Introducing Erlang

Chapter 7. Higher-Order Functions and List Comprehensions

Higher-order functions, functions that accept other functions as arguments, are a key place where Erlang’s power really starts to shine. It’s not that you can’t do higher-order functions in other languages—you can in many—but more that Erlang treats higher-order functions as a native and natural part of the language rather than an oddity.

Simple Higher-Order Functions

Way back in Chapter 2, you saw how to use a fun to create a function:

1> Fall_velocity = fun(Distance) -> math:sqrt(2 * 9.8 * Distance) end.
#Fun<erl_eval.6.111823515>
2> Fall_velocity(20).
19.79898987322333
3> Fall_velocity(200).
62.609903369994115

Erlang not only lets you put functions into variables, it lets you pass functions as arguments. This means that you can create functions whose behavior you modify at the time you call it, in much more intricate ways than is normally possible with parameters. A very simple function that takes another function as an argument might look like Example 7-1, which you can find in ch07/ex1-hof.

Example 7-1. An extremely simple higher-order function

-module(hof).
-export([tripler/2]).

tripler(Value, Function) -> 3 * Function(Value).

The argument names are generic, but fit. tripler/2 will take a value and a function as arguments. It runs the value through the function, and multiplies that result by three. In the shell, this might look like the following:

1> c(hof).
{ok,hof}
2> MyFunction=fun(Value)->20*Value end.
#Fun<erl_eval.6.111823515> ...