Improving Fault Detection Capability Using Coverage Based Analysis

Abstract

A test suite for a system consists of several test cases and as the number of test cases is more, execution of all of these will take much time. If the number of test cases are reduced the execution time could be reduced. Moreover, if the execution of these is done in an ordered fashion ,it is observed to give an increased rate of fault detection. This can further be beneficial by providing feedback to system developers , improving fault fixing activity and thus ,software delivery. The reduction of test cases can be done in several ways. But the scenario is different when functional dependencies exists between some test cases, that is, one test case is needed to be executed before another. In this paper, two test case reduction techniques will be presented which will be followed by using the dependency information from a test suite. The first reduction technique considers all of the test cases while the second one is applied on the rejected suite obtained after implementing the first technique. This work is actually based on the hypothesis that lesser number of test cases will reduce the execution time, provided that the reduced test suite contains cases which collectively will cover the testing of every statement in the source code. Existing dependencies between different tests represent the interaction in the system under test, and hence, execution of complex instructions earlier is supposed to increase the fault detection rate, compared to arbitrary or untreated test orderings. Thus, the dependency information could further be used for prioritization of the test cases.