I fully agree with that, that's why my point (viii) was: "monadic encapsulation of effects: open". But it's important to realise that (a) the existence of effects and (b) their control at the level of types are two differnt things.