Writing kernel modules in Haskell

10x the effort might still cheap enough for some people.

A factor 10 effort multiplication on some minor addition to an already working system might easily be dwarfed by a) how much effort is already expended daily on regular maintenance and b) is saved further down the chain because a million distributers and kernel devs don't need to understand how to get Haskell going for their kernel build when that module is written in C ;-)