/ EXERCISE, PROGRAMMING, STYLE

Exercises in Programming Style, recursion

This week’s post will be back to fundamentals, as the constraint is to use recursion:

Recursion in computer science is a method of solving a problem where the solution depends on solutions to smaller instances of the same problem (as opposed to iteration). The approach can be applied to many types of problems, and recursion is one of the central ideas of computer science.

— Wikipedia
https://en.wikipedia.org/wiki/Recursion_(computer_science)

This is the 4th post in the Exercises in Programming Style focus series.Other posts include:

  1. Introducing Exercises in Programming Style
  2. Exercises in Programming Style, stacking things up
  3. Exercises in Programming Style, Kwisatz Haderach-style
  4. Exercises in Programming Style, recursion (this post)
  5. Exercises in Programming Style with higher-order functions
  6. Composing Exercises in Programming Style
  7. Exercises in Programming Style, back to Object-Oriented Programming
  8. Exercises in Programming Style: maps are objects too
  9. Exercises in Programming Style: Event-Driven Programming
  10. Exercises in Programming Style and the Event Bus
  11. Reflecting over Exercises in Programming Style
  12. Exercises in Aspect-Oriented Programming Style
  13. Exercises in Programming Style: FP & I/O
  14. Exercises in Relational Database Style
  15. Exercises in Programming Style: spreadsheets
  16. Exercises in Concurrent Programming Style
  17. Exercises in Programming Style: sharing data among threads
  18. Exercises in Programming Style with Hazelcast
  19. Exercises in MapReduce Style
  20. Conclusion of Exercises in Programming Style

Principles of recursion

I already wrote about recursion in the context of applying Functional Programming's approach to the Dijkstra algorithm. Let’s go into the details of its implementation. A recursive function should provide two branches:

  1. A stop branch that returns the final result
  2. A branch that calls the function itself with different arguments

Here’s a simple implementation of the factorial function:

fun fact(n: Int): Int {
    var result = 1
    for (i in 1..n) {
        result *= i
    }
    return result
}

fact(5)

In traditional imperative programming, functions use local variables to accumulate temporary computations. In recursive functions, those variables are replaced by function parameters.

The recursion equivalent of the previous function is the following:

private fun recurseFact(acc: Int, n: Int): Int =
    if (n == 1) acc                                  (1)
    else recurseFact(acc * n, n - 1)                 (2)

fun fact(n: Int) = recurseFact(1, n)                 (3)

fact(5)
1 Stop branch
2 Self-call branch
3 Same as the previous function signature

Note that acc parameter plays the same role as the result local variable in the imperative example.

The following is the function from the exercise. It applies the exact same principles as described in the factorial example:

fun words(rest: List<String>,
                  stopwords: List<String>,
                  words: List<String>): List<String> {
    return if (rest.isEmpty()) words
    else {
        val split = split(rest.last(), stopwords, listOf())
        words(rest.dropLast(1), stopwords, words + split)
    }
}

Issues with recursion

While the recursion code is much more concise than the imperative one, it suffers from a huge issue: function calls are pushed on the thread’s call stack. The infamous StackOverflowError is thrown when the call stack’s size is exceeded.

While the stack size can be set at startup time, it’s nonetheless finite in size.

Command-line options to manage the stack size
-Xss

Standard JVM Hotspot option

-XX:ThreadStackSize

Proprietary option subject to change without notice

To cope with such an issue, Kotlin (and Scala) offers a compiler trick: while the source code is recursive, the compiled bytecode is implemented with standard loops. There are two requirements:

  1. The recursive function call must be the last one. This is called tailed recursion.
  2. The modifier tailrec must be added to the function signature

The words() function above is tail-recursive, hence it’s quite easy to add the tailrec keyword. It can be updated accordingly, so that it will never overflow.

On the opposite, the following function is not tail-recursive because there are two calls using recursion. Hence, the bytecode cannot be optimized.

fun <T> quicksort(list: List<Pair<T, Int>>): List<Pair<T, Int>> =
    if (list.size <= 1) list
    else list.random().let { pivot ->
        val below = filter(list, listOf()) { it.second <= pivot.second }
        val above = filter(list, listOf()) { it.second > pivot.second }
        quicksort(below - pivot) + pivot + quicksort(above)
    }

Conclusion

In general, recursion is a tough nut to crack at first. However, migrating one’s code to use recursion is a simple recipe: just move the local variables to accumulator parameters. The hardest part is to make recursive functions tail-recursive to avoid to overflow the stack. Some functions allow it, some do not.

The complete source code for this post can be found on Github.
Nicolas Fränkel

Nicolas Fränkel

Developer Advocate with 15+ years experience consulting for many different customers, in a wide range of contexts (such as telecoms, banking, insurances, large retail and public sector). Usually working on Java/Java EE and Spring technologies, but with focused interests like Rich Internet Applications, Testing, CI/CD and DevOps. Also double as a trainer and triples as a book author.

Read More
Exercises in Programming Style, recursion
Share this