A collection of caching library implementations built from scratch to develop genuine engineering fluency in core caching strategies, rather than relying on secondhand explanations.
Caching strategies such as LRU, LFU, write through, and cache aside are common in backend and systems design work, but there is a meaningful difference between having read about a strategy and having implemented it correctly, from scratch, more than once. Each language-specific repository in this organization closes that gap directly, one language at a time. A related but distinct effort, mnemosyne-hw, applies the same hands-on philosophy to a different problem: direct control over CPU cache-line behavior rather than eviction policy for application-level data.
| Repository | Language | Status |
|---|---|---|
| mnemosyne-python | Python | MVP complete: LRU and TTL caches, both fully tested |
| mnemosyne-go | Go | Scaffolded, implementation not started |
| mnemosyne-rust | Rust | Scaffolded, implementation not started |
| mnemosyne-hw | Rust, C, Zig, Assembly | Scaffolded, implementation not started |
Each caching implementation is self contained, idiomatic to its
language, and built manually. Shortcuts such as Python's OrderedDict
or Go's container/list are intentionally avoided on the first pass
so the underlying mechanics are properly understood before they are
abstracted away.
MIT