Agile Database Techniques: Effective Strategies for the Agile Software Developer (Wiley Application Development) [ペーパーバック]

Ambler starts laying out the groundwork for the second part of the book by introducing his Agile Data method, UML and data modeling. He also gives a very useful jump-start tutorial on data and class normalization and discusses the infamous object-relational impedance mismatch. Worth noting is that in each chapter (throughout the book) Ambler makes sure that the reader understands how the subjects relate to the role of an agile DBA and what should he be looking out for. The subjects in part one were introduced so well that I more than once thought, "I've never seen such a well-balanced and informative tutorial."

The second part of the book focuses a bit on how evolutionary software development is the reality and how techniques such as Agile Model-Driven Development and TDD can assist in surviving in this climate. The chapter on database refactoring is intriguing and the more comprehensive list of database refactorings, found in the appendices, is something I'll definitely Xerox and keep with me in my traveling tool bag. The second part also covers database encapsulation strategies and approaches to mapping objects to relational databases which, again, is a delightfully comprehensive tutorial, as well as touching on the topic of performance tuning.

The third part is a seemingly random collection of subjects such as finding objects in a relational database, referential integrity, concurrency, security, reports and XML with databases. The text is again of high quality and provides an entertaining walk through the woods, but I feel that these chapters had very little to contribute to the subject of the book. This disconnect is made even more clear when Ambler finishes with a 10-page guide on how to become agile and to introduce agility into the organization.

After reading the book, I feel confused. On the other hand, this book is full of little islands of useful insights that I wouldn't think twice recommending for others to read. On the other, the book as a whole is very fragmented and pretty far from what the book's title might suggest. If you pick up the book with an open mind like I did, you'll gain plenty. Otherwise, you'll probably be disappointed to find out that one half of the book probably talks about something you weren't expecting based on the title. After a long struggle, I'm giving this book "only" fours stars despite that it contains some very, very high quality material.

Should you buy this book? Well, try answering the following questions:

- Could you easily create a physical data model of your schema?
- Can you explain the difference between first and third normal form?
- What about first and third object normalization forms?
- Can you list the challenges in mapping an object model to a data model?

Did you answer honestly? Because these are just some of the items covered in Part One of the book and unless you said yes to all of the above, you will walk away with plenty of ideas for improving your development after a single reading. This is the core reason I would recommend this book. It is full of ideas that you might never have thought to include in your development practices, and probably some that you had thought about but were not sure what the best approach might be.

My complaint with "Agile Database Techniques" is that it could use another iteration. In the introduction, the author states "When I first started writing this book, I intended its focus to be on the agile data method... Because I was taking an iterative and incremental approach to the development of the book, I quickly realized that the real value lay in detailed development techniques instead of yet another methodology." While I can agree with and embrace the merits of applying agility to my development practices, I think there is an overabundance of methodology content in the book that only clouds its real value.

In summary, I think this is one of those books that will sit on my shelf for some time to come as a great reference for how to approach any number of application and database development related items.

More importantly the book has chapters on agile model driven development and test driven development, two topics that are critical to your success as an agile programmer. Anyone doing XP needs to be up to speed on both these techniques. in particular TDD but just as importantly AMDD.

A really cool chapter talks about database refactoring, which is basically refactoring for relational databases. Anyone who wants to be an effective DBA needs to learn this stuff. There's also an appendix with a catalog of DB refactorings. This material alone is worth the price of the book.

I didn't think that the chapters on basic programming issues such as concurrency control, referential integrity, database access, and reporting would be of interest. Man was I wrong. Ambler takes what I thought were dry subjects and shows how many implementation choices you actually have. I can't remember how many arguments I've gotten into with DBAs, or programmers, who thought there was only one way to do these things. Most importantly he discusses the tradeoffs of each strategy and tells you when they work best. If you're designing a new application this material is critical.

If you're struggling to find ways for programmers and DBAs to work together effectively you really need this book. It shows how to overcome the "people impedance mismatch" that you see in most companies. DBAs need to work in agile ways, and this book tells how to do exactly this.