Since the collapse of Lernout & Hauspie, voice has been diminished as a computer interface.

But it makes sense. It's hands-free. It requires training, meaning it brings some security with it by default. I continue to believe in it.

So does IBM.

Igor Jablatov is the man behind IBM's voice strategy. He's based in Charlotte, and has a blog, which mainly prints and links to stories and news release relating to VoiceXML. (Jablatov now heads the VoiceXML Forum.)

The Voice Extensible Markup Language brings voice into the Web standards area, and it's important for that reason. But what's more important is the extension of voice into specific vertical markets. IBM has started with things like cars and consumer electronics, and next plans a move into CRM.

These aren't the markets I would have chosen, but for now voice needs to choose markets based on their money making potential, nothing else. And I trust that IBM has done that kind of analysis here.

Where do we go from here?

With VoiceXML activated in a home-based 802.11 network, you have the perfect interface for Always On applications.

An old person falls down, and their shout, by itself, triggers help from caregivers, maybe even the fire department. Medical monitors can alert a patient, by voice, to take specific actions, and can then issue and alert if those actions aren't taken. Lost your keys? "Where are my keys?" you ask, and the RFID tag on your keys reports to the network, the network responding, much as a good wife (of hyper-organized husband) might.

Voice, as I've said, is the perfect interface for Always On applications, written on a robust, scalable operating system, living in the air of an 802.11 network, using data you create in your daily life, and awaiting your command.

This is how we'll live, in the future. And the future is coming fast.