Greetings all,

This note announces the release of HIRedux v1.0, an IDL package
I have developed for HIRES based on the MIKE pipeline that 
Scott Burles (MIT) and I have built in coordination with R. Bernstein 
(U Michigan).

It is fair to say that IDL novices will have a steep learning 
curve to ride.  Get ready for that..  I highly recommend the 
IDL Introduction Guide.  It is about 100pages of quick reading.

If you want to skip straight to the chase, go here:

http://www.ucolick.org/~xavier/HIRedux/index.html

As this is the first release, there are a lot of growing pains to 
be felt.  If you are likely to get upset when the code 'breaks' and you 
have no idea what to do, then I'd hold off for another 2 weeks or so.  
I'm hopeful, however, that some of you will get your hands dirty and
help me identify the 'bugs'.  If you wish to help build the code,
please contact me.

-----

What's new in V1.0:

Well, this is the first release, so everything is officially new.

The code will 

 aa. Organize the data for a single night
  a. Overscan subtract
  b. Flatfield (ideally with pixel flats)
  c. Trace the order curvature
  d. Automatically wavelength calibrate every science pixel
  e. Trace a single object in the slit
  f. Sky subtract
  g. Extract optimally 

Here is the kind of data that is working well right now:

UV cross-disperser only
C deckers
  -- Am attacking the B deckers right now and am wondering:
     1.  Do people really try to do sky subtraction with these?
     2.  Do people try to do optimal extraction with these?
     3.  Shouldn't we really be using a 5" long slit instead? 
         -- On the last point, I am pushing Keck to put in 5" long
            deckers.  Would anyone resist this if it meant the B
            deckers were replaced??  I believe there would be no 
	    order overlap until below 3100 Ang.
2x2 binning is best but other binning options will work 
Single objects only
Optimal extraction assuming a Gaussian PSF
Wavelength solutions for orders 57-116.
A quick redux mode for 'at the telescope' reduction  (<10min per chip)

Here is what I am working on in the short term (i.e. less than 2 weeks):


 i.  Wavelength solutions for the Red cross-disperser
ii.  Pixel flats for all binnings on the UV and many on the Red

  -- Note this is a new technique and one I highly recommend.  The idea
     is to correct pixel-to-pixel variations by illuminating an entire 
     single chip of the mosaic with light of the appropriate color.  
     A series of frames are taken and a 
     stack is made and normalized.  I consider this to be FAR advantageous
     to using the 'traditional' flats.  The 'traditional' flats should
     only be used for tracing the order curvature and the slit profile.

  -- Also note that I am asking the Observatory to take these calibration
     files on a regular basis (e.g. once per month).  They probably take 
     too long for the visiting astronomer to do in the afternoon. 

iii.  'Fluxing' --  I intend to put in a set of sensitivity functions
  which will provide relative fluxing but not absolute.

iv.  Coadding -- Routines will be added to combine multiple exposures
 at the same setup and also convert the order-by-order extractions into 
 a single 1D spectrum.

v.  Improving the hires_redux GUI which can be used to guide the
reduction steps.

vi.  Double checking the heliocentric velocity correction.  I fear
there may be a bug in this.

vii. Extend the wavelength soultions to at least Order 118.

viii.  Improve the results for the B deckers.

ix.  Install the code at Keck HQ

x.  Update the cookbook to match V1.0

---

Here are somewhat longer term projects (i.e. 1-2 months):

1)  Improving sky subtraction on the brightest sky lines.  Currently
lines like 5575 are not being handled well.

2)  Adding an optimal extraction algorithm with a non-parametric profile
 
3)  Placing the output to the screen into a Log file.  Or at least
important WARNING messages.


Feedback is encouraged.   

Email hiredux@ucolick.org  starting on Monday.

Cheers,

X