[pinhead]

I wonder how these would have compared to Spark http://spark-project.org/ Hadoop isn't really an ideal framework for this type of computation.