speechd-discuss
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Everything about unified interface


From: Michael Pozhidaev
Subject: Everything about unified interface
Date: Mon, 05 Jul 2010 10:12:03 +0700

Hi, Hynek!

I read your messages here and in orca-list. Mostly I agree with
you. I've been incorrect about TTS API position, now understanding
better. No problem to accept and support brailcom initiatives. But some
desirable things: please, do not assume any existing speech server can
be single possible (spd or any other). The reason is not I like my own
developments. Multiple alternatives of any service is very customary
situation and freedom to choose any of them is essential right in
FOSS. You wrote you'are not happy with several providers but we have
emacs and vim, gnome and kde, postgresql and mysql, a lot of examples,
what's wrong here?  The single solution of some task from here to
eternity is the science fiction.

Even more, multiple projects for any task is general
indication this task is very actual. Take a look on development of
distributed file systems (databases) now. I see several projects and I
know they are highly needed. End-user must not suffer from this
situation, so general standarts are our goal. Especially in case we have
good way to do it -- D-Bus.

Next, we must remember two projects can differ only by
they implementation or quality of implementation. I can say aloud, not
everything in spd implementation seems to me very appreciable. OpenTTS
developers have fixed a lot of memory leaks and speak about some other
problems. VoiceMan has long development history (since 2003) and I have
made several attempts to see what I can do with spd sources and found
them very confused. Current VoiceMan has clear and flexible structure
and code quality is the most important thing for me despite some
disadvantages it has. I can easily be wrong, I just write what happened
earlier. In any case, now I have more than appropriate for me to work solution 
and can
offer it to everybody interested.

I agree it is good idea to merge everything created into one common
project and ready to talk. It will be very difficult but not
impossible. If you want to cooperate we need OpenTTS developers
opinion. What they are thinking about common D-Bus interface or
merging. Some more things below.

> the applications to care of many details, if it doesn't want.
> It is also evident that it will need to provide some emulations
> for at least some functionalities which are not provided by lower
> level layers.

Of corse. It is subject of middleware functionality (speech server,
speech provider, anything you want). Middleware must be able to overcome
missed features of TTS engines.

> This is why I say that the work towards a common DBUS
> API has *partly* already been done.

All right. Can you show any fixed results of this partial work? IDL's,
documentation, anything else? Maybe it is time to move from general talk
to exact things.

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




reply via email to

[Prev in Thread] Current Thread [Next in Thread]