Browse Books

Reviewer: Alexandre Bergel

Humble and Farley focus on releasing reliable software through continuous and automated builds, tests, and deployment. Thanks to their multiple exposures to painful and costly maintenance issues, the authors are able to identify (and list) the principles of software delivery. As stated by the authors, this book is primarily recommended for developers, systems administrators, testers, and software managers. It is not recommended for a public that is solely focused on academic and theoretical problems and solutions. The authors adopted a practical stance in order to face problems regularly encountered by practitioners in the field. The book's 15 chapters are organized into three parts. Part 1, "Foundations," covers the bit of theory that introduces the causes of reliability failure and justifies the need for a strong discipline among practitioners. Part 2, "The Deployment Pipeline," covers how to automate the process that pulls software from version control to the deployment phase. The final part, "The Delivery Ecosystem," solves the technical obstacles encountered in practice, including managing the infrastructure and environment, software components, dependencies, and establishing a firm command of the version control system (VCS). By directly addressing the problems faced by today's practitioners, the book describes technical tools such as VCSs. However, these new, emerging tools are fragile against evolution and make the relevance of some chapters unsustainable. For example, while the concurrent versions system (CVS) and Subversion are described, Git, a new distributed VCS, is not. The spread of Git among projects will probably make Subversion obsolete one day, just as the CVS has been replaced by Subversion. Even if some chapters may age (painfully), the problems and solutions given in the book will remain identical. Another weak point is the short description of Hudson, a popular continuous integration server. As it stands, when it comes to installing and using Hudson, readers will have to rely on external information sources. I highly recommend this book. It crystallizes, in an easy-to-read manner, the difficulty of achieving software reliability and maintenance. The problems and the approaches required to solve them apply to a very large range of software types. Even though the authors' biography may suggest otherwise, this book does not apply solely to large industrial software; it can also be applied to academic, open-source, small projects. Online Computing Reviews Service