These examples are very helpful. Now that you have collected them, I
#12363: Type application for infix -------------------------------------+------------------------------------- Reporter: Iceland_jack | Owner: Type: feature request | Status: new Priority: lowest | Milestone: Component: Compiler | Version: 8.0.1 (Parser) | Keywords: Resolution: | TypeApplications Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by Iceland_jack): Replying to [comment:8 goldfire]: think a good next step is soliciting feedback from the community so that we can see if others agree with you. I'll admit to becoming more convinced by your examples, but I'm still leaning against. Other voices in agreement with yours would make a big difference here. I agree that this needs input from the community. The notation often feels clunky and asymmetric, mathematicians would use subscript instead. Visual type application (+ partial type signatures) have been very useful * '''Interactive use''' Toying around, getting ''ghci'' to resolve instances for me {{{#!hs mappend @(_ -> _) :: Monoid b => (a -> b) -> (a -> b) -> (a -> b) quickCheck @(_ -> _) :: (Arbitrary a, Show a, Testable b) => (a -> b) -> IO () quickCheck @(Int -> _) :: Testable b => (Int -> b) -> IO () }}} * '''Teaching''' Using valid Haskell syntax to simplify complex signatures that users can verify on their own {{{#!hs length @[] :: [a] -> Int sequence @[] @IO :: [IO a] -> IO [a] }}} Less so in actual code. The same seems true for this so I am not overly concerned with the syntax. Indeed most of the benefits of the examples don't have to do with actual code, rather presenting concepts with explicit/valid syntax or interactive use: allowing users to quickly specify a type without rearranging/rewriting syntax and adding parentheses (`(.) @A (id @A) f`). No need to muck about with expressions which reduces mental load: if you want `<$>` for lists, write `<$> @[]`. If it's all the same I will post examples until told to stop ;) (for operators, `_` with no spaces looks nicer (`1 +_Int 10`) but I won't seriously suggest it) -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/12363#comment:9 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler