zettel should be atomic
Table of Contents
content of a zettel should only address atomic idea, because:
- atomic notes make reference easy, make discussion with zettel easy, as referring to one propositions is then intuitive
- atomicity - flexibility, making it convenient for rearranging organization and compilation, to reuse notes with zero effort
- atomicity make bad note better and hence ensure works would not be wasted
In practice, when made atomic, a zettel often falls in one of 3 common zettel categories
- use reference only if necessary 1
use only the plainest, most literal terms 2
“I couldn’t reduce it to the freshman level. That means we really don’t understand it.” - Richard Feynman
Backlinks
zettel
basic component of zettelkasten. A zettel is formulated in the following 3 parts:
- unique ID
- used to reference a specific zettel in other zettel
- content
- an idea, a thought.
- reference
- context, or source.
A functional zettel cannot just be your whole notebook with a unique ID like “notes-in-2024”. For a zettel to be functional in a zettelkasten, it has to follow some principles:
navigation zettel
zettel that have lots of links. Could be specially made as entrance to ideas around a topic, like “Artificial Intelligence” to hold link to “Learning is not necessary for Artificial Intelligence”, when I have a good idea on the latter, and the former only as portal to good ideas like the latter.
This is an exception for zettel should be atomic. Clearly, “Artificial intelligence” can be split in many ways; yet this navigation is still useful
navigation zettels often falls into the following categories:
- plain navigation for browsing - index
- incidentally important concept - zettel
- in the process of decomposing a complex problem - “solving XXX with A and B” “convert XXX into YYY” “solve YYY with A to B-compatible form” until the most simple propositions.
- compilation of ideas to teach/present (draft/post of tutorial or artical) - “developmental learning tutorial 1: implementing scholarly thinking” “how one actually find out about useful knowledge from raw experience”
Footnotes:
if something is only “nice to be mentioned”, make another zettel about it. e.g. don’t mention gausian mixed model in the zettel about distribution. Wikipedia has this problem: to understand one term, one has to go through many unnecessarily referenced terms.
don’t resort to technical terms if you can state it in plain words. The word is beaten by a line of formula \(G = R_{t+1} + \gamma R_{t+2} + \gamma^2 R_{t+3} \text{ with } \gamma \in [0,1)\), which contains only primary-school math