i-am-tom / haskell-exercises Goto Github PK

@i-am-tom I am going through the haskell exercises, and in the KindSignatures section I noticed that you explain how class Something a will have a kind of *, and you can change that by adding the KindSignatures pragma.

I think this is great, but I feel there’s a step in between that could be explained to avoid confusion. That is, GHC will correctly infer the type of a based on the method signatures provided, and most of the time you will provide method signatures, so you might not need to jump straight into KindSignatures for it.

class Functor f where
  fmap :: (a -> b) -> f a -> f b

:k Functor
Functor :: (* -> *) -> Constraint


class Stuff a where
   lol :: a b c d

:k Stuff
Stuff :: (* -> * -> * -> *) -> Constraint

This also explains why you mention that it’s a good pragma to make documentation more explicit.

A few notes on the MPTC explanation while I remember:

• https://github.com/i-am-tom/haskell-exercises/blob/master/exercise09/src/MultiParamTypeClasses.hs#L178 - I don't think this series of steps quite makes sense: if we have a winner here (at 3) one way or the other then what are 4 or 5 for?
• https://github.com/i-am-tom/haskell-exercises/blob/master/exercise09/src/MultiParamTypeClasses.hs#L202 - This line compiles just fine, because there's only one Sneaky instance, but the text suggests it will fail (and the explanation for why, on L213, refers to Transform rather than Sneaky)
• https://github.com/i-am-tom/haskell-exercises/blob/master/exercise09/src/MultiParamTypeClasses.hs#L224 - I'm confused by the parenthesised clause in this sentence; it's very abrupt

Thank you so much for the course!
Is there any link to the canonical solutions?