Simon Peyton-Jones <[EMAIL PROTECTED]> writes
>
>> - The simple-context restriction.
>> ...
>> My default position is not to change. Question: who, apart from
>> Ralf, has actually tripped over the lack of contexts of the
>> form (C (a t1 .. tn)) in Haskell 1.4? Is their lack a real
>> problem in practice?
Are you talking about contexts in general, or only contexts in function signatures?
For me, the most powerful argument in favor of generalizing contexts is the
possibility of defining monad transformers, as described in "Monad Transformers and
Modular Interpreters" by Liang, Hudak and Jones. (Because I love this paper so much.
:) I think I convinced you of this once before, when MPC support was being added to
GHC. One has applications in class declarations, and non-variable arguments in
instance declarations:
class (Monad m, Monad (t m)) => MonadT t m where
lift :: m a -> t m a
instance (Monad m, Monad (StateT s m)) => MonadT (StateT s) m where
lift m = \s -> m >>= \x -> return (s,x)
If the definitions from the paper can be turned into valid Haskell 98 w.l.o.g. now,
then I'm happy.
--FC
