Supporting the Learning of Recursive Problem Solving

This research is about the problem solving activities of novice programmers as they learn to create recursive LISP programs. Their problem solving not only includes the issue of mental models, but also how to use these mental models in conjunction with other problem solving techniques. In fact, at various stages of their learning, learners seem to use different packages of problem solving methods. Each of these packages we call a mental method. In this paper, we discuss the PETAL learning environment which assists learners in the use of three of these mental methods: the syntactic method, the analytic method and the analysis/synthesis method. PETAL externalizes each mental method through its own customized interface, called a programming environment tool (PET). Such externalization helps learners internalize concepts, and organize relevant knowledge and generally leads to improved learning. The PETAL System itself is presented. Next we discuss a study where one group of students used PETAL and another group used a standard LISP environment as they learned recursion. The quantitative and qualitative data from learners' protocols demonstrate that the Petal group performed better than the other group during the learning period and also demonstrated higher achievement on a written post test. The paper concludes with a discussion why PETAL may be responsible for improved learning of LISP recursion.