
While working on a project at the University of Waikato, I needed a workflow engine based on Java, that would allow me to implement my own components. After searching the web and a few initial tests, the choice fell on Kepler, an open-source, scientific workflow engine, released under the BSD licence.
Kepler is based on the Ptolemy II system, a mature platform supporting multiple models of computation suited to distinct types of analysis (processing sensor data, for example, or integrating differential equations). Even though it provides a graphical user interface, normally used for modeling, workflows can also be executed from command-line. What I really liked about Kepler, was the ability to run flows on grids and its inherent support for multi-core architectures.
Since the aforementioned project also needed some basic support for WEKA, I started out on creating WEKA actors (an actor is an atomic component in a Kepler workflow). In order to enable WEKA's GenericObjectEditor in Kepler, I needed to patch the underlying Ptolemy II system. My contributions ([[1]](/apps/trac/keplerweka/wiki/CustomQueryBoxesInPtolemy), [[2]](/apps/trac/keplerweka/wiki/CustomQueryBoxesInPtolemy2)) were surprisingly quickly accepted and integrated. Such an active developer community was another bonus for Kepler/Ptolemy II.
Since I thought, that other people might be interested in my WEKA add-ons for Kepler as well, I started up a project page (first on my university homepage and now on sf.net) to make these actors publicly available: [KeplerWeka] was born.
A final word on the license: all WEKA actors are, just like WEKA, released under the GNU General Public License (version 2).
-- FracPete