The biggest single problem I had when I moved from working for myself as a programmer back to the mainstream was dealing with other people's code. For nearly twenty years I only had to deal with my own code. Suddenly I had to understand and change other people's code. It was quite a culture shock, and one of the most difficult things I've done.
I wish I'd found Michael's book earlier, it would have helped ease the transition. Of course, not everyone else's code is legacy code, but even if you are not handling true 'legacy' code, this book will help you deal with the problems you face.
The book is in three parts. The first is a discussion of how you go about changing software that is badly structured and has complex interactions and side effects. Where do you start, and how do you make sure that you don't break things further down the convoluted chain of dependencies. The answer to the latter, of course, is testing, testing, testing, and the author makes a good enough case to persuade even the most sloth like to get into the testing mode.
The second part of the book is organised almost like an FAQ with chapters devoted to common problems like 'I don't understand the code well enough to change it'. This one crops up all too often, and is an excellent example of how the author doesn't avoid difficult questions. The final part of the book is a useful catalog of mostly pattern based techniques that can be used to break dependencies.
The book is well written and features clear examples that are written in either C++, Java, C and C#, and the problems caused by the different features, or lack of them, available to the different languages are discussed and work arounds suggested. I enjoyed reading this book.
9 people found this helpful.
Was this review helpful to you?