Some of the R authors (including Ross Ihaka, one of the R founders) had a similar project, but the home page seems half deserted (the mailman link leads to a 404), and, looking at the waaayback machine, it didn't get a lot of traction (~20 messages on the ML at it's peak in may 2006).<p>It looks like this project has a sounder base, though. Omeaga was supposed to be it's own language, and it's pretty hard to design one properly. Clojure + Parallel Clot sounds like a clean and powerful backend for this kind of platform.<p>And Clojure-the-language is a good fit too. Incanter's use of the Clojure syntax for matrices "literals" is very clever IMO.<p>I think there might be traction towards this, from both the statistical and the scientific computing communities.<p>The R architecture is dated. Matlab is expensive, parallel code requires a custom toolbox, with a per core licensing scheme (and the language is clunky).<p>The main competitor for this project would be the Python-based Sage and NumPy, but I think that the JVM backend of Incanter makes it a better tool for this job.