Shortcomings

This software isn't totally perfect. It does have some shortcomings and assumptions.

Trace: BMEP assumes that your 2D reduction resulted in perfectly rectified data and there is no drift in the trace of an object across the wavelength direction in the slit

Rotation: As of right now, there is no way to handle rotation in the spectra. What I've been doing is just binning all columns with the emission line which results in a fairly broad profile. This gets more flux, but the line widths should be broadened because of the rotation. Ideally it should fit several gaussians along the spatial direction and shift each row accordingly...

Fitting doublets: When estimating redshifts, doublets are just fit to one single gaussian. This is OK if the data is low SN, but may fail if you have nice data.

other things...