Estimating and Managing Project Scope for Maintenance and Reuse Projects

摘要

Estimating project scope is considered by many to be the most difficult part of software estimation. Parametric models have been shown to give accurate estimates of cost and duration given accurate inputs of the project scope, but how do you input scope early in the lifecycle when the requirements are still vaguely understood? How can scope be estimated, quantified, and documented in a manner that is understandable to management, end users, and estimating tools? This talk focuses on scope estimates for maintenance and reuse work, including bug fixes (corrective maintenance), modifications to support changes in the operating system, database management system, compiler or other aspect of the operating environment (adaptive maintenance) and modifications of existing functionality to improve that functionality (perfective maintenance). Reuse includes any case where you are modifying an existing code base to support enhanced functionality, and includes cases where an existing application is translated to a new language. At a high level, maintenance projects consist of three types of work: 1. Maintaining an existing, functioning application; 2. Modifying existing code to support changing requirements; and 3. Adding new functionality to an existing application. A team doing a new build for an existing application would only be concerned with items 2 and 3. A team keeping an existing code base functioning would only do item 1. A project manager may be responsible for both areas and might need to estimate the effort required for all three. We'll deal with each individually.