Persistent incrementality in Rust
Summary
The article explores persistent incremental computations in Rust, compares a naive in-memory approach with Salsa-based incremental recomputation, and introduces Ajika for filesystem-backed persistence. It explains how incremental computations allow reusing previous results to avoid unnecessary recomputation, with illustrative code and practical considerations. It also covers architecture, storage, interning, and the implications of persistence for memory usage and cache validity.