[sixfiveotwo]

Indeed, you can get a lot more from dependent types than Damas-Hindley-Milner inference, yet does it mean that you should use the former everywhere?