Breaking every single Haskell module for some namespace reorganization doesn't seem worth it. I don't think alternative Prelused (one of the justifications) is a good idea to begin with, as programmers will have to first understand which particular version of e.g. map this module uses, instead of knowing it's the same one as every other module uses.
Changes like this will likely cause years worth of pain e.g. see the Python 2/Python 3 failure.
The likely practical result of this is that every module will now read: