Missed that part, sorry! 2 makes more sense with ranged holy as an addition but honestly I still prefer thundercloud procs over such a simple rotation. As for ranged holy that can still be cast on yourself without a target itself, I could take it or leave it. It has it's merits and I can definitely appreciate that hunkering down under the mob stack isn't for everyone. Even with it, I'd still be personally attached at the hip to the tank in dungeon runs though.
Regarding 3, the idea would not be to refresh them after they detonate, but the main focus here is that it just takes the dot damage you've built up as the pack is being positioned (You dot on the run right? If not, you should be), and front loads it all but guaranteeing you a Misery to really get a nice satisfying chunk DPS in a similar way to how DNC front loads so much damage with Technical Step > Standard Step.
I'm not really sure how it makes the optimal rotation impossible to calculate though? The math is really no different to existing multi dotting and is simple to calculate with a timeline.
Lets go with an 6 mob pull. We've detonated our dia bomb and we think the mobs are going to survive long enough to reapply Dia and go for another detonation. The right value is our total potency up to that point refreshing dia on 6 mobs and then holy spamming for a detonate, the middle is our potency just spamming holy, the left is our timeline in seconds. To keep things simple we're going to ignore crits and spell speed to get a feel for how it's going to shape up. As ever, the caveat applies that this is quick off the top of my head math, if there's a fundamental mistake in the sums by all means point it out, but I think it's accurate enough to drive the point home.
2.5 - 900 - 60 - First mob dotted
3.0 - 1800 - 120
5.0 - 2700 - 180 - Second mob dotted
6.0 - 3600 - 300
7.5 - 4500 - 360 - 3rd
9.0 - 5400 - 540
10.0 - 6300 - 600 - 4th
12.0 - 7200 - 840
12.5 - 8100 - 900 - 5th
15.0 - 9000 - 1260 - All 6 are dotted now, the dot tick aligns with the GCD here
17.5 - 9900 - 1410 - If we detonate now, our shortest dot has 5 ticks remaining for 300 potency, the longest has it's full 10 ticks remaining for 600 potency. With 6 dots I make that a total of 2700 potency.
So even if we detonate here, we're miles off the mark multi dotting at this stage vs just mashing Holy. It's not worth going more in-depth because nothing is going to make up that kind of deficit. Really the only optimisation question would be weighing up popping a Lily if you needed to move out of an AoE vs throwing a Dia, but if you're optimising to that level you're either ignoring the AoE outright and just eating it, or using Swiftcast to Holy as you move.
I get that you don't like the idea and this probably isn't going to change a thing in your eyes. But hopefully it clears up the misconception if nothing else.
*Edit* On RNG, I actually disagree, with caveats though
Negative RNG variance in abilities sucks. Going for a button that you expect to use only for it to not be ready/available is a bad time. 4.0 Lilies, Selene haste, AST Lady and the AST haste card are but a few examples of where this has arisen in places.
However, positive RNG where you get more than you were expecting can be great. Go back a few tiers when bosses hit tanks harder and crit heals on said tanks were great, frequently freeing up an entire GCD for more dps. That's why I really want something like Thundercloud for one of the healers, It's literally a sparkling 660 potency proc with an insta cast. You get a juicy number and it breaks up the monotony of DPSing a little by adding in just a touch of RNG with no downsides.
RNG on healers doesn't have to be a bad thing.