30 Oct

Venturing into Open Source GIS: A Global Conference Comes to Our Doorstep

David Howes, David Howes, LLC & Matt Stevenson, CORE GIS

This article was first published in the Washington URISA newsletter The Summit, Issue 36, Autumn 2014 (pages 7, 22-24) and is reproduced here with permission.

A Local Opportunity
In early September, the GIS community of the Pacific Northwest was fortunate to have the largest global gathering focused on open source (OS) geospatial software close at hand as the tenth annual international FOSS4G (Free and Open Source Software for Geospatial) conference took place in Portland, Oregon. Since it was last in our vicinity in Victoria, BC in 2007, the conference has moved as far away as Cape Town, South Africa, Sydney, Australia and Nottingham, England, an indication of the truly international nature of the event and the supporting organization. Not surprisingly, therefore, of the 870 attendees who gathered together this year, 23% came from 36 countries outside of North America and many of those overseas visitors were key figures in the worldwide effort to develop new and innovative open source GIS capabilities.

Seattle’s Open Source GIS Community
The Seattle area is home to an active open source GIS community who assemble frequently as Cascadia Users of Geospatial Open Source or CUGOS and Matt and I have often attended their meetings, typically coming away impressed by the nature of the projects and the interactions of the members. At first, we tended to focus on the clever technological accomplishments, but the way the community operates soon revealed itself as being just as important, if not more so, when compared to the technical output. With their colleagues around the world, this isn’t a group that’s trying to sell you something, it’s a group that’s interested in making the world a better place by working together to come up with alternative solutions to those that many of us have grown accustomed to for many years. These solutions are maturing to the point where they can now serve as the basis for GIS operations at all levels from the individual to the enterprise. As solo consultants witnessing the striking developments in the open source geospatial world, it makes perfect sense for us to take every opportunity to learn more, to enhance our capabilities and better serve our clients and to do our part to support the open source GIS community, hence our interest in FOSS4G.

Our Contribution to the Community
As keen as we were to just go to Portland and learn something new, we were also aware of our duty to support the community. We decided to begin by arranging a session for the 2014 Northwest GIS Conference entitled “Opening the Door to Open Source GIS,” with the idea of exploring the individual, organizational and cultural facets of the open source GIS world as compared to the proprietary GIS world. If we could shed light on some of these facets, it seemed to us that we could help GIS professionals make better decisions with respect to their adoption of technology, help them communicate better and help drive technological innovation. Of course, we can only do so much ourselves, but every little bit helps and, if nothing else, we can promote consideration of the sorts of wider issues that can impact our effectiveness as GIS professionals in ways that are often profound and far-reaching. The level of enthusiasm we saw at the recent Washington GIS Conference in Tacoma seemed to be indicative of a strong appetite for professional development activities of this nature and FOSS4G gave us a perfect opportunity to do some research to assess the value of our endeavor, gather input from a global community and build on our findings at the local level.

The FOSS4G Conference
Overall, this was one of the best conferences that either of us has attended, if not the best, for the quality of the presentations and activities and the nature of the subject matter and the attendees. We expected to witness high quality presentations and we did. We expected to meet a lot of really great people and we did. As a result, we came away thoroughly impressed and intellectually invigorated. One can’t really ask for more than that.

FOSS4G Portland 2014 Logo

A total of 42 workshops over two days run by many of the world’s leading experts in their respective subject matter was an impressive start to a volunteer-run conference of this nature and our attendance focused on administering PostGIS databases, QGIS plugin development using Python and designing beautiful maps with TileMill/Mapbox, Leaflet.js and CartoDB. Following the workshops, the main part of the conference spanned three days with a keynote presentation, an invited presentation and eight session tracks each day. Most of the presentations are available for viewing via the FOSS4G 2014 website and are well worth watching. Here we focus on two of the keynotes and discuss some general observations from other sessions.

Mike Bostock, New York Times and D3.js
In the first keynote presentation, Mike Bostock of the New York Times and author of the D3 JavaScript library for visualization combined an extensive and quite academic discussion on addressing projection representation issues with valuable advice for toolmakers. For example, it’s easy to overlook the fact that people don’t always use the tools we develop in the way we intended them to, but we can try to reduce the likelihood of unintended use cases by focusing on the smallest interesting problem. That involves breaking large problems into small ones and developing and communicating with respect to tools in as focused and explicit a manner as possible. The more clearly we can express what a tool is meant to do, the more likely it is that we can make the tool to do that and only that as well as possible. Furthermore, the less likely it is that a user will find themselves doing things that we didn’t intend them to do and that they didn’t necessarily expect to be able to do. This sort of advice may seem obvious, but it’s not hard at all to find many tools–in the GIS world and beyond–that could be improved significantly. In light of these comments, it makes sense that the most common enhancement requests and suggestions for GeoServer (an open source server for geospatial data) pertain to clarity and simplicity. Since they’re not bound by the sort of release cycles we’re used to in the proprietary world, the developers can move quickly to address such feedback and, thus, heed Mike Bostock’s advice.

Sarah Novotny, NGNIX
Sarah Novotny of NGNIX provided a heartwarming focus on the community aspects of the open source world. As she pointed out, developers of sophisticated software often find it much easier working with computers than with people, which can reveal itself, perhaps unpleasantly, in interactions on chatrooms and forums. For many, the idea of posting questions to an audience scattered around the world is rather intimidating. But, there’s no doubt that learning how to operate in such environments is well worth the effort. For your part, you have to develop a thick skin and establish yourself by helping others as much as possible (answering questions, providing bug reports, developing documentation and so on).

Sarah characterized a FOSS4G project as consisting of three main parts: (1) source code, (2) a license and (3) humility, respect and trust. It’s worth thinking about that for a moment. In the proprietary world we don’t have the source code (although changes are starting to happen), the license is something that’s often associated with large amounts of money and the latter is generally not considered. And yet, as we discussed at the Washington GIS conference, problems with communication (for which humility, respect and trust are critically important alleviators) can do far greater damage to a project than the technology ever could. If we don’t consider them at our professional conferences, where else will we consider them in ways that can have a dramatic impact on the success of our efforts as a GIS community?

Sarah also encouraged us to think carefully about how we go about developing open source solutions. “Build it and they will come” doesn’t really work. “Collectively deciding on a problem to be addressed and then addressing it together” is likely to be a much more successful approach because of the community buy-in from the start, further valuable advice for any situation, not just for open source projects.

Observations from the General Presentations
As one would expect, the topics of the general sessions at the conference varied widely from highly focused technical explanations to descriptions of practical applications and consideration of implementation issues. With respect to the latter, a number of presenters touched on the barriers to open source geospatial adoption, acknowledging that there is still a way to go in allaying typical concerns, such as the lack of a number to call if things go wrong or the resident expert moves on. The continued growth and success of companies like Boundless and AppGeo is, however, serving to lessen this type of concern and demonstrate that open source-based geospatial solutions can often compete with those from the proprietary world and, in many cases, win.

On the technical front, this really is the era of JavaScript, with ever-increasing interest in libraries such as Leaflet.js for web map development, D3.js for visualization and Node.js for server-side processing and networking applications. Geospatial JavaScript Object Notation, or GeoJSON, a format for encoding a variety of geographic data structures, is starting to be particularly widely used and is especially predominant as the basis for GitHub’s growth into the GIS world. GitHub’s platform for collaboration, code review and code management has become the industry standard for version control and publishing to support web applications. It follows, therefore, that GIS professionals are taking note of the ability to store geospatial data in easily understood open formats, to maintain versions of that data and to easily consume it through simple and efficient applications.

Aaron Racicot, a geospatial open source pioneer in our local GIS community, showed the value of a GitHub-based approach through a voluntary project for his home town of Langley on Whidbey Island. In short order, he was able to dramatically improve support for spatial thinking and analysis at very little cost. In front of a sizeable and enthusiastic crowd, he showed how he could use a point-and-shoot camera, a quadcopter and a variety of different software packages to produce quite impressive aerial photos and models of the town center, all for around $700.

Pulling together techniques and capabilities from many different sources, rather than buying into a single platform is a key characteristic of the way the open source geospatial community operates and it’s this sort of agility that explains, in part, the increasing interest from practitioners with more traditional proprietary GIS-based backgrounds. It’s perhaps not surprising that as packages such as GeoServer continue to evolve, they’re being implemented as an alternative to Esri’s ArcGIS Server, often in association with established spatial data storage and manipulation technology from PostGIS, which adds spatial capabilities to the PostgreSQL database. Two primary reasons seem to have emerged for this situation, one being the high cost of the proprietary option(s) and the other being dissatisfaction over quality issues and the time it takes for problems to be remedied. At this point, hybrid approaches, such as using proprietary desktop options and open source server options seem to be increasingly viable, as Ben Sainsbury of Oregon Metro demonstrated quite convincingly.

In Conclusion
For those of us with traditional and proprietary-based GIS backgrounds, going to a conference like FOSS4G is like stepping into a whole new world. It really changes your way of thinking about things and offers many new alternatives to the capabilities that you’ve become accustomed to and somewhat comfortable with. The trick, of course, is figuring out the best approaches for the tasks at hand, wherever they may come from. It’s easy to think of these best approaches in purely technical terms, but as we’ve expressed here and on other occasions, the open source world offers alternate perspectives on how we can work together and potentially enhance our effectiveness as GIS professionals, which is an admirable goal regardless of the specific technological solutions we choose to embrace.