Modula-2 versus Java as the first programming language

Abstract

The choice of the first programming language in computer science curricula is a complex multi-faceted issue that needs to be addressed from multiple points of view. In this paper we adopt the perspective of the teacher, and after surveying factors such as language features, programming paradigms, curricular concerns and students' perceptions, we present extended results of the empirical assessment of student performance at our introductory programming course, where a switch was recently made from Modula-2 to Java, keeping the course structurally unchanged. Our main aim is to investigate the impact of the first programming language to the acquisition of essential programming concepts, techniques and principles. New results involving students' high-school GPA and performance at other courses, backed up by thorough statistical analysis, remove some potential sources of bias regarding students' learning and cognitive abilities, reaffirming the overall conclusion that the programming language used in the introductory programming course does not have a deep impact on the acquisition of essential concepts, techniques and principles of the imperative programming style.