On Feb 7, 2006, at 11:45 PM, John Meacham wrote:
Ashley Yakeley has convinced me this proposal won't work as is. I knew that dropping of the a type parameter in the dictionary passing scheme would bite me :). though, perhaps it will inspire another proposal? In the meantime, I will see about implementing the 'wft' constraints as mentioned in the paper. though, the example ashley gave gives me pause about how useful these will be in general. hmm...
It occurs to me to muse that the "wft" constraints look a whole lot like "Typeable" constraints (ie (Typeable (m (a -> a))) => m Int ). Should there be a class which is implemented by every well-formed type of kind (*)? Should that class have one or more of the following: * Structural decomposition and reconstruction a la Generics? * Scrapped-boilerplate traversal a la Typeable? * Ability to coerce to Dynamic? It would be a special source of amusement to me if such a class were called All or Forall... :-) (though a moment's thought will tell you it's Wrong.) -Jan-Willem Maessen
John
-- John Meacham - ⑆repetae.net⑆john⑈ _______________________________________________ Haskell-prime mailing list Haskell-prime@haskell.org http://haskell.org/mailman/listinfo/haskell-prime