See, this is interesting. I completely agree with Tuco, but it's interesting to examinewhythis is the case. I think, and I think most people agree, that the cause is due to the focus of most games being purely combat oriented. Draegen was talking about it earlier, how leveling systems could be tied to achieving goals, rather than killing monsters, so progress in the game isn't all tied to "reduce HP to zero=win". And while I don't think Brad and Co should focus on exploring that too much, because it would be REALLY pushing the limits of what you can do with programming right now (And from the looks, focus is a key goal this time around)--I think it's an interesting thought exercise on how we perceive progress in MMO's, and it's one being explored in other genres (Especially stealth ones).
The only reason DPS needs to be that ubiquitous is because progress is so tied to health reduction of enemies. In table tops, that's not really the case--and I think the biggest display of how this has affected classes is the difference between the table top rogue, and the "MMO Rogue". You can see how the rogue has been turned from this largely non-combat class (Or a class meant to resolve problems without the opponent fighting), into a class that's a pure offensive combat class--all because of the whole genres technical limitations. And in a lot of cases it feels very constricting and bland because of how ubiquitous DPS needs to be.
Anyway, not sure how I'd fix that without some crazy AI that allowed someone to design all sorts of different paths for completing progression. But it's something that should be kept in mind when designing classes--DPS classes should also have a broad range of utility that works outside of combat but makes encounters easier. Like lock picking, or tracking of feign death--or a host of other utilities that give DPS unique roles to make up for the fact that their primary job has to be shared among all the classes.