This response from James Gosling on relational databases almost sounds reasonable:
In the enterprise space, things like Sandra and Baltimore and some of the NoSQL database. I’ve never got it when it comes to SQL databases. It’s like, why? Just give me a hash table and a shitload of RAM, and I’m happy. And then you do something to deal with failures. And you look at the way things like the NoSQL movement is. It’s various flavors of large scale distributed hash tables and trying to deal with massive scale and massive replication, and you can’t back up the database because no tape farm is big enough. And you find scale and reliability can fit together at the same time. So a bunch of those things are really cool.
Until you work on a real project, that is. I used to have pretty dismissive take on relational databases. Then I noticed (late in the game, sure, but sooner than Gosling, apparently) that people want reports.
Here's the thing: a non-relational database works if you have an isolated project, or don't have data that other people need reports on. If the data you have is of interest to anyone else in your organization, and you don't use a relational database? You just bought a world of report writing along with your oSQL database.