The problem isn't with the wrist movement, per se. The problem is with the teaching concept of the forward wrist "snap." In my opinion, this gives players the wrong image of the actual path of the hand and racket.
If we look at the motion we can see that after contact the hand and arm continue to rotate in a counter-clockwise fashion. This part of the motion has come to be described as "pronation," although it's really just part of a continuum of the rotations that begin when the racket starts upward to the ball. After the contact, the arm and racket continue to rotate or prontate. As they do, they stay in a virtually straight line. This means the racket face is in line with the arm and you can draw a straight line outward from the shoulder to the tip of the racket. There isn't a forward break of the wrist at the contact or in the following part of the motion. This wrist break occurs, to the extent it happens (and it doesn't always happen) well out into the followthrough when the hand is already coming down.
Philipoussis displays a great example of form despite--or because of--the backswing.
This observation of the actual position of the hand in the high speed footage was the basis for challenging the common serving tip: "snap the wrist" for power and spin. In my experience, players who were actually able to follow this advice, and snap the wrist dramatically forward after contact, tended to significantly alter the shape of the motion. If you watch Roger, as his arm and racket continue to rotate they also move forward but also from his left to his right. They reach an angle of about 45 degrees to the baseline before coming down and back across the body to his left. Players who"snap," or try to snap, lack this sweeping left-to-right motion. Instead the motion breaks off sharply in front of them, usually with a restricted followthrough. The whole thing usually looks tight and forced.
There are really two separate issues here to consider. The first issue is what the bio-mechanical components of the motion really are, their quantitative positions in space and time, and their contributions to racket head speed. The second issue is how to make this happen. This has been the focus of my work in high-speed filming. My assumption is that if you can help a player match the physical positions in a great technical motion, then you are probably helping that player maximize the efficiency and the effectiveness of his own stroke.
The quantitative approach developed by Brian can only help us all in our efforts to define model technical positions more clearly. But what happens and how it happens can be separate issues. This is where the art of coaching comes in. The movement of the wrist contributes to racket head speed no doubt. The question is how to help a player to create a motion that naturally includes this, as well as the other equally important, or even more important, parts of the motion. Different players will definitely respond better to different input and to different coaches in their efforts to achieve this. I don't think we have to worry that the debates about how best to do this will be definitively settled in the immediate future.