01
Allocation is work: object churn and memory locality
Measure temporary objects, retained memory, and data layout before trading clarity for reuse.
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.
29 notes, edited for signal.
02
Attribute lookup, from dot to descriptor
Follow one dotted expression through instance state, class namespaces, descriptors, method binding, and fallback hooks.
03
Circular imports are partially initialized modules
Diagnose circular imports as order-sensitive access to incomplete module namespaces, then repair the dependency graph instead of shuffling lines.
04
Closures capture cells, not frozen values
Understand late binding, shared nonlocal state, loop callbacks, and the CPython cells underneath lexical closures.
05
What compile() actually produces
Follow Python source through ASTs and code objects, then watch execution supply the frame and namespaces that compiled code does not contain.
06
super() is a cooperative MRO machine
Replace the parent-call myth with a precise model of MRO-relative lookup, cooperative methods, and safe multiple inheritance.
07
Inside CPython's allocator
Connect Python allocations to pymalloc arenas, pools, blocks, process RSS, tracemalloc domains, and responsible measurement.
08
Crossing the Python boundary
Move useful work across Python, native, I/O, and process boundaries without optimizing the wrong crossing.
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.