At the risk of being a thread diversion, I have a question? If Linux can be programmed to find hardware speech at the console level, why cannot it be programmed to find that same speech from a desktop? I have been told even windows can work, to some extent, with dectalk units for example. Kare