One should make testing a practice so that the deliverable quality is ensured after running succesful tests and removing the bugs.
According to Martin Fowler, "Whenever you are tempted to type something into a print statement or a debugger expression, write it as a test instead." At first you will find that you have to create a new fixtures all the time, and testing will seem to slow you down a little. Soon, however, you will begin reusing your library of fixtures and new tests will usually be as simple as adding a method to an existing TestCase subclass.
You can always write more tests. However, you will quickly find that only a fraction of the tests you can imagine are actually useful. What you want is to write tests that fail even though you think they should work, or tests that succeed even though you think they should fail. Another way to think of it is in cost/benefit terms. You want to write tests that will pay you back with information.
Here are a couple of the times that you will receive a reasonable return on your testing investment:
During Development:
When you need to add new functionality to the system, write the tests first. Then, you will be done developing when the test runs.
During Debugging:
When someone discovers a defect in your code, first write a test that will succeed if the code is working. Then debug until the test succeeds.
One word of caution about your tests. Once you get them running, make sure they stay running. There is a huge difference between having your suite running and having it broken. Ideally, you would run every test in your suite every time you change a method. Practically, your suite will soon grow too large to run all the time. Try to optimize your setup code so you can run all the tests. Or, at the very least, create special suites that contain all the tests that might possibly be affected by your current development. Then, run the suite every time you compile. And make sure you run every test at least once a day: overnight, during lunch, during one of those long meetings.