What's the difference between a Scala function definition using "object myfunction extends (int => int) { def apply(n:int) {...}}" and "def myfunction(n:int) {}"?

I am reading the code here(http://github.com/nkallen/Rowz/blob/master/src/main/scala/com/twitter/rowz/Hash.scala) but don't understand it well.
Answer:

The former is a hack to define a static method without being a member of any other governing class (container). Actually `myfunction' is an Object (a singleton) which implements `apply()' method thus allowing `myfunction()' call semantic. And an illustration of first-class functions in Scala[1] via extending `int => int' function literal. (pure awesomeness). The latter will not compile at all if it's not a member of another object and remember there are no static methods in Scala. (There's something better.) So instead of having a container, why don't we make our static method "self containing". The former is a self contained static method. If we didn't have this flexibility in Scala, the implementation of a static method would be to have a companion object[2] (or a general purpose utility object) and implement the method there. A sample code below. Definition: object Utilities { def myfunction(n: Int): Int = { // doing something really important here. // serious business. } } Usage: val seriousResult = Utilities.myfunction(42); [1] Programming in Scala - Odersky, Spoon, Venners - Page 177, Section 8.3 "First-class functions" [2] Programming in Scala â€“ Odersky, Spoon, Venners â€“ Page 99, Section 4.3 "Singleton objects"

Berk D. Demir at Quora Visit the source

Was this solution helpful to you?