About the Book

Some systems are more difficult to test than others. Software testers contend with undefined or partially defined requirements; outdated, incomplete, or nonexistent documentation; complex logic; a mixture of languages; or worse. All of these factors make a system dirty, or virtually untestable.

In Testing Dirty Systems, authors William Perry and Randall Rice teach testers a six-step process for approaching such systems, using • system diagnosis • test planning • test execution • test analysis • report development • and dirty system repair.

Because of the unknown characteristics of the dirty system, the traditional validation of comparing actual processing results against the expected processing results is often inadequate. Analysis of a dirty system must go much further into describing the expected operational characteristics of the system, including

• probability of failure based on failures during testing
• expected difficulty of making changes based on inadequacy of documentation
• estimate of defects that remain in the system
• operating conditions that will lead to failures
• coverage levels based on code or test cases
• complexity levels based on coding structure

Project leaders, independent testers, quality assurance personnel, and IS auditors will benefit from this book, as well as end-users and customers with a vested interest in the success of their systems.

Insights From the Book

"We hope to give you a framework to build upon, not only to test your dirty systems, but also to improve them, using the results of the testing."

—from the introduction