Think of it - someone changed their name. Artists do it all the time.
Yes, but under completely different circumstances. A trans person’s name change isn’t some Ferry Corsten-esque whim, it’s a process of abandoning what they were expected and raised to be, against who they really were, and having to be referred to by their deadname is often seen as a denial of their true selves by others, and believe me, having to see my deadname on official documents and bank statements is pretty numbing, as is the biases and stereotypes that more often than not come with being perceived as male as a result of my deadname when communicating with others.
yes, musicbrainz is a public site that anyone can browse. we should try to ensure data is presented in a respectful way
this isn’t data layer, it’s presentation layer
if “credited as” was displayed as it is on imdb, the fact that the deadname is a historical name would be obvious
…but that’s not
so i proposed finding a technical solution that would allow for more respectful presentation of “credited as” data, and a stopgap solution of moving data from a structured field that is presented disrespectfully, to a freeform field that where the data can be respectfully presented
that means someone on MB saw an issue that wasn’t getting enough attention, and now there’s a discussion happening about how to change that. seems reasonable to me.
also, using scare quotes around transition is disrespectful.
That this hasn’t been picked up by staff yet is crazy. @Freso? @reosarevok? Just make a call/raise it to wherever it needs to be, a polarised community discussion like this is just plain harmful.
That is a total mischaracterization of the situation, deadnaming is a very specific case that is pretty much unrelated to other artist “name changes”.
What is the “purpose” in seeking out trans artists and making edits that could be harmful to them? Do you have a collection with a lot of trans artists that you are trying to tag?
I want to comment on the two proposals made by @outsidecontext and @briaguya. I favour the former’s proposal with some caveats, but let me first explain what I don’t like about the latter.
Having a singular release with some kind of “credited as” solution, doesn’t reflect the current state of the release. Patricia Taxxon’s releases right now are solely credited to that name. Maybe a satisfactory technical solution for this specific case is possible, but this seems unnecessarily specific to me. It’s also not clear to me how this approach could be translated to physical releases then.
Onto the approach I favour. While I agree that having duplicate releases is a bit unfortunate, this seems like the only decent solution to me, if the old data is to be retained in the DB somehow. Technically no new release took place, but a lot of major things about the release changed (not only the credit but also purchase/download URL). I can see both sides here. Maybe to acknowledge the fuzziness of the situation, the new release could have the same release date as the original, just with all the data adjusted. I would certainly find it weird seeing all those releases of her’s showing up in the same year when browsing my music collection. This approach is also directly compatible with this situation happening for a physical release.
I share @briaguya’s concerns about deadnaming via the “credited as” field, though. Even if the “withdrawn” status for releases is introduced, it doesn’t feel quite right. So my question is, would it be possible to adjust the web interface, so that instead of just printing the direct credits, it does what IMDB does (X (credited as Y))? This might be a bit annoying visually on releases where, for example, the artist name and credit only deviate in capitalisation, but I think it would be worth it. Defaults like this matter a lot, since anyone who intends to build some application on MB, is going to look at how the project handles its own dataset first for inspiration.
The only thing I’m not sure about is if on the original release, the original download/buy URL should be kept. It is kind of deadnaming, though it was an important data point for the release, and more generally I’m not sure if dead links like that should be kept at all.
When a URL relation is no longer valid, which would presumably be the case for the official websites for a “withdrawn” release, it should be marked as ended, with the relevant end date if known.
i think that is one possible way to address the technical solution part of my proposal
i also think that might make a bunch of other releases look really messy (Artist A (credited as artist.A) feat. Artist B (credited as b), Artist C (credited as C the Artist), and Artist D)
since the “credited as” field is plaintext and doesn’t link to aliases, there’s no way for the frontend to know when it’s appropriate to use the “credited as” name exclusively and when it’s not
i think a way to tie aliases to “credited as” would allow for flexibility in presentation.
for example, an alias could have a:
“hidden” mode, where credits could appear as “Artist name [info]” where [info] links to the alias
“credited as” mode, “Artist A (credited as B)”
“replace” mode, where the alias is used exclusively.
all current “credited as” strings could become aliases in “replace” mode, as that is how those strings are currently presented. then, we just switch the deadname aliases to something other than “replace” mode and all those credits would be displayed respectfully
Personally, I think of an artist re-labelling a bunch of digital releases with a new name as being pretty much the same as reissuing a CD/record with updated packaging containing a new name. Of course the obsessive collector who likes to collect a dozen or more different variations of the same release might feel differently.
Rather than wait to be sure of a consensus in this thread, I went ahead and created a ticket requesting functionality for this: https://tickets.metabrainz.org/browse/MBS-11741
I think (most?) everyone in this thread should be able to comment in that ticket.
A trans artist doesn’t necessarily relabel or reissue anything. Transitioning is often an public acknowledgement that they were always x person. Being ‘deadnamed’ can put a transitioned individual in danger from others, as well as trigger dysphoria, which is connected with tragically high suicide rates. The ‘lie’ is often considered to have been the dead name, which was one upon a time used to put society at ease/keep an individual safe/allow them to function in society.
The above doesn’t apply to every person who transitions, it is a generalisation of a possible experience. Also that is my laypersons knowledge of the issue, I am very happy to be corrected.
This difference between ‘relabeling’/changing an artist alias and the issue of deadnaming lies at the core of this discussion - otherwise I would agree and we totally would just apply the MB guidelines for re-issues.
And I’d argue currently it doesn’t. If I look at https://patriciataxxon.bandcamp.com/ I see a lot of releases by Patricia Taxxon. If I’d like her music I might download it, all will appear as being credited to Patricia Taxxon. I might not even know her under any different name. Now I look this up on MB and I see that: Patricia Taxxon - MusicBrainz
This shows a totally different representation. It’s not just misrepresentation of the current situation, it is also bad and messy data. Above have been some suggestions on how we could deal with that, both within the current technical constraints and possible technical enhancements.
So, what is your proposal to fix the data? What do you think of the proposals so far?
I don’t know if I got it right but no, the new physical release date cannot be the original date.
It must be set to the date when this new version was released in shops.
It is a bit tricky here, really, and not as clear as it might look first. Technically those releases have not been freshly “released in shops”. They had already been uploaded and released on the digital stores before. They kept their IDs, track lists, audio files etc. The only change was the “artist” tag being updated. The joy of digital only releases
I can see both arguments. The more I think about it the more I am convinced that we actually should have only a single release with the new name, and if it was released originally under a different name mention it in the annotations (until we get a better way to note that). But I’d personally be ok with the two release solution if the majority would find this the better representation. I find it more messy really.
even in the case of a physical release that has deadnames in ink, we shouldn’t use the deadname in the “credited as” field. i know this goes against standard policy, but until MB can present those credits in a way that provides proper context and respects artists it’s deadnaming.
i don’t see the two release solution as providing a solution to that, there would still be pages showing out of context deadnames
as for digital releases that get updated, i think we’re running into the ship of theseus (edit: if only we had 4D releases)
Catalogers having to do ethics as well as fitting analog realty into digital boxes - good to see.
I’m very much on the side of “Don’t harm people, a recorded music db is just not worth harming people over.”
Well I’m on that side or get back to it usually if I just walk away from the keyboard for a day or a few months.
We currently have a work-artist relationship “previously attributed to”. What if we did something similar for digital releases? In this case that would mean a single release with Patricia Taxxon as the release artist, and a “previously credited to” relationship to “Patricia Taxxon credited as Eric Taxxon”.
The parallel is not exact. But leaving aside the work rel, a “previously credited to” release-artist relationship would let us recognize that (as @outsidecontext notes) we are dealing with a single release that changed attribution.
So maybe a new relationship to capture the old credits? Actually this is something that could work without having to wait for a technical solution that maybe even causes a schema change. Definitely something for @reosarevok to comment on once he is back from vacation.
I don’t see evil intentions here. Quite the opposite, a lot of attempts to improve the data quality on this artist. And yes, this discussion is about this specific case, let’s actually solve it. So how do we actually use these tools you mentioned to update this artist, so that it shows their releases properly reflecting the way the artist makes them currently available while still allowing the historical credits to be available?
What seems rather widely supported is changing release group and recording credits to the current name. Also got some discussion about this on IRC again. I think that is clearly something we should do.
We actually miss a tool to clearly communicate that a release had been originally credited differently. There have been a few proposals above on how to deal with that with what we currently have or if that doesn’t work what we would need.
We could change the credits to the current name, and mention the original name in the annotation. That would mean the release list properly reflects the current releases. But the original name would be unstructured.
We could do the same, but have an explicit release-artist relationship, something like “originally credited as”. That would make this data available in a structured way and also allow for dates to be set. Advantage is that a relationship is rather easy to add.
We could duplicate each release. Now we can argue if that really is a new release when basically nothing was changed, the release still having the same files and IDs in the shops (also same links then). But it would allow for both credits to exist. It would still show the deadname prominently, though.
Improvements could be made to this if we would have the “withdrawn” status for releases. But this looks like a bigger change, maybe even schema change. Other idea was having kind of artist credits aliases, so one can have multiple alternatives to credits. Again likely a bigger database change.
