OpenGeo and Safe Software have been talking about working together to make life easier for users of FME and GeoServer. We’ve both been hearing more about organizations using FME to solve their data conversion challenges and then making the results available to the world using the OpenGeo Stack.
While many people are making things work with the software now, we figure that a few improvements towards tighter integration could be a big win. Our end goal is to enable FME Server and GeoServer to work together seamlessly. This allows each piece to solve the area they excel at - GeoServer in OGC standards and web output formats, FME at complex data conversions and translations. With a few key improvements the combined solution should solve the ‘Community Schema’ problem that Ron Lake recently brought up better than any software in world. In time the GeoServer community will definitely build the capability to handle the full community schema transformation natively, but integrating with FME should provide a transitional path, and FME will always be ahead in terms of the most advanced translations.
If you are a user of both FME and GeoServer we’d love to hear from you with input on how we could work together to make your life easier (or you can just give us encouragement ;). We are still in the early stages, hoping to put together a rough prototype relatively soon, but beta testers in the future will be appreciated. And of course additional funding will enable us to prioritize the work and get it done faster. If you are interested in helping out please get in touch at the OpenGeo contact form.
The GeoServer team is proud to announce the release of GeoServer 1.7.0, available here for download. This is a very big release for GeoServer as the 1.7.x series brings some exciting new features and improvements.
Security in 1.7 has been improved by allowing access control at the layer/feature type level, in addition to the service level. This allows users to secure data in a much more granular way. There has also been much work done to better performance for GeoServer 1.7. The short of which is faster rendering with WMS, as well as faster data access with Shapefile, PostGIS, and ArcSDE. For those interested in the specifics check out the presentation given by Andrea Aime at FOSS4G this year in South Africa. Also of note for this release is a retrofit of the built in OpenLayers map preview. The preview interface now allows for filtering and controlling various rendering aspects such as anti-aliasing and image format.
The folks from GeoSolutions have also been quite busy as usual and have once again made some great improvements on the raster/coverage side of things. GeoServer now supports additional raster formats such as ECW, MrSID, and JPEG 2000. This has been achieved by leveraging the GDAL library. As well as additional formats GeoSolutions has also provided the ability to perform much more powerful coverage portrayal via the SLD RasterSymbolizer construct. With support for RasterSymbolizer users can now do channel selection, define color maps, perform contrast enhancement, and more. Special thanks to GeoSolutions.
As the GeoServer community expands, so does the number of translations to support other languages. 1.7 brings a Dutch and a Russian translation. Thanks to Leon Vanberio and Maxim Dubinin for their great contributions. And special thanks to all users who helped us test the 1.7 release candidates and filed bugs in the tracker.
Not only is GeoServer 1.7.0 a big release for users, it’s also a notable one for developers. Some major improvements have been done to the GeoServer core which will act as building blocks for features in future releases. This involves a change to the internal feature model which now allows the modeling of complex feature relationships; something that gets us closer to achieving full community schema support. There have also been changes to the back-end configuration subsystem which now has a much cleaner API for developers writing plug-ins and user interface components. It will also serve as the base for supporting additional persistence mechanisms such as hibernate and other O/R mappers.
And as usual a heap of bugs were fixed for this release. Over 200 issues were handled for 1.7.0.
Download, try it out, and let us know what you think. Comments and feedback are always welcome on the mailing list. The community is always interested to hear how people are using GeoServer. Also stay tuned for the 1.7.1 release, due out in a month or so, which will include some exciting new improvements to Google Earth and KML support.
I’d like to quickly congratulate the uDig developers team for releasing the newest version of uDig. uDig 1.1 has a venerable history. Quite a venerable history, as releases go. In fact, I think uDig might be in line for some kind of Guinness-style award for having the most release candidates (with 14 RCs and even 3 Sanity Checks). I certainly don’t mean this as a dig (sorry, couldn’t help myself); it just shows that the team wanted to be sure to release a quality product as free of bugs as possible. That said, the GeoServer team does feel slightly better about our growing number of release candidates (currently at 1.7.0-RC4). No matter. uDig is an excellent desktop GIS product and is a great complement to GeoServer. (I personally have always appreciated its ability to generate SLDs quickly and painlessly.) Go check it out.
**Update: **I’ve been asked off-camera about uDig’s editing capabilities. uDig does in fact follow the WFS-T(ransaction) standard, and is one of the best clients out there for doing so. In this way, one can use uDig to edit features served using GeoServer from the safety of a familiar GUI. That’s big news (shapefiles and other data structures can be rather opaque to edit otherwise) and I think it merits a mention here.
There are some questions that are asked quite frequently by people who are using GeoServer, or who are thinking about doing so. And certain ones can’t be so easily answered as those in our FAQ. One of the most common I see is this:
_“GeoServer is nifty, but is it fast?” _
Clearly some benchmarks are needed. Testing was first done at a presentation at FOSS4G 2007 by Justin Deolivera (OpenGeo) and Brock Anderson (Refractions Research). They chose to compare performance of GeoServer against MapServer, another popular open-source GIS. The presentation was well-received and showed off the strengths and weaknesses in both servers.
This year, at FOSS4G 2008, Andrea Aime (also OpenGeo) took over the benchmarking process, this time with an updated test suite including thematic mapping, anti-aliasing, raster data, and tile caching. Since the previous year, MapServer had improved its shapefile rendering to be faster than GeoServer’s render time. But Andrea continues:
_“The hard part started when our results showed GeoServer being in the lead in both PostGIS and raster tests. I could buy the PostGIS results, but I did not believe the raster results. MapServer has had long-standing and well-reputed raster support, so how come the newcomer was doing better?” _
Andrea and Justin worked directly with Steve Lime, Paul Ramsey, and Frank Warmerdam to confirm and discuss the findings. But the results were clear: although MapServer was faster at shapefile rendering, GeoServer was faster at raster and PostGIS rendering.
_“When you look at our results, remember that the MapServer developers are already working hard to improve MapServer performance, just like they did one year ago with the shapefile results. GeoServer developers won’t be sleeping either, as we’re already working on some changes to get PostGIS support both more secure and faster. _
_For next year, we’re looking forward to doing a joint presentation that will allow each group to tune their respective servers to the best of their capabilities, and look into some extra tests. I’m looking forward to it; a bit of friendly competition is benefiting both servers and keeping the audience interested.” _
Sounds like a friendly throwdown to me.
Today we have a new release candidate of GeoServer 1.7.x for everyone’s testing pleasure. We don’t wish to invoke Zeno’s Dichotomy Paradox, but we are in fact very close to the release. In truth this release candidate only includes six new bug fixes (the most notable regarding projection persistence and Google Earth overlays). If it seems like we’re going release candidate crazy, rest assured that it is only because we want to ensure that 1.7.0 will be a stable release and not a beta that was rushed out the door. Give it a try!