Bests & Worsts Reviews from Amazon

To fairly review this book, one must distinguish between the methodology it presents and the actual presentation. As to the presentation, the author attempts to win the reader over with emotional persuasion and pep talk rather than with facts and hard evidence. Stories of childhood and comradeship don't classify as convincing facts to me. A single case study-the C3 project-is often referred to, but with no specific information (do note that the project was cancelled by the client after staying in development for far too long). As to the method itself, it basically boils down to four core practices: 1. Always have a customer available on site. 2. Unit test before you code. 3. Program in pairs. 4. Forfeit detailed design in favor of incremental, daily releases and refactoring. If you do the above, and you have excellent staff on your hands, then the book promises that you'll reap the benefits of faster development, less overtime, and happier customers. Of course, the book fails to point out that if your staff is all highly qualified people, then the project is likely to succeed no matter what methodology you use. I'm sure that anyone who has worked in the software industry for sometime has noticed the sad state that most computer professionals are in nowadays. However, assuming that you have all the topnotch developers that you desire, the outlined methodology is almost impossible to apply in real world scenarios. Having a customer always available on site would mean that the customer in question is probably a small, expendable fish in his organization and is unlikely to have any useful knowledge of its business practices. Unit testing code before it is written means that one would have to have a mental picture of what one is going to write before writing it, which is difficult without upfront design. And maintaining such tests as the code changes would be a nightmare. Programming in pairs all the time would assume that your topnotch developers are also sociable creatures, which is rarely the case, and even if they were, no one would be able to justify the practice in terms of productivity. I won't discuss why I think that abandoning upfront design is a bad practice; the whole idea is too ridiculous to debate. Both book and methodology will attract fledgling developers with its promise of hacking as an acceptable software practice and a development universe revolving around the programmer. It's a cult, not a methodology, were the followers shall find salvation and 40-hour working weeks. Experience is a great teacher, but only a fool would learn from it alone. Listen to what the opponents have to say before embracing change, and don't forget to take the proverbial grain of salt. Two stars out of five for the presentation for being courageous and attempting to defy the standard practices of the industry. Two stars for the methodology itself, because it underlines several common sense practices that are very useful once practiced without the extremity.