- Author
- Artem Zakirullin
- Published
- June 2026 (last update of a living document)
- Kind
- Living GitHub README essay, with code examples and anecdotes
- Original
- github.com/zakirullin/cognitive-load
- Subject
- Software design
What it is
Artem Zakirullin maintains this README as a living document. It aims to replace taste-based best practices with one measure, the effort of understanding. It gives short rules and examples for reducing extraneous load.
What it says
- The README defines cognitive load as how much a developer must think.
- It says cognitive load beyond roughly four chunks makes code much harder to understand.
- It says extraneous load, from how information is presented, can be greatly reduced.
- The README favours deep modules, with simple interfaces, over many shallow ones.
- It recounts a five-developer startup team with seventeen microservices, ten months late.
- It argues layered architecture adds cognitive load and rarely pays off later.
- It suggests that new people confused for about forty minutes signal code to improve.
What evidence it gives
The README offers small code examples marked with brain icons for load. It gives first-person anecdotes: two pet projects, a consulted startup and a storage migration. It ends with endorsements quoted from well-known developers. It reports no measurements.
Interests and limits
The author says cognitive load is used informally and may not match the scientific concept. The anecdotes cannot be checked from the page. The quoted endorsements are selected by the author. The page links to author sponsorship.