I wouldn't be surprised if Hadoop (or another Map/Reduce implementation) becomes a sort of "assembly language" for big data, with higher-level abstractions buil

I wouldn't be surprised if Hadoop (or another Map/Reduce implementation) becomes a sort of "assembly language" for big data, with higher-level abstractions built on top. You can already start to see this happen with Hive and Pig.