Test effort estimation in early stages using Use Case and Class models for web applications

Abstract

The testing phase of Software Development Life Cycle (SDLC) is considered essential for delivering the quality expected from the software product. The quality of the software product determines its acceptability and preserves the reputation of the developing organization. Early estimation of test effort is a significant predicament faced by the software managers. An early estimation enables early test planning which helps the project to meet the timeline, as well as budgetary and quality goals. Early estimation of the test effort is a relatively unexplored territory with limited undertaken research work. Unified Modeling Language (UML) is a widely accepted standard for representing Object Oriented Systems (OOS) during different phases of SDLC. Selected UML models created with adequate amount of information can be utilized for early test effort estimation. This work proposes a simple, semi-automated two-stage approach to estimate system test effort from Use Case and Class models during requirements and design phases by using their inherent attributes as parameters. As per the experiments conducted on a number of Web-based projects, the approach presented in this work predicted the test efforts with reasonable accuracy.