Author: Klaus Haller
Published in: testing experience, June 2010
It was the summer of 2008. Testing suddenly emerged as the Topic in my professional life. I became responsible for the database back-end of a credit rating application. Banks use such applications for estimating whether companies pay back their loans. This impacts whether they get a loan and for which interest rate. So I was interested in how to test “my” application. But the application was different to the ones you read about in testing literature. The application was built on a database (a database-driven application, or, short, a DBAP). The DBAP output of an action not only depends on your input, but also on the history of the DBAP’s usage. The history manifests in rows and data stored in the tables. And this history influenced the present and future behavior of my application. What sounds like a philosophical question
is of high practical relevance. Four questions sum up what I had to answer myself: What do I need DBAP test data for? When is a DBAP correct? What is a DBAP test case? And, finally, which test data is good?