T-Shirt hurling contest: day 2
By jag on Jun 28, 2005
The early prototypes were pretty straightforward. No electronics. Real reliable. What they ended up with was a total victory for creeping featurism.
They managed to launch a reasonable number of shirts, and got great range, but the vast majority of the shirts were randomly spewed all over the stage and backdrop. The infrared sensors were their undoing, and an interesting case study in just how hard testing is.
They had done a lot of testing, in lots of circumstances. They had had problems the night before with the stage lights having enough IR to trigger the device prematurely. They had tuned and corrected for that. But it was the large, professional-grade photographic flashes that did it. No one had fired off one of those in the vicinity of the launcher during any test runs.
Just as in every branch of engineering, "shit happens", and it's incredibly hard to predict.