(serious problem with IO explanation)
|(4 intermediate revisions by 3 users not shown)|
Latest revision as of 13:53, 5 April 2011
Hi, I just stumbled over this IO tutorial, and gave it a quick glance for the usual issues. Unfortunately, it seems to propagate a seriously misleading picture of the IO type: IO cannot, must not be a type synonym. As long as Haskell does not have a uniqueness type system, any functional representation of IO must be hidden behind an abstract type! making only RealWorld abstract is insufficient, as functional programming would still allow us to create RealWorlds out of thin air, or let them disappear: oops :: RealWorld -> (Char,RealWorld) oops rw = let (c1,rw1) = getChar rw (c2,rw2) = getChar rw in ([c2,c1],rw1) whoCares rw = (rw,rw) -- let's make a copy while we're at it fortunately, the fix is simple (but crucial): - make IO a newtype - do handle code inside the abstract type with great care, as many functional equivalences are no longer valid! actually, only the first part is simple, the second part depends on some apparently and happily valid properties of typical optimizations in the implementations. keeping IO abstract also makes it clear that other implementations are possible, and have been used (such as having an interpretation loop for the tree of i/o operations computed by main). also, it might be more useful to think of unsafePerformIO linking into the implicit RealWorld at its point of evaluation, rather than creating an all new one to throw away. this linking in is what makes it so useful as an extension hook for the Haskell implementation.
inlinePerformIO: Don, i've reverted its "definition" to pseudo-form because if we consider this as step-by-step tutorial, at this point reader is yet not know about GHC specific of IO type representation. my definition just duplicates pseudo-definitoon of unsafePerformIO in previous section adding only INLINE instruction. otherwise, reader may be lost in these "differences" between two definitions and GHC-specific details Bulatz 15:47, 18 March 2007 (UTC)
 1 when definition wrong
The definition for when
when :: Bool -> IO () -> IO () when condition action world =
does not compile. The compiler says The equation(s) for `when' have three arguments, but its type `Bool -> IO () -> IO ()' has only two. I cannot fix it, I don't understand it (yet). Franky 15:53, 17 January 2011 (UTC)
 2 The first examples of 6.0 don't work
The first examples of section 6.0 actually prove the opposite of what you are trying to say. The expressions seem to get executed out of the desired order. I've compiled this program on both Windows and Mac.
main = do let get2chars = getChar >> getChar putStr "Press two keys" get2chars return ()
It doesn't print anything until after you hit enter. And then it exits. I tried to compile this version on a Mac and Windows and it doesn't even compile:
main world0 = let get2chars = getChar >> getChar ((), world1) = putStr "Press two keys" world0 (answer, world2) = get2chars world1 in ((), world2)
The output of the compiler is
C:\Users\cwelchmi>ghc getChar2.hs [1 of 1] Compiling Main ( getChar2.hs, getChar2.o )
getChar2.hs:1:1: Couldn't match expected type `IO t0' with actual type `t1 -> ((), t2)' In the expression: main When checking the type of the function `main'
getChar2.hs:2:34: The function `putStr' is applied to two arguments, but its type `String -> IO ()' has only one In the expression: putStr "Press two keys" world0 In a pattern binding: ((), world1) = putStr "Press two keys" world0 In the expression: let get2chars = getChar >> getChar ((), world1) = putStr "Press two keys" world0 (answer, world2) = get2chars world1 in ((), world2)
getChar2.hs:3:38: The function `get2chars' is applied to one argument, but its type `IO Char' has none In the expression: get2chars world1 In a pattern binding: (answer, world2) = get2chars world1 In the expression: let get2chars = getChar >> getChar ((), world1) = putStr "Press two keys" world0 (answer, world2) = get2chars world1 in ((), world2)