For your sanity, keep metadata embedded in each file. I keep k:v pairs. One per line. e.g "topic: D3", "subject: scales", "context: Side Project"

Use grep, awk (whatever). Slice, group, dice, join, merge, backup as and how you want. A bash-like shell is the only dependency. Check into git regularly.

Some scripts I use

  ## group by topic. search for lines that start with "topic". print topic and the file name
  grep -i "^topic" *.txt | awk -F ":" '{printf "%-25s%s\n",$3,$1}' | sort

  ## find all files containing the "topic" tag 
  grep -i "^topic" *.txt | awk -F ":" '{printf "%-25s%s\n",$3,$1}' | sort 

  ## find all files NOT containing "topic". useful for cleaning up 
  grep -iL "^topic" *.txt

  ## find first 10 files not containing "topic" and open each in vi sequentially
  for f in $(grep -ilL "^topic" *.txt | head); do vi $f; done

I find the side-by-side panes of some just a hilariously stupid waste of space. And the editors that at least do live “previews” feel so janky when editing; watching pound signs appear and disappear as you move off a line, or having text shift around as bold and italics markers come and go.

No, I don’t need all the formatting capabilities of a word processor. But there’s a reason why MD is often pretty-fied: it’s damn ugly and noisy as hell.

I get that people want a “pure” text format for longevity’s sake; but just demand a text export-in MD if you like. But working in MD is crazy making; it’s ugly, distracting, and worst of all, the “raw” form is nothing like documents and formatting we’re all used to, so even if you get used to it, all of the other documents in the world you interact with will still be using real bold, italics, and headers of increasing size.