Scala vs. F#: Comparing Functional Programming Features

• facebook
• twitter
• google+
• in
• Email

F# and Scala, two relatively recent programming languages, provide most .NET and Java software developers with new functional programming features that are worth understanding and evaluating. (Read a Short Briefing on Functional Programming for a primer on how functional programming works.)

Scala is a Java-based, general-purpose language that is intended to appeal to both adherents of functional programming as well as devotees of more mainstream imperative and object-oriented programming. It compiles into Java bytecode and runs on top of the Java Virtual Machine (JVM).

While Scala is fundamentally a functional language, it also embodies all the elements of imperative and object-oriented languages, which gives it the promise of introducing functional programming features to a broader programming community.

F# is a general-purpose programming language developed by Microsoft to run as part of .NET's Common Language Runtime (CLR). It is based on another, orthodox functional language, Ocaml. Microsoft introduced F# into the .NET platform because of, among other reasons, the increased interest in functional programming and functional programming's suitability to high-performance computing and parallelism.

Although its syntax is distinctly functional, F# actually is a hybrid functional/imperative/object-oriented language. Its object-oriented and imperative features are present mostly for compatibility with the .NET platform, but F#'s tripartite nature is also pragmatic -- it allows programmers who use any or all of the three programming paradigms to program exclusively in one or to combine all three.

In this article, I will compare and contrast the functional features and related syntax of F# and Scala.

F# vs. Scala: First Order Functions

Functions in F# and Scala are treated as first order types. They can be passed in as arguments, returned from other functions, or assigned to a variable.

In this F# code snippet, I first define a function (increment) that adds 1 to a passed value, and then I define the function handler, which takes type myfunc and applies 2 to it as a parameter. Finally, I invoke the function handler with a parameter incremented to it. The function incrementis passed as a regular value, hence the function is being treated as a first order type:

let increment x = x + 1
let handler myfunc = (myfunc 2)   
printfn "%A" (handler increment)

Notice the type inference in the example above. F# will infer that x is an integer because I add 1 to it, and so x will be treated as an integer (Int) type.

Here is the same example in Scala:

def increment(x:Int) = x + 1
def handler( f:Int => Int) = f(2)
println( handler( increment ))

F# vs. Scala: Lazy Evaluation

F# supports lazy evaluation, but for performance reasons it is not enabled by default. Instead, F# supports so-called eager evaluation: functions can be marked for lazy evaluation by explicitly labeling them with the keyword lazy and running the program with the Lazy.force option specified.

let lazyMultiply = lazy ( let multiply = 4 * 4  )

Like F#, Scala is not lazy by default, but unlike F#, values -- not functions --have to be marked with the keyword lazy and therefore evaluated as call-by-need.

def lazyMultiply(x: => y:) = { lazy val y = x * x }

F# vs. Scala: Lambda Expressions and Currying

Currying is an essential feature of functional programming that allows for the partial application of functions and functional composition. F# supports currying. Here is an example of the curried function add in F#:

Declaration:

val add : int -> int -> int

Implementation:

let add = (fun x -> (fun y -> x + y) )

In Scala, the curried function add looks like this:

def add(x:Int)(y:Int) = x + y

F# vs. Scala: Lambda Expressions

F# also supports Lambda expressions (anonymous functions). In F#, lambda expressions are declared with the keyword fun. In the example below (adopted from F# documentation), an anonymous function is applied to a list of numbers to increment each number in the list and return a new, incremented list:

let list = List.map (fun i -> i + 1) [1;2;3] 
printfn "%A" list

Lambda expressions in Scala are defined in a very succinct fashion. This is how you would increment a function with a Lambda expression (x=>x+1) on a list of numbers (1,2,3) in Scala:

val list = List(1,2,3).map( x => x + 1 ) 
println( list )

F# vs. Scala: Pattern Matching

Pattern matching is a powerful feature of functional programming languages that allows blocks of code within the function to be 'activated' depending on the type of a value or an expression. (Think of pattern matching as a more powerful variation of the case statement.)

In F# , the vertical line character (|) is used to denote a case selector for the function match specification. Here is an F# version of a pattern-matched Fibonacci number function:

let rec fib n =
     match n with
     | 0 -> 0
     | 1 -> 1
     | 2 -> 1
     | n -> fib (n - 2) + fib (n - 1)

Like F#, Scala supports pattern matching on functions. Here is the example of a Fibonacci number calculation in Scala. Notice that Scala uses the keyword case:

def fib( n: Int): Int = n match {

    case 0 => 0
    case 1 => 1
    case _ => fib( n -1) + fib( n-2)
  }

Page 1 of 2