Some working models of note-taking, as informed by practice

2022-05-05

Note that I'm interested in ways that people have written things into [personal] note systems (as opposed to writing for publication), and I want to see practices that are actually useful (in pursuit of an external goal). Pushing paper (or pixels) around is otherwise busywork, or unproductive fluff – a means without an end.

  1. Undirected journalling (append-only)
    • Value is extracted during the process of writing.
    • Results in a linear history of snapshots in time.
    • e.g. most journalling practices.
  2. Writing references for later ("classic" note taking)
    • Value is extracted when you return to something for detailed analysis, or to provide a record of occurrence.
    • Results in a collection of monolithic, independent documents.
    • e.g. meeting minutes, retrospectives, decision records.
  3. Building a case study network for understanding a complex domain (interlinked around a small core of "crossroad" cases)
  4. Recording atomic insights in open-ended research (growing a knowledge base by expanding with references to existing notes)
    • Value is extracted by finding connections to existing ideas, and updating notes with new research.
    • Results in a network of refined ideas clustered around areas of promising research.
    • e.g. Andy Matuschak's "better thinking"

Discussion

Creating the links between documents must contribute value towards a purpose. These are usually either to make explicit something that happens in the domain (in "case study networks" or "open-ended research") or to make referencing easier (in "references" and "case study networks"). Other links are superfluous, or contribute little value, and can force notes to be broken up unnecessarily.

Comparisons

Undirected journalling

This seems like a weak practice, but often, writing forces the processing of vague thoughts into something that can, at least, be expressed. This can work if expression is the goal.

Reference notes

"References" can often be absorbed into SRS if necessary, so for this practice of note-taking, the goal should be to maximise the amount of useful information that will be needed in the future.

If the note collection gets too large, then without some organisation, re-discoverability can become an issue. Links, tags, and search are ways around this.

Case study networks

This practice is a specific technique in applying the Cognitive Flexibility Theory. Each individual note is a specific case study, and links are drawn between the case studies.

Note that this can be considered a non-linear alternative to the literature review.

Atomic evergreen research notes

In this model, each individual note is a specific idea, and links are drawn between ideas. This provides a pattern for growing notes, and allows research to be written in an incremental fashion. In addition, Andy Matuschak makes a particular point about ensuring that each note is "atomic" (minimal).

When viewed this way, this is "just" a research note-taking practice with some more structure.

This practice falls into a broader category of what people associate with modern note taking (e.g. Zettelkasten, BASB, PARA), but is here referring to Andy Matuschak's implementation of the practice. I suspect that a common failure of this note-taking method (when following the hype) is that more effort is spent in the linking and organisation of the notes, rather than the application of the knowledge. If the links are not exposing some meaningful connection between the ideas, then at best they offer a neat way to traverse between ideas that would be "lost", and at worst they're just a distraction.

Interfaces to notes

Knowledge isn't necessarily a granular, discrete thing (indeed, it could be said that complex domain knowledge is the links), but most note-taking tools for interacting with notes are focussed on the vertices of the graph (the note documents) rather than their edges (the links), meaning that it's hard to extract value from connecting graphs. In addition, the communication process usually seems maps better to individual linear documents (blog post, video, article, paper) rather than the graph itself.

Note that mind mapping is the inverse of this – it focusses on the edges rather than the vertices. It makes the connection between things clear, but not necessarily the things themselves.

Ways by which links can provide value:

One large note, or many small notes?

From a practical standpoint, most note editors don't really support multiple notes inlined in a flexible manner. In addition, the fastest search interfaces are often titles (which imply separate notes). Andy's "atomic notes" is one good way to look at it, but a heuristic for using notes (from practice) would be: combine together what you want to read together.

References