Yep, MapReduce isn't a relational database. So what?

About a week ago, MapReduce (and by extension, Hadoop) was broad-sided by some database guys in a rather hyperbolic post: MapReduce: A Major Step Backwards. The MapReduce community seemed quite taken aback, mostly because the attack came out of nowhere and amounted mainly to "MapReduce isn't a relational database, therefore it sucks." I'm not sure how one follows the other, but ok, whatever normalizes your data.
Even though our Hadoop project is over (I'll get to posting the final matrix multiplication code to Google Code soon), and I don't know when I'll get the opportunity to do more MapReducing, I've grown quite fond of this particular parallel processing paradigm that could. Luckily, Mark Chu-Carroll over at ScienceBlogs, has written a very well though-out and funny rebuttal to the rather odd charge of one tool not being some other tool.
Incidentally, if you're not reading Mark's blog every day, you should be. He'll make your brain bigger.