Software cost estimation and project planning can be some of the most difficult tasks in the software industry. Software continues to become increasingly complex, and pressure to release it quickly is also on the rise. This book tries to demystify these issues. It deals with both small- and large-scale software projects. Many graphs are included, which depict the effort and time required based on several parameters. The book also includes fascinating quotes by several leading industry practitioners. The factors that affect cost estimation are explored in great detail, including team structure, size, skills, and quality of the resultant product. Several analytical studies go along with the actual experiments.

Choosing the right model is often a source of contention, and the authors explain the pros and cons of several of them. They outline high-level methods to be used: expert-based methods, data-driven methods, and hybrid methods. Several industry-wide estimation methods are discussed in great detail: statistical regression analysis, the constructive cost model (COCOMO), regression trees, case-based reasoning, Bayesian belief networks, wideband delphi, and a few others. The strengths and weaknesses of each of these are described in great detail. The best practices chapter at the end is one of the most useful in the book. Though the book delves into many analytical studies, I would have preferred a more real-world study; for example, the authors could have approached some large software development firms and asked to see real-world metrics, comparing them to the observations made.

This dense book would serve as a reference in a software library, and would be a great resource for software engineering students, as well as managers, engineers, and architects. Often, cost estimations are rushed and pushed through in companies. To produce good estimates, one needs proper functional specifications and then works on design. Design feeds into implementation costs, which feed into testing and maintenance resources. Companies learn from past mistakes; often, the ones that succeed are those that make constant positive reinforcements to the plans. The book should motivate managers to make proper effort estimations and improve the overall quality of software.