Compression Dictionaries
summary
Claims that HTML and resources shrink upwards of 90% for some sites and that week‑over‑week JS updates can be 96% smaller are backed by concrete numbers (Google Search 40–60% smaller, Vercel 96%, Filament ~45%). Uses existing compressors (zstd and brotli/br‑d), a dictionary generator and build‑time or edge wasm tools, a new Link rel=dictionary preload, and CDN tweaks; Chrome runs an origin trial (Chrome 119) with same‑origin + COEP requirements and a TTL moving toward seven days.
AI-written summary · about
chapters (11)
- 0:00Compression dictionaries intro
- 1:13Why dictionaries matter
- 4:48How dictionary compression works
- 10:34Dictionary compression history
- 13:21Static resources with dictionaries
- 22:10Dynamic HTML compression
- 26:49CDN support config
- 31:44Security privacy limits
- 35:41Tooling dictionary generation
- 41:05Status and future plans
- 43:56Q&A
related talks
36mBreaking The Bloat: Achieving The Fastest Navigations on The Web
47mWeb Compression: How to (Re)Take Control?
46mNot Only Gzip
42mHoney, I Shrunk the Bits. Or how modern HTTP compression works.
58mResponsible JavaScript
03mWhen small is better: Brotli compression in a real case
32mResponsible Javascript
51mWhen JavaScript Bytes
