the less code the better because the more code the higher the maintenance burden

keeping code around isn’t free. it makes refactoring harder, it makes compilation times longer, it makes the kernel larger, it makes it harder to guarantee device compatibility. that’s all part of maintaining software, but it makes no sense to waste work maintaining removed noone is using, work that could’ve been used to implement new features and/or maintain existing code that’s actually in use

what the kernel is doing is the correct approach. unless they’re sure there’s someone using the thing: old, unmaintained code = bin