An evaluation of current approaches to support test coverage analysis

Abstract

Test coverage is an important step in software testing and a good indicator of software quality. It helps in evaluating the effectiveness of testing by providing data on different coverage items. It ensures that testing is effectively carried out without missing out all the key functional areas or the features of the software under development. We have analyzed several approaches to understand the existing test coverage framework. These approaches have been classified in different categories namely instrumentation based (source code and byte code), information retrieval (latent semantic indexing), genetic algorithm and UML based. Moreover, we have compared these approaches based on some criteria such as coverage types, data base support, software traceability support, regression testing support, coverage construction and coverage re-construction techniques. We are of the opinion, although much research effort has been put on how to get coverage information by either code based testing or requirement based testing, not much has been paid to measure the coverage by combining both to get better coverage.