On Mon, Dec 11, 2006 at 12:16:06PM +0000, Simon Peyton-Jones wrote:
[...] Just to summarise, the difficulty is this: I have a dictionary of type (C a b1) I need a dictionary of type (C a b2) There is no functional dependency between C's parameters
PS: the complete program is this: class C a b where op :: a -> a
f :: C a b => a -> a f x = op x
That raises a point I'd wondered about. GHC requires only that each type variable in the context be reachable from the type via a chain of assertions: http://www.haskell.org/ghc/docs/latest/html/users_guide/type-extensions.html... What's the rationale for that, rather than calling types like the above type of f ambiguous? The examples given in the User's Guide involve functional dependencies, albeit obscured by superclasses