iTunes’s “smart playlists” make listening more enjoyable for me; I like to hear songs I have’t heard in a while, and it’s easy to make a list that sorts of selections that haven’t played in the last month. When I want to listen to familiar, favorite music, I can rig that, too.
I look forward to the next step in playlist intelligence: a playlist that distributes the frequency with which I hear selections by my rating of the song (at a crude level, I’d hear five-star songs five times more often than one-star selections) cross-factored against how recently I’ve heard the selection (or how often I’d heard it). I’d thus be most likely to hear a five-star song that I haven’t heard in a few months, for instance, and least likely to hear a no-star song that I just heard yesterday. At the same time, it wouldn’t eliminate the chance that I’d hear a less-favorite selection, or a recent-repeat. Over the long haul, I’d hear my favorites most often, but mixed in with other selections I like well enough, and with occasional less-favored cuts.
It ought to be do-able (it may even be possible now with iTunes’ capacity to nest playlists) — and it would really, really rock.
An option like that doesn’t sound far faetched at all… Take a look a smart shuffle, that’s basically the same thing except it uses the song or album metadata.
Exactly, Nick. I may experiment with putting an array of Smart Playlists into a single folder. You could make a folder that included one playlist that picked, say, five random five-star songs, one that selected four four-star songs, and so on — but it still wouldn’t adjust for flexible play-count selection.