I'm not going to try to explain the technology behind it (that's what the paper for), but it a nutshell it uses a small light source to create an image consisting of thousands of pixels. The pixels are only discernible when a camera is looking at the Bokode while its focus is set to infinity. I hope the next video explains it better:
As the video above shows, there are very nice implications to augmented reality. Aside from coding the identity of the object, it can also encode how's the object positioned in comparison to your camera. Though, if I understood correctly, the demonstration above uses two cameras, one shooting the object in focus, while the other looks at the Bokode.
Another obstacle in the way of wide adoption is that the Bokode currently requires an energy source to operate. Nevertheless, it has already taken a step in the right direction, and currently have a short page on Wikipedia.
More information here and here. Via Augmented.org.