Automatic Evaluation of Interfaces on the Internet

Abstract

Empirical methods in human-computer interaction (HCI) are very expensive, and the large number of information systems on the Internet requires great efforts for their evaluation. Automatic methods try to evaluate the quality of Web pages without human intervention in order to reduce the cost for evaluation. However, automatic evaluation of an interface cannot replace usability testing and other elaborated methods. Many definitions for the quality of information products are discussed in the literature. The user interface and the content are inseparable on the Web, and as a consequence, their evaluation cannot always be separated easily. Thus, content and interface are usually considered as two aspects of quality and are assessed together. A helpful quality definition in this context is provided by Huang, Lee, and Wang (1999). It is shown in Table 1.