Measurement, prediction and risk analysis for web applications

R. M. Fewster and E. Mendes, University of Auckland


Accurate estimates of development effort play an important role in the successful management of larger Web development projects. However, estimating the effort required in developing Web applications can be a difficult task. By applying measurement principles to measure qualities of the applications and their development processes, feedback can be obtained to help understand, control and improve products and processes.

In addition, effort prediction models for Web development can be proposed using the predictor metrics obtained. The objective of this paper is to present a Web design and authoring prediction model based on a set of metrics which were collected using a case study evaluation. The case study used as subjects final year Computer Science students who had to develop Web applications using static and dynamic pages. All the applications had to be structured according to an instructional theory named Cognitive Flexibility Theory.

The paper is organised into three parts: part I describes the case study evaluation (CSE) in which the metrics used in the prediction model were collected. These metrics were organised into five categories: effort metrics, structure metrics, complexity metrics, reuse metrics and size metrics.

Part II presents the prediction model proposed, which was generated using a Generalised Linear Model (GLM), and assesses its prediction power. Finally, part III investigates the use of the GLM as a framework for risk management.

Last updated: 14th August 2000