[bcbrown]

In general, I agree with you. But one counter-example would be my experience with Hive in Spring 2013. Our team of four had to ramp up on a very aggressive deadline without knowing anything about Hive. One of my colleagues picked up the O'Reilly book on Hive, and it was by far the most extensive, comprehensive, and convenient reference. The online resources I found were either outdated, incomplete, or would take much more time to find the answer I was looking for.

Another example would be several months later, when I got a job at a company with a lot of Hadoop work, not having much experience with writing MapReduce code. I had two weeks between jobs, so I picked up 4 books on Hadoop. It's less than a day's salary, and allowed me to feel like I could hit the ground running at the new job (and I did). Those books are all outdated now, but I still feel like I got my money's worth from them.