Comments on: Function Currying in Scala http://www.codecommit.com/blog/scala/function-currying-in-scala (permanently in beta) Sun, 29 Aug 2010 20:01:44 -0700 http://wordpress.org/?v=2.9.2 hourly 1 By: Horst Makitta http://www.codecommit.com/blog/scala/function-currying-in-scala/comment-page-1#comment-4957 Horst Makitta Thu, 01 Apr 2010 21:29:03 +0000 http://www.codecommit.com/blog/scala/function-currying-in-scala#comment-4957 Well a bit late to the game, but maybe i get an answer? As a Java-Programmer i am interested in Scala 'cause it presents all those functional advanced concepts in a very readable manner (unlike Lisp). Or let's say in a form that's much easier to grasp when used to Java. However i still fail to see what the difference in currying and using partially applied functions is (besides having to specify the types of the underscores): Both appear to return the exact same function signatures. You example with process, using partially applied functions: def process[A](filter:A=>Boolean,list:List[A]):List[A] = { lazy val recurse = process(filter, _:List[A]) ... Everything else remains the same and it works as expected: scala> process(even,numbersAsc) res11: List[Int] = List(2, 4) scala> process(even,numbersDesc) res12: List[Int] = List(4, 2) So what is the reason for the special curryable form? Well a bit late to the game, but maybe i get an answer?

As a Java-Programmer i am interested in Scala ’cause it presents all those functional advanced concepts in a very readable manner (unlike Lisp). Or let’s say in a form that’s much easier to grasp when used to Java.

However i still fail to see what the difference in currying and using partially applied functions is (besides having to specify the types of the underscores): Both appear to return the exact same function signatures.

You example with process, using partially applied functions:
def process[A](filter:A=>Boolean,list:List[A]):List[A] = {
lazy val recurse = process(filter, _:List[A])
…

Everything else remains the same and it works as expected:
scala> process(even,numbersAsc)
res11: List[Int] = List(2, 4)

scala> process(even,numbersDesc)
res12: List[Int] = List(4, 2)

So what is the reason for the special curryable form?

]]> By: Landei http://www.codecommit.com/blog/scala/function-currying-in-scala/comment-page-1#comment-3314 Landei Thu, 20 Mar 2008 09:27:45 +0000 http://www.codecommit.com/blog/scala/function-currying-in-scala#comment-3314 Very cool post, you have a new reader :-) Very cool post, you have a new reader :-)

]]> By: Daniel Spiewak http://www.codecommit.com/blog/scala/function-currying-in-scala/comment-page-1#comment-3313 Daniel Spiewak Thu, 20 Mar 2008 08:29:02 +0000 http://www.codecommit.com/blog/scala/function-currying-in-scala#comment-3313 Very true, but it would still be nice to get the inference whenever possible. Very true, but it would still be nice to get the inference whenever possible.

]]> By: Florian Hars http://www.codecommit.com/blog/scala/function-currying-in-scala/comment-page-1#comment-3312 Florian Hars Thu, 20 Mar 2008 07:33:34 +0000 http://www.codecommit.com/blog/scala/function-currying-in-scala#comment-3312 The type of the arguments of add5 can't be inferred in general, because there might be overloaded add methods in scope with argument types (Int, Float, Float) and (Int, Boolean, String). Which one should add(5,_,_) resolve to? The type of the arguments of add5 can’t be inferred in general, because there might be overloaded add methods in scope with argument types (Int, Float, Float) and (Int, Boolean, String). Which one should add(5,_,_) resolve to?

]]> By: Daniel Spiewak http://www.codecommit.com/blog/scala/function-currying-in-scala/comment-page-1#comment-3308 Daniel Spiewak Wed, 19 Mar 2008 01:51:16 +0000 http://www.codecommit.com/blog/scala/function-currying-in-scala#comment-3308 Yes, I am misusing the term "dynamic programming". The problem is that there doesn't seem to be a very good term to describe that particular concept. Dynamic typing is close, but insufficient. Meta-programming perhaps, but meta-programming is possible with entirely static constructs. We really need a new term which encompasses all of the modern techniques which go along with languages like Ruby and Groovy. Yes, I am misusing the term “dynamic programming”. The problem is that there doesn’t seem to be a very good term to describe that particular concept. Dynamic typing is close, but insufficient. Meta-programming perhaps, but meta-programming is possible with entirely static constructs. We really need a new term which encompasses all of the modern techniques which go along with languages like Ruby and Groovy.

]]> By: abulafia http://www.codecommit.com/blog/scala/function-currying-in-scala/comment-page-1#comment-3307 abulafia Wed, 19 Mar 2008 00:19:18 +0000 http://www.codecommit.com/blog/scala/function-currying-in-scala#comment-3307 Hi, i think the term "dynamic programming" is misused in this article. You probably mean something like dynamic typing but not http://en.wikipedia.org/wiki/Dynamic_programming. abu Hi,

i think the term “dynamic programming” is misused in this article. You probably mean something like dynamic typing but not http://en.wikipedia.org/wiki/Dynamic_programming.

abu

]]> By: Daniel Spiewak http://www.codecommit.com/blog/scala/function-currying-in-scala/comment-page-1#comment-3302 Daniel Spiewak Tue, 18 Mar 2008 15:47:24 +0000 http://www.codecommit.com/blog/scala/function-currying-in-scala#comment-3302 Unfortunately, you can't *exactly* curry it in Java. At least, not Java 6 or earlier. If closures make it into Java 7, we're going to see a lot more examples of this sort of thing for Java itself, not just derivative languages. Unfortunately, you can’t *exactly* curry it in Java. At least, not Java 6 or earlier. If closures make it into Java 7, we’re going to see a lot more examples of this sort of thing for Java itself, not just derivative languages.

]]> By: German B. http://www.codecommit.com/blog/scala/function-currying-in-scala/comment-page-1#comment-3300 German B. Tue, 18 Mar 2008 12:36:10 +0000 http://www.codecommit.com/blog/scala/function-currying-in-scala#comment-3300 Good post! So I guess you can curry a Java method, which is nice. Good post!
So I guess you can curry a Java method, which is nice.

]]> By: Daniel Spiewak http://www.codecommit.com/blog/scala/function-currying-in-scala/comment-page-1#comment-3295 Daniel Spiewak Tue, 18 Mar 2008 00:32:09 +0000 http://www.codecommit.com/blog/scala/function-currying-in-scala#comment-3295 Good point, Alex. Currying is the correct term. I tend to use "function currying" because it just sounds more clear. Not really to disambiguate, but to keep the terms associated for the sake of the uninitiated. Good point, Alex. Currying is the correct term. I tend to use “function currying” because it just sounds more clear. Not really to disambiguate, but to keep the terms associated for the sake of the uninitiated.

]]> By: Alex Blewitt http://www.codecommit.com/blog/scala/function-currying-in-scala/comment-page-1#comment-3294 Alex Blewitt Tue, 18 Mar 2008 00:09:23 +0000 http://www.codecommit.com/blog/scala/function-currying-in-scala#comment-3294 I think most of the time this is just referred to as 'currying' rather than 'function currying'. You can't really curry a constant ;-) Otherwise, good post. I think most of the time this is just referred to as ‘currying’ rather than ‘function currying’. You can’t really curry a constant ;-) Otherwise, good post.

]]>