It seems to me the author is not suggesting to update the protocols themselves but rather to stop sending them CR even if the spec requires it. And to patch the corresponding software to it accepts simple newlines.