Re: Forming an Accessibility Steering Committee

[Peter Korn <Peter Korn Sun COM>]

Hi Brian, gang,

> > As Linux becomes a real option for consumers it would be fantastic if
> > those needing accessibility were drawn to it as it provides the best
> > options. e.g an eye tracking system would be a winner as commercial
> > systems are so expensive
>
> Perhaps it would be useful if someone put together some marketing
> information comparing how much it costs to make a Windows or MacOS
> machine as accessible as Linux is for free.  It would be handy to put
> together the cost for different kinds of accessibility use cases.

This is definitely useful.  But cost is not the only key 
dimension/differentiator.  Taking control of ones own destiny is 
another.  Jan Buchal has spoken eloquently on the free-as-in-freedom 
aspects of open source accessibility.  As a long time participant and 
observer of the commercial accessibility marketplace and disability 
community, there has been a very effective dynamic of disability 
advocates and users spending their time "advocating" for access (ranging 
from getting laws passed requiring accessibility to suing folks who 
don't comply with those laws, and everything in between [including a lot 
of consciousness raising]).  To me, FOSS a11y presents this community 
with a new and powerful tool to achieving access ends - the ability to 
participate in technology solutions directly.  I'd like us to ensure 
that this is a prominent part of the UNIX accessibility message.

> I think it would be useful to frame Linux a11y as being a vehicle
> for providing affordable accessibility to the world.  Many people
> with disabilities cannot afford expensive software.  Providing free
> software solutions opens new opportunities for people with disabilities
> to find jobs and to interact with people.  We should make more of an
> effort to push the humanitarian angles of Linux a11y.

Yes, and the participatory aspects of this humanitarian angle is key. 
By the way, you'll hear/see me calling it UNIX accessibility instead of 
Linux accessibility.  I know there are issues with the UNIX trademark, 
but our work is larger than Linux (Solaris, NetBSD, etc.), and larger 
than GNOME (X, KDE, etc.).  We need a term that encompasses all that.

> It is my understanding that MSAA (MicroSoft Active Accessibility)
> doesn't work with common programs like Office, and that users need to
> buy special MSAA-enabled versions of Word Processors, Spreadsheets, etc.
> I would think this would make setting up a functional a11y Microsoft
> machine quite expensive.  No?

MSAA is implemented in MS-Office.  The problem with MSAA is that it is 
only a small subset of ATK/AT-SPI; too much necessary information cannot 
be easily or well conveyed via MSAA (though Aaron Leventhall among 
others can tell you about the many hacks folks have used trying to 
shoehorn info into MSAA).  UI Automation is the MS replacement for MSAA 
(well, one of them; UIA Express is another).  It appears to be rich 
enough to behave as a good accessibility API.

Microsoft has not implemented UIA on much of anything of any importance 
in their software library.  Certainly not on MS-Office or IE.

The result is that to deliver a really rich and powerful experience, AT 
vendors have had to use proprietary and app-unique APIs to get at info. 
 The proprietary COM interface to Word (as distinct from the 
proprietary COM interface to Excel, different again for PowerPoint, 
different again for IE).

Utilizing very much the same API calls, methods, fields, etc. as 
ATK/AT-SPI, IBM introduced IAccessible2 as an extension to MSAA that 
provides a rich accessibility interface to Windows.  They and others are 
in the process of implementing IAcc2 support on a bunch of apps, 
including Lotus Symphony, Firefox, and OpenOffice.org.

In fact, it may be worth noting as part of our messaging to the world 
that the GNOME accessibility framework is a close sibling to the Java 
Accessibility API, the UNO Accessibility API, the XUL accessibility API, 
IAccessible2, and WAI ARIA.  They all have a common root (the Java 
Accessibility API), and together they represent a much more integrated, 
portable, interoperable set of accessibility interfaces than what 
Microsoft and Apple have done.

> Considering that in the U.S.A. Section 508 requires that publicly
> funded institutions (such as libraries and schools) need to provide
> accessible access to the internet, it would be helpful to show the
> cost savings using Linux based a11y solutions.

Absolutely!  Another key benefit is that it is all built-in.  A key 
issue in situations like schools & libraries is that expensive and 
brittle 3rd party AT is only installed on one or two systems (as few as 
they can get away with), and those systems are essentially "kept out of 
circulation".  E.g. in a lab of 10 machines, the JAWS machine is special 
and in order to not "mess things up", commonly kept out of the pool (of 
now, 9 machines), until such time as the blind user needs it.  When the 
second blind user comes along.... oh well.  And when 10 non-blind users 
want to use the lab, well, one gets to wait.

But when every computer has the screen reader installed on it, all or 
none can use it.  And when it is a core part of the OS/platform (that 
isn't replacing keyboard drivers and chaining video drivers and doing 
other system-level hackery), you don't worry so much about "messing 
things up" by allowing others to use the system.  Of course, UNIX in 
general has a much stronger security model and stronger boundaries 
between users, making this less of a concern overall as well.

> I'd think this sort of information would make sense to include or
> reference prominently on the Wiki and in other GNOME marketing
> blurbs.


Absolutely!


Regards,

Peter Korn
Accessibility Architect,
Sun Microsystems, Inc.