It's largely a fuction of encounter design.

Rejuvenation is possibly a little overpowered (the mana cost in particular is insane), but the more important factor is that Blizzard is absolutely dead-set on pushing massive constant raid-wide AOE damage into their fights. The weakness of HoTs has always been overhealing; HoTs on someone who isn't taking damage are worthless. Remove that weakness by making everyone in the raid take constant damage, and it's not so surpising to see Resto Druids dominating on raid healing.

If you look at encounters where damage is much more spiky and random (Jarraxus is a great example), you see Resto Druids performing quite poorly on raid heals.

After all this discussion on Haste and stat ratings... It really started me thinking. I have been of the mindset that throughput and regen need to be measured in their own rights, and for a static score system I still believe that to be the case.

If you use a spreadsheet or some other means where the scoring can be dynamically calculated based on the current gear and by an outlined cast composition and time of encounter... Then you can come up with an accurate current weight. You can download my spreadsheet to see what I am talking about. It currently handles primarily Disc Priests.

So in updating this spreadsheet I was working toward a common single comparison system and used the acronym SHC for Sum Healing Capability. What I came up with is that all stats are throughput stats in some way shape or form.

SP/Haste/Crit
Fairly obvious that they are throughput stats and the means by which they provide the gains.

Spi/Mp5/Int
Not so obvious but here is the reasoning. If you operate with the base mana pool what is your capability to heal? Very limited no matter how much SP/Haste/Crit you stack. Through strict regen stats and through the additional mana pool of Int we increase our ability to produce healing and hence throughput. It is quite obvious that if we have mana we can cast and if we do not we can not. It is critical when converting these stats into a throughput equivalent that we understand the mana required, encounter length, and spell composition as these determine what the ideal level of mana regen to support that throughput would be. Anything above the regen required does not contribute to additional throughput, so in terms of throughput regen stats are soft-capped.

In example, Billy Joe the priest is healing an encounter and would like to cast a FH every 2 seconds (super simplified example). The encounter requires that this level of healing be maintained for 60 seconds. What is the regen required to support the throughput of FH at this rate?

FH costs 591 mana casting one every two seconds is 295.5 mana per second and correlates to a balancing point of 1477.5 Mp5. This is widely variable based on mana pool size, so evaluating at varying int / mana pool sizes we find that....
924 Int -> 17443 Mana -> Requires 0 Mp5 to support this level of throughput for 60 seconds and end OOM.
700 Int -> 14083 Mana -> Requires 303.91 Mp5 to support this level of throughput for 60seconds and end OOM.


Summary
Given a specific encounter you can determine the appropriate mana cost to support the spell composition and throughput required to fulfill your role in the healing lineup. Based on that mana cost and the regen stats you currently have, further regen stats may or may not have value. At any point you can say "I would have" casted more spells or more expensive spells given that my mana could support it, you are in a position where your regen may not be high enough.

From a mathematical evaluation of the stats Mp5/Int are the highest throughput stats for Disc priests until you reach the point that mana regen is no longer required at an additional level. The basic comparison would be Mp5>Int>SP>Haste>Crit=Spi, or very close there about... This is very dynamic so don't take that as the law or a static approach.

For Playing with my spreadsheet to look at the results, it is important to set the % on the 'Healing' tab as the composition of spells cast over the course of the fight based on your playstyle. If they sum to 100% then you are assuming chain casting the entire period of the encounter length set on the 'Talents & Misc' tab. Modifying those values changes the 'Mix Mana Cost' on the 'Healing' tab and impacts the 'Longevity' line that is visible on several tabs. I have it configured such that once the longevity is greater than or equal to 100% then regen stats do not contribute to SHC weights as you have reached the point of excess regen and should only be stacking the other throughput stats. One important note is that on the version currently uploaded Penance is assumed to be casted on CD.

I can't express myself in a mathematical note, and somewhat subjective. But this (The Doctor's post) is why I've always walked around with several trinkets in my bag. Unlike most items which tend to be well-rounded, Trinkets can support a dramatic shift in one direction (they can also be fairly rounded too).

For example, the Dragon Soul can be shifted in for a quick 200 spellpower or Spark of Hope for more regen. This allows me to shift gears on a per-encounter basis without having to carry too much gear. (Although, to be honest I did ended up having a special gear-set for M'uru, god I hope to never go back to that).

It's been pointed out somewhere earlier, but consumables also work in the same manner, it's easy to shift gears from Distilled Wisdom to Frost Wyrm.

Many of the trinkets are not fully working so it may be difficult to play with just the trinket comparisons at this time. I hope to improve that soon. The Spirit portion of Spark of Hope is working but the mana reduction portion is not. If anyone finds problems in the spreadsheet please let me know so I can update it, it is very much a work in-progress.

Trinkets and Consumables(Food, Elixir/Flasks, and Glyphs) are definitely the quickest and easiest way to change your focus.

See my armory page, I have been testing int stacking, and it has been competitive thus far in current Ulduar/ToC content.

Been enjoying reading the discussion here on the relative value of Haste and its changing value in relationship to other stats as those stats increase. I think the top things I take from it all is that the values of our different stats in fact change relative to one another as our gear improves, and that at differing gear progression levels, what should be considered our most important stats to focus on changes.

I would have to brush up on some of my old college math and possibly even learn a bit more to really get into the math discussions, but I can understand the arguments and generally the math sufficiently to take what has been said here and pair it with my own extensive experience as a very active raiding holy priest in the game to arrive at some simplified conclusions I will try to set forth:

Early on as a holy priest in raiding, I think it is most important you work on your mana regen, because unless you have that fuel in sufficient quantity, it matters not how powerful the other stats are for the most part, since without sufficient fuel, your spells do not get cast, and those stats are mostly benched. Thus at the early levels mana and regen seem to have the highest relative value.

Once you get sufficient mana and regen that you can cast at least as much as needed through the end of the encounters you are attempting, past that point mana and regen are relatively useless in additional quantities. Then it makes sense that you would turn your focus to other stats.

I feel that SP is the next stat to focus on because it makes all my spells get a bigger bang for the mana cost, and thus improves my throughput the most, while at the same time in effect further increasing the mileage I get for the same mana. At this point Haste would have more relative value than additional mana and mana regen, but less relatively than SP.

But there seems to come a point at which additional SP translates into less effective healing in actual game conditions and more overhealing, and at some point that in effect diminishes the value of SP relative to Haste.

So now I am at the point I have all the mana and regen I need to play full out to the end of the encounters, and I have so much SP that the effective healing I am getting from adding more is diminishing, and it is at that point where if I add additional Haste, I get the biggest and best return for adding it, relative to adding additional other stats.

Now how to quantify those tipping points for mana and mana regen, and then for SP, is beyond my current math skills without some real refresher math study and perhaps further math learning, but that is not really necessary I think for my point in this post, which is to set out in fairly simple terms some general guidance for developing holy priests, which is as I have stated:

1. First get your mana and regen to that tipping point....THEN
2. Get your SP to the next tipping point.....THEN
3. Go for Haste.

Hope this contributes to the excellent discussion already here, if only making it easier to understand these general principles, and to expose them to constructive critique.

I can definitely buy the gearing strategy :
1. First, go for regen.
2. When you don't need more mana, go for throughput.

But I find it difficult to say that for throughput, you need to focus first on spellpower, then on haste.
You basically need both at the same time. Relative weights might shift, but there is no clear turning point.

Throughput is increased mostly by spellpower and haste. The interest of spellpower (resp. haste) grows as your haste (resp. spellpower) grows.
More than that, you're throughput is roughly affine in both stats. It might be the case that at very low level of gear, one stat is clearly superior to the other, but after some values, the optimum will always be to keep some ratio between spellpower and haste (or, in other words, to increase your haste rating by x when you increases you increases your spellpower by 1, with x >0 ).

Now, the only way to change this is to use granularity argument (ie. to say that there is a specific value where the interest of one stat decreases dramatically). That's what you do when you're saying that there is "a point at which additional SP translates into less effective healing in actual game conditions and more overhealing". That could be meaningfull for one precise fight, with one precise healing team and assignment set. However, the damage pattern is not the same for different boss fights, and the healing requirement is not equivalent. You don't need to heal your target for the same amount on, let's say, Razorscale, XT or Mimiron. You also don't have the same time before your heal is needed.

If you want to gear for a very specific encounter, then you can say that spellpower is to focus on till you reach a certain threshold, because at that threshold, your flashheal (or whatever spell) is going to heal for X HP, which correspond to the boss ability. But that's not true for all fights.

There is a clear threshold : you don't want to heal for more than the player total HP, and a 15k flashheal would overheal in lots of situation. However, we're nowhere close to the situation where our "small" heals are close to "too big". I don't see any reason not to maximize throughput at all time.

Not using haste as Holy is a mistake. For Disc, it highly depends on what you're doing, and what your goal is. If you're primarily a raid healer using lots of PWS, then haste is mostly wasted past a certain point because of Borrowed Time procs. If you're primarily tank healing, then haste can be useful, as you won't be using a PWS more than every ~ 8 seconds (on the tank -> on yourself -> on the tank).

See the last 5 pages if you want more details on haste vs crit vs regen.

I can't comment on holy priests and the relative value of spell power and haste to them, but for discipline priests I'm pretty sure the tipping point between spell power and haste never comes and spell power is always more valuable than haste. With Borrowed Time, Imp. Retribution or Moonkin aura, Wrath of Air Totem and Enlightenment you only need 4.67% haste (153 rating) to hit a 1 second GCD. That means every time you cast two shields in a row the second shield gains no benefit from haste beyond a point that we probably all have whether we like it or not.

I have 42% of my binding heals, 21.5% of my Flash Heals, 52% of my Power Word: Shields and 44% of my Prayers of Mending cast under Borrowed Time. This will vary by play style (I play 10's) but it amounts to about 26% of the time where my haste rating is having no effect at all. That's not to say haste is useless, but even when I turn up my spell power to 4k I still gain more benefit from 2.3 spell power than I do from 2 haste. Since, as you said, we are nowhere near the point of our small heals being "too big" spell power wins. At my current gear level haste is only about 90% as effective as spell power for increasing throughput, completely ignoring the fact that spell power increases throughput in a mana-free way and haste does it at the cost of additional mana.

Spell power has even better returns for holy priests than discipline because they have far more coefficient multiplying talents. In other word, each point of spell power will increase the healing of a holy spell by proportionately more than a discipline priest. Both get 5% healing on many spells from Twin Disciplines. Holy gets 10% on all spells from Spiritual Healing, sometimes 12% from Test of Faith, 10% for many spells from Divine Providence, and if you specced it, 3% on everything from Blessed Resilience. Adding these factors gives a bonus of between 13% and 30% on your healing, which scales the effect of your spell power. Multiplying the factors suggests the bonus is between 13.3% and 46.6%.

In contrast, Discipline gets 4% from Focused Power, a minor boost from Improved Flash Heal, and sometimes 9% from Grace which doesn't even affect the shield. Including the 5% from Twin Disciplines, this is a maximum of 18% if the bonuses add and 19% if they don't.

So if your data says Spell Power is always better than Haste for discipline at obtainable gear levels, we can conclude the same is true for holy.

To expand on this and the previous calculations, the conclusion for holy is this. At higher gearing levels, haste is better than crit. It's still worse than spell power. It just means you should use [Reckless Ametrine] rather than [Potent Ametrine] in yellow sockets where you want the socket bonus. Red sockets get [Runed Cardinal Ruby], as do any sockets where the set bonus is worthless (eg. int or spirit).

On the subject of Borrowed Time, the haste soft cap is 50% after all modifiers. This reduces a 1.5 global cooldown to the 1 second minimum, so getting more haste won't reduce the GCD any more. You get 25% from Borrowed Time, 8% from raid buffs, and 6% from Enlightenment. That's a total of 39% haste, meaning that you can't use more than 11% additional haste when spamming shields. You need 361 haste rating to hit 11%, I believe, so that's the soft cap. Note that it's very easy to hit this as discipline even without gemming for it. This combined with Discipline's atrocious returns from spell crit make it so that stacking [Runed Cardinal Ruby] in nearly every socket is generally the right choice.

I don't know if I would say that spell crit is atrocious for discipline priests. Granted, it is nowhere near as good as Spell Power for increasing throughput and it is almost worthless for PWS, but it's benefit through Divine Aegis is nice while tank healing.

In general, the value of haste and crit as a Discipline priest depends greatly on what role you usually play. If you do mostly PWS spam then the value of crit is pretty low (and haste even lower), but if you do mostly tank healing then crit and haste are still quite valuable.