The Anatomy of a Modern JavaScript Application

Share this article

A woman playing a life-size game of Operation, removing a JS logo from the torso
This article is featured in our book,
JavaScript: Best Practice. Keep on top of the rapidly changing best practices of Modern JavaScript.

There’s no doubt that the JavaScript ecosystem changes fast. Not only are new tools and frameworks introduced and developed at a rapid rate, the language itself has undergone big changes with the introduction of ES2015 (aka ES6). Understandably, many articles have been written complaining about how difficult it is to learn modern JavaScript development these days.

In this article, I’ll introduce you to modern JavaScript. We’ll take a look at recent developments in the language and get an overview of the tools and techniques currently used to write front-end web applications. If you’re just starting out with learning the language, or you’ve not touched it for a few years and are wondering what happened to the JavaScript you used to know, this article is for you.

A woman playing a life-size game of Operation; a metaphor for the many components of a modern JavaScript app"

A Note About Node.js

Node.js is a runtime that allows server-side programs to be written in JavaScript. It’s possible to have full-stack JavaScript applications, where both the front and back end of the app is written in the same language. Although this article is focused on client-side development, Node.js still plays an important role.

The arrival of Node.js had a significant impact on the JavaScript ecosystem, introducing the npm package manager and popularizing the CommonJS module format. Developers started to build more innovative tools and develop new approaches to blur the line between the browser, the server, and native applications.

JavaScript ES2015+

In 2015, the sixth version of ECMAScript — the specification that defines the JavaScript language — was released under the name of ES2015 (still often referred to as ES6). This new version included substantial additions to the language, making it easier and more feasible to build ambitious web applications. But improvements don’t stop with ES2015; each year, a new version is released.

Declaring variables

JavaScript now has two additional ways to declare variables: let and const.

let is the successor to var. Although var is still available, let limits the scope of variables to the block (rather than the function) they’re declared within, which reduces the room for error:

// ES5
for (var i = 1; i < 5; i++) {
  console.log(i);
}
// <-- logs the numbers 1 to 4
console.log(i);
// <-- 5 (variable i still exists outside the loop)

// ES2015
for (let j = 1; j < 5; j++) {
  console.log(j);
}
console.log(j);
// <-- 'Uncaught ReferenceError: j is not defined'

Using const allows you to define variables that cannot be rebound to new values. For primitive values such as strings and numbers, this results in something similar to a constant, as you cannot change the value once it has been declared:

const name = 'Bill';
name = 'Steve';
// <-- 'Uncaught TypeError: Assignment to constant variable.'

// Gotcha
const person = { name: 'Bill' };
person.name = 'Steve';
// person.name is now Steve.
// As we're not changing the object that person is bound to, JavaScript doesn't complain.

Arrow functions

Arrow functions provide a cleaner syntax for declaring anonymous functions (lambdas), dropping the function keyword and the return keyword when the body function only has one expression. This can allow you to write functional style code in a nicer way:

// ES5
var add = function(a, b) {
  return a + b;
}

// ES2015
const add = (a, b) => a + b;

The other important feature of arrow functions is that they inherit the value of this from the context in which they are defined:

function Person(){
  this.age = 0;

  // ES5
  setInterval(function() {
    this.age++; // |this| refers to the global object
  }, 1000);

  // ES2015
  setInterval(() => {
    this.age++; // |this| properly refers to the person object
  }, 1000);
}

var p = new Person();

Improved Class syntax

If you’re a fan of object-oriented programming, you might like the addition of classes to the language on top of the existent mechanism based on prototypes. While it’s mostly just syntactic sugar, it provides a cleaner syntax for developers trying to emulate classical object-orientation with prototypes.

class Person {
  constructor(name) {
    this.name = name;
  }

  greet() {
    console.log(`Hello, my name is ${this.name}`);
  }
}

Promises / Async functions

The asynchronous nature of JavaScript has long represented a challenge; any non-trivial application ran the risk of falling into a callback hell when dealing with things like Ajax requests.

Fortunately, ES2015 added native support for promises. Promises represent values that don’t exist at the moment of the computation but that may be available later, making the management of asynchronous function calls more manageable without getting into deeply nested callbacks.

ES2017 introduced async functions (sometimes referred to as async/await) that make improvements in this area, allowing you to treat asynchronous code as if it were synchronous:

async function doAsyncOp () {
  var val = await asynchronousOperation();
  console.log(val);
  return val;
};

Modules

Another prominent feature added in ES2015 is a native module format, making the definition and usage of modules a part of the language. Loading modules was previously only available in the form of third-party libraries. We’ll look at modules in more depth in the next section.

There are other features we won’t talk about here, but we’ve covered at some of the major differences you’re likely to notice when looking at modern JavaScript. You can check a complete list with examples on the Learn ES2015 page on the Babel site, which you might find useful to get up to date with the language. Some of those features include template strings, block-scoped variables and constants, iterators, generators, new data structures such as Map and Set, and more.


To learn more about ES2015, check out our Premium course: Diving into ES2015.


Code linting

Linters are tools that parse your code and compare it against a set of rules, checking for syntax errors, formatting, and good practices. Although the use of a linter is recommended to everyone, it’s especially useful if you’re getting started. When configured correctly for your code editor/IDE, you can get instant feedback to ensure you don’t get stuck with syntax errors as you’re learning new language features.

You can check out ESLint, which is one of the most popular and supports ES2015+.

Modular Code

Modern web applications can have thousands (even hundred of thousands) of lines of code. Working at that size becomes almost impossible without a mechanism to organize everything in smaller components, writing specialized and isolated pieces of code that can be reused as necessary in a controlled way. This is the job of modules.

CommonJS modules

A handful of module formats have emerged over the years, the most popular of which is CommonJS. It’s the default module format in Node.js, and can be used in client-side code with the help of module bundlers, which we’ll talk about shortly.

It makes use of a module object to export functionality from a JavaScript file and a require() function to import that functionality where you need it.

// lib/math.js
function sum(x, y) {
  return x + y;
}

const pi = 3.141593

module.exports = {
  sum: sum,
  pi: pi
};


// app.js
const math = require("lib/math");

console.log("2π = " + math.sum(math.pi, math.pi));

ES2015 modules

ES2015 introduces a way to define and consume components right into the language, which was previously possible only with third-party libraries. You can have separate files with the functionality you want, and export just certain parts to make them available to your application.

Note: native browser support for ES2015 modules is still under development, so you currently need some additional tools to be able to use them.

Here’s an example:

// lib/math.js

export function sum(x, y) {
  return x + y;
}
export const pi = 3.141593;

Here we have a module that exports a function and a variable. We can include that file in another one and use those exported functions:

// app.js

import * as math from "lib/math";

console.log("2π = " + math.sum(math.pi, math.pi));

Or we can also be specific and import only what we need:

// otherApp.js

import {sum, pi} from "lib/math";

console.log("2π = " + sum(pi, pi));

These examples have been extracted from the Babel website. For an in-depth look, check out Understanding ES6 Modules.

Package Management

Other languages have long had their own package repositories and managers to make it easier to find and install third-party libraries and components. Node.js comes with its own package manager and repository, npm. Although there are other package managers available, npm has become the de facto JavaScript package manager and is said to be the largest package registry in the world.

In the npm repository you can find third-party modules that you can easily download and use in your projects with a single npm install <package> command. The packages are downloaded into a local node_modules directory, which contains all the packages and their dependencies.

The packages that you download can be registered as dependencies of your project in a package.json file, along with information about your project or module (which can itself be published as a package on npm).

You can define separate dependencies for both development and production. While the production dependencies are needed for the package to work, the development dependencies are only necessary for the developers of the package.

Example package.json file

{
  "name": "demo",
  "version": "1.0.0",
  "description": "Demo package.json",
  "main": "main.js",
  "dependencies": {
    "mkdirp": "^0.5.1",
    "underscore": "^1.8.3"
  },
  "devDependencies": {},
  "scripts": {
    "test": "echo \"Error: no test specified\" && exit 1"
  },
  "author": "Sitepoint",
  "license": "ISC"
}

Build Tools

The code that we write when developing modern JavaScript web applications almost never is the same code that will go to production. We write code in a modern version of JavaScript that may not be supported by the browser, we make heavy use of third-party packages that are in a node_modules folder along with their own dependencies, we can have processes like static analysis tools or minifiers, etc. Build tooling exists to help transform all this into something that can be deployed efficiently and that’s understood by most web browsers.

Module bundling

When writing clean, reusable code with ES2015/CommonJS modules, we need some way to load these modules (at least until browsers support ES2015 module loading natively). Including a bunch of script tags in your HTML isn’t really a viable option, as it would quickly become unwieldy for any serious application, and all those separate HTTP requests would hurt performance.

We can include all the modules where we need them using the import statement from ES2015 (or require, for CommonJS) and use a module bundler to combine everything together into one or more files (bundles). It’s this bundled file that we’re going to upload to our server and include in our HTML. It will include all your imported modules and their necessary dependencies.

There are currently a couple of popular options for this, the most popular ones being Webpack, Browserify and Rollup.js. You can choose one or another depending on your needs.


If you want to learn more about module bundling and how it fits into the bigger picture of app development, I recommend reading Understanding JavaScript Modules: Bundling & Transpiling.


Transpilation

While support for modern JavaScript is pretty good among newer browsers, your target audience may include legacy browsers and devices with partial or no support.

In order to make our modern JavaScript work, we need to translate the code we write to its equivalent in an earlier version (usually ES5). The standard tool for this task is Babel — a compiler that translates your code into compatible code for most browsers. In this way, you don’t have to wait for vendors to implement everything; you can just use all the modern JS features.

There are a couple of features that need more than a syntax translation. Babel includes a Polyfill that emulates some of the machinery required for some complex features such as promises.

Build systems & task runners

Module bundling and transpilation are just two of the build processes that we may need in our projects. Others include code minification (to reduce file sizes), tools for analysis, and perhaps tasks that don’t have anything to do with JavaScript, like image optimization or CSS/HTML pre-processing.

The management of tasks can become a laborious thing to do, and we need a way to handle it in an automated way, being able to execute everything with simpler commands. The two most popular tools for this are Grunt.js and Gulp.js, which provide a way to organize your tasks into groups in an ordered way.

For example, you can have a command like gulp build which may run a code linter, the transpilation process with Babel, and module bundling with Browserify. Instead of having to remember three commands and their associated arguments in order, we just execute one that will handle the whole process automatically.

Wherever you find yourself manually organizing processing steps for your project, think if it can be automatized with a task runner.


Further reading: An Introduction to Gulp.js.


Application Architecture

Web applications have different requirements from websites. For example, while page reloads may be acceptable for a blog, that’s certainly not the case for an application like Google Docs. Your application should behave as closely as possible to a desktop one. Otherwise, the usability will be compromised.

Old-style web applications were usually done by sending multiple pages from a web server, and when a lot of dynamism was needed, content was loaded via Ajax by replacing chunks of HTML according to user actions. Although it was a big step forward to a more dynamic web, it certainly had its complications. Sending HTML fragments or even whole pages on each user action represented a waste of resources — especially of time, from the user’s perspective. The usability still didn’t match the responsiveness of desktop applications.

Looking to improve things, we created two new methods to build web applications — from the way we present them to the user, to the way we communicate between the client and the server. Although the amount of JavaScript required for an application also increased drastically, the result is now applications that behave very closely to native ones, without page reloading or extensive waiting periods each time we click a button.

Single Page Applications (SPAs)

The most common, high-level architecture for web applications is called SPA, which stands for Single Page Application. SPAs are big blobs of JavaScript that contain everything the application needs to work properly. The UI is rendered entirely client-side, so no reloading is required. The only thing that changes is the data inside the application, which is usually handled with a remote API via Ajax or another asynchronous method of communication.

One downside to this approach is that the application takes longer to load for the first time. Once it has been loaded, however, transitions between views (pages) are generally a lot quicker, since it’s only pure data being sent between client and server.

Universal / Isomorphic Applications

Although SPAs provide a great user experience, depending on your needs, they might not be the optimal solution — especially if you need quicker initial response times or optimal indexing by search engines.

There’s a fairly recent approach to solving these problems, called Isomorphic (or Universal) JavaScript applications. In this type of architecture, most of the code can be executed both on the server and the client. You can choose what you want to render on the server for a faster initial page load, and after that, the client takes over the rendering while the user is interacting with the app. Because pages are initially rendered on the server, search engines can index them properly.

Deployment

With modern JavaScript applications, the code you write is not the same as the code that you deploy for production: you only deploy the result of your build process. The workflow to accomplish this can vary depending on the size of your project, the number of developers working on it, and sometimes the tools/libraries you’re using.

For example, if you’re working alone on a simple project, each time you’re ready for deployment you can just run the build process and upload the resulting files to a web server. Keep in mind that you only need to upload the resulting files from the build process (transpilation, module bundling, minification, etc.), which can be just one .js file containing your entire application and dependencies.

You can have a directory structure like this:

├── dist
│   ├── app.js
│   └── index.html
├── node_modules
├── src
│   ├── lib
│   │   ├── login.js
│   │   └── user.js
│   ├── app.js
│   └── index.html
├── gulpfile.js
├── package.json
└── README

You thus have all of your application files in a src directory, written in ES2015+, importing packages installed with npm and your own modules from a lib directory.

Then you can run Gulp, which will execute the instructions from a gulpfile.js to build your project — bundling all modules into one file (including the ones installed with npm), transpiling ES2015+ to ES5, minifying the resulted file, etc. Then you can configure it to output the result in a convenient dist directory.

Note: if you have files that don’t need any processing, you can just copy them from src to the dist directory. You can configure a task for that in your build system.

Now you can just upload the files from the dist directory to a web server, without having to worry about the rest of the files, which are only useful for development.

Team development

If you’re working with other developers, it’s likely you’re also using a shared code repository, like GitHub, to store the project. In this case, you can run the build process right before making commits and store the result with the other files in the Git repository, to later be downloaded onto a production server.

However, storing built files in the repository is prone to errors if several developers are working together, and you might want to keep everything clean from build artifacts. Fortunately, there’s a better way to deal with that problem: you can put a service like Jenkins, Travis CI, CircleCI, etc. in the middle of the process, so it can automatically build your project after each commit is pushed to the repository. Developers only have to worry about pushing code changes without building the project first each time. The repository is also kept clean of automatically generated files, and at the end, you still have the built files available for deployment.

Conclusion

The transition from simple web pages to modern JavaScript applications can seem daunting if you’ve been away from web development in recent years, but I hope this article was useful as a starting point. I’ve linked to more in-depth articles on each topic where possible so you can explore further.

And remember that if at some point, after looking all the options available, everything seems overwhelming and messy, just keep in mind the KISS principle, and use only what you think you need and not everything you have available. At the end of the day, solving problems is what matters, not using the latest of everything.

What’s your experience of learning modern JavaScript development? Is there anything I haven’t touched on here that you’d like to see covered in future? I’d love to hear from you!

Frequently Asked Questions (FAQs) about Modern JavaScript Applications

What is the role of a package manager in a modern JavaScript application?

A package manager is a crucial tool in a modern JavaScript application. It helps manage the project’s dependencies, which are the libraries and frameworks that the project relies on to function. The package manager keeps track of the specific versions of these dependencies that your project uses, ensuring consistency and preventing potential conflicts. It also makes it easier to install, update, and uninstall these dependencies as needed. Examples of package managers for JavaScript include npm and Yarn.

How does module bundling work in JavaScript?

Module bundling is a technique used in JavaScript to group together various modules (files) into a single file, or a small number of files. This is done to reduce the number of HTTP requests that a browser needs to make when loading a web page, thereby improving performance. Tools like Webpack and Browserify are commonly used for this purpose. They also allow you to use features like ES6 modules in browsers that don’t support them natively.

What is transpiling in JavaScript and why is it important?

Transpiling is the process of converting code written in one language or language version into another. In the context of JavaScript, it often refers to converting ES6 (or later) code into ES5 code. This is important because not all browsers fully support ES6 features, so transpiling ensures that your code will run consistently across different browsers. Babel is a popular tool for transpiling JavaScript.

How does a task runner like Gulp or Grunt help in JavaScript development?

A task runner automates repetitive tasks in your development workflow, such as minification, compilation, unit testing, and linting. This not only saves time but also reduces the chances of human error. Gulp and Grunt are two popular task runners in JavaScript development. They are configured via a JavaScript file where you define tasks and the order in which they should run.

What is the purpose of a linter in JavaScript?

A linter is a tool that analyzes your code for potential errors and violations of coding standards and style guidelines. This helps maintain code quality and consistency, especially in large projects with multiple developers. ESLint is a widely used linter in JavaScript development. It’s highly configurable and can be tailored to your project’s specific needs.

How does a test runner work in JavaScript?

A test runner is a tool that automatically executes your test cases and generates a report of the results. It’s an essential part of a robust testing strategy. In JavaScript, test runners like Mocha and Jest are commonly used. They support a wide range of testing styles, including unit testing, integration testing, and end-to-end testing.

What is the role of a build tool in JavaScript?

A build tool automates the process of converting your source code into a production-ready, deployable application. This may involve tasks like transpiling, minification, and bundling. Build tools can also automate other aspects of your workflow, such as running tests and deploying your application. In JavaScript, Webpack is a popular build tool, though others like Parcel and Rollup are also used.

How does a JavaScript framework like React or Vue.js fit into a modern JavaScript application?

A JavaScript framework provides a structured way of building complex, interactive web applications. It provides pre-written code for common tasks, allowing developers to focus on the unique aspects of their application. React and Vue.js are two popular JavaScript frameworks. They both use a component-based architecture, which promotes reusability and maintainability.

What is the importance of a version control system in JavaScript development?

A version control system tracks changes to your code over time, allowing you to easily revert to a previous version if needed. It also facilitates collaboration by enabling multiple developers to work on the same codebase without overwriting each other’s changes. Git is the most widely used version control system in JavaScript development.

How does a JavaScript application interact with a backend server?

A JavaScript application interacts with a backend server primarily through HTTP requests. The server processes these requests and responds with data, typically in JSON format. This data can then be used to update the application’s state and UI. JavaScript provides several ways to make HTTP requests, including the Fetch API and libraries like Axios.

James KolceJames Kolce
View Author

Independent Web Developer, Designer and Technical Writer.

build toolsES2015es6ES6 Moduleslearn-modernjsmodernjsmodernjs-hubnilsonj
Share this article
Read Next
Get the freshest news and resources for developers, designers and digital creators in your inbox each week
Loading form