Design suggestion: The server just for synthesis

[Michael Pozhidaev]

Hello, William!

> I don't think it would be a good idea for you to open a new project for
> this on sourceforge.

No problem, I wrote this just as one possible variant.

> Hynek, can you move the ttsapi code and documents to git repositories
> that the reviewers can commit to?

There are some things: 

1. It seems to me it is a bad idea to mix new code in one repository
with any other project. The TTS server can be useful alone and  any
software in the same package will generate needless dependencies, it is
not convenient.

2. New work promises a huge number of changes in the code and performing
them by reviewing in diff format, I suppose, will be very and
very hard. as intermediate approach I can do this work in my own
repository and ask somebody to pull commits when everybody has no
reasons to thing anything is wrong.

> I'm not following what you mean here, unless you are talking about
> automatic module discovery, which is something that is in the plans for
> the project, we just haven't discussed implementing it in detail yet.

Yes, I mean no configuration file edits must be required to enabling new
module. 

> But, I'm also against a new server; I don't think that's needed at
> all.

I suppose this is real and good way to move TTS API, TTS protocol and
other ideas into practical using. I need this result and can do it with
community agreement so result will be useful for everybody.

For some time this project will be developing independently with
speech-dispatcher so none of already created code will be damaged.  When
we are sure it is ready and can work corresponding changes can be made
in speech-dispatcher. This is my point of view how this work can be done.

-- 
Michael Pozhidaev. Tomsk, Russia. E-mail: msp at altlinux.ru
Russian info page: http://www.marigostra.ru/