The main aims of the project were to contribute to a complete simulation of the TESLA accelerator, and to examine the reconstruction of the luminosity spectrum, and its behaviour under different circumstances.
Three broad questions were posed in the project proposal:
a)Does correlated dispersion compensate for disruption?
b)Is there a correlation with event position within the bunch?
c)What can we say about offsets and beam bending?
Overall the aims of the project were extended: much time was spent finding the conditions under which the luminosity spectrum reconstructions produce acceptable results, and their effect on top mass measurement.
Questions a.) and b.) were answered satisfactorily, but it is too early at this stage to address question c.). The code written should easily accommodate for beam bending and offsets, as the preliminary guinea-pig plots show, but until problems with the luminosity spectrum are fully understood, it does not seem constructive to investigate exotic beam shapes. Instead of broadening into an methodical taxonomic investigation of beam pathologies, the project was pushed forward to investigate a simple model of the top cross-section, and the influence of the reconstruction on it.
The main difficulties encountered were technical: the codebase is now in 4 languages (c++/ CERN root macros, python, fortran, as well as some linux bash shell), and the author alone wrote ~4000 lines of code. This means that bugs were harder to find, and languages had to be learnt on the spot. For example we had a bug in the boost code in bhwrun, that took weeks to find: this meant data that had been worked on was often rendered useless, and had to be replaced.
Our approach to discovering bugs was mostly through the physics: scatter and surface plots of all aspects of the data were generated, and its physical significance analysed. An unexpected phenomenon would then mean either new physics or buddy code.
A difficult bug was, for example, an asymmetry in the acollinearity angle distribution ,which came from the bhwide extension which calculated the boost for the particles. When we discovered it, we did not know whether it came from guinea-pig, bhwide, the interfacing scripts, or indeed the physics.