Costs and benefits are recurrent issues that concern all the computing areas, including computer and software engineering. Mastery of optimization algorithms is essential in these fields, but their didactics hardly has received any attention. To fill this gap, the interactive system GreedEx was designed to support the active learning of greedy algorithms by means of experimentation. In this article we describe GreedExCol, a collaborative extension of GreedEx that complements its experimental phase with a discussion phase held by the students in each team. The contributions of the article are threefold. Firstly, we present GreedExCol, a CSCL system aimed at supporting collaborative discussion on experimental results of optimality for greedy algorithms. Secondly, GreedExCol was evaluated with respect to educational effectiveness, obtaining statistically significant improvements of the collaborative, experimental approach over an individual, experimental approach without the support of GreedExCol. Thirdly, GreedExCol was evaluated in the same two groups with respect to motivation, obtaining a statistically significant increase of implicit motivation for students in the experimental group. Overall, we present a medium-term effort for developing an innovative learning system and a comprehensive evaluation of its impact over the students. ß 2015 Wiley Periodicals, Inc. Comput Appl Eng Educ 23:790-804, 2015; View this article online at wileyonlinelibrary.com/journal/cae;