After reading through this thread, I’m in strong agreement with Augmente: If Bandcamp doesn’t provide a UPC on a digital release, the only real info that provides is that the UPC was not entered when the release was published.
Much of this debate seems to boil down to open-world vs closed-world assumptions, and I’d argue that Musicbrainz always needs to assume open-world, especially here.
Within digital distribution, it’s a de facto requirement that all releases have UPCs and ISRCs. This is uniformly enforced to the point where you can treat (UPC, ISRC) as a natural key for a track on a release, and UPC as a natural key for a release. Just because distro happens to work this way doesn’t mean the rest of the world does.
Compare recordings- mb doesn’t treat ISRCs as defining recording identity: ISRCs are overissued, not actively curated, never merged, not public, not resolvable via a centralized registry, have blunders, etc. It shouldn’t be surprising that Musicbrainz’s definition of what constitutes a recording differs from IFPIs.
Regardless, an ISRC is a useful handle for finding and referring to a recording. But it doesn’t define it, and it would be nuts to, say, block recording merges or require splits based on mismatching ISRCs, let alone missing ISRCs.
Getting back to UPCs - I don’t think Bandcamp exposes it, but they do optionally collect works registration information and ISRCs for tracks on releases. Let’s say the info was available -
- Would it be a good assumption that a recording missing ISRC data has never been issued one?
- Would it be a good assumption that a track without works details doesn’t contain any registered works?
Why make the same assumption about UPCs on releases, especially when Bandcamp is basically the only platform where you’ll see releases without UPCs anyways.
