> And you end up writing stuff like main = do { putStrLn "Hello, World!" ; return () } which to me looks like C. There must be a more functional approach.

It looks imperative because that's what Haskell's do-notation was designed to. But it's merely syntactic sugar for the functional equivalent

    main = putStrLn "Hello, World!" >> return ()

which btw is equivalent to simply

    main = putStrLn "Hello, World!"