Evaluation of algorithms to support novice programmer

Abstract

We are interested in supporting learning to program. Problem solving and its translation into a correct algorithm is an important issue. Automating the evaluation of learners' algorithms is a means to face the large and heterogeneous number of learners. In this article we propose an evaluation approach for introductory course programming exercises. This approach is based on a comparison of learner solutions to a model solution. This approach allows us to detect the difficulties of the learners (missing competencies) and lays the ground to propose feedback and adapted exercises.