| nmessenger | ^o^ |
| sioraiocht | LOL |
| dzhefri | It does? I just tried ":t (>>>)" and i'm told "not in scope" |
| nmessenger | Control.Arrow.>>> |
| sorear | dzhefri: Control.Arrow stepcut: so, I've accomplished my goal of attracting attention? |
| dolio | @pl \f g -> runKleisli (Kleisli f >>> Kleisli g) |
| lambdabot | (runKleisli .) . (. Kleisli) . (>>>) . Kleisli |
| sorear | dolio: >=> |
| dolio | Yeah, I know. I was just seeing how worse >>> is. :) How much worse, that is. |
| jcreigh | heh heh...like the {op}= (as opposed to {op}=) form in C...they realize that i=-1 is ambiguous, so they changed it to i-=1, but were reluctant to disallow the =- form because of the then-thousands of lines of code that depended on it. |
| dzhefri | hmm.... now I'm a bit confused i'm told (>>>) :: (Arrow a) => a b c -> a c d -> a b d |
| sorear | dzhefri: that's correct |
| dolio | Yeah, but functions are arrows. |
| nmessenger | dzhefri: yep, so a = (->) |
| sorear | dzhefri: as is (>>>) :: (a -> b) -> (b -> c) -> (a -> c) === |
| dzhefri | I think what i'm thinking of should be (a -> b) -> (b -> c) -> a -> c |
| sorear | (>>>) :: ((->) a b) -> ((->) b c) -> ((->) a c) === (>>>) :: a b c -> a c d -> a b d (where a = (->)) |
| stepcut | sorear: I agree with the sentiment that it would suck to get stuck with what we have today, just because there are a few thousand haskell users, when someday there will be millions of users |
| sorear | <= |
| jcreigh | a few thousand? Is that all? |
| sorear | (>>>) :: Arrow a => a b c -> a c d -> a b d |
| dzhefri | hmmm.... what's this arrow business? |
| dolio | If you use some funky GHC stuff, you can write (I think): (Arrow (~>)) => (a ~> b) -> (b ~> c) -> (a ~> c) |
| sorear | yet another obscure abstraction |
| stepcut | jcreigh: well, depends on who you count -- there are probably thousands of students alone |
| dolio | Which is perhaps somewhat more lucid, since you can replace ~> with -> there. |
| action | sorear changes topic to: #haskell is not #ghc! ... |
| sorear | changes topic to: #haskell is not #ghc! ... |
| nmessenger | dzhefri: an Arrow is just a type that accepts two types, supports some functions, and adheres to laws. The (->) type fits the definition. |
| stepcut | jcreigh: but, perhaps, thousands (10,000 - 100,000) people that would be more severely affected by breaking Prelude |
| sorear | IMO we should introduce breaking changes into the language every 3 months |
| dzhefri | nmessenger, okay, so what does it mean to support a function? |
| jcreigh | "Haskell: thousands of voluntary users" :) |
| sorear | whether or not it is needed |