Test Harnesses for API Testing

Automated testing in most QA organizations involves capture-playback tools in combination with manual testing. But these types of tests often suffer from well-known implementation and execution problems. One way to enrich the set of automated tests is through the use of API-level testing. Because APIs tend to be more stable than GUIs, many of the GUI-related test problems disappear, and API tests can be constructed earlier in the process. Generalizing the testing approach for API testing means running them in a common framework, a “Test Harness.” Michael Sonshine provides an overview of how to add API testing to your current testing mix and implement a minimum useful set of functionality. Find out ways to standardize input and outputs and novel strategies for minimizing the impact of release changes. Learn what inputs should be used, and what outputs should be produced. Find out how much work should be done by the Test Harness, and how much by the tester?

• How API test cases should be structured
• The software tools needed to implement an API test harness (Reflection, .NET remote, Web services, and others)
• How to use instrumentation to complete the testing process