> However, it's often really painful to get that code to typecheck compared with the effort to get the same code correct in python or ruby.
I'd argue that in the overwhelming majority of cases, the Python or Ruby code is equally incorrect to the equivalent Haskell code that doesn't type check, it's just that Python and Ruby don't tell you about it upfront, you find out about it at runtime. Hopefully, "runtime" means "during testing" and not after your code is in production. Or if you want to put a more positive spin on it, Python and Ruby let you run the code in spite of its lack of correctness. If that's the kind of freedom you want, then a dynamically typed language (or even a weak statically typed language) will suit you well. As for me, I've been burned by that "freedom" enough times now that I can't think of it as anything but a negative.