I really donnt buy it is a "trivial idiom" to check unsigned undeflowing by writing 'i > size'. I'd have to check that was right all the time. Also in general

I really donnt buy it is a "trivial idiom" to check unsigned undeflowing by writing 'i > size'. I'd have to check that was right all the time. Also in general saying all we have to do is (I'm paraphrasing a bit) "not make mistakes" is, to me, a bad idea, as all programmers make mistakes. I sometimes assert after a calculation that a value is >= 0, it's much harder to check that if it's underflowed because it was unsigned.