Problem with pure roles is the problem of attempting to differentiate them from existing PHB classes. Lets say the monk is a striker. What is to make him stand out from a rogue or ranger, other than powers with different names, if he ends up playing the same? Likewise, there are already 2, and soon 3 defenders. How might the monk be any different?
The paladin and fighter seem different enough from each other, as are the ranger and rogue. I don't think it takes too much imagination to create new powers that feel different yet have the same goal.
I disagree with the theory that a hybrid class may cram too much into one class. To cite an example, the 4e elf cleric archer archetype. It works fairly well as a leader/striker hybrid. You can still heal well, and deal decent damage, though not as much as a properly optimized striker or leader.
Well, that's not a hybrid class, and it demonstrates that you don't need a hybrid class to play a character with versatility using just the PHB.
This is also a reason why I somewhat despise the idea of roles. They are there simply to idiot-proof the game by allowing new players to quickly get into the feel of how a certain class may be played, but it should not become a straitjacket as to how it must be played (just because a fighter is a defender does not mean that it cannot be built as an effective striker), nor should it needlessly impede the design of future classes (ie: the class must fit into one of the pre-existing role definitions and not be allowed to deviate).
Idiot proof sounds pretty harsh. I want a game that's designed not just for experienced players, but also for newbs and casual players. My friends aren't idiots, but some of them make choices for roleplaying reasons, or they just don't see the point of spending hours min/maxing. I like it when they can't make choices that drag the rest of the group down in combat and result in TPKs.
I realize of course, that some people don't want a game that is "idiot proof" and like that there are strategic challenges in building effective characters. I'm just not one of those people.
My idea of hybrids would that they might not be as effective as a class fully devoted to one of its dual-roles, so it would be more useful as that 5th PC you use to round out your party's existing capabilities, sorta like the bard or warmage in a 3e game. But the advantage would that they allow you a unique playing experience which cannot be replicated using the current available classes (because as you said, a defender will always be a defender). In the end, you would not necessarily be any stronger or weaker, just different.
I think hybrids are pretty much inevitable, to be honest, whether I like the idea or not. If WotC doesn't create them, there will be plenty of third party hybrid classes out there, and some of them will probably be brilliantly designed and fun.
I just don't think that the few good ones will be worth the poorly designed ones. I guess my point is game design needs to be more idiot proofed in 4e, not necessarily game play. Roles help that IMO. By focusing classes into specific roles, it's much simpler to create a balanced fun class because you can compare them directly to existing classes.
I can hope that if the druid or monk is a hybrid in PHB II, they will be well designed hybrids and provide examples for other WotC and third party designers. I'm just not too optimistic about it.