This posting is older than 6 months and can contain outdated information.
Having actually tinkered with the Kinect at the software level myself, this would be a pretty tall order for the developers here.
What the Remote Buddy software does right now is radically different from any of the software hacks out there that would be needed just to even get a kinect unit running, let alone usable enough to function as a reliable input device.
So far, the only reason you see "hackers" toying with the kinect with their pet tricks, rather than end users just plugging it in and getting things done with the kinect is because development hasn't yet advanced far enough to make it safe or even easy to install on an end-user system. Just to even get a kinect up and running requires some programming ability to compile and install the drivers. From there, even calibrating the thing is a complicated and tedious chore. (At least, until someone figures out how to extract this data directly from the kinect itself, so things can be simplified down to the Xbox 360's form of "calibration".
From there though, that's just getting the kinect *running*... the gestures and what not aren't built into the kinect hardware, so all of those "Minority Report"-style fantasies you have aren't going to come easily. All of the gesture data that would be needed for the bulk of Remote Buddy features doesn't even exist yet, so it'd still have to be collected (likely from several sources), and then parsed/processed by some type of visual AI system, like "OpenCV", so whatever software was created to add kinect support could compare footage from a live kinect feed will all of that collected data on file, just to tell what each gesture means and to determine how the software is supposed to respond to each one.
Needless to say, this is anything but simple.
Not that it *couldn't* be done by anyone determined enough to try it... it's just not "practical" yet.
