From just looking at the website SAW looks very interesting. I'll 
reboot to Windows and test it later.

I don't think we would use any of the SAW code directly in SOK since 
it's written in a completely different language. We will certainly look 
at it and take ideas from there though.

We are already going down the same path in some respects. We will be 
adding macro features that can perform repeated tasks in different 
programs and the layout and key selection can also be altered by the user.

I don't think we will do very much with it to improve the accessibility 
of general applications, because those applications should themselves 
have better access. Instead of creating workarounds in SOK, and Orca, 
and Dasher, etc, etc, we would rather contact the maintainers of the 
relevant applications and ask them to improve it. If it's a small thing 
then the response is usually very good and you then also raise awareness 
for the longer term. IMO it's a cleaner way to fix things. If that fails 
though, we can make adaptive scripts like Orca does.

About GOK and SOK: GOK is a mature project, having been around for 5 
years, but it's gotten a bit stale IMO. It does have support for word 
prediction and can be used to control other applications as long as they 
use AT-SPI (as all GTK+ apps should and all KDE4 apps will).

SOK is a much younger project, of only about 3 months, having been 
started in large part as a response to the lack of change in GOK. We've 
focused on basic text entry for now, but its design is flexible. Adding 
macros, images and sounds should be quite straight forward.

So I agree that GOK has many problems (esp. in usability compared with 
alternatives on Windows and Mac. It was in part to improve on that we 
started SOK, though that has a long way to go as well.

Feel free to add feature requests under 'Wishlist items' here: 
https://wiki.ubuntu.com/Accessibility/Projects/SOK

On word-prediction: I don't think that word prediction as it is done now 
is very efficient. I have ideas on improving that but that would require 
a few more pages of text :)

- Henrik

[log in to unmask] wrote:
> Hmm, yes David that's a good question!
>
> I suppose the parts of the code that are not directly dependent on the 
> Windows system APIs might be of interest. But that greatly depends on 
> how well the logics of the program are separated from the native API 
> dependent parts?
> That was questions I asked Jason, Pat and you a while ago : How far 
> did you/we actually get in investigating these issues as part of the 
> SAW5 project? Perhaps that was one of many things that had to be put 
> on the waiting list? If so, it would be an important project in itself 
> to investigate the state of art and how to go from there, ideally 
> together with a couple of programmers experienced in multi platform OS 
> C(++) code development.
>
> Whithout having had much chance to do any more thorough studies or 
> practical testing of the GOK and SOK packages, I miss all more 
> advanced options for language and/or multimodal support in these 
> on-screen keyboards' specs. Such as support for word lists or word 
> prediction, support for graphical symbols and sound in the cells as a 
> complement or alternative for plain letters etc. GOK and SOK are both 
> very basic alphabetic letter keyboards as I have understood. Correct 
> me if I have got this wrong!
> So it's a far way to go to match the specs of SAW5 (and several other 
> on-screen keyboard and communication tools on the Windows and partly 
> also Mac platforms).
>