I don't think Johnson has a point about backward wiring at all.

All questions about Designer's motives are ultimately Theological/Philosophical. Just because Johnson's god would not design "inverted retina" in humans, has no bearing whatsoever on ID.

This way of thinking about biological features also exposes bankruptcy of this ultra reductionist world-view. "Inverted retina" is part of visual system that works very well in humans and birds, so on what basis does Johnson call this "bad design". For reasonable person to conclude this is indeed a bad one, he would need to show that existing design does not work very well and to demonstrate an improved design. I am afraid my imagination is not good enough to see how could this hypothetical new design be scientifically demonstrated.

But even if we agree, for argument sake, that this is bad design and demonstrate that new design would work better, that would still be a bad argument against ID. As far as I know, ID does not say Designer directly designed biological features and it does not argue against evolution.

This seams to me like Johnson is merely parroting the likes of Dawkins and Miller and also showing ignorance about our visual system.

I think it is at least arguable that the vertebrate eye is instead an optimal design considering all the different tradeoffs. It's a biological system, not a video camera. The inverted retina places the photoreceptor ends of the rod and cone layer in direct contact with the choriocapillaris (capillary layer) behind it, which supplies it very efficiently with blood and oxygen. Such close contact is necessary because the photoreceptors have very high energy demands (very high metabolic rate) and need a lot of oxygen and food and immediate evacuation of waste products. If the rod and cone cell layer were reversed with the active photoreceptor ends on the outside, the blood vessel choriocapillaris layer would have to be between the light and the photoreceptors, and because the blood vessels highly absorb light this would drastically reduce both sensitivity and resolution.

The ganglion and optic nerve layer being in front of the photoreceptors still would reduce sensitivity and acuity, so it turns out that there are multitudes of radial Mueller (glial) cells going through the entire retinal thickness in the direction of light propagation, acting as low absorbtion light guides through the nerve layers. So (at least for mammals and reptiles with high metabolic rates) the existing design appears to be highly optimal for achieving maximum resolution combined with maximum sensitivity. This mostly is a summarization of Denton's argument.

This inverted retina design seems only to be optimal for higher and warm-blooded vertebrates with high metabolic rates, so then the question is why it originated in primitive ancestral cold-blooded forms that presumably didn't need it? Just the luck of the draw, or intentional preadaptation (front-loading)?

First, something like this does count against ID, which is one reason why teleologists attempt to account for this type of design. Johnson's mistake is in thinking that it not only refutes ID, but that it refutes ID everywhere.

Second, it helps us appreciate that the blind watchmaker is only "concerned" about making things "work." He writes: Darwinism is usually described as "survival of the fittest," but "survival of the sufficiently fit" might be a better shorthand.

Indeed.

Wouldn't a designer have run them behind the retina, as wires in a camera are run behind the film?

Hey, I just opened my fancy camera and I see no wires. But wait, there is no film either!