Ricky & Stick - Test Before Fielding
Ricky & Stick main page
|
|
|
The stories about the pain and failure caused by inadequate testing are probably the best known tales in the software community; some of them have taken on a near-legendary status. Nor are they all legends, since testing really is a messy issue. It's costly, it's time consuming, and (so the theorists insist) nearly impossible to do perfectly. Even worse, testing tends, whether rightly or wrongly, to come late in the day, and for managers already behind schedule, it's often tempting to cut the testing resources to the bone.
But in yielding to that temptation, you're potentially adding to the painful tales and legends. You may really think that there's a compelling reason for skipping a crucial testing cycle. (Maybe, if you don't hurry up, you'll miss the race.) But chances are that by taking that route, by doing the real-world equivalent of operating a downhill racer without testing the brakes, the eventual crash is almost guaranteed. In retrospect, so were most of the DoD testing failures that have occurred over the years.
To be sure, there's no easy answer to the question: How much testing is enough? But there's a very easy answer when we get into a situation like Ricky and Stick: You've got to do at least some. And it must be real testing of the parts that really need to be tested.
Bottom line: No matter what schedule pressure you may be under, the outcome of a battle may someday depend on the system you're building. So if testing is getting squeezed, you may want to ask the contractor, "What are we risking by skipping this set of tests" When lives are at stake, the "but we're way behind schedule" argument just isn't good enough."
For More Information
Send comments or questions to asp-requests@sei.cmu.edu
return to top
Ricky & Stick main page
Acquisition Support main page