On 8/17/12 5:35 AM, TP wrote:
Hi,
I am currently reading documentation on Generalized Algebraic Data Types:
http://en.wikibooks.org/wiki/Haskell/GADT
I have a question concerning this page. Let us consider the following code proposed in the page:
---------------------------------- -- Phantom type variable a (does not appear in any Expr: it is just a -- dummy variable). data Expr a = I Int | B Bool | Add (Expr a) (Expr a) | Mul (Expr a) (Expr a) | Eq (Expr a) (Expr a) deriving (Show) [...] I don't understand. When we write "eval (I n) = n", as I is a constructor which takes an Int as argument, we are able to deduce that the type of n is Int; so the type of eval should be in this case "Expr Int -> Int". What do I miss?
Perhaps it'd help to rewrite the above ADT using GADT syntax (but note that its the exact same data type): data Expr :: * -> * where I :: Int -> Expr a B :: Bool -> Expr a Add :: Expr a -> Expr a -> Expr a Mul :: Expr a -> Expr a -> Expr a Eq :: Expr a -> Expr a -> Expr a So, when looking at the pattern (I n), since I :: Int -> Expr a we know that n :: Int and that (I n) :: Expr a. -- Live well, ~wren