Design suggestion: The server just for synthesis

[Halim Sahin]

Hi,
I think your sugestion makes at leeast no sense to a11y applications
because it would force all screenreaders to implement their own
audioplayers which would duplicate work and is not useful at this point.

I agree that speech-dispatcher need a way to output the raw samples EG.:
for txt2mp3 creation :-).
Any way this doesn't need to refactor the current server or create a new
one.
It could be enough to create a a file output driver for speechd in
spd_audio library and add an api function for setting the output
filename.

Just my two cents.
Halim