Why not just convert it direct to 709, you do that every time you shoot anyway, the raw sensor data gets converted to 709. The only difference  with raw is that instead of converting the sensor output to 709 in the camera, do it in a computer. If the workflow is properly done, which may mean locking out a lot of the variables, this should be more or less transparent to the user and behave no differently to doing it in camera. The big benefit being that if you do screw up your exposure or white balance post should have a much better chance of recovering something useable.

I think we need to think a bit differently to how we do now. We tend to assume raw must be graded, must have a load of post work done, when really that’s not necessarily the case. We are simply moving the process of converting the sensor data to 709 (or whatever) from the camera to the computer.

This is why ProRes Raw which could/might become a defacto standard is exciting. With a single standard it’s much easier to lock things down so they work as expected every time.