Warren gave a nice detailed overview over his testbeam analysis. He described the set-up and then showed his results of the energy-sharing for various impact points and beam particles: 120 GeV e+ and 150 (120) GeV pions. His energy-reconstruction followed the H8 TB analysis and involved a simple energy sum over all CaloCells as well as 3D topological clustering. (Warren did 3D clustering within each calorimeter type separately.) By plotting energies of different layers (em, had) against each other he was able to identify contaminations by different particles in the beam. During a discussion over some of Warren's results, Rob reminded us to not apply any clustering across different calorimeters, because different e.m. scales in the different types of calorimeter can really confuse things, in fact it is really important to understand the basic energy scales first, and that needs a systematic study.
Rob summarized some talks from the LAr week and showed us his own slides on his OFC study. While most channels are `well behaved', there are several problem channels. These channels are typically EMEC channels for which only high gain data exists and they are typically located right on the edge. According to EMEC experts, the inner wheel of the module 0 is also plagued with severe cross talk.
Rob also summarized the present software issues and features of various Athena releases. The most important problem coming up soon is that with Athena release 10.0.6 a new version of Gaudi comes into effect and it will break the code for the 2002 TB Analysis!
Rolf informed us of his battle and the final success of getting support to implement his code on calculating noise in the calorimeter. It will be implemented as `INoiseTool'. In this framework, the user can chose the noise he/she want to look at over a set of enumerators.
Margret did some more studies of the reconstructed energies of MC events, and was unable to reproduce the discrepancies she found for pions earlier in the year - she is not sure why the distributions look different now. She also looked at the eta/phi distributions of CaloCells of MC TB events and finds that the HEC modules need to be shifted in phi.
Rob then gave us an overview over the upcoming cosmic ray commissioning schedule of Atlas. A MC study of rates should be performed to study first data taking with cosmic rays.
Kai gave us a nice introduction on how to perform a TAG-based analysis. As events are processed, they get attributed `tags', which identify their type and properties to the user. These Tags can be stored e.g. in a collections file. Then, during an analysis, the user can pick and chose events according to their tag.
Rolf outlined changes that will need to be made in order for the TB code to run under the new version of Gaudi. This work is really pressing, as it should go into 10.0.4 which will be closed on May 25.
Rolf also informed us, that with RedHat Linux on it way out, new Athena versions
will now be installed at UVic for SLC3/RHEL3.4 only. He also considers to mirror
the cvs archive at UVic and to run the lxr code-browser locally.