Demand‐driven propagation‐based strategies for testing changes1. (5th June 2013)

• Record Type: Journal Article
• Authors: Santelices, Raul; Harrold, Mary Jean
• Other Names: Harman Mark guestEditor.; Korel Bogdan guestEditor.
• Is Part Of: Software testing, verification & reliability. Volume 23:Number 6(2013)
• Page Start: 499
• Page End: 528
• Publication Date: 2013-06-05
• Subjects: regression testing -- test‐suite augmentation -- change testing -- dependence analysis; Computer software -- Testing -- Periodicals; Computer software -- Verification -- Periodicals; Computer software -- Reliability -- Periodicals; 005.14
• DOI: 10.1002/stvr.1501
• Languages: English
• ISSNs: 0960-0833

Abstract

SUMMARY: Test‐suite augmentation techniques enhance test suites for software changes. In previous work, we introduced an augmentation technique that enumerates the conditions for the propagation of the effects of changes. Empirical studies showed that this technique can test changes effectively but, because of the high complexity of the technique, the experiments were small and the propagation distances from each change were limited. In this paper, we present a new, demand‐driven approach for performing this propagation‐based testing of changes that achieves much greater distances and that enables larger and more significant studies. We implemented this new approach and studied it on a set of changes in Java programs by comparing, to a larger extent than possible before, propagation‐based strategies with other change testing techniques. Our results confirm, with statistical significance, the superiority of propagation‐based strategies over other techniques, and show that these strategies are especially effective for those changes that are the most difficult to test.Copyright © 2013 John Wiley & Sons, Ltd. Abstract : We present a new, demand‐driven approach for propagation‐based testing of software changes that scales considerably more than existing approaches of this kind. We exploit this scalability to perform larger and more comprehensive studies that it was possible before. Our studies indicate, with statistical significance, that propagation‐based strategies for testing changes are more cost‐effective than traditional, coverage‐based approaches.