gaurav
commented
2 years ago We have two options:
- Leave in the current fields, but have them stored in the Phyx file in the
sourcefield. - Replace the curator information with a
sourcefield.
However, we do want to have curator information at the per-phyloreference level (similar to how we have per-term attribution on an ontology). This could be name + identifier or only identifier, so e-mail is probably unnecessary (double-check this with Jim).
So, next steps:
- [ ] Propose changes to Phyx to include basic curator information at the phyloreference level (this would bump it to v1.1.0). Review there and make the change there first.
- [ ] Update Klados to include per-phyloreference curation level.
Goal: stay consistent with user stories in user stories.
At the moment, we record provenance per-file, including curator name, e-mail address and ORCID. I proposed in https://github.com/phyloref/klados/issues/59#issuecomment-994118919 that we should replace this with a citation so we can record multiple curators. We had a larger discussion about provenance today, where we decided that at a minimum we should also record the creation time and last-modified time.
However, thinking about this more broadly, we should look at what other organizations like the GO consortium (or their tools like Noctua, etc.) do and follow their lead -- for example, should we record provenance for each individual phyloreference, or is doing it for the whole file okay? Should we bring back the system of giving phyloreferences states (e.g. "draft", "final draft", "tested", "published") and recording when the state changed and by whom? Should we include a logging system so that we can track all the changes being made by anybody, should those changes be stored in the Phyx file itself, or should we let an external version control system take precedence for dealing with that?
While we have a primary goal of ensuring that Klados is a good curation tool, there is an important secondary goal here -- generating and testing provenance information that we will later incorporate into the Clade Ontology.
I'll do some research into what other organizations and ontologies do, and report back.