Try to keep paramters "same", the more the better to compare!
It will become difficult if you have more than 2 "unknown"

When we discuss DPS, first ask the question what to keep "same"
Then change one parameter and look what happens with the others
You must be careful when you change more than "one"

We can agree that boss lifepoints stay the same.
We can agree that lets say we have the DPS list from a yesterdays bosskill.
Well, we can add many other things too, but lets call it our first data to compare (left side)
What do we change now (becomes right side)?
When we change DPS by keeping everything same (no new run, just pure math) you are limited to phase changes. As result "time" will change, lets say he dies faster or you skip phases...

Whatever, we do not have that sort of "change" in real (ingame) because gear stays same! The Problem you must now think around is clear time, that way you can bring it back to a "nearly" same level.

So, keeping the clear time "same" makes it more realistic to compare different groups or even fights. First bosskill with second bosskill or group A with group B... we just look at clear time and if its "nearly" same you can compare. Do not forget we also keep the other parameters "same" (Jobs, Players, gear and so on)

Ok then lets change DPS again with keeping in mind clear time is "same" (example: first bosskill and second bosskill) *comparing groups is a bit more complex*

You are limited to boss lifepoints as result of same "clear time"

Lets say on first bosskill no one uses any buffs or potions and on second bosskill one player goes all out - what do you think will happen? I can show you many videos or DPS photos from 100 of fights, always different DPS and it clearly shows that when someone having higher DPS one (or more) others will fall in theirs!

I guess a good example is if you have one player death on DPS-check-mechanics, the missing DPS must come from the remaining members (opposite of one doing more DPS argument). You have to have a "seed", something as base reference if you try to compare.

You can ask differently: When you compare the pictures from KarstenS: Which parameters did change?

DPS doesnt have anything to do with total damage done, its strange you bring that up all the time. But it is indeed a fact that the result is influenced of boss lifepoints. We do not try to calculate DPS, we are calculating something that includes DPS and LIFEPOINTS...

When we compare, we do not have the luxury to say we have a formula and we can let all parameters be everything - we have fixed parameters based on the "seed"!

I guess i need to explain the "how do you measure DPS" because its the loophole:

If you change parameter "DPS", it does not automatically result in a "faster" fight because depending on the other parameters. The Problem here is that you dont have enough information to solve the equation. When you add that the clear time was same, then the conclusion is something other as when you add that you cleared it faster. (did you improve/progress or have you lost some DPS)

KarstenS made an example where everyone is doing more DPS but then what does change? It is a physical law that left and right side must be equal (1=1 or 2=2 or 1/2 = 2/4). We can see and calculate: the time is what was changing.

To explain this in detail on the same fight will become very complex and detailed, because when we take above-mentioned example "on first bosskill no one uses any buffs or potions and on second bosskill one player goes all out" it will mostly result in boss skipping phases (faster kill). Please dont force me to explain probabillity calculations...

Thats not the origin of this discussion, we begun with normal "dungeons" and possibility of killing monsters too fast. Calculated from the whole dungeon it sums up to an average win or lose of DPS. It doesnt matter if you speed up because its not one single enemy but the total of them. You can not switch to next monster if there are only 3 and you fight the last one. There are even players who have fun to pull monsters out of range or something other annoying. I used this argument in hope you will not judge low DPS players but look more deeper because there are many reasons why DPS is low.

Special on measuring DPS:
You can view at it from start or you can view on it from end.

When a Monster just have 300 Lifepoints and you kill it in 1s then you have a problem if your DPS-Meter tells you all 8 players did 300 DPS (based on the thought every player can onehit that monster).

There is something wrong because the Monster did not have that many life and someone must have missed - thats where measure interval becomes interesting (but thats a longer story, parsers had a lot trouble with DoTs and HotS)

From Start view: You do not know the end and your DPS can go as high as you like as long when your "End-Parameters" can keep up with it (boss goes faster down)
From End view: You do not know the start and your DPS can go as high as you like as long when other "Start-parameters" compensate for that (other 7 players made less DPS)
While in flow of fight there are many other parameters you have to consider that can influence the outcome but above example was just to show the problem of measuring DPS

At start of a boss players saying "How high did my burst DPS go up this time?"
At end of a boss players asking "what was my average DPS?"

Back to math:
From start: 1+3=?
From end: ?+3=4