There are currently two versions available for installation. A stable (but little slow) version and a much faster development version (but maybe a little instable).

• v0.2.4: The "old" version running a local javadb for its Extendend Javadoc View. It uses a local JavaDB and is a little bit slow on startup, which causes Eclipse to hang for a few seconds when the plug-ins are started.
• v0.2.101: the newer (yet a developer-) version that ships with our new extended javadoc view and a remote database. The remote database will be continously updated with new data to continously grow the recommendations base. Consequently, this version requires you to have an Internet connection to work properly. Please note, that this version is yet not exhaustively tested. Please report any issues you observe using our issue tracker.

Both versions are currently available on different sites.

To install the current SNAPSHOT (v0.2.101) plug-ins for Eclipse 3.5 use the following Eclipse update site:

http://haendel.st.informatik.tu-darmstadt.de/updates/eclipse3.5/

To install the previous version (v0.2.4) for Eclipse 3.4 and 3.5 point you browser this Eclipse update site:

http://cage.st.informatik.tu-darmstadt.de/frameworks/core/

Please note, that the udpate site on cage will be deprecated as soon as the current development stream becomes stable.

When using Eclipse Code Recommenders you might experience a "freezing system" or Out-Of-Memory-Exceptions.To solve this issue you can change the Eclipse configuraton to request more memory on startup. Therefore modify the eclipse.ini file in your eclipse installation folder and add/update the following lines:

• -Xmx512m
• -XX:PermSize=64m
• -XX:MaxPermSize=128m

Future versions might not require this anymore.

• Eclipse 3.5 (win32, linux32, macosx32 or macosx 64)
• Java 5

Please note: we provide an version of Eclipse Code Recommender for Eclipse 3.4 on the update site. However, it will be updated rarely compared to the Eclipse 3.5 version.

A note on the os dependency: We use a native library for creating the recommendations. Yet, we support 32 bit windows and linux and 32/64 bit mac osx.