Why don't we know that? For calculating purposes, we should think of absorb effects as a HP- buff. A player with 20k HP carrying a PW:S which absorbs 5k damage has an effecive HP- pool of 25k. Furthermore, an absorb- effect can't overheal, there is just a possibility that it will expire if there is no damage incoming and therefore becomes useless. Nevertheless, I fully slashsign your comment about overwriting an existing DA.

Thinking about tankhealing, we can just take as granted that we will always have Inspiration and Grace up and have no control over DA or even an overwriting of our own DA. Grace will never run out if you just cast Penance on CD on the tank. And if we now assume that the tank will always have Weakened Soul which provies us with 4% more crit, 29% crit would be enough to have Inspiration always up. Not mathematically backed up, I would then start to lean my attention on aquiring more haste to speed up all my casts and therefore ( never underestimate this ) to be able to heal more while being able to cast for a certain time before being forced to move again. I know, this is not a real min/maxing- argument, but I think that it's got its own value, not knowing the content that awaits us in about 1 or 2 months.

Concerning the discussion of crit versus haste, I think it should be looked from a different viewpoint then this thread suggests. Main concern as I see it so far is not the fact that discipline healing can provide very high HPS only for very limited time and needs a boost by either haste or crit, but the fact that huge number of healing ends up as overheal.

It is basically the same situation as paladins are facing and it is result of direct character of discipline heals, which are overwriten often with quicker aoe heals or with quicker direct heals.

Best tactics to avoid this as I see it, can not be based on haste, as there is not enough haste gear to stack it so high to be quicker then anybody in the raid. Crit can help a bit due the fact it creates shields and shields prevent damage without overhealing, but the inherent problem here is the fact that divine aegis does not last long enough.

So in my opinion, the ultimate answer to the question Crit versus haste is stacking Spell Power and letting Crit and haste to personal preferences.

On a more general theme of discipline healing:

Best tactics for discipline healing in "normal" situation is using shields as often as CD allows it on targets that are almost sure to be damaged . That should provide enough mana for any other action and prevent substantial damage. And that extra action should be usage of cheap, low healing spells on maximum number of targets. I think improved Renew is one of them as it is possible to keep at least three ticking and Flash Heal cast on targets below 50% health is the second one. Third one is PoM. It ends up in 30-40% overhealing anyway, but it is better then 50% that I see quite often during runs to all current instances.
(There are of course different fights, like Sarthorion with 3 drakes, where bursting PW:S, Penance and GH during Flame Breath is priority No. 1 and the rest is just filler, but that is exception)

Discipline build is utility and special purpose build as I see it, it brings to raids spirit buff, unique shields and capability of burst healing. Those three capabilities are good enough reasons for having discipline priest in raid. His/her total effective healing, combined with damage prevented by shields, provides roughly same amount of damage mitigation as any other healer hence having discipline priest does not mean nerfing of raids healing.
Having more then one discipline priest in raid is I would say contraproductive. There is no synergy of discipline buffs, two priests casting shields will probably end up in casting them on same targets and there is hardly any need for double burst healing.

Generally speaking extreme overhealing is bad, though it doesn't matter as much if you can sustain a high level of overhealing without having mana issues. If Blizz makes mana more of a concern this will matter a great deal more. Though I don't devalue Crit at all based on overheal impact.

What I do have concern in my modelling and math is creating an over-value of crit. Which is highly possible. You say that your DA has a "VERY high likelihood of being eaten" but how likely is it really? I would say that the likelihood of a DA getting used on a tank before expiring is pretty much guaranteed. The real concern is overwriting full shields or partially consumed shields.

Looking at Penance with a Flash before and a Flash after... Penance as a 1.5 second window damage will be taken once and in a 3.9 second window of the total cast sequence twice, there are scenarios where damage will be taken twice / thrice respectively however I didn't model those as extra damage taken between crits does not change the result. As was noted previously damage patterns vary per encounter and there is no ultimate general model. This is for showing that there is reason to devalue crit to some degree.

@30% crit..
During the ~1.5 seconds of Penance
18.9% of the time 2 crits - 1 likely overwrite
2.7% of the time 3 crits - 1 guaranteed overwrite and a second likely overwrite
During the ~3.9 seconds of Penance + Pre & Post Flash
0 crit - 16.8%
1 crit - 36%
2 crit - 30.9%
3 crit - 13.2%
4 crit - 2.8%
5 crit - 0.25%

47.2% of the time you are at risk of DA overwriting, 16.25% of the time DA is at extreme risk - possibly overwriting more than once.

@35%
During the ~1.5 seconds of Penance
23.89% of the time 2 crits - 1 likely overwrite
4.29% of the time 3 crits - 1 guaranteed overwrite and a second likely overwrite
During the ~3.9 seconds of Penance + Pre & Post Flash
0 crit - 11.6%
1 crit - 31.2%
2 crit - 33.6%
3 crit - 18.1%
4 crit - 4.9%
5 crit - 0.5%

57.2% of the time you are at risk of DA overwriting, 23.5% of the time DA is at extreme risk - possibly overwriting more than once.

@40%
During the ~1.5 seconds of Penance
28.8% of the time 2 crits - 1 likely overwrite
6.4% of the time 3 crits - 1 guaranteed overwrite and a second likely overwrite
During the ~3.9 seconds of Penance + Pre & Post Flash
0 crit - 7.8%
1 crit - 25.9%
2 crit - 34.6%
3 crit - 23%
4 crit - 7.7%
5 crit - 1%

66.3% of the time you are at risk of DA overwriting, 31.7% of the time DA is at extreme risk - possibly overwriting more than once.

From this 1% crit makes overwriting DA 2% more likely in general and 1.45% of the time almost a guarantee that it happens at least once. It does reduce the chance of not getting a DA obviously which is good... Though this is why if you consider crit solely on the benefit it isn't being considered properly.

Haste
The only negative impact is when you reach 1 second GCD's - statically not possible in current gear and when fast cast spells Flash Heal reach sub 1 second casts (possible with BT). Haste currently has no point without considering hero/PI/BT/trinkets where the threshold is sustainably reached. I float the line a lot due to BT/Trinket but at those times I don't use Flash heal so I am still gaining benefit from the haste.

The above model for crit is based on 25% haste... It should be noted that adding haste compounds the DA overwrite situation. At 50% static haste what I modeled at 3.9 seconds only takes 3.33 seconds and reduces the number of DA's that could be consumed in the window of applications.


Just some further thoughts on the topic.

There is 2 missing points in your post.
First is efficiency. Flash heals triggers rapture, renew not. So the real mana cost of flash heal is heavily decreased.

2nd is the benefit of flash heal :
it can crit, and disc crits heal for about twice the amount (including DA). So that's easily 20% more.
It also gives grace and (when crits) inspiration, which are really good buffs for tanks.

This said, on Patchwerk, I do use renew, to pre-heal when I have penance available and that the tank is full. I don't need to precast (I'll use penance as soon as I see damage), and shields and renew (and POM) helps to mitigate the next strike.

Renew is not absolutely bad. It's not interesting to heal a damaged player for disc priest when standing.
It's still the only cd-free heal we have for preemptive healing and instant heal for movement (or vortex).