I think you would have to work with RJ directly on this one. Coming up with the un-normalized curve would be relatively easy (all you need is a lux meter, a sealed container to test the lights in, and the patience to step through all 256 power levels and record the lumens at each power level - and maybe do it more than twice to ensure consistency in your readings) and you might be able to convince RJ that you can generate the raw data necessary to develop a light curve to the same standard he expects of himself. And he might accept that input from you if he is convinced you consistently do it correctly (because others will use what he loads up).
But the next step (converting the raw data into the curves implemented by the firmware) is probably not going to happen. RJ's process is a lot more involved and technical than the Vixen 2.5 method of just assigning a reference table to a channel. As I understand it, he actually changes the code in the PIC so that it accounts for the uneven dimming (he has 1024 steps as opposed to 256 steps). Given that quality control and configuration management are paramount in his mind with respect to the Lynx line of products, you would be best to probably enter into a PM discssion with him on the topic. I could be wrong - I often am - just ask my wife, but that is my two cents worth.
Steve