Why succ returns Maybe? about purescript-enums HOT 6 CLOSED

I think there might be room for a CyclicalBoundedEnum type class:

I thought about that too, but I think the class would be basically lawless, and I'm not sure it gives you anything that couldn't be defined in terms of BoundedEnum already. Just fewer instances to implement I guess.

from purescript-enums.

For something dynamic like a track list I'd perhaps do something with an infinite (lazy) list, use a zipper, or just a pair of my own functions to handle it.

Yeah, I currently do Track type with rotate :: Track -> Track function. Maybe Enum for Track is not that suitable as my own type, but an interesting idea at start. @garyb and @JordanMartinez thanks for help!

from purescript-enums.

That would break the laws of Enum as defined here, so I guess the answer is "because we chose Enum to mean something that is incompatible with this idea". 😉

Maybe cyclicSucc / cyclicPred functions could be defined in terms of BoundedEnum? That way they're not breaking any laws and they're taking advantage of the existing definitions.

(Also I assume your case isn't actually Boolean, since you could use not for that.)

from purescript-enums.

I think there might be room for a CyclicalBoundedEnum type class:

class CyclicalBoundedEnum where
  skip :: Int -> a -> a
  cyclicalBottom :: a
  cyclicalTop :: a

If something is cyclical, then is it likely that succ/pred could be succeeded with skip?

from purescript-enums.

Also I assume your case isn't actually Boolean, since you could use not for that.

My thoughts were about tracks list, where succ of last track returns first track. With Boolean example I tried to show that Maybe only needs for bounds, when in other realization it can return min bound instead. I also have checked how it is done in Haskell - it has signature succ :: a -> a. I don't have much experience in Haskell, but think it is what I wrote above.

from purescript-enums.

Yeah, Enum in Haskell is quite different, it's only really intended for statically sized things in PureScript. Also its version of succ is still partial, if you try succ True it raises an error, so it's just not exposed in the type the way it is here.

For something dynamic like a track list I'd perhaps do something with an infinite (lazy) list, use a zipper, or just a pair of my own functions to handle it.

from purescript-enums.