Orienting web visitors using landmarks

· Accessibility, Concepts, Design · Patrick Smith

When you visit a new city, one thing you expect to see are landmarks. Statues, botanical gardens, theaters, skyscrapers, markets. These landmarks help us navigate around unfamiliar areas by being reference points we can see on the horizon or on a map.

As makers of the web, we can also provide landmarks to people. These aren’t arbitrary — there are eight landmarks that are part of the HTML standard:

  • Navigation
  • Main
  • Banner
  • Search
  • Form
  • Content info
  • Region
  • Complementary

Some of these seem obvious, but some are odd — what on earth does “contentinfo” mean? Let’s walk through what they are, why they are important to provide, and finally how we can really easily use them.

Navigation

Nearly all websites have a primary navigation. It’s often presented as a row of links at the top of the page, or under a hamburger menu.

Stripe’s navigation bar at the top of its homepage

Stripe’s navigation provides links to the primary pages people want to visit. It’s clear, and follows common practices of showing only a handful of links, and placing the link to sign in up on the far right.

Most visual users would identify this as the primary navigation of the site, and so you should denote it as such in your HTML markup. Here’s what you might write for Stripe’s navigation:

<nav>
  <ul>
    <li><a href="/"> Stripe </a></li>
    <li><a href="/products"> Products </a></li>
    <li><a href="/developers"> Developers </a></li>
    <li><a href="/about"> Company </a></li>
    <li><a href="/pricing"> Pricing </a></li>
    <li><a href="/support"> Support </a></li>
    <li><a href="/login"> Sign in </a></li>
  </ul>
</nav>

Here we use HTML 5’s <nav> element, which automatically has the navigation landmark.

If there’s only one navigation landmark on a page, then people using screen readers can jump straight to it and the links inside. They can visit Stripe’s Support page in a few seconds. It’s like a city subway that connects key landmarks, allowing fast travel between them.

What if you have multiple navigations? Let’s look at GitHub for an example.

A repository on GitHub with two navigations, one for the whole site, and one specific to my repository.

Here we have a black bar offering links to the main parts of the GitHub experience: my pull requests, my issues, the marketplace, notifications, etc.

But I am on the page for a particular repository, and it also has its own navigation: Code, Issues, Pull requests, Actions, etc.

So how do we offer both? And how do users using screen readers know the difference? By attaching labels to each navigation: the top navigation has the label Global, and the repository specific navigation has the label Repository. It’s like a city having multiple sports stadiums: here in Melbourne we have the MCG (used for football and cricket) and the Rod Laver Arena (used for tennis and music). They clearly have different names to identify them by that means people can find them easily and won’t mix them up.

We can add the labels in our HTML like so:

<header>
  <a href="/">GitHub</a>
  <nav aria-label="Global">
    <ul>
      <li><a href="/pulls"> Pull requests </a></li>
      <li><a href="/issues"> Issues </a></li>
      <li><a href="/marketplace"> Marketplace </a></li>
      <li><a href="/explore"> Explore </a></li>
    </ul>
  </nav>
</header>

<main>
  <h1>RoyalIcing / dovetail</h1>
  <nav aria-label="Repository">
    <ul>
      <li><a href="/RoyalIcing/dovetail" aria-current="page"> Code </a></li>
      <li><a href="/RoyalIcing/dovetail/issues"> Issues </a></li>
      <li><a href="/RoyalIcing/dovetail/pulls"> Pull requests </a></li>
      <li><a href="/RoyalIcing/dovetail/actions"> Actions </a></li>
      <li><a href="/RoyalIcing/dovetail/projects"> Projects </a></li>
      <li><a href="/RoyalIcing/dovetail/wiki"> Wiki </a></li>
      <li><a href="/RoyalIcing/dovetail/network/alerts"> Security </a></li>
      <li><a href="/RoyalIcing/dovetail/pulse"> Insights </a></li>
      <li><a href="/RoyalIcing/dovetail/settings"> Settings </a></li>
    </ul>
  </nav>
</main>

Now people using screen readers or similar browser tools can see that there are two navigation to pick from, one named Global and one Repository.

Note also we have an aria-current="page" attribute on the link that represents the page the user is on. This is equivalent to a 🔴 You Are Here mark on a public map.

Main

When watching a show on Netflix, you’ll often be presented with a Skip intro button. This fasts forwards past the intro content that is often the same every time to the part you want to watch: the new episode.

Imagine if that Skip intro button didn’t exist: what would you do? You could watch the minute-long intro every time. Or you could attempt to fast-forward to the spot where the show actually starts. One is tedious and the other is error-prone. It would be a poor user experience.

On the web, our users might find themselves in the same situation. If they use a screen reader, they’ll probably hear all the items in our navigation and header. And then eventually they’ll reach the headline or the part that’s new — the part they are interested in — just like a TV episode. They could fast-forward, but that also would be error-prone. It would be great if we could allow them to skip past the repetitive stuff to the part they are actually interested in.

Enter <main>. Use this to wrap the part of the page where your ‘episode’ actually starts. People using screen readers can then skip past the tedious navigation and other preambles.

Let’s look at that GitHub example again:

<!-- Logo and navigation that’s on every page -->
<header>
  <a href="/">GitHub</a>
  <nav aria-label="Global">
    <ul>
      <li><a href="/pulls"> Pull requests </a></li>
      <li><a href="/issues"> Issues </a></li>
      <li><a href="/marketplace"> Marketplace </a></li>
      <li><a href="/explore"> Explore </a></li>
    </ul>
  </nav>
</header>

<!-- People can skip to here easily -->
<main>
  <h1>RoyalIcing / dovetail</h1>
  <nav aria-label="Repository">
    <ul>
      <li><a href="/RoyalIcing/dovetail" aria-current="page"> Code </a></li>
      <li><a href="/RoyalIcing/dovetail/issues"> Issues </a></li>
      <li><a href="/RoyalIcing/dovetail/pulls"> Pull requests </a></li>
      <li><a href="/RoyalIcing/dovetail/actions"> Actions </a></li>
      <li><a href="/RoyalIcing/dovetail/projects"> Projects </a></li>
      <li><a href="/RoyalIcing/dovetail/wiki"> Wiki </a></li>
      <li><a href="/RoyalIcing/dovetail/network/alerts"> Security </a></li>
      <li><a href="/RoyalIcing/dovetail/pulse"> Insights </a></li>
      <li><a href="/RoyalIcing/dovetail/settings"> Settings </a></li>
    </ul>
  </nav>
</main>

By using <main> we have allowed users to skip the intro.

Banner

We’ve already talked about the top strip on most websites, and these also have a role. Banners hold the primary navigation and also: logo, search field, notifications, profile, or other site-wide shortcuts. The banner often acts as the consistent branding across all pages.

The banner on GitHub when I’m signed in

Here’s GitHub’s banner when I’m signed in. The part I’ve highlighted with the yellow outline is the navigation (using <nav>). The entire element uses <header>, which automatically gains the role of banner if it meets the following (via MDN):

Assistive technologies can identify the main header element of a page as the banner if is a descendant of the body element, and not nested within an article, aside, main, nav or section subsection.

So the following <header> has the role of banner:

<body>
  <header> <!-- This gains the banner role -->
    …
  </header>

  <main>
    <h1>Heading</h1>
    …
  </main>
</body>

While this one doesn’t:

<body>
  <main>
    <header> <!-- This does not gain the banner role -->
      <h1>Heading</h1>
    </header>
    …
  </main>
</body>

And you can use multiple <header> elements for things other than banners, if you nest them inside “article, aside, main, nav or section” as MDN mentions.

Because of this, I might recommend that you add the banner role explicitly, as it will make it easier to identify and also target with CSS (e.g. header[role=banner] selector).

<body>
  <header role="banner"> <!-- Add the role explicitly -->
    …
  </header>

  <main>
    <header> <!-- Because this is nested inside <main>, it won’t gain the banner role -->
      <h1>Heading</h1>
    </header>
    …
  </main>
</body>

Banner’s don’t necessarily have to be a horizontal strip. Twitter has a vertical banner:

My Twitter profile, with a vertical banner & navigation on the left side

The banner here is the entire left hand column containing Home, Explore, etc. It’s also implemented with a <header role="banner">. The HTML 5 elements are named more for their concept that their visual intention.

Search

Search is one of the things that makes the web great. You have an idea of what you are looking for, you type it in, and in seconds you’ll likely be shown it.

Here’s an excerpt of Google’s famous homepage:

Google’s famous homepage (circa April 2020)
<form class="tsf nj" action="/search" style="overflow:visible" data-submitfalse="q" id="tsf" method="GET" name="f" role="search">
…
</form>

See that role="search"? This identifies it as the primary search on the page. If someone needs, they can jump straight to the search form and start typing.

Similarly here’s GitHub search form:

A screenshot of GitHub’s banner with search form
<form class="js-site-search-form" role="search" aria-label="Site" data-scope-type="Repository" data-scope-id="203948748" data-scoped-search-url="/RoyalIcing/dovetail/search" data-unscoped-search-url="/search" action="/RoyalIcing/dovetail/search" accept-charset="UTF-8" method="get">
…
</form>

Again we see a <form> with role="search". If you decide to add a search form to your site, make sure it has the search role.

Form

If you have another form not used for search, say for signing in or creating a new document, then the form role helps out here. The built-in <form> element actually already has the form role implicitly. So what’s left to do?

First, ensure it is labelled so people know what the form is for. That way if there’s multiple forms on a page, they can tell them apart. Also, people can jump straight to the form and start filling it out.

This form on GitHub to create a new repo is labelled Create a new repository.

You can add a label by adding an aria-label attribute (note: avoid title):

<form aria-label="Create a new repository">
  <h2>Create a new repository</h2>
  …
<form>

Or by identifying which heading acts as the form’s label:

<form aria-labelledby="new-repo-heading">
  <h2 id="new-repo-heading">Create a new repository</h2>
  …
<form>

Note in both cases we still have a heading — your forms should probably have a label that is readable by all users, not just those using assistive-tech.

Content info

Ok, so the names have been pretty logical so far. And then we come to contentinfo. What on earth does that mean?

Let’s show some examples of where contentinfo has been used in the wild:

GitHub’s footer with copyright and links to Terms, Privacy, Blog, etc.
Mozilla’s footer with many columns of links, copyright, and legalise.

It’s a footer! With lots of links. And a copyright.

Akin to the banner role and its automatic provider <header>, we can use <footer>:

<body>
  <main>
    …
    <footer> <!-- Because this is nested inside <main>, it won’t gain the contentinfo role -->
      …
    </footer>
  </main>

  <footer role="contentinfo"> <!-- Add the role explicitly -->
    …
  </footer>
</body>

And also like <header>, it only gains the role if it’s a direct child of <body>. However, it’s recommended that you add role="contentinfo" explicitly to the desired element due to long running issues with Safari and Voice Over.

Complementary

Hierarchy is a core principle of visual design. Some parts of a design will be more important than others, and so it is important that the reader is aware of what they should draw their attention to, and what is less important.

Visual users are aided by size, layout, contrast — and so we need a semantic approach too for non-visual users. This might be a user using a screen-reader. Or it might be a search engine’s web crawler, or someone using the reader view available in Safari and Firefox.

A simple hierarchical relationship is primary content supported by complementary content. Some examples of these are:

  • Footnotes to an article
  • Links to related content
  • Pull quotes
  • Comments on a post

Here’s an example article with footnotes, pull quotes, and related links:

<main>
  <h1>Why penguins can’t fly</h1>

  <article>
    <p>Penguins are … </p>
    <p>Their feathers … </p>
    <aside>
      <blockquote>
        <p>
          Penguins swim fast due to air bubbles trapped in their feathers<sup><a href="#footnote-1">1</a></sup>
        </p>
      </blockquote>
    </aside>
    <p>Speeds of … </p>
    <p>They eat … </p>
    
    <aside>
      <h2>Footnotes</h2>
      <ol>
        <li id="footnote-1">
          <a href="https://www.nationalgeographic.com/magazine/2012/11/emperor-penguins/">National Geographic: Escape Velocity</a>
        <li id="footnote-2">…
        <li id="footnote-3">…
      </ol>
    </aside>
  </article>
  
  <aside>
    <h2>Related articles</h2>
    <ul>
      <li><a href="…">…</a>
      <li><a href="…">…</a>
      <li><a href="…">…</a>
    </ul>
  </aside>
</main>

Region

We have covered seven landmarks — what’s left? The generic landmark of region. Use it as a last resort — first reach for one of the above landmarks.

Again, HTML 5 helps us out here: we can use <section>. It’s important that you add an aria-label attribute (or aria-labelledby) to name the landmark, so a user knows why it is important and can tell it apart from other landmarks.

At Smashing Magazine, they used a labelled section for their article intro:

<main>
  <article>
    <section aria-label="quick summary">
      In this Smashing TV webinar recording, join Léonie Watson (a blind screen reader user) as she explores the web…
    </section>
    <p>…</p>
    <p>…</p>
    <p>…</p>
  </article>
</main>

This allowed Léonie (who suggested the change) to identify the summary, and skip it if she liked.

Remember, use navigation, banner, contentinfo roles (<nav>, <header>, <footer>) before using region. The HTML spec suggests for using sections:

Examples of sections would be chapters, the various tabbed pages in a tabbed dialog box, or the numbered sections of a thesis. A Web site’s home page could be split into sections for an introduction, news items, and contact information.

https://html.spec.whatwg.org/dev/sections.html#the-section-element

Article

We’ve been using <article> in some of the examples previously — is this also a landmark? The answer is technically no, but more or less yes. Bruce Lawson goes into detail on why you should use <article> over <section>:

So a homepage with a list of blog posts would be a <main> element wrapping a series of <article> elements, one for each blog post. You would use the same structure for a list of videos (think YouTube) with each video being wrapped in an <article>, a list of products (think Amazon) and so on. Any of those <article>s is conceptually syndicatable — each could stand alone on its own dedicated page, in an advert on another page, as an entry in an RSS feed, and so on.

https://www.smashingmagazine.com/2020/01/html5-article-section/

An article element also helps browsers such as Apple Watch or reader views know what content to jump to with their stripped-back browsers. And many screen readers will surface them as a place-of-interest.

How to observe other websites

The aptly named Landmarks browser extension lets you view, highlight, and navigate around the landmarks on the current page.

The Landmarks extension listing the named landmarks on a page

There’s also a free Accessibility Insights extension from Microsoft that provides similar visualization, and also has automated checks for WCAG’s AA.

I encourage you to view landmarks on news sites, social media such as Twitter, web apps such as GitHub, and everything in between. You’ll find that there’s a fair amount of consistency, and some will be better than others. You’ll also have a bar to meet when building your own.

Conclusion

These landmarks apply to all websites: landing pages, documentation, single-page-apps, and everything in between. They ensure all users can orient themselves to quickly become familiar with and navigate around your creation.

They also provide a consistent language that we can design and build around. Share this and other articles (which I’ll link to below) with developers, designers, and managers on your team. Landmarks provide familiarity, which leads to happier users.

Further reading

People quickly accomplishing ambitious things together

· Concepts, Planning, Product, Teams · Patrick Smith

Patrick Collison of Stripe has started putting together a list of ‘examples of people quickly accomplishing ambitious things together.’ Sometimes an idea comes screaming into the world.

Xerox Alto. Work on the Xerox Alto, the first GUI-oriented computer, started in November 1972 because of a bet: “Chuck said that a futuristic computer could be done ‘in three months’ and a Xerox exec bet him a case of wine that it couldn’t be done”.

Git. Linus Torvalds started working on Git on April 3 2005. It was self-hosting 4 days later. On April 20 2005, 17 days after work commenced, Linux 2.6.12-rc3 was publicly released with Git.

BankAmericard. Dee Hock was given 90 days to launch the BankAmericard card (which became the Visa card), starting from scratch. He did. In that period, he signed up more than 100,000 customers.

Amazon Prime. Amazon started to implement the first version of Amazon Prime in late 2004 and announced it on February 2 2005, six weeks later.

Understanding state

· Concepts, React · Patrick Smith

State is what is known by your app.

In a web app, state refers to local state. It only knows a subset of all information. And that information is from a certain point in time: it might have changed since.

In a JavaScript web app, your user’s web browser talks to an API. The API has its own state, usually stored permanently in a database.

(In a server-side web app like one built with Rails, your application usually talks to the database directly.)

So local state. It’s your web app’s state of the world at a certain time. That world depends on factors, such as which account is signed in, or perhaps whether the user is not signed in, which page they are looking at, and what capabilities their account has.

Loading state

For your web app to display information, it has to first load it.

So your web app loads this initial information, and this becomes its state of the world. If five minutes later, the latest information is desired, it must load it again.

Making things more complicated is the fact that there is not just one set of information.

Take Instagram as an example. There is the feed of photos. As you scroll through the feed, more photos are loaded. For a good user experience, photos are loaded before you reach them. But still, only a subset of photos are loaded. It would make little sense to load your feed from the top to the very bottom, as it would likely be hundreds of megabytes of information in total, and be a huge strain on Instagram’s servers to collate all that information.

Instagram offers more than just a feed. Your can explore and search for photos. You can take or upload a photo. You can see a list of activity aimed at you, such as recent comments or likes on your posts. You can also look at your own profile, with the photos you have posted.

Depending on which section you look at, and how you interact within that section (scroll, tap to see details, tap to go to someone’s profile), new information will have to be loaded. The app’s state of the world expands. It goes from a small subset of data, to a larger subset of data.

Managing this state, and coordinating the loading of additional or fresher state, is one of the key skills in building web apps.

A step back. A command line app.

Any app that has an interactive user interface has state. Think of PowerPoint and the file that is being viewed, the currently active slide, whether that slide is being edited, viewed with slides listed to the side, or is being presented. All of those variables are state.

A command line app also has state. The less command efficiently reads from a file, only presenting a slice of it that fits in the terminal window. As the user scrolls and down, the app’s state is updated with the offset slice into the file.

A command line app is a useful starting point, because many operate in the same manner as web apps based on React. In React, the application is built by deriving the entire user interface from state. The application developer declares their intent for what should be displayed on screen given a certain state. If the state changes, then that codified intent is used again, but with the updated state. And so on, every user interaction usually affects the state in some way, and the user interface is updated quickly to respond. This pipeline approach is popular with games, where the displayed image is rebuilt again and again many times a second, fast enough to produce fluid motion.

The alternative is to intertwine the state with what is being presented. As new information comes in, it is not a simple change to state. It is a manual change to what’s on screen too.

A pipeline is like to correct a typo in a printed document, making the change in software, throwing the old copy away, and printing a new fresh copy. Fortunately, in an app built totally with software means that nothing is wasted.

An intertwined approach is more akin to correcting a typo by using whiteout and a fine pen. It’s much less wasteful, but takes much more effort and skill than just printing a new copy would be. The same is true for apps.