This paper reports on the experiences of using an early assessment intervention, specifically employing a Use-Modify-Create scaffold, to teach first-year undergraduate functional programming. The particular intervention that was trialled was the use of an early assessment instrument, in which students had to use code given to them, or slightly modify it, to achieve certain goals. The intended outcome was that the students would thus engage earlier with the functional language, enabling them to be better prepared for the second piece of assessment, where they create code to solve given problems. This intervention showed promise: the difference between a student’s score on the Create assignment improved by an average of 9% in the year after the intervention was implemented, a small effect.

A Computer Science introduction course ought to focus on exciting students about the subject matter and on problem solving through the methodical design of programs. An effective way to achieve both is through the development of functional video games. As most students are interested in video games, their development adds an exciting domain to any introduction to programming. This article advocates that an exciting crowning achievement for students in such a course is the design and implementation of a multiplayer distributed video game. By exploiting a domain that is popular with students, they are taught about design principles, communication protocols, and pitfalls in distributed programming. This article puts forth a successfully used design recipe that places distributed programming well within the reach of beginning students and outlines the use of this design recipe in the classroom. For those teaching beginners, this article presents a model for developing their own distributed programming module. The success of the presented methodology is measured through student feedback on their distributed programming experience. The empirical results suggest that the design and implementation of distributed functional video games is effective and well-received by students. Furthermore, the data suggests that the presented methodology fails to exhibit the gender gap common in Computer Science and is effective regardless of the programming experience CS1 students bring to the classroom.

We challenge the idea that a course intended to convey principles of languages should be structured according to those principles, and present an alternate approach to teaching a programming language course. The approach involves teaching emerging programming languages. This approach results in a variety of course desiderata including scope for instructor customization; alignment with current trends in language evolution, practice, and research; and congruence with industrial needs. We discuss the rationale for, the course mechanics supporting, and the consequences of this approach.

The observation that program structure follows data structure is a key lesson in introductory programming: good hints for possible program designs can be found by considering the structure of the data concerned. In particular, this lesson is a core message of the influential textbook “How to Design Programs” by Felleisen, Findler, Flatt, and Krishnamurthi. However, that book discusses using only the structure of input data for guiding program design, typically leading towards structurally recursive programs. We argue that novice programmers should also be taught to consider the structure of output data, leading them also towards structurally corecursive programs.

Elliot Soloway’s Rainfall problem is a well-known and well-studied problem to investigate the problem-solving strategies of programmers. Kathi Fisler investigated this programming challenge from the point of view of functional programmers. She showed that this particular challenge gives rise to five different high-level solution strategies, of which three are predominant and cover over 80% of all chosen solutions. In this study, we put forward the Segments problem as an alternative challenge to investigate the problem-solving skills of functional programmers. Analysis of the student solutions, their high-level solution strategies, and corresponding archetype solutions shows that the Segments problem gives rise to seven different high-level solution strategies that can be further divided into 17 subclasses. The Segments problem is particularly suited to investigate problem-solving skills that involve list processing and higher-order functions.