Tamara showed some plots of the normalized energies measured in the EMEC and HEC versus time.
In order to produce these plots, she took the time-intervals covered by various runs and aligned them such that for each run the `time' is from 0 to 22 ns. The normalized energies are obtained by dividing the measured energy for a particular bin in time by the energy averaged over all time bins for a particular run.
She plotted the results of a number of runs into the same histogram and finds that the shape of the time-dependence is independent of the beam energy. This means, that a global correction can be performed, regardless of energy!

Tayfun reported that he created a list of runs that are tagged `good' or `bad', depending on the global timing and the beam-chamber reconstruction. If a run is has more than 60% bad global time or bad beam chamber reconstruction he classifies it as `bad'.
Michel pointed out that it would be useful of he would add the good/bad acceptance/rejection criteria into the list to form a table or a poor man's data-base. This would make it easier to study variation on the criteria and to select for different criteria.

Margret showed some plots of her longitudinal shower shape analysis. She now used the approach, where she fits for the shower-shape parameters in the EMEC and for the amount of material that is always in front of the EMEC. These fits are done runs that have no additional material in front.
Then she fixes all shower shape parameters and fits for the amount of material in front. This results in somewhat poor and inconsistent fits. The thickness of the individual Pb plates of extra material is known to be about 0.45 Xo. With that, the fit-results indicate either less and less material that is always in front of the EMEC or a plate-thickness that is significantly less than the one given. Neither case is very likely.
So, instead of fitting for the material thicknesses, Margret fixed all shower shape parameters to the ones obtained with the runs that have no additional material in front, and she also fixes the amount of material in front to be the amount of material obtained in the initial fits plus the known material of the additional plates. Then she overlays the measured EMEC energies from the first, second and third EMEC compartment with the theoretical energies, obtained from the initial shower-shape parameters. She finds that there is good agreement in the second and third compartment of the EMEC. She also finds that for the runs with lots of additional material in front, the energy measured in the first compartment tends to be significantly lower than the energy that should have been measured, if the theory is correct.
The conclusion that was reached by everyone is that the most likely scenario is, that in a significant number of events the shower start early and radiates energy far off to the side, where it is not detected. It was suggested that Margret should look at the lateral shower size, and, if possible see whether there are any hits in the front veto wall. Michel pointed out that without MC studies it will not be possible to obtain any further understanding on this. Richard pointed out that the understanding of the longitudinal shower shape might be useful to separate em showers from pion showers e.g. in tau analyses in ATLAS. Furthermore the shower-shape might help to assign depth-weights to cells that are used in position resolution studies.

Michel reported that he talked to Rob on the phone and learned the very important information that Naoko now uses FFT to obtain the digital filtering weights for the EMEC. After she obtained the new weights, she finds that the need for timing corrections goes away!