Designing Object Oriented C++ Applications Using The Booch Method

Rating:
Summary: By far, the best OO book I've read to date.
Review: A must have for anyone doing OOA and OOD! If you areprogramming in any strongly typed OO language you oweit to yourself to read this book. Truly excellent!

Rating:
Summary: My "Aha!" Moment for O-O Programming
Review: Even if you don't know C++ or use the Booch method, this is the best work I've read on how to engineer software systems. The principles presented in this book apply regardless of computer language or design notation. And, while the book is geared to O-O, those priciples, I think, apply to software design in general. When you finish this book you will think "Of course! Why would you do it any other way?!". Despite over 20 years experience is software design and programming (including 10 in O-O), this book was the "Aha!" moment that made me truely understand how to design and build software systems. I can't reccommend it highly enough. Now, if Martin would just do the same book using Java and UML...

Rating:
Summary: Essential, thorough and immensely pracitcal guide to OOD
Review: I have used this book in the years since it was first pubished and have reccommended it on numerous occasions.It tackles the chasm that seperates idealistic design from realistic implementation.I refer to it no matter which OO design methodology I am using, no matter which OO language a system is being implemented in. I have found that the difference between the designs of people who exercise the principles of this book and those who don't are noticeable - especially during the design of large systems

Rating:
Summary: Well done for its time, but showing its age
Review: Of course, anyone up on the world of software development can tell that this book is long-in-the-tooth just by the title. Even Grady Booch doesn't use the Booch method anymore.

Robert Martin includes one of the best sections on software metrics you are going to find anywhere. His discussion of coupling is clear and usable. He was a little too zealous when he "defined" representational objects as "not-OO" because they didn't fit on his "main sequence." His main sequence analysis is useful, but he's a little too rigid about it.

His definition of the famous "Open-Closed" principle is obscure, but he makes up for it with copious examples.

I really must praise his Security System example as one of the best thought out examples in the literature. His use case analyses, though, are really not up with the times, and were not very well explained in the first place.

Unfortunately, there aren't a lot of folks writing better books in the time since. For use cases, Check out Alistair Cockburn. For modeling, pretty much anything post 1999 from Rational Software will give you better method introductions. Read this book, though, for examples of how even big projects can be done well.