I don't think that goes against the rule. Your code doesn't impose an arbitrary limit on the data, it just internally represents it differently based on size.