Supposedly while building the backend, they realised the actual summary data for a reasonable breadth of tracks (say, anything you'd likely hear on the radio or on a jukebox) was tiny and so, why build a service at all when you can just ship the data to phones ?
Recently for whatever reason I was listening to the twist/ cover "Stacy's Dad" and Now Playing recognised it as the rather more famous original Fountain's of Wayne "Stacy's Mom". So yeah, it doesn't know everything. It also doesn't recognise lots of obscure stuff I own that's like B-sides or special editions that never saw radio play, or bands that my friends were in (but everybody I know has read both Steve Albini's "Some of your friends are probably already this fucked" and the KLF's "The Manual" and so none of them signed a recording contract and thus you've never heard of them) but I've never had a situation where I heard something I liked at like a bar or somewhere and Now Playing didn't know what it is.
yeah packaging the data and updating it async does make a lot more sense. Also, I guess its fine that it doesn't know it all but covers a good percentage of requests
Recently for whatever reason I was listening to the twist/ cover "Stacy's Dad" and Now Playing recognised it as the rather more famous original Fountain's of Wayne "Stacy's Mom". So yeah, it doesn't know everything. It also doesn't recognise lots of obscure stuff I own that's like B-sides or special editions that never saw radio play, or bands that my friends were in (but everybody I know has read both Steve Albini's "Some of your friends are probably already this fucked" and the KLF's "The Manual" and so none of them signed a recording contract and thus you've never heard of them) but I've never had a situation where I heard something I liked at like a bar or somewhere and Now Playing didn't know what it is.