Points requiring a discussion between developers and users

  • passing to netcdf4 with compression of the diagnostics (proposed by Bertrand)
  • the surface assimilation seems to be separated into three distinct parts (SODA in HIRLAM, OI_MAIN in GMAP, VARASSIM in VEGEO) that evolve independently of each other. It could be interesting to discuss if a further common version could be possible.