The Linux Screen Reader is a screen reader application. The core of the system is the AccessEngine, a message pump that receives events via the Assistive Technology Service Provider Interface (AT-SPI) on the GNOME desktop. The commands available for controlling the screen reader are determined by cascading scripts that completely define the user experience.

You will recall IBM committed to solving the disability issue for MA. I don't know that there is a connection, but if enough folks help, I can see how there could be. If you are so inclined and you are not currently working on another assistive application, you might want to have at it and take it for a spin. Perhaps you will wish to add Festival support or braille definitions, or maybe you have an idea for extensions, and if so, the authors ask you to contact them. I note they mention joysticks. The homepage has contact info at the bottom of the page.

They say the license is free for noncommercial use; a license needed for commercial use. Look into that yourself carefully, as usual, as IANAL. Even if I were, I wouldn't be your lawyer, and I don't know your personal views on licenses, so I leave that part to you. PeteS notes that it says: "The scripts can be loaded and unloaded by the user at run time, and custom scripts can be written to improve the usability of certain applications and interactions."

The longer description from alphaWorks, Overview tab begins like this:

What is the Linux® Screen Reader?

The Linux Screen Reader (LSR) is an application that transforms the contents of the computer screen to other media, enabling non-visual access to the graphical Gnome desktop environment. The basic function of the screen reader is to report changes in the state of an application as a user interacts with it using the keyboard or other system input device. What is reported to the user and what commands are available for controlling the screen reader are determined by cascading scripts that completely define the user experience. The scripts can be loaded and unloaded by the user at run time, and custom scripts can be written to improve the usability of certain applications and interactions. How information is reported to the user and how the user gives input are determined by one or more configured input and output devices. Like scripts, devices can be loaded and unloaded by the user at run time, and custom device profiles can be written to support new methods of input and output (e.g. Braille, speech, switches, and joysticks).

The source code is currently at alpha level 0.1.0. The current implementation defines a keyboard device for input and a speech device for output using the IBM ViaVoice speech engine. A default script that responds to focus, selection, and caret changes in the active application is included. The default script also defines some basic keyboard commands for navigating applications, setting the speech rate, and reporting where the user focus currently lies. Documentation is included with the source code.

They are working on a user's guide as well as a run-time help facility. I also note that on IBM's "Human Ability and Accessibility Center" page, you can download a browser called Home Page Reader, which is described like this: "A powerful, talking Web browser for users who are blind or have low vision, Home Page Reader also helps developers test the accessibility and ease of use of Web sites."

You can find instructions from alphaWorks for the new LSR by clicking on the Requirements tab on the alphaWorks page, and there is a readme file in the download also. The instructions go like this:

******************************

Software: IBM text-to-speech (TTS) synthesizer (for speech output). Single user
 licenses of the IBM TTS run-time environment can be purchased from 
Capital Accessibility, LLC. Capital will provide the gnome-speech-ibmtts 
package required in order to make gnome-speech work with IBM TTS. 

Installation instructions for Linux Screen Reader 

The following instructions assume that the user is running Fedora Core 4 or 
above. LSR will work on other Linux systems, but it is untested and will 
require slight modifications to the installation procedure below. 

The following RPMs must be installed before LSR will work: 

gnome-common 
pyorbit-devel
gnome-speech 


1.	The first two required RPMs (above) can be gotten using yum, as 
follows: 
su
yum install gnome-common
yum install pyorbit-devel
exit 


2.	As of 3/1/06, the Gnome-speech package provided by Fedora Core 4 is 
missing a required binary file. The equivalent RPM from the Polished Linux 
Distribution can be used in its place. Download it from 

ftp://ftp.pld.org.pl/dists/ac/ready/i586/gnome-speech-0.3.6-1.i586.rpm and 
install it as follows: 
su
rpm -Uvh --force gnome-speech-0.3.6-1.i586.rpm
exit 


3.	Unpack the LSR source from the tar.gz file by typing the following: 
tar xzf lsr-0.1.0.tar.gz 


4.	Navigate into the lsr directory and type the following to complete the installation: 
./autogen.sh
su root
make install
exit 

For usage instructions, please see the included "readme" file.