Aaron Leventhal observes:
> Only links are key navigable or exposed via MSAA in the Firefox Flash
> player.
Actually, the links aren't exposed by Adobe Flash Player. They're just
part of the ordinary HTML content of the page that the Flash object is
embedded in, and exposed by Firefox itself. So it's even worse than this
description would suggest.
If you examine the page in both Internet Explorer and Firefox with the
Accessible Explorer from the Microsoft Active Accessibility 1.3 SDK it
becomes clear that while Flash Player 9 exposes lots of fun objects
(wombat, butterfly, scrapbook, and so on) in Internet Explorer 6, it
exposes *nothing at all* in Firefox 2.0. So it's hardly surprising that
it doesn't work with the same Window-Eyes screen reader used in Firefox
rather than Internet Explorer! Here's a screenshot of Accessible
Explorer dissecting the page which illustrates the scandalous difference
in the plugin's behaviour depending on the browser:
http://tinyurl.com/ymqxdz
In "Best practices for accessible Flash design" (August 2005), Bob Regan
wrote:
> All accessible Flash content must be tested using Microsoft Internet
> Explorer. At the time of publishing this document, Internet Explorer
> was still the only accessible browser available. The Mozilla Project
> has made some improvements with the Firefox Browser, and support for
> screen readers will soon be available. However, the version of the
> Flash Player that runs in Firefox is not yet accessible.
> Designers often question the wisdom of testing accessible content using
> only one browser and one platform. While the future of cross platform
> and cross browser accessibility looks promising (with the recent
> improvements at Apple and Mozilla), the reality is that today, people
> with disabilities are almost exclusively restricted to using Windows
> with Internet Explorer. For now, designers should feel comfortable
> testing their content using only this configuration.
(See
http://tinyurl.com/ybo9dm for the PDF of Regan's whitepaper.)
Even if that were true back in August 2005, it could hardly be said to
have been true in June 2006, when Adobe Flash Player 9 was released. In
October 2005, the 7.0 release of JAWS (the most widely used screen
reader and second on Regan's list of Flash-ready readers) added
supported for Firefox 1.5, then still in beta. In November 2005, Dolphin
released map files for the Firefox beta for HAL (fourth on Regan's
list). Later that month, Firefox 1.5 was released with MSAA support
arguably superior to Internet Explorer's:
http://www.mozilla.org/access/windows/at-apis#knowndif
Finally in the same November, the 5.5 release of GW-Micro Window-Eyes,
first on Regan's list, included first-class support for Firefox. JAWS
7.10, released like Flash Player 9 in June 2006, included even better
support for Firefox.
And since then, the rationale for exposing Flash content in Gecko-based
browsers like Firefox has only grown. While Internet Explorer 7 is still
struggling even to expose ordinary HTML like Q to screen readers via
MSAA, Firefox 3 (slated for release in May 2007) is making Ajax and
XForms accessible!
http://www-306.ibm.com/able/resources/firefox.html
And whereas Internet Explorer can only offer accessibility for Flash
content on one platform, Firefox 3 (it is hoped) will be integrated with
three: MSAA on Windows, AT-SPI on Linux, and Mac OS X's accessibility
APIs. Amazingly, there's free Firefox extension that turns Firefox into
a talking browser on Mac, Windows, and Linux:
http://www.firevox.clcworld.net/
Unlike any commercial screen reader (as far as I know), Fire Vox even
includes some support for aural CSS.
In conclusion, the Gecko engine and the Linux and Mac accessibility
frameworks deserve Adobe's support, and integration with them is crucial
if Flash "accessibility" is not to be properly considered a token
effort. As accessibility becomes more and more a legal requirement, if
Adobe fails to move, won't Flash lose out to accessible DHTML web
applications and video plugins?
Flash Player 9 for Linux has entered the Beta stage and is being shown
off on Ubuntu Linux:
http://tinyurl.com/yah5fx
Ubuntu is devoted to improving accessibility and usability generally,
and this is an excellent opportunity for Adobe to show their own
commitment to disabled users by beginning to add support for AT-SPI,
allowing free and open source screen readers to access their content for
the first time.
Moving on, Jason White asks:
> are the people working on the GNU Flash client planning to implement
> support for the Gnome Accessibility API or other, comparable
> mechanisms?
Gnash (GNU Flash) development obviously lags somewhere behind Adobe
Flash Player. Implementing the Accessibility class has been on their
TODO list for some time. I have now added a suggestion to their
bug-tracker that they should support these accessibility frameworks,
particularly AT-SPI and MSAA. Interested parties may wish to comment at:
https://savannah.gnu.org/bugs/index.php?18194
