> I won't get tired of repeating the same comment in every topic that suggests Rust [...] we should compare it to other existing solutions, like ATS [1] One of

> I won't get tired of repeating the same comment in every topic that suggests Rust [...] we should compare it to other existing solutions, like ATS [1] One of the problems with repeating comments is that the previous replies are not incorporated into the discussion; for example, mine: >Back when it was on the PL shootout, ATS was consistently in the top five for performance. However, it was also consistently the worst in program size I looked into learning ATS some years ago before Rust existed, and shied away due to the complexity. It's been around much longer than Rust, and I think there's a reason it hasn't caught on: it's just too onerous. ATS might be appropriate for formally verified systems, but language ergonomics does matter, especially for a project with as many contributors as the Linux kernel.