On 03/23/2015 04:43 PM, martin wrote:

Hello all

I skimmed several Arrow tutorial and looked at the pretty diagrams and thought to myself "Hmm, these guys seem to take input and produce output".

But that's what a function does. So what's the fundamental difference between an arrow and a fuction?

They're basically functions, and you won't steer yourself wrong thinking about them like that. By analogy, in Haskell, if you wanted to add two floating point numbers you would need to define a floating point addition: floatAdd :: Float -> Float -> Float Then if you wanted to add two integers, you'd need a second function: intAdd :: Int -> Int -> Int The prelude already abstracts this away for you -- both Int and Float are instances of the Num class, which gives you, (+) :: (Num a) => a -> a -> a So at that point, you might ask yourself, what's the point of a Num? Aren't they just (floating point, int, etc.) numbers? The answer is "yes," and you don't want to have twenty different addition functions for everything that is some kind of number, so it makes sense to group all number-things into one typeclass. Likewise, all function-things are grouped into the Arrow typeclass so that you can compose them without worrying about the exact type. Everything in the Arrow typeclass is some kind of function, and you can compose them with the Arrow class methods.