Tamara reported that she spent quite some time reading up on Minuit during the past week. She has also performed Legendre fits on the asymmetry distribution and found that they were producing similar Chisquareds and errors as the polynomial fits. She also performed a Gaussian-like fit and finds that that is a little better, and she used it to apply a correction on the middle segment of the second HEC layer. After the correction, she finds that the response is improved but the resolution remains largely unchanged - there is a very small improvement at high energies, but it doesn't look really significant.
Next on Tamara's list of things to do are the energy-dependent depth-weights.

Tayfun has now looked at all runs and energies of his data sample and plotted E/(Emean) vs. phi. He has binned phi such, that each point represents the width of one cell. In this plot, the error bars on the points in the distribution represent the intrinsic resolution of each cell: about 0.9%. Also clearly seen on the plot is the reduced response of the impact points on the left. He has 12 points and performed a 9-parameter polynomial fit on the distribution. Then he used the results of the fit to correct for the reduced response. The effect of the correction on the resolution is not really significant. At lower energies, there appears to be some improvement, but at high energies it is pretty much the same.
Tayfun's next step will be to look at the global timing

Margret finally managed to compile LArG4. What had been tripping her up was the fact that g++ was pointing to a version number 2.96, while gcc was pointing to 3.2.3. The root of the cause a link set wrong in the system. Jan fixed that and from then on things started to work. Margret compile LArG4, ran an exercise with it and produced a picture of en electron hitting the barrel calorimeter. Then she attempted to compile LArG4TB, but finds it is not really quite ready yet. With some fair amount of guesswork she finally managed a compile and is now looking some more at the code to find out what it does and what not. She also tried to do a cvs checkout, but it turns out that we don't have a Kerberos ticket for CERN on the fate for the kserver, and the pserver seemed to cause some hurdles also.