Abstract:  Entropy is one of the most interesting and confusing topics in physics. Gibbs himself said that the concept "may repel beginners." And John von Neumann is said to have told Claude Shannon to call his information measure "entropy, because nobody knows what entropy really is, so in a debate you will always have the advantage."

Of course, scientific debates are best settled by experiments. I will discuss what we have learned about entropy from experiments on colloids, systems of classical, interacting nanoparticles that can be seen under an optical microscope. We find that entropy has a complex relationship with order and symmetry, but one that can be understood within a simple framework that does not "repel beginners" -- so long as we allow that entropy is a measure of the number of states that we as observers /choose/ not to distinguish, rather than of those that are fundamentally indistinguishable. I will show how that understanding allows us to use entropy to design new materials and nanomachines.