Computer programmers have a wide variety of protocols to guide the creation of software, including top-down design, decomposition, procedural programming, and object-oriented programming. Debugging has traditionally been resistant to attempts to normalize it because popular theories of debugging have not matched the actual practice of debugging, which is more intuitive than cognitive.

The authors of this paper propose a remedy by applying information foraging theory to debugging to enable the development of better navigation tools for programming environments. This application of foraging theory is based on a hunter-prey analogy, where the hunter is the programmer and the prey is the bug. The key feature of the theory is a concept the authors call “information scent” and the intuitive perception of this “scent” that guides the programmer on the path to find the prey.

The usefulness of the paper extends beyond the articulation of the theory, because the authors tested the theory against the debugging practices of experienced programmers working with actual programs. The empirical study confirmed that information scent was more descriptive of actual practice than hypothesis processing; as a result, the authors were able to provide guidance for enhancing software development packages. The information foraging approach was more predictive of programmer behavior than competing approaches.