Voice recognition has been the idealized computer interface mechanism for years: Speak to your computer, it operates on your inputs and provides a response or action.
In actual practice, though, it has been a general failure. It’s an input method that sees little usage.
Microsoft, for instance, has been pushing voice recognition for years, making it a primary selling feature of Office XP, and then Office 2003. How many people actually used that functionality?
While it has been accurate enough for a demo, it fell below the payback threshold to make it worth dealing with the individual voice training and then finding and correcting the endless mistakes.
Google has passed that return threshold. The voice recognition functionality of Android (more correctly of the Google cloud service performing it, thus it’s not an Android-only feature and hopefully will come to other platforms) is simply spectacular.
I’m not a good test subject given that I have both a big tongue (Jamie Oliver-esque), and a Canadian accent to boot (it’s aboot time I got a Canadian joke in there): My enunciation is far from perfect. Yet for me this thing is amazingly accurate time and time again — closing on 100% — in a context where voice recognition is most useful. Rumor is that iOS 5 will add voice recognition throughout that platform, so it is a facet of mobile devices that will soon be the norm.
It really is a feature of the platform that is underappreciated, and I suspect many simply never click on the little microphone, or they don’t realize the platform control they have with their voice. It is an amazing bit of functionality.