Although much information is available on test design techniques, very little is written on how to select which techniques to use for the job at hand. Derk-Jan de Grood believes that many testers find it difficult to select the right techniques and very often use a technique simply because they know it. Instead, the best reason is that the technique is likely to discover the most important errors quickly. Derk-Jan shares his insights on test technique selection and poses three questions you should ask yourself when selecting a technique: What types of errors do I want to find? What impact do these errors have in production? Is the needed information to perform these tests available? He then lays out a list of common test techniques and discusses which error types they are most likely to discover. Take back a new understanding of test technique choice and selection to become a better software defect sleuth.

It seems that developers have as many definitions of “done” as Eskimos have words for “snow.” But without a clear definition of done, it is difficult to gauge progress on a project. Menlo Innovations has a simple solution. Instead of declaring a story card or feature done on its own, developers collaborate with the business analyst and testing teammates to determine when the application meets their requirements. Tracy Beeson highlights the power of asking one deceptively simple question that has multiple, complex answers that can, if not implemented with care, upset the status quo and lead to new problems. Using loosely orchestrated role playing, you'll practice this approach for discovering when done is really done.

In the past several years, virtualization has dramatically improved tester productivity. A virtual machine is a useful abstraction for encapsulating the entire software stack. Roussi Roussev presents proven techniques that no modern test environment is complete without. Running multiple virtual machines on a single host maximizes hardware resource utilization and reduces operating costs. Strong isolation facilitates building security testing and multi-tenant environments. With the help of snapshots, virtual machines can quickly travel in time and space. Virtual hardware makes simulating machine, cluster or entire datacenter failure scenarios a whole lot easier. Deterministic record/replay helps track hard to reproduce bugs, comparing outputs allows for measuring the impact of small configuration and binary changes.

The islands that many development and test teams live on seem far apart at times. Developers become frustrated by defect reports with insufficient data to reproduce a problem; testers are constantly retesting the same application and having to reopen "fixed" defects. Chris Menegay focuses on two areas for improvement that will help both teams: a better build process to deliver a more testable application to the test team and a defect reporting process that delivers better defect data back to the developers. From the build perspective, he explores ways for the development team to identify which requirements are completed, which defects were fixed, and how to guide testers on which test cases to execute. Chris details the components of a good defect report, illustrating ways for testers to provide accurate reproduction steps, demonstrating video capture tools, examining valuable log files, and discussing test environment issues.