As mentioned previously, FRUGOS (Front Range Users of Geospatial Open Source) held an "unconference" in Boulder, CO, today. I went along together with over twenty others, and it was a really excellent event. Many thanks to Sean Gillies and Brian Timoney for organizing things (insofar as an unconference can admit to having organizers!), and to Tom Churchill of Churchill Navigation for hosting us in a beautiful location right at the foot of the Flatiron Mountains in Boulder.
Having never been to an unconference before (as was the case for most people, I think), I wasn't sure what to expect, but an intentionally fairly random process of having people volunteer to speak, demo or lead discussions, with very minimal organizing of the agenda, produced a set of sessions that were of a higher quality than most highly structured conferences I have been to. We had some sessions with the whole group and some where we broke into two groups. As well as strictly open source topics, there were several sessions which talked about Google Earth and Maps, and I talked about general geospatial future trends in a shortened version of my recent GITA presentation.
Gregor Allensworth-Mosheh of HostGIS talked about his HostGIS Linux distribution, which comes with all sorts of open source geospatial goodies installed and ready to run right out of the box, with a nice set of examples. I have a copy of the CD and plan to give it a try when I have the time. Secondly he talked about "how to display 10,000 points in Google Maps", which I thought was great. He had two approaches, one which downloaded all the points to the client in JSON format (which is much more compact than other options like KML), and did all the processing on the client to combine multiple points which were close together into a single marker. The other used a WMS service which combined points on the server and rendered a raster image, but still allowed selection of points using a mechanism which went back to the server. Both these approaches overcome one of my pet peeves, which is displaying large result sets in multiple "pages", which I really think is a lazy solution which is not at all useful in most circumstances. For example, if I look at the photos I have geocoded in flickr, the result is split into 12 pages, so I just get a random one twelfth of the 1200 or so photos I have geocoded, with no idea what the real geographic distribution of the whole set of photos is. Come on flickr guys, you can do better than this!
Tom Churchill ended up the day by showing us his very cool touch table with an application which overlaid live video coming from a helicopter on top of a map with both imagery and vector data - it was very dynamic and very cool! This is just a side project for them, their main efforts are concentrated on producing a new generation of in car navigation system, which aims to do to that category what Google Earth did to online map display - make it much more dynamic and fun. You can get a flavor of what they are up to from these videos, but they really don't do full justice to what Tom showed us. I look forward to seeing how their system develops!
There is certainly a great energy about all that is going on with this new generation of geospatial systems right now, and it was good to meet a number of the people in the Front Range area who are making things happen in this area. I look forward to future FRUGOS events.