>> If I'm ripping an album, the obvious date I'd expect to get from the DB is the date for the album, not individual tracks on the album.
Doesn't this depend on the nature of the album? Sometimes the album is the original release of a song, but sometimes not. After a song ages a bit, it can reappear on all kinds of other albums and formats, ranging from the "Now" series to greatest hits to box sets and special editions and topical theme albums and other various artist collections. Which of these is the "real date" of the song?
I'd argue that the key date is when the song was originally recorded/first-released, same as the date when a photo was taken or a movie was filmed. This "real" date is the one used by reference works and popularity charts, and connects the recording to music and performer history.
When the song happened to be put into the marketplace on given piece of plastic might be of interest, but it doesn't indicate the age/vintage/history of the track. They can repackage "The Sound Of Music" 99 times (and probably have) but it's still a 1965 movie.
I always laugh when I do an online lookup and see a song date that is years after the song was recorded, often after the group split up, and even after the artist died.
Just as there are separate fields for Album name and Track name, how about a separate field Album Date, distinct from Track Date?
Sure it would be nice to have a consistent Artist name and Album, but which of the tracks could be presumed correct?