Developer News

Focus trapping is a term that refers to managing focus within an element, such that focus always stays within it:

• If a user tries to tab out from the last element, we return focus to the first one.
• If the user tries to Shift + Tab out of the first element, we return focus back to the last one.

This whole focus trap thing is used to create accessible modal dialogs since it’s a whole ‘nother trouble to inert everything else — but you don’t need it anymore if you’re building modals with the dialog API (assuming you do it right).

Anyway, back to focus trapping.

The whole process sounds simple in theory, but it can quite difficult to build in practice, mostly because of the numerous parts to you got to manage.

Simple and easy focus trapping with Splendid Labz

If you are not averse to using code built by others, you might want to consider this snippet with the code I’ve created in Splendid Labz.

The basic idea is:

1. We detect all focusable elements within an element.
2. We manage focus with a keydown event listener.

import { getFocusableElements, trapFocus } from '@splendidlabz/utils/dom' const dialog = document.querySelector('dialog') // Get all focusable content const focusables = getFocusableElements(node) // Traps focus within the dialog dialog.addEventListener('keydown', event => { trapFocus({ event, focusables }) })

The above code snippet makes focus trapping extremely easy.

But, since you’re reading this, I’m sure you wanna know the details that go within each of these functions. Perhaps you wanna build your own, or learn what’s going on. Either way, both are cool — so let’s dive into it.

Selecting all focusable elements

I did research when I wrote about this some time ago. It seems like you could only focus an a handful of elements:

• a
• button
• input
• textarea
• select
• details
• iframe
• embed
• object
• summary
• dialog
• audio[controls]
• video[controls]
• [contenteditable]
• [tabindex]

So, the first step in getFocusableElements is to search for all focusable elements within a container:

export function getFocusableElements(container = document.body ) { return { get all () { const elements = Array.from( container.querySelectorAll( `a, button, input, textarea, select, details, iframe, embed, object, summary, dialog, audio[controls], video[controls], [contenteditable], [tabindex] `, ), ) } } }

Next, we want to filter away elements that are disabled, hidden or set with display: none, since they cannot be focused on. We can do this with a simple filter function.

export function getFocusableElements(container = document.body ) { return { get all () { // ... return elements.filter(el => { if (el.hasAttribute('disabled')) return false if (el.hasAttribute('hidden')) return false if (window.getComputedStyle(el).display === 'none') return false return true }) } } }

Next, since we want to trap keyboard focus, it’s only natural to retrieve a list of keyboard-only focusable elements. We can do that easily too. We only need to remove all tabindex values that are less than 0.

export function getFocusableElements(container = document.body ) { return { get all () { /* ... */ }, get keyboardOnly() { return this.all.filter(el => el.tabIndex > -1) } } }

Now, remember that there are two things we need to do for focus trapping:

• If a user tries to tab out from the last element, we return focus to the first one.
• If the user tries to Shift + Tab out of the first element, we return focus back to the last one.

This means we need to be able to find the first focusable item and the last focusable item. Luckily, we can add first and last getters to retrieve these elements easily inside getFocusableElements.

In this case, since we’re dealing with keyboard elements, we can grab the first and last items from keyboardOnly:

export function getFocusableElements(container = document.body ) { return { // ... get first() { return this.keyboardOnly[0] }, get last() { return this.keyboardOnly[0] }, } }

We have everything we need — next is to implement the focus trapping functionality.

How to trap focus

First, we need to detect a keyboard event. We can do this easily with addEventListener:

const container = document.querySelector('.some-element') container.addEventListener('keydown', event => {/* ... */})

We need to check if the user is:

• Pressing tab (without Shift)
• Pressing tab (with Shift)

Splendid Labz has convenient functions to detect these as well:

import { isTab, isShiftTab } from '@splendidlabz/utils/dom' // ... container.addEventListener('keydown', event => { if (isTab(event)) // Handle Tab if (isShiftTab(event)) // Handle Shift Tab /* ... */ })

Of course, in the spirit of learning, let’s figure out how to write the code from scratch:

• You can use event.key to detect whether the Tab key is being pressed.
• You can use event.shiftKey to detect if the Shift key is being pressed

Combine these two, you will be able to write your own isTab and isShiftTab functions:

export function isTab(event) { return !event.shiftKey && event.key === 'Tab' } export function isShiftTab(event) { return event.shiftKey && event.key === 'Tab' }

Since we’re only handling the Tab key, we can use an early return statement to skip the handling of other keys.

container.addEventListener('keydown', event => { if (event.key !== 'Tab') return if (isTab(event)) // Handle Tab if (isShiftTab(event)) // Handle Shift Tab /* ... */ })

We have almost everything we need now. The only thing is to know where the current focused element is at — so we can decide whether to trap focus or allow the default focus action to proceed.

We can do this with document.activeElement.

Going back to the steps:

• Shift focus if user Tab on the last item
• Shift focus if the user Shift + Tab on the first item

Naturally, you can tell that we need to check whether document.activeElement is the first or last focusable item.

container.addEventListener('keydown', event => { // ... const focusables = getFocusableElements(container) const first = focusables.first const last = focusables.last if (document.activeElement === last && isTab(event)) { // Shift focus to the first item } if (document.activeElement === first && isShiftTab(event)) { // Shift focus to the last item } })

The final step is to use focus to bring focus to the item.

container.addEventListener('keydown', event => { // ... if (document.activeElement === last && isTab(event)) { first.focus() } if (document.activeElement === first && isShiftTab(event)) { last.focus() } })

That’s it! Pretty simple if you go through the sequence step-by-step, isn’t it?

Final callout to Splendid Labz

As I resolve myself to stop teaching (so much) and begin building applications, I find myself needing many common components, utilities, even styles.

Since I have the capability to build things for myself, (plus the fact that I’m super particular when it comes to good DX), I’ve decided to gather these things I find or build into a couple of easy-to-use libraries.

Just sharing these with you in hopes that they will help speed up your development workflow.

Thanks for reading my shameless plug. All the best for whatever you decide to code!

A Primer on Focus Trapping originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

From making apps, browsing the Web, to creating files, today's AI agents today can take on an increasing number of computing tasks on their own. But the software underlying these capabilities, wasn't made for agents. It was designed and built for people to use. As such there's an opportunity, and perhaps an increasing need, to rethink these systems for agent use.

When building agent-based AI applications, you'll likely butt up against a number of situations where existing software isn't optimized for what thinking machines can do. For instance, Web search. Nearly every agent-based AI application makes use of information on the Web to get things done. But Web Search APIs weren't written with agents in mind.

They provide a limited number of search results and a condensed snippet format that lines up more with how people use Web search interfaces. We get a page of ten blue links and scan them to decide which one to click. But AI agents aren't people. Not only can they make sense of many more search results at once, but their performance usually improves with larger document summaries and contents. People on the other hand, are unlikely to read through all search results before making a decision. So search APIs could certainly be rethought for agents.

Similarly, when agents are developing applications or collecting data, they can make use of databases. But once again databases were designed and built for people to use not AI agents. And once again they can be rethought for agents, which is what we did with our most recent launch: AgentDB.

Agents can (and do) produce 1000x more databases than people every day, so the process of spinning up and managing any database for an agent needs to be as easy and maintenance-free as possible. Most of the databases AI agents create will be short-lived after serving their initial purpose. But some databases will be used again and others still will be used regularly.

With this kind of volume costs can become an issue, so keeping that many databases available needs to be as cost effective as possible. Last but not least, the content of databases needs to work well as context for AI models so agents can use this data as part of their tasks.

AgentDB is a database system designed around these considerations. With AgentDB, creating a database only requires a Universally Unique Identifier (UUID). There's no setup or configuration step. So whenever an AI agent decides it needs a database, it has one simply by creating a UUID. No forms or set-up wizards involved.

Databases in AgentDB are stored as files not hosted services requiring compute and maintenance. If an AI agent needs to query a database or append to it, it can. But if it never needs to access it again, the database is just a file. That means you're only paying for the cost of storage to keep it around and because AgentDB databases are just files, they scale. Meaning they can easily keep up with the scale of AI agents.

To make data within each AgentDB database easily accessible as context for AI models, every AgentDB account is also an MCP server. This makes the data portable across AI applications as long as they support MCP server connections (which most do).

Altogether this example illustrates how even the most fundamental software infrastructure systems, like databases, can be rethought for the age of AI. The AgentDB database system doesn't look like a hosted database as a service solution because it's not designed and built for database admins and back-end developers. It's built for today's thinking machines.

And as agents take on more computing tasks for people, it won't be the only software made with agents as first class users.

From making apps, browsing the Web, to creating files, today's AI agents today can take on an increasing number of computing tasks on their own. But the software underlying these capabilities, wasn't made for agents. It was designed and built for people to use. As such there's an opportunity, and perhaps an increasing need, to rethink these systems for agent use.

When building agent-based AI applications, you'll likely butt up against a number of situations where existing software isn't optimized for what thinking machines can do. For instance, Web search. Nearly every agent-based AI application makes use of information on the Web to get things done. But Web Search APIs weren't written with agents in mind.

They provide a limited number of search results and a condensed snippet format that lines up more with how people use Web search interfaces. We get a page of ten blue links and scan them to decide which one to click. But AI agents aren't people. Not only can they make sense of many more search results at once, but their performance usually improves with larger document summaries and contents. People on the other hand, are unlikely to read through all search results before making a decision. So search APIs could certainly be rethought for agents.

Similarly, when agents are developing applications or collecting data, they can make use of databases. But once again databases were designed and built for people to use not AI agents. And once again they can be rethought for agents, which is what we did with our most recent launch: AgentDB.

Agents can (and do) produce 1000x more databases than people every day, so the process of spinning up and managing any database for an agent needs to be as easy and maintenance-free as possible. Most of the databases AI agents create will be short-lived after serving their initial purpose. But some databases will be used again and others still will be used regularly.

With this kind of volume costs can become an issue, so keeping that many databases available needs to be as cost effective as possible. Last but not least, the content of databases needs to work well as context for AI models so agents can use this data as part of their tasks.

AgentDB is a database system designed around these considerations. With AgentDB, creating a database only requires a Universally Unique Identifier (UUID). There's no setup or configuration step. So whenever an AI agent decides it needs a database, it has one simply by creating a UUID. No forms or set-up wizards involved.

Databases in AgentDB are stored as files not hosted services requiring compute and maintenance. If an AI agent needs to query a database or append to it, it can. But if it never needs to access it again, the database is just a file. That means you're only paying for the cost of storage to keep it around and because AgentDB databases are just files, they scale. Meaning they can easily keep up with the scale of AI agents.

To make data within each AgentDB database easily accessible as context for AI models, every AgentDB account is also an MCP server. This makes the data portable across AI applications as long as they support MCP server connections (which most do).

Altogether this example illustrates how even the most fundamental software infrastructure systems, like databases, can be rethought for the age of AI. The AgentDB database system doesn't look like a hosted database as a service solution because it's not designed and built for database admins and back-end developers. It's built for today's thinking machines.

And as agents take on more computing tasks for people, it won't be the only software made with agents as first class users.

A while back, our man Geoff Graham treated us to a refresher on the CSS initial-letter property, but how can you style drop and initial caps to reflect a brand’s visual identity and help to tell its stories?

Here’s how I do it in CSS by combining ::first-letter and initial-letter with other unexpected properties, including border-image, and clip-path.

Patty Meltt is an up-and-coming country music sensation.

My brief: Patty Meltt is an up-and-coming country music sensation, and she needed a website to launch her new album. She wanted it to be distinctive-looking and memorable, so she called Stuff & Nonsense. Patty’s not real, but the challenges of designing and developing sites like hers are.

First, a drop cap recap. Chris Coyier wrote about drop caps several years ago. They are a decorative letter at the beginning of a paragraph, often spanning several lines of text. It’s a typographic flourish found in illuminated manuscripts and traditional book design, where it adds visual interest and helps guide a reader’s eye to where they should begin.

Study manuscripts from the Middle Ages onwards, and you’ll find hand-decorated illuminated capitals. The artists who made these initial letters were fabulously called “illuminators.” These medieval versals went beyond showing someone where to start reading; historiated letters also illustrated the stories, which was especially useful since most people in the Middle Ages couldn’t read.

A basic drop cap

On the web, drop caps can improve readability and reflect a brand’s visual identity.

A brief refresher on properties and values

In CSS, drop caps are created using the ::first-letter pseudo-element in combination with initial-letter. As you might expect, ::first-letter targets the very first letter of a block of text, enabling you to style it independently from the rest of a paragraph. The first number sets how many lines tall the letter appears, and the second controls its baseline alignment — that is, which line of text the bottom of the cap sits on.

p::first-letter { -webkit-initial-letter: 3 3; initial-letter: 3 3; }

Because browser support still varies, it’s common to include both the unprefixed and -webkit- prefixed properties for maximum compatibility. And speaking of browser support, it’s also sensible to wrap the initial-letter property inside an @supports CSS at-rule so we can check for browser support and provide a fallback, if needed:

@supports (initial-letter:2) or (-webkit-initial-letter:2) { p::first-letter { -webkit-initial-letter: 3 3; initial-letter: 3 3; } }

The initial-letter property automatically calculates the font size to match the number of lines a drop cap spans. On its own, this can make for quite a first impression. However, drop caps really start to come to life when you combine initial-letter with other CSS properties.

Tip: Interactive examples from this article are available in my lab.

Shadows Text shadows applied to first letters (live demo)

When I want to lift a drop cap off the page, I can add a single text-shadow. Shadows can be colourful and don’t have to be black. I created a full live demo you can check out.

p::first-letter { /* ... *// text-shadow: 6px 6px 0 #e6d5b3; }

But why use just one shadow when two hard-edged shadows will turn a cap into a classic graphic typographic element?

p::first-letter { /* ... */ text-shadow: -6px -6px 0 #7d6975, 6px 6px 0 #e6d5b3; } Examples showing unstyled, single text shadow, and two text shadows (live demo) Strokes A text shadow applied to a first letter (live demo)

The text-stroke property — shorthand for text-stroke-width and text-stroke-color — adds an outline to the centre of the text shape. It’s a Baseline feature and is now widely available. I can make the cap text transparent or colour it to match the page background.

p::first-letter { /* ... */ text-stroke: 5px #e6d5b3; } Backgrounds Solid and gradient backgrounds applied to first letters (live demo)

Adding a background is a simple way to start making a cap more decorative. I could start by adding a solid background-color.

p::first-letter { /* ... */ background-color: #97838f; }

To add a lighting effect, I could apply a conical, linear, or radial gradient background image (here’s a demo):

p::first-letter { /* ... */ background-color: #e6d5b3; background-image: linear-gradient(135deg,#c8b9c2 0%, #7d6975 50%); }

And even an image URL to use a bitmap or vector image as a background (and here’s that demo):

p::first-letter { /* ... */ background-color: #e6d5b3; background-image: url(...); background-size: cover; } Background images and a background clipped to text

Things become even more interesting by clipping a bitmap, gradient, or vector background image to the text while setting its colour to transparent. Now, the image will only appear inside the text space (demo).

p::first-letter { /* ... */ background-clip: text; color: transparent; } Borders Two examples of borders applied to first letters, one square and one rounded

You might think borders are boring, but there’s plenty you can do to make them look interesting. I could start by applying a solid border to surround the cap box (demo).

p::first-letter { /* ... */ border: 5px solid #e6d5b3; }

Then, I could apply border-radius to slightly round all its corners (demo).

p::first-letter { /* ... */ border-radius: 1rem; }

Or, I might round individual corners for a more interesting look (demo):

p::first-letter { /* ... */ border-top-left-radius: 3rem; border-bottom-right-radius: 3rem; } A border radius applied to the first letter, where the top-left and bottom-right edges are rounded (live demo)

And then there’s the border-image property, a powerful, yet often overlooked CSS tool. By slicing, repeating, and outsetting images, you can create intricate borders and decorative drop caps with minimal code.

A CSS border image applied to a first letter (live demo)

You can insert a bitmap or vector format image, or drop a CSS gradient into the border space:

p::first-letter { /* ... */ border-style: solid; border-width: 10px; border-image: conic-gradient(...) 1; } Clipping Clipping first letters

The clip-path property lets you define a custom shape that controls which parts of an element are visible and which are hidden. Instead of always showing a rectangular box, you can use clip-path to crop elements into circles, polygons, or even complex shapes defined with SVG paths. It’s an effective way to create visual effects like this right-facing arrow. Clipping the drop cap into an arrow shape isn’t just decorative — it reinforces direction and hierarchy, literally pointing readers to where the story begins. Here’s a demo of the following example.

p::first-letter { /* ... */ padding-inline: 1rem 2rem; background-color: #e6d5b3; clip-path: polygon(...); }

Or a glossy sticker shape cap, made by combining clip-path with a gradient background image and a text shadow (demo).

Transforms Two examples of transforming first letters, one rotated (demo) and one scaled (demo)

You can transform a drop cap independently from the rest of a paragraph by rotating, scaling, skewing, or translating it to make it feel more dynamic:

p::first-letter { /* ... */ margin-inline-end: 2.5em; transform: skew(20deg, 0deg); }

And with a little trial and error to arrive at the correct values, you could even flow the remaining paragraph text around the cap using the shape-outside property (demo):

p::first-letter { /* ... */ display: block; float: left; shape-outside: polygon(0 0, 0 200px, 250px 600px); shape-margin: 50px; transform: skew(20deg, 0deg) translateX(-60px); }

Drop caps don’t just help guide a reader’s eye to where they should begin; they also set the tone for what follows. A well-designed drop cap adds visual interest at the start of a block of text, drawing attention in a way that feels intentional and designed. Because it’s often the first element the reader sees, caps can carry a lot of visual weight, making them powerful tools for expressing a brand’s identity.

Designing for Patty Meltt

Patty Meltt wanted a website packed with design details. Every element added to a design is an opportunity to be expressive, and that includes her drop caps.

Her biography page is presentable, but we felt a focus on where someone should start reading was lacking.

Patty Meltt’s biography without a drop cap

From the selection of designs I showed her, she felt the sticker-style cap best suited her brand.

To implement it, first, I added a cursive typeface which matches her branding and contrasts with the rest of her typographic design:

p::first-letter { font-family: "Lobster Two", sans-serif; font-weight: 700; }

I changed the cap colour to match the page background and added a semi-transparent text shadow:

p::first-letter { /* ... */ color: #140F0A; text-shadow: 6px 6px 0 rgba(163,148, 117, .8); }

Next, I clipped the cap box to a visible area shaped like a sticker:

p::first-letter { /* ... */ clip-path: polygon(...); }

…before applying two background images — a noise-filled SVG and a radial gradient — that I blended using a background-blend-mode:

p::first-letter { /* ... */ background-image: url(img/cap-noise.svg), radial-gradient(circle, #e6d5b3 0%, #cdaa65 100%); background-blend-mode: soft-light, normal; } Patty Meltt’s biography with a stylsh new drop cap (demo)

The result is a drop cap that’s as stylish as cut-off jeans and a pair of gator-skinned boots.

Conclusion

Styling drop caps isn’t just about decoration — it’s about setting a tone, drawing readers in, and using every detail to express a brand’s voice. CSS has the tools to go beyond the default: gradients, textures, borders, and even complex shapes all help transform first letters into statements. So don’t waste the opportunities that drop caps give you. Make ’em sing.

Getting Creative With Versal Letters originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

Folks have a lot to say about “liquid glass,” the design aesthetic that Apple introduced at WWDC 2025. Some love it, some hate it, and others jumped straight into seeing how to they could create it in CSS.

There’s a lot to love, hate, and experience with liquid glass. You can love the way content reflects against backgrounds. You can hate the poor contrast between foreground and background. And you can be eager to work with it. All of those can be true at the same time.

Image credit: Apple

I, for one, am generally neutral with things like this for that exact reason. I’m intrigued by liquid glass, but hold some concern about legibility, particularly as someone who already struggles with the legibility of Apple’s existing design system (notably in Control Center). And I love looking at the many and clever ways that devs have tried to replicate liquid glass in their own experiments.

So, I’m in the process of gathering notes on the topic as I wrap my head around this “new” (or not-so-new, depending on who’s talking) thing and figure out where it fits in my own work. These links are a choice selection of posts that I’ve found helpful and definitely not meant to be an exhaustive list of what’s out there.

WWDC Introduction

Always a good idea to start with information straight from the horse’s mouth.

In short:

• It’s the first design system that is universally applied to all of Apple’s platforms, as opposed to a single platform like Apple’s last major overhaul, iOS 7.
• It’s designed to refract light and dynamically react to user interactions.
• By “dynamic” we’re referring to UI elements updating into others as the context changes, such as displaying additional controls. This sounds a lot like the Dynamic Island, supporting shape-shifting animations.
• There’s a focus on freeing up space by removing hard rectangular edges, allowing UI elements to become part of the content and respond to context.

Apple also released a more in-depth video aimed at introducing liquid glass to designers and developers.

In short:

• Liquid glass is an evolution of the “aqua” blue interface from macOS 10, the real-time introduced in iOS 7, the “fluidity” of iOS 10, the flexibility of the Dynamic Island, and the immersive interface of visionOS.
• It’s a “digital meta-material” that dynamically bends and shapes light while moving fluidly like water.
• It’s at least partially a response to hardware devices adopting deeper rounded corners.
• Lensing: Background elements are bended and warped rather than scattering light as it’s been in previous designs. There’s gel-like feel to elements.
• Translucence helps reveal what is underneath a control, such as a progress indicator you can scrub more precisely by seeing what is behind the surface.
• Controls are persistent between views for establishing a relationship between controls and states. This reminds me of the View Transition API.
• Elements automatically adapt to light and dark modes.
• Liquid glass is composed of layers: highlight (light casting and movement), shadow (added depth for separation between foreground and background), and illumination (the flexible properties of the material).
• It is not meant to be used everywhere but is most effective for the navigation layer. And avoid using glass on glass.
• There are two variants: regular (most versatile) and clear (does not have adaptive behaviors for allowing content to be more visible below the surface).
• Glass can be tinted different colors.

Documentation

Right on cue, Apple has already made a number of developer resources available for using and implementing liquid glass that are handy references.

• Introduction to Liquid Glass
• Adopting Liquid Glass
• Landmarks: Building an app with Liquid Glass
• Applying Liquid Glass to custom views

‘Beautiful’ and ‘Hard to Read’: Designers React to Apple’s Liquid Glass Update

This Wired piece is a nice general overview of what liquid glass is and context about how it was introduced at WWDC 2025. I like getting a take on this from a general tech perspective as opposed to, say, someone’s quick hot take. It’s a helpful pulse on what’s happening from a high level without a bunch of hyperbole, setting the stage for digging deeper into things.

In short:

• Apple is calling this “Liquid Glass.”
• It’s Apple’s first significant UI overhaul in 10 years.
• It will be implemented across all of Apple’s platforms, including iOS, macOS, iPadOS, and even the Vision Pro headset from which it was inspired.
• “From a technical perspective, it’s a very impressive effect. I applaud the time and effort it must have taken to mimic refraction and dispersion of light to such a high degree.”
• “Similar to the first beta for iOS 7, what we’ve seen so far is rough on the edges and potentially veers into distracting or challenging to read, especially for users with visual impairments.”

Accessibility

Let’s get right to the heart of where the pushback against liquid glass is coming from. While the aesthetic, purpose, and principles of liquid glass are broadly applauded, many are concerned about the legibility of content against a glass surface.

Traditionally, we fill backgrounds with solid or opaque solid color to establish contrast between the foreground and background, but with refracted light, color plays less a role and it’s possible that highlighting or dimming a light source will not produce enough contrast, particularly for those with low-vision. WCAG 2.2 emphasizes color and font size for improving contrast and does provide guidance for something that’s amorphous like liquid glass where bending the content below it is what establishes contrast.

“Apple’s “Liquid Glass” and What It Means for Accessibility”:

• “When you have translucent elements letting background colors bleed through, you’re creating variable contrast ratios that might work well over one background, but fail over a bright photo of the sunset.”
• “Apple turned the iPhone’s notch into the Dynamic Island, Android phones that don’t have notches started making fake notches, just so they could have a Dynamic Island too. That’s influence. But here they are making what looks like a purely aesthetic decision without addressing the accessibility implications.”
• “People with dyslexia, who already struggle with busy backgrounds and low-contrast text, now deal with an interface where visual noise is baked into the design language. People with attention disorders may have their focus messed up when they see multiple translucent layers creating a whole lot of visual noise.”
• “It’s like having a grand entrance and a side door marked ‘accessible.’ Technically compliant. But missing the point.”
• “The legal landscape adds another layer. There’s thousands of digital accessibility lawsuits filed in the U.S. yearly for violating the ADA, or the American Disabilities Act. Companies are paying millions in settlements. But this is Apple. They have millions. Plus all the resources in the world to save them from legal risks. But their influence means they’re setting precedents.”

“Liquid Glass: Apple vs accessibility”:

• “Yet even in Apple’s press release, linked earlier, there are multiple screenshots where key interface components are, at best, very difficult to read. That is the new foundational point for Apple design. And those screenshots will have been designed to show the best of things.”
• “Apple is still very often reactive rather than proactive regarding vision accessibility. Even today, there are major problems with the previous versions of its operating systems (one example being the vestibular trigger if you tap-hold the Focus button in Control Centre). One year on, they aren’t fixed.”
• “State, correctly, that Apple is a leader in accessibility. But stop assuming that just because this new design might be OK for you and because Apple has controls in place that might help people avoid the worst effects of design changes, everything is just peachy. Because it isn’t.”

“Liquid Glass” by Hardik Pandya

• “The effect is technically impressive, but it introduces a layer of visual processing between you and your memories. What was once immediate now feels mediated. What was once direct now feels filtered.”
• “While Apple’s rationale for Liquid Glass centers on ‘seeing’ content through a refractive surface, user interface controls are not meant to be seen—they are meant to be operated. When you tap a button, slide a slider, or toggle a switch, you are not observing these elements. You are manipulating them directly.”
• “Buttons become amorphous shapes. Sliders lose their mechanical clarity. Toggle switches abandon their physical affordances. They appear as abstract forms floating behind glass—beautiful perhaps, but disconnected from the fundamental purpose of interface controls: to invite and respond to direct manipulation.”
• “The most forward-thinking interface design today focuses on invisibility – making the interaction so seamless that the interface itself disappears. Liquid Glass makes the interface more visible, more present, and more demanding of attention.”

“Liquid glass, now with frosted tips”:

• It’s easy to dump on liquid glass in its introductory form, but it’s worth remembering that it’s in beta and that Apple is actively developing it ahead of its formal release.
• A lot has changed between the Beta 2 and Beta 3 releases. The opacity between glass and content has been bumped up in several key areas.

Tutorials, Generators, and Frameworks

It’s fun to see the difference approaches many folks have used to re-create the liquid glass effect in these early days. It amazes me that there is already a deluge of tutorials, generators, and even UI frameworks when we’re only a month past the WWDC 2025 introduction.

• Create this trendy blurry glass effect with CSS (Kevin Powell)
• Liquid Glass design using CSS (Nordcraft)
• Adopting Apple’s Liquid Glass: Examples and best practices (LogRocket)
• Liquid Glass Figma File
• CSS Liquid Glass Effects (DesignFast)
• Liquid Glass UI Framework
• Liquid Glass CSS Generator

Experiments

Let’s drop in a few interesting demos that folks have created. To be clear, glass-based interfaces are not new and have been plenty explored, which you can find over at CodePen in abundance. These are recent experiments. The most common approaches appear to reach for SVG filters and background blurs, though there are many programmatic demos as well.

Using a CSS-only approach with an SVG filter with backdrop-filter with a series of nested containers that sorta mimics how Apple describes glass as being composed of three layers (highlight, shadow and illumination):

CodePen Embed Fallback

Same sort of deal here, but in the context of a theme toggle switch that demonstrates how glass can be tinted:

CodePen Embed Fallback

Comparing a straight-up CSS blur with an SVG backdrop:

CodePen Embed Fallback

Contextual example of a slider component:

CodePen Embed Fallback

Using WebGL:

CodePen Embed Fallback Assorted links and coverage

A few more links from this browser tab group I have open:

• “Apple’s Liquid Glass is exactly as ambitious as Apple” (Fast Company)
• “Apple unveils iOS 26 with Liquid Glass” (9to5Mac)
• “Apple Announces All-New ‘Liquid Glass’ Software Redesign Across iOS 26 and More” (MacRumors)
• “Apple just added more frost to its Liquid Glass design” (The Verge)
• “Apple tones down Liquid Glass effect in iOS 26 beta 3” (The Apple Post)
• “More assorted notes on Liquid Glass” (Riccardo Mori)
• A bunch of CodePen Collections

Getting Clarity on Apple’s Liquid Glass originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

State of Devs 2025 survey results are out! While the survey isn’t directly related to the code part of what we do for work, I do love the focus Devographics took ever since its inception in 2020. And this year it brought us some rather interesting results through the attendance of 8,717 developers, lots of data, and even more useful insights that I think everyone can look up and learn from.

I decided to look at the survey results with an analytical mindset, but wound up pouring my heart out because, well, I am a developer, and the entire survey affects me in a way. I have some personal opinions, it turns out. So, sit back, relax, and indulge me for a bit as we look at a few choice pieces of the survey.

And it’s worth noting that this is only part one of the survey results. A second data dump will be published later and I’m interested to poke at those numbers, too.

An opportunity to connect

One thing I noticed from the Demographics section is how much tech connects us all. The majority of responses come from the U.S. (26%) but many other countries, including Italy, Germany, France, Estonia, Austria, South Africa and many more, account for the remaining 74%.

I mean, I am working and communicating with you right now, all the way from Nigeria! Isn’t that beautiful, to be able to communicate with people around the world through this wonderful place we call CSS-Tricks? And into the bigger community of developers that keeps it so fun?

I think this is a testament to how much we want to connect. More so, the State of Devs survey gives us an opportunity to express our pain points on issues surrounding our experiences, workplace environments, quality of health, and even what hobbies we have as developers. And while I say developers, the survey makes it clear it’s more than that. Behind anyone’s face is someone encountering life challenges. We’re all people and people are capable of pure emotion. We are all just human.

It’s also one of the reasons I decided to open a Bluesky account: to connect with more developers.

I think this survey offers insights into how much we care about ourselves in tech, and how eager we are to solve issues rarely talked about. And the fact that it’s global in nature illustrates how much in common we all have.

More women participated this year

From what I noticed, fewer women participated in the 2024 State of JavaScript and State of CSS fewer women (around 6%), while women represented a bigger share in this year’s State of Devs survey. I’d say 15% is still far too low to fairly “represent” an entire key segment of people, but it is certainly encouraging to see a greater slice in this particular survey. We need more women in this male-dominated industry.

Experience over talent

Contrary to popular opinion, personal performance does not usually equate to higher pay, and this is reflected in the results of this survey. It’s more like, the more experienced you are, the more you’re paid. But even that’s not the full story. If you’re new to the field, you still have to do some personal marketing, find and keep a mentor, and a whole bunch of stuff. Cassidy shares some nice insights on this in a video interview tracing her development career. You should check it out, especially if you’re just starting out.

Notice that the average income for those with 10-14 of experience ($115,833) is on par with those with between 15-29 years of experience ($118,000) and not far from those with 30+ years ($120,401). Experience appears to influence income, but perhaps not to the extent you would think, or else we’d see a wider gap between those with 15 years versus those with more than double the service time.

More than that, notice how income for the most experienced developers (30+ years) is larger on average but the range of how much they make is lower than than those with 10-29 years under their belts. I’m curious what causes that decline. Is it a lack of keeping up with what’s new? Is it ageism? I’m sure there are lots of explanations.

Salary, workplace, and job hunting

I prefer not drill into each and every report. I’m interested in very specific areas that are covered in the survey. And what I take away from the survey is bound to be different than your takeaways, despite numbers being what they are. So, here are a few highlights of what stood out to me personally as I combed through the results.

Your experience, employment status, and company’s employer count seem to directly affect pay. For example, full-timers report higher salaries than freelancers. I suppose that makes sense, but I doubt it provides the full picture because freelancers freelance for a number of reasons, whether its flexible hours, having more choice to choose their projects, or having personal constraints that limit how much they can work. In some ways, freelancers are able to command higher pay while working less.

Bad management and burnout seem to be the most talked-about issues in the workplace. Be on guard during interviews, look up reviews about the company you’re about to work for, and make sure there are far fewer complaints than accolades. Make sure you’re not being too worked up during work hours; breaks are essential for a boost in productivity.

Seventy percent of folks reported no discrimination in the workplace, which means we’re perhaps doing something right. That said, it’s still disheartening that 30% experience some form of discrimination and lowering that figure is something we ought to aim for. I’m hoping companies — particularly the tech giants in our space — take note of this and enforce laws and policies surrounding this. Still, we can always call out discriminatory behavior and make corrections where necessary. And who’s to say that everyone who answered the survey felt safe sharing that sort of thing? Silence can be the enemy of progress.

Never get too comfortable in your job. Although 69% report having never been laid off, I still think that job security is brittle in this space. Always learn, build, and if possible, try to look for other sources of income. Layoffs are still happening, and looking at the news, it’s likely to continue for the foreseeable future, with the U.S., Australia, and U.K. being leading the way.

One number that jumped off the page for me is that it takes an average of four applications for most developers to find a new job. This bamboozles me. I’m looking for a full-time role (yes, I’m available!), and I regularly apply for more than four jobs in a given day. Perhaps I’m doing something wrong, but that’s also not consistent with those in my social and professional circles. I know and see plenty of people who are working hard to find work, and the number of jobs they apply for has to bring that number up. Four applications seems way low, though I don’t have the quantitative proof for it.

Your personal network is still the best way to find a job. We will always and forever be social animals, and I think that’s why most survey participants say that coworker relationships are the greatest perk of a job. I find this to be true with my work here at CSS-Tricks. I get to collaborate with other like-minded CSS and front-end enthusiasts far and wide. I’ve developed close relationships with the editors and other writers, and that’s something I value more than any other benefits I could get somewhere else.

Compensation is still a top workplace challenge. JavaScript is still the king of programming (bias alert), taking the top spot as the most popular programming language. I know you’re interested, that CSS came in at third.

To my surprise, Bluesky is more popular amongst developers than X. I didn’t realize how much toxicity I’ve been exposed to at X until I opened a Bluesky account. I hate saying that the “engagement” is better, or some buzz-worthy thing like that, but I do experience more actual discussions over at Bluesky than I have for a long time at X. And many of you report the same. I hasten to say that Bluesky is a direct replacement for what X (let’s face it, Twitter) used to be, but it seems we at least have a better alternative.

Health issues

Without our health, we are nothing. Embrace your body for what it is: your temple. It’s a symbiotic relationship.

— Mrs. N.

I’m looking closer at the survey’s results on health because of the sheer number of responses that report health issues. I struggle with issues, like back pains, and that forced me to upgrade my work environment with a proper desk and chair. I tend to code on my bed, and well, it worked. But perhaps it wasn’t the best thing for my physical health.

I know we can fall into the stereotype of people who spend 8-12 hours staring at two big monitors, sitting in a plush gaming chair, while frantically typing away at a mechanical keyboard. You know, the Hackers stereotype. I know that isn’t an accurate portrayal of who we are, but it’s easy to become that because of how people look at and understand our work.

And if you feel a great deal of pressure to keep up with that image, I think it’s worth getting into a more healthy mindset, one that gets more than a few hours of sleep, prioritizes exercise, maintains a balanced diet, and all those things we know are ultimately good for us. Even though 20% of folks say they have no health issues at all, a whopping 80% struggle with health issues ranging from sleep deprivation to keeping a healthy weight. You are important and deserve to feel healthy.

Think about your health the way you think about the UI/UX of the websites you design and build. It makes up a part of the design, but has the crucial role of turning ordinary tasks into enjoyable experiences, which in turn, transforms into an overall beautiful experience for the user.

Your health is the same. Those small parts often overlooked can and will affect the great machine that is your body. Here’s a small list of life improvements you can make right now.

Closing thoughts

Diversity, representation, experience, income, and health. That’s what stood out to me in the 2025 State of Devs survey results. I see positive trends in the numbers, but also a huge amount of opportunity to be better, particularly when it comes being more inclusive of women, providing ample chances for upward mobility based on experience, and how we treat ourselves.

Please check out the results and see what stands out to you. What do you notice? Is there anything you are able to take away from the survey that you can use in your own work or life? I’d love to know!

What I Took From the State of Dev 2025 Survey originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

First, what is line length? Line length is the length of a container that holds a body of multi-line text. “Multi-line” is the key part here, because text becomes less readable if the beginning of a line of text is too far away from the end of the prior line of text. This causes users to reread lines by mistake, and generally get lost while reading.

Luckily, the Web Content Accessibility Guidelines (WCAG) gives us a pretty hard rule to follow: no more than 80 characters on a line (40 if the language is Chinese, Japanese, or Korean), which is super easy to implement using character (ch) units:

width: 80ch;

The width of 1ch is equal to the width of the number 0 in your chosen font, so the exact width depends on the font.

Setting the optimal line length

Just because you’re allowed up to 80 characters on a line, it doesn’t mean that you have to aim for that number. A study by the Baymard Institute revealed that a line length of 50-75 characters is the optimal length — this takes into consideration that smaller line lengths mean more lines and, therefore, more opportunities for users to make reading mistakes.

That being said, we also have responsive design to think about, so setting a minimum width (e.g., min-width: 50ch) isn’t a good idea because you’re unlikely to fit 50 characters on a line with, for example, a screen/window size that is 320 pixels wide. So, there’s a bit of nuance involved, and the best way to handle that is by combining the clamp() and min() functions:

• clamp(): Set a fluid value that’s relative to a container using percentage, viewport, or container query units, but with minimum and maximum constraints.
• min(): Set the smallest value from a list of comma-separated values.

Let’s start with min(). One of the arguments is 93.75vw. Assuming that the container extends across the whole viewport, this’d equal 300px when the viewport width is 320px (allowing for 20px of spacing to be distributed as you see fit) and 1350px when the viewport width is 1440px. However, for as long as the other argument (50ch) is the smallest of the two values, that’s the value that min() will resolve to.

min(93.75vw, 50ch);

Next is clamp(), which accepts three arguments in the following order: the minimum, preferred, and maximum values. This is how we’ll set the line length.

For the minimum, you’d plug in your min() function, which sets the 50ch line length but only conditionally. For the maximum, I suggest 75ch, as mentioned before. The preferred value is totally up to you — this will be the width of your container when not hitting the minimum or maximum.

width: clamp(min(93.75vw, 50ch), 70vw, 75ch);

In addition, you can use min(), max(), and calc() in any of those arguments to add further nuance.

If the container feels too narrow, then the font-size might be too large. If it feels too wide, then the font-size might be too small.

Fit text to container (with JavaScript)

You know that design trend where text is made to fit the width of a container? Typically, to utilize as much of the available space as possible? You’ll often see it applied to headings on marketing pages and blog posts. Well, Chris wrote about it back in 2018, rounding up several ways to achieve the effect with JavaScript or jQuery, unfortunately with limitations. However, the ending reveals that you can just use SVG as long as you know the viewBox values, and I actually have a trick for getting them.

Although it still requires 3-5 lines of JavaScript, it’s the shortest method I’ve found. It also slides into HTML and CSS perfectly, particularly since the SVG inherits many CSS properties (including the color, thanks to fill: currentColor):

CodePen Embed Fallback <h1 class="container"> <svg> <text>Fit text to container</text> </svg> </h1> h1.container { /* Container size */ width: 100%; /* Type styles (<text> will inherit most of them) */ font: 900 1em system-ui; color: hsl(43 74% 3%); text { /* We have to use fill: instead of color: here But we can use currentColor to inherit the color */ fill: currentColor; } } /* Select all SVGs */ const svg = document.querySelectorAll("svg"); /* Loop all SVGs */ svg.forEach(element => { /* Get bounding box of <text> element */ const bbox = element.querySelector("text").getBBox(); /* Apply bounding box values to SVG element as viewBox */ element.setAttribute("viewBox", [bbox.x, bbox.y, bbox.width, bbox.height].join(" ")); }); Fit text to container (pure CSS)

If you’re hell-bent on a pure-CSS method, you are in luck. However, despite the insane things that we can do with CSS these days, Roman Komarov’s fit-to-width hack is a bit complicated (albeit rather impressive). Here’s the gist of it:

• The text is duplicated a couple of times (although hidden accessibly with aria-hidden and hidden literally with visibility: hidden) so that we can do math with the hidden ones, and then apply the result to the visible one.
• Using container queries/container query units, the math involves dividing the inline size of the text by the inline size of the container to get a scaling factor, which we then use on the visible text’s font-size to make it grow or shrink.
• To make the scaling factor unitless, we use the tan(atan2()) type-casting trick.
• Certain custom properties must be registered using the @property at-rule (otherwise they don’t work as intended).
• The final font-size value utilizes clamp() to set minimum and maximum font sizes, but these are optional.

<span class="text-fit"> <span> <span class="text-fit"> <span><span>fit-to-width text</span></span> <span aria-hidden="true">fit-to-width text</span> </span> </span> <span aria-hidden="true">fit-to-width text</span> </span> .text-fit { display: flex; container-type: inline-size; --captured-length: initial; --support-sentinel: var(--captured-length, 9999px); & > [aria-hidden] { visibility: hidden; } & > :not([aria-hidden]) { flex-grow: 1; container-type: inline-size; --captured-length: 100cqi; --available-space: var(--captured-length); & > * { --support-sentinel: inherit; --captured-length: 100cqi; --ratio: tan( atan2( var(--available-space), var(--available-space) - var(--captured-length) ) ); --font-size: clamp( 1em, 1em * var(--ratio), var(--max-font-size, infinity * 1px) - var(--support-sentinel) ); inline-size: var(--available-space); &:not(.text-fit) { display: block; font-size: var(--font-size); @container (inline-size > 0) { white-space: nowrap; } } /* Necessary for variable fonts that use optical sizing */ &.text-fit { --captured-length2: var(--font-size); font-variation-settings: "opsz" tan(atan2(var(--captured-length2), 1px)); } } } } @property --captured-length { syntax: "<length>"; initial-value: 0px; inherits: true; } @property --captured-length2 { syntax: "<length>"; initial-value: 0px; inherits: true; } CodePen Embed Fallback Watch for new text-grow/text-shrink properties

To make fitting text to a container possible in just one line of CSS, a number of solutions have been discussed. The favored solution seems to be two new text-grow and text-shrink properties. Personally, I don’t think we need two different properties. In fact, I prefer the simpler alternative, font-size: fit-width, but since text-grow and text-shrink are already on the table (Chrome intends to prototype and you can track it), let’s take a look at how they could work.

The first thing that you need to know is that, as proposed, the text-grow and text-shrink properties can apply to multiple lines of wrapped text within a container, and that’s huge because we can’t do that with my JavaScript technique or Roman’s CSS technique (where each line needs to have its own container).

Both have the same syntax, and you’ll need to use both if you want to allow both growing and shrinking:

text-grow: <fit-target> <fit-method>? <length>?; text-shrink: <fit-target> <fit-method>? <length>?;

• <fit-target>
  • per-line: For text-grow, lines of text shorter than the container will grow to fit it. For text-shrink, lines of text longer than the container will shrink to fit it.
  • consistent: For text-grow, the shortest line will grow to fit the container while all other lines grow by the same scaling factor. For text-shrink, the longest line will shrink to fit the container while all other lines shrink by the same scaling factor.
• <fit-method> (optional)
  • scale: Scale the glyphs instead of changing the font-size.
  • scale-inline: Scale the glyphs instead of changing the font-size, but only horizontally.
  • font-size: Grow or shrink the font size accordingly. (I don’t know what the default value would be, but I imagine this would be it.)
  • letter-spacing: The letter spacing will grow/shrink instead of the font-size.
• <length> (optional): The maximum font size for text-grow or minimum font size for text-shrink.

Again, I think I prefer the font-size: fit-width approach as this would grow and shrink all lines to fit the container in just one line of CSS. The above proposal does way more than I want it to, and there are already a number of roadblocks to overcome (many of which are accessibility-related). That’s just me, though, and I’d be curious to know your thoughts in the comments.

Conclusion

It’s easier to set line length with CSS now than it was a few years ago. Now we have character units, clamp() and min() (and max() and calc() if you wanted to throw those in too), and wacky things that we can do with SVGs and CSS to fit text to a container. It does look like text-grow and text-shrink (or an equivalent solution) are what we truly need though, at least in some scenarios.

Until we get there, this is a good time to weigh-in, which you can do by adding your feedback, tests, and use-cases to the GitHub issue.

Setting Line Length in CSS (and Fitting Text to a Container) originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

Scroll-driven animations are great! They’re a powerful tool that lets developers tie the movement and transformation of elements directly to the user’s scroll position. This technique opens up new ways to create interactive experiences, cuing images to appear, text to glide across the stage, and backgrounds to subtly shift. Used thoughtfully, scroll-driven animations (SDA) can make your website feel more dynamic, engaging, and responsive.

A few weeks back, I was playing around with scroll-driven animations, just searching for all sorts of random things you could do with it. That’s when I came up with the idea to animate the text of the main heading (h1) and, using SDA, change the heading itself based on the user’s scroll position on the page. In this article, we’re going to break down that idea and rebuild it step by step. This is the general direction we’ll be heading in, which looks better in full screen and viewed in a Chromium browser:

CodePen Embed Fallback

It’s important to note that the effect in this example only works in browsers that support scroll-driven animations. Where SDA isn’t supported, there’s a proper fallback to static headings. From an accessibility perspective, if the browser has reduced motion enabled or if the page is being accessed with assistive technology, the effect is disabled and the user gets all the content in a fully semantic and accessible way.

Just a quick note: this approach does rely on a few “magic numbers” for the keyframes, which we’ll talk about later on. While they’re surprisingly responsive, this method is really best suited for static content, and it’s not ideal for highly dynamic websites.

Closer Look at the Animation

Before we dive into scroll-driven animations, let’s take a minute to look at the text animation itself, and how it actually works. This is based on an idea I had a few years back when I wanted to create a typewriter effect. At the time, most of the methods I found involved animating the element’s width, required using a monospace font, or a solid color background. None of which really worked for me. So I looked for a way to animate the content itself, and the solution was, as it often is, in pseudo-elements.

CodePen Embed Fallback

Pseudo-elements have a content property, and you can (kind of) animate that text. It’s not exactly animation, but you can change the content dynamically. The cool part is that the only thing that changes is the text itself, no other tricks required.

Start With a Solid Foundation

Now that you know the trick behind the text animation, let’s see how to combine it with a scroll-driven animation, and make sure we have a solid, accessible fallback as well.

We’ll start with some basic semantic markup. I’ll wrap everything in a main element, with individual sections inside. Each section gets its own heading and content, like text and images. For this example, I’ve set up four sections, each with a bit of text and some images, all about Primary Colors.

<main> <section> <h1>Primary Colors</h1> <p>The three primary colors (red, blue, and yellow) form the basis of all other colors on the color wheel. Mixing them in different combinations produces a wide array of hues.</p> <img src="./colors.jpg" alt="...image description"> </section> <section> <h2>Red Power</h2> <p>Red is a bold and vibrant color, symbolizing energy, passion, and warmth. It easily attracts attention and is often linked with strong emotions.</p> <img src="./red.jpg" alt="...image description"> </section> <section> <h2>Blue Calm</h2> <p>Blue is a calm and cool color, representing tranquility, stability, and trust. It evokes images of the sky and sea, creating a peaceful mood.</p> <img src="./blue.jpg" alt="...image description"> </section> <section> <h2>Yellow Joy</h2> <p>Yellow is a bright and cheerful color, standing for light, optimism, and creativity. It is highly visible and brings a sense of happiness and hope.</p> <img src="./yellow.jpg" alt="...image description"> </section> </main>

As for the styling, I’m not doing anything special at this stage, just the basics. I changed the font and adjusted the text and heading sizes, set up the display for the main and the sections, and fixed the image sizes with object-fit.

CodePen Embed Fallback

So, at this point, we have a simple site with static, semantic, and accessible content, which is great. Now the goal is to make sure it stays that way as we start adding our effect.

The Second First Heading

We’ll start by adding another h1 element at the top of the main. This new element will serve as the placeholder for our animated text, updating according to the user’s scroll position. And yes, I know there’s already an h1 in the first section; that’s fine and we’ll address it in a moment so that only one is accessible at a time.

<h1 class="scrollDrivenHeading" aria-hidden="true">Primary Colors</h1>

Notice that I’ve added aria-hidden="true" to this heading, so it won’t be picked up by screen readers. Now I can add a class specifically for screen readers, .srOnly, to all the other headings. This way, anyone viewing the content “normally” will see only the animated heading, while assistive technology users will get the regular, static semantic headings.

CodePen Embed Fallback

Note: The style for the .srOnly class is based on “Inclusively Hidden” by Scott O’Hara.

Handling Support

As much as accessibility matters, there’s another concern we need to keep in mind: support. CSS Scroll-Driven Animations are fantastic, but they’re still not fully supported everywhere. That’s why it’s important to provide the static version for browsers that don’t support SDA.

The first step is to hide the animated heading we just added using display: none. Then, we’ll add a new @supports block to check for SDA support. Inside that block, where SDA is supported, we can change back the display for the heading.

The .srOnly class should also move into the @supports block, since we only want it to apply when the effect is active, not when it’s not supported. This way, just like with assistive technology, anyone visiting the page in a browser without SDA support will still get the static content.

.scrollDrivenHeading { display: none; } @supports (animation-timeline: scroll()) { .scrollDrivenHeading { display: block; } /* Screen Readers Only */ .srOnly { clip: rect(0 0 0 0); clip-path: inset(50%); height: 1px; overflow: hidden; position: absolute; white-space: nowrap; width: 1px; } } Get Sticky

The next thing we need to do is handle the stickiness of the heading. To make sure the heading always stays on screen, we’ll set its position to sticky with top: 0 so it sticks to the top of the viewport.

While we’re at it, let’s add some basic styling, including a background so the text doesn’t blend with whatever’s behind the heading, a bit of padding for spacing, and white-space: nowrap to keep the heading on a single line.

/* inside the @supports block */ .scrollDrivenHeading { display: block; position: sticky; top: 0; background-image: linear-gradient(0deg, transparent, black 1em); padding: 0.5em 0.25em; white-space: nowrap; }

Now everything’s set up: in normal conditions, we’ll see a single sticky heading at the top of the page. And if someone uses assistive technology or a browser that doesn’t support SDA, they’ll still get the regular static content.

CodePen Embed Fallback

Now we’re ready to start animating the text. Almost…

The Magic Numbers

To build the text animation, we need to know exactly where the text should change. With SDA, scrolling basically becomes our timeline, and we have to determine the exact points on that timeline to trigger the animation.

To make this easier, and to help you pinpoint those positions, I’ve prepared the following script:

@property --scroll-position { syntax: "<number>"; inherits: false; initial-value: 0; } body::after { counter-reset: sp var(--scroll-position); content: counter(sp) "%"; position: fixed; top: 0; left: 0; padding: 1em; background-color: maroon; animation: scrollPosition steps(100); animation-timeline: scroll(); } @keyframes scrollPosition { 0% { --scroll-position: 0; } 100% { --scroll-position: 100; } }

I don’t want to get too deep into this code, but the idea is to take the same scroll timeline we’ll use next to animate the text, and use it to animate a custom property (--scroll-position) from 0 to 100 based on the scroll progress, and display that value in the content.

If we’ll add this at the start of our code, we’ll see a small red square in the top-left corner of the screen, showing the current scroll position as a percentage (to match the keyframes). This way, you can scroll to any section you want and easily mark the percentage where each heading should begin.

CodePen Embed Fallback

With this method and a bit of trial and error, I found that I want the headings to change at 30%, 60%, and 90%. So, how do we actually do it? Let’s start animating.

Animating Text

First, we’ll clear out the content inside the .scrollDrivenHeading element so it’s empty and ready for dynamic content. In the CSS, I’ll add a pseudo-element to the heading, which we’ll use to animate the text. We’ll give it empty content, set up the animation-name, and of course, assign the animation-timeline to scroll().

And since I’m animating the content property, which is a discrete type, it doesn’t transition smoothly between values. It just jumps from one to the next. By setting the animation-timing-function property to step-end, I make sure each change happens exactly at the keyframe I define, so the text switches precisely where I want it to, instead of somewhere in between.

.scrollDrivenHeading { /* style */ &::after { content: ''; animation-name: headingContent; animation-timing-function: step-end; animation-timeline: scroll(); } }

As for the keyframes, this part is pretty straightforward (for now). We’ll set the first frame (0%) to the first heading, and assign the other headings to the percentages we found earlier.

@keyframes headingContent { 0% { content: 'Primary Colors'} 30% { content: 'Red Power'} 60% { content: 'Blue Calm'} 90%, 100% { content: 'Yellow Joy'} }

So, now we’ve got a site with a sticky heading that updates as you scroll.

CodePen Embed Fallback

But wait, right now it just switches instantly. Where’s the animation?! Here’s where it gets interesting. Since we’re not using JavaScript or any string manipulation, we have to write the keyframes ourselves. The best approach is to start from the target heading you want to reach, and build backwards. So, if you want to animate between the first and second heading, it would look like this:

@keyframes headingContent { 0% { content: 'Primary Colors'} 9% { content: 'Primary Color'} 10% { content: 'Primary Colo'} 11% { content: 'Primary Col'} 12% { content: 'Primary Co'} 13% { content: 'Primary C'} 14% { content: 'Primary '} 15% { content: 'Primary'} 16% { content: 'Primar'} 17% { content: 'Prima'} 18% { content: 'Prim'} 19% { content: 'Pri'} 20% { content: 'Pr'} 21% { content: 'P'} 22% { content: 'R'} 23% { content: 'Re'} 24% { content: 'Red'} 25% { content: 'Red '} 26% { content: 'Red P'} 27% { content: 'Red Po'} 28%{ content: 'Red Pow'} 29% { content: 'Red Powe'} 30% { content: 'Red Power'} 60% { content: 'Blue Calm'} 90%, 100% { content: 'Yellow Joy'} }

I simply went back by 1% each time, removing or adding a letter as needed. Note that in other cases, you might want to use a different step size, and not always 1%. For example, on longer headings with more words, you’ll probably want smaller steps.

If we repeat this process for all the other headings, we’ll end up with a fully animated heading.

CodePen Embed Fallback User Preferences

We talked before about accessibility and making sure the content works well with assistive technology, but there’s one more thing you should keep in mind: prefers-reduced-motion. Even though this isn’t a strict WCAG requirement for this kind of animation, it can make a big difference for people with vestibular sensitivities, so it’s a good idea to offer a way to show the content without animations.

If you want to provide a non-animated alternative, all you need to do is wrap your @supports block with a prefers-reduced-motion query:

@media screen and (prefers-reduced-motion: no-preference) { @supports (animation-timeline: scroll()) { /* style */ } } Leveling Up

Let’s talk about variations. In the previous example, we animated the entire heading text, but we don’t have to do that. You can animate just the part you want, and use additional animations to enhance the effect and make things more interesting. For example, here I kept the text “Primary Color” fixed, and added a span after it that handles the animated text.

<h1 class="scrollDrivenHeading" aria-hidden="true"> Primary Color<span></span> </h1>

And since I now have a separate span, I can also animate its color to match each value.

CodePen Embed Fallback

In the next example, I kept the text animation on the span, but instead of changing the text color, I added another scroll-driven animation on the heading itself to change its background color. This way, you can add as many animations as you want and change whatever you like.

CodePen Embed Fallback Your Turn!

CSS Scroll-Driven Animations are more than just a cool trick; they’re a game-changer that opens the door to a whole new world of web design. With just a bit of creativity, you can turn even the most ordinary pages into something interactive, memorable, and truly engaging. The possibilities really are endless, from subtle effects that enhance the user experience, to wild, animated transitions that make your site stand out.

So, what would you build with scroll-driven animations? What would you create with this new superpower? Try it out, experiment, and if you come up with something cool, have some ideas, wild experiments, or even weird failures, I’d love to hear about them. I’m always excited to see what others come up with, so feel free to share your work, questions, or feedback below.

Special thanks to Cristian Díaz for reviewing the examples, making sure everything is accessible, and contributing valuable advice and improvements.

Scroll-Driven Sticky Heading originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

Layout. It’s one of those easy-to-learn, difficult-to-master things, like they say about playing bass. Not because it’s innately difficult to, say, place two elements next to each other, but because there are many, many ways to tackle it. And layout is one area of CSS that seems to evolve more than others, as we’ve seen in the past 10-ish years with the Flexbox, CSS Grid, Subgrid, and now Masonry to name but a few. May as well toss in Container Queries while we’re at it. And reading flow. And…

That’s a good way to start talking about a new online course that Ahmad Shadeed is planning to release called The Layout Maestro. I love that name, by the way. It captures exactly how I think about working with layouts: orchestrating how and where things are arranged on a page. Layouts are rarely static these days. They are expected to adapt to the user’s context, not totally unlike a song changing keys.

Ahmad is the perfect maestro to lead a course on layout, as he does more than most when it comes to experimenting with layout features and demonstrating practical use cases, as you may have already seen in his thorough and wildly popular interactive guides on Container Queries, grid areas, box alignment, and positioning (just to name a few).

The course is still in development, but you can get a leg up and sign up to be notified by email when it’s ready. That’s literally all of the information I have at this point, but I still feel compelled to share it and encourage you to sign up for updates because I know few people more qualified to wax on about CSS layout than Ahmad and am nothing but confident that it will be great, worth the time, and worth the investment.

I’m also learning that I have a really hard time typing “maestro” correctly. &#x1f913;

The Layout Maestro Course originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

They say context is king and that's certainly true in AI products where the content, tools, and instructions applications provide to AI models shape their behavior and subsequent results. But if context is so critical, how do we allow people to understand and manage it when interacting with AI-driven software?

In AI products, there's a lot of stuff that could be in context (provided to an AI model as part of its instructions) at any given point, but not everything will be in context all the time because AI models have context limits. So when getting results from AI products, people aren't sure if or how much they should trust them. Was the right information used to answer my question? Did the model hallucinate or use the wrong information?

When I launched my personal AI two years ago, context was much simpler than it is today. In Ask LukeW, when people ask a question about digital product design, the system searches through my writings, finds and puts the most relevant bits into context for AI models to use and reference, then cities them in the results people see. This is pretty transparent in the interface: the articles, videos, audio, and PDFs used are shown on the right with citations within each response to where these files were used the most.

The most complicated things get in Ask LukeW is when someone opens one of these citied articles, videos, or PDFs to view its full contents. In this case, a small "context chip" is added to the question bar to make clear questions can be asked of just this file. In other words, the file is the primary thing in context. If someone wants to ask a question of the whole corpus of my writings and talks again, they can simply click on the X that removes this context constraint and the chip disappears from the question bar. You can try this out yourself here.

Context chips are pretty common in AI products today because they're a relatively easy way to both give people a sense of what's influencing an AI model's replies and to add or remove it. When what's in context expands, however, they don't scale very well. For example, Augment Code uses context chips for retrieval systems, active files, selected text, and more.

Using a context chip to display everything influencing an AI model's response begins to break down when many things (especially different things) are in context. Displaying them all eats up valuable space in the UI and requires that their names or identifiers are truncated to fit. That kind of defeats the purpose of "showing you what's in context". Also when AI products do automatic context retrieval like Augment Code's context retrieval engine: does that always show up as a chip? or should people not worry about it and trust the system is finding and putting the right things into context?

With AI products using agents these issues are compounded because each tool call an agent makes can retrieve context in different ways or multiple times. So showing every bit of context found or created by tools as a context chip quickly breaks down. To account for this in earlier versions of Bench, we showed the context from tools used by agents as it was being created. But this turned out to be a jarring experience as the context would show up then go away when the next tool's context arrived (as you can see in the video).

Since then, we've moved to showing an agent's process of creating something as condensed steps with links to the context in each step. So people can click on any given steps to see the context a tool either found or created. But that context isn't being automatically flashed in front of them as it's made. This lets people focus on the output and only dig into the process when they want to understand what led to the output.

This approach becomes even more relevant with agent orchestration. When agents can make use of agents themselves, you end up with nested amounts of context. Told you things were a lot simpler two years ago! In these cases, Bench just shows the collective context combined from multiple tool calls in one link. This allows people to examine what cumulative context was created by sub agents. But importantly this combined context is treated the same way - whether it comes from a single tool or a subagent that uses multiple tools.

While making context understood and manageable feels like the right thing to provide transparency and control, increasingly people seem to focus more on the output of AI products and less on the process that created them. Only when things don't seem "right" do they dig into the kinds of process timelines and context links that Bench provides. So if people become even more confident using AI products, we might see context management UIs with even less presence.

They say context is king and that's certainly true in AI products where the content, tools, and instructions applications provide to AI models shape their behavior and subsequent results. But if context is so critical, how do we allow people to understand and manage it when interacting with AI-driven software?

In AI products, there's a lot of stuff that could be in context (provided to an AI model as part of its instructions) at any given point, but not everything will be in context all the time because AI models have context limits. So when getting results from AI products, people aren't sure if or how much they should trust them. Was the right information used to answer my question? Did the model hallucinate or use the wrong information?

When I launched my personal AI two years ago, context was much simpler than it is today. In Ask LukeW, when people ask a question about digital product design, the system searches through my writings, finds and puts the most relevant bits into context for AI models to use and reference, then cities them in the results people see. This is pretty transparent in the interface: the articles, videos, audio, and PDFs used are shown on the right with citations within each response to where these files were used the most.

The most complicated things get in Ask LukeW is when someone opens one of these citied articles, videos, or PDFs to view its full contents. In this case, a small "context chip" is added to the question bar to make clear questions can be asked of just this file. In other words, the file is the primary thing in context. If someone wants to ask a question of the whole corpus of my writings and talks again, they can simply click on the X that removes this context constraint and the chip disappears from the question bar. You can try this out yourself here.

Context chips are pretty common in AI products today because they're a relatively easy way to both give people a sense of what's influencing an AI model's replies and to add or remove it. When what's in context expands, however, they don't scale very well. For example, Augment Code uses context chips for retrieval systems, active files, selected text, and more.

Using a context chip to display everything influencing an AI model's response begins to break down when many things (especially different things) are in context. Displaying them all eats up valuable space in the UI and requires that their names or identifiers are truncated to fit. That kind of defeats the purpose of "showing you what's in context". Also when AI products do automatic context retrieval like Augment Code's context retrieval engine: does that always show up as a chip? or should people not worry about it and trust the system is finding and putting the right things into context?

With AI products using agents these issues are compounded because each tool call an agent makes can retrieve context in different ways or multiple times. So showing every bit of context found or created by tools as a context chip quickly breaks down. To account for this in earlier versions of Bench, we showed the context from tools used by agents as it was being created. But this turned out to be a jarring experience as the context would show up then go away when the next tool's context arrived (as you can see in the video).

Since then, we've moved to showing an agent's process of creating something as condensed steps with links to the context in each step. So people can click on any given steps to see the context a tool either found or created. But that context isn't being automatically flashed in front of them as it's made. This lets people focus on the output and only dig into the process when they want to understand what led to the output.

This approach becomes even more relevant with agent orchestration. When agents can make use of agents themselves, you end up with nested amounts of context. Told you things were a lot simpler two years ago! In these cases, Bench just shows the collective context combined from multiple tool calls in one link. This allows people to examine what cumulative context was created by sub agents. But importantly this combined context is treated the same way - whether it comes from a single tool or a subagent that uses multiple tools.

While making context understood and manageable feels like the right thing to provide transparency and control, increasingly people seem to focus more on the output of AI products and less on the process that created them. Only when things don't seem "right" do they dig into the kinds of process timelines and context links that Bench provides. So if people become even more confident using AI products, we might see context management UIs with even less presence.

This is the fourth post in a series about the new CSS shape() function. So far, we’ve covered the most common commands you will use to draw various shapes, including lines, arcs, and curves. This time, I want to introduce you to two more commands: close and move. They’re fairly simple in practice, and I think you will rarely use them, but they are incredibly useful when you need them.

Better CSS Shapes Using shape()

1. Lines and Arcs
2. More on Arcs
3. Curves
4. Close and Move (you are here!)

The close command

In the first part, we said that shape() always starts with a from command to define the first starting point but what about the end? It should end with a close command.

But you never used any close command in the previous articles!?

That’s true. I never did because I either “close” the shape myself or rely on the browser to “close” it for me. Said like that, it’s a bit confusing, but let’s take a simple example to better understand:

clip-path: shape(from 0 0, line to 100% 0, line to 100% 100%)

If you try this code, you will get a triangle shape, but if you look closely, you will notice that we have only two line commands whereas, to draw a triangle, we need a total of three lines. The last line between 100% 100% and 0 0 is implicit, and that’s the part where the browser is closing the shape for me without having to explicitly use a close command.

I could have written the following:

clip-path: shape(from 0 0, line to 100% 0, line to 100% 100%, close)

Or instead, define the last line by myself:

clip-path: shape(from 0 0, line to 100% 0, line to 100% 100%, line to 0 0)

But since the browser is able to close the shape alone, there is no need to add that last line command nor do we need to explicitly add the close command.

This might lead you to think that the close command is useless, right? It’s true in most cases (after all, I have written three articles about shape() without using it), but it’s important to know about it and what it does. In some particular cases, it can be useful, especially if used in the middle of a shape.

CodePen Embed Fallback

In this example, my starting point is the center and the logic of the shape is to draw four triangles. In the process, I need to get back to the center each time. So, instead of writing line to center, I simply write close and the browser will automatically get back to the initial point!

Intuitively, we should write the following:

clip-path: shape( from center, line to 20% 0, hline by 60%, line to center, /* triangle 1 */ line to 100% 20%, vline by 60%, line to center, /* triangle 2 */ line to 20% 100%, hline by 60%, line to center, /* triangle 3 */ line to 0 20%, vline by 60% /* triangle 4 */ )

But we can optimize it a little and simply do this instead:

clip-path: shape( from center, line to 20% 0, hline by 60%, close, line to 100% 20%, vline by 60%, close, line to 20% 100%, hline by 60%, close, line to 0 20%, vline by 60% )

We write less code, sure, but another important thing is that if I update the center value with another position, the close command will follow that position.

CodePen Embed Fallback

Don’t forget about this trick. It can help you optimize a lot of shapes by writing less code.

The move command

Let’s turn our attention to another shape() command you may rarely use, but can be incredibly useful in certain situations: the move command.

Most times when we need to draw a shape, it’s actually one continuous shape. But it may happen that our shape is composed of different parts not linked together. In these situations, the move command is what you will need.

Let’s take an example, similar to the previous one, but this time the triangles don’t touch each other:

CodePen Embed Fallback

Intuitively, we may think we need four separate elements, with its own shape() definition. But the that example is a single shape!

The trick is to draw the first triangle, then “move” somewhere else to draw the next one, and so on. The move command is similar to the from command but we use it in the middle of shape().

clip-path: shape( from 50% 40%, line to 20% 0, hline by 60%, close, /* triangle 1 */ move to 60% 50%, line to 100% 20%, vline by 60%, close, /* triangle 2 */ move to 50% 60%, line to 20% 100%, hline by 60%, close, /* triangle 3 */ move to 40% 50%, line to 0 20%, vline by 60% /* triangle 4 */ )

After drawing the first triangle, we “close” it and “move” to a new point to draw the next triangle. We can have multiple shapes using a single shape() definition. A more generic code will look like the below:

clip-path: shape( from X1 Y1, ..., close, /* shape 1 */ move to X2 Y2, ..., close, /* shape 2 */ ... move to Xn Yn, ... /* shape N */ )

The close commands before the move commands aren’t mandatory, so the code can be simplified to this:

clip-path: shape( from X1 Y1, ..., /* shape 1 */ move to X2 Y2, ..., /* shape 2 */ ... move to Xn Yn, ... /* shape N */ ) CodePen Embed Fallback

Let’s look at a few interesting use cases where this technique can be helpful.

Cut-out shapes

Previously, I shared a trick on how to create cut-out shapes using clip-path: polygon(). Starting from any kind of polygon, we can easily invert it to get its cut-out version:

CodePen Embed Fallback

We can do the same using shape(). The idea is to have an intersection between the main shape and the rectangle shape that fits the element boundaries. We need two shapes, hence the need for the move command.

The code is as follows:

.shape { clip-path: shape(from ...., move to 0 0, hline to 100%, vline to 100%, hline to 0); }

You start by creating your main shape and then you “move” to 0 0 and you create the rectangle shape (Remember, It’s the first shape we create in the first part of this series). We can even go further and introduce a CSS variable to easily switch between the normal shape and the inverted one.

.shape { clip-path: shape(from .... var(--i,)); } .invert { --i:,move to 0 0, hline to 100%, vline to 100%, hline to 0; }

By default, --i is not defined so var(--i,)will be empty and we get the main shape. If we define the variable with the rectangle shape, we get the inverted version.

Here is an example using a rounded hexagon shape:

CodePen Embed Fallback

In reality, the code should be as follows:

.shape { clip-path: shape(evenodd from .... var(--i,)); } .invert { --i:,move to 0 0, hline to 100%, vline to 100%, hline to 0; }

Notice the evenodd I am adding at the beginning of shape(). I won’t bother you with a detailed explanation on what it does but in some cases, the inverted shape is not visible and the fix is to add evenodd at the beginning. You can check the MDN page for more details.

Another improvement we can do is to add a variable to control the space around the shape. Let’s suppose you want to make the hexagon shape of the previous example smaller. It‘s tedious to update the code of the hexagon but it’s easier to update the code of the rectangle shape.

.shape { clip-path: shape(evenodd from ... var(--i,)) content-box; } .invert { --d: 20px; padding: var(--d); --i: ,move to calc(-1*var(--d)) calc(-1*var(--d)), hline to calc(100% + var(--d)), vline to calc(100% + var(--d)), hline to calc(-1*var(--d)); }

We first update the reference box of the shape to be content-box. Then we add some padding which will logically reduce the area of the shape since it will no longer include the padding (nor the border). The padding is excluded (invisible) by default and here comes the trick where we update the rectangle shape to re-include the padding.

That is why the --i variable is so verbose. It uses the value of the padding to extend the rectangle area and cover the whole element as if we didn’t have content-box.

CodePen Embed Fallback

Not only you can easily invert any kind of shape, but you can also control the space around it! Here is another demo using the CSS-Tricks logo to illustrate how easy the method is:

CodePen Embed Fallback

This exact same example is available in my SVG-to-CSS converter, providing you with the shape() code without having to do all of the math.

Repetitive shapes

Another interesting use case of the move command is when we need to repeat the same shape multiple times. Do you remember the difference between the by and the to directives? The by directive allows us to define relative coordinates considering the previous point. So, if we create our shape using only by, we can easily reuse the same code as many times as we want.

Let’s start with a simple example of a circle shape:

clip-path: shape(from X Y, arc by 0 -50px of 1%, arc by 0 50px of 1%)

Starting from X Y, I draw a first arc moving upward by 50px, then I get back to X Y with another arc using the same offset, but downward. If you are a bit lost with the syntax, try reviewing Part 1 to refresh your memory about the arc command.

How I drew the shape is not important. What is important is that whatever the value of X Y is, I will always get the same circle but in a different position. Do you see where I am going with this idea? If I want to add another circle, I simply repeat the same code with a different X Y.

clip-path: shape( from X1 Y1, arc by 0 -50px of 1%, arc by 0 50px of 1%, move to X2 Y2, arc by 0 -50px of 1%, arc by 0 50px of 1% )

And since the code is the same, I can store the circle shape into a CSS variable and draw as many circles as I want:

.shape { --sh:, arc by 0 -50px of 1%, arc by 0 50px of 1%; clip-path: shape( from X1 Y1 var(--sh), move to X2 Y2 var(--sh), ... move to Xn Yn var(--sh) ) }

You don’t want a circle? Easy, you can update the --sh variable with any shape you want. Here is an example with three different shapes:

CodePen Embed Fallback

And guess what? You can invert the whole thing using the cut-out technique by adding the rectangle shape at the end:

CodePen Embed Fallback

This code is a perfect example of the shape() function’s power. We don’t have any code duplication and we can simply adjust the shape with CSS variables. This is something we are unable to achieve with the path() function because it doesn’t support variables.

Conclusion

That’s all for this fourth installment of our series on the CSS shape() function! We didn’t make any super complex shapes, but we learned how two simple commands can open a lot of possibilities of what can be done using shape().

Just for fun, here is one more demo recreating a classic three-dot loader using the last technique we covered. Notice how much further we could go, adding things like animation to the mix:

CodePen Embed Fallback Better CSS Shapes Using shape()

1. Lines and Arcs
2. More on Arcs
3. Curves
4. Close and Move (you are here!)

Better CSS Shapes Using shape() — Part 4: Close and Move originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

Four years ago, I wrote an article titled Minding the “gap”, where I talked about the CSS gap property, where it applied, and how it worked with various CSS layouts.

At the time, I described how easy it was to evenly space items out in a flex, grid, or multi-column layout, by using the gap property. But, I also said that styling the gap areas was much harder, and I shared a workaround.

However, workarounds like using extra HTML elements, pseudo-elements, or borders to draw separator lines tend to come with drawbacks, especially those that impact your layout size, interfere with assistive technologies, or pollute your markup with style-only elements.

Today, I’m writing again about layout gaps, but this time, to tell you all about a new and exciting CSS feature that’s going to change it all. What you previously had to use workarounds for, you’ll soon be able to do with just a few simple CSS properties that make it easy, yet also flexible, to display styled separators between your layout items.

There’s already a specification draft for the feature you can peruse. At the time I’m writing this, it is available in Chrome and Edge 139 behind a flag. But I believe it won’t be long before we turn that flag on. I believe other browsers are also very receptive and engaged.

Displaying decorative lines between items of a layout can make a big difference. When used well, these lines can bring more structure to your layout, and give your users more of a sense of how the different regions of a page are organized.

Introducing CSS gap decorations

If you’ve ever used a multi-column layout, such as by using the column-width property, then you might already be familiar with gap decorations. You can draw vertical lines between the columns of a multi-column layout by using the column-rule property:

article { column-width: 20rem; column-rule: 1px solid black; }

The CSS gap decorations feature builds on this to provide a more comprehensive system that makes it easy for you to draw separator lines in other layout types.

For example, the draft specification says that the column-rule property also works in flexbox and grid layouts:

.my-grid-container { display: grid; gap: 2px; column-rule: 2px solid pink; }

No need for extra elements or borders! The key benefit here is that the decoration happens in CSS only, where it belongs, with no impacts to your semantic markup.

The CSS gap decorations feature also introduces a new row-rule property for drawing lines between rows:

.my-flex-container { display: flex; gap: 10px; row-rule: 10px dotted limegreen; column-rule: 5px dashed coral; }

But that’s not all, because the above syntax also allows you to define multiple, comma-separated, line style values, and use the same repeat() function that CSS grid already uses for row and column templates. This makes it possible to define different styles of line decorations in a single layout, and adapt to an unknown number of gaps:

.my-container { display: grid; gap: 2px; row-rule: repeat(2, 1px dashed red), 2px solid black, repeat(auto, 1px dotted green); }

Finally, the CSS gap decorations feature comes with additional CSS properties such as row-rule-break, column-rule-break, row-rule-outset, column-rule-outset, and gap-rule-paint-order, which make it possible to precisely customize the way the separators are drawn, whether they overlap, or where they start and end.

And of course, all of this works across grid, flexbox, multi-column, and soon, masonry!

Browser support

Currently, the CSS gap decorations feature is only available in Chromium-based browsers.

The feature is still early in the making, and there’s time for you all to try it and to provide feedback that could help make the feature better and more adapted to your needs.

If you want to try the feature today, make sure to use Edge or Chrome, starting with version 139 (or another Chromium-based browser that matches those versions), and enable the flag by following these steps:

1. In Chrome or Edge, go to about://flags.
2. In the search field, search for Enable Experimental Web Platform Features.
3. Enable the flag.
4. Restart the browser.

To put this all into practice, let’s walk through an example together that uses the new CSS gap decorations feature. I also have a final example you can demo.

Using CSS gap decorations

Let’s build a simple web page to learn how to use the feature. Here is what we’ll be building:

The above layout contains a header section with a title, a navigation menu with a few links, a main section with a series of short paragraphs of text and photos, and a footer.

We’ll use the following markup:

<body> <header> <h1>My personal site</h1> </header> <nav> <ul> <li><a href="#">Home</a></li> <li><a href="#">Blog</a></li> <li><a href="#">About</a></li> <li><a href="#">Links</a></li> </ul> </nav> <main> <article> <p>...</p> </article> <article> <img src="cat.jpg" alt="A sleeping cat."> </article> <article> <p>...</p> </article> <article> <img src="tree.jpg" alt="An old olive tree trunk."> </article> <article> <p>...</p> </article> <article> <p>...</p> </article> <article> <p>...</p> </article> <article> <img src="strings.jpg" alt="Snow flakes falling in a motion blur effect."> </article> </main> <footer> <p>© 2025 Patrick Brosset</p> </footer> </body>

We’ll start by making the <body> element be a grid container. This way, we can space out the <header>, <nav>, <main>, and <footer> elements apart in one go by using the gap property:

body { display: grid; gap: 4rem; margin: 2rem; }

Let’s now use the CSS gap decorations feature to display horizontal separator lines within the gaps we just defined:

body { display: grid; gap: 4rem; margin: 2rem; row-rule: 1rem solid #efefef; }

This gives us the following result:

We can do a bit better by making the first horizontal line look different than the other two lines, and simplify the row-rule value by using the repeat() syntax:

body { display: grid; gap: 4rem; margin: 2rem; row-rule: 1rem solid #efefef, repeat(2, 2px solid #efefef); }

With this new row-rule property value, we’re telling the browser to draw the first horizontal separator as a 1rem thick line, and the next two separators as 2px thick lines, which gives the following result:

Now, let’s turn our attention to the navigation element and its list of links. We’ll use flexbox to display the links in a single row, where each link is separated from the other links by a gap and a vertical line:

nav ul { display: flex; flex-wrap: wrap; gap: 2rem; column-rule: 2px dashed #666; }

Very similarly to how we used the row-rule property before, we’re now using the column-rule property to display a dashed 2px thick separator between the links.

Our example web page now looks like this:

The last thing we need to change is the <main> element and its paragraphs and pictures. We’ll use flexbox again and display the various children in a wrapping row of varying width items:

main { display: flex; flex-wrap: wrap; gap: 4rem; } main > * { flex: 1 1 200px; } main article:has(p) { flex-basis: 400px; }

In the above code snippet, we’re setting the <main> element to be a wrapping flex container with a 4rem gap between items and flex lines. We’re also making the items have a flex basis size of 200px for pictures and 400px for text, and allowing them to grow and shrink as needed. This gives us the following result:

Let’s use CSS gap decorations to bring a little more structure to our layout by drawing 2px thick separator lines between the rows and columns of the layout:

main { display: flex; flex-wrap: wrap; gap: 4rem; row-rule: 2px solid #999; column-rule: 2px solid #999; }

This gives us the following result, which is very close to our expected design:

The last detail we want to change is related to the vertical lines. We don’t want them to span across the entire height of the flex lines but instead start and stop where the content starts and stops.

With CSS gap decorations, we can easily achieve this by using the column-rule-outset property to fine-tune exactly where the decorations start and end, relative to the gap area:

main { display: flex; flex-wrap: wrap; gap: 4rem; row-rule: 2px solid #999; column-rule: 2px solid #999; column-rule-outset: 0; }

The column-rule-outset property above makes the vertical column separators span the height of each row, excluding the gap area, which is what we want:

And with that, we’re done with our example. Check out the live example, and source code.

Learn more

There’s more to the feature and I mentioned a couple more CSS properties earlier

• gap-rule-paint-order, which lets you control which of the decorations, rows or columns, appear above the other ones.
• row-rule-break / column-rule-break, which sets the behavior of the decoration lines at intersections. In particular, whether they are made of multiple segments, which start and end at intersections, or single, continuous lines.

Because the feature is new, there isn’t MDN documentation about it yet. So to learn more, check out:

• CSS Gap Decorations Module Level 1 (First Public Working Draft)
• Microsoft Edge Explainer

The Edge team has also created an interactive playground where you can use visual controls to configure gap decorations.

And, of course, the reason this is all implemented behind a flag is to elicit feedback from developers like you! If you have any feedback, questions, or bugs about this feature, I definitely encourage you to open a new ticket on the Chromium issue tracker.

The Gap Strikes Back: Now Stylable originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

Adam Wathan has (very cleverly) built Tailwind with CSS Cascade Layers, making it extremely powerful for organizing styles by priority.

@layer theme, base, components, utilities; @import 'tailwindcss/theme.css' layer(theme); @import 'tailwindcss/utilities.css' layer(utilities);

The core of Tailwind are its utilities. This means you have two choices:

1. The default choice
2. The unorthodox choice

The default choice

The default choice is to follow Tailwind’s recommended layer order: place components first, and Tailwind utilities last.

So, if you’re building components, you need to manually wrap your components with a @layer directive. Then, overwrite your component styles with Tailwind, putting Tailwind as the “most important layer”.

/* Write your components */ @layer components { .component { /* Your CSS here */ } } <!-- Override with Tailwind utilities --> <div class="component p-4"> ... </div>

That’s a decent way of doing things.

But, being the bad boy I am, I don’t take the default approach as the “best” one. Over a year of (major) experimentation with Tailwind and vanilla CSS, I’ve come across what I believe is a better solution.

The Unorthodox Choice

Before we go on, I have to tell you that I’m writing a course called Unorthodox Tailwind — this shows you everything I know about using Tailwind and CSS in synergistic ways, leveraging the strengths of each.

Shameless plug aside, let’s dive into the Unorthodox Choice now.

In this case, the Unorthodox Choice is to write your styles in an unnamed layer — or any layer after utilities, really — so that your CSS naturally overwrites Tailwind utilities.

Of these two, I prefer the unnamed layer option:

/* Unnamed layer option */ @layer theme, base, components, utilities; /* Write your CSS normally here */ .component { /* ... */ } /* Named layer option */ /* Use whatever layer name you come up with. I simply used css here because it made most sense for explaining things */ @layer theme, base, components, utilities, css; @layer css { .component { /* ... */ } }

I have many reasons why I do this:

1. I don’t like to add unnecessary CSS layers because it makes code harder to write — more keystrokes, having to remember the specific layer I used it in, etc.
2. I’m pretty skilled with ITCSS, selector specificity, and all the good-old-stuff you’d expect from a seasoned front-end developer, so writing CSS in a single layer doesn’t scare me at all.
3. I can do complex stuff that are hard or impossible to do in Tailwind (like theming and animations) in CSS.

Your mileage may vary, of course.

Now, if you have followed my reasoning so far, you would have noticed that I use Tailwind very differently:

• Tailwind utilities are not the “most important” layer.
• My unnamed CSS layer is the most important one.

I do this so I can:

• Build prototypes with Tailwind (quickly, easily, especially with the tools I’ve created).
• Shift these properties to CSS when they get more complex — so I don’t have to read messy utility-littered HTML that makes my heart sink. Not because utility HTML is bad, but because it takes lots of brain processing power to figure out what’s happening.

Finally, here’s the nice thing about Tailwind being in a utility layer: I can always !important a utility to give it strength.

<!-- !important the padding utility --> <div class="component !p-4"> ... </div>

Whoa, hold on, wait a minute! Isn’t this wrong, you might ask?

Nope. The !important keyword has traditionally been used to override classes. In this case, we’re leveraging on the !important feature in CSS Layers to say the Tailwind utility is more important than any CSS in the unnamed layer.

This is perfectly valid and is a built-in feature for CSS Layers.

Besides, the !important is so explicit (and used so little) that it makes sense for one-off quick-and-dirty adjustments (without creating a brand new selector for it).

Tailwind utilities are more powerful than they seem

Tailwind utilities are not a 1:1 map between a class and a CSS property. Built-in Tailwind utilities mostly look like this so it can give people a wrong impression.

Tailwind utilities are more like convenient Sass mixins, which means we can build effective tools for layouts, theming, typography, and more, through them.

You can find out about these thoughts inside Unorthodox Tailwind.

Thanks for reading and I hope you’re enjoying a new way of looking at (or using) Tailwind!

Using CSS Cascade Layers With Tailwind Utilities originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

Alongside an increasing sameness of features and user interfaces, AI applications have also converged on their approach to primary calls to action: "What Do You Want To ___?" But is there a better way... especially for more domain specific applications?

Looking across AI products today, most feature an open-ended text field with an equally open-ended call to action:

• What do you want to know?
• What can I help with?
• What do you want to create?
• What do you want to build?
• What will you imagine?
• Ask anything...
• Ask a question...
• Ask [AI tool]...

So many questions. I've even turned them into a running joke. When a financial company integrates their AI: "What do you want to bank?" or "What do you want to accountant?" Silly I know, but it illustrates the issue. People often don't know what AI products can do nor how to best instruct/prompt them. Questions just exacerbate the issue.

It may be a small detail but instead of asking, how about instructing? Reve's image creation call to action says: "Describe an image or drop one here...". Bench's AI-powered workspace starts with: "Describe the task you want Bench to do...". Both calls to action are still open ended enough that so they can capture the kind of broad intent AI models can handle. But perhaps there's something to having a bit more guidance beyond "What Do You Want To AI?"

Alongside an increasing sameness of features and user interfaces, AI applications have also converged on their approach to primary calls to action: "What Do You Want To ___?" But is there a better way... especially for more domain specific applications?

Looking across AI products today, most feature an open-ended text field with an equally open-ended call to action:

• What do you want to know?
• What can I help with?
• What do you want to create?
• What do you want to build?
• What will you imagine?
• Ask anything...
• Ask a question...
• Ask [AI tool]...

So many questions. I've even turned them into a running joke. When a financial company integrates their AI: "What do you want to bank?" or "What do you want to accountant?" Silly I know, but it illustrates the issue. People often don't know what AI products can do nor how to best instruct/prompt them. Questions just exacerbate the issue.

It may be a small detail but instead of asking, how about instructing? Reve's image creation call to action says: "Describe an image or drop one here...". Bench's AI-powered workspace starts with: "Describe the task you want Bench to do...". Both calls to action are still open ended enough that so they can capture the kind of broad intent AI models can handle. But perhaps there's something to having a bit more guidance beyond "What Do You Want To AI?"

Blob, Blob, Blob. You hate them. You love them. Personally, as a design illiterate, I like to overuse them… a lot. And when you repeat the same process over and over again, it’s only a question of how much you can optimize it, or in this case, what’s the easiest way to create blobs in CSS? Turns out, as always, there are many approaches.

To know if our following blobs are worth using, we’ll need them to pass three tests:

1. They can be with just a single element (and preferably without pseudos).
2. They can be easily designed (ideally through an online tool).
3. We can use gradient backgrounds, borders, shadows, and other CSS effects on them.

Without further ado, let’s Blob, Blob, Blob right in.

Just generate them online

I know it’s disenchanting to click on an article about making blobs in CSS just for me to say you can generate them outside CSS. Still, it’s probably the most common way to create blobs on the web, so to be thorough, these are some online tools I’ve used before to create SVG blobs.

• Haikei. Probably the one I have used the most since, besides blobs, it can also generate lots of SVG backgrounds.
• Blobmaker. A dedicated tool for making blobs. It’s apparently part of Haikei now, so you can use both.
• Lastly, almost all graphic programs let you hand-draw blobs and export them as SVGs.

For example, this is one I generated just now. Keep it around, as it will come in handy later.

<svg viewBox="0 0 200 200" xmlns="http://www.w3.org/2000/svg"> <path fill="#FA4D56" d="M65.4,-37.9C79.2,-13.9,81,17,68.1,38C55.2,59.1,27.6,70.5,1.5,69.6C-24.6,68.8,-49.3,55.7,-56,38.2C-62.6,20.7,-51.3,-1.2,-39,-24.4C-26.7,-47.6,-13.3,-72,6.2,-75.6C25.8,-79.2,51.6,-62,65.4,-37.9Z" transform="translate(100 100)" /> </svg> Using border-radius

While counterintuitive, we can use the border-radius property to create blobs. This technique isn’t new by any means; it was first described by Nils Binder in 2018, but it is still fairly unknown. Even for those who use it, the inner workings are not entirely clear.

To start, you may know the border-radius is a shorthand to each individual corner’s radius, going from the top left corner clockwise. For example, we can set each corner’s border-radius to get a bubbly square shape:

<div class="blob"></div> .blob { border-radius: 25% 50% 75% 100%; } CodePen Embed Fallback

However, what border-radius does — and also why it’s called “radius” — is to shape each corner following a circle of the given radius. For example, if we set the top left corner to 25%, it will follow a circle with a radius 25% the size of the shape.

.blob { border-top-left-radius: 25%; } CodePen Embed Fallback

What’s less known is that each corner property is still a shortcut towards its horizontal and vertical radii. Normally, you set both radii to the same value, getting a circle, but you can set them individually to create an ellipse. For example, the following sets the horizontal radius to 25% of the element’s width and the vertical to 50% of its height:

.blob { border-top-left-radius: 25% 50%; } CodePen Embed Fallback

We can now shape each corner like an ellipse, and it is the combination of all four ellipses that creates the illusion of a blob! Just take into consideration that to use the horizontal and vertical radii syntax through the border-radius property, we’ll need to separate the horizontal from the vertical radii using a forward slash (/).

.blob { border-radius: /* horizontal */ 100% 30% 60% 70% / /* vertical */ 50% 40% 70% 70%; } CodePen Embed Fallback

The syntax isn’t too intuitive, so designing a blob from scratch will likely be a headache. Luckily, Nils Binder made a tool exactly for that!

Blobbing blobs together

This hack is awesome. We aren’t supposed to use border-radius like that, but we still do. Admittedly, we are limited to boring blobs. Due to the nature of border-radius, no matter how hard we try, we will only get convex shapes.

Just going off border-radius, we can try to minimize it a little by sticking more than one blob together:

CodePen Embed Fallback

However, I don’t want to spend too much time on this technique since it is too impractical to be worth it. To name a few drawbacks:

1. We are using more than one element or, at the very least, an extra pseudo-element. Ideally, we want to keep it to one element.
2. We don’t have a tool to prototype our blobby amalgamations, so making one is a process of trial and error.
3. We can’t use borders, gradients, or box shadows since they would reveal the element’s outlines.

Multiple backgrounds and SVG filters

This one is an improvement in the Gooey Effect, described here by Lucas Bebber, although I don’t know who first came up with it. In the original effect, several elements can be morphed together like drops of liquid sticking to and flowing out of each other:

CodePen Embed Fallback

It works by first blurring shapes nearby, creating some connected shadows. Then we crank up the contrast, forcing the blur out and smoothly connecting them in the process. Take, for example, this demo by Chris Coyer (It’s from 2014, so more than 10 years ago!):

CodePen Embed Fallback

If you look at the code, you’ll notice Chris uses the filter property along the blur() and contrast() functions, which I’ve also seen in other blob demos. To be specific, it applies blur() on each individual circle and then contrast() on the parent element. So, if we have the following HTML:

<div class="blob"> <div class="subblob"></div> <div class="subblob"></div> <div class="subblob"></div> </div>

…we would need to apply filters and background colors as such:

.blob { filter: contrast(50); background: white; /* Solid colors are necessary */ } .subblob { filter: blur(15px); background: black; /* Solid colors are necessary */ }

However, there is a good reason why those demos stick to white shapes and black backgrounds (or vice versa) since things get unpredictable once colors aren’t contrast-y enough. See it for yourself in the following demo by changing the color. Just be wary: shades get ugly.

CodePen Embed Fallback

To solve this, we will use an SVG filter instead. I don’t want to get too technical on SVG (if you want to, read Luca’s post!). In a nutshell, we can apply blurring and contrast filters using SVGs, but now, we can also pick which color channel we apply the contrast to, unlike normal contrast(), which modifies all colors.

Since we want to leave color channels (R, G and B) untouched, we will only crank the contrast up for the alpha channel. That translates to the next SVG filter, which can be embedded in the HTML:

<svg xmlns="http://www.w3.org/2000/svg" version="1.1" style="position: absolute;"> <defs> <filter id="blob"> <feGaussianBlur in="SourceGraphic" stdDeviation="12" result="blur" /> <feColorMatrix in="blur" mode="matrix" values="1 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 18 -6" result="goo" /> <feBlend in="SourceGraphic" in2="blob" /> </filter> </defs> </svg>

To apply it, we will use again filter, but this time we’ll set it to url("#blob"), so that it pulls the SVG from the HTML.

.blob { filter: url("#blob"); }

And now we can even use it with gradient backgrounds!

CodePen Embed Fallback

That being said, this approach comes with two small, but important, changes to common CSS filters:

1. The filter is applied to the parent element, not the individual shapes.
2. The parent element must be transparent (which is a huge advantage). To change the background color, we can instead change the body or other ancestors’ background, and it will work with no issues.

What’s left is to place the .subblob elements together such that they make a blobby enough shape, then apply the SVG filters to morph them:

CodePen Embed Fallback Making it one element

This works well, but it has a similar issue to the blob we made by morphing several border-radius instances: too many elements for a simple blob. Luckily, we can take advantage of the background property to create multiple shapes and morph them together using SVG filters, all in a single element. Since we are keeping it to one element, we will go back to just one empty .blob div:

<div class="blob"></div>

To recap, the background shorthand can set all background properties and also set multiple backgrounds at once. Of all the properties, we only care about the background-image, background-position and background-size.

First, we will use background-image along with radial-gradient() to create a circle inside the element:

body { background: radial-gradient(farthest-side, var(--blob-color) 100%, #0000); background-repeat: no-repeat; /* Important! */ } CodePen Embed Fallback

Here is what each parameter does:

• farthest-side: Confines the shape to the element’s box farthest from its center. This way, it is kept as a circle.
• var(--blob-color) 100%: Fills the background shape from 0 to 100% with the same color, so it ends up as a solid color.
• #0000: After the shape is done, it makes a full stop to transparency, so the color ends.

The next part is moving and resizing the circle using the background-position and background-size properties. Luckily, both can be set on background after the gradient, separated from each other by a forward slash (/).

body { background: radial-gradient(...) 20% 30% / 30% 40%; background-repeat: no-repeat; /* Important! */ } CodePen Embed Fallback

The first pair of percentages sets the shape’s horizontal and vertical position (taking as a reference the top-left corner), while the second pair sets the shape’s width and height (taking as a reference the element’s size).

As I mentioned, we can stack up different backgrounds together, which means we can create as many circles/ellipses as we want! For example, we can create three ellipses on the same element:

.blob { background: radial-gradient(farthest-side, var(--blob-color) 100%, #0000) 20% 30% / 30% 40%, radial-gradient(farthest-side, var(--blob-color) 100%, #0000) 80% 50% / 40% 60%, radial-gradient(farthest-side, var(--blob-color) 100%, #0000) 50% 70% / 50% 50%; background-repeat: no-repeat; }

What’s even better is that SVG filters don’t care whether shapes are made of elements or backgrounds, so we can also morph them together using the last url(#blob) filter!

CodePen Embed Fallback

While this method may be a little too much for blobs, it unlocks squishing, stretching, dividing, and merging blobs in seamless animations.

Again, all these tricks are awesome, but not enough for what we want! We accomplished reducing the blob to a single element, but we still can’t use gradients, borders, or shadows on them, and also, they are tedious to design and model. Then, that brings us to the ultimate blob approach…

Using the shape() function

Fortunately, there is a new way to make blobs that just dropped to CSS: the shape() function!

I’ll explain shape()‘s syntax briefly, but for an in-depth explanation, you’ll want to check out both this explainer from the CSS-Tricks Almanac as well as Temani Afif‘s three-part series on the shape() function, as well as his recent article about blobs.

First off, the CSS shape() function is used alongside the clip-path property to cut elements into any shape we want. More specifically, it uses a verbal version of SVG’s path syntax. The syntax has lots of commands for lots of types of lines, but when blobbing with shape(), we’ll define curves using the curve command:

.blob { clip-path: shape( from X0 Y0, curve to X1 Y1 with Xc1 Yc1, curve to X2 Y2 with Xc21 Yc21 / Xc22 Yc22 /* ... */ ); }

Let’s break down each parameter:

• X0 Y0 defines the starting point of the shape.
• curve starts the curve where X1 Y1 is the next point of the shape, while Xc1 Yc1 defines a control point used in Bézier curves.
• The next parameter is similar, but we used Xc21 Yc21 / Xc22 Yc22 instead to define two control points on the Bézier curve.

I honestly don’t understand Bézier curves and control points completely, but luckily, we don’t need them to use shape() and blobs! Again, shape() uses a verbal version of SVG’s path syntax, so it can draw any shape an SVG can, which means that we can translate the SVG blobs we generated earlier… and CSS-ify them. To do so, we’ll grab the d attribute (which defines the path) from our SVG and paste it into Temani’s SVG to shape() generator.

This is the exact code the tool generated for me:

.blob { aspect-ratio: 0.925; /* Generated too! */ clip-path: shape( from 91.52% 26.2%, curve to 93.52% 78.28% with 101.76% 42.67%/103.09% 63.87%, curve to 44.11% 99.97% with 83.95% 92.76%/63.47% 100.58%, curve to 1.45% 78.42% with 24.74% 99.42%/6.42% 90.43%, curve to 14.06% 35.46% with -3.45% 66.41%/4.93% 51.38%, curve to 47.59% 0.33% with 23.18% 19.54%/33.13% 2.8%, curve to 91.52% 26.2% with 62.14% -2.14%/81.28% 9.66% ); }

As you might have guessed, it returns our beautiful blob:

CodePen Embed Fallback

Let’s check if it passes our requirements:

1. Yes, they can be made of a single element.
2. Yes, they can also be created in a generator and then translated into CSS.
3. Yes, we can use gradient backgrounds, but due to the nature of clip-path(), borders and shadows get cut out.

Two out of three? Maybe two and a half of three? That’s a big improvement over the other approaches, even if it’s not perfect.

Conclusion

So, alas, we failed to find what I believe is the perfect CSS approach to blobs. I am, however, amazed how something so trivial designing blobs can teach us about so many tricks and new CSS features, many of which I didn’t know myself.

CSS Blob Recipes originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

KelpUI is new library that Chris Ferdinandi is developing, designed to leverage newer CSS features and Web Components. I’ve enjoyed following Chris as he’s published an ongoing series of articles detailing his thought process behind the library, getting deep into his approach. You really get a clear picture of his strategy and I love it.

He outlined his principles up front in a post back in April:

I’m imagining a system that includes…

• Base styles for all of the common HTML elements.
• Loads of utility classes for nudging and tweaking things.
• Group classes for styling more complex UI elements without a million little classes.
• Easy customization with CSS variables.
• Web Components to progressively add interactivity to functional HTML.
• All of the Web Component HTML lives in the light DOM, so its easy to style and reason about.

I’m imagining something that can be loaded directly from a CDN, downloaded locally, or imported if you want to roll your own build.

And that’s what I’ve seen so far. The Cascade is openly embraced and logically structured with Cascade Layers. Plenty of utility classes are included, with extra care put into how they are named. Selectors are kept simple and specificity is nice and low, where needed. Layouts are flexible with good constraints. Color palettes are accessible and sport semantic naming.

Chris has even put a ton of thought into how KelpUI is licensed.

KelpUI is still evolving, and that’s part of the beauty of looking at it now and following Chris’s blog as he openly chronicles his approach. There’s always going to be some opinionated directions in a library like this, but I love that the guiding philosophy is so clear and is being used as a yardstick to drive decisions. As I write this, Chris is openly questioning the way he optimizes the library, demonstrating the tensions between things like performance and a good developer experience.

Looks like it’ll be a good system, but even more than that, it’s a wonderful learning journey that’s worth following.

KelpUI originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

We’ve known it for a few weeks now, but the CSS if() function officially shipped in Chrome 137 version. It’s really fast development for a feature that the CSSWG resolved to add less than a year ago. We can typically expect this sort of thing — especially one that is unlike anything we currently have in CSS — to develop over a number of years before we can get our dirty hands on it. But here we are!

I’m not here to debate whether if() in CSS should exist, nor do I want to answer whether CSS is a programming language; Chris already did that and definitely explained how exhausting that fun little argument can be.

What I am here to do is poke at if() in these early days of support and explore what we know about it today at a pretty high level to get a feel for its syntax. We poke a little harder at it in another post where we’ll look at a more heady real-world example.

Yes, it’s already here!

Conditional statements exist everywhere in CSS. From at-rules to the parsing and matching of every statement to the DOM, CSS has always had conditionals. And, as Lea Verou put it, every selector is essentially a conditional! What we haven’t had, however, is a way to style an element against multiple conditions in one line, and then have it return a result conditionally.

The if() function is a more advanced level of conditionals, where you can manipulate and have all your conditional statements assigned to a single property.

.element { color: if(style(--theme: dark): oklch(52% 0.18 140); else: oklch(65% 0.05 220)); } CodePen Embed Fallback How does if() work?

Well before Chrome implemented the feature, back in 2021 when it was first proposed, the early syntax was like this:

<if()> = if( <container-query>, [<declaration-value>]{1, 2} )

Now we’re looking at this instead:

<if()> = if( [<if-statement>: <result>]*; <if-statement>: <result> ;? )

Where…

• The first <if-statement> represents conditions inside either style(), media(), or supports() wrapper functions. This allows us to write multiple if statements, as many as we may desire. Yes, you read that right. As many as we want!
• The final <if-statement> condition (else) is the default value when all other if statements fail.

That’s the “easy” way to read the syntax. This is what’s in the spec:

<if()> = if( [ <if-branch> ; ]* <if-branch> ;? ) <if-branch> = <if-condition> : <declaration-value>? <if-condition> = <boolean-expr[ <if-test> ]> | else <if-test> = supports( [ <ident> : <declaration-value> ] | <supports-condition> ) media( <media-feature> | <media-condition> ) | style( <style-query> )

A little wordy, right? So, let’s look at an example to wrap our heads around it. Say we want to change an element’s padding depending on a given active color scheme. We would set an if() statement with a style() function inside, and that would compare a given value with something like a custom variable to output a result. All this talk sounds so complicated, so let’s jump into code:

.element { padding: if(style(--theme: dark): 2rem; else: 3rem); }

The example above sets the padding to 2rem… if the --theme variable is set to dark. If not, it defaults to 3rem. I know, not exactly the sort of thing you might actually use the function for, but it’s merely to illustrate the basic idea.

Make the syntax clean!

One thing I noticed, though, is that things can get convoluted very very fast. Imagine you have three if() statements like this:

:root { --height: 12.5rem; --width: 4rem; --weight: 2rem; } .element { height: if( style(--height: 3rem): 14.5rem; style(--width: 7rem): 10rem; style(--weight: 100rem): 2rem; else: var(--height) ); }

We’re only working with three statements and, I’ll be honest, it makes my eyes hurt with complexity. So, I’m anticipating if() style patterns to be developed soon or prettier versions to adopt a formatting style for this.

For example, if I were to break things out to be more readable, I would likely do something like this:

:root { --height: 12.5rem; --width: 4rem; --weight: 2rem; } /* This is much cleaner, don't you think? */ .element { height: if( style(--height: 3rem): 14.5rem; style(--width: 7rem): 10rem; style(--weight: 100rem): 2rem; else: var(--height) ); }

Much better, right? Now, you can definitely understand what is going on at a glance. That’s just me, though. Maybe you have different ideas… and if you do, I’d love to see them in the comments.

Here’s a quick demo showing multiple conditionals in CSS for this animated ball to work. The width of the ball changes based on some custom variable values set. Gentle reminder that this is only supported in Chrome 137+ at the time I’m writing this:

CodePen Embed Fallback The supports() and media() statements

Think of supports() the same way you would use the @supports at-rule. In fact, they work about the same, at least conceptually:

/* formal syntax for @supports */ @supports <supports-condition> { <rule-list> } /* formal syntax for supports() */ supports( [ <ident> : <declaration-value> ] | <supports-condition> )

The only difference here is that supports() returns a value instead of matching a block of code. But, how does this work in real code?

The <ident>: <declaration-value> you see here is, in this case, the property name: property value e.g. display: flex.

Let’s say you want to check for support for the backdrop-filter property, particularly the blur() function. Typically, you can do this with @supports:

/* Fallback in case the browser doesn't support backdrop-filter */ .card { backdrop-filter: unset; background-color: oklch(20% 50% 40% / 0.8); } @supports (backdrop-filter: blur(10px)) { .card { backdrop-filter: blur(10px); background-color: oklch(20% 50% 40% / 0.8); } }

But, with CSS if(), we can also do this:

.card { backdrop-filter: if( supports(backdrop-filter: blur(10px)): blur(10px); else: unset ); }

Note: Think of unset here as a possible fallback for graceful degradation.

That looks awesome, right? Multiple conditions can be checked as well for supports() and any of the supported functions. For example:

.card { backdrop-filter: if( supports(backdrop-filter: blur(10px)): blur(10px); supports(backdrop-filter: invert(50%)): invert(50%); supports(backdrop-filter: hue-rotate(230deg)): hue-rotate(230deg);; else: unset ); }

Now, take a look at the @media at-rule. You can compare and check for a bunch of stuff, but I’d like to keep it simple and check for whether or not a screen size is a certain width and apply styles based on that:

h1 { font-size: 2rem; } @media (min-width: 768px) { h1 { font-size: 2.5rem; } } @media (min-width: 1200px) { h1 { font-size: 3rem; } }

The media() wrapper works almost the same way as its at-rule counterpart. Note its syntax from the spec:

/* formal syntax for @media */ @media <media-query-list> { <rule-list> } /* formal syntax for media() */ media( <media-feature> | <media-condition> )

Notice how at the end of the day, the formal syntax (<media-query>) is the same as the syntax for the media() function. And instead of returning a block of code in @media, you’d have something like this in the CSS inline if():

h1 { font-size: if( media(width >= 1200px): 3rem; media(width >= 768px): 2.5rem; else: 2rem ); } Again, these are early days

As of the time of this writing, only the latest update of Chrome supports if()). I’m guessing other browsers will follow suit once usage and interest come in. I have no idea when that will happen. Until then, I think it’s fun to experiment with this stuff, just as others have been doing:

• The What If Machine: Bringing the “Iffy” Future of CSS into the Present (Lee Meyer)
• How To Correctly Use if() In CSS (Temani Afif)
• Future-Proofing Indirect Cyclical Conditions (Roma Komarov)
• The new if() function in CSS has landed in the latest Chrome (Amit Merchant)

Experimenting with early features is how we help CSS evolve. If you’re trying things out, consider adding your feedback to the CSSWG and Chromium. The more use cases, the better, and that will certain help make future implementations better as well.

Now that we have a high-level feel for the if()syntax, we’ll poke a little harder at the function in another article where we put it up against a real-world use case.

Lightly Poking at the CSS if() Function in Chrome 137 originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.

In a previous article, I showed you how to refactor the Resize Observer API into something way simpler to use:

// From this const observer = new ResizeObserver(observerFn) function observerFn (entries) { for (let entry of entries) { // Do something with each entry } } const element = document.querySelector('#some-element') observer.observe(element); // To this const node = document.querySelector('#some-element') const obs = resizeObserver(node, { callback({ entry }) { // Do something with each entry } })

Today, we’re going to do the same for MutationObserver and IntersectionObserver.

Refactoring Mutation Observer

MutationObserver has almost the same API as that of ResizeObserver. So we can practically copy-paste the entire chunk of code we wrote for resizeObserver to mutationObserver.

export function mutationObserver(node, options = {}) { const observer = new MutationObserver(observerFn) const { callback, ...opts } = options observer.observe(node, opts) function observerFn(entries) { for (const entry of entries) { // Callback pattern if (options.callback) options.callback({ entry, entries, observer }) // Event listener pattern else { node.dispatchEvent( new CustomEvent('mutate', { detail: { entry, entries, observer }, }) ) } } } }

You can now use mutationObserver with the callback pattern or event listener pattern.

const node = document.querySelector('.some-element') // Callback pattern const obs = mutationObserver(node, { callback ({ entry, entries }) { // Do what you want with each entry } }) // Event listener pattern node.addEventListener('mutate', event => { const { entry } = event.detail // Do what you want with each entry })

Much easier!

Disconnecting the observer

Unlike ResizeObserver who has two methods to stop observing elements, MutationObserver only has one, the disconnect method.

export function mutationObserver(node, options = {}) { // ... return { disconnect() { observer.disconnect() } } }

But, MutationObserver has a takeRecords method that lets you get unprocessed records before you disconnect. Since we should takeRecords before we disconnect, let’s use it inside disconnect.

To create a complete API, we can return this method as well.

export function mutationObserver(node, options = {}) { // ... return { // ... disconnect() { const records = observer.takeRecords() observer.disconnect() if (records.length > 0) observerFn(records) } } }

Now we can disconnect our mutation observer easily with disconnect.

const node = document.querySelector('.some-element') const obs = mutationObserver(/* ... */) obs.disconnect() MutationObserver’s observe options

In case you were wondering, MutationObserver’s observe method can take in 7 options. Each one of them determines what to observe, and they all default to false.

• subtree: Monitors the entire subtree of nodes
• childList: Monitors for addition or removal children elements. If subtree is true, this monitors all descendant elements.
• attributes: Monitors for a change of attributes
• attributeFilter: Array of specific attributes to monitor
• attributeOldValue: Whether to record the previous attribute value if it was changed
• characterData: Monitors for change in character data
• characterDataOldValue: Whether to record the previous character data value

Refactoring Intersection Observer

The API for IntersectionObserver is similar to other observers. Again, you have to:

1. Create a new observer: with the new keyword. This observer takes in an observer function to execute.
2. Do something with the observed changes: This is done via the observer function that is passed into the observer.
3. Observe a specific element: By using the observe method.
4. (Optionally) unobserve the element: By using the unobserve or disconnect method (depending on which Observer you’re using).

But IntersectionObserver requires you to pass the options in Step 1 (instead of Step 3). So here’s the code to use the IntersectionObserver API.

// Step 1: Create a new observer and pass in relevant options const options = {/*...*/} const observer = new IntersectionObserver(observerFn, options) // Step 2: Do something with the observed changes function observerFn (entries) { for (const entry of entries) { // Do something with entry } } // Step 3: Observe the element const element = document.querySelector('#some-element') observer.observe(element) // Step 4 (optional): Disconnect the observer when we're done using it observer.disconnect(element)

Since the code is similar, we can also copy-paste the code we wrote for mutationObserver into intersectionObserver. When doing so, we have to remember to pass the options into IntersectionObserver and not the observe method.

export function mutationObserver(node, options = {}) { const { callback, ...opts } = options const observer = new MutationObserver(observerFn, opts) observer.observe(node) function observerFn(entries) { for (const entry of entries) { // Callback pattern if (options.callback) options.callback({ entry, entries, observer }) // Event listener pattern else { node.dispatchEvent( new CustomEvent('intersect', { detail: { entry, entries, observer }, }) ) } } } }

Now we can use intersectionObserver with the same easy-to-use API:

const node = document.querySelector('.some-element') // Callback pattern const obs = intersectionObserver(node, { callback ({ entry, entries }) { // Do what you want with each entry } }) // Event listener pattern node.addEventListener('intersect', event => { const { entry } = event.detail // Do what you want with each entry }) Disconnecting the Intersection Observer

IntersectionObserver‘s methods are a union of both resizeObserver and mutationObserver. It has four methods:

• observe: observe an element
• unobserve: stops observing one element
• disconnect: stops observing all elements
• takeRecords: gets unprocessed records

So, we can combine the methods we’ve written in resizeObserver and mutationObserver for this one:

export function intersectionObserver(node, options = {}) { // ... return { unobserve(node) { observer.unobserve(node) }, disconnect() { // Take records before disconnecting. const records = observer.takeRecords() observer.disconnect() if (records.length > 0) observerFn(records) }, takeRecords() { return observer.takeRecords() }, } }

Now we can stop observing with the unobserve or disconnect method.

const node = document.querySelector('.some-element') const obs = intersectionObserver(node, /*...*/) // Disconnect the observer obs.disconnect() IntersectionObserver options

In case you were wondering, IntersectionObserver takes in three options:

• root: The element used to check if observed elements are visible
• rootMargin: Lets you specify an offset amount from the edges of the root
• threshold: Determines when to log an observer entry

Here’s an article to help you understand IntersectionObserver options.

Using this in practice via Splendid Labz

Splendid Labz has a utils library that contains resizeObserver, mutationObserver and IntersectionObserver.

You can use them if you don’t want to copy-paste the above snippets into every project.

import { resizeObserver, intersectionObserver, mutationObserver } from 'splendidlabz/utils/dom' const mode = document.querySelector(‘some-element’) const resizeObs = resizeObserver(node, /* ... */) const intersectObs = intersectionObserver(node, /* ... */) const mutateObs = mutationObserver(node, /* ... */)

Aside from the code we’ve written together above (and in the previous article), each observer method in Splendid Labz is capable of letting you observe and stop observing multiple elements at once (except mutationObserver because it doesn’t have a unobserve method)

const items = document.querySelectorAll('.elements') const obs = resizeObserver(items, { callback ({ entry, entries }) { /* Do what you want here */ } }) // Unobserves two items at once const subset = [items[0], items[1]] obs.unobserve(subset)

So it might be just a tad easier to use the functions I’ve already created for you. &#x1f609;

Shameless Plug: Splendid Labz contains a ton of useful utilities — for CSS, JavaScript, Astro, and Svelte — that I have created over the last few years.

I’ve parked them all in into Splendid Labz, so I no longer need to scour the internet for useful functions for most of my web projects. If you take a look, you might just enjoy what I’ve complied!

(I’m still making the docs at the time of writing so it can seem relatively empty. Check back every now and then!)

Learning to refactor stuff

If you love the way I explained how to refactor the observer APIs, you may find how I teach JavaScript interesting.

In my JavaScript course, you’ll learn to build 20 real life components. We’ll start off simple, add features, and refactor along the way.

Refactoring is such an important skill to learn — and in here, I make sure you got cement it into your brain.

That’s it! Hope you had fun reading this piece!

A Better API for the Intersection and Mutation Observers originally published on CSS-Tricks, which is part of the DigitalOcean family. You should get the newsletter.