Addressing the response to the rogue vs. goblins example from a page or so back, since I haven't been able to get to it until now:
Yes, absolutely.
Because your example doesn't work. The rogue kills goblin 1 and then trips Goblin 2. Goblins 3 and 4 run away and you don't get the chance to trip them. Why are they standing around waiting for you to trip them?
OTOH, if I have an encounter, or perhaps a daily power, that is considerably more powerful than a single trip attempt, I can trip targets in an area attack effect, tripping the entire lot of them in a single round.
But, if I do that, and then do it again, and then do it again, that's boring as hell.
You're assuming 1 attack and no ability to chase; a 3e rogue could have TWF to trip the two closer goblins and then charge after the third after it runs away (the goblin opens the door and moves 30 feet, and the rogue runs 120 feet past him; if the goblin tries to run past, the rogue can charge and trip). Yes, if you have exactly 1 attack and are on a featureless plain some combat maneuvers are less worthwhile...but then, featureless plains are fairly rare whether you're in dungeons or outside them.
And the above example brings up another advantage of an a la carte maneuver system: you can use them on attacks. If you have 3 attacks per round, you can attack/attack/attack normally, or you can attack/trip/attack, or disarm/trip/attack, or any other combination. Meanwhile, powers are atomic, so you can't combine them like that unless a power is specifically made to modify other powers.
The basic point here is that some attacks should be more powerful. Not every attack that the character attempts should be the same power. That's what we had in 3e and earlier. And, you get exactly that - the non-caster characters do the exact same thing, over and over and over again, round after round, level after level, with the occasional bright spot where they got to do something funky.
All because no attack option could ever be better than a basic, vanilla attack.
3e went some ways towards mixing things up - they added decent mechanical resolutions for various common maneuvers. However, the problem here is that the common maneuvers become better than a regular attack, thus, you simply shift from making a standard attack all the time to making a trip attack all the time. And, of course, this generally only applied to fighters because they were the only ones who could afford the feats to be able to spam special maneuvers.
It was an improvement over what came before, but it didn't go far enough.
4e says, as Pemerton rightly points out, "Look, some attacks are just flat out better than other attacks. This power here? It does twice as much damage as your regular attack. But, we can't really let you spam that because then, well, you don't need a regular attack. So, we make it an encounter power."
No one is claiming that martial characters should be spamming the same thing over and over. What people want in a general maneuver system that a power system doesn't have is combinatorics, as mentioned above. You should
not have a Trip That Guy power in your system, because tripping a guy is something you should just be able to
do--whether or not it is reliable or worth it in a particular situation is another story. You
should have either a Be Better At Tripping Guys power (which makes the base maneuver better) or a Trip That Guy Differently power (which alters the way the maneuver is used, such as making it an AoE or a reaction).
Only things that legitimately should be rare would be good power candidates. Pushing a guy 20 squares is a power; pushing a guy 3-5 is not. Tripping something 4 sizes larger than you is a power; tripping something your size is not. Slamming someone into a wall and breaking their back is a power; slamming someone into a wall and doing some damage is not. One of those options in each case is a dramatic power-up, a scene-defining ability, and/or a powerful option to be used sparingly; one is a basic tactic that should be usable as the situation warrants.
Like Pemerton points out, it's empowering players to be able to decide when they want to be awesome. Or at least cool.

If we go back to the idea that no attack can ever be better than a basic attack, but, you can do it all day long, we go back to the idea that fighter types basically do one thing most of the time, level after level.
Hey, if you like that, cool. Me, I don't. I don't want my moment of cool to be dependent on random chance. I want to be able to set it up. Now, I might fail, and that's fair enough. In the last session, I set up a nice little chain of powers with my human warlock, that utterly failed due to poor rolls. C'est la vie. But, the attempt was still under my control, not the DM nor random chance.
I see this sentiment in many threads. The 4e rules do not prohibit combat creativity, nor do they say what cannot be done. Instead, the rules say what can be done reliably. The rest has to be judged by the DM, similar to AD&D or OD&D. The DMG (p. 42) and the DMG2 (terrain powers section) give excellent advice on adjudicating these situations. I do agree that the rules could have been more explicit about trying things besides specific powers. I think that is a victim of moving combat rules from the DMG to the PHB back in 3e.
The new edition of the game need only provide a set of clear guidelines, similar to the 4e p.42 and the DMG2, and combat maneuvers will be used. They could even throw in some combat stunt rules such as in Iron Heroes. Didn't Mearls work on that or something?
You know, I hate page 42. I really
hate page 42. The general concept is good, telling the DM what appropriate math is for various things at various levels, but the execution is terrible. Rant spoilered for length:
[sblock]First off, the numbers provided are weaker than powers. The rationale given is that you don't want to make situational abilities better than powers, but (A) situational abilities
should be better than powers, just like dailies should be better than encounters, otherwise there's no point in using the brazier or the acid pit or the spiky wall or whatever in place of your powers and (B) it reinforces the message that using powers is better than improvising.
Second off, "have the DM make it up" is not a
system, it's a
suggestion. You can't codify every last bit of math and write up pages and pages and pages and pages of powers and then say "Oh, the DM can wing that stuff." Either the DM can make up everything, in which case you don't need powers, or you should provide rules for everything so you don't have such a stark disparity between powers and "everything else."
Third, having rules is essential for combat creativity. As Hussar said, players should be able to choose when to be awesome, not just hope that the DM won't put too many checks in their way to make them fail. If you have rules for tripping, jumping, shoving, chandelier-swinging, etc., a player knows what his chances of success are. If you have to play Mother May I with the DM, you're better off not bothering. Many people have praised powers for codifying rules and empowering players, for letting you do something, period, instead of having the DM ask you for 5 different checks when you try something out of the ordinary. Precisely the same problem occurs here: lack of codification is the enemy of creativity, because players can't accurately judge risk and reward and because choosing between reliable powers and unreliable DM fiat is a no-brainer.
Fourth, any rules are better than no rules. How well would combat work if the books gave no rules for difficult terrain and AoOs, instead saying "If a player moves into difficult terrain, assess appropriate penalties according to these guidelines" and "If a player does something that would probably be distracting, give their enemies AoOs"? It would be clunky as heck and wouldn't work well. If you
have rules, DMs can use them, ignore them, or modify them as they see fit, other rules can reference them for enhancement or modification, and players are empowered; even bad rules provide a starting point for houserules. If you don't have any rules, their quality will vary proportionally to the DM's, and they live in a vacuum because no other rules can say how often or how well they'll be used. It is well within a DM's right to change the rules, and he should make rulings when needed, but he should not be required to continually come up with rulings on the fly for common situations when rules can be given instead.
Finally, page 42 is a cop-out. When people start talking about the various minutiae that the rules don't cover in enough detail, it is guaranteed that someone will say "Oh, just use page 42!" or "Just grab the numbers from page 42 and wing the rest!" or "You don't need rules; guidelines like page 42 are fine!" No, guidelines are not fine; D&D hasn't been codifying more and more rules with every single edition for kicks and giggles. Guidelines should be reserved for things like monster frequency, encounter composition, and other things that will vary by party, setting, and campaign theme; basic gameworld laws of physics like "How do I knock something over?" should have concrete answers.[/sblock]
TL;DR for this post: If you're going to standardize powers, you should standardize maneuvers instead of relying on rough guidelines; not everything should be a power, only legitimately special things; leaving holes in the rules for DM fiat to fill is bad.