On 6/6/2013 5:37 AM, Martijn van der Pauw wrote:
Hello,
I am developing an application that handles GRIB files and we use
netcdf Java for much of that. One of the demands is that this
application must also output (modified) GRIB files. I know this isn't
possible with the netcdf library. Is there a technical reason netcdf
doesn't support that?
And, do you have suggestions for tools (preferably in java) that could
handle the wring of GRIB files.
Thanks,
Martijn
_______________________________________________
netcdf-java mailing list
netcdf-java@xxxxxxxxxxxxxxxx
For list information or to unsubscribe, visit:
http://www.unidata.ucar.edu/mailing_lists/
Hi Martijn:
The technical reason is just that its hard to do well, and we have
limited resources.
At a deeper level, I personally think that GRIB and BUFR are not a good
format for long-term data storage, because the external table management
practices are inadequate, there is no reference library, and the
standard itself is difficult and has dark corners with no user group to
help explain. This results in files that are not actually interoperable
amongst various reading and writing libraries. So I would advocate not
using these formats if possible. See
http://www.unidata.ucar.edu/staff/caron/papers/GRIBarchivals.pdf for
more details.
Having said that, if you, or someone else wanted to develop GRIB writing
classes as part of netCDF-java library, we would be happy to publicize
and refer others to. I realize that GRIB and BUFR are mandated standards
and many have no choice but to use them. Unidata's strategy is to read
these formats and provide translation tools to netCDF/CF. This is an
ongoing effort that we are very committed to.
John
