- even such an organisation as LAST, performing and rotating around socially\cooperatively producing first, then also tagging, classifying, measuring and clustering all the informational strings set attached to musical units (songs, tracks),
- EVEN LAST falls into such a stupid trap as, apparently, the proposed but useless measure of compatibility across any 2 users. Or: not knowing how to deal with such a basic, so primary task.
Instead of being informatve & synthetic, what LAST suggests here is disinformative and disturbing, it just creates noise. 2 ways out are possible amogst MANY; they should B both produced, in place of the current rubbish:
(a) a (so -called, although strictly nothing ever is) objective compatibility, using some reasonable light weight algorithm; the most interesting experiment would be to introduce here the general average of "self-declared" individual weightsd sub-b; MUCH better than the totally false and cumulatively misleading "one counts for one" assumption (a paradox in music: were it true, a typical music lover should listen all his life to the entire universe of songs, but nener twice the same song he might prefer).
(b) A dual inter-subjective compatibility. For example: each one DECLARES the weights (I prefer this to teh bserved behaviour within last, for a vast array of reasons: 1st of all since I'm in search of pure subjectivity here). What matters most for him: I might say 1\3 Amy WIno, 1\3 Glenn Gould and then attribute the other 1\3 to all the other performers. When I am compared with e.g. a Lady A., the self-declared weights of the two matter equally and determine together how much matters the presence\absence of one artist or another (if she doesn't metion Glenn Gould at all, he will affect 1\6 of our inter-subjective compatibility index). THIS WOULD BE THE "REAL THING", in association with some attempts to quasi-objective (a) measures.