A Field Study on the Elicitation and Classification of Defects for Defect Models

TL;DR

This study explores a framework for eliciting and classifying software defects within specific contexts, aiming to improve defect modeling and quality assurance processes through practical case studies.

Contribution

It introduces a defect model lifecycle framework and evaluates its effectiveness across multiple contexts through qualitative case studies.

Findings

Framework effectively elicits context-specific defects

Defect classification aids in operationalizing defect models

Practitioners perceive immediate benefits from the approach

Abstract

Defect models capture faults and methods to provoke failures. To integrate such defect models into existing quality assurance processes, we developed a defect model lifecycle framework, in which the elicitation and classification of context-specific defects forms a crucial step. Although we could gather first insights from its practical application, we still have little knowledge about its benefits and limitations. We aim at qualitatively analyzing the context-specific elicitation and classification of defects to explore the suitability of our approach for practical application. We apply case study research in multiple contexts and analyze (1) what kind of defects we can elicit and the degree to which the defects matter to a context only, (2) the extent to which it leads to results useful enough for describing and operationalizing defect models, and (3) if there is a perceived additional immediate benefit from a practitioner's perspective. Our results strengthen our confidence on the suitability of our approach to elicit defects that are context-specific as well as context-independent. We conclude so far that our approach is suitable to provide a blueprint on how to elicit and classify defects for specific contexts to be used for the improvement of quality assurance techniques.