Re: 20030719: hdf-eos to netcdf conversion

>To: address@hidden
>From: Suzana Djurcilov <address@hidden>
>Subject: Re: 20030719: hdf-eos to netcdf conversion
>Organization: UC Santa Cruz
>Keywords: 200307190018.h6J0I0Ld012879 HDF to netCDF

Suzana,

> I know you're spending your precious time on a problen that you don't HAVE
> TO  spend time and and I want you to know that I aprreciate it.
> 
> I did replace the YDim:Grid variable with YDim-Grid and that seems to have
> gotten me past that hurdle. It now croaks on line 52:
>                 :Miami Format = 14 ;
> with a syntax error message, and I have a feeling there'll be many more.

That's another special character in a component name, in this case
blank " ".  If you changed it to ":Miami_Format = 14;", for example,
ncgen wouldn't see a syntax error.  If you looked through the other
attribute names and replaced the embedded blanks with "_", ncgen just
might work.  This is the we changed the allowable characters in netCDF
component names to be match what could be parsed in CDL files, and to
more closely match what characters are allowed in variable names in
programming languages.

> The real question is probably why ncdump from netcdf3.5 doesn't even
> recognize the file.

Since the file was generated by the version of netCDF in HDF, it
produces HDF files.  The netCDF layer in HDF4 was an attempt to
provide some interoperability, so people with programs that used the
netCDF data model and library could link them with the HDF-embedded
library and read/write HDF files.  We're now working with NCSA on a
similar merging of netCDF-4 with HDF5 by providing an enhanced netCDF
interface using the HDF5 file format:

  http://www.unidata.ucar.edu/proposals/NASA-AIST-2002/abstract.html

> I'm now in touch with a guy who wrote an HDF-EOS5 to netcdf converter,
> which is also not going to work for me because MODIS files are HDF-EOS4
> (apparently a different animal). I thought you might be interested. He
> (Bob Bane <address@hidden>) is going to post his converter on
> http://hdfeos.gsfc.nasa.gov/hdfeos/softwarelist.cfm in a few days, he
> says.

Yes, thanks, I'm very interested.  And this response will go in our
email support database so people who search for HDF-EOS will see it
also ...

--Russ