Home :: Contact :: Sitemap ::

 

:: Pheasant

Abstract

For the physicist, the analysis phase of High Energy Physics (HEP) is the culmination of years of work on an experiment. In this phase, physics experimentalists look at the sheer volume of data collected by detector machines and try to infer statistical physics results.

The software systems in these areas have been growing in line with the complexity of the experiments. Unfortunately, for the software mining the stored data, the growth took an unstructured way. This is reflected negatively in the whole process performance, including both user productivity (in terms of man/hours) and the efficiency of the query systems (in terms of speed, space and cpu usage).

The aim of our work is to increase the user productivity and introduce a framework that allows computer experts to investigate efficient ways to optimize the High Energy Physics analysis process without requiring to be physics experts.

We achieve this by introducing an engineering methodology and making use of a declarative Domain Specific Visual Query Language (DSVQL) to raise the abstraction level in the existing query systems and to give room to new optimizations of different levels. In order to corroborate our argument, we implement a prototype framework, called PHEASANT, and a visual language named PHEASANT QL. This framework is developed in the context of the last big experiment, HERA-B in DESY(Deutsches Elektronen Synchroton in Hamburg, Germany), running in Europe before the LHC era. It is an interesting case study, since it has real data to study and users to interact with.

Project Website

Link to the Project Website

Reports & Publications

See also