Hello again,
It has been pointed out that my earlier note (included below) can be
interpreted as a criticism of WCS for not being capable of dealing with the
various categories of observational data collections we deal with in the GALEON
community. That was not my intention.
In fact, I believe the first order of business is for the GALEON/CF/netCDF
community is to come to agreement on extensions to the CF conventions that
apply to collections of observational data. John Caron has proposed a set of
such extensions. In parallel, we need to work with various OGC groups to
determine how best to convey those types of data. For my part, I plan to
continue working with the WCS Standards Working Group on a set of use cases
that capture the requirements of the GALEON community for serving and accessing
these observational data collections. The leadership of the WCS SWG has made
it clear that they welcome and anxiously await this sort of input from GALEON.
In American sports jargon, the ball is in GALEON's court.
My apologies for any misunderstandings my earlier note caused.
-- Ben
On Wed, Dec 10, 2008 at 4:07 AM, Ben Domenico <Ben@xxxxxxxxxxxxxxxx> wrote:
Hi Steven et al.,
Following up on an action item for me at the Coverages session at the
Valencia Technical Committee meetings, I'd like to mention two
GALEON-related items.
First, we are working on a WCS extension for CF-netCDF encoding. The
GALEON group made a a few valuable suggestion for augmenting the draft.
encoding spec that was circulated by Stefano Nativi to the GALEON email
list. Those changes are being incorporated into the next draft that will
be submitted to the SWG. It also seems the draft should be recast in terms
of the encoding extensions template that is being developed.
Second, the topic of non-gridded coverages came up at both the joint
session with SWE and the Coverages session at the TC meetings. In GALEON,
the issue of how to deliver non-gridded observational data is an important
question. The relationship to WCS is not entirely clear, but GALEON 1
showed the value of a simple WCS use case in which the client defines a 3D
bounding box, a time frame, and a property (temperature, pressure, etc.) and
gets back a coverage encoded in CF-netCDF. However, with WCS we are
currently constrained to coverages that are gridded at regularly spaced
points in some Coordinate Reference System. In the GALEON community, this
works for some forecast model output and some geostationary satellite
imagery. A key question facing GALEON now is how to deliver the wide
variety of collections of observational data via a similarly
straightforward request -- based on OGC and ISO standards. And
specifically, might WCS evolve (perhaps via extensions to the core) so that
it would a viable protocol?
For the true gluttons for punishment among you, I'm attempting to capture
some of these issues related to the collections of non-gridded data in an
airport weather use case at:
http://sites.google.com/site/galeonteam/airport-weather-use-case-and-standards
As you will note, this is a very rough compilation drawn from presentations
by several GALEON team members. Many standard specifications come into play
-- WCS possibly among them. As time permits, I plan to refine this document
into something more complete and coherent, but I think it captures the
salient issues even in its current form.
-- Ben