I highly enjoyed some of the original ideas put forth in which ability scores played a heavier role in skills and doing things. I get the impression (and some of the designers have said things which make me believe) that some of those original ideas I liked will not be making it into the final version of the game though.
I thik some of what they are talking abut with ability scores and skills related to background is interesting.
Simply decoupling skills from class is a fairly basic move (altough with potentially interesting consequences on its own), but building up backgrounds to be a richer feature of the PC and the PC's embedding in the gameworld seems a good development. Where I'm a bit dissapointed at the moment is that the designers aren't giving any indication of how they might leverage this into a rich and rewarding non-combat resolution system. (Though [MENTION=54877]Crazy Jerome[/MENTION] has pointed to hints of this in the DM Guidelines document: on p1, RHS at the bottom of the column they talk about the importance of intents and stakes. So at least they are using the right vocabulary, even if, to date, they're not doing much with it.)
Generally, I like a lot of what I hear in Pem's posts. Unfortunately, my experiences with 4E have gone the exact opposite way. I found the group I normally game with and our sessions becoming (imo) less creative. Playing a game built more on the idea of narrative logic and game logic rather than in-game world logic seemed to have the effect of the group viewing the game more as game and less as rpg. I am in no way disputing what Pem says; I am only putting forth that 4E had the exact opposite impact on my gaming.
I think the relationship between playgroups, mecahnical techniques, genre expectations, preferences for how narration is handled, etc are all very idiosyncratic. So maybe it's not that surprising that different people have different experiences with the same game.
I also came into 4e looking for certain things, based on my growing dissatisfaction with Rolemaster and especially its lack of conflict resolution mechanics outside combat. One clear example of frustration, for me, was GMing Bastion of Broken Souls (having converted it to Rolemaster, and tweaked it to fit it into the cosmology and thematic flavour of my campaign, which was roughly "mortal freedom vs heavenly-mandated karma"). In that module, there is a scene where the PCs wish to meet with an exiled god, and the gate to that god's prison plane is not only guarded by, but actually takes the form of, an angel. To open the gate you have to kill the angel.
Now the module says that the angel won't talk, and has to be killed. (The module, as written, is annoying like that in several places.) I naturally disregarded that, and one of the PCs - nicely played by his player - tried to persuade the angel that the only way to stop massive suffering in the mortal world was to let herself be killed, even if this meant going against the decree of heaven that she was carrying out.
Resovling this involved an impassioned performance by the player (at least impassioned by my table's standards! we're not taking out any drama awards), mediated via skill checks (I can't remember now what the relevant skills were, but maybe Leadership, Etiquette and Conciliation). The problem was, I had no real basis as a GM for deciding when they PC (and player) had said enough. Allow the first roll to be adequate, and the scene is cut short and its dramatic potential undermined. Keep insisting on more rolls, though, and I'm making success here impossible (and pushing the players to default to combat, which has more objective adjudication procedures). The whole situation is getting very close to "mother may I" and an ultimately arbitrary choice by me to let the player suceed or not.
For me, the skill challenge framework (or anything similar) solves this problem by establishing boundaries, and a pacing dynamic within that, that let's me play my NPC and evolve the situation in interesting ways without having to decide, essentially arbitrarily (as it seems to me) when enough skill checks, or a high enough skill roll, do the job.
If I hadn't come to 4e with these sorts of dissatisfactions in mind, and with a conscious desire for a different sort of game more indie-ish in the sorts of ways I've described (but still gonzo fantasy in its underlying tropes and themes), I might have had a pretty different experience.
For me personally, I moved from playing D&D 4E to learning GURPS. The negotiations with traders and things of that nature which have been mentioned were things which happened when I branched into different games. I would never suggest those things never happened in 4E. I ran a few games which I feel were very successful, but -every step of the way- I felt as though I (as the DM) was fighting against some sort of preconceived notion which the players held about how a 4E game should go.
I mention my transition to a different game because it is my understanding that RQ is more similar to a game like GURPS (or at least far more similar to it than D&D in so far as the games are built upon different ideals about gaming.) In Pem's experience, there were things that would not have happened in RQ that did happen in D&D 4E. I moved from D&D 4E to a different game (GURPS in my case) because I found that the things I wanted to happen and the style of game I wanted was easier for me to achieve with the game I moved to. It's interesting to me that the same game can have what I feel is the opposite effect on two different people and their gaming habits.
I hope what I've said above explains why Rolemaster (or Runequest - it doesn't significantly differ from RM in the fundamentals of its treatment of non-combat skill checks; and I agree with you that are both are closer to GURPs than any of those three games is to 4e) put limits on what could be done in a satisfactory way.
Now I think there is probably more combat in my 4e game than in my Rolemaster game, just because 4e PCs are so primed for combat when you look down a character sheet, and the combat mechanics beckon! For 4e (as, in my view, for other versions of D&D as well) I think that combat is the primary site of conflict resolution.
But I haven't felt I've had to push against the game to introduce the non-combat stuff that has been part of my campaign. I don't know if this is because of my own experiences and preconceptions as a GM, or something about my players, or the fact that I have mostly ignored the WotC modules, or some combination or influence of other factors I'm not thinking of.
And even within combat, my players have done creative stuff from very early on - like the player of the paladin using his religion skill to say prayers when in combat with undead in a very early session. (I allowed a check for combat advantage staked against a modest amount of damage from the undead's backlash against an ineffective prayer.) And as we've got more familiar with the system, the players continue to try more interesting stuff, especially using their attack powers out of combat. The page 42 framework, and other things I've extrapolated from that (like the importance of keywords), plus a freedom to rely on "genre logic" and metagame/narrative "causation", have been key here.
If we changed to GURPS or back to RM or whatever, the issue for me wouldn't be a squelching of player creativity, but rather a lack of GM tools to adjudicate non-combat resolution in the ways I've become used to. And hence, I think, less unexpectedness in resolution than what we've been experiencing in 4e.
A final comment: from other posts of yours, it seems that your 4e game got bitten badly by scaling problems which undermined both the narrative and the mechanical integrity of play. Whether through good luck, good management or maybe just not having pushed so hard against the fragile areas (like destruction of objects, say) my group hasn't had that experience.
But I can see how it is a real problem inherent in the system. I think bounded accuracy is one of the better things being suggested for D&Dnext.