[Accessibility] TTS API interface description + reqs update

[Jonathan Duddington]

say_char (wchar_t* char)

   Should the parameter type be  wchar_t  rather than  wchar_t* ?
   Isn't the parameter always a single character code whose name is to
   be spoken?


voice_description_t
   wchar_t* name;
   wchar_t* dialect;
set_voice_by_name(wchar_t* voice_name)
 
   Would  UTF-8 char *  be a more convenient type for these rather than
   wchar_t* ?  The voice name and dialect are constant strings which are
   only copied and matched by S.D. and synthesizers.  UTF-8 is the form
   in which such names are generally stored in configuration files,
   etc.


say_key (wchar_t* key_name)

  Similarly to the above, would  UTF-8 char * be more suitable than
  wchar_t* ?

--