[Accessibility] Accessibility Workgroup aim / Multiple toolkit strategy

[Olaf Schmidt]

Hi!

At the last phone conference we discussed releasing a statement that the 
approach of this working group is not to pick a single desktop or toolkit, 
but rather to promote cooperation between the various desktops and toolkits.

I believe this is necessary because some people are spreading the claim that 
GNOME is about to be adopted as the standard desktop by both the FSG and the 
OSDL. This is often coupled with the misinformation that all GTK and GNOME 
applications are perfectly accessible and that Qt and KDE are totally 
inaccessible.

I have often stressed the good work of the GNOME accessibility team and the 
good spirit of cooperation with us, and I am grateful that Peter and Bill 
have often returned the compliments in talks or when posting to Slashdot 
discussions. I have therefore long hesitated to bring this topic up to avoid 
unneccessary bad feelings.

Unfortunately some misleading sentences on the OSDL and LSB sites have led to 
a lot of confusion during the last weeks (more on this at the end of this 
long email). I believe a clear statement issued by us might help to get those 
fixed. I am also sure that our decision to standardise the GNOME-authored 
AT-SPI will be misinterpreted as adoption of GNOME by some people, unless we 
make clear statement avoiding the wrong impression and safeguarding against 
deliberate misinterpretations.

Picking a single toolkit and desktop would of course run totally contrary to 
our aim to standardise the assistive technology API that is currently shared 
by Java, GTK, and Qt4. Moreover, it would be harmful to accessibility, 
because the GNOME and KDE accessibility teams have a kind of unofficial 
agreement to work on different kinds of assistive technologies to avoid 
duplication and reach as many users as possible. Using only GTK would make 
the Unix desktop less accessible for a number of users.

The direction of this misinformation gives a bad name to accessibility, to 
standardization efforts, and to free software in general because it gives the 
wrong impression that we are unable to work together. (Just as the claim that 
Novell would force SuSE to drop KDE after they where bought led to a lot of 
bad feelings. And my impression is that this are the same people who are now 
spreading other misinformation after the first attempt didn't work.)

Some people believe that both the OSDL desktop working group and the LSB 
desktop working group are dominated by people who have an anti-Qt-agenda, and 
who would use any available argument to go for a GTK-only route. I don't know 
any of the members, so I give them the benefit of the doubt. I hope that a 
statement by our working group would convince them to revise those web pages 
that are being misinterpreted as opposition to Qt/KDE.



Some example for such pages:

http://osdl.org/lab_activities/desktop_linux/DTL-PS/document_view

"Based upon these criteria, the current implementation of Gnome appears to be 
the best fit, and will be used as the current reference desktop system for 
the purposes of working toward the DTL working group’s strategic goal of 
accelerating Linux on the desktop. This selection does not imply any 
endorsement or, by implication, any rejection, of Gnome."

Even if the page is clearly saying that they do endorse GNOME, this page is 
used by a number of people as a proof of the claim that KDE is officially 
disfavoured by the OSDL because of being "inaccessble". Accessibility is one 
of the three mentioned criteria. Of course cooperation and integration 
between the toolkits is needed for accessibility, not dropping all assistive 
technologies that use the "wrong" toolkit.


http://www.linuxbase.org/LSBWiki/DesktopWG

"Following is the list of libraries currently under investigation for their 
inclusion in the first release of Desktop module: [...] GTK libraries"
"Libraries not under consideration due to lack of resources: [...] libQt 
libraries"

The page does not say that GTK will indeed be standardized, because it still 
needs to pass the "best practice" test. If however, GTK is really included in 
the next version of the LSB, while excluding Qt, then it is likely that both 
the FSG will quickly loose support among free software developers - at least 
outside of the USA. I have already heard comments such as "ignore the FSG" 
from some people. An adoption of GTK-only by the LSB would also make it 
difficult to justify my membership in this working group.

But I am still optimistic that we can avoid this harmful road by making the 
purpose of this working group clear.

Olaf Schmidt