I agree that if enmity is to be at all a concern for tanks, there should never be a situation (except by specific fight mechanics) that they are entirely freed from that duty, or that even that more than roughly half of it should be performed by others.
However, I do think that others' contributions, if given at all, should not be so negligible as, say, Second Wind's contribution to survival, nor should it come as a maintenance skill. Rather, it should add a skillgap component that additionally affects how one performs their normal tasks, rather than replacing or interrupting it.
To be more concrete, it should not be
- Insignificant (let's say, less than some 20% of contribution towards optimized enmity management);
- Wholly situational (used specifically and at too significant a cost to warrant in the vast majority of situations, as that would just feel bloated); nor
- Optimal, yet irritating (forcing jobs to interrupt their rotations or integrate into them seemingly unrelated and incohesive skills).
Of course. And yes, GCDs would be included in such (and obviously contributing more over time to damage-mitigation variance than shared-recast time or gauge skill oGCDs, if those are even used on the given job). The difference is that one builds its margin during periods of reduced danger (prepping for mitigation events) and the other around periods of reduced damage potential (prepping for raid buffs). I find the first to be more iconic for tanks, and thus focused on that over simply "make sure to have enough enmity to max out the next trick attack". Strictly speaking, the latter would still apply, but adding the other focus should help, however much, to make tanks feel less like mere "Blue DPS".
Let's not break the notion of the quote from the requirements that immediately followed it, shall we?
At present tanks have an incredibly small performance range relative to other roles. This is due primarily to both having as great of offensive uptime as any other melee, unlike healers, and fewer optimizations to be made (or, less effort required in making those optimizations) than most DPS. Decreasing the gap in complexity by increasing the effort required for optimal tank play would greatly diminish the chances of a tank replacing a dps simply due to being easier to perform with as is currently the case.
But, let's consider as well -- what you're describing has already been a historical norm until Stormblood. A 90th percentile tank could often outperform a 50th percentile DPS in ARR and HW. Was that so horrible? If tanks then had possessed the passive mitigation they carry now, I'd argue so, but, they didn't. So where does the problem really lie, in a good player being able to outperform a poorer player even outside the obviously optimal role, or simply when one role carries too many simultaneous advantages? The latter, I'd have to say.
A good DPS should obviously outperform a good Tank by a large margin (already the case); a fair DPS should likewise outperform a fair tank by a large margin (not so much the case, as the relative gap is greatly diminished). The issue to me is that there is little reward, and in many ways little to reward, in optimal tank play relative to DPS, and that if tanks were allowed more proportionate reward (i.e. damage range) they would be overpowered so long as they keep so much passive indirect contribution.