That’s Why We Actually DO the Tests
When I was running testing teams at Microsoft, we often would get pushback on why we need to run our full test suites at the end of the project. While time consuming and tests that almost always passed, we often would find small bugs that had “regressed”—functions that used to work but were broken in the process of fixing or changing something else. In the midst of my team finding a showstopper issue with something that should’ve worked, couldn’t possibly fail, etc., we would again illustrate that’s why we actually do the lengthy regression and stress tests.
Boeing is running into a similar situation with the development of their 787 Dreamliner. While computer models had simulated stress testing of the wings without problem, actual testing by bending the wings identified areas that needed extra reinforcement. From the Seattle Times:
Scott Fancher, the head of the Dreamliner program, explained the structural problem in a conference call early Tuesday morning. He said that late in May engineers, performing wing bend tests on the airplane that is set aside inside the factory specifically for ground testing, found that strain gauges showed higher stress than predicted by the computer models at multiple points along the upper part of the wing-to-body join.
:
When engineers checked the areas flagged by the instruments measuring the strain, they found evidence that the structure was indeed stressed. Asked if that meant delamination, meaning separation of layers in the composite material, Fancher didn't specify exactly but hinted that visible damage was apparent.
That’s why we actually do the tests.
