What a difference 3 years makes - Cassandra 2.x, CQL, Java, Clojure, and Node APIs
It has been a few years since I had a need for Cassandra (I now have a pilot project for which I would like a scalable data store). I grabbed the latest Cassandra source from github and I have been looking at some of the code, and running it locally also with the CQL shell. I know that CQL has been around for a while, but this is my first experience with it. Huge improvement. I really like it.
I give up some tunable degree of consistency (the "C" in "ACID") with Cassandra but I can easily live with that in my project. For now "developer friendliness" is really important to me since I am just experimenting with some ideas but if I need to I would like to be able to scale without changing my code.