The non-archimedean CCR ratio for efficiency analysis: A rejoinder to Boyd and Färe☆