How Unix Spell Ran in 64kB RAM
Summary
This article recounts how Unix spell-checking was engineered to fit 64kB of RAM on a PDP-11, tracing the evolution from disk-based lookups to memory-efficient structures. It covers the use of Bloom filters, hash-based dictionaries, Golomb coding, and binning to achieve near-optimal compression and fast lookups. Useful for readers interested in resource-constrained systems and historical data-structure design.