Buff Broker 

PRETTY hopeful that 1.6.14B fixes some of the bugs with re-buffing the wrong buff, and suggesting the same buff be re-applied too often.

Please report any errors with 1.6.14B; if I get no word back, this will probably turn into 1.7.0

Tried the latest version in the 25-man tonight; was pretty disappointed. Looks like cross-class buffs could use some work (ie: paladin module picking up on shaman totems...still).

I'll look into this & more

As an aside, I don't want to come off as an insensitive ass who's justifying terrible decisions and ignoring valuable feedback.

I'm putting 99% of the responses & comments into a list of things to fix, or alleviate. Even if you get a bullshit defensive response. Especially if you get a bullshit response.

Understandable; that's a common sentiment from melee DPS, I find, who are more prone to wtf AOE dead scenarios. Kings vs nothing or might vs nothing is a big deal; the difference between Kings or Might is noticeable, but justifiable.

I'm planning a new two-tier role system which will help alleviate this, as well as allow me to hook into the LFG assignments. It would give players a finer granularity than just "DPS": "Prudent" dps who want kings, for example, or "min/max" dps who want might. I'm still mulling over how to expose that in the UI.

I'm still sticking to my guns that might > kings in every conceivable scenario where you can only have one, but that's from a pure dps perspective. gaining marginal stamina for a marginal dps loss is pretty hard to quantify, or at least, garner consensus on.

Sanctuary > Kings > Might is only suggested for "Tanks"...and you need Righteous fury active to be considered a Tank. Try activating RF before questing, and see if that changes anything.

Conversation sake, the AP from might is a greater DPS contribution than the stats from Kings or Sanctuary, even accounting for scaling block value on shield of the righteous. Despite this, I use Sanctuary myself when questing. You can pick and choose: activate RF for tankish suggestions, or leave it off for max-dps suggestions.

This is on the feature request list. It's something I tried to hack in back around 1.3, and never got around to finishing. Scrolls/Drums being acknowledged is pretty high up there, but actually using them is a bit lower. Especially since you want to use them in, say, raids, but not while doing stuff in the-world/arenas/battlegrounds/instances. Maybe heroics; picking where, via config, is probably required before a "scroll of fort kk thx" would be useful checkbox.

Also, while I agree casting scrolls would be cool (pending time)... it's probably not as cool as weapon buffs for shamans/rogues/warlocks, or fixes to the "greater buff is active" bug priests are suffering, or per-character-spec role selection, or zone-detection... well maybe cooler than zone-detection. Still, I'm doing my best to pick through the request list with what time I have

Hunter pets should share blessings with Warriors, not Hunters. Buffing pets should not affect the assignment on hunters. That could have changed in the latest patch; I'll attempt to verify.

Thanks for the feedback, seebs. By default, group buffs are only suggested when 90% of targets are nearby, and you have at least 2 people to cast it on. You can tweak these numbers in the settings.

It was probably suggesting MotW while stragglers came in range of the buff location, then up-scaled to GotW When your last raid members showed up. Either you had 2 players out of range during preliminary buffs (8/10), or 2 players didn't have pets out yet (10/12), or 1 pet class wasn't near the buff spot (9/11). Those are scenarios I'd call "bad" but "expected".

I'll make a note to add more feedback in the tooltip, to indicate why GotW isn't ready to be cast yet. I agree that your scenario isn't intuitive :P

Haha, yeah...I think imps get rocked to the bottom of the suggest list after that though. I'll try to find a way to detect their stupid phase shift.

I bet hunter "camouflage" is going to be equally dumb

Those are some really interesting metrics; I'll have to try out this Option House. Thanks for the link!

Early in development, I didn't have the tools or background or implementation available to know what (WoW/LUA) operations were "expensive". There's actually a lot going on every ~5 seconds; and from past experience, I think the two big hits are from the periodic check (server call), and from rebuilding the suggest list (memory reallocation). These operations aren't linked in any way, but they do occur at the same interval. A lot of this behavior is disabled while in combat, exactly because of the unknown risk/performance.

I'm hopeful that addon you linked can narrow it down for me (I'll play with it this week). My gut says rebuilding the suggest list (every ~5 seconds) is the expensive operation, and there are things I can (and need to) fix that will allow all the logic to run while in combat. My eye is on re-sorting without re-building, which would eliminate the periodic memory reallocation. This hasn't been a priority until now, because (a) it shouldn't happen in combat, and (b) I'd rather have an addon that works slowly outside combat, than be super efficient at not working right. Now that the big bugs are resolved, it's just a matter of deciding which is more important to my users; in-combat suggestions, or say, weapon enchants.

Anyway, if you're feeling frisky, try your test while in combat (i.e. let it run while killing Chillmaw, or a heroic). My suspicion is that the spike won't be there, even in a raid. I believe this because Blizzard has a vested interest in making efficient server calls, and the resources to follow through on it...whereas I'm just some schmuck who doesn't know how efficient LUA is at garbage collection & array reuse :P