First: Haskell isn't dead. People are looking at consolidating libraries; for signs of life see the Haskell mail list.|
But why is it important that Haskell become a "real-world language"? That's not the only measure of success. It's also important to have places to develop new ideas without having to worry too much about backwards compatability, for example.
While Haskell has some practical advantages over OCaml (mainly syntax), almost(?) everything that's possible in Haskell can be done in OCaml (or CL for that matter) with little extra effort - and those languages also support OOP, imperative programming, and other features (higher order functors in OCaml, for example) that make them more practical for larger projects.
Haskell is just fine as it is for exploring the issues associated with a very pure approach to functional programming. Turning it into a more general language would duplicate what we already have and lose something unique.
Of course, I'm not saying it's perfect - more libraries would be nice - but it is an excellent tool for learning about some fun ideas in computer science: exactly what it was intended to be.
[compare this with Ada that, afaik, was intended to be a popular, large-scale language...]