When we first imported music data, we made some data modeling decisions based on not wanting to over-stress the database. These considerations are no longer relevant.
One of those decisions was that we don’t have a clean division between albums and releases. We identify a group of releases as all being of the same album; we then pick one of those releases and designate it as the album. The other releases are then releases of that album, but the primary release is implicit, bound up with the album itself.
I’d like to change this, making the album explicit and separate from the release. The way this would happen is that every album would be duplicated; the new entity would be a release, and would be marked as the first release of the album. The MusicBrainz identification code—which is really tied to the release, not the album—would be migrated to the newly-created release.
Any comments?