This is a type of issue I’ve already encountered a few times, but the most recent is a release of John Coltrane’s Africa/Brass: https://musicbrainz.org/release/e4d4c059-3700-4826-895a-74d7c477484a
This specific release is lacking in uniquely identifying information: no label, catalog#, or barcode. There are no external links specific to the release, only at the release group level. What data is there matches a CD in my collection (including the disc ID and the specific title, which has varied a lot across this release group). I’m trying to decide which course of action is better:
- Assume this release represents the CD I have, and fill in the missing data accordingly; or
2, Assume it doesn’t, and create a new release for my CD, leaving the incomplete release alone.
I’m leaning towards #1, because otherwise this particular release will never get improved. If the data I enter doesn’t match a CD someone else owns, they can create a new release with their specific details.
I’d like to hear if more experienced editors have recommendations.
I have run into releases of which I couldn’t verify the existence, and in those circumstances I usually set the release quality to Low with an annotation explaining the situation.
EDIT: In your case since it’s a match I’d probably suggest filling in the rest of the data using the release in your possession.
Yes, if all the data matches yours, I would assume that it is yours. Otherwise no-one will ever know and it will remain useless. Perhaps by making your edits votable you could give anyone who cares about it the chance to speak up.
This is a very valid question. It points at a substantial problem of the musicbrainz data: The high number of low quality entries, lacking info allowing us to identify the exact release (often dating back to some freedb entry) or even having contradicting info. I have to say that discogs is better in this respect.
I agree: If all the thin info of a low-info MB release matches yours, just hijack it and make something consistent out of it. Otherwise, setting the release quality to low is always a good idea. In the past I even merged such low-info releases into better ones if everything was consistent. It should be a long term goal to improve as many of those low-info releases as possible.
The disappointing thing about the “release data qualtity” feature right now is that it is well hidden. I think we should put a lot more emphasis on rating and displaying release data quality as I wrote there. If you agree, please consider voting for this ticket!
If nothing in the existing data, or the edit history is enough to identify the exact release feel free to add data to convert it to a proven release.
If another release has exact same data, you may merge to it.
In worst cases, when data is very bad, set data quality to low.
If the release is very dubious, or lacking essential information, delete it.
Usually edit history gives hints about original editor intent, but in this case, there are plenty of releases.
We know it was added from freedb, it has a discid attached, and pre-date 18 Dec 2002, same tracklist as most CD releases in the RG.
Since the discid isn’t shared by any other release atm, i would just set this one to low quality, until someone with the matching discid complete the infos.
That’s interesting, I dared to do this only once in the past IIRC. Would you mind specifying what do you consider “essential information” here?
See few examples of cases where release removal was the best option from my edit history at https://goo.gl/TCTbzS (see edit notes)
Among cases are:
- Obvious crap
- empty release
- release proven as incorrect by artist and/or label
- release data heavily incoherent and no way to merge safely
- release added before actual release and never released after years