EETimes.com - Intel CTO rethinks analog as computational problem

I like the Camera example he uses. But this is an idea being pursued by a TI as well, why is it continuously being packaged as a new idea. I can understand  trumpeting new milestones along achieving this goal.

EE Times --- Rattner used the evolution of photography as an example. "Except for the light from the complex lens creating a digital image, nothing much has changed in the camera since film has gone away. Now if you replace the lens apparatus with a computational 'light field', then you can use software to choose which image among many possible light field images you want to capture. "The problem moves to the computational domain and the focus and depth of field can be adjusted at will at the time of image capture," added Rattner. The camera and lens can then be simplified, a good thing since today's mechanical lenses are at the limit of their optical performance. If radio design is considered as a "computational problem," this would allow one radio to act as many. Rattner said this would dramatically simplify radio architectures by utilizing fewer transistors to implement many radio modulation schemes.