To me the benefit of Rust is less safety, than better abstractions and modularity. Linux has been growing in complexity without much of a breaking point, and t

To me the benefit of Rust is less safety, than better abstractions and modularity. Linux has been growing in complexity without much of a breaking point, and the discipline and separation of concerns high level languages could help with is a huge long-term benefit. (For example, say we could one day make a bunch of basic Linux drivers somewhat kernel agnostic? That would be amazing.) Now, ATS is I hear a fine language, but I think the ability to write good, and especially abstract, ATS is sadly too far outside of most kernel dev's skill set. Even among PL researchers, there is a tendency to write monolithic Coq---to wit, is there a Hackage or crates.io for Coq?---because abstraction are hard and the academic paper economy doesn't really reward it. Rust is just easy enough, and with enough crates.io momentum, that I hope the relative cost of reusable vs unreusable code will be less. And finally, it takes a lot of political will to introduce a new language in any existing project, let alone one as big and storied as Linux. Like it or not, but general popularity of the language absolutely does help with that.