* you need to have pgplot installed as well as all usual 
libraries that i'm using (see other tar file if needed)  
to compile the files you need here:  
  make gdata
  make gdata2
  make gscale 
  make zdicam
  make zdimag 
  make new

* all lsd profiles (computed with a standard K1 mask) 
are stored in subdirectory lsd according to their dates.  
their names relate to their rotation phase according to 
ephemeris:  
  jd0 = 2449997.223  per = 17.769426 (kovari et al 07)
(phase 0.0 : secondary star in front)  

* to extract the file containing all lsd profiles in zdi format:  
  cd lsd
  make extract 
  extract < zetaandI_aug08.in 
  extract < zetaandV_aug08.in

the numbers at the top of the .in files mean:  
  25.5  : K1 sini (velocity amplitude of primary)
  0.508 : phi0 (phase at which primary is in front) 
 -24.0  : gamma (system rv) 
 1.003  : continuum level of lsd profile
  10.0  : magnitude diff between stars (10.0 = big)  
  12    : number of spectra available 
-49.2 to 49.2  : size of velocity window for zdi file 
the first 4 were minimised to obtain minimum spot coverage (see below)  
(these are slightly different from values of kovary et al)
for each spectrum, the file mentions the rotation cycle, the sn level 
and the spectral resolution 

it produces 2 .s files, one for stokes i profiles and 
one for stokes v profiles respectively, both stored in 
the subdirectory zetaand: 
  zetaandI_mean_aug08.s
  zetaandV_mean_aug08.s

note that i got rid of the 12th stokes v profile that looks weird 
(abnormally large noise in spectrum, and not similar to stokes v profile 
one rotation cycle earlier)

* goto subdirectory zetaand 
to rescale all stokes i profiles to the same equivalent width: 
  rescale aug08 
it produces: 
  zetaandI_mean_aug08.rs
containing normalised stokes i profiles in zdi format 

* to reconstruct the brightness image 
  ../zdicam -b < zetaandI_aug08.in 
it starts from a red chi2 of 3.97 and ends up at a red chi2 of 1.0 
for a total spot coverage of about 2.2%
it assumes inc=65deg (kovary et al) and vsini=38.5 km/s (different from 
kovary et al) - also assumes zero differential rotation (only one 
rotation cycle covered)  

the reconstructed image is stored in file
  zetaandI_mean_aug08.m1
and the reconstructed profiles in file 
  zetaandI_mean_aug08.s1

the profile table it uses is (abdor.tab) contains the reference 
photospheric and spot profiles (number of points, wavelengths, 
normalised profile for both the photosphere and spot ref profiles) 
abdor.tab is the old abdor table, but tables with gaussian profiles 
work just as well.  

the 'multiplication factors for line Mean' are the factors by which 
you have to stretch the line to fit the observed equivalent width : 
you can either set them manually (by changing them in the .in file) 
or ask the code to adjust the equivalent width (by answering y to 
the 4th question in the list) - in the latter case, you get one more 
column in the output (right) telling you by how much you have to 
multiply the multiplication factors...  the code converges better 
and smoother when the ew is not fitted (so fit it first then freeze 
it to the proper value).  

by rerunning the whole process for various values of the continuum 
level, K1 sini, phi0 and gamma, you obtain the image with minimum 
spot (and also presumably optimal values of the 4 fitted parameters).  
vsini was fitted visually in this example 

* to reconstruct the magnetic image 
  ../zdicam -b < zetaandV_aug08.in
it starts from a red chi2 of 4.0 and converges at red chi2 of 1.0.  
the final field (Bmod) is 17.7 G in average 

the parameters describing the line profile (assumed gaussian) are 
in ../simple_zetaand.c (in particular, dl is deltalambda, ie 0.6 times 
the fwhm of the local profile, and str is the relative strength of 
the local gaussian profile at line centre) 

* to plot the resulting maps
  build aug08 
  display output.ps 
should be the same as map.ps 

to plot the spectra on the screen  
for stokes i
  show aug08 
for stokes v
  showv aug08 

to get a ps plot
for stokes i
  show2 aug08
for stokes v
  showv2 aug08 
should be the same as fiti.ps and fitv.ps 

* looks like the polar spot of zetaand is fairly small 
(may be the inclination angle is not favourable) 
hope it will still be useful for the zdi/interferometry comparison 
experiment