Well, there’s an interesting wrinkle. When the linear coefficient b is large relative to the other coefficients, the quadratic formula can give wrong results when implemented in floating point arithmetic.
No, there is no wrinkle whatsoever: your computer, guided by your assumptions and given inadequate input (let alone processing logic) got it wrong.
The day you start to blame inputs for errors in output, for which you have complete control is the one that you should mark as a good point to really consider whether you are in the right game.
He’s not blaming the inputs. He’s simply saying that some inputs can produce incorrect outputs. He’s not saying the incorrect outputs are the _fault_ of the inputs. He goes on to explain where the blame actually lies, in floating point precision.
No, there is no wrinkle whatsoever: your computer, guided by your assumptions and given inadequate input (let alone processing logic) got it wrong.
The day you start to blame inputs for errors in output, for which you have complete control is the one that you should mark as a good point to really consider whether you are in the right game.