This topic is designed to provoke curiosity.

How many of you have a hard science background? You know who you are: Chemists, Biologists, Physicists, Geologists… Recall the test plans from your time in hard science classes. These documents left very little to interpretation or imagination as you were headed to the lab. All of the background information regarding Hypothesis being examined, the tests to be conducted, how they would be conducted, how the integrity of the test was checked and cross checked, all of this information is typically included. After all we don’t want chemistry majors to be graduating with bad burns or physics majors to lose limbs in interpretive pendulum testing. Plus we want repeatability in our tests, both by the individual tester and anyone else who would want to audit our tests by repeating the same effort. All things being equal they second party should obtain the same result as the first, if the documentation is spot on, clear and precise.

The closest we tend to get to a reference sample for performance testing is not so much an application test, but it is an industry standard benchmark, in this case the TPC-C benchmark suite, . Download it and take it for a spin. It leaves little to the imagination in terms of what is to be done during the test. How does your test plan stack up? Could someone else take your plan and wind up testing the system under test in exactly the same way? Does this pose issues for you if your test was to be anonymously audited by an outside party? Or is this a compromise on the time allowed by your management for test plan documentation versus what should be in there?

And then there are those who have no test plan. Well, some items are best left unsaid in polite company…