REPORT MINI-WORKSHOP 4-5/12/2002 IN LYON

• Due to various issues raised by the FITS experts in their comments to the current Euro3D data format and worries expressed by the Lyon team concerning large-file handling, it had been decided in early november to use binary tables instead of 2D images to store the data. Markus presented the latest version of the data format. An updated version of the Format Definition document was not available at that time. It was emphasized that the changes affected only the way the data are stored on disk. No change in the format functionalities was introduced. Summarizing, the new FITs file contains three mandatory extensions : one containing only the initial fits header, one containing a binary table and one containing the group information. The binary table now contains both the data (signal, data quality and statistical error) and position information, merging four extensions of the "old" format into a single one. Spectra are now stored into vector elements of the binary table. This storage strategy will ensure much higher performances in case of huge data cubes. 
• The work on the EURO3D I/O library is in progress, despite delays introduced by the format changes. It is based on the former Lyon I/O library, that works very well. It includes test programs that check the ability to read/write Euro3D format files. A preliminar version, not including group handling  has been provided.
• Different solutions on how to handle the data in the C-programs based on the library were discussed (definition of the C-structures and of the basic I/O functions).  We converged toward a solution, which had, in addition, the advantage of minimizing the  changes in the actual version on the I/O library.
• Arlette has presented in detail the contents of the existing library (functions, structures...) as well as its philosophy. A detailed documentation is generated semi-automatically from the source code, allowing an easy updating of the documentation of the routines (purpose, syntax...). The documentation will also include a description of the the Euro3D format  C-structures. 
• Sebastian presented his ideas about the technical requirements for the visualization tool, and on which graphics library could be used. He has shown the latest capabilities of PGPLOT, comparing them with the abilities of other graphics libraries (more extended, easy to transport between platforms, easy to manage, well documented, free...). Both Arlette and Pierre have agreed that, taking into consideration the time constraints, it will be nice to use a well tested library. Arlette has seen as an advantage the fact that PGPLOT interacts with TCL/TK, and different posibilities for the GUI have been explored.
• It has been agreed to present, at the visualization mini-workshop, a simple prototype of the visualization tool that would use the I/O library to read a raw Euro3D frame and PGPLOT to show the data. This prototype could, as an example, allow to plot a spectrum when clicking on a frame. All agreed that, to present this as an starting point for the next January meeting, would be a good idea. The Lyon group acceptedto provide the current version of the I/O library to start the work on the prototype and detect possible problems. Sebastian compromises to test the I/O library and provide the prototype for the January meeting.
• All agreed in the gradual development of the visualization tool, from a command line based tool to a more complex GUI (if possible), starting from the basic things and then add more and more functionalities. Discussion on which capabilities to include in the visualization tool will would be at the heart of the January Potsdam meeting.