10 Easy Ways to Craft More Readable CSS
Always code as if the [person] who ends up maintaining your code will be a violent psychopath who knows where you live. Code for readability. —John Woods
Diving into a large, old piece of CSS typically is neither easy nor pleasurable. I find that the biggest challenges in working with old CSS often lie in understanding the purpose and interactions of the styles.
When styling new elements, we have the entire context of the implementation immediately available, and it is easy to write styles that make sense to us at that very moment. However, in a few weeks or to a fresh pair of eyes, what made a lot of sense at first might end up being a lot more cryptic. Without a clear understanding of the purpose and interactions of the styles, modifying stylesheets can be dangerous, tedious, and cumbersome. Therefore, it is important to communicate enough context so that future developers will be able to grok the code easily and make informed decisions.
At Causes, we have adopted the following practices which we believe have improved the maintainability of our stylesheets, reduced bugs, and increased developer velocity. When you have finished reading this, I hope that you will have a few more tools to help move your codebase toward greater maintainability.
1. Show Your Work
My math teachers would always require me to show my work. And, for good reason—this exposed the steps that led up to my answer, allowing my teacher to observe my thought processes and understand the point at which things went horribly wrong (or magnificently right). In other words, showing my work communicated a greater context with which to understand the reason something was done.
Similarly, styles that show the work behind mysterious numbers communicate more context, making them more readable and maintainable. We can easily apply this when using a relative unit like ems, where we determine the value by taking the target size in pixels divided by the context size in pixels:
ems = target px / context px
Imagine that we want to have an element with a font size of 10px, but defined in ems, inside a container with a font size of 16px. The quick and easy way would be to run through the equation above and simply drop in the result:
However, if someone were to come along a few weeks later and see this strange number, they might be confused about its exact meaning. Using a CSS preprocessor such as Sass, we can use a little math to show our work and communicate more context:
Or, in plain CSS, a little comment can go a long way:
Now, when someone sees this style, it is plain that it was expected to target a 10px font size in a 16px font size context, and it will be easier to make an informed decision as to how to modify it.
Think of our styles as machines with many parts. If the machine isn’t working as we’d like, it will be easier to repair it if we can actually see the moving parts.
2. Two or More, Use a For
Spend some time establishing patterns and defining characteristics generally, to promote reusability. This can be accomplished through setting up presentational classes that can be mixed and matched to achieve a unified visual appearance. It might be a good idea to set up a library of these patterns that you can share with your team or start with a good framework like Typeplate or Bootstrap that kickstart this process for us.
With a CSS preprocessor like Sass, we have some very useful tools that allow us to kick this up a notch, more easily take a DRY approach to CSS, and do more with less code.
Dispel magic numbers with variables. Similar to showing our work, instead
of leaving a hex color like
#00c5cd for somebody to grok, naming it
$turquoise helps others immediately recognize what it is. In this case, the
reusability of variables is just the icing on the cake.
Perform common calculations with functions. Sass has a bunch of built-in functions that can make our lives easier. To fill in the gaps, Sass allows us to define our own functions. For example, at Causes we added a Sass function that facilitates calculations of ems:
1 2 3 4 5 6 7
Use mixins and loops to abstract out common patterns. Packages like Compass and Bourbon give us a lot of mixins right out of the box that make cross-browser compatibility a breeze. So, instead of this mess:
1 2 3 4 5
Similarly, our first pass at a responsive design implementation added a stylesheet on top of all of our styles to collect media queries. However, there was too much distance between the original style and the style being modified in the media query, which resulted in the styles quickly becoming out of sync as developers made modifications. So we sprinkled these styles throughout our code, moving our media queries adjacent to the styles themselves to increase locality, and used mixins to simplify the implementation, improve readability, and reduce repetition:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
However, mixins and loops are a double-edged sword that can easily result in
generating too many styles. Always consider how your styles can be generalized
and reused. There may also be occasions where
@extend makes a lot of sense,
but make sure to use it with placeholders.
3. Use Percentages for (Most) Widths
Have you ever beautifully styled an element, only to come back a few months later to discover that the width needs to be adjusted due to an unrelated change to a container?
I’ve discovered that if I start with percentages for horizontal dimensions (widths, margins, and paddings) instead of pixels, even when I’m not working with a flexible design, my styles are more resilient and require much less babysitting. As a bonus, they tend to work better when dropped into new and unexpected contexts and transition very well to flexible and responsive layouts.
When using Sass, lean heavily on the percentage function to show your work (or leave a comment):
Browsers use a default content-box sizing model to determine the sizes of
elements. This means that the
width style will apply to the content box, and
then the padding, border, and margin get added on top of that. This is a
frequent gotcha when using percentages for widths, causing the element to take
up more space than you intended. Thankfully, we can tell the browser to use a
different box sizing model that calculates the size of an element including its
padding and border. Using the
1 2 3 4 5 6 7 8 9 10
Or, using Compass:
1 2 3 4 5 6
There are circumstances that require some fixed-width styles. For instance, imagine a search box widget that consists of a text input and a button side-by-side. If we want the entire widget to occupy the full width of its container, but the button to be always as narrow as possible, we might want to define the width of the button using a fixed unit and the input using a percentage.
1 2 3 4 5 6
1 2 3 4 5 6 7 8 9 10 11 12 13 14
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
4. Express Intention
At the core of the previous tips is this: write styles that most closely and clearly express the intention of the implementation. This is my single, overarching principle for writing readable CSS.
For example, let’s say we have a
heading and a
blockquote and we want the
heading to be larger and the
blockquote to be smaller than the body copy.
Using pixels, we could write:
1 2 3 4 5 6 7
While this will work pretty well for our small, local implementation, the styles themselves live in isolation and they don’t communicate to the reader how they relate to one another. In this case, using a relative unit like ems allows us to define how these styles relate to each other and communicate that what we really want is for the heading to be twice as big as the body copy and for blockquotes to be slightly smaller than the body copy:
1 2 3 4 5 6 7
Similarly, styles that use
rgba appropriately are often more
intention-expressing and resilient.
1 2 3 4 5 6 7 8
Class and ID names that are consistently formed and semantically meaningful go
along with the spirit of this point. Standardize on how to treat spaces. At
Causes, we use hyphens, but underscores and camelCasing are also reasonable
choices if you are consistent. Tend to pick names that describe the purpose or
meaning of the content, and not how it happens to appear. For instance,
.local-navigation is a better name than
.left-column, especially when you
start using media queries to reflow major elements or decide that your local
navigation should be on the right side one day.
5. Be Less Specific
Ideally, styles behave consistently across your site. Behavioral consistency allows developers to quickly and easily understand the fundamentals of your classes and apply them confidently in new situations.
One way to achieve this consistency is by defaulting to less-specific selectors—write shorter selector chains, or if using Sass use minimal nesting. Less-specific selectors allow you to define general style patterns that are easier to reuse, override where necessary, and may be more performant. Qualifying an ID with any preceding selectors is a big code smell.
This means that we should tend to prefer classes over IDs and avoid
!important. Strive for a flatter structure—the worst I’ve seen is a
stylesheet whose structure directly mirrored the markup it was being applied
to. Overly-qualified selectors are more susceptible to the whims of markup
changes; the structure of our stylesheets should be as markup-agnostic as
possible. Just keep things simple.
Additionally, stylesheets and markup are two sides of the same coin—many styling problems are actually markup problems in disguise, and solving these types of problems on the CSS side means you now have two problems. Sometimes a little markup simplification can go a long way toward fixing a styling issue.
6. Love the Cascade
Reset stylesheets, such as Eric Meyer’s Reset CSS, can be good because
they cancel out all default styles, putting all elements at a common
starting-point. However, it is important to define good base styles for
typographical elements (e.g.
ol) on top of the
defaults so your site can have some consistency and you avoid repetition. When
styling an element, ask yourself how many of those styles can be applied
globally, and when writing a style that overrides another, ask yourself if your
style override can be implemented more generally or moved up the chain.
At Causes, we were using a reset stylesheet but had failed to set up some general, global styles. As a result, each time we wanted to use a new heading on a page, we had to write some styles to make it look like the headings on other types of pages. Not only did this slow development, but over time these styles became inconsistent as some were modified but others were forgotten, leaving us with a subtly fragmented and unmaintainable visual design. So, we refactored these styles into sane defaults that are applied sitewide discovered taht we could move faster and break fewer things.
However, be careful when applying styles to generic container elements (e.g.
nav). For example,
header element can be easily used in many different contexts, just
spans, it might not be a good idea to give them all a certain
background color or font size. This messes up a sane style reset, makes it
difficult to use these general elements in new contexts, and ends up requiring
more complicated overrides that lead to styles with confusing purposes and
7. Leave a Comment
Occasionally, it is not possible to write intention-expressing styles or invent class names that perfectly encapsulate an idea. These situations may easily result in unconventional practices or styles whose purposes are not immediately apparent. Although your code may seem incredibly obvious to you in the moment, that doesn’t mean it will make sense to you or someone else in the future. When I find myself in this situation, I leave a little comment to help explain my intention.
For instance, although !important is best avoided because it violates the
principle of least astonishment and is difficult to override, sometimes
it is necessary such as when overriding an inline style that was added via
leave a comment explaining why it is important. This helps prevent a future,
well-intentioned developer from coming along and “fixing” your code. Bonus
points for leaving findable keywords in comments, which help people identify
and remove unusual styles as they are no longer needed.
I typically prefer same-line comments for this purpose because they more easily stay attached to the style they reference. Sass provides one-line comments:
Standard CSS can still use block-style comments:
8. Maintain Order
Similar to structuring our DOM with an eye toward hierarchy, the order of our selectors should represent some order of hierarchy or use. For instance, when using Sass at Causes, we follow this order:
- Properties (styling the element)
- Chained pseudo-selectors (styling states of the element)
- Chained selectors (styling variants of the element)
- Nested selectors (styling children of the element)
This way, the styles that are most closely related to the element are vertically closer to the selector itself.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
We also alphabetize our properties. Although it may seem like lukewarm OCD, alphabetizing helps us
easily spot duplicates and assists developers in finding the properties they
want to modify, thanks to a standard ordering. As a side-benefit, according to
Google, properties that consistently appear in the same order will compress
better when gzipped. Many modern code editors provide a sorting
functionality, which makes alphabetizing styles a snap. (In Vim command mode I
9. Use Whitespace Effectively
Whitespace allows us to more clearly see where one selector set ends and another begins. Indentation helps communicate the hierarchy. Standardize on these things across your codebase for maximum readability.
You should be minifying and gzipping your CSS already, so there is no penalty for blank lines in your code.
At Causes, we indent with two spaces, leave a blank line between selector sets of the same indentation level, and give each selector its own line. These coding styles not only improve readability of the styles on the filesystem, they also simplify diffs which eases code review and merge conflict resolution.
10. Enforce Coding Style
Maintaining a readable codebase can be challenging if not everyone on your team has fully bought into these ideas. To help these tactics permeate our team and persist throughout our codebase, we lean heavily on automated linting and code review.
A linter serves as a low-cost first pass to catch most of the basic issues. At Causes, we are developing a SCSS linter (pull requests welcome) that runs as one of our Git pre-commit hooks. We have trained this linter to enforce basic coding styles such as using one line per selector and alphabetizing properties. Also, a tool like EditorConfig might be worth taking for a spin.
For problems that are more nuanced and difficult to detect through static analysis, such as how well a set of styles expresses the intention of the implementation, code review serves as a valuable second pass. When reviewing stylesheet changes, I try to identify styles whose purpose is not immediately apparent and explain some small steps that can be taken to make it better.
I hope that this article has informed your underlying philosophy and will assist you on your journey of writing beautiful code. There are plenty of other relevant areas that I did not touch upon, such as the structure and organization of your files, but I hope that what I did cover can be part of a strong and sustainable foundation for your stylesheets.
CSS is like storytelling. There are often many ways to achieve a desired outcome; some are more coherent than others.