Nicolas Goaziou <n.goaz...@gmail.com> writes: > I favor [cite:PROPERTIES] over [[cite:PROPERTIES]], because the latter > (link syntax) implies a (optional) description part. I don't think > a description is ever meaningful in citations.
I have been holding on to this for a while. Just typing it out as it comes to my mind. I favor the above syntax. I view Citations as closer to Footnotes. The syntax should parallels footnotes syntax. 1. PROPERTIES should be opaque to Org. It is a key or a list of keys possibly bibtex but Org doesn't take stand on how it looks like. 2. There will be a org-BACKEND-citation-reference. 3. There will be a org-BACKEND-bibliography. 2, 3 more likely with interface with respective citation processor (citation processor as opposed to a database) via CLI. Citation processor could be whatever org-exp-bibtex interfaces with right now. I also have some proof-of-concept - see zotcite - for zotero. 2, 3 will parallel footnote-reference and footnote-section callbacks in HTML backend. 4. Footnotes can be introduced with either fn: prefix or cite: prefix. There should be a way to put fn: and cite: in same enumeration context. There should be a way to put fn: and cite: in different enumeration context. The former case could be a degenerate mode where Org can transcode what is seen in the buffer where everything is footnotes. The latter case will result in Citations and Bibliography being generated by the above backend transcoders. 5. Citation definitions in Org buffer will be *ignored*. (It could be considered when the exporter works in a degenerate footnote only mode where plain text transcoding is resorted to because there is no suitable application available for the backend format.) Plain text citation definitions are only to help the author have a glimpse of what he is doing, it has only UI-value but no contents value. 6. There may be an advisory citation style - say APA, Chicago etc - which the backends may honor or ignore. I am not clear about: 1. How multiple keys are to be handled. 2. What prenotes or postnotes mean. 3. Chicago note style etc. I think the community should answer and articulate 1, 2, 3 clearly. With my proposal, there could be some minor changes in Footnotes normalization and some minor changes in existing transcoders. The Org syntax for citations should *at this point* in time should NOT make any assumptions about the Citation Database or the Citation Application. As far Org is concerned, there should be a way for Emacs to interact with these engines and have them return Citation Refernece and Citation Defintion contents in required backend format. --
