OGC/GALEON January 4, 2008 Teleconference Summary

(Ben's sketchy notes for starters)

Original Agenda Topics:

Participants:

• John Cartwright
• Aaron Braeckel
• Rob Weingruber
• Oliver Newell
• Greg Rappa
• Bruce Wright
• Peter Baumann
• George Percivall
• Stefano Nativi
• Lucia Lovison
• Ethan Davis
• John Caron
• Ben Domenico

Liaison with OGC WCS.RWG (soon to be WCS.SWG)

Peter summarized his suggestion that GALEON participate on a monthly basis in a special segment of the OGC Web Coverage Service Revision Working Group (WCS.RWG) telecon.  This segment of the telecon would be dedicated to GALEON-related topics.  Several important points were clarified.  First of all, the more GALEON participation in WCS.RWG the better, so GALEON is encouraged to participate as much as possible in all the WCS.RWG deliberations.  The special monthly telecon segment would be there to insure this minimum interaction between the two groups.   Several GALEON participants are already members of the WCS.RWG.  These include Aaron Braeckel, Stefano Nativi, Andrew Woolf, and Peter Baumann.  It was decided that the best approach to start with is to have rotating representation at the WCS.RWG GALEON sessions.  Here again, designating a representative for one telecon is not meant to discourage others from participating; it's just a mechanism for insuring that one GALEON person is there.

The WCS.RWG is in the midst of a transition to becoming the WCS Standards Working Group (WCS.SWG).  (Note that all active OGC revision working groups are going through such a transition), so the the Jan 23 was proposed as the first WCS.SWG telecon with a dedicated GALEON discussion.  This will give the WCS standards folks time to set up the logistics of the new SWG.  Peter will bring this proposal to the WCS.SWG.  Peter will contact the WCS.SWG with this proposal

GALEON Status Update

George mentioned two NASA initiatives with WCS components that might be good additions to GALEON.  It was not clear whether the GMU GALEON participants were tied to these initiatives.  George agreed to send an email with contact information about the groups so we can follow up on them.

Lucia indicated that her group at Harvard is working on an implementation of a WCS server and expects to have something running in the next month.  Likewise Aaron pointed out that his group at the UCAR Research Applications Lab is developing a WCS 1.1 server.   Lucia and Aaron will post information about their efforts to the GALEON list.

Conference Participation

Ben mentioned that upcoming conferences present an opportunity for GALEON-related presentations and face-to-face discussions:

• OGC TC in St. Louis March 24 - 27 http://www.opengeospatial.org/event/0803tc
• EGU in Vienna April 13 - 18 http://meetings.copernicus.org/egu2008/

Stefano described the Earth and Space Science (ESSI) sessions that are part of the EGU agenda for the first time this year and indicate that GALEON presentations would fit very well into several of the proposed ESSI sessions.  The deadline for EGU ESSI abstract submission is 14 January 2008. Submissions can be made online at: http://www.cosis.net/members/meetings/programme/view.php?m_id=49&p_id=323

George suggested that GALEON discussions would fit into several working group sessions at the St. Louis OGC Technical Committee meetings.  The Coverages and WCS.SWG sessions are perhaps the most obvious, but the Earth Observations (EO) and Natural Resources and Environment (NRE) working groups recently joined forces and that would be another good venue for GALEON presentations and discussions. 

WCSplus Discussion

The WCSplus initiative came up in several places during the telecon and George asked for a clarification of WCSplus.  Several people described this initiative from different points of view:

Ben indicated WCSplus grew out of follow up discussions at the Unidata OGC Interoperability Day at the September OGC TC.  From this perspective, the concern that arose was that GALEON participants had developed working WCS 1.0 client and server implementations.  Because of added complexity, updating those to WCS 1.1 proved more difficult than anticipated.  Moreover, it appeared that WCS 1.2 was heading in a directions different from 1.1 in that it would have a specification for a simpler "core" with a number of extensions.  Hence the WCSplus initiative was formed to move in the direction of the newer specifications but to do so by starting with the working WCS 1.0 implementations and adding needed functionality incrementally.  The WCSplus focus is to be on grounding the work in actual working implementations in a step wise fashion.

Peter clarified that WCS 1.1 was actually an effort to spell out elements of WCS 1.0 that had been left unspecified or unclear.  The result of course was a much larger specification that appears to be more complex.  But in many instances, it simply describes components of the specification that were unclear in the 1.0 specification, but the additional detail is needed to insure interoperability.  Because of its lack of specificity, it is possible to implement WCS 1.0 servers that meet the spec but are not interoperable with one another.

Stefano pointed out a recurring theme in the Unidata OGC Interoperability Day presentations was the need for a general abstract model for coverages along with specific detailed encoding specifications for communities of practice.   Stefano noted that the ISO 19123 specification is useful as the general abstract model, but the problem is that the concept of the coverage is so general that it can be made to fit almost any conceivable data type.  But this sort of broad generality at the abstract level alone does not foster interoperability. But attempting to develop a detailed specification for WCS that encompasses all the generality of the abstract coverage model results in a very complex specification that is very difficult to implement.  Hence, there is also a need for detailed but limited specifications at the encoding level which should be consistent with the abstract model but should be developed bottom up from needs of communities of practice working to serve and access the datasets representative of their discipline via the specification.  Starting with these specific implementations, we can then work toward a more general specification.  This is what WCSplus is aiming toward.

George mentioned this was analogous to what had happened with GML and WFS.  The GML 3.1 specification was very general but nearly impossible to implement for many of the groups attempting to use it to serve data via WFS.   As a result, that community moved toward a Simple Web Feature Service specification using specific encoding profiles of the general GML spec.

Further Discussions and Conclusions?

Follow on discussions centered on:

• the relationship between WCSplus and GALEON.  Current thinking is that WCSplus is a part of GALEON as a specific, immediate goal, but that GALEON encompasses a broader range of topics -- for example whether it makes sense to serve coverages via other protocols.
  
• the need to get GALEON input into the specification of the "core" component of the next version of WCS.  Decisions have to be made on topics important to GALEON.  One example that generates considerable discussion is whether the WCS core coverages should be confined to 2D.  (E.g., many WCS participants are mainly interested in serving satellite and other Earth imagery and would rather not have to deal with implementing clients and servers that can handle full 4D datasets like those generated by weather and climate forecast models.)
  
• the need to develop of a set of specific goals for GALEON for the coming year.  Things have settled out now to a point where this can be done.  In fact some outline level attempts have already been made as a starting point.