A Look at 25 Years of Data

Is 25 years enough time to build up a coherent body of knowledge that can help point to useful principles? As a testbed for helping us answer this question, software testing techniques are a good place to start. Few software practices are as important as testing, and testing techniques are amenable to measurement and reasoning about their effectiveness. Because they're aimed at removing faults, measuring the number and type of such removed faults seems like a natural part of applying these techniques. To make sense of this data, Universidad Politecnica de Madrid researchers have spent some time worrying about how to put 25 years' worth of work together usefully. © 2009 IEEE.