● Atomic Notes

279 2026-07-27 note ♋ ♊

One idea per note, self-contained enough to survive being read outside the context it was written in. Luhmann's slip-box already worked this way, but the zettelkasten methodNiklas Luhmann kept a slip-box of roughly 90,000 index cards over his working life, each one a single idea, each one cross-referenced to others by hand-written note numbers. He credited it, half-jokingly, with most of his output as a sociologist.… frames atomicity as a discipline more than a rule, a compass rather than a litmus test.

Andy Matuschak writes that a note should be about one thing, but capture the entirety of that thing.

The practical test isn't length, it's whether a note holds up once it's pulled out of the piece it came from. StubsA problem that wiki-linked digital garden run into is figuring out what happens when you reference a note that doesn't exist on disk yet. A traditional website would 404. This site turns it into a stub instead. A stub is… grow into it over time.

See Evergreen Note TitlesAndy Matuschak argues that evergreen note titles work like an API: a well-factored, well-titled note lets you reference its entire argument through the title alone, the same way a function signature lets you use code without re-reading its body. Separation… for what a well-factored atomic note's title looks like, and Maps of ContentA map of content is a note that's mostly links, a curated index into a cluster of atomic notes on one topic, sitting alongside folders and tags rather than replacing them. Automatic groupings can't make judgment calls the way a… for what to do once enough of them accumulate to need an index.