Hi Nicolas, As I was working on adding a function to generate the new cite syntax, an idea occurred to me that I'd like to raise, that would simplify this model and syntax.
I ran the idea by the pandoc list just to confirm my thinking was reasonable, since that has a lot of experience using these plain-text citations. https://groups.google.com/g/pandoc-discuss/c/zbntclaBWDc/m/DCx-mpfcDQAJ Let's review: In org-cite, we have citations, and citation items. Citations may have optional styles. Citation items may have an optional "suppress-author" "variant", indicated with a "-". So: [cite:-doe19;doe20] The idea: This distinction is wrong, and we should treat "suppress-author" as a citation style instead. The result would be: [cite/supress-author:doe19;doe20] Or even can keep the shorthand: [cite/-:doe19;doe20] In fact, with the change, could even remove it entirely If you read the discussion in the pandoc list, there is mild, but I don't think convincing, pushback. Part of that pushback is an artifact of the fact that pandoc citations don't have a notion of citation style, where you specify things like OTOH, pandoc and (haskell) citeproc author John MacFarlane agrees that in retrospect the pandoc decision was probably wrong, in part because of another related issue. Supress-author is a convention that CSL implementations settled on, arguably a hack, to get around lack of explicit support for natbib citet-like citations (though this is being added in the next version). So to get "Doe (2019) argues" you would need to write "Doe [@doe19]" in pandoc, for example. But if you can do this in org-cite, "suppress-author" is no longer needed: "[cite/text:@doe19] argues ..." Both should probably be possible, but the textual (aka narrative) style is what most users are looking for there IMO. I think a positive side-effect is this model would match better LaTeX output models (natbib and biblatex namely). WDYT? Bruce
