Speech Dispatcher and output modules (was: Speech Dispatcher under alsa)

[Hynek Hanke]

Thomas Stivers píše v Pá 17. 03. 2006 v 09:13 -0600:
> Perhaps there could be a way to send ssip commands using
> /proc/speakup/synth_direct although a change may have to be made to
> speechd-up to support it.

I don't think this is the easiest solution. I think the easiest
and cleanest way how to do this is to retrieve the ID of the
connection based on its name and set the parameters from a userspace
controll application, as I suggested before. This is exactly one of the
cases this feature of SSIP was designed for. (Another case is when you
want to change a parameter globally for more applications and you don't
want to need to go to every one of them to use its own interface to do
it.)

By ``inconvenience'' I meant the fact that you can't use Speakups
controll keys to do that and you need to use some application or script
for that. I understand Speakup is not designed for that and I agree that
adding a code to Speakup especially to do this is not a very good idea.

Maybe there could be a generic mechanism in Speakup of how to map keys
to user defined commands and then we could use this. That would make
more sense. But I think it is difficult to implement in a kernel space
application in a nice well configurable way.

> Something like that seems the easiest way to
> let userland scripts use the full power of speech-dispatcher without
> having to add lots of stuff to the speakup code.

A completely external solution is easier, since you don't need
to modify anything. You don't even need to care about the socket
communication part of SSIP, just use spdsend or one of the clients
libraries.

If someone wants to try, he has my complete support.

Have a nice day,
Hynek