Users can write their own analyzer from scratch or use the standard MIDAS analyzer framework which uses the HBOOK package for histogramming. Using the MIDAS analyzer framework has following advantages: Events are received automatically, only a user routine has to be written to process the events. This concept is similar to the frontend. The analyzer is structured into "stages", where each stage analyzes a part of the event and adds some calculated data to it, which can be read by later stages. This simplifies the design of complex analyzers. The analyzer framework can receive events from a MIDAS buffer (online analysis) or from a file (off-line-analysis) without recompilation. The analyzer framework can produce output files which may contain a combination of raw and analyzed data. Output files can be in different formats like HBOOK RZ files which can be directly analyzed with PAW or .root for ROOT access. An ODB dump contained in a data file can be retrieved and copied to the current ODB. This ensures that the same configuration values are used online and off-line. Additionally, parameters can be overloaded from off-line configuration files. Several files can be analyzed off-line each having its own configuration file. While HBOOK histograms have to be booked and filled manually from the user code, N-tuples can be booked automatically from one or more banks. This works also online where "live" N-tuples can be used to monitor an experiment with PAW. The following paragraphs explain these features in more detail and show how to use them.
To be written
To be written
Midas DOC Version 1.9.3 ---- PSI Stefan Ritt ----
Contributions: Pierre-Andre Amaudruz - Suzannah Daviel -
Doxygen - Peter Green -
Greg Hackman - Gertjan Hofman - Paul Knowles - Rudi Meier - Glenn
Moloney - Dave Morris - Konstantin Olchanski - Renee Poutissou
- Andreas Suter - Piotr Adam Zolnierczuk