The GeoServer Team is pleased to announce the latest 1.6.x release: 1.6.0-beta2. We are hoping that our next release in this series will be a release candidate, so please help us out with testing as we work towards complete stability and reliability. This release should actually be quite solid, but it’s also got some major changes so we want another round of testing to be completely sure.

As for improvements, we’ve been making strides to allow GeoServer to play more nicely with Java Enterprise Edition containers. The biggest is that we’ve swapped out our connection pool code to be more Java EE compliant, enabling the use of JNDI configurations (though we still lack a UI for it), and defaulting to DBCP. This is a huge improvement over our old code, as it gives much more control over number of connections and can revalidate lost connections. It also again points to the strength of the Java open source world, as we’ve gotten a huge improvement by just leveraging an existing library.

This release also includes better logging options, giving users settings for the logging levels and where the output goes. Reprojection in WFS 1.1 is also working again, so GeoServer can now give you raw data reprojected on the fly. Other improvements include integrated demos for Google Maps, Yahoo Maps and Virtual Earth, dateTimes are working again, and there are numerous small bug fixes. The full log is on our JIRA.

Mick Wilson, who I had the pleasure of meeting while I was doing a fellowship in Zambia, has been doing some great blogging lately about the progress in building a Spatial Data Infrastructure in East Africa.  His latest entry talks about a hands on training they ran, setting up GeoServer with PostGIS, uDig and qGIS.  It sounds like it was quite successful, and it’s really great to see GeoServer being used to help build capacity and give people the ability to share data.  We’re hoping to get feedback from his group and others on how to run such a training day, as it sounds like a great way to share skills and spread knowledge about GeoServer and the SDI stack around it.  If you’re interested in running your own training day or have done so in the past, please get in touch.

I’ve just uploaded a new GeoJSON plug-in for GeoServer, enabling JSON output for any WFS request.  The plug-in is definitely compatible with all versions of GeoServer 1.5.x, and will likely work with 1.4.x as well.  This was a breeze to write, which is a testament to the nice design of GeoServer and to the strength of the Java open source world.  I made use of json-lib, and extended the JSONBuilder to handle geometries.  Then hooked that up as an output format plugin for WFS, and we can now plug it in to existing GeoServer installations.  Please give us feedback on how it’s working, and if we get a positive response we’ll likely include it in the default release.

The plug-in can be downloaded from sourceforge, and there’s a bit of a write-up in the GeoServer documentation.  For more information on GeoJSON, see http://geojson.org.  The spec is not yet finalized as 1.0, but is a first release candidate.  So hopefully it will hold up, but if not we’ll definitely stay up to date with the latest versions.

The long awaited first beta release of the 1.6.x series has just been uploaded to SourceForge, go to the download page to grab your copy. The main thrust of this release is full support for WFS 1.1, completed as part of the CITE thread in OGC’s OWS-4 testbed. A number of other cutting edge features are also landing on the 1.6.x series, many getting backported to 1.5.x, but a number only available in this release. The one’s that will make it on to 1.5.2 include many small improvements to our KML support, along with GeoRSS output. Both also make use of our new templating system to customize output of the pop-ups. We’ve also got a new Google Maps demo, some nice syntax highlighting for SLD editing, and some improvements on the sample request screen (thanks to Jubal Harpster for the suggestion).

As for features that will not make it back to 1.5.2, the big one is our first beta release of the Versioning WFS we’ve been working on. It adds three new operations to the WFS – GetLog, GetDiff, and Rollback – to allow wiki type editing of maps. This deserves its own post, so soon we’ll get something up about our progress and how to try it out. The other big improvement is we’ve now got a really nice security sub-system called Acegi, fully integrated. We’re not really using it a whole lot yet, as we don’t have a good UI to add users and the like. Our first initial goal is to be able to authenticate any user who edits a versioned map, so we can store who did what in the database, allowing rollbacks on particular users. But it provides a framework to allow fine grained control over who can view and edit which featureTypes, and eventually even particular areas of the map. If anyone has funding available to advance this that would be great. The one place you can see the new security system in action already is when you log in to the web admin tool. There is now a check box to ‘remember me’, which adds a cookie so you never have to log in again, even after system restarts. This may actually be my favorite new feature, since I used to spend so much time typing in the passwords when debugging.

The final major improvement is increased performance of shapefiles, thanks to the Shapefile Renderer developed by Refractions for use in uDig. This is a great example of the virtuous circle of open source, where we can take advantage of their improvements. So shapefiles should be a bit faster in this release. So please download the release, try it out, and tell us what you think. We love getting feedback on where the pain points are, and hearing your ideas on how to make it better.

Mike Frumin, who has been actively working with GeoServer and OpenLayers the past few months, has just released the culmination of his work for the Regional Planning Association. In his write up on his blog he includes a number of pictures that were generated by GeoServer, and links to a really nice OpenLayers application, as well as the same information in Google Earth. Both made extensive use of GeoServer, with some custom processing for a couple things that were pushing the envelope (that we hope to completely support in GeoServer eventually), including KML customization and SLD extensions. Read his post, it’s a great demonstration of using maps to tell a story, and we’re excited that they’re powered by GeoServer.