Let me summarize my questions in one post:
I don't think "snap" refers to what Brian mentions. He uses two terms:
"ulnar deviation": movement of the hand and wrist left to right
"flexion": bringing the wrist forward.
When we think of snap, we think of the wrist coming forward and down towards the forearm. Like when you dribble a basketball, the hand breaks down towards the forearm.
If you look at top servers, however, you don't see the wrist break down on a flat serve. Check out the clip of Johannson on my homepage (scroll all the way down, lower right corner):
http://www.hi-techtennis.com/
I definitely see his wrist flex (it comes forward to line up with the forearm), but the wrist doesn't ever move farther forward than being perfectly inline with the forearm. After contact, the wrist and hand and forearm are in line. This, interestingly, is what I see with groundstrokes as well. The wrist "catches up" to the rest of the arm on some balls, but doesn't then break past the arm. If you look on my homepage, check out the Nalbandian wiper clip. You can see how his wrist doesn't break forward of the forearm. That's why the tip of the racket never points towards the net in this clip.
From my experience, this is the big difference between top servers and lower level servers. I see people "snap their wrist" on the serve, the wrist breaks forward towards the forearm, and the shot just isn't very good. All the pro first serves I study show the forearm, hand and wrist in line on and after contact.
So to conclude - I don't think "snap" is a good word - but I think it is fine to say that "wrist flexion" is a contributor to the overall motion (which invovles the entire arm).
I'd be interested to hear Brian's thoughts on the term "snap", and if he sees flexion and deviation as being different from a wrist "snap".