Re: accessibility question

["Eric S. Johansson" <esj harvee org>]

Sven Neumann wrote:
> Hi,
>
> I found the link I gave you is pretty much outdated. Try these
> instead:
>
>   http://developer.gnome.org/projects/gap/
>   http://developer.gnome.org/doc/API/2.0/atk/atk.html

thank you for these links.  Unfortunately, they confirmed my fears which 
is that is really focused on blind accessibility first, extreme 
handicapped user second (i.e. quadriplegics) and speech recognition 
dependent users are an afterthought.  And it's (understandably) only 
Linux focused.

for example, in the slides they talk about integrating with Sphinx 4 (a 
 research class recognizer) for command and control.  If they had 
talked to any significantly disabled person especially those with RSI, 
they would find out that command and control is the last thing on their 
minds.  It's important especially for those of us who are hurting but 
full English (or wherever language) dictation is far more important 
because then you can communicate with other people through tools like 
instant messenger or e-mail.  It's the foundation from which handicapped 
users can work and make money on which to live.  Command and control 
will not cut it as a stand-alone accessibility aid.

my hypothesis is that properly done, a blind accessibility and speech 
recognition accessibility interface would have about a 90 percent 
overlap.  We both need the same type of information, it's just how we 
use it that changes.

my initial query was motivated first by trying to find out if it's 
possible to connect NaturallySpeaking to GTK and make NaturallySpeaking 
features like select and say available to all applications using GTK. 
The second was to help define a project better for OSSRI.

The project I am advocating is moving NaturallySpeaking to Linux via 
wine and then bridging NaturallySpeaking to local accessibility tools. 
By advocate this because it will provide us with the level of 
functionality we need relatively quickly while open source speech 
recognition products eventually get to the level of NaturallySpeaking.

OSSRI(Open Source Speech Recognition Initiative) is a recently formed 
nonprofit (nonprofit status pending) to provide a framework for 
shepherding speech recognition projects down the path that serves speech 
recognition dependent handicap users.

---eric