Hi Simon,
>>>>> "Simon" == Simon Peyton-Jones <[EMAIL PROTECTED]> writes:
> | Subject: Type signatures in instance declarations?
> ......
> | I think it would be even useful to the compiler when one has a
> | polymorphic recursion. ISTR occasions when I wasn't able to define
> | fmap directly (eg. for a nested datatype for lambda calculus terms),
> | but had to define another name for it with a polymorphic declaration,
> | then equate fmap to it in the instance declaration.
> Interesting: so you wanted to write a type signature that was *more general*
> than the one for the instance decl? Right?
Not exactly. Actually, I was wrong. Here's how you can write some
code that needs polymorphic recursion *without* an explicit type
declaration!
data L x = Var x | App (L x) (L x) | Abs (L (Maybe x))
instance Functor L where
fmap f z = case z of Var x -> Var (f x)
App e1 e2 -> App (fmap f e1) (fmap f e2)
Abs e -> Abs (fmap (fmap f) e)
instance Monad L where
return = Var
e >>= f =
case e of Var x -> f x
App e1 e2 -> App (e1 >>= f) (e2 >>= f)
Abs bd -> Abs (bd >>= f')
where f' m = case m of
Just x -> fmap Just (f x)
Nothing -> Var Nothing
Well, you live and learn...
Cheers,
Peter
