Re: 19991130: XML support for ncdump(1), netCDF and XML

>To: address@hidden
>From: Bear Giles <address@hidden>
>Subject: XML support for ncdump(1)
>Organization: .
>Keywords: 199911300239.TAA25453

Hi Bear,

> Attached are a few source files for ncxdump(1), a program that generates
> a generic XML ("xCDL") instead of the standard CDL.  I won't work on the 
> corresponding ncxgen(1) program until I get feedback on the XML format, 
> although I don't expect there to be a problem.
> 
> In the long run, as everyone becomes comfortable with XML, both ncxgen(1)
> and ncxdump(1) should be able to handle domain-specific DTDs, but that 
> will require a complete rewrite of both programs.  OMF (Weather observation
> markup format) is an example of this type of domain-specific XML which 
> might be tied to netCDF files.
> 
> In any case, I've checked this DTD with 'nsgmls'.  While it reports a 
> few errors, they don't appear to interfere with that application properly 
> parsing the file.
> 
> Bear Giles
> address@hidden

Wow, thanks, that looks very useful.  I'm forwarding your message to a
few other users who indicated an interest in an XML version of netCDF
data.  I've made your sources and man page documentation available from:

  ftp://ftp.unidata.ucar.edu/pub/netcdf/contrib/xml.tar.gz

(Probably the easiest way to build ncxdump is to just untar these into
 the ncdump source directory and build ncxdump from there with
 something like:
    c89 -c -O -I../libsrc varxdata.c 
    c89 -c -O -I../libsrc ncxdump.c
    c89 -oncxdump -O ncxdump.o varxdata.o dumplib.o ../libsrc/libnetcdf.a
)

I tried it out with the "-c" option on the output of the RUC model the
way we decode it, and then made the resulting CDL and XML versions
available at

  http://www.unidata.ucar.edu/staff/russ/bs/metadata/ruc.cdl
  http://www.unidata.ucar.edu/staff/russ/bs/metadata/ruc.xml

for comparison.  I like the XML, it's quite readable and not that much
more verbose than the CDL.

Thanks again for making this available.  With an ncxgen as well, XML
tools could easily be used with netCDF data.

--Russ