A type error isn't some kind of punishment, it's a discussion with the compiler. Type errors are good -- they mean that there's a discrepancy with your idea of

A type error isn't some kind of punishment, it's a discussion with the compiler. Type errors are good -- they mean that there's a discrepancy with your idea of the program that you're writing, and the way that you've actually written it, and a good typechecker will help you track down that discrepancy. So when GHC says "you're doing side effects outside of IO", it's not trying to harsh your vibe. It's explaining to you the inconsistency between your program and the "type" (description) that you've written for it. You might not want to have to explain yourself to the compiler, and that's fine -- go write in an untyped language instead. But it's hardly authoritarian to offer a different way of programming. (Keep in mind that typechecking meets many of the same needs that debugging does. For someone who already tries to typecheck in their head to ensure that their programs are correct, having an explicit, external typechecker can be a relief. That might not be you, but it's at least some people; when I program in Lisp I'm always running into problems that a typechecker would have solved at compile time, and I find that frustrating. To each their own.)