Stealing from Biologists to Compile Haskell Faster
Summary
The piece explains how Haskell's GHC do-notation optimizations switch from sequential binds to parallelized applicative combines, using dynamic programming ideas to minimize round trips. It compares compiler scheduling to RNA folding, describes the greedy and optimal strategies, and shows how extreme-cut and longest-chain bounds keep compile times practical. It argues that while sub-cubic algorithms exist in theory, real-world constraints favor the simpler, local fixes that improve GHC performance today.