Use ssml to tell speech synthesis to speak as character
In some cases the speech synthesis can not know whether it should speak one-character text as a word or as a character (e.g. "y" in French which may be pronounced \i\ as a word, or \i ɡʁɛk\ as a character). Now that we pass ssml to speech-dispatcher, we could use it to mark one-character text. It seems that it could be just implemented in src/orca/speechdispatcherfactory.py 's speakCharacter, what do you think?
I was however surprised that getCharacterName returns "y" for "y", and that's indeed the default value in chnames.py when not(charName or mathName), shouldn't it return None, so that we know whether we really got a name, and not just the character itself?