Overflow can happen and we test for it wherever possible (at least that's the hope). By using unsigned lengths, the overflow tests are simple: x + y < x. We als

Overflow can happen and we test for it wherever possible (at least that's the hope). By using unsigned lengths, the overflow tests are simple: x + y < x. We also don't have to worry about negative values which, in my experience, are at risk of being forgotten more often than excessively large ones.