The change to make Monad a special case of Applicative has been a long time coming, and it has been a long time coming precisely because it's going to break things. I feel ambivalent about this. As soon as the question was raised, it was clear that it *would have been* nicer had Haskell been this way originally. I'm not looking forward to the consequences of the change. On 7/10/2015, at 12:32 am, Henrik Nilsson <Henrik.Nilsson@nottingham.ac.uk> wrote:
While we can discuss the extent of additional breakage MRP would cause, the fact remains it is a further breaking change. A survey of breakage to books as Herbert did is certainly valuable (thanks!), but much breakage will (effectively) remain unquantifiable.
This suggests one way in which one of the costs of the change could be reduced. Not just to survey the breakage, but to actually put revised material up on the Haskell.org web site, with a prominent link from the home page.