As I just stated, the purpose of this proposed (and wholly self-engendered) redesign is not to sneak out from the hybrid tax. Neither is it to try and push warriors ahead of other tanks. Rather, the goal is for each spec to have that thematic feel that Blizzard stated they wanted: with Mastery coming and each tree being redesigned to free up options and allow players to pick talents they like rather than talents purely because (as in the case of Cruelty) they're boring but necessary DPS or tank boost talents (Deflection in arms is another example) now's a good time.
The danger, of course, is that we'd veer dangerously close into 'DK lite' territory. To some degree I think protection warriors already feel too much like protection paladins: they both use shields, hit things with their shields, TC and Shockwave are used with Deep Wounds for a poor man's Consecrate, we're getting a Revenge cleave that's similar to their seals, we both debuff attack speed, we strip off armor and they just ignore it with holy damage, both classes use block, parry and dodge (the only two classes that do as tanks) etc etc. Even if you can make the argument that prot warriors were doing a lot of that stuff first, it's fair to say the difference between a prot warrior and paladin are currently centered around a few special abilities each class has that the other wants.
The easiest way to avoid homogenization of feel is to again go back to the idea of talent trees not as role defined, but as themes that define aspects of the class. Warriors are purely martial: they have no magic, no sneaky tricks or stealth, they don't tap into abhorrent powers of the undead or transformative magics or what have you. Warriors are build around three aspects: weapons, rage, and raw force.
Bornakk - Class Q&A Series: Warrior
We also need to make some decisions about the difference between Arms and Fury. Traditionally, Arms was the PvP tree and Fury was the PvE tree. We understand some players prefer that model, but we don't like the way it cuts off such a big chunk of the class from players who might not have much interest in the PvP or PvE parts of the game. However, we would like to reinforce a little more the kits of Arms and Fury.
By making this emphasis (and including protection, or whatever you'd want to call it) in the process, you'd finally move towards a modern warrior class design. Frankly, the warrior feels like what it is, the first viable tanking class designed and implemented. Warrior tanking mechanics (and to some extent druids, who are admittedly designed to mimic warriors) are antiquated now. In the modern WoW
experience, where DPS scales much better than tanks and the great majority of DPS use either mana or energy, systems that start at full and either don't really run out due to MP5 mechanics or natural regeneration, we have a situation where the DPS can go whole hog from the get go while the warrior needs to build up his resource system. (This is also why DK tanking can often be stressful, as they not only have a complicated resource system to manage, but they're also often GCD or rune locked while tanking. This isn't a DK column, but in any post discussing moving warriors to a more open talent system it's a pitfall we need to avoid.) In PvE, it's not like the infinite resource replenishment of rage even matters, since very few other physical DPS ever have to worry about running out of mana and this have effectively 'blue rage' but with the added benefit of a full bar up front.
One possible way to do this would be to move arms and protection away from rage. It doesn't really fit the 'disciplined, soldierly feel' that's been proposed for arms to have them build a resource that's called 'rage' after all. However, the amount of redesign that would be required here would be so great that it starts to move past the purview of an expansion, and in general rage is so much a part of warriors after five years that I don't want to see it gone. However, changing the way rage works
for arms is certainly feasible.
Imagine if, instead of fury's constant cycle of rage in/rage out, arms was based instead around the idea of building and maintaining rage: that the various attacks and abilities of an arms warrior did more damage or gained other benefits related to rage pool. At the same time, each ability would still cost rage, so you couldn't simply sit on a full pool and have your DPS increase. You'd have to acquire and spend rage, but rather than simply following a rotation and burning through rage as fast as possible (dumping if you appear in danger of capping out) you'd want to save rage for big flourishes and finishing moves that would do more damage the higher your rage bar got, but also cost more rage to use when you did, leading you to a tactical use of rage.
Meanwhile, the protection tree would be based around stoking rage like a furnace. A high rage pool would make you inherently more threatening, reduce the amount of damage you took, but still rage would be spent to generate threat and reduce/mitigate incoming damage. Having high rage would increase your outgoing white damage but hoarding would be discouraged by making active tanking require steady rage expenditure. Neither the fury warrior's frenzied spend, spend, spend nor the arms warrior's maneuver and riposte, protection would be about holding rage in until you just can't hold any more and then letting it go in massive, ponderous bursts of pure force, slamming your body into enemies like a rampaging bull and letting your raw mass bowl them over.
In the end, I am the worst seer you're likely to meet. Warriors will be whatever they will be, and I have no more claim on knowing what that is than anyone else. But I firmly believe that the time to redesign rage and the talents is now: warriors aren't a bad class, but they're definitely showing signs of age and older design that need to be reconsidered. The goal is not to make warriors dominant at either of their roles, but rather, to let a warrior perform either of them based not on locking them into choices in a narrow field but rather by giving options useful for either in each toolkit.