Let me be clear here that I don't want to see tanks further homogenized. My issue is simply when "specific" advantages are given so straightforwardly from basic categories that they become simply differences in broad performance for the given type of content.
We saw Warrior and Paladin suffer from this in 1.9 and 2.0:
- In 1.9, Warrior self-healed for so much in dungeons that any disadvantage in heal-efficiency or eHP was moot, since there was nothing Paladin could survive that Warrior could not yet Warrior required far less due to being able to Benediction itself with each Steel Cyclone.
- In 2.0, those self-heals were neutered when the healing from damage dealt was moved from Steel Cyclone to solely Inner Beast and Warrior's damage was reduced in general. That said, it could still be the more efficient tank... so long as someone else could kite mobs between Bloodbaths and one was hugely overgeared. Resultantly, before the 2.1 changes, we were on a path for Warrior to end up weak until overgeared and preferable thereafter -- broad categorical differences that left no competitive tank choice except at mid-high gear progression unless there was an AoE DPS check (since Paladin's only AoE damage at the time was Circle of Scorn, its only spammable AoE a suppressive debuff).
And for the last few expansions, we've see that among Dark Mind*, Dark Missionary, and Heart of Light.
(*Dark Arts could have aided against this but almost never made room for enough healer offensive relative potency through its extra sustain generated to make up for the loss to the DRK's own relative potency otherwise spent augmenting Souleater's damage [and therefore its healing, too, as it healed for damage dealt back then].)
Now, apart from being almost wholly self-scaled vs. scaling with incoming damage,
almost all of these can theoretically be fixed through context, yes. You can ensure that each fight has magic damage at amounts and spacing just enough for Dark Mind, for instance, to
feel useful and for it and DRK as a whole to provide merely even performance. You can give each fight something to Cover with just enough value to make up for its gauge cost. You can give just the right amount of adds at just the right times for Bloodwhetting to compete on the whole even if its per-hit value were tuned to better fit its AoE potential (instead of just scaling like any other AoE in the game). All of that is theoretically doable.
But doing so would also add a burden to content creation and fetter to its available creativity.
That's not to say that content shouldn't give a damn about individual jobs kits or that it should just leave it a happy coincidence for iconic skills to see good use, of course. They should, likely far, far more than they do now, but that's a lot easier to do in varying ways if the skills themselves do not scale so wildly across broad categories (AoE vs ST, magic vs. physical, single hit vs. many, etc.) or there are at least counterweights across the kits that have those more variedly performing abilities. Such allows those interactions to focus more on utilities instead of just broad summative power likely from a given context.
For instance, Bloodwhetting's healing might scale with damage/hits dealt and therefore primarily with one's own stats, but maybe Warrior would take less damage and heal for more through its vampirics as its health decreases, which allows it to leverage higher amounts of incoming damage. And while Warrior could put out additional healing and damage in AoE via Raw Intuition/Bloodwhetting and Vengeance/Damnation, perhaps Paladin could debuff enemies that strike its shield during Shelltron or Bulwark with a slowing DoT, which would likewise scale better with single-target, while Camouflage might offer something through parry scored, etc.
That is, the means can still be different enough for tanks to feel different even if no tank is advantaged noticeably for a whole content type in its typical form or overall across a specific instance. Mixing up content types (having more adds in raids, more and less scripted damage to non-tanks in dungeons, etc., etc.) would still do a ton of good in that regard, but mixing up who's good at what by allowing them similar capacities in highly differing ways both does not homogenize what really matters (the gameplay) and better allows for freedom in design among each existing and potential future content types.