't have a strong opinion.

[sbliven]

I asked the OBO community about licenses. One issue that was brought up is that using CC-BY ontology terms in a CC0 schema might not be allowed, although there is debate over whether identifiers are copyrightable (it hasn't been tested in court). The community preference seems to be for CC0, but with some members arguing for CC-BY if we use CC-BY ontologies (eg cryoem, although I'm sure the author wouldn't mind in that case 😉 ).