SMB:  Need to put in real gains, header is wrong
        Need to try the difference method for the gain.

SMB:  Try to shift ordr_str to each Arc.

SMB:  Use Kelson's method to find order centers.

SMB:  Take out slope in TFLAT, not sure why there is such a huge
        slope in the internal Quartz lamp. 
      Fit for a slope in the first call to bspline_extract

   


JXP: 8/?? --
1. Remove OV files from Bias step

JXP: 9/23/03 --
7.  Allow combining of files prior to object finding and tracing

SMB:  2/19/04
9.  Match skyspectra to UVES line list to check for pixel shifts.
   
10. Measure Arc line widths and centroids, after intial tracearc.
11. Deal with Partial Orders reasonably

SMB:  2/24/04
12. Improve milky flat fitting to produce better pixel flats

13. Attempt to automate look-up of standard star calibration
    --Add a check during autoid to see if a cataloged standard was observed.

14. Merge trcflat and fittflat to always get matching edges
15. Find a better way to skip orders with no lines found

JXP: 4/07/04

16.  Add a GROW option to mike_objcr
17.  Improve box car extraction post-optimal


----------------------------
COMPLETED

2. Add order guessing for Blue side to mike_trcflat (considerd necessary)
  JXP 4/4/04
3. Update Arc names in setup
  Completed 8/18/03 JXP
4. Update Arc routines to take an index value instead of just obj_id
  Completed 8/18/03 JXP
5. Add CR procedure
  Completed 8/19/03 JXP

JXP:  8/20/03 --  done, at least first iteration
6. Understand and remove the residual count on the blue side.  Probably
   an issue of incorrect variance estimation.   

8.  Make sure the vacuum vs air wavelengths are all done 'right'