Yeah it's just weird that a new day starts at 12am, and that the jump from am to pm happens from 11 to 12. If we made the day start at 0am, the noon start at 0pm, things are just logical. If you take x am to be x hours since the new-day marker, (or x pm to be x hours since noon), then you'd naively expect 12am to be noon. You could argue that we work modulo 12, but the extra am/pm bit means that it's not fully the same as working in Z/12Z.