There's a bit of a chicken-and-egg problem here; the COM tools are not well maintained, so that discourages use, which in turn makes it less rewarding to work on them. What I don't know is the level of suppressed demand: if there were good tools, would lots of people start using them?
I think it will be beneficial for many Windows user if there was a good Haskell to COM integration. HDirect can be a real nightmare for a large projects. The first step could be to build a new COM library designed in more FFI addendum fashion. I already have started the first step with: http://darcs.haskell.org/packages/hscom/ It is far from complete of course. The second step should be to reimplement the IDL->Haskell translator. The existing translator often generate code that isn't type correct and even if you change the code by hand there is still a chance to have an implementation errors which can't be captured from the typechecker. The whole task isn't that hard but still requires a volunteer willing to spend 2-3 months of full time work. Regards, Krasimir