The design tree: a visual approach to top-down design and data flow

Abstract

Top-down design, an accepted technique for program development in most teaching environments, is an integral part of the introductory computing courses taught at Carnegie Mellon University. Although this planning technique works well for experts, it's application among less experienced users is limited: many novices abandon this technique as soon as implementation begins, focusing their attention instead on low-level details. This paper proposes a solution to this problem: the Design Tree, a simple graphic tool used to represent the top-down decomposition of a problem. This tool can be used by instructors, as well as students, independent of the software used to build their programs. The Design Tree not only facilitates top-down decomposition of problems, but also help students decide what type of control structures and data flow to use in implementation.