Always Be Composing
summary
Shows a practical spectrum: starting with a cinsky macro that expands immediately, moving to higher-order function wrappers, then to a pure-data representation that generates shapes, and argues tradeoffs behind the mantra data > functions > macros. Implements a regex-style automata library called automat and discusses transducers, a reservoir sampler versus sum/count tuples for means, and a concurrency primitive let-flow that derives dependency order to maximize parallelism while preserving composability.
AI-written summary · about
chapters (16)
- 0:00Intro
- 2:21sierpinski sierpinski draw-triangle
- 7:27THE TROUBLE WITH
- 7:50THE VALUE OF
- 10:38TRANSDUCERS
- 13:39YOUR DATA ISN'T UNIFORM
- 16:29REGULAR EXPRESSIONS
- 18:291 (* 2) + 3
- 20:40CONVEYING DATA
- 22:16BACKPRESSURE IS
- 23:59CAUSALITY
- 26:05STREAMS VS DEFERREDS
- 26:52STREAMS YIELD DEFERREDS
- 27:20ASYNC COMPOSITION
- 28:40DATAFLOW
- 30:42CODE IS DATA
