HDF: an update and future directions

HDF (Hierarchical Data Format) is a scientific data management project providing basic tools for managing scientific data, and support for scientists who use HDF software (http://hdf.ncsa.uiuc.edu/). HDF products include two distinct data formats, HDF4 and HDF5, supporting libraries for reading and writing HDF data, and utilities for working with HDF. Available since 1988, HDF has become a standard for scientific data archiving and data sharing. The range of users of HDF is illustrated in tabular form. HDF is the underlying format for the HDF-EOS standard (http://hdfeos.gsfc.nasa.govn/) and consequently is the target format for a large amount of remotely sensed data. HDF5 is a completely new version of HDF that provides an improved format and more general data model, supports efficient storage of very large datasets, and is supported by an I/O library that runs on both serial and parallel platforms. The home of the HDF project is the National Center for Supercomputing Applications (NCSA) at the University of Illinois at Urbana-Champaign (www.ncsa.uiuc.edu.)