Automatic Testing for Link Correctness on the Internet

The user-friendly convenience of the WWW promotes its rapid application delivery, but the technical complexities of a website and variances in the browser make web testing and quality control much more difficult. This paper proposes an automatic framework for testing hyperlink correctness in the Internet environment. We first investigate the underlying features of Internet applications and then discuss quality issues and a performance index. We study how to employ the technique of statistical usage testing in such a domain and develop an integrated validating mechanism for web-based applications. All possible navigation paths are first formulated to represent a scenario model with the Markov chain property, which is then used to statistically generate a navigation-script file. After tracing these navigation scripts, we may then perform numerical computations of link validation based on Markov chain theory. It is noted that the proposed automatic mechanism is not only systematically effective on certifying hyperlinks, but also highly efficient for designing a test process.