Admittedly, I hadn’t thought about that.
I see three options:
- do nothing (
idwas designed, from the beginning, with this case in mind)
- do nothing with csl-data.json, but add a note in the spec (what I suggested above)
- add a new variable to hold a human-friendly id (say
citekey), that may-or-may-not be the same as the
id(what you suggested at top)
There are problems with all three, but I think we should do 2 or 3, and still lean towards 2.
I agree we should have feedback, but the only project developers who’ve been around here and active recently is @Dan_Stillman, who I am certain has thought about this issue, given that’s come up in the zotero forums since the beginning.
On implementers: we really need to get them added to a PR review team, and a commitment to actually provide feedback, but I don’t even know at this point who we contact.