Note / August 17, 2026
Measure the failure, too
A short argument for recording the dimensions of parts that did not work.
A failed part often contains a more useful measurement than the final one. “Too flexible” becomes actionable when the material, span, thickness, and load are still attached to it.
I now mark rejected prototypes with the same number as the drawing revision and record the reason before changing the model. This takes about a minute and prevents a surprising amount of circular work.