Constraints Liberate, Liberties Constrain
summary
Argues that tighter interfaces and earlier symbolic representations preserve compositionality: printing via raw dot‑matrix codes fails, whereas keeping a syntax tree for composition works; embedding SQL as strings loses structure. Scala’s compiler discards ASTs, so shipping tasty trees recovers composition. Shows numeric example (sqrt(8) → doubles) and contrasts applicative vs monad composition: applicatives compose mechanically, monads do not. Advocates least privilege, free/forgetful adjunctions, and delaying compilation to retain analysis power.
AI-written summary · about
chapters (35)
- 0:00Intro
- 2:01Printing
- 2:59Strings
- 4:00David Wheeler
- 5:23Mistake
- 6:10Example
- 8:10Moral
- 8:35Binary Compatibility
- 9:18Binary Representation
- 10:17Tasty Trees
- 11:15Premature Optimization
- 11:42Compilation or Folding
- 11:52Unconstrained and Too Powerful
- 15:01Type Signature
- 17:13Constraints
- 18:09Applicative Functor
- 19:04Monas
- 20:06Composite
- 21:17Monads
- 23:00Actors vs Futures
- 23:12Algebraic Reasoning
- 25:06Scala Concurrency
- 26:20Actors
- 26:47Lord Acton
- 27:11Side effects
- 29:09Least privilege
- 34:24Leaky abstraction
- 35:22Logical semantics
- 35:34Syntax and semantics
- 35:54Game
- 42:06Galois Connection
- 44:03Scala Junction
- 47:06List Junction
- 47:59Summary
- 48:57Questions
related talks
37mWhen is an ADT not an ADT?
50mPrinciples of Elegance
34mHow I learned to stop worrying and love Functional Design
01h 01mThe power of composition
33mExtensible Effects: A Leaner Cake for Purely Functional Code
43mThe Tyranny of Structurelessness
01h 04mKeynote - What to Leave Implicit
46mThoughtful Software Engineering