01
Asyncio cancellation is part of your API
Design coroutine cleanup, deadlines, and cancellation propagation as observable interface behavior.
Independent field notes / Vol. 01
Python,
below the surface.
Precise explanations of the machinery, patterns, and tradeoffs that make Python feel deceptively simple.
Read the cover storyThe index
Choose a rabbit hole.
6 notes, edited for signal.
02
Processes make serialization part of the design
Design process workers around explicit values, importable callables, transfer costs, and start-method realities.
03
Task groups and failure propagation
Use structured concurrency to make child lifetime, sibling cancellation, and multiple failures explicit.
04
Testing concurrency and cancellation without timing roulette
Replace sleeps and scheduler luck with barriers, events, bounded waits, and observable lifecycle assertions.
05
Threads, tasks, and processes solve different waits
Choose a concurrency mechanism by locating blocked resources, ownership boundaries, and coordination costs.
06
The GIL, without the folklore
What CPython locks, what it does not guarantee, and how free-threaded Python changes the decision.
Editorial policy
No listicles. No folklore. No just trust me abstractions.
Every note should leave you with a better model of the system, not just another snippet to paste.