Any old math with consistent types and incorrect calculation? (If only we'd diff'd the output with some known result.) No doubt we're at cross purposes. I'm a

Any old math with consistent types and incorrect calculation? (If only we'd diff'd the output with some known result.) No doubt we're at cross purposes. I'm a little surprised that you haven't mentioned — "Program testing can be used to show the presence of bugs, but never to show their absence!" (or existential versus universal quantification).