Or maybe we have come to the point where Haskell's lack of a 'real' module system, like e.g. in SML, actually starts to hurt? Can associated types come to the rescue?
The organic growth of Haskell resulted in many compromises to the core design. As a result Haskell inherits inadequate module system. Also all the bells and whistles, like constraint handling rules, overlapping instances, associated types, show that the pre-GADT type-class solutions do not allow for too much (e.g. DoCon failure). I believe that true dependent sums and products at the core level should allow to capture modules at the core and be first lass. Likewise functors should be the fundamental unit of composition at all levels. But that would mean no back compatibility. Cheers, -Andrzej