WAHESE 2015 - Workshop on Applications of Human Error Research to Improve Software Engineering

The workshop theme is to investigate the usefulness of employing the perspective and contributions of human error research to understand software engineering errors and to improve software quality. Human error refers to a defect in the human thought process that can occur during the perceiving and processing information, making decisions, or taking some action. Analysis of early human error models showed that errors can happen when choosing a method (referred to as mistakes) or while carrying out an intended method (referred to as slips or lapses).
Human error research (focusing on how normal psychological processes go awry) has been instrumental in reducing aviation accidents and adverse drug events, and improving process control. Considering the human-centric nature of software development, it is not surprising that software failures can be traced back to mistakes in the human thought process (referred to as a human error).
Workshop Goals
The high-level goal of this workshop is to discuss the appropriate ways to apply human error research to better understand the psychological processes (of human reasoning, planning, and problem solving) and how they fail during software development. A better understanding of software errors, informed by human error models will support the use of error information in improving software quality. To have the greatest impact on software quality, the workshop is focused on the errors committed during the requirements development (the first and most critical phase of software development). Natural language (NL) requirements documents are especially defect-prone due to inherent imprecision, ambiguity, and vague nature of NL. Current methods of detecting requirement defects focus on the manifestation of the problem (i.e. the fault) rather than underlying problem (i.e. the error). Therefore, the major goal of the workshop is to understand and structure human errors so that the frequency and gravity of requirement errors can be reduced. More specifically:
Goal 1 ? Collecting Error Information: Understand the nature and type of human errors by collecting examples of errors. Through inclusion of SE (accounting for domain specific errors and defects) and psychology experts (through human error models), we aim to find common failure processes, and interpret the identified errors in light of human information processing limitations and known error patterns.
Goal 2 ? Developing and Refining an Error Taxonomy: The next goal is to structure the error information in a categorization scheme (e.g., an error taxonomy) based on the underlying cognitive mechanism that caused the error so that the errors are clearly structured and useful.
Goals 3 ? Guide the Development of Interventions: The final goal of the workshop is to guide the development of software engineering practices that use the error taxonomy to help developers detect and prevent requirement errors. These interventions can range from changes in task or changes in practice, training, and tool-based assists. The goal here is to establish a research agenda and foster collaboration among member of the the SE and psychology communities to develop tools and techniques that will improve software quality and reduce development costs.
For more information contact Gursimran Walia or Jeffrey Carver.