Third Geodata Committee weekly summary

[Jo Walsh]

dear all,

Thanks to those who attended this week's meeting; I'm sorry I had to
run off rather suddenly at the end. Minutes are here:
http://wiki.osgeo.org/index.php/Geodata_Committee_Meeting_20060405
http://logs.qgis.org/osgeo/%23osgeo.2006-04-06.log from 07:02

As this week's discussion was a followup from last week's, when we'd
had confirmed news of hosting facilities available for OSGeo at
telascience, it's worth recapping on that here. 

At last week's meeting, the second, we had a very broad discussion of
what kind of data sets OSGeo members would be interested in hosting in
a repository. The repository discussion came out of telascience's
generous offer of hosting facilities; data collections will support
the software stack in offering demos and public interfaces on the
telascience systems. John Graham and Norm Vine, who are supporting
those systems, are particularly interested in better metadata
management processes, and in OSGeo helping to establish a workable
metadata standard both for well-known collections of geodata and
collectively contributed ones.

https://geodata.osgeo.org/servlets/ReadMsg?listName=geodata&msgNo=10
http://wiki.osgeo.org/index.php/Geodata_Committee_Meeting_20060329
http://wiki.osgeo.org/index.php/Geodata_Repository - collected after
the broad ground-staking discussion at that meeting.
http://wiki.osgeo.org/index.php/Geodata_Discovery_Working_Group#Public_Datasets_for_Possible_Mirroring_or_Distribution
- the list collated by Markus this week of what public geographic data
sets that are in the world now, we could beneficially consider hosting
at OSGeo in the future.

Having a basic repository will provide a 'seed project', to establish
basic infrastructure that can be offered in support of other future
projects that involve geodata+metadata collections. 

This week focused on deciding on a few specific data sets that OSGeo
can usefully host in a repository at Telascience, that will fit will
with software project members' existing needs and provide a good basis
for a showcase of the different OSGeo project's capacities and
interfaces.
http://wiki.osgeo.org/index.php/Geodata_Committee_Meeting_20060405#Data_Hosting
This came down to an agreement on VMap0, Blue Marble NG, and SRTM30
as the most useful core data sets for projects. At telascience we can
use Mapserver for a WMS/WCS, Geoserver for a WFS. 

There's also a lot of interest in offering bittorrent downloads for
geodata, complementary to the OGC standard interfaces, and Frank
offered to talk to the maintainer of http://geotorrent.org about whether
they'd consider expanding the very constrained list of file formats
they offer, or whether OSGeo would need to run its own tracker. 

We also talked over the metadata requirements that data sets will
need; both those that are well-known and in the public domain; and
those that are collaboratively contributed. Norm Vine insists on "no data
without metadata!" from the start. I agreed to work out a basic
metadata standard for us that cross-sects with existing standards, and
if possible do a bit of prototyping before that.
http://wiki.osgeo.org/index.php/Geodata_Committee_Meeting_20060405#Metadata_Standards

This all will need to be correlated with what's happening in the 
http://wiki.osgeo.org/index.php/Infrastructure_Working_Group that's
responsible for coordinating, under Norm's direction, what goes onto
the telascience systems. 

best wishes,


jo