Voice support

I am currently working on an XForms implementation
that also supports Voice. My approach so far was to
use XSchema in the model to generate grammars and
datatypes to aid the recognizer, but effective use of
XSchema to do this requires writing really complex
Schemas (such as large enumerations) and isn't really
appropriate since the Schema represents the canonical
value, not what is recognized, which are two different
things.


Voice grammars appear to be largely a UI issue, so
right now, I am working on an implementation that uses
the new <extension> field to attach grammars to UI
elements, such as

<input ...>
  <extension>
   <grammar xlink:href="location of extension
grammar">
       internal XML grammar (see W3C
http://www.w3.org/TR/speech-grammar/)
   </grammar>
</input>

For mixed-initiative dialogs, I use a top-level
<extension> inside <xform> to include global grammars.
Grammar rules can have their "slot" attach to the ID
of of a UI element (or a binding), so that the normal
navigation sequence algorithm is overridden and
multiple bindings or UI elements can be filled in via
a single utterance.

Now, this works, but I abhor doing this in a
proprietary fashion and I was wondering if the Xforms
WG is considering Voice support, perhaps in a follow
on side specification as one of the profiles (XForms
Voice Profile?). Also, is there going to be a side
spec for dealing with UI layout and CSS issues? I
noticed these were removed from the latest draft.



__________________________________________________
Do You Yahoo!?
Make a great connection at Yahoo! Personals.
http://personals.yahoo.com