There is another perspective to take as well.
BryonD, you are absolutely right, I think. 4e is more homogeneous than earlier editions. But, and you knew there was going to be a but here didn't you

, there's a question of degree.
If the range in 3e is 0 to autowin and the range in 4e is 20% to 80% success rates, that means that 4e has a flatter curve. Certainly. But, that ignores the effects of the rules in play.
IME, what generally happened was you had players who would never bother taking a few ranks in a skill because they were just wasted slots. If you focused, you would reach that autowin state, with autofails in everything else. The problem was, at higher levels, you couldn't spread the points around to give yourself a decent chance at a number of things, simply because you didn't have enough points (barring a few classes of course).
So, the range actually led to fewer options being exercised at the table. Everyone laser beam focused on a few skills and, across the group, you usually had someone who autowin'd every time.