(Note: this is a long-standing idea of mine, but the latest dev blog about feat taxes sparked this actual writing.)
I will write a more full-fledged blog post (in my "Class Action" series) to go into further detail on some of the classes, but I also wanted to write about the problem in general as briefly as I can, to the lurking developers and the fellow playtesters that read in here.
I see basically only one problem in Next design guidelines and it's summarized in one motto I've seen used: "Fighters should be the best at fighting". It's a motto that assumes that "Wizards should be the best at casting spells" and other similar mottos are also design decisions. This is so wrong and it's not just my opinion, it's logic: there are many classes that fight, many classes that cast spells. And if they're not called Fighter or Wizard, their players are going to have less fun, if these mottos become mechanics. And no, I'm not saying classes should have a combat role like in 4e. Because in 4e, the combat role enforced a tactic. Tactics should be things that can vary from round to round. Defending, Leading, Controlling, and Striking should be things that every single character COULD be able to do (in different rounds), if built to do so. But this is another matter. One may also want his character to be be good at a single tactic, that's fine. I'm just saying that classes shouldn't enforce a tactic like in 4e (good idea to make themes handle those things in a way). BUT shouldn't also steer in the opposite direction and extreme of "fighters are the best at fighting". For the above reason. Because it's a direction that will spark many wars and more importantly, lessen the fun at the tables.
What's my solution to this? Simple: each class should be able to be very accurate and effective (with effective meaning damaging OR NOT) when following a class-driven STRATEGY. Not tactic. I'm talking of something much broader than 4e role, and much narrower than current Next's philosophy of "being best at weapons". The problem is finding a good, defining strategy for each class. Something I see very akin to defining the class' true SOUL. And I think the soul of certain classes and especially the basic ones, hasn't been understood yet.
As I told you, I'm ok with "Defender" and "Striker" being things of the past. they're to gamist, and they do not belong to class: themes make them available to anybody who loves that style of play and that's good. So what would be the Fighter's specialty if it's not to be too broad as "weapons", and too narrow as "defender/striker"? For me it's simple: Fighters are the best in the front-line. War-like situations. When outnumbered. I'm for making classes more situational, yes. But also able to create the situations they're good at. Wizards shouldn't just be "best with spells" (cause where would that leave Warlocks? Sorcerers? Shamans? Whatever?), or "controllers" (anybody should be able to control, that's a player's choice). They are the "engineers" of D&D, the guys that deal with complex stuff. So in battle, they should be best when dealing with complex situations, they change the battlefield, they morph things around them. That's their soul. Their strategy.
The fixed bonuses, the +1s and +2s that as of now the Fighter only gets with weapons, should be accessible to it and all the other weapon-using classes, but in different situations. Fighters have them when in the thick of melee, Rogues when fighting dirty, Rangers when taking advantage of terrain, Paladins when they face the strongest enemy maybe, or when sacrificing themselves, or whatever their creed says. Wizards should be good with complex spells, Sorcerers with simple spells, Warlocks should be good with darker things, things that isolate them from their very allies (I picture warlock magic as tainting, forcing allies to leave them alone). Spellcasters are a different problem actually, because they need balancing on their spell effects and resources usage, and you can read many other blog posts I wrote about this. With weapon users, the "situational math solution" I'm suggesting is much more straightforward and it really just requires developers to understand these never well-understood SOULS of each class.
And to represent better the classic specialties, such as "fighters are good with weapons", instead of the math bumps, add more possibilities. The Paladin and the Fighter may be just as accurate and strong with their swords in their relative "specialty situation". But the fighters knows how to do MORE things with that sword. Maneuvers maybe, whatever. And more importantly, the Fighter may know how to exploit the strengths of every weapon. Take the 4e new Expertise feats. What I'm saying is give the accuracy bonus to everybody (situationally), but the Fighter may be the one to get automatically the secondary benefit associated with that weapon expertise. A fighter with a polearm may be better at stopping enemies in their tracks with a spear better at charges and resisting charges maybe and so on. THIS is something that would make the Fighter "the best with weapons", without making other melee classes mathematically worse. Also, other classes could also specialize in weapons, but would never be specialized in ALL of them as the Fighter may be. Other classes of course would have other benefits that the Fighter will never have, and here you have balance, not through numbers, but through different possibilities, and keeping tradition.
Here it is. Lengthy but this in my opinion a solution for the old problem of balance VS simulationism/tradition.