Paul Ramsey's Thoughts on FOSS4G 2013

FOSS4G 2013 in Nottingham has come and gone, so it’s time to collect my thoughts about how the event reflected on the open source geospatial community.

• Local colour and character. The Nottingham organizers added a wonderful local flourish and character to the event. It was not just FOSS4G 2013, it was very clearly FOSS4G 2013 in Nottingham. The enthusiasm was infectious and transmitted to the attendees, making for a very friendly and high-energy event throughout.

• Open source mandates. The UK is amongst a number of European countries that have adopted “open source first” IT procurement mandates — as our work with Ordnance Survey attests. While the mandates generally include large “special features unique to proprietary” loopholes, the very existence of the mandate has to be shifting the mindset of managers and architects preparing new systems designs.

• Open source at the edges, moving inexorably inwards. If there was one pattern that repeated for every public agency I talked with, it was that open source adoption starts at the edges of the enterprise, in web publication and internet access tools, and then slowly works its way inwards, first by fully conquering the edges, then slowly taking over roles in the data production and management areas as old systems age out and are replaced.

• The desktop is dead, long live the desktop! The general IT trend of moving functionality onto the web continues, and open source geospatial is taking advantage of new web features like WebGL, HTML5, Canvas and so on to make web-based workflows more seamless and interactive. OpenLayers3 fits right into this trend. Simultaneously, the QGIS community has never been more active: a troop of over 30 developers held a code sprint in Brighton the week before FOSS4G, then bussed north to Nottingham for the conference. As QGIS jumps the 80/20 feature/functionality hurdle for GIS desktop software it is becoming very easy for administrations to pare down their proprietary desktop licensing to just a handful of special case installations.

• Git, GitHub and GeoGit. The majority of open source geospatial development now uses git for source code control, and GitHub as the internet site for sharing and synchronization (yes, PostGIS is an exception to the overall rule so far). The wide knowledge of git features and workflow in the crowd leads to excitement when people demonstrate git-related insights, like: GitHub can manage and publish small GIS data sets in the same way it manages code, as versioned text files; the computer science underpinnings of git (full snapshots tracked with acyclic graphs and identified with SHA1 hashes) can be applied to managing large collections of GIS data, a concept being proved out by GeoGit.

• Generational tipping points. I’ve long thought that open source adoption would only really pick up speed as old-guard decision makers simply aged out and retired, replaced by a new generation comfortable with open source development principles. The youthful faces at FOSS4G no longer belong exclusively to students and start-up hackers, they belong to employees of public sector administrations, some of which sent substantial delegations to FOSS4G.

• One-way valve. So far, my core theory of open source domination remains intact: while open source adoption may occasionally (or often) be slow, it’s an inexorably one-way process. Organizations increase their use of open source over time, but they rarely (ever?) decrease it.

Reprinted with permission from the Boundless Blog.