a very long time. The very first attempt to make X applications accessible with assistive technologies, so far as I know, was the UltraSonix project, originally developed for Sun workstations.
Wow. That's a blast from the past, and it was really developed for Xt-based widget sets that ran on Sun workstations and Digital workstations. I think you're right in that it was the first, and I think it may also be the first attempt at doing an assistive technology using a service-based model.
If you dig deep enough, you'll see that I was involved in the infrastructure side of Mercator/UltraSonix: the remote access protocol (RAP). RAP, the ICE X Rendezvous Protocol, and AccessX were some of the first collaborations I did with accessibility 'way back when.' :-) As a result, I have somewhat of a twisted view of the modern world. To me...it's all the same stuff just reinvented several times, with mostly the same complaints and challenges.
One of the most promising changes that I've seen, however, is the warming up that has come from the developer community. It used to be the incomprehensible arguments of "Sure, you want me to 'dumb down' the world for people who don't really matter" and "I want to play pinball with my Shift keys. Stupid AccessX and stupid people with physical impairments. I hate them all." Now, I have intelligent conversations with developers, most of whom "get it" and understand what it means to do the right thing. That is great.
Will