Hi, everyone.
The official styleguide has constatnly avoided this issue. I think it would serve everyone if an official instance was taken.
I have coded a little php script that’s publicly available in https://etc.marlonob.info/mb-importer/ in an attempt to more accurately gather data for digital releases (the code was written for personal use, and I must clarify that I’m not a programmer). But I think in the process of codding it I may have gather some relevant insight on the matter.
My instance on this issue when I first joined MB was that most digital releases should be [worldwide], since in those cases, for example, when a release is available in every iTunes’ store, is clear that the intention is for it to be available everywhere and that the limit of some territories is a technical issue of Apple and not the artist/distributor’s.
But then it was pointed to me (and that make sense) that technical and legal issues have always determined different releases. There are albums released almost world-widely but in every country they have a slightly different copyright notice an probably a barcode also. Everyone of those releases have a place in the database. The case is, in theory, every release group should have dozens of different releases, one for each country in which it was licensed and sold.
Now, is it better for the database to have only a couple of countries or to have all of them? I think, the answer is the latest. The important thing is to have as much particular data as possible.
For those more interested in the album concept, the release-group should be enough for that. A concept of main release may be useful, similarly to how we have a main cover image automatically assigned to the first release on the list, but that can be mapped to a particular release in special cases.
But, first: what defines a particular digital media release? For physical releases, this is relatively clear: any difference in the packaging or the content suffice to make a different release. But with digital media, those differences are less available to the consumer.
Now, sometimes the vendors make little changes even to the track or artist names. Some of them include every credited artist as a track artist, some only the album artist, …. I have find that the the only reliable identifier is the UPC. The problem: most vendors make this data hard to access, if at all.
But, back to the matter of the thread. My script function under the assumption that a digital release is the same if, and only if, they share a UPC, and query Spotify, Deezer, and iTunes APIs looking for specific information on the release.
So, about the countries. Only Spotify provide a list of available markets. For iTunes, the only way of knowing is to request every iTunes country-store to see if they have it available or not. Deezer doesn’t provide that information.
Specifically about the Spotify API: No, the countries are not always the same. Most of the time, they are all the countries in which Spotify is available, but not always. Sometimes is only one country, sometimes, territories like North America (mx-ca-us) or or North + Central America or Latin America, etcetera. Now, what appears in those lists is the intersection of two sets: A) Countries where Spotify has pressence; and B) Countries where the specific-upc album was released. So, that may not be all the countries where it was released, but all of those countries have a valid release event, and that is what gets registered. For now, this is the most complete info that we have access to.
So, the question is: If not all knowed release events, what should the criteria be? Before I automated that task, I would (manually) test for three kind of countries: 1.- The artist’s area; 2.- The register’s area (ideally, the country in which the ISRC was registered), and; 3.- Mexico, which is my country.
Some people have a problem with seeing Mexico so prominently present (even tho I have never seen a comment when the us is the only country for releases by artists or labels from uk, au, or basically any other country) in an otherwise “not-mexican” release. The thing is, for example, maybe for a Norwegian person is more important to know which releases of Mutter by Rammstein are available in Norway for streaming that to know that there was a German release for an album by a German band. I don’t think that it has to be reminded that browsing it is not the only —not even the main— object of MB.
What I do find to be terribly missing from MB for digital media is a service for snapshots of vendors APIs / album pages. In the same manner in which the CAA is useful as a source to find the particularities of a physical release, having a snapshot for the sources of digital releases may serve the same propose and, given that the CAA is already maintained by the Internet Archive, I think this may not be so complicated.