Dawkins should probably read a recent article on Phys.org for a more updated view on our eyes. The article is a report on research published in Nature Communications:
“Having the photoreceptors at the back of the retina is not a design constraint, it is a design feature. The idea that the vertebrate eye, like a traditional front-illuminated camera, might have been improved somehow if it had only been able to orient its wiring behind the photoreceptor layer, like a cephalopod, is folly. Indeed in simply engineered systems, like CMOS or CCD image sensors, a back-illuminated design manufactured by flipping the silicon wafer and thinning it so that light hits the photocathode without having to navigate the wiring layer can improve photon capture across a wide wavelength band. But real eyes are much more crafty than that."
The article goes on to explain:
“A case in point are the Müller glia cells that span the thickness of the retina. These high refractive index cells spread an absorptive canopy across the retinal surface and then shepherd photons through a low-scattering cytoplasm to separate receivers, much like coins through a change sorting machine. A new paper in Nature Communications describes how these wavelength-dependent wave-guides can shuttle green-red light to cones while passing the blue-purples to adjacent rods. The idea that these Müller cells act as living fiber optic cables has been floated previously. It has even been convincingly demonstrated using a dual beam laser trap. In THIS case (THIS, like in Java programming meaning the paper just brought up) the authors couched this feat as mere image transfer, with the goal just being to bring light in with minimal distortion.”
In other words, the bad design argument of the human eye is a bad argument. The eye itself is the product of wonderful workmanship.
Hewitt, John. 2014. Fiber optic light pipes in the retina do much more than simple image transfer. Phys.org (July 21).