[Kde-accessibility] [MA ODF Accessibility Team] Report from a11y meeting 29 Nov

[Gary Cramblitt]

Mark Bucciarelli and I attended a meeting 29 Nov in Boston, Massachusetts.  
The meeting was sponsored by the Disabilities Law Center (DLC), Massachusetts 
Office on Disability (MOD), and the Disability Policy Consortium.

The following description is mostly mine with input from Mark.  It is taken 
from memory and the few notes we made.  Nothing in this report should be 
taken as fact, completely accurate, or official policy.  It is our 
interpretation of what was said.  Where Mark and I differ in interpretation 
or viewpoint, I've so indicated.

It was not a large meeting.  About 30 people I'd say.  Some of the people 
present had no idea what Gnome or KDE are.

The title of the meeting was "What is Open Source?", but from what Mark and I  
could  tell, the real purpose of the meeting was:

  1.  To smooth the ruffled feathers of the disabled community in 
Massachusetts and give them a chance to be the focus.

  2.  To inform the various groups in MA that represent users with 
disabilities as to what the MA Open Document Format (ODF) decision means.

  3.  To describe the current state of accessibility (a11y) in applications, 
desktops, and operating systems that support ODF and PDF.

  4.  To describe the process that will be used to implement the decision with 
respect to a11y.

  5.  Identify the issues that need to be solved in order to fully support 
users with disabilities in MA using ODF and PDF.

The formal speakers were:

Peter Korn of Sun.

Aaron Leventhal of IBM.

Peter Quinn, CIO of MA Information Technology Division (ITD).

Greg Pisocky of Adobe Systems.

John Chappel of the Massachusetts  Rehabilitation Commission (MRC).

Leon Shiman of Shiman Associates, member of the X.org Board of Directors.  
Leon is our main contact and the reason we were able to attend.

I (Gary) also spoke *very* briefly, representing KDE e.V.

We haven't seen a full attendees list, but some names we did get were:

Christine ("Chris") Griffin, Disabilities Law Center (convened the meeting)

Bill Allan, Head of Disability Policy Consortium, not a state organization, 
rather a lobby/advocate for the disabled.  25% of his board requires 
accessibility technology to use computers.  Bill has spoken out against the 
ODF decision.

Jerry Berrier, President of the Bay State Council for the Blind.  He has 
spoken out against the ODF decision before.

Alan Yates, Product Manager at Microsoft.

Myra Berloff, Massachusetts Office of Disabilities (MOD), ADA Coordinator for 
the State (Americans with Disabilities Act--big Federal Law).  She is 
working on Memorandum of Understanding between the various MA agencies as to 
how a11y will be addressed.

Barbara Lyberger, an attorney for the Massachusetts Office on Disability.

Paul Snayd, IBM.

Cynthai Ice, IBM.

I won't report everything that was covered in detail (I couldn't do that even 
if I wanted to).  Peter Korn recorded audio of the meeting and once he has 
received an OK from everyone who spoke, intends to put it on the web.  At the 
very least, he intends to publish his part of the meeting, and by far, he had 
the longest part.  Let me try to summarize what was covered briefly.

1.  Peter Korn gave two talks.  The first talk was about standards in general, 
open standards, how open standards relate to open source (most people confuse 
the two), and how all these relate to accessibility, accessibility standards, 
and ODF.

2.  Peter Korn gave a second talk on Open Source Accessibility & Open Office 
Accessibility.  Peter described how the Microsoft Active 
Accessibility API (MSAA) has stifled progress in a11y on the Windows 
platform.  For instance, the MSAA is not a robust enough API for assistive 
screen readers to get at the contents of documents.  It also doesn't permit 
scripting of applications via the API.  Most of the "success" in Microsoft 
Office has been through the "heroic" efforts of AT vendors such as Freedom 
Scientific (makers of JAWS) and Ai Squared (makers of ZoomText) who have had 
to reverse engineer the MS Office applications in order to achieve the 
admittedly excellent support they have.  He pointed out this is limited to MS 
Office applications, not the entire Windows desktop.  He pointed out that 
these implementations often require specialized hardware drivers that don't 
support all hardware.  He pointed out that the AT vendors are going to have 
to do this all over again with MS Office 12, that upgrades will be expensive 
for the MA government and for users, and that significant retraining will be 
needed.  He described this as the "bolt-on" approach.  Then he described the 
a11y infrastructure approach, where a11y is built into the system from the 
ground up.  (Though he didn't say it, he's talking about AT-SPI.)  He 
described how such a robust infrastructure benefits all applications running 
on the system, and provides opportunities to enhance a11y that simply are not 
possible with the bolt-on approach.  Peter then gave a demonstration under 
Gnome of Gnopernicus screen reader, Gnome On-screen Keyboard (GOK), and 
Dasher.  Peter mentioned that a scripting screen reader named Orca will be out 
in beta sometime in April or May.

3.  Aaron Leventhal described the current state of a11y in Mozilla and 
Firefox.  Firefox 1.5, which was released the same day, has significant a11y 
enhancements.  There are many technical issues to deal with however.  In 
particular, how should browsers deal with pages when part of the page has 
dynamically changed (timed javascript and flash for example)? Firefox has 
incorporated accessibility in their project by having a team that reviews all 
commits to ensure they comply with accessibility issues.  This is now a 
project policy.  Not a lot of rules but they stick to them.  (Could KDE 
develop similar rules and script them?)

4.  Peter Quinn directly addressed the disabled community.  Some points:

- categorically will move the date if the chosen ODF product(s) cannot give at 
least the current level of a11y functionality.

- is very reticent to change the date because there is a lot of industry 
attention and focus now (IBM + Sun together) and it is wise to keep the  
pressure up.

- spoke about the Microsoft application to ECMA to have their XML schema 
ratified as an open standard: 9 - 18 months, then they have to go through 
international standards body.   Basically a no-go if he can stick to  the 
1/2007 date.

5.  Greg Pisocky described the PDF/A standard and the enhancements for a11y in 
both Adobe Writer and Reader.  He gave a demonstration of OpenOffice.org 
running under JAWS (which currently works but with lots of fixable problems), 
had oo.org write out a PDF/A file and speak it in Reader.  Since Adobe Reader 
has its own TTS capability, the point of this demonstration was that ODF 
supports a11y and since oo.org exports to PDF, it can create highly 
accessible documents (at least as far as PDF goes).  The PDF "TouchUp 
ReadOrder Tool" looked really nice.  Issues it dealt with:

    - Logical Read Order (for example, in a doc with two columns, don't 
      read across from one column to the other, like Jaws does now in 
      OOo).

    - Alt text for graphics

    - Logical tab order (for forms)

    - Short field descriptions (for forms)

    - Tags on figures and tables

Mark felt Greg's demonstration was very impressive.  His was the only demo 
that showed JAWS and Mark was impressed with its capabilities.  It was amazing 
to watch and see how a blind person must interact with a computer.  As 
someone who had not seen that before, Mark appreciated that he could use it 
and showed how it worked.  I thought it was less so because it showed the 
current problems between oo.org and JAWS (airing dirty laundry).

6.  I was only allowed to speak for about 2 minutes, in which I stated that 
Mark and I were volunteer developers representing the e.V., that KDE is 
committed to accessibility and desires to cooperate with accessibility groups 
and users in MA and elsewhere to improve a11y in KDE and KOffice, that KDE is 
an alternative graphical desktop, that KDE includes a suite of office 
applications called KOffice, that the suite includes word processor, 
spreadsheet, presentation, project planning, database and graphical 
applications, that KOffice reads and writes ODF, that KDE and Gnome are 
cooperating and not competing, that KDE 4 will share the very same a11y 
infrastructure as Gnome, and that this means the AT tools in Gnome will work 
with KDE apps and vice versa.

7.  John Chappel spoke briefly about the CRM's plans.  The CRM, together with 
the MOD, are preparing a set of a11y software standards.  The standards will 
be published on the web with a public comment period.  The CRM and MOD are 
also looking at what a11y training requirements will be needed as a result of 
the ODF decision.


Some important points that were raised in questions and discussion:

1.  ODF software not only needs good user documentation and training, but that 
documentation and training itself must be accessible.  Something to keep in 
mind is that most of the trainers are themselves disabled.

2.  MA doesn't have a good sense of exactly what a11y functionality is 
currently being used within the MA government and therefore would be a 
priority when migrating to new software.  There are certain features, for 
instance, that are available in JAWS and ZoomText, but are these features 
actually being used?  A survey to collect that information is underway.

3.  Should Adobe Writer, oo.org, and KOffice "nag" document authors to put 
a11y information in the document?  From an a11y standpoint, yes, they must.  
From a usability standpoint, such a feature would just be turned off and 
ignored  by most users, even if were the default.  At the very least, the 
capability should be in the software, and it should be possible for admins to 
force it on.

4.  Disabled government employees are concerned that a) they are going to need 
extensive training in using the new software (one blind person stated that he 
required 6 months of training with MS Office and JAWS), b) if a variety of 
software is used within the MA government, their ability to move jobs from 
one agency to another would be impaired, c) since most disabled people in MA 
use Windows, their ability to get a MA government job will similarly be 
impaired, d) the current state of a11y in ODF software is horrible and just 
won't do, and e) promises are being made to fix this, but they don't have 
confidence those promises will be fulfilled.

5.  Because of #4, not only should MA have a standard document format, but it 
should have a standard application user interface as well, so that users with 
disabilities who are forced to change applications do not have to learn a 
completely new interface.

6.  A blind person's desktop will be the same as a sighted persons.  No 
reduced capability.


The following are my (Gary's) personal observations and thoughts.

Peter Korn did a fantastic job at explaining how the bolt-on approach impedes 
a11y progress (and is expensive), and that the AT-SPI approach offers many 
opportunities to not only meet the current level of a11y under Windows, but 
to exceed it.  He barely mentioned KDE or KOffice, but I still thought the 
talk was an excellent one.  If you can listen to his presentation (I'll post a 
link when he puts it up), I urge you to listen.  Despite his efforts, the 
message just didn't seem to get through to the people with disabilities in 
the room.  They kept saying (not in these words exactly, but in essence), "we 
don't want to have to change".  They are happy with what they have and view 
the ITD decision as a) forcing them to accept a level of a11y that is much 
less than what they currently have, and b) raising huge roadblocks to job 
mobility and availability for disabled persons.  Despite Peter and others 
mentioning several times that Office 12 will force change on them whether 
they like it or not, they keep saying "don't make us change."  We need to 
keep hammering home the message that they are going to have to accept change 
and that non-choice is coming from Microsoft.  We also need to address the 
training issue in a big way.

IMHO, the demonstrations were very impressive and should have 
"sold" the users with disabilities on the possibilities for the future that 
AT-SPI creates.  Alas, that didn't seem to happen.  Mark thought Peter's talks 
were too technical, too long, pushed UNIX too much, and not appropriate for 
the audience.  Given the questions that followed, Mark may be correct.

I was also disturbed by #4, #5 and #6 above.  In essence, this is the pro- 
"software monoculture" argument.  If everybody in the world uses the same 
software and it never changes, then users with disabilities are much better 
off.  It was particularly distressing to me to hear these comments coming 
from the MOD and MRC people, who will be writing the software a11y standards.  
When pressed, most backed down somewhat to restrict the requirement to "all 
ODF software should have common functionality and a common keyboard 
interface, i.e., the same shortcuts across all ODF applications."  This 
"requirement" could have a profound impact on whether KDE and KOffice are 
ever used in MA.  Think about the frame-based interface in KWord for instance 
and how this differs from the style-based interface in oowriter.  If the 
software monoculture requirement wins the day, in essense it means KDE must 
*be* Windows.  argh!

OTOH, we must not dismiss the very real and valid concerns about job mobility 
for disabled users.  I think we can address this in two ways.

1.  MA and the open source community (us) need to develop a robust strategy 
for a11y training.  We should develop an "accessibility kit" that agencies, 
governments, schools, universities, etc can easily obtain to help them 
implement training.  Obviously, a11y training cannot be fully automated, but 
we can provide most of the electronic materials that would be needed.

2.  We need to keep hammering home the message that a) software keeps 
changing, particularly when only one vendor is controlling it, and b) the 
a11y infrastructure approach offers opportunities to extend functionality 
beyond what is currently implemented, and therefore, offers positive change, 
not negative change.

In addition, Peter Korn pointed out that if MA develops a11y training in a 
variety of applications and platforms, it will open up job opportunities to 
people with disabilities that they might not otherwise have had.  Jobs in 
UNIX system administration and website/services creation, for instance.

IMHO, this "software monoculture" argument could be the greatest obstacle to 
adoption of ODF and open source software in MA, second only to the dirty 
politics.


In a separate private conversation, Leon asked us for the following 
information.  He needs specific information, not generalizations.  I'd like 
whoever can supply the information, in as detailed and comprehensive manner 
as you can manage, to send it to Leon (leon at magic dot shiman dot com).  He 
needs this info by 25 Dec 2005.

1.  What capabilities are needed in X.org in order to implement a11y in KDE?  
(I know almost nothing about X.org, but I think this means composite and 
transparency?)  What specific problems exist and are they fixable by Jan 
2007?  Leon is on the board of X.org, so he can help to make things happen.

2.  What platforms does KOffice run under?  Sun/Solaris?  BSD?  All Linux 
distros?  If not, which ones?

3.  What are the issues and problems with D-BUS that need to be solved before 
KDE and AT-SPI can migrate to it?  He has heard that D-BUS is "not ready for 
primetime" and would like to know details.

4.  What are the current issues and limitations of the Gnome ATK and AT-SPI 
with respect to KDE that would have to be solved?

In addition, Leon wants to perform usability testing of KOffice.  He wants us 
to supply him with the "best" version of KOffice we can manage to put 
together and also provide someone who can "hand hold" while they do the 
testing.  That person would not have to be physically present, but they must 
be reachable by phone or email and knowledgeable.  He wants to do this 
testing very soon (around end of year).  This is also an opportunity for 
KOffice to get some valuable and detailed testing results.

He also wants to perform accessibility testing of KOffice.  I'm trying to talk 
him out of this one, since I've already written a detailed accessibility 
report myself.

-- 
Gary Cramblitt (aka PhantomsDad)
KDE MA ODF Accessibility Team Chairman