Hi all,
_From my point of view, this has been an excellent and productive discussion in
a topic area that I've found difficult to fathom since I became involved in the
OGC. But, it's been very far reaching and I'd like to propose that we break it
up into sub-topics that might fit better into different email lists and
discussion groups.
1. GALEON I.E.: Since the discussion started in the GALEON context, I'll start
with my idea for the GALEON topics that fit in as part of an interoperability
experiment:
-- Starting with the kinds of CF-netCDF coverages we worked with in GALEON 1,
does it make sense to represent those coverages in GML encodings such as
ncML-GML, CSML? Given that approach, is it practical to represent all the data
in the GML or should the payload still be binary encoded in CF-netCDF or
perhaps GML-JP2K? And given that approach, should the data be served via WFS
as well as WCS?
-- For collections of stations observation data taken over time (time series of
data taken at a large number of fixed points in space, such as weather stations
or river gaging stations), does it make sense to represent those as discrete
grid point coverages? If so, how should they be encoded and served?
To me these are important areas of practical experimentation for GALEON 2 and I
plan to put them onto the agenda for our next GALEON telecon.
2. OCEANS I.E.: The issue of the relationship between the coverages of GALEON
1 and SWE/SOS could be addressed as part of the Oceans interoperability
experiment. Certainly the output of weather forecast models is of interest to
the ocean sciences. Can the CF-netCDF encoded coverages of such datasets be
served via SOS to that community of clients? My respectful suggestion is that
the Oceans I.E. consider this as a part of the experiment. Since I'm part of
the Oceans IE, I'll see if there's time to broach the subject at that telecon
later this morning.
3. ARCHITECTURE?: Much of the theoretical, abstract level discussion of what
are the fundamental data objects we are dealing with and how the data and
metadata access protocols relate to one another seems to me to be a question of
overarchihng architecture and should be taken up there. But, since I'm not
part of that group, I leave it to others to decide how to mov e forward on
those issues. Of course, these abstract discussions will be grounded by the
practical experiences gained in the interoperability experiments.
4. OTHER: Aspects of the interaction have touchs on topics that relate to
coverages in general, catalogs, etc. and one hopes those facets of the
discussion will be carried to the appropriate working groups so that we can all
continue our focussed efforts while remaining aware of the bigger picture
issues of how our work fits with that of others.
Many thanks to all for a wonderful discussion. I hope it continues to flourish
as a set of lively interactiions on the sub-topics as well as a set of
practical implementations and experiments for each of the key components .
I should add that I have no real objection to continuing the interactions here,
but I'd like to make sure that some of these specific topics be taken up in the
appropriate groups.
-- Ben
