[Accessibility] TTS API document updated

[Hynek Hanke]

Hello all,

I've updated the TTS API draft with the latest suggestions from
the list.

Project page:
http://www.freebsoft.org/tts-api

HTML:
http://www.freebsoft.org/doc/tts-api/tts-api.html

Other formats (PDF, PS, info, DVI):
http://www.freebsoft.org/doc/tts-api/

Changes:

* I've added a function from speaking the given text from a
given character position.

* defer() description now states there is no upper limit on the
number of simultaneously deffered() messages unless forced by
administrator or available resources. A partial defer(), for
one message only, for example, would not be very much useful
because the application would need to use say_text() anyway
to provide consistend behavior across the various synthesizers. 

This point should be further discussed and possibly changed
if someone disagrees.

I've also made clear previously deferred messages must be discarded
by the application when no longer needed.

* I've changed all references in the API requirements
from UTF-8 to Unicode.

* I've added a note that errors due to processing SSML
(not syntax, but functional, like ``language not available'')
are currently not reported.

* Several minor corrections (strings are null-terminated arrays,
typos, etc.) please see ChangeLog in CVS for more details.

If I ommited some change that we discussed and you think it should
be done, please protest.

It is still left for us to write the Use Case section as suggested
by Gary Cramblitt.

With regards,
Hynek Hanke