Quote Originally Posted by Tunda View Post
There are instances when we don't require healer in dawntrail
This issue exist and there is no salvation to it
I'm aware. But someone bleeding isn't a reason to shoot them.

Supports dps jobs are just better in all cases
No. They're better than an alternative when the possible average clear speed with an otherwise optimal party is better with them than the alternative. For short, though, without any utility*, a job with "support" is better than jobs without support when it produces more rDPS+aDPS, and non-redundant rHPS+aHPS than its competing option(s). I.e., when its straight-up more powerful.
No, raid buffs should not even be thought of as utility; if anything they further constrain the party, therefore creating a loss to utility (difficult-to-quantify contribution towards an encounter's ultimate goal), in this case by reducing opportunities to push out a bit more relative potency over the fight as compared to less constrained means of reaching the same rDPS).

If neither its likely nor theoretical value in offensive or sustain is difficult to quantify for a given composition, there is no more reason to call it "utility" than to call Heavy Swing or Cure "utility"; it's just damage or sustain, respectively.

A final note: rDPS alone never tells you something's full strength. If something completely lacks any indirect contribution, then aDPS will give you a basically full picture, but rDPS will never do so, since it completely (and purposely) removes how well that job makes use of others' buffs.
But we have clear issue and clear solution
Your "clear issue" is, judging from your discussion of it, poorly understood, and your "clear solution" would require mangle the identities of a handful of jobs, constraining job identity around an arbitrary and unnecessary bimodal (degree of indirect contribution), and a tremendous amount of developer work for... more than likely a worse result for each issue you've used as reason to make said suggestion.

That means it is optimal to play 2 support 2 healer 2 dps 2 tank
This is the truth
You're both somewhat misunderstanding and overcomplicating the matter.

A job that contributes a portion of its rDPS through a single other player will always pair best with the best possible exploiter of its finite (highest 20s per 120s burst) and constant (highest would-be aDPS overall) buffs. Because that's just math.
Apart from that, you're just following Role stat bonuses, whoever's generally strongest, and equilibrium (the more relative potency per minute there is to buff, the more each scalar is worth, but --assuming balance-- the more jobs with scalars you have, the less original potency you can bring).

We tend to take Viper because it's OP. Because Viper is OP, or if for whatever reason the static has a SAM or BLM, Dancer is competitive for the single Ranger we're forced to take, but otherwise we take Bard... because, atop any extra utility, it's simply brings more total damage than MCH. Etc., etc.

That said, the 2/2/2/2 assumption you have isn't particularly accurate right now, let alone historically:
  • 4 support-heavy DPS has at times been stronger (old NIN/DRG/MCH/BRD). Taking just 1 support-heavy DPS job has, at other times, been strongest.
  • Even now, unless you consider both MNK and PIC "support" jobs, we don't tend towards 2/2/2/2 for the fastest runs at time of writing.

There is no hard 2/2/2/2 rule, just the wax and wane of balance and some basic math, ultimately bounded by bonuses to primary stat and Limit Gauge generation.