Introduction
Like any other programming language, CSS has functions. They can be inserted where you’d place a value, or in some cases, accompanying another value declaration. Some CSS functions even let you nest other functions within them!
More
In programming, functions are a named portion of code that performs a specific task. An example of this could be a function written in JavaScript called sayWoof():
function sayWoof() {
console.log("Woof!");
}
We can use this function later in our code, after we have defined our desired behavior. For this example, any time you type sayWoof()
in your website or web app’s JavaScript it will print “Woof!” into the browser’s console.
Functions can also use arguments, which are slots for things like numbers or bits of text that you can feed into the function’s logic to have it modify them. It works like this in JavaScript:
function countDogs(amount) {
console.log("There are " + amount + " dogs!");
}
Here, we have a function called countDogs()
that has an argument called amount
. When a number is provided for amount, it will take that number and add it to a pre-specified sentence. This lets us create sentences that tell us how many dogs we’ve counted.
countDogs(3); // There are 3 dogs!
countDogs(276); // There are 276 dogs!
countDogs("many"); // There are many dogs!
Some programming languages come with baked-in functions to help prevent you from having to reinvent the wheel for every new project. Typically, these functions are built to help make working with the main strengths and features of the language easier.
Take libraries, for example. Libraries are collections of opinionated code made to help make development faster and easier, effectively just curated function collections — think FitVids.js for creating flexible video elements.
Basics of CSS Functions
![Anatomy of a CSS declaration. Inside of a selector class called .selector there is a declaration of background-image: url(‘parchment.jpg’); Arrows label the property (background-image), the function (url), and the argument (parchment.jpg).](png/lv2oi0tm24e6.png)
Unlike other programming languages, we cannot create our own functions in CSS, per se. That kind of logic is reserved for CSS selectors, which allow you to create powerful conditional styling rules.
As opposed to other programming languages — where the output of a function typically invisibly affects other logic down the line — the output of CSS functions are visual in nature. This output is used to control both layout and presentation of content. For example:
.has-orange-glow {
filter: drop-shadow(0.25rem 0 0.75rem #ef9035);
}
The CSS filter function drop-shadow()
uses the arguments we provide it to create an orange outer glow effect on whatever it is applied to.
In the following demo, I have a JavaScript function named toggleOrangeGlow
that toggles the application of the class .has-orange-glow
on the CSS-Tricks logo. Combining this with a CSS transition, we’re able to create a cool glowing effect:
You may be familiar with some CSS functions, but the language has a surprisingly expansive list!
Much like any other technology on the web, different CSS functions have different levels of browser support. Make sure you research and test to ensure your experience works for everyone, and use things like @supports
to provide quality alternate experiences.
Common CSS Functions
url()
.el {
background: url(/images/image.jpg);
}
Using url()
url()
allows you to link to other resources to load them. This can include images, fonts, and even other stylesheets. For performance reasons, it’s good practice to limit the things you load via url()
, as each declaration is an additional HTTP request.
attr()
/* <div data-example="foo"> */
div {
content: attr(data-example);
}
Using attr()
This function allows us to reach into HTML, snag an attribute’s content, and feed it to the CSS content property. You’ll commonly see attr()
used in print stylesheets, where it is used to show the URL of a link after its text. Another great application of this function is using it to show the alt description of an image if it fails to load.
calc()
.el {
width: calc(100vw - 80px);
}
Using calc()
If there’s one function you should spend some time experimenting with, it’s calc()
. We have a complete guide just on calc()
.
This function takes two arguments and calculates a result from the operator (+
, -
, *
, /
) you supply it, provided those arguments are numbers with or without an accompanying unit.
Unlike CSS preprocessors such as Sass, calc()
can mix units, meaning you can do things like subtract 6rem
from 100%
. calc()
is also updated on the fly, so if that 100%
represents a width, it’ll still work if that width changes. calc()
can also accept CSS Custom Properties as arguments, allowing you an incredible degree of flexibility.
lang()
p:lang(en) {
quotes: "\201C" "\201D" "\2018" "\2019" "\201C" "\201D" "\2018" "\2019";
}
Using lang()
Including a lang
attribute in your HTML is a really important thing to do. When present in your HTML, you’re able to use the lang()
function to target the presence of the attribute’s value and conditionally apply styling based on it.
One common use for this selector is to set language-specific quotes, which is great for things like internationalization.
Clever designers and developers might also use it as a hook for styling translated versions of their sites, where cultural and/or language considerations mean there’s different perceptions about things like negative space.
:not()
h3:not(:first-child) {
margin-top: 0;
}
Using not()
This pseudo-class selector will select anything that isn’t what you specify. For example, you could target anything that isn’t an image with body:not(img)
. While this example is dangerously powerful, scoping :not()
to more focused selectors such as BEM’s block class can give you a great deal of versatility.
Currently, :not()
supports only one selector for its argument, but support for multiple comma-separated arguments (e.g. div:not(.this, .that)
) is being worked on!
CSS Custom Properties
There is only one function specific to CSS custom properties, but it makes the whole thing tick!
The var()
function is used to reference a custom property declared earlier in the document.
html {
--color: orange;
}
p {
color: var(--color);
}
It is incredibly powerful when combined with calc()
.
html {
--scale: 1.2;
--size: 0.8rem;
}
.size-2 {
font-size: calc(var(--size) * var(--scale));
}
.size-2 {
font-size: calc(var(--size) * var(--scale) * var(--scale));
}
More on using var()
Another example of this is declaring a custom property called --ratio: 1.618
; in the root of the document, then invoking it later in our CSS to control line-height, like line-height: var(--ratio);
.
Here, var()
is a set of instructions that tells the browser, “Go find the argument called --ratio
declared earlier in the document, take its value, and apply it here.”
Remember! calc()
lets us dynamically adjust things on the fly, including the argument you supply via var()
.
This allows us to create things like modular scale systems directly in CSS with just a few lines of code. If you change the value of --ratio
, the whole modular scale system will update to match.
In the following CodePen demo, I’ve done exactly just that. Change the value of --scale
in the Pen’s CSS to a different number to see what I mean:
It’s also worth mentioning that JavaScript’s setProperty
method can update custom properties in real time. This allows us to quickly and efficiently make dynamic changes to things that previously might have required a lot of complicated code to achieve.
Color Functions
Another common place you see CSS functions is when working with color.
rgb()
and rgba()
.el {
color: rgb(255, 0, 0);
color: rgba(255, 0, 0, 0.5);
color: rgb(255 0 0 / 0.5);
}
Using rgb()
and rgba()
These functions allow you to use numbers to describe the red (r
), green (g
), blue (b
), and alpha (a
) levels of a color. For example, a red color with a hex value of #fb1010
could also be described as rgba(251, 16, 16, 1)
. The red value, 251
, is far higher than the green and blue values (16
and 16
), as the color is mostly comprised of red information.
The alpha value of 1 means that it is fully opaque, and won’t show anything behind what the color is applied to. If we change the alpha value to be 0.5, the color will be 50% transparent. If you use an rgb()
function instead of rgba()
, you don’t have to supply an alpha value. This used to mean you couldn’t supply an alpha value, but that function will take one now whether you use the old comma-syntax or the new slash-syntax.
hsl()
and hsla()
.el {
background: hsl(100, 100%, 50%);
background: hsla(100, 100%, 50%, 0.5);
background: hsl(100 100% 50% / 0.5);
}
Using hsl()
and hsla()
Similar to rgb()
and rgba()
, hsl()
and hsla()
are functions that allow you to describe color. Instead of using red, green, and blue, they use hue (h
), saturation (s
), and lightness (l
).
I prefer using hsla()
over rgba()
because its model of describing color works really well with systematized color systems. Each of the color level values for these functions can be CSS Custom Properties, allowing you to create powerful, dynamic code.
New Color Functions
In the upcoming CSS Color Module Level 4 spec, we can ignore the a
portion of rgba()
and hsla()
, as well as the commas. Now, spaces are used to separate the rgb
and hsl
arguments, with an optional /
to indicate an alpha level.
We’ll also start seeing new functions like lab()
and lch()
that will use this new format
Pseudo Class Selector Functions
These selectors use specialized argument notation that specifies patterns of what to select. This allows you to do things like select every other element, every fifth element, every third element after the seventh element, etc.
Pseudo class selectors are incredibly versatile, yet often overlooked and under-appreciated. Many times, a thoughtful application of a few of these selectors can do the work of one or more node packages.
:nth-child()
.el:nth-child(3n) {
background-color: #eee;
}
nth-child()
allows you to target one or more of the elements present in a group of elements that are on the same level in the Document Object Model (DOM) tree.
![](png/ysqyqnhdf670.png)
In the right hands, :nth-child()
is incredibly powerful. You can even solve fizzbuzz with it! If you’re looking for a good way to get started, Chris has a collection useful pseudo selector recipes.
:nth-last-child()
.el:nth-last-child(2) {
opacity: 0.75;
}
.el:last-child {
opacity: 0.5;
}
This pseudo class selector targets elements in a group of one or more elements that are on the same level in the DOM. It starts counting from the last element in the group and works backwards through the list of available DOM nodes.
Demo
:nth-of-type()
h2:nth-of-type(odd) {
text-indent: 3rem;
}
:nth-of-type()
matches a specified collection of elements of a given type. For example, a declaration of img:nth-of-type(5)
would target the fifth image on a page.
Demo
:nth-last-of-type()
section:nth-last-of-type(3) {
background-color: darkorchid;
}
This pseudo class selector can target an element in a group of elements of a similar type. Much like :nth-last-child()
, it starts counting from the last element in the group. Unlike :nth-last-child
, it will skip elements that don’t apply as it works backwards.
Demo
Animation Functions
Animation is an important part of adding that certain je ne sais quoi to your website or web app. Just remember to put your users’ needs first and honor their animation preferences.
Creating animations also requires controlling the state of things over time, so functions are a natural fit for making that happen.
cubic-bezier()
.el {
transition-timing-function:
cubic-bezier(0.17, 0.67, 0.83, 0.67);
}
Instead of keyword values like ease
, ease-in-out
, or linear
, you can use cubic-bezier()
to create a custom timing function for your animation. While you can read about the math that powers cubic beziers, I think it’s much more fun to play around with making one instead.
![A custom cubic bezier curve created on cubic-bezier.com. There are also options to preview and compare your curve with CSS’s ease, linear, ease-in, ease-out, and ease-in-out transitions.](png/k3gttbgo3a5b.png)
path()
.clip-me {
clip-path: path('M0.5,1 C0.5,1,0,0.7,0,0.3 A0.25,0.25,1,1,1,0.5,0.3 A0.25,0.25,1,1,1,1,0.3 C1,0.7,0.5,1,0.5,1 Z');
}
.move-me {
offset-path: path("M56.06,227 ...");
}
This function is paired with the offset-path
property (or eventually, the clip-path
property). It allows you to “draw” a SVG path that other elements can be animated to follow.
Demo
Both Michelle Barker and Dan Wilson have published excellent articles that go into more detail about this approach to animation.
steps()
.el {
animation: 2s infinite alternate steps(10);
}
This relatively new function allows you to set the easing timing across an animation, which allows for a greater degree of control over what part of the animation occurs when. Dan Wilson has another excellent writeup of how it fits into the existing animation easing landscape.
Sizing & Scaling (Transform) Functions
One common thing we do with animation is stretch and squash stuff. The following functions allow you to do exactly that. There is a catch, however: These CSS functions are a special subset, in that they can only work with the transform
property.
scaleX()
, scaleY()
, scaleZ()
, scale3d()
, and scale()
.double {
transform: scale(2);
}
Scaling functions let you increase or decrease the size of something along one or more axes. If you use scale3d()
you can even do this in three dimensions!
translateX()
, translateY()
, translateZ()
, translate3d()
, and translate()
.center {
position: absolute;
top: 50%;
left: 50%;
transform: translate(-50%, -50%);
}
Translate functions let you reposition an element along one or more axes. Much like scale functions, you can also extend this manipulation into three dimensions.
perspective()
.cube {
transform: perspective(50em) rotateY(50deg)
}
This function lets you adjust the appearance of an object to make it look like it is projecting up and out from its background.
rotateX()
, rotateY()
, rotateZ()
, rotate3d()
, and rotate()
.avatar {
transform: rotate(25deg);
}
Rotate functions let you swivel an element along one or more axes, much like grasping a ball and turning it around in your hand.
skewX()
, skewY()
, and skew()
.header {
transform: skew(25deg, 15deg);
}
Skew functions are a little different from scaling and rotation functions in that they apply a distortion effect relative to a single point. The amount of distortion is proportionate to the angle and distance declared, meaning that the further the effect continues in a direction the more pronounced it will be.
Jorge Moreno also did us all a favor and made a great tool called CSS Transform Functions Visualizer. It allows you to adjust sizing and scaling in real time to better understand how all these functions work together:
As responsible web professionals, we should be mindful of our users and the fact that they may not be using new or powerful hardware to view our content. Large and complicated animations may slow down the experience, or even cause the browser to crash in extreme scenarios.
To prevent this, we can use techniques like will-change
to prepare the browser for what’s in store, and the update
media feature to remove animation on devices that do not support a fast refresh rate.
Filter Functions
CSS filter functions are another special subset of CSS functions, in that they can only work with the filter
property. Filters are special effects applied to an element, mimicking functionality of graphics editing programs such as Photoshop.
You can do some really wild things with CSS filter functions, stuff like recreating the effects you can apply to your posts on Instagram!
brightness()
.avatar:hover {
filter: brightness(150%);
}
This function adjusts how, um, bright something appears. Setting it to a low level will make it appear as if it has had a shadow cast over it. Setting it to a high level will blow it out, like an over-exposed photo.
Demo
blur()
.ghost {
filter: blur(50%);
}
If you’re familiar with Photoshop’s Gaussian Blur filter, you know how blur()
works. The more of this you apply, the more indistinct the thing you apply it to will look.
Demo
contrast()
.wow {
filter: contrast(200%);
}
contrast()
will adjust the degree of difference between the lightest and darkest parts of what is applied to.
Demo
grayscale()
.no-color {
filter: grayscale(100%);
}
grayscale()
removes the color information from what it is applied to. Remember that this isn’t an all-or-nothing affair! You can apply a partial grayscale effect to make something look weathered or washed out.
An interesting application of grayscale()
could be lightly applying it to images when dark mode is enabled, to slightly diminish the overall vibrancy of color in a situation where the user may want less eye strain.
Demo
invert()
While invert()
can be used to make something look like a photo negative, my favorite technique is to use it in a inverted colors media query to invert inverted images and video:
@media (inverted-colors: inverted) {
img,
video {
filter: invert(100%);
}
}
This ensures that image and video content looks the way it should, regardless of a user’s expressed browsing mode preferences.
opacity()
.filter-visibility {
filter: opacity(0);
}
This function controls how much of the background is visible through the element (and child elements) the function is applied to.
Demo
An element that has 0% opacity will be completely transparent, although it will still be present in the DOM. If you need to remove an object completely, use other techniques such as the hidden attribute.
saturate()
.full-color {
filter: saturate(100%);
}
Applying this filter can enhance, or decrease the intensity of the color of what it is applied to. Enhancing an image’s saturation is a common technique photographers use to fix underexposed photos.
Demo
sepia()
.is-old-timey {
filter: sepia(1);
}
There are fancier ways to describe this, but realistically it’s a function that makes something look like it’s an old-timey photograph.
Demo
drop-shadow()
.fit-shape-shadow {
filter: drop-shadow(3rem 0 0.5rem #e486da);
}
A drop shadow is a visual effect applied to an object that makes it appear like it is hovering off of the page. There’s a bit of a trick here, in that CSS also allows you to apply drop shadow effects to text and elements. It’s also distinct from the box-shadow property is that it applies drop shadows to the shape of an element rather than the actual box of an element.
Skilled designers and developers can take advantage of this to create complicated visual effects.
Demo
hue-rotate()
.change-color {
filter: hue-rotate(180deg);
}
When a class with a declaration containing hue-rotate() is applied to an element, each pixel used to draw that element will have it’s hue valued shifted by the amount you specify. hue-rotate()
‘s effect is applied to each and every pixel it is applied to, so all colors will update relative to their hue value’s starting point.
This can create a really psychedelic effect when applied to things that contain a lot of color information, such as photos.
Demo
SVG filters
filter()
also lets us import SVGs filters to use to create specialized visual effects. The topic is too complicated to really do it justice in this article — if you’re looking for a good starting point, I recommend “The Art Of SVG Filters And Why It Is Awesome” by Dirk Weber.
![The word “West!” rendered in a Wild West-style font, with layered teal drop shadows giving it a 3D effect. Behind it is a purple starburst pattern. Screenshot.](jpg/odiiclhpda8f.jpg)
Comparison Functions
The idea with these functions is that we can compare multiple values and apply (or, represent, as the spec says) of the values when they’re evaluated.
clamp()
.page-wrap {
width: clamp(320px, 80%, 1200px);
}
body {
font-size: clamp(12px, 1rem + 2vw, 18px);
}
When providing minimum, maximum, and preferred values as arguments, clamp()
will honor the preferred value so long as it does not exceed the minimum and maximum boundaries.
clamp()
will allow us to author things like components whose size will scale along with the size of the viewport, but won’t shrink or grow past a specific size. This will be especially useful for creating CSS locks, where you can ensure a responsive type size will not get so small that it can’t be read.
max()
and min()
.minimum-of-these {
width: max(500px, 50%);
}
.maximum-of-these {
width: min(320px, 90%);
}
These functions allow you to select either the maximum or minimum value from a range of values you provide. Much like clamp()
, these functions allow us to make things responsive up until a certain point.
Logical Combinations
The spec files these functions under a “Logical Combinations” heading, but it might be easier to think of them as functions for matching selectors.
:is()
and :where()
:is()
has had a bit of an identity crisis. Previously referred to as both matches()
and vendor prefixed as :-webkit-any
/:-moz-any
, it now enjoys a standardized, agreed-upon name. It is a pseudo class selector that accepts a range of selectors as its argument.
This allows an author to group and target a wide range of selectors in an efficient way. :where()
is much like :is()
, only it has a specificity of zero, while the specificity of :is()
is set to the highest specificity in the provided selector list.
:is(section, article, aside, nav) :is(h1, h2, h3, h4, h5, h6) {
color: #BADA55;
}
/* ... which is equivalent to: */
section h1, section h2, section h3, section h4, section h5, section h6,
article h1, article h2, article h3, article h4, article h5, article h6,
aside h1, aside h2, aside h3, aside h4, aside h5, aside h6,
nav h1, nav h2, nav h3, nav h4, nav h5, nav h6 {
color: #BADA55;
}
:is()
and :where()
allow us a good deal of flexibility about how we select things to style, especially for situations where you may not have as much control over the web site or web app’s stylesheet (e.g. third-party integrations, ads, etc.).
#CSS
— Adam Argyle (@argyleink) November 7, 2019
:is() selector 🎉
the successor to :any() and :matches()
sneak peak into our talk, here’s a neat gif I made with XD showing what the :is() selector syntax can do. be excited for Chrome Dev Summit y’all!https://t.co/0r2CcUx9Hv pic.twitter.com/wSuGOsDLvZ
Gradient Functions
Gradients are created when you transition one color to one or more other colors. They are workhorses of modern user interfaces — skilled designers and developers use them to lend an air of polish and sophistication to their work.
Gradient functions allow you to specify a whole range of properties, including:
- Color values,
- The position on the gradient area where that color comes in,
- What angle the gradient is positioned at.
And yes, you guessed it: the colors we use in a gradient can be described using CSS color functions!
linear-gradient()
and repeating-linear-gradient()
Linear gradients apply the color transformation in a straight line, from one point to another — this line can be set at an angle as well. In cases where there’s more area than gradient, using repeating-linear-gradient()
will, er, repeat the gradient you described until all the available area has been filled.
Demo
radial-gradient()
and repeating-radial-gradient()
Radial gradients are a lot like linear gradients, only instead of a straight line, color transformations radiate outward from a center point. They’re oftentimes used to create a semitransparent screen to help separate a modal from the background it is placed over.
Demo
conic-gradient()
and repeating-conical-gradient
Conic gradients are different from radial gradients in that the color rotates around a circle. Because of this, we can do neat things like create donut charts. Unfortunately, support for conic gradients continues to be poor, so use them with caution.
![A teal and red donut chart set to 36%. To the right of the chart is a range slider, also set to 36%. Screenshot.](png/ywuglaxd4c90.png)
Grid Functions
CSS Grid is a relatively new feature of the language. It allows us to efficiently create adaptive, robust layouts for multiple screen sizes.
It’s worth acknowledging our roots. Before Grid, layout in CSS was largely a series of codified hacks to work with a language originally designed to format academic documents. Grid’s introduction is further acknowledgement that the language’s intent has changed.
Modern CSS is an efficient, fault-tolerant language for controlling presentation and layout across a wide range of device form factors. Equipped with Grid and other properties like flexbox, we’re able to create layouts that would have been impossible to create in earlier iterations of CSS.
Grid introduces the following CSS functions to help you use it.
fit-content()
This function “clamps” the size of grid rows or columns, letting you specify a maximum size a grid track can expand to. fit-content()
accepts a range of values, but most notable among them are min-content
and max-content
. These values allow you to tie your layout to the content it contains. Impressive stuff!
Demo
minmax()
minmax()
allows you to set the minimum and maximum desired heights and widths of your grid rows and columns. This function can also use min-content
and max-content
, giving us a great deal of power and flexibility.
Demo
repeat()
You can loop through patterns of grid column and rows using repeat()
. This is great for two scenarios:
- When you do know how many rows or columns you need, but typing them out would be laborious. A good example of this would be constructing the grid for a calendar.
- When you don’t know how many rows or columns you need. Here, you can specify a template that the browser will honor as it propagates content into your layout.
Demo
Shape Functions
Like filter()
and transform()
, shape CSS functions only work with one property: clip-path
. This property is used to mask portions of something, allowing you to create all sorts of cool effects.
circle()
This function creates a circular shape for your mask, allowing you to specify its radius and position.
Demo
ellipse()
Like circle()
, ellipse()
will draw a rounded shape, only instead of a perfect circle, ellipse()
lets you construct an oblong mask.
Demo
polygon()
With polygon()
, you are able to specify an arbitrary number of points, allowing you to draw complicated shapes. polygon()
also takes an optional fill-rule
argument, which specifies which part of the shape is the inside part.
Demo
inset()
This function will mask out a rectangle inside of the element you apply it to.
Demo
Miscellaneous Functions
These are the un-categorizable CSS functions, things that don’t fit neatly elsewhere.
element()
Ever pointed a camera at its own video feed? That’s sort of what element()
does. It allows you to specify the ID of another element to create an “image” of what that element looks like. You can then apply other CSS to that image, including stuff like CSS filters!
It might take a bit to wrap your head around the concept — and it has some support concerns — but element()
is a potentially very powerful in the right hands.
Preethi Sam‘s “Using the Little-Known CSS element() Function to Create a Minimap Navigator” demonstrates how to use it to create a code minimap and is an excellent read.
Here, she’s created a minimap for reading through a longform article:
image-set()
.responsive-background {
background-image:
image-set("image.png" 1x,
"image-2x.png" 2x,
"image-print.png" 600dpi
);
}
This function allows you to specify a list of images for the browser to select for a background image, based on what it knows about the capabilities of its display and its connection speed. It is analogous to what you would do with the srcset
property.
::slotted()
::slotted(.marker) {
background: lightyellow;
}
This is a pseudo-element selector used to target elements that have been placed into a slot inside a HTML template. ::slotted()
is intended to be used when working with Web Components, which are custom, developer-defined HTML elements.
Not Ready for Prime Time
Like any other living programming language, CSS includes features and functionality that are actively being worked on.
These functions can sometimes be previewed using browsers that have access to the bleeding edge. Firefox Nightly and Chrome Canary are two such browsers. Other features and functionality are so new that they only exist in what is being actively discussed by the W3C.
annotation()
This function enables Alternate Annotation Forms, characters reserved for marking up things like notation and annotation. These characters typically will be outlined with a circle, square, or diamond shape.
Not many typefaces contain Alternate Annotation Forms, so it’s good to check to see if the typeface you’re using includes them before trying to get annotation()
to work. Tools such as Wakamai Fondue can help with that.
![he numbers 1 and 2 enclosed in hollow and solid-filled circles. Following them are the letters B and R enclosed in hollow and solid-filled squares. Screenshot.Stylistic Alternates.](png/revvazxkb775.png)
counter()
and counters()
When you create an ordered list in HTML, the browser will automatically generate numbers for you and place them before your list item content. These pieces of browser-generated list content are called counters.
By using a combination of the ::marker
pseudo-element selector, the content property, and the counter()
function, we can control the content and presentation of the counters on an ordered list. For browsers that don’t support counter()
or counters()
yet, you still get a decent experience due to the browser automatically falling back to its generated content:
For situations where you have nested ordered lists, the counters()
function allows a child ordered list to access its parent. This allows us to control their content and presentation. If you want to learn more about the power of ::marker
, counter(), and counters()
, you can read “CSS Lists, Markers, And Counters” by Rachel Andrew.
cross-fade()
This function will allow you to blend one background image into one or more other background images. Its proposed syntax is similar to gradient functions, where you can specify the stops where images start and end.
dir()
This function allows you to flip the orientation of a language’s reading order. For English, that means a left-to-right (ltr
) reading order gets turned into right-to-left (rtl
). Only Firefox currently has support for dir()
, but you can achieve the same effect in Chromium-based browsers by using an attribute selector such as [dir="rtl"]
.
Demo
env()
body {
padding:
env(safe-area-inset-top)
env(safe-area-inset-right)
env(safe-area-inset-bottom)
env(safe-area-inset-left);
}
env()
, short for environment, allows you to create conditional logic that is triggered if the device’s User Agent matches up. It was popularized by the iPhone X as a method to work with its notch.
That being said, device sniffing is a fallacious affair — you shouldn’t consider env()
a way to cheat it. Instead, use it as intended: to make sure your design works for devices that impose unique hardware constraints on the viewport.
has()
has()
is a relational pseudo-class that will target an element that contains another element, provided there is at least one match in the HTML source. An example of this is be a:has(> img)
, which tells the browser to target any link that contains an image.
![A diagram showing how the CSS selector a:has(> img) targets only links that contain images in a collection of links that contain either images or paragraphs.](png/l75srocu2f25.png)
Interestingly, has()
is currently being proposed as CSS you can only write in JavaScript. If I were to wager a guess as to why this is, it is to scope the selector for performance reasons. With this approach has()
is triggered only after the browser has been told to process conditional logic, and therefore query the state of things.
image()
.help::before {
content: image("try.webp", "try.svg", "try.gif");
}
This function will let you insert either a static image (referenced with url()
, or draw one dynamically via gradients and element()
.
Trigonometry functions
These functions will allow us to perform more advanced mathematical operations:
- Sine:
sin()
- Cosine:
cos()
- Tangent:
tan()
- Arccosine:
acos()
- Arcsine:
asin()
- Arctangent:
atan()
- Arctangent:
atan2()
- Square root:
sqrt()
- The square root of the sum of squares of its arguments:
hypot()
- Power:
pow()
I’m especially excited to see what people who are more clever than I am will do with these functions, especially for things like animation!
:host()
and :host-context()
To be honest, I’m a little hazy on the specifics of the jargon and mechanics that power the Shadow DOM. Here’s how the MDN describes host()
:
The
:host()
CSS pseudo-class function selects the shadow host of the shadow DOM containing the CSS it is used inside (so you can select a custom element from inside its shadow DOM) — but only if the selector given as the function’s parameter matches the shadow host.
And here’s what they have to say about :host-context()
:
The
:host-context()
CSS pseudo-class function selects the shadow host of the shadow DOM containing the CSS it is used inside (so you can select a custom element from inside its shadow DOM) — but only if the selector given as the function’s parameter matches the shadow host’s ancestor(s) in the place it sits inside the DOM hierarchy.
:nth-col()
and :nth-last-col()
These pseudo-classes will allow you to select one or a specified series columns in a CSS grid to apply styling to them. A good mental model for how these functions will work is how CSS pseudo class selectors operate. Unlike pseudo class selectors, :nth-col()
and :nth-last-col()
should be able to target implicit grid columns.
symbols()
This function allows you to specify a list of different kinds of characters to use for list bullets. Much like annotation()
, you’ll want to make sure the typeface you use contains a glyph you want to use as a symbol before trying to get symbols()
to work.
Deprecated Functions
Sometimes things just don’t work out the way you think they will. While deprecated CSS functions may still render in the browser for legacy support reasons, it isn’t recommended you use them going forward.
matrix()
and matrix3d()
These functions were turned into more discrete sizing and scaling functions.
rect()
This function was part of the deprecated clip
property. Use the clip-path
property and its values instead.
target-counter()
, target-counters()
, and target-text()
These functions were intended to help work with fragment URLs for paged (printed) media. You can read more about them on the W3C’s CSS Generated Content for Paged Media Module documentation
Typography
The web is typography, so it makes sense to give your type the care and attention it deserves. While CSS provides some functions specifically designed to unlock the potential of your website or webapp’s chosen typefaces, it is advised to not use the following functions to access these advanced features.
Instead, use lower-level syntax via font-feature-settings. You can figure out if the font you’re using supports these features by using a tool such as Wakamai Fondue.
character-variant()
, styleset()
, and stylistic()
Many typefaces made by professional foundries include alternate treatments for certain letters, or combinations of letters. One example use case is providing different variations of commonly-used letters for typefaces designed to look like handwriting, to help make it appear more natural-looking.
![Two examples of the sentence, “Easy Sunday morning & my fox. The first sentence does not have Stylistic Alternates enabled. The second sentence does, with the alternate characters (a, “un”, “m, “rn” g, &, m, f, and x) highlighted in green. Screenshot.](png/cvntsekr5bfc.png)
Utilizing these functions activates these special alternate characters, provided they are present in the font’s glyph set.
Unfortunately, it is not a standardized offering. Different typefaces will have different ranges of support, based on what the typographer chose to include. It would be wise to check to see if the font you’re using supports these special features before writing any code.
format()
When you are importing a font via the url()
function, the format()
function is an optional hint that lets you manually specify the font’s file format. If this hint is provided, the browser won’t download the font if it does not recognize the specified file format.
@font-face {
font-family: 'MyWebFont';
src: url('mywebfont.woff2') format('woff2'), /* Cutting edge browsers */
url('mywebfont.woff') format('woff'), /* Most modern Browsers */
url('mywebfont.ttf') format('truetype'); /* Older Safari, Android, iOS */
}
leader()
You know when you’re reading a menu at a restaurant and there’s a series of periods that help you figure out what price is attached to what menu item? Those are leaders.
The W3C had plans for them with its CSS Generated Content for Paged Media Module, but it unfortunately seems like leader()
never quite managed to take off. Fortunately, the W3C also provides an example of how to accomplish this effect using a clever application of the content
property.
local()
local()
allows you to specify a font installed locally, meaning it is present on the device. Local fonts either ship with the device, or can be manually installed.
Betting on someone installing a font so things look the way you want them to is very risky! Because of this, it is recommended you don’t specify a local font that needs to be manually installed. Your site won’t look the way it is intended to, even moreso if you don’t specify a fallback font.
@font-face {
font-family: 'FeltTipPen';
src: local('Felt Tip Pen Web'), /* Full font name */
local('FeltTipPen-Regular'); /* Postscript name */
}
ornaments()
Special dingbat characters can be enabled using this function. Be careful, as not all dingbat characters are properly coded in a way that will work well if a user does something like change the font, or use a specialized browsing mode.
swash()
Swashes are alternate visual treatments for letters that give them an extra-fancy flourish. They’re commonly found in italic and cursive-style typefaces.
![An example of a swash being applied to a script-style typeface. There’s two versions of the phrase, “Fred And Ginger”. The first version doesn’t have swashes activated. The second example does. In the second example, the letter F, and, and the letter G are highlighted to demonstrate them being activated. Screenshot.](png/fc2d_nlhe513.png)
Why so many?
CSS is maligned as frequently as it is misunderstood. The guiding thought to understanding why all these functions are made available to us is knowing that CSS isn’t prescriptive — not every website has to look like a Microsoft Word document.
The technologies that power the web are designed in such a way that someone with enough interest can build whatever they want. It’s a powerful, revolutionary concept, a large part of why the web became so ubiquitous.
Examples involving dogs are inherently clearer! :]
I am always grateful to css-tricks. This article is written in a great way and helpful for anyone.
Ver good article to understant what functions in CSS are.
But I was hoping for the answer to the question: will we ever be able to write our own functions in CSS?
Possibly! You might want to check out Houdini. It grants access to the CSSOM, which is incredibly powerful stuff.
Very nice article please more of this please
calc()
is absolutely fantastic! That is all.Nice, especially has and other functions that are not supported by any browser: https://developer.mozilla.org/en-US/docs/Web/CSS/:has
re: fit-content() function
what am I missing here? 250 px vs 350 px in your example…
In this example, we are using an argument of 250px for the fit-content() function. This tells the browser to restrict the first grid column to a width of 350px,
I forgot to include in my 1st comment that I think this article is a wonderful piece of work. Hats off to the person(s) who did the work!!
A few notes:
translate[x|y|z|3d]
(translation),rotate[x|y|z|3d]
(rotation),skew[x|y]
(shearing) don’t perform scaling either. They are transform functions, just likematrix
andmatrix3d
, which, as far as I know, are not deprecated. At least the spec says nothing about them being deprecated, so I’m not sure where that info comes from. The source linked in that section gives me a 404. What has been deprecated for almost a decade however is theskew(x, y)
function (spec). It was even removed from Firefox in August 2012, but it was later reintroduced for backwards compatibility reasons. Also on skewing: “amount of distortion is proportionate to the angle and distance declared” – there’s no distance in theskew[x|y]
functions. This section explains how skewing works, as well as the difference between how it’s applied on HTML and SVG elements.And note that stretching and squashing is only a tiny part of what can be done with transform functions.
sqrt()
,pow()
andhypot()
are not trigonometric functions. Together withlog()
andexp()
, they’re classed as exponential functions. There are also sign-related functions (abs()
andsign()
) and stepped value functions (round()
,mod()
,rem()
). Here’s the spec for that.conic-gradient()
actually has decent support now if we are to ignore pre-Chromium Edge. Chromium browsers support it. Safari supports it. Even Firefox supports it behind a flag – go toabout:config
, search for thelayout.css.conic-gradient.enabled
flag and set its value totrue
(it’sfalse
by default, double-clicking sets it totrue
).Same goes for
min()
,max()
andclamp()
. Leaving aside old Edge/ IE, which surely won’t evolve to support them, they are actually well-supported.Thanks for the reply, Ana. I really respect your work and appreciate the time you spent pointing out the above. I’ll try and get updates from your feedback incorporated as soon as possible.
Thanks for this awesome article!
(Gosh CSS functions sure have ballooned since I learned in the early 2000s. This guide brought me up-to-date in just a few minutes.)
Just wanted to point out a little formatting errors under Trigonometry Functions. For the functions
acos()
andhypot()
, the code formatting starts earlier than it should.May The Source be with you!
Thanks for the heads up! And for reading all that way down.