Caius' study showed that the next affect of all the drifts together is a ~2% effect at the FD over 2 years, so we feel that the best thing to do is simply not correct at all and eat it into our error budget.
As for linearity:
(a) at the FD, we have uncovered new nonlinearities (probably in the PINs) which we don't understand. Peter is working hard do try to work around these problems by doing simultaneous fitting LG, HG, and N/F linearity curves.
(b) at the ND, we continue to fight hardware problems (we still don't have high-gain LEDs) and Tingyun and Simona are working hard at getting the first iteration of the linearity constants ready.
I've been trying to say this for some time: we do NOT currently have a drift calibration. We do NOT currently have a linearity calibration. Our strip-to-strip calibration is not in full production mode, although R&D is continuing. Our attenuation calibration is still undergoing validation and improvement. Or stopping-muon calibration is mature, but has unexplaned ~2% systematic effects that we still don't understand. On top of that, our PE calibrations are antiquated and the timing calibrations have lapsed and are being picked up again.
We are working very hard on all of these things, but none of them are easy. Many problems exist at the ND and FD which were either not present at CalDet, or were unseen because they were masked by larger problems (like the blue LEDs).
Add it all up and we THINK that this means we have a ~5% shower energy uncertainty if we apply the best current s2s, attenuation, and d2d numbers.. and that's the best we can do for your reprocessing schedule.