Work in progress — analyzing the gap between diagrams and code in computer science

Students in sophomore computer science (ldquoCS 2rdquo) are required to study the properties of a number of standard data structures; that is, common patterns of organizing data in a computer program. Typically, students are first presented with diagrams that graphically depict the data structure, and then shown sample code that actually implements it. We have observed, however, that there is a sizable gap between these two representations, and that many students who master the former have great difficulty translating that knowledge into the latter. We suspect that our pedagogy could be made more effective by treating diagrams themselves as formal entities, and providing students with a way of mapping operations on the ldquoeasyrdquo (pictorial) domain into the ldquohardrdquo (programmatic) domain. To help develop this technique, we carried out a semester-long experiment in which students demonstrated their understanding of the material both in diagrams and in code. The goal was to ascertain the kinds of mistakes that are often made, and how a technique like this could be most effective.