when i use the superb org-capture extension in firefox, there is only one icon, so any decisions about formatting have to be one-size-fits-all at capture time in elisp. here are my ideas, but not implementation. the decisions i'd make are:
1. fill long unfilled paragraphs vs. preserve verbatim lines if i capture code, i don't want it filled. if i capture prose that is unfilled, as in a long line, i probably want it filled at capture. heuristic: loop through every line. if the line qualifies by some sub-heuristic, and is more than 2x (window-width), then fill at capture time, but only that line, not contiguous ones that do not meet the sub-heuristic (different prefix etc.). what does one do to loop through paragraphs? is this a reasonable heuristic? anybody tried it? this is all too much for me atm but i thought i'd pu out the idea in case anybody has thought of these things. alternative: figure out and use an emacs minor mode that makes unfilled paragraphs look like filled ones, including fill-column, filling word boundary heuristics, fill prefix, etc. this would still interfere with diff and such. it might be best reserved for this purpose rather than ridding me of my m-q filling habit. that would be drastic. 2. region on org heading vs. in body heuristic: if selected text is zero length, i.e. nonexistent, then link on org heading. if it is short, "text -- [link]". if long, then i'm not sure which i like better: short part of first line on header and then rest and then "-- [link]", or link on header. -- The Kafka Pandemic A blog about science, health, human rights, and misopathy: https://thekafkapandemic.blogspot.com
