assembleCoadd.py performs a step where it scales the input images to a "common zero point" using ScaleZeroPointTask.
DM-15254 implemented PhotoCalib.calibrateImage() which photometrically calibrates images directly. DM-15751 uses the new calibrateImage method to return those calibrated images during coaddition, removing the need for ScaleZeroPointTask.
Note that this is not really a change in what "units" coaddition is working in: after scaleZeroPoint.run(), the images should be in a physical flux unit, divided by some common scale factor. This RFC just proposes to remove that common scale factor and have the makeCoaddTempExp output all have a photometric calibration of 1.