Learning Path: Learn Functional Programming with JavaScript

This video provides an overview of the entire course.

JavaScript has traditionally been taught by the "Imperative or Procedural Programming" paradigms where functions ("procedures") perform tasks and modify a shared state. This can cause unpredictable behavior and hard to track bugs, especially as programs grow. Pure functions alleviate many problems of traditional programming paradigms and form the foundations of Functional Programming.

Pure functions, like mathematical functions, always return the same value when given the same inputs, so the two are interchangeable. Functional Programming is expressive, we think less about what's happening inside functions and focus more on relations between function inputs and their outputs.

Every program has impure functions; they are the only ticket in town for reading, writing and displaying information. In order to protect the benefits, we've gained by writing pure functions; we'll begin looking at how we can isolate impure code while still using it to feed us information.

Pure functions seem to restrict what we’re able to do in our apps, higher-order functions open our eyes to new ways we can write and use functions. They give us powerful tools and reusable utilities.

There are constructs in our programs we have grown accustom to always using, loops are one of these constructs. We can stop writing loops and instead use higher-order functions. This will make our code more readable, expressive and open our thinking to new ways of using higher-order functions.

JavaScript added the higher-order map and filter methods to Arrays in ES5.1 but they don’t work in the way we as functional programmers need. We can make our own implementations that will certainly prove to be more and more useful the further we get into functional programming.

Reduce is another very popular higher-order function used for working over collections of data. It can be a little more difficult to reason about than map and filter, however it can have a lot more uses than map and filter as well.

Higher-order functions did some neat tricks for us working with Arrays, and we aren’t close to finish. Rendering to the DOM can get a lift from higher-order components by allowing us to perform generalized work over arbitrary components

As programs grow, creating higher-order nested functions won’t be as flexible as we once thought. There is a better way to pass a couple input values into a function and still get as return a higher-order function with those values in a closure without wiring nests of functions like Russian dolls.

Now that we understand what partial application and currying are, we’ll want to see how we can use some common partial application utilities to write higher-order functions without writing entire new functions or refactoring old ones. We’ll see how a few different utilities operate.

Partial Application and Currying can take time to adjust to but they become tools we don’t want to live without. It’s nice to be able to quickly write partial application utilities so we can add them to any project. Writing utilities gives the added benefit of gaining stronger understanding.

It’s time again to combine the techniques we’ve been learning and see how they work well with each other. We should be able to use a lot more functions with our higher-order functions map, filter, and reduce as well as instantly improve any function that works over arrays with the “curry” utility.

Making objects immutable could add complexity to an application. Persistent data structures might be complicated to add into a project and many time might not even be needed. Lenses can simplify working with data and make adding new data structures into a project easier.

A function to update the entire DOM at once is more deterministic, but less performant. Virtual-DOM keeps a representation of the DOM in JavaScript so it can figure out which changes to make on updates and only render parts of the DOM that need to change rather than overwriting everything.