Quantum light (i.e. photons), linear optics, adaptive measurements and feed forward can be used to build an efficient universal quantum computer. ref: Knill et. al. 2001.
Clearly the above resources can be also be used to build an efficient classical computer (since a QC can efficiently simulate a CC).
Can an efficient classical computer be built using classical light (instead of photons), linear optics and some detection model? I expect the answer is yes.
I am looking for a constructive proof. Particularly
- bits: degree of freedom of classical light
- universal gate set: what operations and measurements need to be performed to implement one such set.
- bonus: do a complexity analysis to show efficiency in time and resources.
Failing a constructive proof, a non-constructive proof will be welcome.