Is Haskell suffering from a research language problem where the only things considered valuable are things that can be made into papers, i.e. only new additions?