As an aside to what you posted, RDM (and SMN, to a lesser extent... and even healers) have this strange property that they scale with encounter design. That is, fights centered around body checks make the rez irrelevant. Fights not centered around body checks and with lax dps checks make the rez extremely powerful. I've posted on another thread that my personal thoughts on the matter is that while caster balance before DT isn't ideal, it's not terrible either, and I think the rez should stay since it gives some semblance of a distinction between casters.
For the remainder of my discussion, I'm excluding SMN, because SMN isn't really a caster right now. Further, I have not leveled and played RDM yet, so I cannot say anything conclusively. But, from what you're saying, it feels to me all three casters have the idiosyncrasy of being able to sacrifice burst potential for uptime. BLM has it the worst out of the three (needing to consume its nuke and most powerful dps steroid, Triple), PCT has it the best out of the three. I don't want to go into PCT optimizations much, but beyond the obvious cost of delaying your next CYM combo, Holy in White also interferes with optimal PCT burst (in which you start mid CYM combo on Y). These are small losses, but you do consider them if you try to push the job to 100%. This seems similar to the argument that RDM's mobility has some cost in burst (although I did play RDM in EW, I progged Dragonsong with it and did some Abyssos with it too, so if it's anything like that, you have some liberty on where you put your "filler" melee combo between burst phases). This is so ubiquitous to the caster role it almost feels like their version of "positionals"- a unique aspect of their optimization (the ability to trade some amount of burst for uptime).
I think currently PCT has adequate mobility for the fight design they're trying to push. I think it's very much comparable to EW BLM (although EW BLM was harder to setup and execute in practice, it's a similar feeling in terms of shifting the pieces of your rotation around to place the mobility parts where you need them while minimizing any losses of uptime or losses of damage in your burst window). I don't know how RDM compares, but if it suffers in comparison, I think studying what makes PCT work and apply it to RDM might be warranted (just like I think Smudge should be the blueprint for all caster dashes). Similarly, I feel that, ideally, we'd revert BLM back to its EW state (which plays somewhat comparably to PCT).
My entire tirade loops back to the original point- I think the last dps should be potentially a melee, caster or physical ranged with sub 5% rdps difference between each comp. PCT is the only caster and ranged job that can match the melees on some metrics in some percentiles. The only way to "break" melee supremacy is by allowing other roles to come within a small margin of their damage potential. If PCT is overtuned, it's by less than 1%, which is a different picture than what some people paint when they say it's "massively overtuned"- and the logs seem to back this assessment up, as far as I can see. I concede your point that fights should be designed assuming a melee might be in that 4th dps slot, much like they should also assume a caster might fill it. If the design is such that the 4th dps slot incurs a small loss as a melee and/or caster, I might live with it. I'm slightly bothered by the fact double melee has been the optimal choice for 5 years after playing a meta where we had 3 best comps depending on fight.