A case study of the class firewall regression test selection technique on a large scale distributed software system

Summary, in English

Regression testing is expensive and may consume much of organizations’ software development budgets. Thus, it is of interest to reduce the total time devoted to test execution by using test selection techniques. Many techniques have been proposed but few have been evaluated on real-world, large scale systems. In this paper we report on an empirical evaluation of using the class firewall regression test selection technique, in combination with scenario testing, on a large scale industrial software system using the Java byte code in the analysis. The study was performed on a large complex distributed software system in one of Sweden’s largest banks. Effects of using scenario testing together with regression test selection are reported. The results are that not all test cases were selected by the class firewall selection technique. Using scenario testing, where test cases are dependent, affects the number of test cases selected, as do the location and number of changes in the system.

• Author: Mats Skoglund; Per Runeson
• Department/s: Department of Computer Science; Software Engineering Research Group
• Publishing year: 2005
• Language: English
• Pages: 72-81
• Publication/Series: Proceedings of the International Symposium on Empirical Software Engineering
• Links: http://dx.doi.org/10.1109/ISESE.2005.1541816
• Document type: Conference paper
• Publisher: IEEE - Institute of Electrical and Electronics Engineers Inc.
• Topic: Computer Science
• Keywords: authorisation; Java; distributed processing; program testing; regression analysis
• Conference name: International Symposium on Empirical Software Engineering, 2005.
• Conference date: 2005-11-17 - 2005-11-18
• Status: Published
• ISBN/ISSN/Other: ISBN: 0-7803-9507-7