Is there place for perfectionism in the NIR spectral data reduction?

"Despite the crucial importance of the near-infrared spectral domain for understanding the star formation and galaxy evolution, NIR observations and data reduction represent a significant challenge. The known complexity of NIR detectors is aggravated by the airglow emission in the upper atmosphere and the water absorption in the troposphere so that up until now, the astronomical community is divided on the issue whether ground based NIR spectroscopy has a future or should it move completely to space (JWST, Euclid, WFIRST). I will share my experience of pipeline development for low- and intermediate-\nresolution spectrographs operated at Magellan and MMT. The MMIRS data reduction pipeline became the first example of the sky subtraction quality approaching the limit set by the Poisson photon noise and demonstrated the feasibility of low-resolution (R=1200-3000) NIR spectroscopy from the ground even for very faint (J=24.5) continuum sources. On the other hand, the FIRE Bright Source Pipeline developed specifically for high signal-to-noise intermediate resolution stellar spectra proves that systematics in the flux calibration and telluric absorption correction can be pushed down to the (sub-)percent level. My conclusion is that even though substantial effort and time investment is needed to design and develop NIR spectroscopic pipelines for ground based instruments, it will pay off, if done properly, and open new windows of opportunity in the ELT era."