AbstractsComputer Science

Empirical studies on exploratory software testing

by Juha Itkonen




Institution: Aalto University
Department:
Year: 2011
Keywords: Computer science; software testing; exploratory testing; defect detection; effectiveness; experience; domain knowledge; case study; controlled experiment; field observation
Record ID: 1136310
Full text PDF: https://aaltodoc.aalto.fi/handle/123456789/5054


Abstract

Exploratory software testing (ET) is a practically relevant approach to software testing that lacks scientific knowledge. In ET, the tester's work is not based on predesigned and documented test cases. Instead, testing is guided by a higher-level plan or mission, and the testing work involves parallel test design, test execution, and learning. One of the distinct characteristics of ET is that the tester designs the tests during ET and uses information gained to design new and better tests continuously. The ET approach relies on testers' skills and experience. The main claimed benefits of ET are the tester's ability to apply personal knowledge and creativity during testing as well as effectiveness, efficiency, and agility in terms of adapting to changes and working with imperfect documentation. In this thesis, the ET approach has been studied using empirical research methods. Two case studies, one controlled experiment, and two field studies were performed to address three research goals: defining ET and understanding its applicability based on the literature; empirically investigating the benefits and shortcomings of ET; and providing empirically based results on how the ET approach is applied in practice. This research identifies different approaches to ET in industry and describes concrete testing practices. The role of the tester's personal knowledge is identified in the literature, and this research provides a detailed analysis of the application of personal knowledge in failure detection using ET. The main conclusions of this work are that ET can be as effective as test case-based approaches and even more efficient in certain contexts. The testers are capable of utilizing their personal knowledge in failure detection, and the role of personal knowledge is important in the ET approach. In addition, software testing in product organizations seems to involve multiple diverse organizational groups, and ET was found to be an applicable approach to engage domain experts in testing. The main implications of this thesis are introducing the exploratory testing approach to the research community and motivating its relevance by providing empirical studies in industry. In addition, the results of the effectiveness and efficiency of ET as well as the qualitative data on exploratory testing practices and the detailed analysis of knowledge in exploratory testing work are valuable for the research community. The main practical implications include presenting the benefits and applicability of the ET approach along with the potential shortcomings and providing empirical evidence regarding the benefits of ET.