Re: Proposal: more general unionWith for Data.Map

Oh yeah, I didn't mean we should actually replace the implementations, I was just demonstrating that merge is truely a more general operation and provide evidence it was useful. The most recent use case for me dealt with a lattice (in the mathematical sense) class, where the result of a meet or join could be a different type than the arguments... or maybe it was a boolean algebra class... hrmm.. or was it an implemenation of approximations of probability distributions where the map contained representative samples... sigh.. I juggle too many projects at once.. :) John