Extrapolation of Extinction Curve

[karllark]

Extrapolating the extinction curve may cause weird and interesting results.
For example, knowledge from dust grain modeling tells us that the extinction curve at wavelengths shorter than 912 A does not continue to rise, but peaks around 800 A and then drops. This is only an example.
Basically, it might be good to have checks on the wavelength range in the spectrum and raise an error if it is beyond the tabulated extinction curve wavelength range.

[pllim]

Just for the record, currently the following snippet does not raise an error (but it should according to the spec above):

import pysynphot as S
import numpy as np

wave = np.arange(50, 80)
flux = np.ones_like(wave)
sp = S.ArraySpectrum(wave, flux)

ext = S.Extinction(0.1, 'mwavg')

sp2 = sp * ext