I was very lucky to have been able to participate at FOSS4G in Seoul Korea this year as both a presenter and avid listener. While the first and only other FOSS4G I attended (in Denver 2011) was a huge eye opener for me from a technical perspective this experience was much different. In Seoul I felt a whole new sense of the open source geo communities strength and cohesiveness. In the past I have always approached these groups as an avid user where I was mostly concerned with technical details but this time was a little different. This time I came to the conference as a presenter, as an open source contributor, and with a strong desire to build relationships with this inspiring community. As a result I had a fantastic and productive time that will likely affect my path ahead in this dynamic and kinda crazy industry.
A Bit About The Venue
Seoul is a wonderful city of around 10 million people where ancient temples sit below modern sky scrapers. The local transit system is one of the best I have ever experienced making it ridiculously easy to travel all around the city in comfort. Seriously, their subway is cleaner than your computer desktop. The food is amazing as well. Everywhere you look are bunches of small restaurants offering different flavors of Korea. And finally, the people are incredibly nice. Many people speak English or at least can figure out what you need with a little effort. All and all I found this to be a great venue to bring people from around the world.
Sensors and the Internet of Things (IOT)
I had the great opportunity to spend time with Ray from SensorUp who is doing some great work with Internet of Things (IOT) which I think will be a huge part of geo-tech in the not so distant future. The basic idea of IOT is that there will be (or are) sensors broadcasting to a common network from which applications can access signals to provide more intelligent decisions/behavior based off of that live environmental feedback. Right now IOT seems to be in it’s earlyish stages but after seeing how connected, automated, and sensor driven so much of Seoul’s day to day activity is I think it wont be long before we are all working with sensor data in real-time to help add rich features to our applications.
* His business partner couldn’t make it because he was needed pushing their involvement in IOT through as an OGC standard!
Volker Mische demonstrated an amazingly simple and scalable distributed system of document stores with spatial support. It was pretty amazing to see how simple it is to distribute geo data across a cluster and then redistribute data across a changed cluster if a machine is added or removed. The ease of use and flexibility of this project will likely lead to more people deploying apps on advanced scalable architectures as the barriers to entry are removed.
The GeoServer team has been pushing lots of improvements to raster based performance and functionality including creating new supporting core components to replace old project dependencies. Andrea Aime was busy running around to all the many presentations he had throughout the conference. The list of new features is pretty big but here is a short snapshot.
- Faster, Faster, Faster, Faster!
- Clipping rasters by vector or raster mask
- Image overlay averages to combine images rather than just using alphas
- … I’m forgetting a lot but the videos will be out soon
Also, the Boundless Geo folks gave me a sneak peak into an upcoming GeoServer UI which looks awesome and much more useful than what we are all used to using. It takes a more map designer perspective where layers are styled in a map studio of sorts.
The PostGIS team is busy as usual making lots of improvements to the spatial database we all love so much. It seems like most of the new work has been done on the raster side of things where in general we will increasingly be able to do more crazy robust computation with less code. You’ll have to watch the videos for details because I didn’t write down the syntax details. Needless to say this team, and especially Paul Ramsey, is not slowing down and will continue to build incredible functionality for the open source community!
OpenLayers 3 is coming along nicely with lots of new developments in all kinds of areas. The pieces that stuck out to me the most include increasing support for all kinds of formats (including Vector Tiles and WebGL). They also mentioned that support for WebGL is currently limited to points but progress for supporting polgons and lines are in the works. I really liked their demonstration of advanced topological editing as it demonstrated how sophisticated OpenLayers 3 is while giving some indication that this library is going to kick even more ass in the future.
MapZen is doing great things with WebGL and data visualization in general but they stirred up some thought and conversation by talking about a global Gazetteer for geographic places. You may not think this is interesting but the difficulties in achieving such a goal range from both technical to social and if successful could affect all of us as geographers so pay attention. You can read more about it here.
Free vs. Open
While not exactly a technology or specific presentation there where numerous discussions about the difference between the meaning of ‘free’ vs ‘open’. While it may seem like semantics I was convinced that this does matter but is not so easy to measure. This is actually not a new debate. You can read about a long running discussion on the GNU website here for more details but in general you can think of free meaning the freedom to use software as you want while open simply meaning the ability to see it but not necessarily use it for your intended purpose. Some may argue that this is just a matter of what licence you use but the issue is more complicated. This discussion reaches deep down to the very core of what we do in the open source community and should continue to be discussed.
The importance of OSGeo, LocationTech, and fostering open source
Easily one of the biggest lessons I was reminded about at FOSS4G this year is that community is important. It’s very important. In fact, I think it’s so important that at some point it’s what makes or breaks an open source project. It was very clear that the work that OSGeo and LocationTech are doing to support open source groups is valuable to both developers and users. For those that don’t know, these groups provide resources, community, and support for projects to grow and thrive. Projects that are part of these groups get branded as stable and reliable which provides a mechanism for reaching beyond the open source community to show others that these tools are not just a basement project.
I gave a talk about how we implement ontologies and geo-ontologies to enhance data quality, data accessibility, and mapping but I was excited to see another group was also talking about ontologies. Specifically Hamish Mcnair and Paul Goodhue where talking about how they used ontologies to ensure data quality in biological surveys. It was great to see others using these techniques as they are proving to be very useful for projects we work on at TerraFrame.
Other points of note
- People have been pushing lots of data into Vector Tiles with reasonable performance (look for “Stuffing your vector tiles full of data”).
- Changing data on the fly in already loaded Vector Tiles is still really tricky.
- OL3-Cesium is a pretty impressive library for adding a Cesium 3D globe to applications based on OpenLayers 3 (look for “OL3-Cesium: 3D for OpenLayers maps”).
- Tile Reduce is really great for handling larger data sets in the browser (look for “Big data analysis with Tile Reduce and Turf.js”).
- OGC GeoPackage… Use it! It’s great.
- Huge point clouds in the browser with Potree (look for “Point Clouds in a Browser with WebGL”).
- Write more documentation! (look for “Everybody wants (someone else to do) it: Writing documentation for open source software”).
That’s about all I have for now. Keep an eye out for the session videos which should come out soon.
- A Programmers Guide To Cheating The Clock For Basic Climbing Fitness
- Creating a Node.js & PostGIS App on OpenShift