What is the following methods' difference?
def sum1() = 1+2 def sum2(a:Unit) = 1+2 I think they are semantically identical, is it right?
Add a comment |
3 Answers
With sum1, you can call it with or without parentheses:
val x = sum1 // x: Int = 3 val y = sum1() // y: Int = 3 But with sum2 you are forced to provide parentheses.. I think that if you call sum2(), you are actually calling sum2 with () as the argument a.
val x2 = sum2 // error val y2 = sum2() // y2: Int = 3 
1 Comment
Dan Burton
A good rule of thumb is to require parens when the computation involves a side-effect, but don't require them when the computation is pure. I guess you could use parens this way in the same way that Lisp and Ruby use
!.Note that passing unit as an argument to an expression lets you simulate lazy evaluation in a strict language. By "moving evaluation under a lambda" you ensure that the expression isn't eval'd until the () gets passed in. This can be useful for e.g. auto-memoizing data structures, which collapse from a function to a value the first time they're inspected.
Comments
These methods are not identical. Once receives a parameter, the other does not. See here:
scala> sum1(println("Hi, there!")) <console>:9: error: too many arguments for method sum1: ()Int sum1(println("Hi, there!")) ^ scala> sum2(println("Hi, there!")) Hi, there! res1: Int = 3 