Hi there,
hmm, why?
How does Speech Dispatcher limit used modules?
Like yes, SD is licenced under GPL, but that doesn't apply for independend parts like speech modules, which can have their own licences.
I personally think Speech Dispatcher is a very well made system,
it has even a great documentation (unlike Orca for example).
The only speech related problem I see on Linux is, that Orca and may be other screenreaders? don't provide, or at least I'm not avare of a way to speak through them.
That means I have to implement speech configuration in every
single application which wants to speak, what is a drag,
especially if doing the settings part wouldn't even be necessary
without the speech stuff.
This however is not a problem of Speech Dispatcher, but an Orca
issue.
Fixing it requires fixing Orca, which is problematic for a new
contributor due to the above mentioned lacking documentation.
So, I'd say documenting Orca is something we currently need much more than an obsolete speech system.
In fact, I'd personally go even further, and say, that we need
documenting Orca and Linux accessibility stack in general the most
out of everything else.
Just look at NVDA and its addon repositories. How many great plugins can you count there?
Well, tens, if not hundreds of them, I didn't really count them
myself. :) NVDA has a great addons-developing documentation, which
allows anyone with at least basic skills with Python and object
oriented programming to easily get in and develop a plugin with
minimal time spent on finding out how to do it.
And how many addons do we have for Orca?
Well, I know about one, aiming to make Orca able to work with NvdaRemote, and even thatone is incomplete and seems rather hacky to me.
What's totally understandable, since there are no resources to
teach the proper way to do it.
If we want Linux accessibility to grow and not get fractioned, we need to get in more people. And to get in more people, we need to give them an easy way to enter, without need for a lengthy Orca reverse engineering process.
Well, I've ended up probably a bit off of the original topic, so
returning to the original idea. There is quite a lot relevant work
to do here. I recommend investing time in them, it will help a
hlot.
Best regards
Rastislav
Source code for the TTS modules is already available on Github. The full system will be running by January 1, and will be an alternative to Speech Dispatcher.
I would like to thank the Voxin and Orca authors for all their efforts over the years. I rely heavily on both of these tools to remain productive at work.
Voxin broke again for me on Monday after a Linux update was pushed to my corporate laptop. This is not the fault of the Voxin authors! They do this work out of charity, knowing that blind and low vision folks around the world rely on their efforts. The problem is the architecture of Speech Dispatcher, which is hostile to commercial TTS engines. I will this with Speech Switch.
In the future, any Linux user, blind or sighted, will be able to easily install Voxin and likely several other commercial TTS engines. You'll be able to buy a TTS license once, and use that same binary installer on any Linux distro, for years to come. Updates to your Linux installation will no longer break the connection to your favorite TTS engine, so long as your distro tests that Orca works with Espeak as part of their release process. This is my promise with Speech Switch.
Speech Switch currently supports Voxin, Espea, and PicoTTS. Over the next few days, I am going to write the initial version of Speech Switch's command interpreter for SSIP, which is the command language used by Speech Dispatcher. By January, you will be able to replace /usr/bin/speech-dispatcher with a symbolic link to /usr/bin/speechswitch, and everything should just work.
I could use help from volunte who would like to contribute! In particular, I'd like some help writing a new Python library to allow Orca users to select either Speech Dispatcher or Speech Switch. Replacing /usr/speeech-dispatcher with a symlink to /usr/speechswitch is just a temporary hack. Writing wrappers for other TTS engines would also be appreciated. I've tried to make writing a TTS backend module as simple as possible, and I link only to libc, so your compiled modules should port across distros. I also could use help from testers on various Linux distros.
If you're interested, send me an email to waywardgeek gmail com
Thanks,Bill