Comment by ddxxdd

The difference is that we know for a fact that the proper implementation works for integers up to 2^31, whereas the improper implementation deceived us into thinking that the code would work in situations where the code actually doesn't work.

I find it valuable to understand when my code will crash.