This is pretty meaningless -- there's already a THRIFT interface which allows easy job creation and control as well as HadoopStreaming which allows access to cr

This is pretty meaningless -- there's already a THRIFT interface which allows easy job creation and control as well as HadoopStreaming which allows access to creating map-reduce jobs for anything using stdin/stdout. This has dubious benefit, and just adds another unnecessary layer into this process. I'm not sure why this is news.