> > Basically, quantified types can't be given as arguments to type >> constructors (other than ->, which is its own thing). I'm not entirely >> sure >> why, but it apparently makes the type system very complicated from a >> theoretical standpoint. By wrapping the quantified type in a newtype, the >> argument to IO becomes simple enough not to cause problems. > > Thank you, I have read about predicative types and it seems I understand the origin of the problem now.
> GHC has an extension -XImpredicativeTypes that lifts this restriction, > but in my experience, it doesn't work very well. > Yes, it didn't help in my case. Thank you, Dmitry
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://www.haskell.org/mailman/listinfo/haskell-cafe
