In spite of the rapid pace of experimental progress in designing and running human computation and crowdsourcing applications, there is a lack of underlying tools and methods to systematically address issues such as efficiency, robustness and scalability in these systems.