[Haskell-cafe] Why MTL and Transformers?

Something that I have been wondering for a while is: why are there *still* two monad transformer libraries with seemingly identical functionality, especially given that they have conflicting namespaces? It creates stupid problems like the one that Erik encountered and had to work around. I recognize that this situation most likely came about via historical accident rather than by intention, but why does it continue to be true? Cheers, Greg