Re: accessebility suggestion for Ubuntu 6.06 LiveCD

[Henrik Nilsen Omma <henrik ubuntu com>]

Petra Ritter wrote:
> Hello,
>
> My suggestion is to turn on the accessibility Technology by default
> and to have accesskeys to start Gnopernicus or what ever program they
> use for persons with motor difficulties. I assuming that the for 'Lesser
> visual impairment' there use just another theme to change the
> appearance of the Screen. To find a way to do this with a accesskey
> should be not a problem. The option "Mmoderate visual impairment'is a
> problem anyway because the magnifire option in Gnopernicus don't work
> very well.
>
> On this way eg a blind person can start Ubuntu and when the system is
> up and runing start Gnopernicus using a accesskey.
>   
Hi Petra,

That seems like a valid approach, but there are reasons why we didn't do 
it that way. Ubuntu is a mainstream distro used by hundreds of thousands 
or (hopefully) millions. The physical CDs from ShipIt certainly get 
distributed in the millions. Making such a distro for a wide range of 
user groups involves a long list of compromises. The trick is to look 
for ways to meet the needs of as many groups as possible without those 
getting in each others way. That's not just related to AT support. You 
only have to look at any of our development related mailing lists to see 
that compromises are happening constantly.

In that light, there are two problems with what you are describing. 
Problems that would affect the bulk of users. 1) Extra computer resource 
use: Loading the assistive technology modules uses extra resources in 
the form of CPU cycles, memory, desktop and application startup, etc. 
You could argue that it's minimal on a modern system, but most people 
prefer to have a lean setup where they can add what they need. For 
people with older systems it can make a significant difference. 2) Bugs: 
The AT stack is much less tested than anything else in our main software 
repository simply because very few people have traditionally been using 
it. Obviously this needs to be improved, but ATM it is a reality. If you 
turn on AT support by default you will cause serious problems and 
crashes for a large number of people, which the wider community would 
not and should not accept. I have serious issues with OpenOffice 
crashing when AT-SPI running for example. That's probably an OpenOffice 
bug (and I've filed it), but it will sting people just the same. Also 
remember that accessibility is not only for the visually impaired, but 
also motoric and others. Our current solutions aims to be generally 
useful to several groups.

So it sounds like the right solution would be to make a separate 
derivative with these tools enabled by default, an Ubuntu-AT derivative. 
We've had several debates on that early in the project and I've actually 
switched sides in that debate several times :) (first being against a 
derivative, then thinking it was the way forward, and now being strongly 
against).

The advantages of being 'in the main distro' are significant on several 
levels.

* First of all, the CDs get distributed in the millions, where as a 
derivative would likely be a download-only item available from just a 
few mirrors. Now if a blind user gets a standard Ubuntu CD at a Linux 
fair or from a friend it actually has all the AT tools on it. Being on 
the default CD will reach many more potential users. This will also 
result in more testing.

* The access tools are also installed on every HD-installed Ubuntu 
system. It just needs to be activated to work. That means that if you 
walk up to some random computer in a Library with Ubuntu you can use 
that machine directly without having to locate some system administrator 
with root access who can install it first. (but yes, the methods of 
activation are a problem)

* A very important, but less obvious factor is developer-mindshare. 
Ubuntu attracts some brilliant developers, both as volunteers and 
Canonical employees,  who can do some real magic if given the scent of 
the issue. By being in the main distro we involve these people in a 
natural way and can work with them to create solutions that are better 
for everyone.
Our next release has 91 feature goals, or 'specifications'. 8 of those 
are new accessibility targets. See: 
https://launchpad.net/distros/ubuntu/edgy/+specs


So that was a long explanation of why we've done it they way we have :)  
It's also the case that we didn't quite get it working the way we had 
hoped for in 6.06, but I still feel that it was important to get this 
ball rolling and establish these features. The good news is that we are 
working to improve it for 6.10 and that we do listen too feedback :) We 
also appreciate very much the efforts of early adopters like yourself 
who are making the effort to test these early and imperfect attempts at 
getting it right. Feedback from the community is very valuable!

Now, finally getting to your point: I agree that the method of 
activation is not ideal. We need to discuss what is the best way to 
improve on that. Let me start by explaining how we are planning to 
improve things for 6.10 and then air some alternative or possibly 
additional improvements we should consider.

* Improved Live CD boot menu. Many people have commented that the 
boot-press-F5 system is difficult to use for the visually impaired. 
That's true, so we need to improve it further. We plan to play a sound 
when the system is ready to take boot options so that the user will know 
that it's time to press F5. That will take you to a new boot menu with 8 
large high-contrast icons that you can select with the arrow keys. Each 
time you change the selection a unique sound will be played so that you 
know where you are in the menu. (link: 
https://wiki.ubuntu.com/Accessibility/Specs/LiveCdAccess)

* Using Orca. We will move from gnopernicus to Orca on the Live CD. 
Hopefully this will be a more reliable solution, removing the startup 
failures we had. The Orca team have also created a keep-alive daemon 
that brings Orca and related speech architecture back up should it fail 
for any reason.

* The administration system and password problems are being addressed. 
The AT apps were not playing nicely with the system admin tools, and 
that is being addressed here: 
https://wiki.ubuntu.com/Accessibility/Specs/SudoAdminAtspi

* Improving the installer. The above point will allow the installer to 
work with AT tools and we also need to fix a few rough AT edges on it as 
well. The next version will also remember the AT settings you used so 
that when you boot your newly installed system it will start those tools 
by default.

* Better web-based documentation. We have recently set up 
http://access.ubuntu.com which now contains a basic introduction, but 
will also have guidance on how to use these tools. You be able to hear 
the sounds on-line that will be used in the boot menu, with 
explanations. We have also discussed making audio tutorials, but it 
would be great to get some help with that :)


There are probably other things we should do as well.

* The GDM login manager has some access support, which we haven't 
utilised. This could provide an alternative way of activating the AT 
features. This would be esp. useful on public systems.

* An AT system activation daemon with the functionality you are talking 
about. One problem is that you have to log out an in again to activate 
AT-SPI and we cannot leave it running for all users. This seems like a 
fundamental AT-SPI bug to me that should be addressed.

* The Gnome accessibility settings are currently spread in many 
different places and should be collected together and cleaned up in my 
opinion.


Hopefully we'll see some of these improvements appearing on the Edgy 
Live CDs in the next few weeks and we would appreciate if the community 
here could give them a spin and report back. Thanks for your patience so 
far!

Henrik
Ubuntu Accessibility Coordinator