I will nitpick the word selection that "capable" is nearly useless in a user interface. It needs to be nearly 100% reliable or its useless. 99% of my interaction with Kinect is my daughter crying that she can't navigate menus in her dance games and is all frustrated, followed by me being all frustrated and swearing about how if only I could bypass this POS and use the buttons on the controller I would be done twenty seconds ago and I hate Kinect with a passion. It works fine for gross motor like my daughters dance games but useless for fine motor. Perhaps in the future I will write SQL statements by performing an interpretive dance at work, but I hope not.
The failure rate is vital... If I'm typing this at 100 WPM, which is probably about right, then a 99% motion detection success rate means I'd swear and hate motion detection and have to stop and fix an error, what, every six seconds or so? All day long? Forget that, I'm sticking to the keyboard and mouse, I don't have the patience for 99% success.