Customer Reviews

The Art of Software Testing

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

 

 

......... this is probably not the first book that you would buy about software testing.

However, the book is a classic and it deserves a place on a serious tester's bookshelf.

Its examples are dated, I think its description of cause-effect graphing is incomprehensible, and its catalog of test types in the pages from 103 forward is sketchy.

The book is valuable because its presentation of the basic issues is clear, concise, and persuasive. The discussion of equivalence classes and boundaries is remarkably clear. When we wrote Testing Computer Software, one of our goals was to handle this important topic as clearly and crisply as Myers. That was a challenge, and I'm not sure we succeeded. (Jorgensen's Software Testing: A Craftsman's Approach does a great job with this topic.) The discussion of bias (one of the issues in the psychology of testing) is also well done.

In short, the first 103 pages of the book are some of the best writing in the field and have had a powerful influence on the writers who came later. Reading them in the original will often, I suspect, make subsequent presentations clearer and more meaningful.

-- Cem Kaner (senior author: Testing Computer Software)

This is by far the most concise and insightful book I've ever read about code level testing. It does not have all the nitty gritty details of every which method ever invented, nor does go into details about the paperwork. But the lists of principles and checklists are priceless.

I would not recommend this book for beginners since it is hard for inexperienced testers to pick out the gems from the dated items.

I agree with a previous review that stated that the first hundred or so pages are must reads. Don't be put off by the $ per page ratio. This book is worth every penny.

Like all standard references, one can argue that the book is dated. In details perhaps yes, though my professional experience indicates that a 25 year old book still defines a higher level of software testing than the current state of the practice.

The first 75 pages lay out the basic approaches to software testing and make for easy technical reading. More detailed graph theory follows and helps take the reader to a more rigorous approach.

Please feel free to contact me for more detailed insght and opinions.

Doug Claflin; President - Maritime Design

This book was a noble effort for its time. It's a good starting place, especially since it's the one book on software testing that all the other authors seem to have read. And the chapter on test cases is pretty good.

The book has some problems, though. First and most obviously, the book has a lard factor of about 60%. Short as it is, it needs a lot of pruning. (See Richard Lanham: Revising Prose.)

Second, here and there it is just wrong. E.g., in its answers to the chapter 1 neophyte tester test, there are six, not three, permutations of three distinguishable objects.

Third, he doesn't adequately sum up or review what he's said or done. Nowhere does he have a survey of what techniques one should use to generate all the test cases for his chapter 1 test. In chapter 2 he gives a horrible example of why you can't do exhaustive branch testing of a program with even fairly simple logic inside a loop, but he never comes back to say how you *should* test such a thing. Later on he talks about multiple decision/condition coverage, but he never explains how that differs from exhaustive branch testing.

Also -- OK, it was a first try, so perhaps one shouldn't expect too much. But by now someone should be applying complexity theory to the problem. Given *this* program and *this* test coverage criterion, how many test cases will it take to test this program? And the coverage criteria should be stated in terms of the language definition, i.e. probably in terms of Backus-Naur form.

Buy this book. Read it, with a red pencil in hand. But expect it to raise as many questions as it answers.

There is a recent update of The Art of Software Testing, but I thought I start with the previous version.
The $ 0,01 price tag certainly helped with that.

Although this book is now a bit dated, the fundamentals presented are still very helpful.

The quality of this old librarybook is as stated: good.

It is a classic. Many new testing books came out later, but frankely speaking, they were much worse this classic. Talking mistakes in the book, well, there are more mistakes in other books.