If you are familiar with the “The Pragmatic Programmer”(written by Andy Hunt and Dave Thomas) then you will probably recognise the Pragmatic Bookshelf series of books, set up by Hunt and Thomas, in which Ship It! is included. This series of books sets itself apart by emphasising the practical aspects of delivering software, and also doing it well.
The book is split into the following chapters:
• Tools and Infrastructure
• Pragmatic Project Techniques
• Tracer Bullet Development
• Common Problems and How to Fix Them
From the very first chapter, the authors make it clear ‘Ship It!’ is not another methodology, “There is no single, right way to develop software. There are a lot of wrong ways…”. Instead, they have gathered together the ‘best’ bits’ of various styles and methodologies they have been directly involved with, and combined them into a practical approach with the focus on delivering a project. The authors do not expect you to necessarily implement everything they suggest all at once. Adopt one or two at time and determine if they work in your environment. This practical stance is reinforced throughout the book. A selection of the topics examined can be broadly listed as follows:
• Use source control (for anything you want to keep)
• Automate your build (one click)
• Use continuous integration
• Track issues and features
• Write tests and automate them
• Use mocking
• Build end-to-end stubbed systems (Tracer Bullet Development)
• Build teams and foster communication
It is easy to read and feels fresh. As another reviewer pointed out, “It's a rare book that speaks convincingly to both developers and managers, but this one does a good job”. There is a definite emphasis on the positive; it is about how to make projects succeed rather than a post-mortem of why they fail. The chapter “Common Problems and How to Fix Them” is a gem, with practical advice for developers, managers and customers.
In Appendix E: Issue Tracking Systems, I was surprised to see that Gemini from CounterSoft (http://www.countersoft.com/) was not mentioned. I have implemented it at several sites with minimal effort and it has been used by developers and users with great success.
In Tracer bullet development, I agree with the approach to the end-to-end working ‘stubbed‘ systems, but I would also add that methods should also be written in such a way that the release version will throw an exception in any method that has not been implemented (rather than just returning a dummy value). Otherwise, years from now, you might find yourself mentioned on the DailyWTF! In some ways, it would have been nice to have had a worked code example for this section, but I realise the authors did not want to limit themselves to one particular language and restrict the reach of this book; I think they made the right decision.
If you are an experienced developer, then you probably will not find anything revolutionary or truly new amongst its 200 pages, but you might find something that your current routine is missing. Is this book worth reading? Absolutely. I would not hesitate for a moment in saying go out and buy it. Highly recommended.
Disclosure: The Perth .NET User Group is a member of the O'Reilly User Group and Professional Association Program. O’Reilly make copies of their books available for user group libraries, and the copy reviewed here was kindly donated by O’Reilly. That said, it was not a factor in reviewing this book