You find zero indexing logical most likely because you learned programming on languages which are zero-based. But most of the rest of the population, including statisticians, for which R is the intended audience, likely start at one and aren't used to OB1 errors.