Design goals
open data format for open science projects
self describing data
flexible data structure layout
hierarchical data structure (nesting groups, dictionaries)
(posix path syntax support?)
extendable data structure
platform independent
simple object model
support of standard metadata formats (key/value, …)
support of standard dataset formats (hdf5, netcdf, csv, …)
support of standard dataset types (datacubes, tables, series, …)
support of physical units (conversion of units)
transparent, optional data compression (zlib, blosc, …)
support of (de-)serialization of every dataset type (group, data, metadata)
easy defineable (project) standards, e.g. for a uniaxial tension test (UT)
(optional data encryption (gpg, …))
change management support?
Enable use of data structures from existing tensor libraries transparently
(single writer/ multiple reader (swmr) support)
(nested table support)