The 2.0 Release of Carrot2 brings a lot of
additions and improvements to the core framework, components and
applications. Major improvements include:

Brand-new web application
A brand-new
search clustering web application (see on the right) featuring a
more usable user interface, much faster back-end and better support for
adding new data sources.
-
Major refactoring and simplification of the core API, removal of
obsolete interfaces.
-
Attractive cluster visualization added to the standalone GUI clustering
tool based on the Aduna ClusterMap Library (see on the right).
-
A major speed improvement of the build scripts.

New visualization (WebStart)
A new look for the project website.
-
Adding an input component (carrot2-input-msnapi) for fetching search
results from the MSN Search Engine.
Smaller improvements include:
-
Autoloading of component descriptors. Now the majority of component
definitions can sit in their home projects and don't need
to be declared in the applications that use them.
-
Added support for incremental document pushing to the carrot2-input-lucene component.
-
Added support for incremental document pushing to the carrot2-opensearch-lucene component.
-
Added a simple GUI for configuring a Lucene index as a data source for
the standalone GUI clustering tool.