John Berryhill, writing on the Registrars' Mailing List: "The notion that performing a data processing task in 2005 substantially identical to the same data processing task on a per unit basis performed in 1998 would be more costly is a proposition that could only be believed by an idiot."