NetCDF is a format developed at Unidata that was
inspired by the CDF format from NASA. NetCDF
stands for "Network Common Data Form" and is a
self-describing data format, commonly used in
scientific and engineering applications.
Build options have been fixed to conform to standards. Bugfixes include szlib linking, support for large blocksize filesystems, backwards compatibility, DAP support, unsigned longlong, support for platforms without uint, parallel I/O in F77, and assorted F90 issues. The code was upgraded to work with HDF5 1.8.7. The documentation was updated.
This release modifies the OPeNDAP constraint parser to be more compatible with a Java version of the parser. It modifies ncgen to utilize iterators internally; there should be no user visible effect. It adds libdispatch, which decides whether to call a netcdf classic, netcdf-4, or OPeNDAP version of each function in the API. It has a fix for a make bug that caused the documentation to be unnecessarily rebuilt after make clean. It fixes a bug in accessing a multidimensional variable with more than 4 billion values on a 32-bit platform. It fixes memory performance problems in parallel I/O. There are minor bugfixes, speedups, and documentation updates.
Updates include remote data access with a built-in OPeNDAP client, a new utility nccopy, ncgen that works with netCDF-4 enhanced data mode, the ability to read some HDF4 and HDF5 data files, use of the parallel-netcdf library for parallel I/O to classic format files, bugfixes, and portability and performance enhancements.
This release adds optional arguments to the F90 API so that all netCDF-4 settings may be accomplished with optional arguments, instead of separate function calls. It adds control of the HDF5 chunk cache to allow for user performance tuning. It changes default chunking to better handle very large variables. Performance refactoring of core netCDF-4/HDF5 read/write code. Additional parallel I/O tests and benchmarks. Experimental OPeNDAP support based on a new C-only implementation. Improved portability, especially on high-performance platforms.