it seems that the problem you want to solve is not really clear, and suggesting a big architectural change based on some gut feelings is hard to develop against.
Maybe instead of suggesting solutions it would be better to describe the problem you are having.
Why do you believe that?
Don’t you think there will be an ACES 3 and 4 etc…
In a fast-changing world, only an agile system will survive.
(ACES 3.0 will have a DRT that is differentiable function, I am sure 
)
And I was wrong.
From ACEScg docs:
Problem with inconsistency can (and I hope, will) be solved by one standard CAT.
New working space probably is not an option because of a lot of textures created in AP1, as one example that comes to my mind.
The rest in not strictly a problem, let’s call it a one user’s feedback on user experience.
ACES 1.0 was released in the end of 2014, it’s almost 9 years ago now
I think ACES 2 is going to stay for many years
Yes-ish… There are other reasonable choices for imaginary primaries too. But AP0 is fine.
My (least) favorite subject: DNG (DJI Drone) debayering…
With the assumption that Resolve debayers natively to Rec.709/Linear (with a manual override of +1.4 exposure to match D-Log), what is the correct (well, maybe not correct, but “agreed-upon)” chromatic adaptation to go from DNG RAW (Rec709 D65??) to ACES?
From what I can tell, if you’re ACES color managed in Resolve, it’s using CAT02 (akin to the following:
