Since when I started my career as a web designer until now, as a front-end developer, CSS is part of my professional identity. And I can’t imagine my day to day development without thinking about styling with CSS. But the question is: how do I usually create my CSS these days?
The way we think about styling and using CSS is changing and some discussions about the death of traditional CSS have appeared. So, in this article, I'll show the evolution of CSS since its creation and give you some personal reflections on the current state of that language.
Well, to understand the current state of CSS, let's quickly go back to the old days of web development and understand why CSS, as a native language, was created and analize its evolution to this day.
- 1994: Among some style language proposals, the first draft of the CSS (Cascading Style Sheets) was released by Håkon Wium Lie as a solution to style web documents - the same year that Tim Berners-Lee founded the World Wide Web Consortium (W3C) and HTML has established itself as a universal document format.
- 1996: CSS level 1 finally emerged as a W3C Recommendation in the end of 1996 and Microsoft Internet Explorer 3 became the first commercial browser to support it. At that time, CSS1 allowed web designers to set just basic properties, such as fonts, colors and other minimal features.
- 1998: W3C published CSS level 2 which added support for positioning properties and table content. It also included support for media types, web fonts (the polemic ability to embed fonts in a web document) and some other properties related to user interface.
- 1999: W3C started working on CSS Level 3 - the version we are now familiar with. The point here is also the introduction of modules to release independent features instead of new versions of the language.
- 2011: CSS 2.1 was released to correct a few errors in the CSS2 specification: some parts have been altered, and some parts removed.
- 2011+: CSS level 3 modules were released and added many new features that allowed developers to style HTML elements with less CSS code, such as: box shadow, text shadow, new ways to describe colors (RGBA, for example), gradients, opacity, animations, transforms, media queries and more. CSS3, the latest version of CSS, is totally compatible with former CSS versions and continues focusing on developing new independent features through its modules.
“There has never been a CSS4. There will never be a CSS4. CSS4 is not a thing that exists.” - Tab Atkins, 2012 - member of the CSS working group at W3C. He also works for Google on the Google Chrome Team.
While for a long time people referred to CSS new features as CSS 3, now we no longer need to do it. It’s just CSS as Rachel Andrew explained in her article.
But, if a new version of CSS existed, should we put the following features under the CSS4 flag? Grid, Variables, Flexbox…
CSS solutions for specific problems
As part of CSS evolution, many tools have been created to solve some specific CSS problems:
- CSS Resets: Eric Meyer's solution was one of the most famous CSS Resets - a group of CSS declarations used to reduce the differences between the browser’s default styles: heights, margins and font sizes of heading for example.
- Pre-processors: CSS preprocessors are amazing tools that work with a pre-processing engine that implements nesting selectors, variables, mixins, extends and logic in stylesheets. However, as the CSS has been improved a lot, some of these features are no longer needed at the moment. And here are some of the most popular CSS preprocessors: Sass, LESS, Stylus and PostCSS.
- Component based thinking: a lot of concepts have emerged to help us to organize our styles in order to avoid CSS conflicts and improve the scalability and maintenance of them. Atomic Design System, BEM, SMACSS and OOCSS were some of those tools.
- Modular CSS and scope: By nature, CSS doesn't work with local scope and it’s one of the biggest problems of this language. Because of this, CSS Modules was created to make sure that the class names will be unique to avoid style conflicts.
“For three years, I have styled my web apps without any
Advantages of CSS-in-JS
There are many benefits to using CSS-in-JS and I really think it can help us a lot in our projects, because we won't have to deal with the painful application style maintenance in the same way that we used to do with preprocessors and with the manual definition of classes using BEM.
I think there are three key points to keep in mind when using this new approach to working with styles:
- Dead code elimination: CSS-in-JS safely removes redundant and unused CSS from our code, reducing the file size without side effects.
- CSS isolation: CSS-in-JS automates the local scope, generating unique class names by default when compiling CSS to avoid style conflicts - similar to what we used to do manually using BEM. It also abstracts the CSS model to the component level.
- State-based styling: due to modularity and isolation, properties can be easily shared between CSS and logic sides in order to add dynamic functionalities to your style rules. Let’s imagine we are building a component that needs to present a score. So, depending on the result, we can define different colors for positive and negative numbers just by observing the HTML content directly in the CSS declaration. For this, we used to create different classes to handle those different colors. But now it's much easier.
The most famous CSS-in-JS Libraries
To deal with CSS-in-JS there are several similar libraries, such as:
Among all these libraries one that has gained a lot of popularity is Styled Components due to some advantages:
- Automatic CSS Critical: the CSS for above-the-fold (the content before scrolling) is loaded first, in order to render the content to the user as fast as possible.
- Removing unused CSS: great to performance because it reduces the size file.
- Defined scope: it handles conflicts by not allowing different component classes to have the same name.
- Dynamic styles: we are able to change the styles according to the component's props and themes.
- Automatic vendor prefixing: it handles differences between browsers by adding the correct prefixes - which can be pre-configured according to your needs, BTW.
- Simplified and safe maintenance: since the scope is reduced, we get more confidence to maintain the code.
- And much more features.
As you could see, the traditional CSS architecture isn’t designed to be used with larger projects. And in a world of React, Vue, Angular and similar frameworks, that have been widely used, we need better solutions to solve some painful CSS problems.
CSS preprocessors have added a lot of power to stylesheets and things like CSS modular, nesting and variables have helped me in many different projects - BTW, I created this boilerplate using Gulp and Stylus in the past to develop my web applications. But now the world of front-end development is different and solutions based in React (like GatsbyJS, NextJS and React Native) are part of my current daily development routine and I need something more powerful.
And because of these changes in the way we think about styling and create CSS, several discussions have been emerging about the death of traditional CSS as you can see in this Samson Zhang article and in this other Chris Coyer article as well.
I showed a perspective on the current state of CSS, its default problems and some solutions that have been created over time. It's not my intention to convince you that CSS-in-JS is the best solution ever and that you should use it on all projects. But, in my perspective, if you're working with some component-based JS structures (like React, for example), I think this new way of writing CSS might be a good idea to deal with old and complex style problems.
I love CSS and for me writing CSS is an art. I fully agree that tradicional CSS will still be used depending on the project.
What about you? What do you think is happening with CSS? Have you ever tried CSS-in-JS?