A new mechanism will be defined for plug-ins to register themselves and to define their dependencies. This way, the Geomajas plug-in system becomes a plug-and-play system where the back-end automatically checks if all dependencies and requirements are met.
A new check will be built in that ensures backward compatibility between releases. The Hudson server will run this check on a daily basis, so the Geomajas team is notified in time when a certain commit breaks compatibility. This way such problems can be detected and fixed in time, to better guarantee backward compatibility.
Release date: 22/04/2010
Geomajas will be transformed into a fully modular framework with a clear definition of a public API and it's extension points. This improves the overall system design and increases clarity and consistency.The modular architecture also benefits extensibility because it allows for easy creation of additional modules (plug-ins).
GWT face: Google Web Toolkit client
In addition to the Dojo face, a GWT face will be added. This allows engineers to develop, test and debug applications in one single programming environment (Java). It also helps in establishing cross-browser compatibility, because GWT resolves many of the browser specific differences in behavior.
With 1.6 we will also introduce advanced security features. Access and restrictions can be configured and applied on a fine grained level. There will be support for both functional constraints (commands, tools) as well as data constraints (faces, layers, geographical extent, features, attributes).
System Configuration using Spring
Leveraging an industry standard framework, Spring, the overall flexibility and extensibility will substantially improve. The goal is to deliver configuration contexts for the server side, client side and security.
The Rendering Pipeline will enable developers to configure different strategies for caching and rendering of vector and raster data. Performance is realized through server side creation of SVG and VML, and on-the-fly rasterization. On top of this, a lazy feature loading approach will be supported to minimize bandwith and further enhance performance.