Sorry for the slow response--still digging out after SHOT.
Although I have been shooting for a long time, my perspective on personal defense began with empty-hand training. And one of the first and most important rules of that is to maintain a guard. That lesson was learned the hard way during early training (getting hit in the head is outstanding motivation to learn).
When I got deeper into shooting and started exploring all the different technique options, I ran into what I feel are some serious contradictions. For example, everyone acknowledges that handguns are imperfect "stopping" weapons and that bigger calibers, better bullet designs, and multiple shots are all recommended to stop an attacker decisively. Everyone also acknowledges and "people who need shooting" include all types of threats--from contact-distance knife and club wielders to more distant gun toters. Knowledgeable people also acknowledge that many incidents occur well within the mythical 7-yard range.
Assuming that we're in basic agreement on all these points, it would seem that to be prepared for a likely type of attack, one would need to be prepared to shoot effectively against a at short range against an attacker potentially armed with a contact-distance weapon. Yet despite the logic of that assumption, most shooting techniques actively and purposely advocate putting your hand on your chest or some other place where it does you absolutely no good and you are totally vulnerable to being clanged in the head or stabbed in the neck as you attempt to draw.
There are two basic reasons for this:
1) The techniques were developed by shooters to work well in the static environment of a shooting range. The traditional Speed Rock is a perfect example of this. If you were fighting WITHOUT a gun, the idea of hitting an armed attacker once with your weak hand and then leaning back to put yourself on the edge of your range of balance would be ludicrous. Similarly, the belief that you'll get instant stops from your pistol because you happen to be closer to the target doesn't make sense. Putting two flawed thoughts together doesn't solve the logic problem. You still get hit in the head with a pipe and, if you're quick enough, he gets shot. The Japanese have a term for that--"mutual slaying." That's not good enough for me and sure makes me wonder about them.
2) Liability in training. When students shoot themselves in the hands, that's bad for business. Since "the business" for most of the people who practice this stuff is teaching shooting-oriented tactics, they had to find ways to address close-quarter shooting while minimizing liability. The way to do that is to create techniques that get your hands well away from the muzzle and, consequently, teach you habits that work well on cardboard IPSC targets but don't make sense in a real fight.
In basic terms, reholstering is the exact reverse of your drawstroke. Neurologically, it's actually a great way to reinforce body mechanics and familar task transfer (a better term for what is normally known as "muscle memory"). My drawstroke assumes the likelihood that I might be shooting someone at very close range. As such, after the initial startle response that brings my hands up (the "recognize" phase of Combat Focus), my left hand CONTINUES to guard my head (most people are right handed, most violent attacks use forehand gross-motor-skill strikes that would target the left side of my body, so this is a good default), and my right hand goes to the pistol. The initial grip, lift, and orientation are exactly as shown on the show and are consistent with Rob's approach. At that point, I decide to either shoot from the weapon-retention position (which I have assumed by default as part of the drawstrooke) or that I have the distance to extend the gun safely. If I decide to do that, my left hand comes back to touch my chest for a momentary physical index so my hands can come together safely (this is a point where Rob and I differ, as he advocates bringing the hands together as part of a continuous flow of both arms). I prefer the idea of touching the chest, but only do so when I know I'm not sacrificing my guard against a potential contact-distance attack.
When putting the gun away, I'm rewinding that thought process. Since the effects of SNS activation include tunnel vision and loss of near-vision acuity, I need to make sure that things are safe around me. That's why I scan AND put a guard back up. As for the belief that the guard would obscure my vision, it doesn't. If you notice the position of my hand, it's about a foot or slightly less from my head. Binocular (two-eyed) vision and the scanning movement of my head allows me to see "around" it without a problem and the benefit of having the guard up far outweighs any minor visual obstruction it might cause.
As for the danger of shooting that hand, in addition to insisting on the chest touch before extending the gun, there is an entire set of body mechanics based on empty-hand fighting structure that help keep that from happening. Hopefully, we'll get to that in a future show or a DRTV web-only video (a concept that is in the works).
Finally, I actually believe it is better to "mitigate" close-range attacks empty handed and then go to the gun, rather than fighting or guarding one handed as you try to bring the gun out. This sequences skills rather than trying to duct-tape them together into a bundle that is often a recipe for disaster. An example of this is coming up in Episode 7 of The Best Defense so please keep tuning in.
I hope this covers everything and answers your question thoroughly. If not, you certainly know where to find us.
Stay safe,
Mike