RE: Thoughts on GALEON Phase 2

• To: "Ron Lake" <rlake@xxxxxxxxxxxxx>, <rhusar@xxxxxxxxxxxx>, "Ben Domenico" <ben@xxxxxxxxxxxxxxxx>
• Subject: RE: Thoughts on GALEON Phase 2
• From: "Woolf, A (Andrew)" <A.Woolf@xxxxxxxx>
• Date: Wed, 1 Mar 2006 08:26:45 -0000

At the most simplistic level, GALEON-1 has been about demonstrating the
use of netCDF within WCS. At a deeper level, it has been about layering
the 'coverage' meta-model (ISO 19123) over a range of netCDF-based earth
science datasets (WCS then provides a convenience API for accessing such
data).

Insofar as GML provides a rich standard language for constructing
explicit semantics, it has a key role within GALEON in the broader
sense.

Our view is that the main enabling contribution of the raft of TC211
(and OGC) standards is to provide a common reference model for semantics
- captured in the Domain Reference Model of ISO 19101. According to the
ISO program, you explicitly model key concepts of your universe of
discourse ('feature types'), and draw in the semantic bricks of the
other standards (and GML) to construct them.

This is precisely the approach we've adopted in CSML - it's an
application schema for "climate science"
(atmosphere/meteorology/oceanographic), with feature types and a GML
encoding. (Perhaps the least interesting bit is that which describes the
'construction' of feature instances from netCDF and other files.)

Being explicit about semantics also has implications for the use of
netCDF within WCS. It means, for instance, that you could use the
facilities already provided through the 'xlink' mechanism to minimally
wrap a netCDF file in a GML response. We have the following
(hypothetical) example in our GALEON report
(http://galeon-wcs.jot.com/WikiHome/Implementation+Progress+Page/GALEON-
NERC-report.doc):

<gml:RectifiedGridCoverage gml:id="netcdf_ID0">
        <gml:RectifiedGridDomain
xlink:href="http://myserver/myfile.nc#temperature";
xlink:role="http://netcdf/GMLOps/Grid/spatialDomain"/>
        <gml:rangeSet>
                <gml:File>
                        <gml:rangeParameters>
                                <gml:Quantity uom="degC"/>
        
<fileReference>http://myserver/myfile.nc#temperature</fileReference>
                                <fileStructure>Record
Interleaved</fileStructure>
        
<mimeType>application/x-netcdf</mimeType>
                        </gml:rangeParameters>
                </gml:File>
        </gml:rangeSet>
</gml:RectifiedGridCoverage>

Here, the 'xlink:role' attribute is used to indicate an assumed
well-known procedure for dereferencing netCDF Grid axes to a GML
domainSet. This - as well as having WCS provide netCDF as an output
format - is only possible if 'netCDF' has explicit well-defined
semantics that allow it to be interpreted. But it doesn't! Hence our
recommendation on CF-compliant netCDF as an additional WCS format.

GALEON-2 hopefully will continue to investigate the application of the
ISO/OGC semantics framework to earth science data in the broadest sense
- looking at the full range of governance issues and components
(registries etc) needed to support a scalable, interoperable
infrastructure. A number of related organisations (WMO, IOC, INSPIRE)
are beginning to do this. We look forward to contributing our
CSML-related thoughts and experiences, and agree the OGC Network
approach sounds appropriate.

Regards,
- Andrew

-----Original Message-----

From: owner-galeon@xxxxxxxxxxxxxxxx [mailto:owner-galeon@xxxxxxxxxxxxxxxx] On Behalf Of Ron Lake

Sent: 24 February 2006 21:45
To: rhusar@xxxxxxxxxxxx; Ben Domenico
Cc: GALEON email list
Subject: RE: Thoughts on GALEON Phase 2

What is the current role of GML in GALEON?

Cheers
Ron

• Follow-Ups:
  • Re: Thoughts on GALEON Phase 2
    • From: John Caron