You know where I came down on that question!
Although I do concede that there is still a place for subjective evaluation when it comes to speakers. Not that measurements are incapable of telling nearly all where they are concerned, but that doing complete, valid measurements is very difficult, time-consuming, and beyond the means of all but a handful of laboratories (and probably techs). The Harman Int'l. facility described in the recent speaker articles is one.
And hey, some folks genuinely prefer a bit of boost or attenuation here and there in the frequency response. Perfecly legit even if it isn't my thang. But few (including me) have the training to say, "oh, I'll probably like the sound of that speaker because it has an attenuated response by 3dB between 2000Hz and 4000Hz and a 2dB/oct rolloff above 10kHz". So the ears still have a role there...at least for everyone who doesn't have their own anechoic chamber and calibrated test gear. (My wife keeps saying "no!")
Oh, and for the "is audio science or art" question: Science, period. In the olden days when hi-fi was aborning and technology and measurement were cruder, I'd say maybe there was some art in there as a matter of necessity, but not today. Ditto for the disciplines on which audio engineering is based, from the "hard" sciences like physics to the "soft" domains like psychoacoustics. Only the music we listen to is art. Audio, as Dr. Floyd Toole says, is science in the service of art.