Perhaps I'm just jaded from NSDI a few weeks ago, but I'm tired of papers about Hadoop and systems like it. If you're going to do work with Hadoop, at least sta

Perhaps I'm just jaded from NSDI a few weeks ago, but I'm tired of papers about Hadoop and systems like it. If you're going to do work with Hadoop, at least start comparing what you're doing with other work that improves Hadoop. There's been a bunch of papers about improving the scheduling or the shuffle phase or whatever but they all compare to vanilla Hadoop and not each other.