I also really like Rust, and I see it as a good replacement of C++, except for one reason: the monoculture around it. You are expected to use one specific comp

I also really like Rust, and I see it as a good replacement of C++, except for one reason: the monoculture around it. You are expected to use one specific compiler. And there's very little (apparent) interest in actually writing a spec, so that other people can actually try to implement their own compilers with a working borrow checker, compilers capable of compiling rustc without sacrificing soundness (so far the compilers I've seen are lacking borrow checker, for example). My only remaining hope rests on the Ferrocene people. So I can see why Rust is being used for user applications, but until that monoculture issue is fixed, I don't really see it replacing C, C++, much less SPARK, as a language for critical systems. It is my sincere hope that having Rust in the Linux kernel, and in GCC, results in Rust getting an actual spec, so that we can finally know what is expected compiler behavior and what is a bug in rustc, other than "what the rustc from this specific repository does". Like, for example, Rust 2018 edition is frozen (AFAIK), so that would be a good target for a baseline spec. Could be a more recent edition, even; it doesn't really matter as long as it's something.