Re: Orca access to apps in wine?

[Peter Korn <Peter Korn Sun COM>]

Hi Joanie,

> Hi all.  There's one Windows application that I'd really like to be able
> to use in Linux:  Duxbury.  I am not surprised that Orca sees it as an
> "inaccessible".  My question, however, is this:  Is the reason it is an
> "inaccessible" due to how Wine is implemented or how Duxbury is
> implemented?

The answer is "both and", and stems from the fact that all graphical 
accessibility in GNOME/UNIX is API-based access: the AT makes queries of 
the objects in the applications, and based on the answers it gets it 
re-presents the application in an alternate modality -> in speech, 
Braille, magnification, etc.

In order for Duxbury to work in WINE on Linux, we would need the 
following things:
1. Windows to have defined a rich accessibility API that includes more 
or less everything our AT-SPI contains
2. Duxbury to have implemented that non-existent Windows accessibility API
3. WINE to have included a bunch of glue code that bridges the 
non-existent Windows accessibility API from those Windows apps that 
implement it to the UNIX AT-SPI.

By contrast, Java apps do work in UNIX because:
1. Java has defined a rich accessibility API that includes nearly 
everything in AT-SPI (and in fact, the Java accessibility API is the 
'parent' of AT-SPI)
2. Java Swing apps, and a variety of other apps, implement the Java 
accessibility API
3. The gnome-java-bridge, which we load into the Java Runtime, is the 
glue code that bridges the Java accessibility API to the UNIX AT-SPI


Regards,

Peter Korn
Accessibility Architect,
Sun Microsystems, Inc.