Listen, by keep repeating that "4e does not need...etch" you insult my intelligence. SC are really irrelevant to 4e's design. They designed 4e encounter mechanics and rules and then they saw how they could fit something like SCs. 4e is really nothing without its encounter mechanics. While it is what it is even without SCs.
And SCs are a failure. Primarily, errata has nothing to do with it. They do not fail just at implementation: they fail at their design goal. SCs as designed could never provide a mechanical way to do what they wanted them to do. Which is to provide to the GM solid possibilities for creating interesting "problems" for the group to solve in a cooperative fashion where everyone needs to participate.
But even if you want to say that they were not meant for this -but rather just as mathematical guidelines, they still fail at their mathematical design premise which is the DM to be able to control with a D20 rng the difficulties and success rates of various possible SC efforts. Neither the DM nor the players can control the knowledge of the difficulties without assistance from some software where you have to input all the possible SC scenarios and let it spell out the probabilities for you. And this not for every SC but rather for every SC roll as with every roll things change. The latest thread I can remember about SC fundamental and core design is this one:
http://www.therpgsite.com/showthread.php?t=17278&highlight=challenges+frank
EDIT: I am not contrary to SC design goal. In fact I think it is a valid one. Only that you cant achieve it by just using SCs the way they are designed.
I don't think the Skill Challenge system is irrelevant to 4E or its design. Yeah, they didn't implement it flawlessly. In fact, it seems the designers just threw it in there because the felt *obliged* to do the "social combat system" they had promised. It's not that it wouldn't work as written; it does, if the DM understands how conflict resolution works in indie games (as we know, the system is a "hybridization" of conflict resolution and complex skill checks in UA). Furthermore, if the DM is willing to relinquish at least *some* control over the story to the players -- and the players are creative team-players who share the same concept of what a "good story" is -- it works brilliantly. In fact, it makes the game *shine*.
However, as you wrote, 4E is a very "gamist" system. The mechanical nature of the game may prove to be problematic; both the DM and the players may feel limited by the game mechanics, and it doesn't help that the whole SC system is vaguely written. Even though 4E is very flexible, many DMs may feel restricted by the explicit skill challenge format -- regardless of those SC articles written by Mearls. Also, if the group has clashing goals or gaming styles, SCs may feel *very* artificial and awkward during play. It doesn't help that there are no "real" stakes there; the most common "penalty" for failing an SC seems to be an extra combat encounter. For many players that is, in fact, a *reward* in the form of additional XP. In this regard the SC system is limited by the very nature of D&D adventures; it's almost unthinkable that the players could suffer a serious setback in the story -- or even "lose" the adventure -- by failing a single SC. Ergo, there is usually no "real" penalty for failure; at worst it's either some extra XP, a healing surge or a minor obstacle (such as having to find an alternate solution to reach their goal). So there's usually very little at stake. From this perspective the SC system, as a whole, may indeed feel irrelevant. Nothing prevents the DM from "raising the stakes", however; it may require rewriting and improvising a bit, but it's possible.
Yet how is this different from "nar" games? Yes, there is usually at least *SOME* sort of impact on the story (and the fate of the protagonists) for failing a conflict. Occassionally the stakes may even involve the life of the protagonist(s), and often a dramatic failure may result in a better story. Still, even the most hallowed indie RPGs require the same things as SCs do to work; even a single player who doesn't get the game's theme and/or share the same concept of what is a "good story" may easily ruin the game for everyone. For example, I've played with people who thought it'd be cool to use narration rights to reveal that someone else's character is a sodomizing pervert (go figure). If the only official rule is that you need to enter a conflict to alter something another player just narrated (with the usual caveat "This game may not suit everyone"), it reminds me of the worst problems with SCs. Except that with SCs, the DM still has the final say over what happens.
So, my point is that even though people may scoff at SCs as a badly-written version of conflict resolution, I feel that it reflects narrow-minded thinking; there's nothing wrong with SCs as written. If the whole group consists of team-players and the DM understand how SCs should work, I think it's a flexible system that bring fresh elements and outside-the-box thinking into D&D. Whatever problems you would have with SCs are the same that would feel just as awkward in most indie RPGs, too.
YMMV, of course.