[Haskell-cafe] Trouble with currying and uncurrying...

[Angel de Vicente]

Hi,

On 25/04/11 14:21, Stephen Tetley wrote:
> On 25 April 2011 14:11, Angel de Vicente<angelv at iac.es>  wrote:
>
>> curry :: ((a,b) ->  c) ->  (a ->  b ->  c)
>> curry g x y = g (x,y)
>
> Is expressing curry this way more illuminating?
>
> curry :: ((a,b) ->  c) ->  (a ->  b ->  c)
> curry g = \x y ->  g (x,y)
>
> That is, curry is a function taking one argument that produces a
> result function taking two arguments.
>
> In Haskell - the type signature:
>
>> curry :: ((a,b) ->  c) ->  (a ->  b ->  c)
>
> ... can mean either form.
>
> Though this is a "quirk" of Haskell, in the language Clean, for
> example, the parens in the type signature "mean what they" say so only
> only my second definition is allowed, the first version won't compile.

This was my first encounter with this syntax, but it was a bit confusing.
curry :: ((a,b) -> c) -> a -> b -> c

is much clearer to me, and once partial application is understood it is 
no problem to see what curry f  or curry f x y mean. But to me (at this 
point in my Haskell trip) it would make more sense that once you put 
the parentheses as in the original definition, no partial application is 
allowed and so the lambda notation is required.

Thanks a lot,
Ángel de Vicente
-- 
http://www.iac.es/galeria/angelv/

High Performance Computing Support PostDoc
Instituto de Astrofísica de Canarias
---------------------------------------------------------------------------------------------
ADVERTENCIA: Sobre la privacidad y cumplimiento de la Ley de Protección de Datos, acceda a http://www.iac.es/disclaimer.php
WARNING: For more information on privacy and fulfilment of the Law concerning the Protection of Data, consult http://www.iac.es/disclaimer.php?lang=en