It's rarely wise to reach 100% code coverage. How do you decide where to spend your limited testing capacities? What are high value tests that you should focus on early in your projects?
