I was taught this in my fourth CCR class (SF2). Kudos to @Capt Tom McCarthy for his thoroughness, even on a crossover. I think understanding linearity is great, but there's more than one way to skin this cat. On the surface I calibrate my cells at %100 and check that they all go down to %21 with a dill flush and the mV should be within specs. Obviously, if the cells weren't linear, then they would show a skewed value for air. Then, at 20 ft I do an O2 flush, raising my PPO2 to 1.6. and then a dil flush which takes it back to 0.3. Again, if the cells were not linear or even limited, I would be getting skewed values, right? If my PDC reads all PPO2s correctly, then why is it important to do the mV math? Moreover, checking that my cells are able to go over 1.3 (my setpoint) is pretty damn important. Without any mV math, I've checked cell linearity as well as if they are limited. Why over-complicate a pretty simple process? Flame suit on, so fire away! 
