Customer Reviews

A Practical Guide to Testing Object-Oriented Software

5 star:		(2)
4 star:		(1)
3 star:		(0)
2 star:		(1)
1 star:		(1)

 

 

I purchased this book to use as a guide for developing a strategy to validate directory-enabled security infrastructures. It met my requirements and more. I discovered that this is also a great introduction to object-orientation for test professionals new to the paradigm. Also, because this book covers every facet of O-O testing it is also useful to testers who have worked in the O-O environment.

It starts out with a good description of the differences and unique challenges of testing O-O software. It then covers the basics. Starting with test planning it provides a clear approach to performing this important task in the O-O environment. This material can also be used as best practices for testing in any environment. I especially liked the chapter on testing analysis and design models. In particular, this chapter validated my personal theories on the way to approach design validation of directory services, such as LDAP, that are becoming a prevalent component of enterprise security infrastructures. The next three chapters address testing classes, interactions and class hierarchies. These chapters cover the foundation of O-O testing and provide those who are new to O-O testing with a clear view of what's important and how to approach developing a test strategy and cases, and performing the actual testing.

The book moves into advanced topics in the chapter on testing distributed objects. This is a complex subject and developing a test strategy in this area is a daunting task. The authors cover this material in depth, and provide a clear and complete description of all issues and factors. I was impressed with thoroughness and attention paid to the most minute detail. The authors include life cycle testing, which makes this book also important to software quality assurance (SQA) professionals. Note, SQA and testers are two different groups with two different sets of objectives. Where the tester is concerned with trying to break software, SQA collects metrics and devises process improvement with respect to defect prevention. The authors' approach contains something for each of these groups. The next level of testing granularity that is addressed is testing whole systems. This chapter covers the entire range, including stress testing, the wider considerations of system life cycle testing, embedded software, multi-tiered systems and more. Because my interest was testing security infrastructures I was pleased to see that the authors included security testing in this chapter. This is an added bonus, and shows the thoroughness with which the book approaches testing in the whole.

Other topics that make this a valuable book for test professionals are included in the chapter on components, frameworks and product lines. The authors show the differences between testing objects and components, and cover topics such as enterprise java beans, inspecting and testing frameworks, and interface testing.

Overall this is a comprehensive book that covers every conceivable task and topic related to O-O testing. It is well written and contains many pleasant surprises, such as security testing, and well developed test strategies for any environment with which an O-O tester will be confronted. It provided me with the exact answers to some thorny questions about validating and testing directory-enabled security, and taught me a lot more about O-O testing along the way.

The book describes object orientation. Oh, come on, folks, there are tons of books on that topic and they do it better.
The contents are very object orientation centered. In my experience as a tester, I don't care about the use language, paradigm or what ever (except for reviews of course). To me it seems that OO was new to the authors and they thought it would demand many new testing strategies, which it doesn't (after the JUnit testing of course).
The proposed procedures look very much like something used at a university, but are not practical in the real world.
I am a tester and test manager and I certainly did not found any help in here.

As you might expect from a couple of academic professors this book is full of formal processes that are easy to teach and write exams around but don't scale to the real world.

The book spends 47 pages covering the basics of object oriented design (class, object, inheritance etc) and UML diagrams (class diagrams, state diagrams, sequence diagrams) yet completely ignores basic testing concepts such as white and black box testing. The book also has zero coverage of popular testing tools such as JUnit and NUnit that are making a real difference in test productivity and code quality these days.

I found only two useful ideas in the book, guided inspection and orthogonal array testing. Guided inspection is documented in mind numbing detail but unfortunately the book does such a poor job of explaining orthogonal array testing that I had to go and research it on the web.

Surprisingly for a book that claims to be a practical guide the exercises are largely ambiguous and open-ended essay style questions. The authors provide snippets of a breakout-style game that they use as a running example throughout the book. The exercises could have been so much better if they had included a design and implementation of a simpler application and set practical problems based on that code. If, like me, you learn through doing then you won't get much help from this book.

I doubt I will pick this book up again anytime soon and neither will I be recommending it to any of my friends or coworkers.

All i can say is that I have read this book at least 4 times. Evertime it enlightened me and allow me to see OO testing in different perfectives. This is not a techie book but a practical one (as the book title claimed). After I read this book, i realized company who claimed to have top of the quality class has many missing components along the quality process.

EQ

MM. McGregor et Sykes ont écrit dans ce livre les méthodes qu'ils ont développé pour tester du code orienté objet. Les diagrammes en UML abondent et aident beaucoup à illustrer leurs propos.
Les auteurs commencent par expliquer avec beaucoup de détails et code à l'appui comment créer des classes de pilotes pour tester les invariants, les interfaces et les méthodes de classes simples, qu'elles soient codées en C++ ou en Java. Après ce point, il n'y a plus de code en C++ ou en Java. L'inévitable sujet du test des interactions entre objets est traité copieusement et avec suffisamment de diagrammes UML, et il est aisé de s'inspirer du code du chapitre précédent pour trouver une manière de coder ce genre de tests. Ensuite, il est question de la construction de hiérarchies de classes de pilotes pour tester des hiérarchies correspondantes de classes. Plus loin, les auteurs décrivent avec bon nombre de détails les enjeux des tests de programmes multiprocessus, que les tâches en question, éphémères ou durables, soient réparties dans le même procédé, dans la même machine ou dans un réseau. La manière de coder ce genre de test est plus complexe en C++ qu'en Java, et j'aurais aimé avoir sous les yeux un exemple en C++.
En résumé, je trouve que ce livre est un bon point de départ pour un programmeur qui veut tester son code orienté objet plus à fond et d'une manière automatique avant de soumettre une nouvelle version. Ce livre parle très peu des tests de boîte noire et qui consistent à identifier des entrées ou des conditions externes qui causent des problèmes avec le logiciel étudié.