Reality Check: We DO fly on airplanes with systems that are coded exactly as the author describes.

Unless you're claiming that all version one software is perfect and therefore remains in service forever.

Stop trying to be clever and argumentative, and think. Maybe give the article another read as if you didn't already know better.

Software development -- all of it -- is in most cases a long endeavor of learning and discovery. Unanticipated problems are discovered. We learn from them, fix them, and things get better.

The entire point of the article, which you seem either to have missed or to be working very hard to avoid, is that it is impossible to determine mathematically whether code is defective, AND it is impossible to test everything; THEREFORE, we here in the land of reality, are stuck with learning and improving.

I'm all for rigor. And in the case of Boeing you certainly have a point. But the point of the article is much more general, nor does the article EVER advocate sloppiness. It's an article about how to improve DESPITE a lack of absolute certainty.