A binary version of the Java based jKlustaKwik, rev 1.0 was released today. This version is now working to sort the test case properly.
The source code for a java version of KlustaKwik is now available as the jKlustaKwik directory on the trunk.
This initial version includes the option for initial sorting of a random subset of the data, resulting in significant speed improvement.
In general, this runs in about 130% the time of an optimized C++ compilation, but is multi-platform.
Two new versions released today.
R1.8 corrects problems in the line including saving state to use standard C++ headers.
"Subsetter" release incorporates Ken Harris's changes to speed up clustering dramatically by using a randomly selected subset of the data points.
This release makes the line-endings in the archive unix style, which should fix some compiler complaints.
Also changed building to use optimization and Altivec support on the MacOS X Codewarrior build.
KlustaKwik version 1.6 released today. This version providing BIC and AIC cluster splitting penalties and saving of cluster models.