What I'm trying to say is simple: personal aesthetic preference, empirical results and research questions are all very important, and every one of them is a reasonable motivation for making decisions, but they are very different from another, and should not be conflated.

As to Haskell, it is definitely production-ready, and usable in a variety of circumstances. Also, its design and evolution are just as definitely primarily guided by research goals. These two simple, obvious, facts are not in opposition to one another. I have read numerous articles on why some people like Haskell, or many other languages. They are very important: some people like the tracking of "side effects"; some enjoy the succinctness. None of those articles, however, have any bearing on other questions. Which are not necessarily more important, but they are very different.