@kota1 - thanks for your reply. I think the argument perhaps needs to be separated out into two strands. If your essential argument is that - measurements don't reveal how a component sounds and that the reveal even less how a component will sound in a given system - then by and large, I agree with you. At component level (with some exceptions as set out below ( and for clarity, I mean for example, a CD player and not the individual components in the player) unless measurements reveal gross defects then they say little enough about sound quality and certainly do not describe sound quality in substantive way. The same applies cumulatively at system level.
The exceptions to the above are that certain measurements of amplifiers do correlate pretty well with subjective sound quality. For example, SET amps tend to measure pretty similarly in regard to their levels of harmonic distortion and that is subjectively audible. Secondly, there is a reasonable correlation between loudspeaker measurements and aspects of subjective sound quality. But for completeness, I am not arguing that those correlations are in any way comprehensive.
Going back to the basic argument, I still feel that it's a leap of logic from there to concluding that measurements are not useful. The progress of science depends on the formulation of hypotheses and repeatably testing and validating. The more engineers and designers measure, the better they can understand the relationship between their design choices and the subjective performance of their designs.