Copyright © 2006 W3C® (MIT, ERCIM, Keio), All Rights Reserved. W3C liability, trademark and document use rules apply.
The Roadmap for Accessible Rich Internet Applications (WAI-ARIA Roadmap) addresses the accessibility of dynamic Web content for people with disabilities. The roadmap outlines the technologies to map controls, AJAX live regions, and events to accessibility APIs, including custom controls used for Rich Internet Applications. The roadmap also outlines new navigation techniques to mark common Web structures as menus, primary content, secondary content, banner information and other types of Web structures. These new technologies can be used to improve the accessibility and usability of Web resources by people with disabilities, without extensive modification to existing libraries of Web resources.
This roadmap identifies some technology gaps to achieve the desired access. Some of the technologies required to fill these gaps are under development in other W3C Groups, such as the XHTML Role Attribute Module [XHTML-ROLE] being developed by the HTML Activity. The remaining gaps are planned to be filled by two companion specifications: Roles for Accessible Rich Internet Applications [ARIA-ROLE] and States and Properties Module for Accessible Rich Internet Applications [ARIA-STATE].
This section describes the status of this document at the time of its publication. Other documents may supersede this document. A list of current W3C publications and the latest revision of this technical report can be found in the W3C technical reports index at http://www.w3.org/TR/.
This document is a First Public Working Draft by the Protocols & Formats Working Group of the Web Accessibility Initiative. The Working Group expects to advance this document to a Working Group Note.
The PFWG seeks feedback on the requirements, gap analysis, and plan set out in this document. In particular, the Working Group would like input about whether:
Please provide comments to assist the group in improving this document. Comments on this document may be sent to public-pfwg-comments@w3.org (Archive).
Publication as a Working Draft does not imply endorsement by the W3C Membership. This is a draft document and may be updated, replaced or obsoleted by other documents at any time. It is inappropriate to cite this document as other than work in progress.
The disclosure obligations of the Participants of this group are described in the charter.
According to the SecuritySpace Technology Penetration Report, JavaScript is found on over 50% of all Web sites today, dramatically affecting the ability for persons with disabilities to access Web content. New Rich Internet Web applications are being created to render custom widgets modeling rich desktop componentry and to perform UI updates without having to reload the entire page - much like a GUI. Legacy GUI accessibility frameworks address these issues through a comprehensive accessibility API and infrastructure to foster interoperability with assistive technologies. These API's constitute a contract between applications and assistive technologies, such as screen readers, to allow them to gain access to rich dynamic content with the appropriate semantics needed to produce a usable alternative.
Unfortunately, HTML and other markup does not provide adequate markup to support accessible dynamic content. Until now, the W3C WAI has discouraged the use of JavaScript per Web Content Accessibility Guidelines (WCAG) 1.0 ([WCAG1], Checkpoint 6.1). The W3C has a number of initiatives in the pipeline to address this problem using a declarative markup approach. This roadmap is designed to create a bridge to fix the interoperability problem with assistive technologies today by incorporating the appropriate meta data in today's XHTML markup to support today's accessibility API. It will incorporate some of the advanced accessibility features, originally designed, in technologies like XHTML2. The intent of XHTML 2 is to make the enablement of Web applications and documents easier for authors. This roadmap will create cross-cutting technologies which may be incorporated by XHTML authors in their markup to describe GUI widgets and document structure to assist assistive technology vendors in providing accessible, usable solutions. The W3C WAI PF working group will work with User Agent manufacturers and assistive technology vendors to ensure we have a working solution.
Accessibility for traditional HTML suffers from a number of problems that make it difficult to support dynamic content:
Authors of JavaScript generated content do not want to tie themselves to the standard tag elements that define the actual user interface element such as tables, ordered lists, etc. JavaScript also make extensive use of tag elements such as DIVs where they will dynamically apply a UI through the use of style sheets and dynamic content changes. An HTML DIV provides no semantic information. The page author may define a DIV as the start of a pop-up menu or even an ordered list. No HTML facility provides a mechanism for the page author cannot convey the role of the DIV as being a pop-up menu. Consequently, a page author cannot tell an assistive technology when these elements have focus. There is no mechanism to convey accessibility property information as to whether the pop-up menu is collapsed or expanded. There is also no mechanism to define what the actions are that can be formed on the element other than through a device-dependent means through the event handler type (onmouseover, onclick, etc.)
In short JavaScript needs an accessibility architecture it can write to such that a solution could be mapped to the accessibility frameworks on the native platform by the user agent.
Figure 1.0 Accessibility
Interoperability at a DOM Node without JavaScript
Figure 1.0 illustrates a typical DOM node in a Model-View-Controller architecture. On the node, data, or the "Model", which should include semantic information, is separated from the user interface presentation, the "View." Here, the document element is managed by the user agent based on the default behavior of the element. The user agent's default behavior at the document element forms the controller. Placed between the DOM node and the assistive technology is a box containing the contract provided by the user agent to the assistive technology. This data includes typical accessibility information found in the accessibility API for many of our accessible platforms for GUIs (role, state, caret, selection, text, hypertext, name description, parent/child information, parent/child information, and relations). For HTML and other W3C markup, the accessibility information provided is dependent on that which is provided by the element's tag name and any accessibility attributes that map to that provided. For example, the accessible role of a table is table. The accessible description would be provided by the author by assigning a title attribute.
Figure 2.0 Accessibility
Interoperability at a DOM node with JavaScript
Figure 2.0 shows the same DOM node provided in Figure 1.0 but with JavaScript acting as the new controller. JavaScript overrides the default user agent behavior at the DOM node. It manipulates the data, content, and style in response to events caused by user interaction to produce Custom widgets. In this situation the default accessibility information is no longer valid and therefore the contract is now invalid. Figure 2.0 shows the contract with asterisks in front of role, state, actions, value, event changes, and relations. These asterisks represent potential accessibility errors and gaps in the base markup. These gaps result from the author's inability to provide the new semantic data needed to support the contract.
At this time, the primary effort in the W3C WAI Protocols and Formats working group is to focus on an extension to XHTML 1.1. It should be noted that XHTML 1.0 could also be enhance to support this extension through the use of a hybrid DTD which incorporates the extensions. The reason for focusing on XHTML 1.X is that XHTML is extensible through the use of namespaces and because it is renderable by today's browses. For this work, we will be producing a document called the Accessible Adaptive Applications (AAA) document and a GUI role taxonomy document. The AAA will constitute extensions to XHTML 1.1 to fill most of the gaps needed to support accessibility API infrastructures and dynamic XHTML content. It will also borrow a new role attribute from XHTML for which a set of accessible role values may be applied from the WAI GUI Role Taxonomy specification. The Role taxonomy specification will provide the necessary core roles found in Accessibility API sets for Windows and Linux. Additionally, it will contain roles which are representative of document structure. The use of document structure is necessary for assistive technologies to navigate complex documents and to know when they have entered active areas of a Web page such as in the case of a dynamic scripted Web application. It will be designed to assist the user agent or authoring tool in determining what properties are supported by a given role and assist with accessibility api mapping of these properties. The taxonomy will be like a class hierarchy used to convey semantics and structure and therefore knowledge about each role. At this time that taxonomy will be written in RDF. Note: Although this work targets XHTML markup, IBM has contributed a W3C technique to embed role and state attributes into HTML documents allowing authors to avoid validation issues. This is may be of particular importance to Web sites needing to support this work in HTML.
The GUI Role Taxonomy specification is meant to be cross-cutting. XHTML 2 introduces the role attribute as part of its built-in accessibility support. The common roles defined in this taxonomy may also be applied to XHTML 2 content and other markup such as SVG should those efforts allow for a mechanism to reference these common roles.
Required Components | Who does what today? (HTML) | Plan to fill the gaps for XHTML 1.X | Plan to fill gaps for XHTML 1.X + XForms | XML Events | Plan to fill the gaps for XHTML 2.0 |
---|---|---|---|---|---|
Events: | |||||
FocusChange | DOM 2 UI Events ([EVENTS], Section 1.6.1) | ||||
Activation | DOM Level 2 Device Independent UI Events ([EVENTS], Section 1.6.1, UIEvent) | ||||
Caret Change | User Agent API | User Agent API | User Agent API | User Agent API | |
* Value Change | User Agent to monitor specific accessibility properties in the States and Properties Module [ARIA-STATE]. For properties representing platform accessibility API values as specified by a given role, it will generate an accessible value change event. (See State Change Event below) | User Agent to monitor change from WAI State Taxonomy information supplied by the Author and fire the platform specific Accessibility API event | |||
* State Change | User Agent to monitor specific accessibility properties in the States and Properties Module [ARIA-STATE]. For properties representing platform accessibility API states, it will generate an AccessibleState Change event. (See Value Change above) | User Agent to monitor change from WAI State Taxonomy information supplied by the Author and fire the platform specific Accessibility API event | |||
Selection Change | User Agent API | User Agent API | User Agent API | User Agent API | |
Mutation | DOM Events | ||||
Accessible Actions: | |||||
* Event Handler functional information and descriptive shortcuts |
|
|
|||
* Access to the available event handlers for enumerating the actions | DOM 3 Events | ||||
* State Information: | States and Properties Module [ARIA-STATE]. | User Agent to map information found in XForms data model with limited use of States and Properties Module. User Agent of XForms data model to accessible state meta found in: States and Properties Module [ARIA-STATE]. | User Agent to map information found in XForms data model with limited use of States and Properties Module. User Agent of XForms data model to accessible state meta found in: States and Properties Module [ARIA-STATE] | ||
* Role Information: | Limited to standard HTML tag names. (Mix Content/presentation) | ||||
Relationships: Parent/child | Limited DOM (affected by style) (Mix Content/presentation) | ||||
Relationships: (Label, MemberOf - Group, ControllerFor) | Limited to HTML (Title, alt, label) | States and Properties Module [ARIA-STATE] |
|
label, group acquired from XForms incorporated into XHTML 2 | |
Text: | Core DOM 1 from parsed HTML | ||||
Content selection: | User Agent | ||||
Font/Font Style Information: | Can set but can't get final format | User Agents final format styling using CSS2 APIs | User Agents final format styling using CSS2 APIs | User Agents final format styling using CSS2 APIs | |
* Description/Help: | limited to HTML 4.0 - Alt Text, title text | States and Properties Module [ARIA-STATE] | Acquired from XForms help, XForms hint incorporated into XHTML 2 | Acquired from XForms help, XForms hint incorporated into XHTML 2 | |
* Accessible value: | limited to form elements | States and Properties Module [ARIA-STATE] | User Agent acquires from XForms data model | User Agent Acquires from XForms data model incorporated into XHTML 2 | |
Coordinates (Bounding rectangle, etc.): | User Agents.platform accessibility API | ||||
Accessible Name: | User Agent determines from HTML tag | ||||
* Respond Desktop Font/Color Changes | partial (conflicts with CSS and JavaScript) | CSS3 - Requires User Agent support | CSS3 - Requires User Agent support | CSS3 - Requires User Agent Support | |
* Device independent navigation: | Use of xhtml2 landmarks through the <link rel""> | XHTML 2 Role land marks | XHTML 2 Role land marks | XHTML 2 Role land marks, and Access Key replacement | |
*Accessibility API Mapping: | partial - User Agents | User Agent to map new AAA properties and role meta data to the platform accessibility API | |||
Provide focus to all active elements | TABINDEX=-1 best practice | TABINDEX | TABINDEX | All XHTML 2 elements shall be focusable. |
Table 2.0 shows the roadmap for filling the gaps in Table 1.0. The effort is designed to fill the gaps for HTML content. Required components marked with an asterisk indicate gaps to fill in the roadmap. As you move to the right in the table a transition to more declarative markup will take place. The use of declarative markup reduces the enablement effort required while better addressing accessibility. Declarative markup is used to build accessibility in rather than bolting it on. Subsequent sections highlight the requirements and standards efforts set to address the gaps.
A way to summarize this chart is that the critical extensions needed to make accessible dynamic Web content accessible for XHTML 1.x, that may be made interoperable with assistive technologies, will be provided through the use of:
In short, these changes will be used to fix the dynamic accessibility of scripted Web content, in particular the use of JavaScript with HTML markup. The next section will describe how these are used all together. Less critical for XHTML 1.X but will be very helpful for accessibility will be the descriptive extensions to XML events and the new access element for XHTML 2.
It should be noted the working group may shift the contents of what is in each namespace at the time of last call.
Each platform accessibility API has the notion of a role for a GUI object. This is the case for Java Accessibility API [JAPI], Microsoft Active Accessibility [MSAA], Apple Accessibility for COCOA [AAC], and the Gnome Accessibility Toolkit (ATK) [GAP], or UI Automation for Longhorn [UIAUTOMATION] (called content type in UI Automation). To add this information we use namespaces to pull in the role attribute from XHTML2 and assign it a value from the GUI Role Taxonomy. Since the XHTML 2 role attribute takes a qname we can use a qname to reference the role attribute from the W3C WAI GUI Role Taxonomy. In the following example we use namespaces to assign a role of menu to a div. Menu is defined in the GUI role taxonomy.
<?xml version="1.1" encoding="us-ascii"?> <!DOCTYPE html PUBLIC "Accessible Adaptive Applications//EN" http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> <html xmlns="http://www.w3.org/1999/xhtml" xmlns:wairole="http://www.w3.org/2005/01/wai-rdf/GUIRoleTaxonomy#" > <body> <div role="wairole:menu"> File </div> </body> </html>
A decision by the HTML working group was to make role a qname
and allow others to define sets of standard roles. Furthermore,
these roles should be extensible. It is the intent of the HTML
working group to have these qnames refer to RDF which defines the
element. The RDF should ultimately define the object. The Roles for Accessible Rich Internet Applications [ARIA-ROLE], covered later in this roadmap,
introduces the purpose of this taxonomy and how RDF can be used to
describe these roles.
Note: For developers using Firefox 1.5, the role attribute
was used from the XHTML 2 namespace. Creation of an XHTML role
module for XHTML 1.X only occurred recently. For those doing
development on Firefox 1.5, the role attribute from XHTML 2
must be used. This being corrected in Firefox 2.0 although Firefox
2.0 will support role in both the old XHTML 2 and 1.1 namespaces.
For Firefox 1.5 supporters, this is the equivalent
markup:
<?xml version="1.1" encoding="us-ascii"?> <!DOCTYPE html PUBLIC "Accessible Adaptive Applications//EN" http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> <html xmlns="http://www.w3.org/1999/xhtml" xmlns:wairole="http://www.w3.org/2005/01/wai-rdf/GUIRoleTaxonomy#" xmlns:x2="http://www.w3.org/2002/06/xhtml2" > <body> <div x2:role="wairole:menu"> File </div> </body> </html>
<?xml version="1.1" encoding="us-ascii"?> <!DOCTYPE html PUBLIC "Accessible Adaptive Applications//EN" http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> <html xmlns="http://www.w3.org/1999/xhtml" xmlns:wairole="http://www.w3.org/2005/01/wai-rdf/GUIRoleTaxonomy#" xmlns:waistate="http://www.w3.org/2005/07/aaa" > <body> <div role="wairole:menu" waistate:haspopup="true"> File </div> </body> </html>
A Windows user agent may now map this property to the Microsoft Active Accessibility state of STATE_SYSTEM_HASPOPUP. Adding or removing this state would result in the Windows user agent sending an EVENT_OBJECT_STATECHANGE event to the assistive technology. The task of the JavaScript page author would be to maintain this state attribute which can easily be done through the use of Document Object Model calls.
Virtually all adaptive technology solutions, such as screen readers and onscreen keyboards, need to know what object currently has focus. For example, I might want to insert text into the current object with focus or I might want to announce information about the object that has focus. Today, HTML 4.01 and XHTML 1.x limit script authors to only being able to provide focus to form and anchor elements yet the Document Object Model Specification allows all elements to receive events including keyboard events. This means that HTML, by design prohibits script authors from making all HTML elements keyboard accessible. This single problem has impacted the usability of Web pages whereas gaining access to all elements means using the tab key on desktop browsers. This slow, unproductive, approach makes it difficult for portal navigation whereas all active elements must be tabbed through to get to an active element in the last portlet in a document. To solve this problem in XHTML 1.x we are incorporating a feature in Firefox and IE to define the tabindex for -1. This allows a script author to give an element focus without placing it in the tab order: The following table describes these changes which will be incorporated into the new Accessible Adaptive Application specification.
tabindex attribute | Focusable with mouse or JavaScript via element.focus() | Tab navigable |
---|---|---|
not present | Follows default behavior of element (yes for form controls, links, etc.) | Follows default behavior of element |
Negative, e.g. tabindex="-1"
|
Yes | No, author must focus it with element.focus() as a
result of arrow or other key press |
Zero, e.g. tabindex="0"
|
Yes | In tab order relative to element's position in document |
Positive, e.g. tabindex="33"
|
Yes | Tabindex value directly specifies where this element is positioned in the tab order. These elements will be positioned in the tab order before elements that have tabindex="0" or that are naturally included in the tab order (form elements and links) |
The following example shows the introduction of TABINDEX to provide focus to a DIV having the new accessibility meta data:
<?xml version="1.1" encoding="us-ascii"?> <!DOCTYPE html PUBLIC "Accessible Adaptive Applications//EN" http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> <html xmlns="http://www.w3.org/1999/xhtml" xmlns:wairole="http://www.w3.org/2005/01/wai-rdf/GUIRoleTaxonomy#" xmlns:waistate="http://www.w3.org/2005/07/aaa" > <body> <div role="wairole:menu" waistate:haspopup="true" tabindex=-1> File </div> </body> </html>
XHTML 2 will allow all elements to receive focus, eliminating the need for this tabindex modification.
In addition to the common roles which will reside in the GUI Role Taxonomy, both XHTML 2.0, and the XHTML Role module defines a collection of common roles ([XHTML-ROLE], Section 26.1) that define pertinent parts of a document for the purpose of accessibility. User agents may incorporate device equivalents, such as key mappings in the case of a desktop user agent, to navigate to these sections of a document independent of the Web site. The addition of these semantics allows the user agent to provide standardized navigation to common document sections. This is especially important for portals to improve the usability. These may be used as attributes in XHTML 1.x by applying them to sections of the document as shown in this example.
<?xml version="1.1" encoding="us-ascii"?> <!DOCTYPE html PUBLIC "Accessible Adaptive Applications//EN" http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> <html xmlns="http://www.w3.org/1999/xhtml" xmlns:wairole="http://www.w3.org/2005/01/wai-rdf/GUIRoleTaxonomy#" > <body> <div role="navigation"> /* navigation is defined in the XHTML role attribute module */ <ul> <li> <a href="www.w3.org/WAI">WAI</a></li> <li> <a href="www.ibm.com/able">IBM Accessibility Center</a></li> <li> <a href="www.opera.com">Opera Software</a></li> <li><a href="www.mozilla.org/access">Mozilla Accessibility</a></li> <li><a href="www.ubaccess.com">UB Access</a></li> </ul> </div> </body> </html>
An an alternative to placing these landmarks on the XHTML elements is to place them in the header through the use of link which is also applicable to HTML 4.01. Here, Link's rel attribute may take an string value which may be used to represent the landmark. Additionally, a description may be added to the title as shown in this example:
<html lang="en-us"> <head> <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1" > <title>Career Center Home Page</title> <link rel="x2:banner" href="#title" title="Career Center"> <link rel="x2:navigation" href="#nav1-links" title="Career Counseling Resources"> <link rel="x2:navigation" href="#nav2-links" title="Resources by Audience"> <link rel="x2:navigation" href="#leftnav-links" title="Career Center Services"> <link rel="x2:navigation" href="#bottomnav-links" title="Quick Links"> <link rel="x2:main" href="#content-col-1" title="Welcome to the Career Center"> <link rel="x2:secondary" href="#content-col-2" title="Featured Events"> <link rel="x2:secondary" href="#contact-info" title="Contact Info"> <link rel="x2:search" href="#search" title="Search Career Center website"> <link href="styles/career-main.css" type="text/css" rel="stylesheet"> </head>
This is being worked by the WAI coordination group as a best practice. The example above was taken from the header from the Career Center Web page at the University of Illinois at Urbana-Champaign. Students form this university, under the Jon Gunderson's guidance created Accessibility extensions for Mozilla/Firefox, in part, allow a page author or user to launch a table of contents to viewer to list the landmarks. Keyboard navigation of the table of contents causes the corresponding document section to be highlighted.
Figure 3.0 Table of Contents
generated from navigation landmarks in the header
Figure 3.0 shows the Accessibility extensions for Mozilla/Firefox from the University of Illinois at Urbana-Champaign to render the document landmarks. This picture shows the Firefox Browser rendering the University of Illinois Career Career Center home page. In this example. The "Table of Contents" viewer is shown launched from the extension on the toolbar and the secondary "Featured Events" is selected resulting in that section of the document to be highlighted in yellow. The Table of Contents viewer lists the following section type/title pairs:
It will be necessary to describe how to interact with a document element having a specific role. This is critical when defining roles not included in the common set specified previously. It should be possible to describe UI widgets and how to interact with them. RDF is a way for using data to describe data and provides a W3C standards-based approach to represent this information.
Figure 4.0 Partial RDF Map for a GUI Role Taxonomy
Figure 4.0 Shows a basic RDF mapping which defines a set of terms and relationships defining an object. At the center is a Widget object that defines common states and properties for all GUI widgets. The Button object extends Widget and inherits defined accessibility properties from the superclass Widget. It also defines a relatedConcept property to a Link object. The ButtonUndo role extends Button and an instance of extends of Button. It has a relatedConcept of an HTML input object. ButtonUndo will introduce Dublin Core meta data such as the description of the object. The terms relatedConcept and requiredState are terms which which will be defined as part of the corresponding RDF schema. Each role instance will represent standard Roles found in platform accessibility APIs platforms like Windows and Gnome as well as content structure. These roles will form the taxonomy. Each RDF role may be referenced using a qname from Host XML markup such as XHTML:
<div role="wairole:ButtonUndo"> wherebywairole:ButtonUndo
expands to:
http://www.w3.org/2005/01/wai-rdf/GUIRoleTaxonomy#ButtonUndo in the
button markup
The power of this design is it allows user agents to go back into the corresponding RDF markup and determine what properties it supports for Accessibility API mapping. Also, new Web authoring tools can use the self describing RDF to help test for accessibility conformance to semantic Web markup representing the role. Additional, middleware solutions can now make intelligent transformations of Web Content by processing the semantics behind rich browser and rich structured frameworks to adapt accessible solutions for a broader user base. Our immediate goal is to fix the accessibility problem with scripted Web content. Assistive technologies will use the standard roles to determine how to render most content. For custom widgets they should be able to determine how to interact with it using the RDF document defining the roles.
Figure 5.0 shows a DHTML example of using XHTML, JavaScript, and CSS to produce a GUI-like application. This example of a DHTML Web page being developed in IBM shows a notebook tab with a data grid that behaves like you would expect a GUI to. Keyboard navigation within the data grid is done with arrow keys. Navigation among the page tabs is done with the arrow key. Navigation between the notebook tab, the edit fields, buttons, and the data grid is done using the tab key. Accessible role and state meta data from the GUI role taxonomy and Accessible Adaptable Application specifications, respectively are added as attributes to each of the XHTML elements repurposed as GUI widgets dynamically. The user agent, in this case Firefox, maps this information to the platform accessibility API. Figure 6.0 shows the Microsoft Active Accessibility rendering of the new accessibility markup provided on the DataGrid page tab which has focus.
Figure 6.0 Microsoft Inspect Tool rendering of the page tab
DataGrid
Figure 6.0 is a Microsoft Inspect 32 rendering of the DataGrid Page page tab in Figure 5.0. Inspect32 provides Microsoft Active Accessibility Information. It shows the accessible role of "page tab" and accessible state information of focused, focusable, and linked. There are no page tab elements in XHTML. Here, an XHTML DIV element is repurposed by a JavaScript controller to look like a notebook tab. It is now able to receive focus unlike in standard XHTML 1.X. It can do so without requiring tabbing. With these specifications the script author may now add the accessibility properties to support platform accessibility API. Accessible state properties for the DataGrid page tab are shown as focused, focusable, and linked. Unlike a GUI application, the author need only enable their application once for multiple operating system platforms.
HTML is a device dependent solution. An impaired user needs to make an intelligent decision as to what actions they can perform on a document element. An impaired user wishing to enumerate the available actions that may be performed on a page element is confronted with the non-informative information provided by HTML's intrinsic events ([HTML], Section 18.2.3). The information provided is very device specific and non-informative. For example, what action does an event handler perform in response to an onclick event? This is problematic for blind, low vision, and mobility impaired users. There may be more than one action to be performed on a given element. A mobility impaired user would like to enumerate the handlers and make a decision. Generating the appropriate mouse event is problematic. Blind users and low vision users would like to make informed decisions before they accidentally make a purchase. Some information can be gleaned from the UI by sited users, however some of this information may be obscured from view due to the use of a magnifier by low vision users.
The HTML working group is working to address this problem through the use of XML events [XML-EVENTS]. XML Events provide XML languages with the ability to uniformly integrate event listeners and associated event handlers with Document Object Model (DOM) Level 2 event interfaces [DOM-EVENTS]. The result is to provide an interoperable way of associating behaviors with document-level markup. Through the use of XML this is done declaratively. Having the ability to use XML to integrate listeners and handlers will allow us in in future versions of the XML event specification to tie a descriptive purpose to the handlers.
This problem also manifests itself when encountering event handlers authored via element script. XML Events and XML Handlers solves the first half of the accessibility problem by introducing a clean declarative syntax while encouraging authors to use high-level event names and descriptions. This gives the author the flexibility to assign device specific mappings to event handlers while providing for added semantics. In XForms 1.0, we have gone one step further by introducing declarative action handlers for commonly used event handlers, e.g., setfocus. The combination of REL navigation semantics, XML events, XML handlers, and XForms declarative markup set the stage for an access key replacement. Access key suffers from a lack of semantics and support for device independence. Dynamic Web content implies the use of event handlers. From an accessibility perspective we need to assist the user in determining how to interact with a Web application in a device independent way. The provision for named events and the ability to bind event handlers to document element events in the XML Event specification is critical. At the time of binding we can incorporate the additional Role information as well as make use of the name information to provide semantics about what the event handler does. In the future, this information should be mapped to the available accessibility infrastructure by a, UAAG conforming, user agent on the platform. Once complete, a user agent conforming to the DOM 3 event specification will allow for enumeration of event handlers with corresponding descriptions. The ability to ask this information has been added to the EventTarget interface in DOM3 events through the hasEventListenerNS method.
Platform accessibility API like the Gnome ATSPI Accessible Action interface may now enumerate these actions for assistive technologies so that the user may be provided with descriptive information to make an informed decision as to which action to take.
Screen Reader user would be able to pop-up a list of event handlers in a list such as follows:
XForms [XFORMS] addresses the accessibility problem in the following ways:
The first benefit is less JavaScript. XForms provides a common set of available actions that can be performed on a form. XForms is a declarative markup. The engineers who designed XForms took the standard use cases for how we create forms and built in support for the standard actions performed in a Web form. From an accessibility perspective, this means these functions should be available from the browser independent of platform. So, rather than guess what an onclick does on a form, the user should be able to access the available functions and Notification events that are used in that specific form as defined by the XForms standard.
You are also notified of events which actually occur during form operation. These types of events allow applications, like a screen reader, to provide alternative modalities such as speech.
The use of a standard data model reduces extensive author enablement work in the area of adding accessibility properties to content. The model, when bound to XForms elements, may be used by the user agent to directly support the platform accessibility API as shown in this abridged table:
This saves authoring time by levering the XForms model to support accessibility directly. IBM is beginning development platform accessibility support for XForms in 2005 as a contribution to the open source community.
XHTML 2 [XHTML2] is being designing to build accessibility in from the ground floor. It is clear from the roadmap that some XHTML 2 accessibility features are borrowed through the use of name spaces to extend XHTML 1.1 through modularization. XHTML 2 incorporates the role attribute, document landmarks, XForms, XML Events, and a number of other accessibility features. It introduces an access element designed to address device independence, semantic navigation, and activation. It also benefits from the WAI GUI Role Taxonomy which is a cross-cutting technology. The XHTML 2 specification is still in working draft form. A future list of the extensive accessibility features of XHTML 2 will be necessary.
User agents are required to provide interoperable interfaces to assistive technology run times to produce a usable experience for users with a specific disability. A task of the roadmap will be to define requirements for UAAG. This is necessary provide a cohesion between user agent and author supplied state information.
User agents may need to adjust their user interface in response to state information stored in document elements. For example, an element having the focusable attribute set to true should result in a desktop user agent's placing the element in the keyboard tabbing order.
The WAI PF working group and IBM are working with assistive technologies to ensure support of the roadmap.
The WAI PF working group will be looking at ways to address authoring tools requirements for testing the accessibility of scripted Web content. The development of authoring tools to assist the author in testing dynamic Web content for accessibility is important. Today's accessibility fixers and checkers test to static Web content. Accessibility properties will change during Web page use. Web authoring tools should consider reusable accessible widgets and prompt for variable accessibility information such as title text. The use of RDF in the WAI GUI Role Taxonomy should be used to ensure which properties may be supported by which roles. W3C Authoring Tools Accessibility Guidelines 2.0 [ATAG2] is developing requirements that will better address authors will creating accessible, dynamic, scripted Web content.
At this time, IBM is doing much of the prototyping of this work. Sample markup has been contributed by IBM and may be found on the Mozilla accessibility Web site. IBM is working on a best practices document and is contributing to WCAG 2.0 implementation techniques. UB Access is contributing much of the RDF design supporting the Role taxonomy work. At this time these guidelines are not yet made available.
The following table is the technology deliverable Timeline for the Dynamic Accessible Web Content Roadmap.
Specification. | WD 1.0 | WD 2.0 | LC | CR | Implementation Report | PR | REC |
---|---|---|---|---|---|---|---|
States and Properties Module for Accessible Rich Internet Applications [ARIA-STATE] | Q3 2006 | Q4 2006 | Q4 2006 | Q1 2007 | Q1 2007 | Q2 2007 | |
Roles for Accessible Rich Internet Applications [ARIA-ROLE] | Q3 2006 | Q4 2006 | Q4 2006 | Q1 2007 | Q1 2007 | Q2 2007 | |
XHTML Role Attribute Module [XHTML-ROLES] | Q3 2006 | Q3 2006 | |||||
XHTML2 Specification
[XHTML2] (Includes XForms) Note: In last call we point out that we may shift the namespace of the role attribute, and its standard role landmarks, from XHTML 2 to our own, if XHTML2 is not in last call. In CR we can include both namespaces, with the explicit note that whichever is most useless is definitely at risk of being dropped. |
Q2 2005 | TBD | TBD | TBD | TBD | TBD | |
Forward Looking Efforts: | |||||||
XML Events [XML-EVENTS] with handler descriptions (not critical for XHTML) | TBD | TBD | TBD | TBD | TBD | TBD | |
XForms 1.0 [XFORMS] | 3Q03 |
The following table provides implementations of the new accessibility specifications for supporting dynamic accessible Web content.
Implementation | Category of Implementation |
---|---|
Firefox 1.5 (including MSAA support) | user agent |
GW Micro Window-Eyes 5.5 - with Firefox support | assistive technology (screen reader) |
JAWS 7.0 with Firefox support(partial) | assistive technology (screen reader) |
Windows Magnifier (with IE or Firefox 1.5) | assistive technology (magnifier) |
Internet Explorer from DOM API | user agent |
Mozilla Firefox Test samples | content |
Dojo JavaScript Toolkit | content - in Development |
Mozilla/Firefox Accessibility Extensions Test Tool and test suite | content and accessibility test tool |
Bindows Object Oriented Platform for AJAX Development | content - in Development |
WAI Role Taxonomy Extension Tool | Tool - in Development |
This roadmap is designed to address the accessibility of dynamic, scripted, Web content that may be rendered in today's browser while bridging to future declarative standards, such as XHTML2, in the W3C. The extensions being created for XHTML 1.X are intended to be cross-cutting. This roadmap clearly indicates that WCAG, ATAG, and UAAG should be in lock step for a common strategy. All groups are impacted and are interdependent. Additionally, close work with mainstream working groups, like HTML, CSS, and Device Independence will not only produce a working solution but it will produce a working solution that will have a broader impact than just XHTML.
The following table indicates the specifications to track which address the gaps in the Roadmap.
Specification. | XHTML 1.0, XHTML 1.1 | XHTML 2.0 |
---|---|---|
Accessible State Meta Data | States and Properties Module for Accessible Rich Internet Applications | XHTML2 Specification (includes XForms) |
TABINDEX Correction for focus | States and Properties Module for Accessible Rich Internet Applications | All elements are focusable in XHTML 2 |
Accessible Role Meta Data | Roles for Accessible Rich Internet Applications | Roles for Accessible Rich Internet Applications |
Role Attribute | XHTML Role Attribute Module | XHTML2 Role Specification |
This section is informative.
The following contributed to the development of this document.
Jim Allan (TSBVI)
Judy Brewer (W3C/MIT)
Michael Cooper (W3C/MIT)
Donald Evans (AOL)
Kentarou Fukuda (IBM)
Alfred S. Gilman
Andres Gonzalez (Adobe)
Georgios Grigoriadis (SAP AG)
Jeff Grimes (Oracle)
Jon Gunderson (UIUC)
Alex Li (SAP AG)
Charles McCathieNevile (Opera)
Dave Pawson (RNIB)
David Poehlman (State of MD)
Janina Sajka (FSG)
Richard Schwerdtfeger (IBM)
Lisa Seeman (UB Access)
Ryan Williams (Oracle)
Gottfried Zimmermann (Access Technologies Group)
Special thanks to Aaron Leventhal for effort and insight as he implemented a working prototype of accessibility API bindings.
Christian Cohrs, Becky Gibson, Ari Gruen, Barbara Hartel, Jael Kurz, Vitaly Sourikov.
This publication has been funded in part with Federal funds from the U.S. Department of Education under contract number ED05CO0039. The content of this publication does not necessarily reflect the views or policies of the U.S. Department of Education, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.