[Kde-accessibility] KOffice 1.4.1 Accessibility Assessment

[Gary Cramblitt]

I have completed an accessibility self-assessment for KOffice version 1.4.1.  
You may view the assessment by downloading one of the following files:

http://accessibility.kde.org/reports/koffice-1.4.1-assessment.sxw  (oo.org v1)
http://accessibility.kde.org/reports/koffice-1.4.1-assessment.odt   (ODF)
http://accessibility.kde.org/reports/koffice-1.4.1-assessment.pdf  (PDF)

The assessment has been coordinated with the KOffice development team and 
includes notes about fixes in the upcoming 1.5 release (early 2006).

Comments, corrections, etc. welcome.  In a couple of days, barring major 
changes or objections, I intend to publish it on the KDE Accessibility 
website at

http://accessibility.kde.org/reports/

The reasons for the assessment are:

  1.  The Massachusetts Commission for the Blind requested accessibility 
information concerning software that supports the Open Document Format (ODF), 
i.e., KOffice.  (See 
http://lists.kde.org/?l=kde-accessibility&m=112653249424930&w=2)

  2.  It needed to be done.

I expected blind access to be poor, since KDE does not have a screen reader, 
but the results for low-sighted and motor impaired users were surprisingly 
disappointing.  That's the bad news.  The good news is that the KOffice devs 
have been very supportive and many (but not all) of the problems will be 
fixed in 1.5.

I've reached the following conclusions:

  1.  Some of the problems are fundamental to Qt3 and kdelibs3.  QSplitter, 
QDockWindow, and QToolbar, for instance, are 3 problem areas.

  2.  We need to get to Qt4/kdelibs4 as soon as possible.  But there will 
still be many issues to address.  Accessibility won't happen automagically.

  3.  We need to do this sort of thing more often and hold developers feet to 
the fire.  Many of these problems are easy to fix if caught early.  Others 
are not so easy and will need changes to core functionality to fix properly.  
Some are basic design flaws, such as providing functions that can only be 
performed using a mouse.

  4.  The timing of this assessment is really bad because KDE 3.5 is in freeze 
and KDE4 is a long way off.  Had this assessment been done 6 or 9 months ago, 
we probably could have solved most of the problems (except for blind access) 
by making needed changes to kdelibs3.

If anyone wants to do similar assessments for other KDE applications, please 
do so and send them to me.  They are not hard to do, and you don't have to be 
a programmer to do them.

Regards
-- 
Gary Cramblitt (aka PhantomsDad)
KDE Text-to-Speech Maintainer
http://accessibility.kde.org/developer/kttsd/index.php