Notes on ARIA use in HTML

First rule of ARIA use

If you can use a native HTML element [[!HTML5]] or attribute with the semantics and behaviour you require already built in, instead of re-purposing an element and adding an ARIA role, state or property to make it accessible, then do so.

Under what circumstances may this not be possible?

Second rule of ARIA use

Do not change native semantics, unless you really have to.

For example: Developer wants to build a heading that's a button.

Do not do this:

< h1 role=button >heading button</ h1 >

Do this:

<h1> <button> heading button </button> </h1>

Or if you can't possibly use the correct element, do this:

<h1> < span role=button >heading button </span> </h1>

Note: if a non interactive element is used as the basis for an interactive element, developers have to add the semantics using ARIA and the appropriate interaction behaviour using scripting. In the case of a button, for example, it is much better and easier to Just use a (native HTML) button.

Note: it is OK to use native HTML elements, that have similar semantics to ARIA roles used, for fallback. For example, using HTML list elements for the skeleton of an ARIA enabled, scripted tree widget.

Third rule of ARIA use

All interactive ARIA controls must be usable with the keyboard.

If you create a widget that a user can click or tap or drag or drop or slide or scroll, a user must also be able to navigate to the widget and perform an equivalent action using the keyboard.

All interactive widgets must be scripted to respond to standard key strokes or key stroke combinations where applicable.

For example, if using role=button the element must be able to receive focus and a user must be able to activate the action associated with the element using both the enter (on WIN OS) or return (MAC OS) and the space key.

Refer to the keyboard and structural navigation and design patterns sections of the WAI-ARIA 1.0 Authoring Practices

Fourth rule of ARIA use

Do not use role="presentation" or aria-hidden="true" on a visible focusable element .

Using either of these on a visible focusable element will result in some users focusing on 'nothing'.

Do not do this:

< button role=presentation >press me</ button >

Do not do this:

< button aria-hidden="true" >press me</ button >

Note: If an interactive element cannot be seen or interacted with, then you can apply aria-hidden, for example:

button {display:none}

< button aria-hidden="true" >press me</ button >

Fifth rule of ARIA use

All interactive elements must have an accessible name.

An interactive element only has an accessible name when it's Accessibility API accessible name (or equivalent) property has a value.

For example, the input type=text in the code example below has a visible label 'user name' , but no accessible name:

    user name <input type="text">

<!-- label element used, but not associated with the control 
it is supposed to label -->
    <label>user name</label> <input type="text">

The control's MSAA accName property is empty:

example control with MSAA name and role information displayed. The accName property has no value, the accRole property is 'editable text'.

In comparison, the input type=text in the code example below has a visible label 'user name' and an accessible name. This example has an accessible name because the input element is a labelable element and the label element is used correctly to associate the label text with the input.

<!-- Note: use of for/id or wrapping label around text 
and control methods will result in an accessible name -->

<label>user name <input type="text"></label> 


<label for="uname">user name</label> <input type="text" id="uname">

The control's MSAA accName property has a value of "user name":

example control with MSAA name and role information displayed. The accName property has a value of 'user name', the accRole property is 'editable text'.

The label element cannot be used to provide an accessible name for custom controls, unless the label is referencing a native HTML labelable element.

     <!-- HTML input element with combox role -->

      user name <input type="text"  role="combobox";>

The control's MSAA accName property has a value of "user name":

example input element with MSAA name and role information displayed. The accName property has a value of 'user name', the accRole property is 'combo box'.

A div element regardless of what role is assigned is not a HTML labelable element.

     <!-- HTML div element with combox role -->

      user name <div  role="combobox"></div>

The control's MSAA accName property is empty:

example div element with MSAA name and role information displayed. The accName property is empty, the accRole property is 'combo box'.

5th rule is a work in progress


What does adding a role do to the native semantics?

Adding an ARIA role overrides the native role semantics in the accessibility tree which is reported via the accessibility API, and therefore ARIA indirectly affects what is reported to a screen reader or other assistive technology.

For example, this code in the HTML tree:

<h1 role=button>text</h1>

Becomes this in the accessibility tree:

button  with a label of 'heading text'

What adding a role does not do

Adding an ARIA role will not make an element look or act differently for people not using assistive technology. It does not change the behaviours, states and properties of the host element but only the native role semantics.

For example, this code in the HTML tree:

<button role=heading aria-level=1>text</button>

Becomes this in the accessibility tree:

a heading

But it can still be pressed, it is still in the default tab order, still looks like a button and still triggers any associated actions when pressed. That's why it is a HTML5 conformance error to change a button into a heading.

Note: Changing the role of an element does not add behaviors, properties or states to the role used. ARIA does not change the way it looks or acts in a browser. For instance, when links are used to behave like buttons, adding role=button alone is not sufficient. It will also be necessary to make act like a button, by including a key event handler that listens for the space key which native buttons do, because native buttons can be activated using the enter key or the spacebar.

Add ARIA inline or via script?

If the ARIA role or aria-* attribute does not rely on scripting to provide interaction behaviour, then it is safe to include the ARIA markup inline. For example, it is fine to add ARIA landmark roles or ARIA labelling and describing attributes inline.

If the content and interaction is only supported in a scripting-enabled browsing context, i.e. Google docs (its applications require JavaScript enabled to work), it is also safe to include the ARIA markup inline as the application simply will not work (for anyone) without JavaScript enabled.

Otherwise insert, change and remove ARIA via scripting. For instance, a collapsed section of a tree widget might look like this:

<li role=treeitem aria-expanded=false ...

When the user opens the section, it is changed to this using JavaScript :

<li role=treeitem aria-expanded=true ...

ARIA validation

The easiest method is to use the HTML5 DOCTYPE with ARIA markup and validate using the W3C Nu Markup Checker. ARIA works equally well with any other DOCTYPE, but validation tools will produce errors when they encounter ARIA markup as the associated DTDs have not been updated to recognise ARIA markup and it is unlikely they ever will be.

These validation errors in versions of HTML prior of HTML5 are in no way indicative of ARIA creating any real world accessibility problems nor do they mean there will be a negative user experience. They are merely the result of old automated validation tests that do not accommodate ARIA accessibility annotations.

Note: The W3C Nu Markup Checker support for ARIA checking is a work in progress, so cannot be wholly relied upon (though it is pretty darn good!)to provide the correct results. It is recommended that if you encounter a result that conflicts with the ARIA conformance requirements in the ARIA specification or the HTML5 specification, please raise a bug report.

Use of role=presentation

role=presentation removes the semantics from the element it is on.

For example, this code in the HTML tree:

<h1 role="presentation">text</h1>

Becomes this in the accessibility tree:

text, no heading

In other words, it is just reported in the accessibility tree as a text string with no semantic meaning.

For elements with no required children, any elements nested inside the element with role=presentation preserve their semantics.

For example, this code in the HTML tree:

<h1 role="presentation"><abbr>API</abbr></h1>

Becomes this in the accessibility tree:

abbr with text of API

For elements with required children (such as ul or table) any required child elements nested inside the element with role=presentation also have their semantics removed.

For example, this code in the HTML tree:

<table role="presentation">

Becomes this in the accessibility tree:

abbr with text of API

Note: Any elements that are not required children of the element with a role=presentation keep their semantics. This includes other elements with required children such as nested lists or nested tables.

For example this code, consisting of a table with another table nested inside it, in the HTML tree:


Becomes this in the accessibility tree:

outer ttable with 1 row and 1 cell containg another table with 1 row and 1 cell containing an abbr element.

By adding role=presentation to the outer table element, this code in the HTML tree:

<table role="presentation">

Becomes this in the accessibility tree, the semantics of the outer table including its required children (the tr and td elements) are removed by the addition of role=presentation:

table with 1 row and 1 cell containing an abbr element

Examples of role=presentation use

Use in fixing an incorrect table structure

<div aria-readonly="true" role="grid">
<table role="presentation">
<tbody><tr role="row">
<th role="columnheader">Dog Names</th>
<th role="columnheader">Cat Names</th>
<th role="columnheader">Cow names</th>
<table role="presentation">
<tbody><tr role="row">
<td role="gridcell">Fido</td>
<td role="gridcell">Whiskers</td>
<td role="gridcell">Clarabella</td>
<tr role="row">
<td role="gridcell">Woofie</td>
<td role="gridcell">Claws</td>
<td role="gridcell">Glenn</td>

aria-labelledby and aria-describedby

Currently aria-labelledby and aria-describedby are more robustly supported for associating text content to a subset of interactive content elements. As of this writing they do not work correctly on links , support on embedded content is unknown, but can be safely used on form controls including the many input types.

In Internet Explorer, if you use aria-labelledby with multiple id references or aria-describedby with single or multiple id references, the referenced elements must be what Microsoft terms as accessible HTML elements.

The following example of aria-labelledby with multiple references uses a span with a tabindex=-1 added. Refer to Making Non accessible Elements Accessible.

<label id="l1" for="f3">label text</label>

<input type="text" id="f3" aria-labelledby="l1 l2" >

<p>other content</p>

<span tabindex="-1" id="l2" >more label text</span>

Elements also become accessible HTML elements in Internet Explorer when the element has an ARIA role. For example:

<div aria-describedby="test">text</div>

<div id="test" role="tooltip" >tooltip text</div>

Using ARIA role=application

How does role="application" affect a screen reader?

On many popular screen readers today, most keystrokes are captured by the screen reader and not the web page when the user is in browse mode. This is necessary for efficient navigation of a page. As of this writing, when application mode is set, many screen reader stop intercepting keystrokes, and pass all keystrokes directly to the browser. Then the user won't be able to navigate the page as easily. For instance they won't be able to skip around the page by headings or read a paragraph of static text line-by-line. However, several screen readers do not behave differently when there is an application role set.

So when should I use it, and when not?

In determining when to use role=application, one should consider, among other things, the advantages of screen reader keyboard shortcuts weighed against the loss of those features. It generally should not be used, and if it is, usability testing with screen reader users should be conducted.

You do not use role="application" if a set of controls only contains these widgets, that are all part of standard HTML. This also applies if you mark them up and create an interaction model using WAI-ARIA roles instead of standard HTML widgets:

NOTE: It's not recommended that authors develop custom text input widgets. It's almost always best to use the native inputs for these.

You also do not use the application role if your widget is one of the following more dynamic and non-native widgets. Screen readers and other assistive technologies that support WAI-ARIA will support switching between browse and focus modes for these by default too:

You only want to use role=application if the content you’re providing consists of only focusable, interactive controls, and of those, mostly advanced widgets that emulate a real desktop application. Note that, despite many things now being called a web application, most of the content these web applications work with are still document-based information, be it Facebook posts and comments, blogs, Twitter feeds, or even accordions that show and hide certain types of information dynamically. We primarily still deal with documents on the web, even though they may have a desktop-ish feel to them on the surface.

It is not necessary to use role=application to have control-specific keyboard shortcuts while the user is in forms (focus) mode on their screen reader. For instance, a custom control with ARIA role=listbox can easily capture all keys pressed including arrow keys, while the user is interacting with it.

In short: The times when you actually will use role=application will probably be very rare!

So where do I put role=application in the rare cases it is useful?

Put it on the closest containing element of your widget, for example, the parent div of your element that is your outer most widget element. If that outer div wraps only widgets that need the application interaction model, this will make sure focus mode is switched off once the user tabs out of this widget.

Only put it on the body element if your page consists solely of a widget or set of widgets that all need the focus mode to be turned on. If you have a majority of these widgets, but also have something you want the user to browse, use role=document on the outer-most element of this document-ish part of the page. It is the counterpart to role=application and will allow you to tell the screen reader to use browse mode for this part. Also make this element tabbable by setting a tabindex=0 on it so the user has a chance to reach it.

As a rule of thumb: If your page consists of over 90 or even 95 percent of widgets, role=application may be appropriate. Even then, find someone knowledgeable who can actually test two versions of this: One with and one without role=application set to see which model works best.

NEVER put role=application on a widely containing element such as body if your page consists mostly of traditional widgets or page elements such as links that the user does not have to interact with in focus mode. This will cause huge headaches for any assistive technology user trying to use your site/application.

For further information on the use of role=application refer to If you use the WAI-ARIA role "application", please do so wisely!

Custom Control Accessible Development Checklist:

Check your custom control against the following design considerations. If the answer to any of them is 'No' then consider fixing before release or at least documenting the issues to inform other developers that your custom control will not be usable by some people due to limited accessibility support.

Custom Control Design Considerations
design consideration description Yes/No
focusable Can you get to the control via the keyboard? Refer to Providing Keyboard Focus
operable Can you use the control with the keyboard? Refer to Keyboard Navigation
expected operation Can you use the standard keys for the control type to operate it. Refer to ARIA Widget Design Patterns
clear indication of focus Can you easily see it when the control has focus? Refer to Visible Focus (WCAG2)
label The control has a text label that is exposed as an accessible name in accessibility APIs
role The control has an appropriate role exposed in accessibility APIs
states and properties The control has any UI states and properties that it has exposed in accessibility APIs
color contrast The control label/description/icon is perceivable/usable for low vision users (Use a color contrast checker.)
high contrast mode The control is perceivable/usable when High Contrast Mode is enabled (e.g. Windows HC mode)

Recommendations Table:

Refer to the Document conformance requirements for use of ARIA attributes in HTML table in the ARIA in HTML specification.

ARIA Role, State, and Property Quick Reference

(Reformatted and reorganized information from: Accessible Rich Internet Applications (WAI-ARIA) 1.0)

In addition to the states and properties shown in the table, the following global states and properties are supported on all roles.

Global states and properties

ARIA Roles, States and Properties
Role Description Required Properties Supported Properties -
Global +
alert A message with important, and usually time-sensitive, information. See related alertdialog and status. NONE
alertdialog A type of dialog that contains an alert message, where initial focus goes to an element within the dialog. See related alert and dialog. NONE
application A region declared as a web application, as opposed to a web document. NONE
article A section of a page that consists of a composition that forms an independent part of a document, page, or site. NONE
banner A region that contains mostly site-oriented content, rather than page-specific content. NONE
button An input that allows for user-triggered actions when clicked or pressed. See related link. NONE
checkbox A checkable input that has three possible values: true, false, or mixed.
columnheader A cell containing header information for a column. NONE
combobox A presentation of a select; usually similar to a textbox where users can type ahead to select an option, or type to enter arbitrary text as a new item in the list. See related listbox.
complementary A supporting section of the document, designed to be complementary to the main content at a similar level in the DOM hierarchy, but remains meaningful when separated from the main content. NONE
contentinfo A large perceivable region that contains information about the parent document. NONE
definition A definition of a term or concept. NONE
dialog A dialog is an application window that is designed to interrupt the current processing of an application in order to prompt the user to enter information or require a response. See related alertdialog. NONE
directory A list of references to members of a group, such as a static table of contents. NONE
document A region containing related information that is declared as document content, as opposed to a web application. NONE
form A landmark region that contains a collection of items and objects that, as a whole, combine to create a form. See related search. NONE
grid A grid is an interactive control which contains cells of tabular data arranged in rows and columns, like a table. NONE
gridcell A cell in a grid or treegrid. NONE
group A set of user interface objects which are not intended to be included in a page summary or table of contents by assistive technologies. NONE
heading A heading for a section of the page. NONE
img A container for a collection of elements that form an image. NONE
An interactive reference to an internal or external resource that, when activated, causes the user agent to navigate to that resource. See related button. NONE
list A group of non-interactive list items. See related listbox. NONE
listbox A widget that allows the user to select one or more items from a list of choices. See related combobox and list. NONE
listitem A single item in a list or directory. NONE
log A type of live region where new information is added in meaningful order and old information may disappear. See related marquee. NONE
main The main content of a document. NONE
marquee A type of live region where non-essential information changes frequently. See related log. NONE
math Content that represents a mathematical expression. NONE
menu A type of widget that offers a list of choices to the user. NONE

A presentation of menu that usually remains visible and is usually presented horizontally.

Authors SHOULD ensure that menubar interaction is similar to the typical menu bar interaction in a desktop graphical user interface.It is NOT really intended to mark up site navigation list items

menuitem An option in a group of choices contained by a menu or menubar. NONE
menuitemcheckbox A checkable menuitem that has three possible values: true, false, or mixed.
menuitemradio A checkable menuitem in a group of menuitemradio roles, only one of which can be checked at a time.
navigation A collection of navigational elements (usually links) for navigating the document or related documents. NONE
note A section whose content is parenthetic or ancillary to the main content of the resource. NONE
option A selectable item in a select list. NONE
presentation An element whose implicit native role semantics will not be mapped to the accessibility API. NONE
progressbar An element that displays the progress status for tasks that take a long time. NONE
radio A checkable input in a group of radio roles, only one of which can be checked at a time.
radiogroup A group of radio buttons. NONE
region A large perceivable section of a web page or document, that the author feels is important enough to be included in a page summary or table of contents, for example, an area of the page containing live sporting event statistics. NONE
row A row of cells in a grid. NONE
rowgroup A group containing one or more row elements in a grid. NONE
rowheader A cell containing header information for a row in a grid. NONE
scrollbar A graphical object that controls the scrolling of content within a viewing area, regardless of whether the content is fully displayed within the viewing area.
search A landmark region that contains a collection of items and objects that, as a whole, combine to create a search facility. See related form. NONE
separator A divider that separates and distinguishes sections of content or groups of menuitems.
  • NONE
slider A user input where the user selects a value from within a given range.
spinbutton A form of range that expects the user to select from among discrete choices.
status A container whose content is advisory information for the user but is not important enough to justify an alert, often but not necessarily presented as a status bar. See related alert. NONE
tab A grouping label providing a mechanism for selecting the tab content that is to be rendered to the user. NONE
tablist A list of tab elements, which are references to tabpanel elements. NONE
tabpanel A container for the resources associated with a tab, where each tab is contained in a tablist. NONE
textbox Input that allows free-form text as its value. NONE
timer A type of live region containing a numerical counter which indicates an amount of elapsed time from a start point, or the time remaining until an end point. NONE
toolbar A collection of commonly used function buttons represented in compact visual form. NONE
tooltip A contextual popup that displays a description for an element. NONE
tree A type of list that may contain sub-level nested groups that can be collapsed and expanded. NONE
treegrid A grid whose rows can be expanded and collapsed in the same manner as for a tree. NONE
treeitem An option item of a tree. This is an element within a tree that may be expanded or collapsed if it contains a sub-level group of treeitems. NONE

Definitions of States and Properties (all aria-* attributes)

Below is an alphabetical list of ARIA states and properties to be used by rich internet application authors. A detailed definition of each ARIA state and property can be found by following the attribute links (to their definitions in Accessible Rich Internet Applications (WAI-ARIA) 1.0.

Identifies the currently active descendant of a composite widget.
Indicates whether assistive technologies will present all, or only parts of, the changed region based on the change notifications defined by the aria-relevant attribute. See related aria-relevant.
Indicates whether user input completion suggestions are provided.
aria-busy (state)
Indicates whether an element, and its subtree, are currently being updated.
aria-checked (state)
Indicates the current "checked" state of checkboxes, radio buttons, and other widgets. See related aria-pressed and aria-selected.
Identifies the element (or elements) whose contents or presence are controlled by the current element. See related aria-owns.
Identifies the element (or elements) that describes the object. See related aria-labelledby.
aria-disabled (state)
Indicates that the element is perceivable but disabled, so it is not editable or otherwise operable. See related aria-hidden and aria-readonly.
Indicates what functions can be performed when the dragged object is released on the drop target. This allows assistive technologies to convey the possible drag options available to users, including whether a pop-up menu of choices is provided by the application. Typically, drop effect functions can only be provided once an object has been grabbed for a drag operation as the drop effect functions available are dependent on the object being dragged.
aria-expanded (state)
Indicates whether the element, or another grouping element it controls, is currently expanded or collapsed.
Identifies the next element (or elements) in an alternate reading order of content which, at the user's discretion, allows assistive technology to override the general default of reading in document source order.
aria-grabbed (state)
Indicates an element's "grabbed" state in a drag-and-drop operation.
Indicates that the element has a popup context menu or sub-level menu.
aria-hidden (state)
Indicates that the element and all of its descendants are not visible or perceivable to any user as implemented by the author. See related aria-disabled.
aria-invalid (state)
Indicates the entered value does not conform to the format expected by the application.
Defines a string value that labels the current element. See related aria-labelledby.
Identifies the element (or elements) that labels the current element. See related aria-label and aria-describedby.
Defines the hierarchical level of an element within a structure.
Indicates that an element will be updated, and describes the types of updates the user agents, assistive technologies, and user can expect from the live region.
Indicates whether a text box accepts multiple lines of input or only a single line.
Indicates that the user may select more than one item from the current selectable descendants.
Indicates whether the element and orientation is horizontal or vertical.
Identifies an element (or elements) in order to define a visual, functional, or contextual parent/child relationship between DOM elements where the DOM hierarchy cannot be used to represent the relationship. See related aria-controls.
Defines an element's number or position in the current set of listitems or treeitems. Not required if all elements in the set are present in the DOM. See related aria-setsize.
aria-pressed (state)
Indicates the current "pressed" state of toggle buttons. See related aria-checked and aria-selected.
Indicates that the element is not editable, but is otherwise operable. See related aria-disabled.
Indicates what user agent change notifications (additions, removals, etc.) assistive technologies will receive within a live region. See related aria-atomic.
Indicates that user input is required on the element before a form may be submitted.
aria-selected (state)
Indicates the current "selected" state of various widgets. See related aria-checked and aria-pressed.
Defines the number of items in the current set of listitems or treeitems. Not required if all elements in the set are present in the DOM. See related aria-posinset.
Indicates if items in a table or grid are sorted in ascending or descending order.
Defines the maximum allowed value for a range widget.
Defines the minimum allowed value for a range widget.
Defines the current value for a range widget. See related aria-valuetext.
Defines the human readable text alternative of aria-valuenow for a range widget.

Abstract roles

Do not use the following abstract roles as they do not do anything!

The following roles are used to support the WAI-ARIA role taxonomy for the purpose of defining general role concepts. Abstract roles are used for the ontology. Authors MUST NOT not use abstract roles in content.