How Do Programming Students Read and Act upon Compiler Error Messages?

Abstract

The research on understanding how programmers read and comprehend error messages is still limited. Through an eye-tracking study, we hope to add to what is known about how student programmers read and act upon error messages. In this paper, we elaborate on the methodology that we plan to use in the hopes of eliciting feedback from the research community that can help us improve upon our research design. We propose the study to enable us to collect data about whether student programmers read error messages, how they parse the code to find the source of the error, and how students with different ability and confidence levels vary in the way they process the errors. We will be examining where the participants fixate and how they read through the code to find the error. We are also interested in comparing the scan patterns of students who are proficient versus less proficient as well as those with high and low self-efficacy. We hope that the findings will contribute to the discourse about how student programmers read and parse code for errors, as well as to the discourse about the design of better error messages.