@magicalclick: Actually, now that I think about it, you probably shouldn't use Kinect to fire Pointer events, and there's probably no way to properly abstract a single event model for all forms of interaction.
Because each type of interaction is so different, you have to build a specifically targeted event model for each one, otherwise you might be limiting what can be done with the new features of the device. Even the Pointer events are limited to points on a 2D surface. What if a future version of the Surface tablets use PixelSense? The capabilities would require a separate event model. Kinect requires an event model that understands 3D gestures. I don't think you are going to see a unified interaction event model. What you might see is like what they've done with Pointer, common events, and then device specific events that can be included if enhanced functionality is desired.