Thoughts on Functional Javascript

Written by Eric Rasmussen on September 9, 2013.

I have long struggled with the necessity of JavaScript and the pain it brings, and was naturally very interested in what Michael Fogus’s new book had to offer. Functional JavaScript is a book that appears to serve two very different goals:

  1. teaching functional programming to JavaScript programmers
  2. teaching JavaScript programming to functional programmers

Balancing the two is no easy task, but Fogus guides the reader expertly, never dwelling so long on a particular concept as to disengage readers from one camp or the other. I approached the book from a functional background, but was surprised at how much I learned from Fogus’s explanation of functional concepts in the context of a language that doesn’t have native support for my favorite abstractions.

The biggest selling point for Functional JavaScript is the book finds a way to combine the two more elegantly than I could have imagined. I was about ready to give up on vanilla JavaScript and move on to an altJS language1, but Fogus has helped restore some of my faith in being able to write reasonable software in JavaScript despite its failings. This is a good thing, since we’ll be stuck with it for the foreseeable future2.

Functional Interfaces

As a functional programming evangelist (read: someone who talks way too much about functional programming), one of the most common questions I get from object oriented (OO) programmers is how you can reuse code without objects. Here’s my table for that:

Language Code Reuse
Haskell Types
Scala Types
OCaml Types
JavaScript mumbles about higher order functions and composition

Fogus’s answer, it turns out, is a whole lot better than mine. He guides the reader through building functional APIs that can:

  • minimize state (using purity where applicable)
  • leverage functions as units of abstraction (encapsulation)
  • write property-based tests for impure functions
  • view programs in terms of data flow (pipelines)

The examples used throughout the book are illustrative and concise. He even manages to build functions for currying and partial application that let me employ the functional abstractions I know and love, without mangling JavaScript to do it.

Embracing the Craziness

Whenever I write JavaScript I have to tread carefully and silently rage at HOW DO YOU PEOPLE PROGRAM THIS WAY. JavaScript offers limitless ways to shoot yourself in the foot, and is missing a number of constructs I find increasingly difficult to program without, including algebraic data types. To avoid the error-prone nature of manually managing inputs and outputs that can be many different types of objects, I typically create restrictive APIs that only expect and return well-defined objects or classes of objects.

Many of Fogus’s examples run contrary to my intuition here, and fully embrace JavaScript’s ability to pass around objects of any type. The book’s samples are meant to be illustrative of course, but you will see examples like a flat function that is designed to take an array and recursively flatten any arrays it contains. One consequence is that the function’s sole parameter, ary, will eventually be a non-array value that needs to be wrapped in an array when returned.

But setting aside my usual predilections, if you squint hard enough, you can see these values as sum types and accept this as a very reasonable and flexible way to write extensible JavaScript. There is certainly no question that it works, is understandable, and takes better advantage of the language than my usual rigid adherence to pretending the issue doesn’t exist.

Many of the other chapters focus on collections-based programming with higher order functions, leveraging primitive types, and encapsulation through functions. Fogus also answers the question of how to take advantage of objects without giving into inheritance, and he manages to weave an object compositional style (mixins) with a functional API in a very slick way.


I would have liked to see more on how to use these ideas to build rich user interfaces, but there’s a lot of disagreement on that topic in general, and it’s an open problem for the functional community. However, even though it’s not a central idea for the book (nor should it be), Fogus takes care in the appendix to touch on several alternative libraries and altJS languages.

Functional JavaScript primarily uses underscore.js in its examples. The ideas in the book are in no way dependent on the library, but it offers stable implementations of many higher order and collections-based functions to make the functional programmer feel at home. As nice as it is to not have to write for loops for everything, underscore.js isn’t quite what I’m looking for in a functional library, but there’s no real consensus or widely accepted library for the displaced functional programmer writing JavaScript. Next on my list is giving a go to see if it addresses some of my personal issues (which, in any case, would be quite a feat).

(function conclusion () {

If you write JavaScript, read this book!

You’ll learn a lot about writing clean, maintainable code that is applicable to any imperative language, but especially encouraging for JavaScript, a language that is easily, and frequently, abused.


1. Any language that compiles to JavaScript. There’s a nice list at

2. Largely on the web, but expect the list of non-web uses on Wikipedia to continue growing.

Tagged: code, javascript