MENDONÇA, Andréa P.; http://lattes.cnpq.br/0938462047218130; MENDONÇA, Andréa Pereira.
Abstract:
Traditionally, introductory programming courses emphasize the development of logical reasoning and the construction of programs, more than any other skill. In theses couses, well-defined problems are resources used to assist students in their learning of programming. However, this teaching approach has a problem: it fails to expose students to a realistic problem solving process, involving the problem space (understanding the problem and requirements specification) along with the solution space (program construction). Thus, there have been too many reports in the literature concerning the
novice students’ difficulties in understanding ill-defined problems and requirements specification. To address the problem presented, we defended that introductory programming courses should (i) use teaching strategies that take programming as part of Software Engineering, and (ii) develop activities that enable students to go from requirements specification to the program by solving of ill-defined problems. Based on this, we have designed a teaching methodology called Problem Oriented Programming
(POP). POP is characterized as a set of activities that can be inserted to the introductory
programming courses. It consists of a cycle of problem solving that puts into practice typical activities of the Software Engineering, such as: requirements elicitation and specification, testing and programming. POP has been evaluated by means of case studies and controlled experiments. For the experiments, we have adopted a set of variables found to be most representative of the students’ performance when dealing with the problem space together with the solution space. In 83:3% of all observed variables, POP students exhibited far greater efficiency than non-POP students. To exemplify this: in both experiments, POP students documented, on average, over 79% of the requirements
that the non-POP students. Furthermore, POP students implemented four versions less of the program to meet all the requirements compared with the non-POP students. Despite the fact that the results cannot be generalized, the way we treated the threats to validity allow us to ensure the quality of our results.