Web application fault classification - an exploratory study

Abstract

Controlled experiments in web application testing use seeded faults to evaluate the effectiveness of the testing technique. However, the classes of seeded faults are not always experimentally supported by real-world fault data. In this paper, we conduct an exploratory study on two large open source web systems to identify a fault classification that is representative of and supported by real world faults. Through our study we provide support to several categories of an existing web application fault classification, and identify one new fault category and six new sub-categories. Researchers and experimenters will find the proposed fault classification useful when evaluating techniques for testing web applications.