Hi all, In the light of the recent discussion, I've updated DData as such 1. "Seq" has been reverted to what it was (strict append). If anyone is interested in the fully lazy version of sequences that I've hacked, I keep it available under "DData/LazySeq.hs". 2. I've added {# UNPACK #} as suggested. I understand that it suffices to keep DData on par with FiniteMap/Set as it exists now. Please correct me if I'm wrong. 3. I've modified IntSet and IntMap in the same fashion. data IntMap a = Nil | Tip {-# UNPACK #-} !Key a | Bin {-# UNPACK #-} !Prefix {-# UNPACK #-} !Mask !(IntMap a) !(IntMap a) -- notice that the Tip is unpacked as well. That change speeds up the following "benchmark" twofold. import DData.IntSet s = fromList [1..1000000] main = print $ all (`member` s) [1..1000000] I did not try to be precise in the measurement, since it seemed obvious that there was an improvement. The source is at the usual URL (users.skynet.be/jyp/DData/ddata.tar.gz) but I have not updated the docs. Cheers, JP. __________________________________ Do you Yahoo!? Yahoo! Domains � Claim yours for only $14.70/year http://smallbusiness.promotions.yahoo.com/offer