Andreas J. Wicenec (European Southern Observatory, Germany), Mario Nonino (Osservatorio Astronomico di Trieste, Italy), Erik Deul (Leiden Observatory, The Netherlands)
Reference URL: http://www.eso.org/eis
ESO has commited itself to build up an Science Archive Research Environment (SARE) which should, besides other projects, support public survey projects such as the ESO Imaging Survey (EIS). The requirements of surveys imposed to the Observatory, to the Archive and eventually to the reduction system are very different than for usual observing programs. The speed with which raw EIS data have been converted to deliverable products is quite unprecedented and should be maintained for upcoming projects as well. By using new telescopes and wide field instruments the amount of data will grow by a factor of 4 to 10, making the pragmatic approach of the EIS data reduction inappropriate.
In EIS areas like status control, data flow control and versioning control for software, pipeline configuration and data products turned out to be essential for an efficient and flexible treatment of the large amount of data. We show and describe an object oriented data model which not only couples data and methods but also different versions of data and methods. Such a model supports cases which are very difficult to handle with an usual system. For instance the optimisation procedure of the parameters and methods of the object detection, astrometry and photometry tasks or multiple observations of the same area or even cross-correlation with external data sets. We introduce the abstract concept of generalized 'Reduction Blocks' (RB) as the atomic quantities that 'live' within the model, where the basic RB concept has already been used to build up the VLT Data Pipeline System.