I wonder if there would be any interest in a fast and compact associative data structure (a.k.a. finite map)? I did a quick writeup here: http://biohaskell.org/GSOC/kmerindex and could also volunteer to mentor it (but would be happy to see somebody else do it, too). As I am often working with large data sets, the cost of the standard data structures (Data.Map and friends) are often causing problems. Judy arrays get me some of the way, but I'd like something more general. Although I want it for shamlessly selfish reasons, I think it would make a worthwhile addition to the Haskell ecosystem at large. Anyway - feel free to contact me if anything is unclear. -k -- If I haven't seen further, it is by standing in the footprints of giants