In this paper we report about a large-scale measurement programme concerning the static quality of student-written Java code. The goal of the programme is two-fold: we investigate what metrics are useful for measuring static quality in an educational setting, and we investigate what conclusions can be drawn from the measurement results.

We describe here an implemented small programming language, called Alma-O, that augments the expressive power of imperative programming by a limited number of features inspired by the logic programming paradigm. These additions encourage declarative programming and make it a more attractive vehicle for problems that involve search. We illustrate the use of Alma-O by presenting solutions to a number of classical problems, including α-β search, STRIPS planning, knapsack, and Eight Queens. These solutions are substantially simpler than their counterparts written in the imperative or in the logic programming style and can be used for different purposes without any modification. We also discuss here the implementation of Alma-O and an operational, executable, semantics of a large subset of the language.