> So how do you avoid getting nerd-sniped when something really is just bad? This is not logic, where existence of a counterexample invalidates the general arg

> So how do you avoid getting nerd-sniped when something really is just bad? This is not logic, where existence of a counterexample invalidates the general argument. One reason for looking at even genuinely bad code is that it may express domain knowledge that you are currently unaware of. Jon Bentley told the story of a Bell Labs researcher who found that his fast and correct algorithm for something or other had been "fixed" and "improved" so much that it was no longer efficient (a few comments may have helped here - though not necessarily - but that is a discussion for another time.)