I disagree. I feel the parent did a great job of bridging explaining why the math is the way it is, and why our "intuition" is wrong.
Perhaps that was my fault. I don't think I was very clear. I was disagreeing with the statement that 'NaN' means the output of a function is not defined. That was the only thing I was disagreeing with. The rest of my comment tries to address some of the differences between the mathematical side and the computational side.
Is your point of disagreement down to the difference between "f(0) is undefined" versus "f can not be evaluated at 0"?