Data formats and Archiving

• Should we use FITS all the way ? If so, do we need to get extensions to the standard approved? Don't we need a support for hierarchical keywords ? Is there a way to compress data and not headers ?
• Do we need a specific format for raw data? check VLBA FITS format at http://www.aoc.nrao.edu/~cflatter/FITS-IDI.html . This would need some additions for millimeter-wave.
• How much data will need to be archived? Where will this be done? What access is to be provided to the community-all data via the net? only catalogs? snapshots?
• Archive calibrated, adequately compressed, uv data, from which images can be easily produced.
• Archive also on-line produced images when available (this provides a quick-look facility to the archive). In that case keep only significant data (cut out the noise at e.g. 3$\sigma$) ?
• How is the archive to be used ?