Re: Mercator Project

From: Brian Buhrow (buhrow@moria)
Date: Mon Apr 12 1993 - 23:07:32 PDT

        Well, Mike, I do remember the lady and, in fact, saw her here in
California last summer. She was working as a summer researcher at Sun
MicroSystems and was going to give me a demonstration of their X-windows
project. Unfortunately, due to transportation troubles I was unable to get
there in time to actually see a demonstration. I did, however, get to talk
with her and address some of the issues of accessability and usability with
her. First, the three-dimensional sound stage had to go. Currently, (as
of September 1992), their Talking X-windows manager ran on the DECTalk
through the serial port that all Sun workstations have. The driver was
written in C and, according to her, could be easily modified to drive any
speech synthesizer. They were working on plans to put in an internal
Sun-Produced speech algorithm which would be used through the internal
audio port, but they were having a problem with the speech response time.
Beth expressed the feeling that the Sun speech wasn't going to be a reality
for a good bit of time.
        The window manager, she said, had to be re-written from the ground up.
and so that was a problem they had gotten around, but not one that they
had totally solved. One of the problems with their new window manager was
that if you opened a window, then that window spawned another window, you
couldn't go back up the window tree to get to any windows you may have had
opened on the top level. I said that this was unacceptable, considering
the fact that I can run 9 windows at a time on any Unix box that I'm
attached to and get to any one of those windows in less than three
key-strokes. She also said that they had opted to fore-go any mouse
control and that mouse control was simulated with the arrow keys on the Sun
keyboard. Of course, sighted colleagues could use the mouse, and so could
we, but at a great loss of efficiency.
        In terms of what they were going to support, she said that they were
going to support the athena X window widgit set, and, with the grace and
funding of Sun, their Ollit widgit set. Between these two widgit sets,
most of today's text-based X-windows applications should be relatively
accessible. She had hoped to let me play with xmh-mail, but as I say, I
didn't have enough time.
        In terms of availability, she said that they hoped to have something
working by December. It would be primative, but it would be something we
could hack around with on the net, considering that it was going to be
produced by Georgia Tech. I expressed an interest in trying it out and
seeing what sorts of tricks I could make it do in our Sun environment. We
use the athena widgit set on a daily basis and have access to all the
necessary compilers.
        Up till now, I have had no contact with her again. I've looked for
files in anonymous ftp sites with references to talking X-windows
applications, but nothing has come to light. I'm glad to see someone else
was pestering as well. Perhaps we can stage a strategically planned
approach whereby one of us will eventually get an answer. And, with any
luck, a piece of software to compile and play with to our heart's content.

P.S. I've kept in touch with Gregory Pike of IBM, who came to our New
Orleans convention to help Jim Thatcher show Screen Reader for OS/2 for the
first time and, who, as you may remember, was working on the Screen Reader
for X-Windows under AIX (IBM's version of Unix). He tells me that they
have tty access and access to the basic X screen, but that they have a good
bit of distance to go. It hasn't been getting as much attention as the
OS/2 project has, because the OS/2 thing is an actual product that needs
updating and the "need" for supporting the X-windows Unix project hasn't
come to fruition yet. With any luck, it will soon. Still, I'm rooting for
Beth because I don't have an RS-6000 and because I think there will be some
sources available from that. That means, of course, that we won't have to
wait a year to make a fix after the first bugs surface.

This archive was generated by hypermail 2b29 : Sun Dec 02 2012 - 01:30:03 PST