Hi again.
Here's one more item of background information for our GALEON telecon
on Wednesday. This one is a note from Carl Reed in which he indicates
the OGC Architecture Board has been discussing some of the issues
raised at the Unidata OGC Interoperability Day.
He also suggests a possible GALEON presentation at the WCS.RWG session
in Stresa. My own thoughts were for a presentation at the Coverages
session, but perhaps we need to provide an update at both sessions.
Please give it some thought before the telecon.
-- Ben
---------- Forwarded message ----------
From: Carl Reed OGC Account <creed@xxxxxxxxxxxxxxxxxx>
Date: Nov 15, 2007 3:04 PM
Subject: RE: WCS and WCS Plus
To: Ben Domenico <bendomenico@xxxxxxxxx>
Hi Ben .
Hope all is well. Have not interacted since the Boulder meetings.
About a gazillion emails cross my virtual desk each day. Among them
was a GALEON related email that spoke to a number of issues and
suggestions regarding WCS 1.0, including mention of a WCS plus. So,
thinking that this was a juicy topic of discussion for the OGC
Architecture Board, I put it on the agenda for today's discussion.
Turned out to be juicer than I thought as it brought to light a
variety of issues and like minded discussion in the OAB. Obviously,
the OAB is strongly behind the core-extension model. From that
perspective, one issue with WCS 1.1 (and WCS 1.0 for that matter) is
that the core - if one can call it that - is way to heavy and all
encompassing.
Anyway, the following is a synopsis, which is by no means binding, of
the OAB discussion. This topic needs additional discussion and perhaps
someone from GALEON can appraise the WCS RWG.
I "think" we came to this view:
WCS 1.1 / 1.2 combines changes to the actual WCS 1.0 interface "and"
changes towards a core-extension specification practice. This appears
to create concern in Galeon as to exactly what can and should be
implemented in the near future.
A historical view of WCS shows two distinct audiences, those who want
simple image subsets, and those who want to query and download complex
multi-dimensional coverages.
A core-extension pattern should work well for this. The core service,
access by domain to coverages, is well accepted. Two extensions can
cater to the two WCS audiences:
1) Raster Access, e.g. supports 2-3D domain subselect with a single
possibly not query-able range parameter, returning GeoTIFF or JPG2000
2) Galeon, supporting multi-dimensional filters, netCDF +
Support for this pattern by Galeon and others depends on getting to
these implementable, testable, "named" extensions as quickly as
possible, so that vendors can see a productive result besides
specification elegance.
"WCS" and "WCS-G" ?
Thoughts and comments most welcome.
Thanks and regards
Carl