In your original posting you mentioned a goal of creating better ISO
support for netCDF in Geotoolkit. I'm assuming that the focus is on
supporting local NetCDF files that geotoolkit reads and then maps to the
ISO metadata classes, is that correct?
If this is the case, I don't think that using UnidatataDD2MI.xsl is a
good initial reference point for your task as there are some
implementation details that are creating some confusion. ncISO has two
implementations, in the first it acts as a client to a THREDDS server
and accesses data/metadata contained in NetCDF via the NetcdfDataset
object (using OpenDAP) AND has access to THREDDS metadata. In the
second implementation, it acts as a service embedded within THREDDS and
again has access to the local file (or virtual dataset) and THREDDS
If my assumption is correct about your goal of providing better ISO
support for local netCDF files in Geotoolkit, then I would rely mostly
on the ACDD convention mapping located here:
The xsl mapping could then be useful in terms of looking at any
precedence issues, when for example, you have spatial and temporal
extent information coming from the NetCDF axis that conflicts with
information contained in the file attributes.
On 10/6/2011 9:18 AM, Martin Desruisseaux wrote:
Hello again Dave
Sorry to bother you for what may be a trivial question (I'm also
unfamiliar with XPath, in case my question is related to XPath syntax)...
On the following page:
For example we can read for the "creator_name" attribute a
"metadata/creator/name" THREDDS path. But I'm not sure how to
interpret this path. Is it a "name" attribute inside a "creator" group
which is itself inside a "metadata" group? What means the @ character
that we can read in some place? (for example
Later on the same page, there is a NcML example which contains
<attribute name="id" value="crm_v1" />
<attribute name="creator_name" value="David Neufeld"/>
<attribute name="creator_email" value="David.Neufeld@xxxxxxxx"/>
But this doesn't seem consistent with my previous paragraph:
* The group name is "THREDDSMetadata", not "metadata"
* There is no "creator" sub-group with a "name" attribute, but
instead a "creator_name" attribute.
file seems to give yet an other picture, since the path is defined as
(omitting the global attribute):
Which I interpret as "name" attribute inside a "creator" group, itself
inside a "creators" group (does it means that we can have many groups
named "creator" inside "creators" in a NetCDF file? I know we can do
that in XML, but is it allowed in NetCDF too?), itself inside
So I'm a little bit confused... Maybe my understing is not correct?
What are the THREDDS attributes we are supposed to look at?