Re: GNOME A11y: where do we need to improve? (Want input by 25-Jan)

[Luke Yelavich <themuso themuso com>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I originally sent this directly to Willie, but I think it wouldn't hurt to also send it to this list.

On Wed, Jan 23, 2008 at 03:23:36PM GMT, Willie Walker wrote:
> Hi All:
> 
> I've seen some good activity on the WIKI (thanks!).  I also need your 
> help with identifying the top 5 to 10 things that need doing on this 
> list.  Please be brave and speak up, especially those of you who have 
> expressed concerns about where the community might be focusing or 
> spending its energy.  This really is your chance to help influence where 
> things go.

Hi Willie
Here are my top 5 things I would like to see happen:
1. gnome-speech/speech infrastructure.
2. gnome-panel (No point having it, if we can't access all of it)
3. TTS localization (Jonathan, the author of ESpeak would be the best person to help here)
4. Rhythmbox. The computer is a jukebox these days, so lets make sure it works well with a11y.
5. Evince. We need to be able to read PDFs etc. We shouldn't rely on Acrobat Reader where at all possible.

My involvement in general for GNOME/Linux accessibility will be to keep Ubuntu at the forefront of Linux Accessibility, and make sure the best tools are as well integrated for users as possible, to make things as seemless as possible.

In light of the above list of attention items however, I feel I can help the most with what I think is THE most important issue we need to resolve, and need to resolve ASAP. Yep, you guessed it, speech. Firstly, there is the nice fact that KDE will be coming online with accessibility some time in the coming years, and they said early on, at least for use with the existing kde accessibility tools, that they favoured speech-dispatcher's model. Secondly, speech-dispatcher's model is flexible enough, that with some small speech-dispatcher changes, third parties could easily build drivers for synthesizers outside the speech-dispatcher source tree. This would especially be useful for building support for proprietary speech synthesizers, and would then make it easier for users to set up such synthesizers, without having to rebuild speech-dispatcher, with the possible effect of screwing up their system.

There is also the changing face of audio infrastructure to be considered. Luckily, just about all speech synthesizers support returning audio to the calling application via a callback. Speech-dispatcher makes good use of this with espeak, festival, flite, and ttsynth, and then can send the audio to one of many audio outputs, such as OSS, NAS, ALSA, and recently, PulseAudio. At this point, I am very much grappling with having to deal with what direction to go for Ubuntu hardy, in relation to speech audio output, and PulseAudio. At this stage, I am looking at wrapping the espeak-synthesis-driver gnome-speech binary in a call to padsp, so that all OSS calls from espeak+portaudio v19 are sent to PulseAudio. For next release, I will likely implement a speech-dispatcher solution, as that is what users are asking for on the Ubuntu accessibility lists.

if upstream could also support speech-dispatcher, that would be much appreciated. I could then easily add scripts to the speech-dispatcher package to allow for distributions to make use of speech-dispatcher per user session, as that is how PulseAudio is being run, at least in Ubuntu. All this would be possible, due to me now having speech-dispatcher CVS access.

In the longer term future, I would then write a nice GUI, to allow users to more easily adjust speech-dispatcher's options, or even better, help in implementing a hook into Orca, so that Orca users could directly alter speech-dispatcher settings from clicking a button in the speech preferences pain.

I would urge to consider my proposal, and my offer of help. Together with the speech-dispatcher developers, we can sort out any remaining issues you think exist, and move forward to using a completely GNOME abnostic framework for speech output.

Thanks for your time.
- --
Luke Yelavich
GPG key: 0xD06320CE 
	 (http://www.themuso.com/themuso-gpg-key.txt)
Email & MSN: themuso themuso com
Jabber: themuso jabber org au
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFHma1jjVefwtBjIM4RAjIqAKCkZM4byCXE45/gxTfRrvOY+twEtACfZc86
LA35kWsei4gJ9/OYBkcBrUk=
=m6DO
-----END PGP SIGNATURE-----