> Organization: .
> Keywords: 199209242313.AA25254
Hi Rich,
> Part of the driving force behind netCDF is the possibility of writing
> applications that access generic netCDF files and operate on them to do
> whatever they will. Graphics applications seem like an obvious
> possibility. However, it is likely that novices (like myself) will leave out
> "necessary" attributes or other information that effectively limit that use
> of generic applications. Have there been writings or discussions concerning
> what a "well-behaved" netCDF files look like? If so, I'd like to hear about
> it!
A draft document is available via anonymous FTP suggesting some conventions
for conventional attributes, attribute and variable naming, units syntax,
and time conventions in the file
host: unidata.ucar.edu
file: pub/netcdf/conventions.info
It may be undergoing some revisions in the next few months as a local group
works to achieve agreement on additional standards for netCDF files
containing certain specific kinds of meteorological data.
We also make a directory available via anonymous FTP of some example netCDF
file structures, in the form of ASCII CDL files. This is mostly just a
collection of structures for files we currently use for gridded model data,
geo-referenced satellite images, and point observations. If anyone out
there has some particularly good examples of CDL files that might be
instructive to others, we would be happy to include them also. The
directory is pub/netcdf/cdl/ on the host unidata.ucar.edu.
Finally, you can look in the archive for this mailing list for other
discussions of what should be in a netCDF file. It's available for
anonymous FTP in the file mail-archives/netcdfgroup on unidata.ucar.edu, but
be warned that it has now grown to almost 1 megabyte. We will soon be
anouncing a WAIS server for the mailing-list archive which will permit
easier access to information in past postings.
--Russ
