In my case, I didn't find 4E to be coherent at all. The mechanics seemed (in my opinion) to be very much at odds with the picture that the fluff and background information was trying to paint in both the preview books and the early 4E books.
Well, I tried to word my post to make clear that I was talking about rules consistency and coherence *as a set of rules*. What you are talking about here is coherency between the rule set and the game world, which I agree is a different thing (and I don't think that this was well done in early 4E, FWIW).
Still, if we're talking strictly rules and nothing else, I would agree that 4E is a coherently designed game.
Yes, internal coherency in the rules is exactly what I'm talking about. I think it's important because, for me, if the rules themselves aren't coherent I don't even get as far as worrying about the coherency they have with the game world. It's like someone saying "here is a world where the physics as defined don't work out"; the fact that the people in the world seem to be assuming a different set of physics again is secondary, to me - I get stuck at the "this world's physics don't work" stage!
For me, then, internal rules consistency comes first. That is NOT to say that it's the only important thing, but its what is required as a precondition for the rest to be worth tackling. And I find the rest (fitting the world to the rules effects, fitting the characters' expectations to the world "realities" and so on) to be relatively easy to fix once I have a coherent rule set.
Now, I am someone who enjoys 4E. As I've said other places, I've learned the strengths and weaknesses of the game, and I've learned what it's meant for and what it isn't meant for. I've also added other rpgs to my library, so, when I want something outside of 4E's scope, I have games available to me to have fun with instead of getting frustrated trying to make 4E into something it isn't.
I think I maybe had a head start in this area, because I moved from running D&D to other RPGs around 1980 or so because of the trouble I was having with rules incoherency and D&D not doing what I wanted out of an RPG at that time*. As a result I tried many systems and started to discover that systems are often better when they set out to do one thing well; find that one thing and you can generally enjoy the system. Try to bend the system for what it's not designed for and it often falls short. What it seemed to me many D&D players were doing was making up their own systems - from rulings and house rules - that fitted what they wanted to do. The downside with that is that it is much harder for the players (as opposed to the GM) to learn the system when it's not written down. As an aside, I think this is one reason many players dislike the idea of learning new rules sets - they have the expectation that reading the rules will not be what is required to learn the rules and that, in a sense, reading the rules is only a (partially misleading) preliminary to "really learning the rules" that will come from "learning the DM". Rebellion against this same "feature" might factor into the phenomenon of "rules lawyers", as well.
I do find that I enjoy the more balanced approach to 4E classes; even the worst ones are close enough to the same general ballpark to still possibly be fun. In contrast, I'm not sure if I could ever play a non-caster in 3E again -even if I wanted to.
I'm not someone who is heavily tied to balance, but I think at least some balance is necessary. I don't mind playing a character who isn't the best or most optimized, but I shouldn't constantly feel like I'm tricked by a game into playing DC's Robin or Marvel's Maggot while everyone else is getting to play the part of Superman, Hulk, and Thor. There are times when 3E (and even Pathfinder) can feel that way. A lot of options are presented as being valid options, but, in truth, they really aren't... at least not beyond a certain level.
I don't want to risk turning this into another long and rambling thread full of tired old (non-)arguments and
ad hominem attacks like a certain other hereabouts, but I agree 100% with all of this.
*: I didn't see the issues clearly then, I should say. I thought in terms of "realism" being needed and looked for what, in retrospect, I recognise as coherency in game world physics through making them the same as real world physics. When I eventually realised that sufficient understanding of the real world to fully model it was functionally impossible, I moved on to realise that what was important was working, coherent game world "physics", not a rule set that closely modelled the real world. If I talk of "my philosophy" for roleplaying games I think it gives the impression that I have a static, unchanging "philosophy"; I don't, but what I have now has evolved slowly over a great many years (and is almost certainly evolving still).