Asserting that it's impossible or meaningless to measure the quality of a UI is to deny the entire field of UX research. Yes, it's subjective, but you can meaningfully measure the average across many humans. And that's exactly what you should do if you're trying to decide on a UX that will be used by many humans.
Rather than it being meaningless to measure the quality of a UI, I meant that even a UI that tests well generally might not be the best UI for a particular person.
