Some spectra have relevant covariance matrices which should be stored in the FITS file.
from @kbwestfall:
I would envision it as two different tables in 1 file. The current I/O basically works directly with HDUs. So you could have the main code in core (maybe in NDData actually) that constructs the HDUs that then get added to the spectrum HDU(s).
Some spectra have relevant covariance matrices which should be stored in the FITS file.
from @kbwestfall:
I would envision it as two different tables in 1 file. The current I/O basically works directly with HDUs. So you could have the main code in core (maybe in NDData actually) that constructs the HDUs that then get added to the spectrum HDU(s).