Re: [Haskell-cafe] Class Quantification

1 Oct 2008

      Bas van Dijk wrote:
...
On Wed, Oct 1, 2008 at 3:01 AM, Reiner Pope  wrote:
...
I believe there is no way to simply express this "abstraction over classes",
but the Scrap your boilerplate with class[1] paper discusses this same
problem and present a workaround by defining the class's dictionary of
methods as an explicit type.
What follows is the code to implement their workaround for your example.
First some fairly standard extensions:
...
{-# LANGUAGE Rank2Types, EmptyDataDecls, FlexibleInstances, KindSignatures
#-}
And some more controversial, but necessary ones:
...
{-# LANGUAGE UndecidableInstances, OverlappingInstances #-}
Just an observation:

That terrifying list of extensions is needed for this 'full' solution.

To just literally encode what the OP wanted you just need a concrete 
dictionary and existentials, AFAIK.

Jules

Re: [Haskell-cafe] Class Quantification

Jules Bean