OpenAerialMap Future Directions Meeting 2009
On 12 November 2009, nine of us met face-to-face in Paso Robles California USA, and six more folks dialed in. Discussion lasted about 2 hours. This was an excellent meeting with lots of great ideas and contributions.
- David Bitner,OSGeo/Sahana
- Jessica Block, UCSD
- Don Brutzman, NPS
- Jon Crowley, STAR-TIDE
- Stefan de Konink, OpenStreetPhoto
- Schuyler Erle
- Cristiano Giovando, EC-JRC
- Todd Huffman, STAR-TIDE Strike Force
- Jeff Johnson, Open Solutions Group
- Don McGregor, NPS
- Hugo Meiland, Leiden University
- Jill Olen, San Diego
- Marc Pfister, ENPLAN
- Travis Pinney, Open Solutions Group
- Charles Schmidt, Terrapan Labs
OAM Reboot Technical Proposal: Catalog Quickstart
Schuyler first discussed his OAM Reboot Technical Proposal and the existing SVN repository at http://svn.openaerialmap.hypercube.telascience.org
Catalog server does seem like most important initial task facing us.
- We are building a Web application
- This will be front-end for submitting and retrieving OAM data
- Catalog capabilities will be cross-connected to other GIS server capabilities
- Hosting site must offer reliability, security, up-time, capacity and access
- Can we compare software alternatives to accomplish this?
- Application Frameworks: two candidates seem pertinent.
Functionality seems approximately equivalent. Each can talk to an SQL relational database on the back end. Might need a REST interface.
Might need to swap in a different application layer later, so it might be smart to quickly put up something workable as quickly as possible to learn lessons and reach conclusions, throwing away first-round application isn't necessarily bad.
Random Hacks of Kindness is a big opportunity. It would be cool to have a prototype catalog server up and running by next Monday for discussion purposes. Perhaps other versions will be built later using other application frameworks later for comparison - nothing wrong with that. Communications will be via irc://#firstname.lastname@example.org
Achieving Consensus, Initial Governance
Public discussion is REALLY IMPORTANT. Back-channel discussion is helpful for initially figuring out some things, but more critical is that the knowledge gets disseminated and considered by the group.
Achieving group consensus through open discussion is the most critical thing we can do. Posing questions, options, tradeoff alternatives, etc. means that people can understand what is going on. If we act with mutual trust and openness during these critical reboot efforts, we are setting the pattern for effective progress.
Figuring out who is part of the OAM developers and looking at other exemplars will help us figure out how to best establish longer-term patterns for effective cooperation and self governance.
Communities of Interest (with many overlaps)
- Geographic Information System (GIS) developers and users
- Remote Sensing and scientific users
- Emergency Management, support for first responders and disaster relief
- Web developers and users
Exemplars for group management
- Open Street Map Foundation
- OpenLayers steering committee
- Other organizations might also work as hosts for an OAM working group:
- Open Source Geospatial Foundation (OSGeo)
- Open Geospatial Consortium (OGC)
- Web3D Consortium
Project steering committee
- Keep coherent shape to many concurrent efforts
- Role includes ensuring that consensus is followed when possible, variations are certainly OK but should be documented and further discussed
- How to bootstrap? Mailing list self-nominations, initial volunteers for interim period, voting?
- Both coders and community domain expertise are needed for effective decision making
- Committee primarily guided by consensus, lack of consensus usually means more work is needed (and forced decisions often turn out to be ineffective anyway)
Our goal for OAM is to accept and serve open-source code and open-access data. Several important questions pertain.
- What open-source licenses are acceptable for code?
- What open-access licenses are acceptable for data?
- Likely different licenses for code and for data
- Attribution and redistribution requirements
- Can we prevent malicious misuse or repurposing of data?
- How do we verify identities of submitters?
- Can we embed or link license metadata within image files themselves (e.g. JPEG2000 or GeoTiff)
Enough is known about our licensing goals at this point that we can proceed with the prototype. Further discussion and documentation will be needed.
- Telascience server SDSU-UCSD - John Graham
- NPS - Don Brutzman and Don McGregor
- NPS is both .EDU and .MIL. The .EDU side is quite open, while .mil can be restricted.
- Open aerial map is international in nature. Our priority is open, international access.
- NPS has connections to CalREN, has ROCKS cluster. The cluster is not publicly exposed, but can act as a behind the scenes resource
- Other sites can also host the data
Goal is to create a distributed data model with multiple sites providing various degrees of caching. This is important discussion topic for this weekend.
- Updates might be via hard-disk (sneakernet) but direct connection over network needs to be possible.
- Also need to be able to download/cache/disconnect data on demand, in either modest or large quantities.
Data Model Considerations
There are many relevant data standards available. Nevertheless there is also a need to keep the number of _required_ standards to a bare minimum, so that minimum functionality can be maintained.
- Critical standards bodies and groups for geospatial metadata include
- What are critical data standards that we can start with?
- Support for GDAL: gdalinfo
- Dublin Core for document metadata
- Geography Markup Language (GML)
- License name, link
- Need to list others here, many exist
- Are there any that would help initial catalog-server standup effort?
- What metadata conventions does OpenStreetMap follow? (Reportedly no metadata requirements are used, need to confirm...)
- Additional data standards are OK but probably considered extra
- What are criteria for including a standardized set of metadata on the short list of required inputs?
- Might even want to allow well-formed but unstructured user-defined metadata?
Likely there are plenty more action items that will flow from all of the things we discussed. Here are some.
- Integrating these points into the OAM Reboot Technical Proposal
- Schuyler will post proposal about how to establish an initial interim Project Steering Committee
- What else do we need to consider?
- How will the massive amount of data be stored. The previous infrastructure couldn't handle all the available data now. How will OAMv2 address this hosting and storage (distribution) wise? --Skinkie 17:25, 12 November 2009 (UTC)
- License-based serving; if a user is only interested in CC-BY, it should only serve all compatible licenses under it and not CC-BY-SA (above it). --Skinkie 17:30, 12 November 2009 (UTC)
- Once an initial catalog server has been stood up, what do we do next? Another workshop meeting perhaps?
- Might a monthly teleconference/IRC meeting be helpful? Maybe in January...
This was a productive session today. A lot of productive discussion occurred. We hope this weekend's code-a-thon goes well!
Additional edits to the minutes are welcome to ensure that they are accurate and complete.
New points and ongoing discussion on these topics can best occur on the OAM talk mailing list.
Thanks to all participants, and thanks for considering these powerful possibilities.