Earlier this week I did a couple of presentations for the Association for Geographic Information (AGI) Welsh Group, along with my colleague Matt and a couple of people from Dotted Eyes, another company doing a lot of work with Open Source here in the UK. I did an introductory presentation on open source and the OSGeo “stack”, and then one demonstrating the capabilities of Quantum GIS. You can see my slides here and here. Be warned, the QGIS talk was actually a live demo, so the slides are just emergency screenshots in case the live demo crashed and burned! (It didn’t crash and burn, which was nice!)

We had a short round-table set of introductions before kicking off the talks properly, and it was really interesting to see what people hoped to get out of the day. A lot of the people there were very keen to identify desktop GIS alternatives, whereas in the past the discussion has tended to be about the database and web components. People are now relatively familiar with using PostgreSQL/PostGIS, and Mapserver/Geoserver, and usually OpenLayers or something similar. This battle has been won, I think, whereas for desktops it’s a different matter.

It’s quite clear now that Quantum GIS is a perfectly acceptable dekstop GIS package. I don’t accept any of the FUD that still goes around about its capabilities or stability. The people at the AGI Cymru event were definitely impressed with what it can do these days. Some of them had tried it a few years ago when, let’s be honest, it wasn’t so good, and are now prepared to take another look, which is a big win. However, in the majority of cases there was concern about how you’d go about deploying a totally new GIS package in an organisation where users may have worked with their current package for decades or more. This, to me, is not an open source issue, though it’s often portrayed as one. This issue will be a problem every time a new piece of software is deployed, be it open or closed source, and indeed if a new version of an old favourite has a re-design (Microsoft Office, I’m thinking of you).  I’m not sure how we get around this really, other than by understanding the motivations and concerns that people have, and trying to work within those limitations. Softly, softly, catchy monkey…

So, what else did we talk about? Linked data, both consuming and producing. Strategies for deploying open source within an organisation. IBM’s advice on Open Source in 2006 and whether or not it still holds true (yes, mostly). A good day!