� Piracy by necessity | Main | The Responsibility to Protect �

July 24, 2005

Speech Recognition - Who cares?

I often wonder if vendors develop products in a vacuum. Voice recognition software has been available to the public for several years. To date, I do not know a single person who uses the software. Conceptually, it is very cool to be able to talk and have your PC/handheld or whatever device and have it understand and execute the construct as appropriate. However, is it practical for the mainstream, joe average or better yet M$ user? I'm not certain that it is. Consider that you must teach the machine to understand your dialect, intonations, inflections, etc. The computer is only as intelligent as the space between the monitor and keyboard.

Granted it would be nice if I could simply speak my SMS into my Treo600, but I could only use that feature for one or two applications. Clearly, I would not want to utter my passwd, to gain access to email or AIM.

I suppose the bigger issue is that vendors like IBM haven't figured out how to offer a compelling value proposition to the consumer. Clearly if there was a huge demand for the product, hordes of vendors proprietary and open source vendors would be running to develop the applications. Yes, I do know that there are niche markets (ie vision impaired and other physically challenged users), but unfortunately, they are not the majority of computer users. So obviously, there are technical solutions available (albeit various levels of success), but demand is moderate at best.

As the author states in the article below, Speech recognition takes a huge amount of clever programming to perform in a capable fashion. Hopefully, there will soon be some compelling value proposition to warrant widespread acceptance.

How Speech Recognition Works And Doesn't Work--ExtremeTech Feature

Posted by AG at July 24, 2005 5:30 AM