I'm getting patches from folks for new tests for Rubicon just about every day (many thanks!), so rather than generate release after release, I've made its CVS repository publicly accessible. You can get the latest version using: % cvs -d :pserver:anonymous / pragprog.com:/usr/local/repository login CVS password: guest % cvs -d :pserver:anonymous / pragprog.com:/usr/local/repository co rubicon Right now it fails on two tests on 1.6.2 (sort! on an array of length 1, and an arity test). I'm not too sure what the official word on what the correct interpreter behavior in these cases should be, so I've left the failures in as a reminder to ask... Dave