This is a browser extension (for Firefox, Chrome and Opera) that enables navigation of WAI-ARIA landmarks, via the keyboard or a pop-up menu.
Landmarks broadly signpost the areas of a page (e.g. navigation, search, main content and so on). They can make navigation considerably easier for people who use the keyboard to navigate and those using assistive technologies such as screen-readers, because they make it much quicker to get an overview and to navigate to (and between) areas of interest.
The following sections explain how to install and use the extension.
If you’re a web author/developer, check out the information below on why landmarks rock, and how easy they are to put into your site—in fact, if you’re using HTML 5, you probably already have landmarks on your site, but there are some ways to make them even more helpful, as discussed below.
Table of Contents
- Navigating Landmarks
- Border Preferences
- This Extension’s Support for Landmarks
- Information for Web Authors, Designers and Developers
- Firefox: Install via Mozilla Add-ons
- Chrome: Install via the Chrome Web Store
- Opera: Install via Opera add-ons
If you need support, please check the known issues for Landmarks and, if necessary, file a new issue using the “New Issue” button on that page.
Via Shortcut Key
You can use shortcut keys to navigate between landmarks. By default, the keys are:
- Next landmark: Alt+Shift+N
- Previous landmark: Alt+Shift+P
- Main landmark: Alt+Shift+M
- Open the landmarks pop-up: Alt+Shift+L, then use Tab and Shift+Tab to move between buttons, Space or Enter to move focus to a landmark and Escape to close the pop-up.
(On a Mac, use the Option key, which is equivalent to Alt.)
Landmarks will be focused, and a border shown according to your border preferences.
You can change the keyboard shortcuts in the following browsers.
- Chrome: More ⋮ → Settings → Extensions [on the left-hand side] → Keyboard shortcuts [at the bottom of the page]
- Opera: Opera Menu / Speed Dial → Extensions → Extension Keyboard Shortcuts
Via Toolbar Pop-up
If landmarks are found on the page, the Landmarks button in the toolbar (which looks like an “L”) will be badged with the number of landmarks found.
Activate the Landmarks toolbar button. A pop-up will appear. If landmarks are present on the page, they will be shown in a nested list, reflecting the structure of the landmarks on the page.
Activate a button in the list to move to that landmark. The landmark will be focused, and a border shown according to your border preferences.
To close the pop-up, press Escape, or click outside of the pop-up.
A border will be drawn around the landmarks as you navigate them. The border contains a label that displays the landmark’s type and name (if one was provided by the author of the page). You can customise various aspects of this in the extension’s preferences/options page. The available settings are:
- If the border should be displayed, and for how long.
- Momentary (default): the border remains visible for a short time, then disappears.
- Persistent: the border remains visible at all times.
- None: no border is drawn.
- The border’s colour.
- The font size used in the landmark label.
You can get to the extension’s settings as follows.
- Firefox: Menu ☰ → Add-ons → Extensions [on left-hand side, if not already activated] → Preferences [under Landmarks, then scroll down]
- Chrome/Opera: Right-click on, or activate the context menu of, the Landmarks button in the toolbar → Options
This Extension’s Support for Landmarks
The extension supports WAI-ARIA landmark roles, both as supplied via the
role attribute and as implicit landmarks via HTML 5 elements. All landmark roles are supported, with some caveats, as per the relevant specifications, which are described below.
If a landmark label is present (via the
aria-label attributes), they’ll be shown in the pop-up. As per the accessible name calculation algorithm used by browsers, the
aria-labelledby attribute takes precedence over
contentinfo) elements are not considered landmarks unless they are the page-wide header/footer elements. (As per the HTML element role mappings.)
regionlandmarks are intended to be labelled. Ideally, this should be done with a visual label and an
aria-labelledbyattribute (so all users can perceive the label). However, if a label is only provided by the (non-visual)
aria-labelattribute, this extension will recognise it.
There is ambiguity in the WAI-ARIA specification as to whether they might still be counted as landmarks even if they are unlabelled. Most assistive technologies do not count unlabelled
region landmarks, because doing so could add a lot of noise to landmark navigation. Therefore this extension also ignores them.
Digital Publishing ARIA Landmarks
The following additional landmark roles defined in the Digital Publishing WAI-ARIA Module 1.0 are also supported.
doc-index(is a landmark via
doc-pagelist(is a landmark via
doc-toc(is a landmark via
Information for Web Authors, Designers and Developers
As described at start of this document, landmarks can really help various people get a quick overview of your site, and navigate it much more effectively. This can save them a lot of time, so please consider implementing landmarks on your site; here is some information to help you do so…
- Léonie Watson demonstrates landmarks (video)
- W3C ARIA landmarks example and advice page
- Easy content organisation with HTML5 (The Paciello Group ‘blog article)
- Using WAI-ARIA landmarks (The Paciello Group ‘blog article)
- W3C WCAG technique ARIA11: Using ARIA landmarks to identify regions of a page
Please bear in mind the following when implementing landmarks…
It’s important that landmarks are not over-used, because their power comes from providing a concise overview of the content of the page. The heading hierarchy for the page can be relied upon for more fine-grained navigation within a particular area of a page.
Rule of thumb: Use as few landmarks as possible, but ensure that all of the content on the page is within a landmark region.
If you’re using HTML 5 elements such as
<main>and others, then your page will inherit some landmarks automagically. However it can be really helpful to label them (especially if there’s more than one of a landmark on a page, such as a separate site-wide and page-local
<nav>). The W3C documentation has all the details, but essentially you would use either the
Rule of thumb: If you’ve more than one type of landmark, then be sure to label them, so their purpose is clear.
You can build and run the current code locally as follows.
Clone the Landmarks repository on GitHub to your computer.
Ensure you have all the required build tools with
npm install(you will need Node.js).
Run the build script to build one or all of the extensions:
npm run build:firefox
npm run build:chrome
npm run build:opera
npm run build:edge(Edge support is in development, but not fully ready yet.)
npm run build:all
The built versions of the extension are placed in the
build/<browser>/directories and ZIP files for each will be created in the root of the checked-out repository.
To load and use the extension locally in your browser…
- Firefox: either:
- use Mozilla’s instructions on temporarily loading extensions from disk, or
- if you have
npm run start:firefoxto open Firefox with Landmarks loaded. It will keep itself up-to-date when you re-build.
- Chrome: follow Google’s instructions on loading the extension.
- Opera: refer to Testing and Debugging.
- Edge: use the Adding, moving, and removing extensions for Microsoft Edge instructions on Microsoft’s developer site.
- Firefox: either:
Some further info on the test/build process:
Automated tests are run as a pre-requisite part of the build process; you can also run them with
You can remove the
build/<browser>/directories and ZIP files with
npm run clean:<browser>and
npm run clean:all, as with the build scripts above.
Because the process of rasterising the SVG to variously-sized PNGs is slow, the PNGs are cached so they only need to be re-generated when the SVG changes. You can clean out the cache with
npm run clean:cache.
pre-commithook is used to ensure only code that passes tests is committed (it does this by running a build, which, in turn, runs the tests). Husky manages this so that a build is run before you are asked for a commit message.
build:chrome:testscript is provided for making an alpha/beta/test build for Chrome, which is the same as a normal build, but the extension is retitled to “Landmarks (test version)”. A separate extension listing is required for publishing test versions in the Chrome Web Store. For Firefox Add-ons, a version number such as “2.1.0beta1” can be used and the built package can be uploaded to the extension’s beta channel.
The following pages are incorporated into the automated test suite, but you can also visit them in-browser to try out the extension’s UI.
This is a fork of the original landmarks extension written by davidtodd at IBM. Thanks to stevefaulkner for suggesting I work on this, and for feature suggestions (and again to davidtodd for supporting me doing so), and to The Paciello Group for donating a significant chunk of the initial development time.
- 2.3.0 - 17th of May 2018
- Add landmark labels to the border, which is now drawn more robustly and has customisable colour. [#158, #162]
- Options are saved as they’re changed by the user, and borders get updated to reflect settings changes immediately. [#160]
- Fix text sometimes overflowing buttons in Firefox. [#163]
- Minor tweaks to documentation, build process; library package bumps. [#159, #161, #164]
- 2.2.0 - 18th of February 2018
- 2.1.1 - 19th of January 2018
- Improve performance on sites/apps that change rapidly (such as Google Docs) by limiting how quickly Landmarks responds to further changes when they’re being made in quick succession. [#139]
- Add a keyboard shortcut to show the landmarks pop-up. [#135]
- Give the options page a title that shows up for Opera users. [#136]
- Behind-the-scenes code and infrastructure improvements for improved quality. [#124, #128, #129, #130, #142, #143]
- 2.1.0 - 6th of November 2017
- Landmarks are now updated when pages change dynamically (not just when the whole page is loaded). This should make the extension much more useful when working with web-apps and pages with pop-ups and slide-out menus, for example. [#111]
- Add a keyboard shortcut to skip to the main landmark. [also #111]
- Fix a bug whereby sibling landmarks may not be identified as such. [#112]
- Adopt more browser-like UI on Firefox (pop-up and options) and Opera (options). [#115]
- Use Mozilla’s ‘addons-linter’ to check the built extension. [err, also #111]
- 2.0.8 - 18th of September 2017
- Landmarks now ignores hidden regions. [#85]
- Fix a bug that caused the pop-up to incorrectly report nesting that changes by more than one level between landmarks. [#102]
- Correctly restore elements’ outlines after they are highlighted. [#94]
- Automatically disable the extension on browsers’ extensions store pages. [#97]
- Start exploring what’s needed for Edge support in future. [#99]
- Improvements to the SVG to PNG process. [#95]
- Other more minor tweaks and fixes.
- README updates.
- 2.0.7 - 11th of May 2017
- Officially support Opera.
- Make the landmark highlight more visible.
- Open a help page when the extension is installed/updated on Firefox (this was already supported on Chrome, and is on Opera).
- Make use of Firefox’s synching of settings across devices.
- More tests, and numerous other code improvements behind the scenes.
- Partly works on Edge; still a few things to sort out before it’s robust (also, the extensions store is not yet immediately open to submissions from allcomers).
- 2.0.6 - 2nd of February 2017
- Add a test suite to ensure landmarks are identified correctly.
- Various large internal code-quality improvements.
- 2.0.5 - 5th of December 2016
- No user-facing changes.
- Fix error in packaging (the new build system was not actually compressing the ZIP file, which different parts of the submission process for Chrome and Firefox didn’t like—oops!)
- Add more code robustosity checks with ESLint.
- 2.0.4 - 4th of December 2016
- Clean up the appearance of the pop-up.
- Increase ‘momentary’ highlight duration to two seconds, from one second.
- Remove a workaround for a bug in Firefox pop-up sizing that was fixed in Firefox 50.
- Drop Grunt and switch to just using NPM and scripts for building the extensions.
- Track builds with Travis CI.
- Use ESLint and EditorConfig code standards and quality tools.
- 2.0.3 - 23rd of September 2016
- When installed/updated on Chrome, show the web page, with a (hopefully) helpful notice about the install/upgrade.
- Automatically re-inject the content script on Chrome when the extension is updated (or inject it when the extension is installed), as users would expect it to start working straight away. (Firefox does this itself.)
- Locale is now en_GB (instead of en).
- Switch to using grunt-phantom-rasterize for converting the SVGs to PNGs.
- 2.0.2 - 12th of August 2016
- First WebExtension Release