semantic information

Hi all,

I'm Antonio and I'm new to this list, so sorry if I'll ask something stupid.

This is a proposal I discussed with Aurelio De Rosa after listening to his
talk about the Speech API at jsday in Verona.

Is there a semantic way to inform the client that the application supports
speech synthesys or recognition?
If not, would it be useful to mark in example an HTML tag to inform the
interpreter that the speech interaction is available? So that the
interpreter can inform in the most efficient way the user about the
possibilities.

Cheers,
Antonio