What stops a producer from reusing a barcode for subsequent versions of a product?
When you started to sell wine in 2019, you first acquired an EAN. Then in 2020, you decide to change the packaging and continue to use the EAN.
For MusicBrainz, the release of wine with new packaging would be interesting enough to create a new release entry for. However we can’t use the EAN to infer the specific release date in 2020, unless there is a database that would allow two separate records for the same EAN and the wine producer would care to submit this data.
Searching with the UPC today reveals two entries but none has date information.
But let’s assume the “1991-07-01” date is correct for that CD.
Digital Media release
This one I just added using a-tisket (cached version). I don’t know which of the three data sources provided the UPC, but I have to assume it is correct.
An ISRC search reveals “2007-04-19”.
Trying to make conclusions
Different barcodes help to distinguish releases, but one barcode can refer to multiple releases. Inferring release dates is not that simple, even impossible.
I assume the isrcsearch.ifpi.org site keeps only one record per UPC/EAN and it is up to the labels which release date they enter. This means the date can refer to ANY of various releases that may exist.
Maybe there is yet another CD release of the “Commodores” album, which is actually truly implied by this UPC. In this case it would be wrong to assign “2007-04-19” to the digital media release.
But we cannot know that.
Well no. The only conclusion I can make right now is this: “The UPC provided for a digital media release allows you to look-up a release date and regard that as valid, until somebody eventually comes along with a different version that has the same UPC. When that happens, everything becomes guesswork.”