General considerations (no decision can be made right away):
- AMNS DB base... (IDS, UDA, CernFS, ...)
- AMNS "expert" level - git-like version tracking, not open-source data
- Data licensing issues are to be addressed.
a) EIRENE should consider reading AMNS data directly following the B2 experience (SOLPS-ITER).
b) We need a clear catalogue of data available inside the AMNS (with quality assessment).
c) AMNS should take care of (changing) data formats (ADAS, SDTrimSP etc.) - it should provide the end result (excessed as a function call).
d) AMNS should provide a way to find the preferable data, but also one should be able to reproduce any earlier simulations.
e) Installation of infrastructure should be easy, well-documented, automatized (best way - sript...).
f) AMNS should provide returning of vectors/matrices of data instead of singular values.