What is the future of CSS?

Utility-first CSS is all the rage these days. Does anybody have feelings about what the next evolution of CSS (or web styling in general) will be?

What is on the horizon? Or, at least, what problems still need solving?

Comments

[Nick Taylor]

I think a lot of things that folks like about tooling are coming to CSS or are at least being discussed.

For example, in SASS, you can nest CSS rulesets.

.some-container {
  .some-child-component {
    // insert awesome CSS
  }
}

There is a working draft of this for CSS.

Scoped styles which some CSS tooling does and a lot of CSS in JS frameworks do has a working draft as well.

Aside from that, there are more pseudo selectors and classes. The :has() pseudo-class gives us something folks have wanted for a while. A parent selector.

una.css 👩🏻‍💻🇺🇦

@una

Literally me any time I remember that :has() is in Safari's latest stable release *and* behind a flag in Canary

:has() is REALLY GOOD Y'ALL

Need to make time to explore this one more and make demos!

bram.us/2021/12/21/the…

14:33 PM - 24 Mar 2022

Just like the browser APIs adopted jQuery APIs or were inspired by them, the same appears to be happening for CSS.

All these improvements mean that you can do a lot more in CSS without relying on JS for a lot of things you had to in the past.

[𒎏Wii 🏳️‍⚧️]

That working draft on CSS scoping is very outdated. Afaict it was very much dead for the longest time, and only recently got revived in the working draft of CSS Cascading and Inheritance Level 6

[Thomas Hansen]

Nested selectors is about time. I'm not sure about :has() since it makes the rendering engines more complex algorithmically, possibly degrading performance, or ...?

[𒎏Wii 🏳️‍⚧️]

At the very least, I'd expect :has() to have no performance impact where it isn't use. If that can be achieved, and it just turns into one of these features that shouldn't be overused, but won't instantly kill your whole website, then I wouldn't mind having it as a feature tbh.

[Nicholas Stimpson]

This is the fundamental problem. :has() does have a performance impact even where it isn't used. When a DOM element is mutated, the cascade needs to be recomputed. Without :has() implemented in browsers, only the cascade of the descendant elements of the parent of the mutated element needs to be recomputed. With :has() implemented, it was believed that the cascade of every element in the DOM would need recomputing, even if :has() wasn't used in any selector. The work done recently to implement :has() was essentially about finding a way to short cut that process to minimise the amount of the cascade that needs to be recomputed. But it can't be entirely eliminated.

[Thomas Hansen]

Similar ideas is the reason why Hyperlambda is so blistering fast, because parsing only implies "forward operations", and nothing during parsing can change anything backwards in the file, which of course allows the parser to exclusively look forward ignoring everything it's already seen. Thank you for re-iterating that fact ...

[𒎏Wii 🏳️‍⚧️]

It's super easy:

• list all your selectors with :has() somewhere in them
• keep track of all elements that match the rest of the selector
• follow the selector up to whatever ancestor (or descendant) the :has() applies to
• keep track of these elements and only re-compute styles for changes inside them
• spend another three to four decades fixing the countless bugs in your implementation, and the bugs you create by fixing those bugs, and the bugs resulting from optimisations of your previous bugfixes...

/s

[Temani Afif]

utility-first it not an evolution. It's only a trending as it doesn't add anything new to CSS.

The future of CSS is:

• The houdini project where you and me can extend the CSS language with new feature (houdini.how/about/)
• Cascade Layers that bring another level of flexibility related to CSS selectors: css-tricks.com/css-cascade-layers/
• CSS variables, calc(), clamp(), etc. They are quite old but still not used a lot

And many more

[Nicholas Stimpson]

One thing I read recently that amazed me:

Set the left margin to be 1/10 of the widow[sic] width, but always more than 2em and less than 20em:

html : margin.left = 2em < WIDTH/10 < 20em

From: w3.org/Style/CSS/draft1.html#arith...

This is a proto- clamp()! In the very first draft of CSS in May 1995. Before declaration blocks had even been invented, or CSS syntax had been established.

[𒎏Wii 🏳️‍⚧️]

calc() combined with var() and @property can be used to build amazing things without any need for pre-processors.

[Timothy Foster]

Sometimes I feel like CSS is basically the Assembly language of styling, writing lots of code to achieve small "ideas" such as overlap, gridding, theming, and whatnot. Flexbox, Grid, etc I think have gotten us closer to this more-or-less declarative approach, where our code better reflects what the design really wants rather than just serve as a means to achieve that design.

[𒎏Wii 🏳️‍⚧️]

Old-school CSS was very much about compiling esoteric idioms for simple design concepts into actual designs (float et al.), but these days the language has idioms for most of the common design patterns, so writing CSS is much more about actually describing the design and less about fighting the available paradigms.

[Timothy Foster]

Yep, exactly! Another example I thought of is aspect-ratio which is a great idiom for ideas like "I want a square", and superior to 100% bottom padding jank.

I do still find myself occasionally wishing for a better idiom (such as for overlap where you have to choose between absolute positioning, negative margins, or things sharing a grid space), but it's way better than it was a decade ago, and why I think it'll keep getting better.

[𒎏Wii 🏳️‍⚧️]

Yes, a content-overlap property would be pretty nice.

[FJones]

I think it's worth noting that CSS was, initially, effectively a language for print-style layouting. No one really intended for it to work with viewport-covering areas or boxed layouts in the sense we're used to now. We abused the living hell out of it by exploiting the underlying rendering algorithms, e.g. using float to create column layouts, or using table layouts. The rest was what HTML and framesets were meant to take care of.

[leob]

Good observation, sometimes CSS feels incredibly low level and a lot of non-intuitive hocus pocus is needed to achieve the simplest things ... a lot of that is due to its origins as a page/text styling language, it was never designed with application development in mind ... but (as you already noticed) recent developments like Flexbox and CSS Grid have indeed tried to improve on this situation!

[Daniel Schulz]

CSS is getting really good at describing reactive, context-aware layouts. New stuff like :is(), :where() and :has(), as well as cascade layers solve scoping across components. Container Queries solve compenent-scope layouting. Auto-fill and masonry grids provide layouting algorithms. There's a spec for full-on if/else. We have variables. Is described some of that in my post about it: dev.to/iamschulz/writing-logic-in-...

I hope we can overcome the current trend of utility-first css. That approach really limits the possibilities of those new features.
Being a pessimist though, I see more tailwind and CSS-inJS.

[Alain D'Ettorre]

• SASS will be fully integrated into CSS (nesting, more functions etc)
• There will be an easier way to style web components
• Maybe people will re-discover that naming classes the right way then style them is better than drowning HTML and SASS with utility-only classes, who knows

[Ben Sinclair]

More and more terrible things will become "all the rage".
The web will get incrementally worse until it's only understood by AI parsers.
All hope will be lost.

[Liviu Lupei]

Junior Developer mentioning Tailwind CSS in 3, 2, 1 ...

[Racheal Walker]

Hey, I feel this is the future of CSS
Web developers who know CSS are in high demand for businesses. In fact, according to the Bureau of Labor Statistics, web development is a profession expected to grow much faster than the average occupation over the next decade.

[Sherry Day]

It's still hard to test styles in an effective way. I'd like that to be an evolution.

[spO0q 🐒]

dunno if it's the horizon, maybe a part of it at least, but I look forward to using Container Queries. Context and container that react to changes could have a significant impact on component-based projects.

Still early stage, though.