The corpus by language
A language is read off the code, so nothing here rests on anybody declaring anything.
99% of measured codebases (7,073 of 7,124) are on this page. The rest are codebases carrying no primary language this reading could read, which is a limit of the indexing behind it and never a fact about the codebases, or languages holding too few codebases to read as a group.
- Rust — 566 codebases, median 58.9
- Swift — 529 codebases, median 55.3
- C# — 519 codebases, median 46.5
- Ruby — 489 codebases, median 59.0
- JavaScript — 478 codebases, median 46.0
- Scala — 478 codebases, median 56.7
- TypeScript — 449 codebases, median 47.7
- Python — 436 codebases, median 52.0
- PHP — 416 codebases, median 54.6
- Kotlin — 412 codebases, median 54.6
- Java — 381 codebases, median 51.4
- Dart — 349 codebases, median 54.6
- Go — 338 codebases, median 56.7
- Erlang — 326 codebases, median 59.5
- Elixir — 315 codebases, median 61.4
- F# — 314 codebases, median 56.3
- VB.NET — 278 codebases, median 41.8