STARWEST 2016 - Process & Metrics

Anyone who has ever attempted to estimate software testing effort realizes just how difficult the task can be. The number of factors that can affect the estimate is virtually unlimited. The keys to good estimates are understanding the primary variables, comparing them to known standards, and normalizing the estimates based on their differences. This is easy to say but difficult to accomplish because estimates are frequently required even when we know very little about the project—and what we do know is constantly changing. Throw in a healthy dose of politics and a...

In performance testing we run a suite of tests, modify the system in an attempt to improve its performance, and then repeat the tests. We want to know if the performance is “better.” However, no single performance measure exists; we must consider and evaluate many variables. Although viewing a full histogram of various test runs is more helpful, quantifying the change between the tests becomes the challenge. Parvez Ahammad introduces the relative divergence index (RDI), a multidimensional statistical method to compare differences between two sets of performance...