On Sun, Feb 6, 2011 at 8:48 PM, Andrew Coppin
In particular, I get strange looks from people in the OOP community when I say I'm using a language that doesn't have any abstractions at all for dealing polymorphically with containers. In general, it's almost impossible to write a Haskell function that will work with a list, an (immutable) array, a hash table or a map, polymorphically.
I guess it's the case that containers has been there so long now that changing it would break everything. Still, I find myself hungry for something better.
Than again, the Prelude itself leaves several things to be desired...
I'm working on a new map data type at the moment, which is 2x faster than Data.Map for all key types I've tried (i.e. Ints, Strings, and ByteStrings). As part of that work I might try to define a map class using associated data types. Data.Map can be made an instance of that class without breaking any old code. I think one reason we haven't seen a type class for containers is that it isn't easy to create one using vanilla type classes (see Simon PJ's paper on the topic.) Johan