M. Shupe, TB03 Update -- September 16, 2003 

>         I've updated the analysis code on pcfcal01 /home/shupe/tb03 to the
>         version I am running here--used to produce the linearity plots
>         shown by John.  Linking and running are done as follows
>         (for example):
>      
>         source tb03_analyze.lnk
>         nice tb03_analyze.out <run_e_200_4L_3128_6.inp >log_3128.txt &
>      
>         This particular run file run_e_200_4L_3128_6.inp, and its
>         corresponding parameter file cuts_e_200_4L_3128.inp are an example
>         of the latest settings.  This runs on pcfcal01, and the run*.inp
>         file can be edited to point to files elsewhere.
>      
>         In the past month we have been focusing on the following:
>      
>         (1)  There are several methods used to reconstructed cell energies,
>              in order to study the systematics and biases among the
>              different approaches.  But our primary method (used in making
>              the linearity plots), is labelled "EOpt".  John Rutherfoord
>              has been doing a lot of pulse shape studies and has
>              produced a method using "optimal coefficients", where the
>              information from six samples is used to determine the
>              time and amplitude of pulse-max.  When using this, sample
>              zero is part of the fit, so an initial scan is made through
>              the input file to determine the pedestals for the run from
>              pedestal triggers.
>      
>         (2)  The emphasis has been on electron response, and we have not
>              had time to do much with the pions yet.  At the moment,
>              the electron signal is taken from FCal module 1 only, but
>              we are working toward including those events where the
>              shower penetrates into FCal 2.  Since this summer, the
>              clustering algorithm has been improved by calculating
>              the area overlap fraction for each cell at the edge of the
>              cluster, and including that fraction of the cell's energy
>              in the cluster sum.  A cluster radius scan was then done,
>              leading us to set the EM clustering radius at 6.5 cm.
>              This is the same radius setting found by Peter Loch in
>              earlier testbeam runs--so no surprises.
>      
>         There is a long list of things still to do, but this is where
>         it stands at the moment.  I have not had time to update the
>         guide, but quick_guide_2.doc (a text file, not an MSWord file),
>         is still helpful in understanding the mechanics of the program.
>         It's in the same area as the code.  (So are input files, libraries,
>         and everything else!)
>      
>         I hope this helps.  Let me know if you encounter difficulties
>         or need more information.