Emergent Design by Scott Bain is a interesting book. The title is very promising, when I first heard about it, I got very excited! Finally a book about how designs emerge, how designs emerge from multiple people and how designs evolve over time compared to specifying. After reading the book, I felt the book was good, but disappointing. It did not cover the topics I would like to have seen.
The general idea of the book is that software should grow better over time instead of decay over time and that the optimal design will emerge. An idea I strongly agree with. The author links this to software development needing to change to become a profession. If SW development is a profession, then people will use proper practices and design will emerge. The practices (in a broad sense) are principles of design, patterns and disciplines. After the first couple of chapters the book was having a good start, though I started wondering if the author didn't bite of more than he could chew. Those are huge topics by themselves!
From chapter 7 to chapter 14 the author just describes good practices. He starts with qualities of code and qualities of designs. He moves to unit-testing, refactoring and then Test-Driven-Development. He ends with the pattern chapter. The last chapter puts all things together in a case study. Scott does a reasonable job in describing all practices. There are a couple of weird things, like the recommendation that every class has exactly one test class. The TDD chapter also seems to have very little TDD in it :)
As a catalog of best practices, this book perhaps does the best of all the current agile related books. Great job by the author.
However, there are some things that personally bothered me. The book seems to be very pattern focused. Scott seems to be of the opinion that patterns is what hold everything together (probably everything in the world). Though, I agree that patterns are an important concept in modern software development, I wouldn't put so much pattern focus in e.g. a chapter on test-driven development. Maybe the title of the book would better be "Scott on SW design and patterns".
That brings me to another issue with the book, the title. Emergent design is an immensely important topic. How does a design start with the first requirement. How does it evolve. How do multiple people work with the design. How can the overall architecture evolve. What about items that evolve difficult, like different programming language usage etc. So much to talk about and the book doesn't do this. It misses a huge opportunity to talk about emergent design & architecture and instead (although important) decides to talk about design principles, patterns and practices. (in that sense, the book is similar to Bob Martin's "Agile Software Development: Principles, Patterns and Practices", which I would recommend over this book).
But again, the content of the book is good and useful and normally I would go for a 4 star rating, but I decided to go for 3 stars. This is because the book IMHO contains things that really turned me off.
One of the examples is the talk about professionalism. Don't get me wrong, I do agree with the author on this subject. The point is, we are not alone. In fact, IEEE has been working on certification for many years. In 1999, Steven McConnell wrote a book called "After the Gold Rush" with the subtitle "Creating a True Profession of Software Engineering". Scott talks about finally making a profession out of SW development, but he seems to have not done any research on this topic and seems to not be involved in other attempts to make it a profession. It would have increased his credibility a lot if he would have said "the earlier attempts are different because ..." or something similar.
Another item that was a huge turn-off was the constant promotion of Net Objectives. The book, at times, almost felt like a commercial. Personally, I didn't need to know about what courses Net Objectives teaches, I want to know about Emergent Design!
Anyhow, all these negative points aside, Emergent Design is a good introduction to modern agile development practices. Especially if you are not yet familiar with topics like Refactoring, TDD and patterns, this book is certainly worth reading. Next to that, Scott's writing style is funny and easy to read. So, if you belong to that group of people, recommended! Otherwise, skip it.
