Granted, the challenge is not extreme.
I just always have difficulties to provide some numbers that are true in specific context, and that people will always generalize as a well-known fact latter. I've seen senior researchers doing this, I've seen good Wow player doing this, and I don't have any hope that the majority will be careful when dealing with numbers ;-) That's why I prefer to emphasize that healing is not hps before giving numbers.
Please note that wrath of air totems provides also a static 5% boost to haste. And as far as I am concerned, I always have one or 2 shamans in heroic raids (and almost always 1 in 10s), whereas I can't count for sure on elemental shaman or boomkin.
But what you wrote is mostly true. Haste will be better for a throughput stat once you have infinite mana, and has diminushing returns.
There is just a few points in favor of crit I'd like to re-asset, just in case :
1) Inspiration is a huge bonus to crit. You don't need tons of crit to have it nearly full-time, but you need some crit. That part of crit, you don't want to miss.
2) Divine aegis is also a good bonus, compared to TBC case. Basically, it is close to meaning that crit heals heal for twice the amount of normal heals (whereas it was only 50% more during TBC). It would be better if Divine Aegis wouldn't replace itself, but that's still one good bonus for crits. Arguably, shielding is also a interesting feature (both for damage prevention and overhealing reduction).
3) Haste has also some limits, due to cooldowns for penance / PW:S / POM. Therefore, 1% more haste is less than 1% more heal. Therefore, it behaves comparably to crit as far as scaling is concerned (see later for detailed point).
4) Haste has a soft cap, which is the 1s gcd under bloodlust / power_infusion / borrowed time. That soft cap is not really far.
Detailed point on scaling :
Ignoring overhealing and overwriting of DA, each point in crit (all other things being equal) grants the same amount of hps : 1% more crit gives 1% (considering crit heals, including DA, heals for 200%, for simplicity) more of your non-crit healing. For that reason, depending on terminology, one can say its return is constant (in absolute value) or diminishing (in relative value : going for 0 to 1% crit is relatively better that for 99% to
Haste divide casting time (and gcd). If one has x% of haste, the casting time is divided by (1 + x/100). In other words, the number of heals you throw in the same time-frame is multiplied by (1+x/100).
Now, disc best heals are on cd. This means that even if they do be hasted, you can't throw more penance or PW:S. It just leave more room for flash or greater heal.
If you have 100% haste, the same heals you cast without haste take half of the time. During the other half, you can now cast fh / gh. They are hasted, which means that you can throw during that half of the time the same number of heals you could throw unhasted during the full time). Basically, you've gain the hps of fh/gh unhasted. That's nice, but that's not doubling you hps either.
More, if you consider x% of haste only, time to cast the same rotation takes you (1/(1+x/100) ) of the normal time. It leaves you x/100 / (1+x/100) ) for new heals. They are hasted by x%, so in that time frame, you can cast x/100 new heals. You've still gained x/100 of the hps of unhasted fh/gh.
Once again, the return is either constant (in absolute value) or diminushing (in relative value).
[B}Conclusion : [/B]
Blizzard made sure that nearly all the stats have a constant absolute return. That's true for haste / crit / AP / spellpower and hit. The one case that was not constant in TBC was ArP and maybe spell penetration, which I've never been interested in (I don't know whether ArP has been "linearized" in 3.0 or not) . It was a common mistake during TBC for (french) hunts to consider that AP was "bad" because of static absolute return (the famous unmodified 14AP = 1dps), when crit / haste/ hit had relative returns and were not interesting. At least in EJ, I'd would be glad we don't make the same mistake in scaling.
NB : I'm not native english speaker. Please be indulgent for my poor english.
Firstly, thank you for the clarification on static haste buffs, an oversight on my part. I have bolded the main point I want to discuss here.
Let me first give a situation that will hopefully conclude playstyle B as superior as well as a greater value given to haste assuming playstyle B. Both assume you are healing MT and spotting raid or OT
Playstyle A: Your WWS may have Penance as the number one heal and you like to use this on CD, keeping all the nice buffs disc has to offer up on the tank, DA,grace,inspiration etc.
Playstyle B: you mostly flash heal the MT and reserve penance for emergencies. Your WWS may have flash heal as 1 and PoM as 2 with penance as 3. What playstyle B allows is for a penance when a dps pulls aggro or when a boss soft enrages or a sarth deep breath. Basically reserving penance as you would pain suppression but without the long CD. You should never really have a situation where 2 players need penance at the same time(assuming raid is functioning as intended) and due to this the value of haste is supreme as you want a fast flash heal. The worries of penance being on CD is negated. Also while flash heal is not as effective as penance for keeping inspiration at el on the tank it still suffices enough to warrant playstyle B in my opinion.