Geodata Committee Seventh Meeting
Chair: Chris Schmidt Minutes: Jo Walsh IRC: #osgeo channel on irc.freenode.org Date: 2006-July-20 (at least in the Americas and Europe) Time: (fixed time)
This is intended as a fairly low-key and relaxed meeting - checking up on the status of our repository project - and a random brainstorming discussion about interfaces to it that was postponed from last week.
- Checkup on telascience status for geodata processing - see also SAC Service Status
- Any messages to put in a pamphlet for OSCON?
General discussion postponed from last meeting
- Some type of browsing interface for the data neccesary:
- OpenLayers for Raster/Vector data? Easy to set up, might not be full featured enough for some people?
- Perhaps Ka-Map's tile.php as cached layer to feed both kamap and openlayers?
- Single-image screenshots for Raster data as overviews
- Vector dataset descriptions: What various attributes are/mean (better metadata)
- GeoRSS out to get updates, with bounding boxes -- need to get some way to display this data, does anything do GeoRSS rendering with bounding boxes right now?
- Feel free to add...
In attendance: Chris Schmidt, Jo Walsh, Schuyler Erle, Frank Warmerdam, Dave Sampson, Perry Nacionales, Chris Holmes, Allan Doyle. Several of these people are also in a very dry OGC meeting about Catalog Services.
Due to the hard work of Howard Butler and John Graham we now have hardware at telascience whose very high bandwidth we can make use of for establishing a Geodata Repository, one of the core elements of the Geodata Committees's mission. See also SAC Service Status
Chris Schmidt has kindly volunteered to be a coordinator for this effort - getting software set up and helping others to install what they need. Chris is thinking about the interface for exploring and discovering geodata. The repository will also act as a showcase for the web-facing parts of the OSGeo+related software stack - Mapserver, Geoserver and OpenLayers in particular. (Although none of these are "official" OSGeo projects).
Chris outlined his sense of requirements: "We outlined requirements: 1: Works, 2. Is easy, 3. Stores Metadata, 4. Stores files, 5. Can be fixed by someone if I get hit by a bus". Those people wishing to help with this admirable effort are asked to contact Howard Butler about being added to the LDAP database which will grant them shell access to the telascience systems.
http://geodata.telascience.org/ is now pointing to the system we have to provide hosted services on. Jo will also direct CN to point http://dev.geodata.osgeo.org/ at this host. (I'd like to have find.geodata.osgeo.org pointing here too.)
Geodata Metadata Requirements have been described at previous geodata meetings. This is an effort to balance between FGDC metadata compliance, and a level of detail which won't put people off. Jo's implementation of this (in progress) can be browsed at the Geodata Committee subversion repository
Further discussion of data licensing is postponed for now; we agree to focus on public domain data sources, and worry about more complex licensing issues when they arise. (There are some unresolved issues re. ESRI's non-commercial use copyright of elements of VMap0; we are trying to get a more definite word out of the ESRI/NGA about this).
An important part of what the committee can do is to network efforts, both by OSGeo members and others, who are working in this problem space - providing 'safe space' for people who are taking an implementation driven approach to repository/discovery to talk.
Chris Holmes brought up the question of to what extent OSGeo can offer a republishing / tiling service for datasets being maintained by members: "OSGeo could point ka-map/squid/custom caching at their WMS, and point clients at their tiles.". This would fit well with telascience's goals. We need a minimal "catalog" for this (Jo is planning on building this onto the metadata store, using OWSlib).
We plan to have a neat service to demo and publicise by FOSS4G (mid-September); with a service in place in basic form by mid-August, including as much as possible of the following:
- Upload tools for people to upload data in the form of files
- Annotation of those uploads as they happen so we know what the tdata is
- Browsing of the uploaded data in some way that is useful
- Hopefully some tools which work against the uploaded data to display it in a more useful way -- which starts heading towards the earlier described Utopia