Closing the loop
Inserting a-priory knowledge first principle physics
- greatly enhances accuracy even under strict classical calibration concepts
- adds enough predictive power to tackle problematic cases
Physical models
- can be constructed and tuned to sufficient accuracy for the purpose
- add autonomy to pipeline calibration - i.e. no run-away
Next (logical) step Forward Analysis
- all that needs to be done now is to construct a complete model of the observation process - atmosphere, telescope, spectrograph, detector.
- and simulate observations of a range of model targets
- in order to ask statistically sound the question which of those model targets likely produced the actual raw data