[Kde-accessibility] gnome-speech

[Bill Haneman]

On Mon, 2002-09-23 at 19:26, Pupeno wrote:
Hi Pupeno:

It turns out that Marc Mulcahy is going to New York to attend the
presentation of the Helen Keller Achievement Awards and might not be
able to reply for a couple of days.

I had forgotten that the new IDL for gnome-speech is not in CVS yet, the
current CVS codebase is an implementation of a much simpler API (though
it's still useful).  The API currently implemented has commands for
stopping and starting speech, but not for multiple "speakers" and
voices, and not for setting things like speech priority levels (for
interrupting speech) or callbacks to give notification of speech
"progress".  The progress notifications are especially important for
stop-and-start of speech, and for synchronizing speech with other
things, so really you probably would want to see that API instead of
just the simplified one that's already implemented.

We expect to have the new IDL implemented very soon (a couple of weeks).

In the meantime I attach a fairly recent draft of the IDL.  I believe
that rather than keeping the "say", "stop", etc. methods in the
SynthesisDriver class, we will be moving them to a Voice or Speaker
object, to prevent the need to call setCurrentVoice() on a driver before
calling "say()".

The "Say" method takes a UTF-8 unicode string that optionally contains
XML-style markup.  In that way, a single string can contain multiple
languages as well as emphasis, etc, and it's up to the driver
implementation to figure out how to speak the substrings in the
appropriate language, substituting the nearest-available voice if
necessary (for instance if the chosen voice cannot speak in <german>
then the driver could substitute or create a similar voice to say the
work in question).

Best regards,

Bill

> Does anyboyd knows if there's something written about gnome-speech
> architecture ? or do I have to look at the sources ? do I have to get the
> sources from the CVS, right ?
> Thank you.