Sunday, 25 January 2009

The Tortoise and the Hare!

There have not been many updates on the blog or the development mailing list lately because I have been trying to find solutions (or at least workarounds) for some fundamental problems such as getting Speech-Dispatcher and Alsa to work with multi-channel sounds, getting LinuxSpeaks to integrate into the system without causing any problems and trying to get all of the accessibility features to fit on a 700MB CD without removing too many applications. This has been frustrating enough, but on top of this I had a serious neck injury last week which means I have been unable to spend very much time at a computer. On the other hand I have managed to solve the username/password problem, installed and configured YASR and managed to fit a few more useful console based applications onto the CD. I think I will realistically have to leave Speech-Dispatcher, Speakup and LinuxSpeaks until a later version and just get out a new release with some small but significant improvements rather than be over ambitious and release something with more advanced features that may or may not work. I am obviously more of a tortoise than a hare!

Monday, 19 January 2009

New Features in Vinux 1.3!

Here is a list of new features I hope to be able to incorporate in the first release under the new Vinux name: The Username and Password entered during the install process will be retained, the USB Creator application will work with the customised live CD by default, the YASR console screen-reader will be installed allowing users to run without a GUI, the LinuxSpeaks complete audio desktop will be available and a whole host of accessible console based applications including lynx, urlview, joe, mc, oleo, mutt, vlock, partimage, alpine, irssi, mpg123 etc...

I am also trying to configure Speech-Dispatcher with Alsa, Emacspeak, the Linux Screen Reader, Speakup, Screader, Festival and as many other accessibility applications as I can find, but unfortunately I have not been able to get these working satisfactorily yet with Intrepid!

The cost of including these extra accessibility packages is that I have to exclude some standard packages in order to fit them on the CD. I have already stripped Ubuntu down about as far as I can without causing any significant problems, but Open Office is now the only thing left I can safely remove! So the choice is this - do I remove everything I can and provide a bare bones system with all of the accessibility packages ( I suppose I could include a script to reinstall everything I have removed once installed) or do I leave Ubuntu as it is and just add on all of the accessibility applications creating a large iso (possibly up to 1.4Gb) that would have to be burned onto a DVD. This is a problem I am still wrestling with, I may well end up making two different versions one for CD and one for DVD!

Thursday, 8 January 2009

Performance v Functionality!

I have spent the last week unsuccessfully trying to replace PulseAudio and Gnome Speech with ALSA and Speech-Dispatcher. I have been able to get Speech Dispatcher running with ALSA and this does significantly improve the responsiveness of Orca. However, once I enable Speech Dispatcher I have been unable to run any other applications which produce audio output. This is in all probability because I have not configured Speech Dispatcher properly, but I have followed all of the guides I can find online as well as receiving advice from members of the Speech Dispatcher mailing list to no avail. Speech Dispatcher offers many advanced features: It is able to use OSS, ALSA and PulseAudio as well as being able to run as a system service or on an individual user basis. However the fact that it offers so much functionality means it is not easy to configure. I hope that I will be able to resolve this problem but if not I face a tricky decision. Do I install ALSA and Speech Dispatcher and sacrifice multi-channel sound functionality or do I stick with PulseAudio and Gnome Speech and settle for poorer performance?

To be honest I don't think the responsiveness of Orca using PulseAudio is that bad on reasonably modern hardware and adding yourself to the PulseAudio realtime group can improve this further. Hopefully in the long run the Ubuntu, Gnome, PulseAudio and/or Orca will resolve these issues and replacing PulseAudio will be unnecessary. In the meantime if I am unable to resolve the issue than I could install and configure Speech Dispatcher but not enable it by default. This way beginners would have fully functional speech and multi-channel sound, while more advanced users could enable Speech Dispatcher and configure it themselves. I am also intending to install and configure YASR, a console based screen-reader and LinuxSpeaks, a complete console based audio desktop in the next version of Vinux.