This post follows on from the previous two in this series (theory and path coverage). A key theme of these posts is that of modeling paths through an application and then using tests to exercise those paths. What we’ve seen so far is that while many path segments are covered repeatedly, others may only be touched once. You most likely won’t have time to test them all. Along with that, some paths will be more data-dependent than others. So a selection process is critical. Here I’ll talk more about that selection process.