** 1.   Reject lines to improve 1D Arc fit RMS:  Orders 89, 90, 100
  --- I wonder if line centroiding is failing in x_templarc
** 2.   Deal with vignetting at >4000pix in Trace flat
  -- Just blue?
** 3.   Modify hires_slitflat as Scott suggested (extrapolation)
**4.   Worry about scattered light in skysub
**5.   Add header to Final files
** 6.   Subtract scattered light prior to Flat normalization
** 7.   Consider optical mapping of the wavelength solution
** 8.   Worry about DETILT for Internal flats
** 9.   Fit higher order to interpolation in x_fntobj
**10.  Set aper according to seeing (i.e. obj profile)
11.  Wavelengths wrong at edge of Order #68?
**12.  Add an asymmmetric profile for Optimal (non-parametric?)
**13.  Need to improve sky sub for X Optimal
** 14.  Bluest orders are probably dragging things down in x_fittflat; 
**      Consider a separation limit on using it in the fit.  That is, don't
**      fit them but do extrapolate. Only should be important for the blue chip.
** 15.  Normalization in the bluest orders is not so good.   Probably best to
	do NO NORMALIZATION in those orders for now.
     The problem is more scattered light than anything else.  This would
     be solved with a better flat.
**16.  Why are the counts so discrete given the OV sub?
**17.  Fix wavelength zeropoint.  Something is wrong
18.  Enable stacked flat input for hires_qckrdx
19.  Create logfiles of diagnostic output
20.  Consider lower order coeff for Arc solutions
** 21.  Fix helio!!
22.  Set row_shift in fitarc according to Delta(ECHANGL)
## 23.  Change x_templarc so that it can return only 5 lines well spaced on the chip
  with maximum signal
24.  Consider a Bias frame
25.  Get more arc lines for the bluest orders
*26.  Toss out orders on the edge for Edgeflat
27.  Toss out orders for tracing (optional)
	28.  Deal with slit edges for setups without Flats (extrapolate the 
	  normalization out a few pixels or toss out...)
29.  Modify hires_fndtmpl so that ECHANGL is minimized when getting the
   'outlying' orders
30.  Mask out bad columns