Anonymous user
Talk:Partial function application: Difference between revisions
→Proposal for new task description: more explanation of my POV.
(→Proposal for new task description: With Ruby, curry f2.) |
(→Proposal for new task description: more explanation of my POV.) |
||
Line 196:
:: Yes, but currying and partial application are quite closely related and often confused. It may be easier to understand them if they are both used together. Also, if I managed to implement the Ruby example below correctly, currying and partial application seems both to be implemented by Proc#curry. Finally, regarding the complications there seem to be with some of the language: I'm not yet convinced partial application (and currying) can be ''reasonably'' implemented in all languages (as papply and curry are higher-order functions returning a function, I assume you would at least need first-class functions/closures or go through the trouble of emulating those). I'll try to write this task in a few more languages to see if it can be done or not. —''[[User:Ruud Koot|Ruud]]'' 21:56, 15 April 2011 (UTC)
:::Hi again Ruud. I was sure that there would be languages with first-class functions that would not be able to curry/partially-apply. What is new to me is that there may well be languages that are OK with currying, but have problems with partial application as they can't apply a subset of arguments without ''explicitely'' mentioning all arguments in the call to partial. Some manidestly, statically typed languages - i.e. those that require a type signature for everything, would need different and explicit type signatures dependant on how many arguments were being partially applied for example; Haskel works it out for itself.
:::Somehow, I do notice the similarity between currying and PFA, but unlike you, I see it as a strong reason to have two, contrasting tasks. What to do? --[[User:Paddy3118|Paddy3118]] 03:28, 16 April 2011 (UTC)
Here are samples in Haskell and Python:
| |||