<font color='#000000'>Three things come to mind:
1) An unrealistic basis of division between two camps.
To wit, I have not come across anyone who proposes solely to rely on measurements to determine accuracy in audio equipment. Where are these people? As is pointed out, to rely on measurements presupposes the measuring equipment is more sensitive than our hearing--but this already supposed Just Noticeable Differences (JND's), which are, of course, based on Double Blind Tests (DBT's). JND's, of course, vary with the frequency and the level.
Now the DBT does not presuppose much in the way of measurements, except what is necessary for level matching within, ideally, 0.1 dB. Either one can tell the difference or one cannot. However, it is most useful if one can correlate the measurements with measurements. This might look to a subjectivist like relying on measurements, but it in fact implies a correlation with listening tests.
I think the more real basis of division is whether, to determine whether a difference is audible, one relies on uncontrolled, sighted listening comparisons or whether one relies, ultimately, on DBT data. Of course, those who rely on DBT data agree with the mantra "trust your ears," and in fact, are the only ones who do so, as sighted listening relies on human built in biases to make choices, cultural things like reputation, brand name, appearance, and lack of controls.
Of course, there are cultural differences. The subjectivist may want one to do it all oneself, what I call the "hair shirt" approach, which they often oppose to reliance on reading. A real man (most audiophiles are men) will do it himself! As if, in fact, any of us do not rely a great deal on things learned from others. Of course, when I read the similar terminology and thought patterns, I know that the subjectivists views are spread in great part by the written word or other forms of communication, so this is not only irrelevant but silly.
There is also a romantic kind of bias, the reliance on subjective experience rather than soulless machines. But this really relies on a mischaracterization, as the rationalists in fact rely on correlations between DBT results and measurements.
Another objection is that long term listening is needed. Well, this may be a learning process, but once identified, differences should be audible in DBT's. Besides, I think Daniel Shanefield did some long term DBT's, and found they didn't produce different results.
But these are side issues as to whether the most accurate determinations of small audible differences is signted listening or controlled DBT's. Everything else is peripheral.
2) The example using digital cables is not clear, as it seems to mix up concerns regarding regular audio patch cords. Why would a digital cable with a rolled off top end sound "smoother"?
I don't grasp this, but then I am not a techie.
3) The number of trials you suggest seems much too small. From what I have read, those who do such tests suggest at least 16 or 17 trials. For some reason, I can no longer find the ABX website, so I can't find the tables, references, and so one.
So, I think the article needs some modifications.</font>