Localization coverage — every human language × every programming language
One cell per (human language, programming language): how much of that language's vocabulary already reads in that human language. The wide pale frontier isn't a gap to hide — it's the supply NADA exists to build.
Each cell = the share of a programming language's tokens already localized into a human language, computed from the compiled brain. Columns span every in-scope programming language; darker = more complete. Scroll to explore; hover for the value.
Every number here is computed from the live NADA pipeline — nothing is hand-typed.