Sizing maintenance tasks for web applications

Abstract

Web applications are fast becoming the new legacy systems of today. While there is considerable similarity between traditional software systems and Web-based systems, there are also significant differences between them. One area that illustrates this dual nature is cost and effort estimation. There exists a mature body of knowledge for performing such estimates on traditional software systems, but such methods may not appear to be directly applicable to Web applications. This paper presents an effort estimation technique for maintaining a large-scale Web application by measuring and tracking the size and complexity of the Web-based system. Specifically, a combination of function-points and static impact analysis is used, tracing the change requests to different components of the Web application, and then measuring their size and complexity to aid the cost estimation for that particular change request based on function point productivity measurements. To illustrate the use of this technique, a case study from a real-world industrial product is presented. © 2007 IEEE.