One of the things that struck me as a big improvement in 3e compared to prior editions is its ascending AC system. Back in the day, a lower AC was a better AC. You rolled a die, and either looked up the result on a chart or used THAC0 to figure out what AC you hit.
3e used an ascending system, where a higher AC is better. You rolled, added modifiers, and the result was the AC you hit. It seems patently better, but like a lot of things that changed over the course of D&D's history, it's better only within the larger context of the 3e rules, rather than in the context of D&D as a whole.
In 3e and 4e, there are lots of modifiers that go on top of that die roll. Not only do you have modifiers that apply to every roll, like ability score mods and magic items, but spells, conditions, flanking, and so on. There's a decent chance that 25% or so of the attack rolls you make during a session require some additional modifier beyond ability score and a magic weapon's plus.
In that situation, descending AC is a terrible idea. The table lookup or THAC0 math is just an extra step of work. Why not just use the final result?
However, strip away the fluctuating modifiers and the descending AC system comes into its own. At that point, all you need to do is record your to-hit numbers vs. AC on your character sheet. The process of roll and look up is, IMO, much faster when the players work up their own little attack matrices, faster than dealing with any math on the fly.
From the DM's side of things, you can do the same by tracking monsters on index cards. Just write down each critter's line of attack results, and you're done. The real drag with the system, IME, is using the table, but that's easily fixed.
As a side effect, I think this explains why the notorious weapon vs. armor type table in AD&D received so much flack. James Wyatt is the only person I've personally met who used it. It exacerbated the system's shortcomings and pushed it away from its strength.
The question then becomes, do you like lots of potential modifiers or not? And that, IMO, is a matter of taste.
There are a lot of little transitions like that between 2e and 3e, most notably the sudden explosion in power of spellcasters, that I think have a really big effect on D&D's direction in the past 10 years. In a lot of cases, the changes came about because of shifts in mechanics that have subtle effects on distant portions of the system. For instance, IMO the change in initiative made casters into unstoppable beasts, but that's another post for another time.