David Cantrell wrote: > What do people think are *good* tests? Good tests catch bugs.
This might seem obvious and its probably not what you were asking for, but you'd be surprised at how often its forgotten in worries about code coverage, style issues, setup, teardown, portability, black box, glass box, blah blah. If your tests aren't preventing you from breaking things they're no good, I don't care how finely crafted they are. If you're writing lots of tests yet bugs are routinely still slipping through your tests are bad. A classic example is testing the API but not the user interface, for example in a web app testing methods and functions but not the Mason code that uses it.
