Except...
Older games had the exact same mechanic of enmity only ever goes up.
It didn't "Decay". Only certain boss mechanics would force enmity resets (Which would just lead to you doing your pull rotation again. I.e. In games where it mattered, people would stop DPSing and you'd stack your Sunders. In games where it didn't matter, you just spammed your stuff like normal after maybe a taunt)
Even if it DID decay... It would have no effect other than arbitrarily increasing the amount generated to being "X + Y per second" where X is the amount of enmity per second the highest DPS on the list is and Y is the amount of decay.
This is due to the fact that the reason why Enmity is such a non-factor is because the only people that should be getting hit are Tanks. When there's only 1-2 Tanks in a party, then having complex enmity mechanics that can make other people potentially get targeted makes no sense.
For enmity to actually be an interesting mechanic, it would need to be created from the ground up as a unique mechanic as opposed to an independent modifier to justify Tanks doing less DPS/HPS than DPS/Healers (Since that's really the only reason why enmity mechanics actually exist. Otherwise you could simply make it based on DPS/HPS output since that'd have the same effect and then there would be no binary nonsense because as I mentioned, DPS scales infinitely until you one shot an enemy)
Though, it's hard to design something interesting when it still has to focus around the binary system of "Is attacking the Tank" or "Is not attacking the Tank" due to the way Holy Trinity games function.
Games without the Holy Trinity, tend to be able to have more interesting enmity mechanics because it's less important for enemies to be constantly focusing on a singular "Tank" character.
Though... When "So long as everyone does their bit to manage it" happens to be "Press your MP recovery skill on CD to recover MP... Also dump your enmity" "Press Shirk after you Tank Swap" "Press Diversion alongside your other burst CD's" it does heavily mitigate how much it matters.
If it wasn't, then it wouldn't be so easy for SE to simply delete Diversion and give Tanks an enmity toggle while deleting enmity skills and modifiers (Such as on Onslaught)
So... If I'm understanding this right, you can visualize this as some sort of Limit Break gauge type system. Where each bar is a "Floor" of enmity.
So, Tanks generate enmity and fill up the bars which provide DPS boosts. While DPS fill up the bars and then lose the DPS boosts.
With the idea being that, enmity combos can fill up bars without using Provoke, while Provoke takes you just to the next bar.
So that the idea is that the reason why you wouldn't simply just spam enmity combo's 24/7 is because at some point, maybe, you might hit a point where the bar fills up so slowly that you won't be able to fill it until Provoke is back off CD anyway so you may as well just spam DPS until Provoke is back off CD...
But, somehow, swapping to that DPS combo over the enmity combo won't cause you to generate such a low amount of enmity that you start to lose some of your lead... Which of course wouldn't then mean you just spammed DPS combos 24/7 and Provoked on CD because reasons?
To say nothing about how this system works with Tank swaps? Who's going to get aggro? What happens if both tanks are simply spamming Provoke on CD to generate "Floors"? How will the actual basics of what enmity is supposed to do actually function? Or will this literally be just another completely unrelated mechanic while enmity continues to work as current and be a non-factor?