[willtim]

No IO in Haskell is purely functional. Your Haskell program wires up a computation that is eventually run by the runtime. The wiring up, as described in Haskell, supports referential transparency and composition as only pure code can. Your confusion may arise from people using the term "pure" to differentiate from "monadic".

[frik]

All I meant is IO (aka input/output) can have side effects outside of the scope of the language environment. Haskell has the IO type to encoding those this side effects: https://wiki.haskell.org/Introduction_to_IO

[codygman]

I haven't decided yet, check out this link to decide for yourself:

https://leanpub.com/purescript/read#leanpub-auto-the-eff-mon...