Voice recognition has been one of the "coming soon" technologiesn in the telecom industry for a long time. Perhaps the complexity of human vocal interaction has been underestimated. We've all heard the voice response systems that some major companies have employed, and of course there was TellMe, BeVocal, Quack.com, and other voice portals that included elements of voice recognition in them. With 400 million mobile phones being sold annually, and a billion mobile phone users, you'd think that there would be more emphasis here. Not so, this year anyway. Windows XP has VR included, but I for one have really struggled with getting it to work properly. Between these two related pieces of technology there should be a great business somewhere. Two of the bigger companies working on this technology at the core level are the West Coast (re Stanford) team of Nuance, and the East Coast (re MIT) team of SpeechWorks. They are forging new ground and maging progress but there is a long way to go before we are the starship enterprise "computer...set a course of the neutral zone, warp factor 3..and can you bring me a tall skinny latte too?"

Part of the problem could be the old addage - "it's not what you say, but how you say it". In the latest edition of Technology review is an article on just that subject, prosody. Check it out.