Jiwon's Alcove

The Rules of Lazy Note-Taking

Aug 21, 2023

A phenomenon I see in the personal knowledge management (PKM) space is people obsessing over how they manage their notes.

Truth is, notes are only useful when they help us get stuff done.

For none of these use cases is there a requirement to be meticulous and consistent in the ways we organize our notes.

Much like a lazy evaluator which defers expensive computation until it is needed, note-taking should be done lazily until eager effort is needed.

  1. Take notes lazily when a need arises.
  2. Add structure lazily when a need arises.

Rule 1 prevents us from taking notes unnecessarily. Rule 2 prevents us from tweaking the setup unnecessarily.

Let us have a look at some of my notes to illustrate my point.

I don't take notes on topics just because I just learned it.

I take notes on topics if I had a hard time figuring out some details (the reason why Cauchy criterion is more useful than basic convergence definition), or if I can add my "original" thought to it (equivalent definitions that I worked out as an exercise).

Some notes like this are atomic. Others are giant documents. It's okay to be inconsistent and lazy about the atomicity of notes.

If and only if I take repeated notes of a similar structure, will I formalize the structure.

Literature notes is a useful structure since I read and cite a lot of papers.

Notice the use of grammatically incorrect sentences and symbols as a visual shorthand for words (e.g. $\exists$) to expedite writing & reading.

Some notes exist on their own without being linked to other notes.

This note is about an experiment I ran to understand the behavior of clustering algorithms.

You don't have to link every note with every other note.

You don't have to recreate wikipedia.

I didn't create wikilinks to concepts like k-means clustering since I don't have a use for notes like that.

If I wanted to review some facts about k-means clustering, I could just look up wikipedia.

This experiment is mine, and the interpretation of this experiment is mine. I can't google it. As such, I lazily create a note for it.

I make use of bullet points with succint sentences. No one else has to understand this note. There's no point in polishing notes.

There are typos, and honestly, who cares?

If a note doesn't have to be long, then I keep it short.

This note records a question I had when studying mathematics.

Since the title is enough context to remember my original question, I simply tag the note as #question and call it a day.

You don't have to recreate the documentation.

You can just link to the documentation page, or the Stack Overflow page, as needed.

Only when Googling again will be a headache or take way more time than the time it takes to note it down, do I take notes on documentation.