Encoding the language that a work name is written in

[alastairp]

This seems somewhat related to What language to use for classical work names? and Language consistency in work names.

I’m working on a project where we are importing metadata from MusicBrainz into a semantic-web database. In the schema.org ontology, MusicCompositions have a language field which represents the language that the name is written in.

For example, we have the following works:
https://musicbrainz.org/work/17f624e2-97da-45c3-ab54-c3b2daf2ba68 (russian)
https://musicbrainz.org/work/14514e77-f06c-4d24-a4d5-9a5962acae64 (german)
https://musicbrainz.org/work/f334182c-42f7-4808-9576-6c73efe2c632 (english and german)

however, in none of these cases we store the language that the work title is written in.
Sleeping Beauty has an english alias, but nothing to say that the original title is russian.
Das Lied von der Erde is in german, but the wikipedia description has a translation. However, the wikidata page shows that the name for the english page is still the german.
Eine kleine nachtmusik has aliases for english (A little night music, search hint) but the locale isn’t set. Similarly, the german aliases don’t have a locale set either.

My specific questions are:

• In the case that a title is written in a non-english language, should we also add that exact title as an alias with a primary locale in order to be able to identify this language?
• In the case that a title in a non-english language can be translated to english, should that translation be added as an alias with a locale (with or without primary)?

If we can come to an agreement on a policy here then it’s possible that our project can find some development/crowd-sourcing resources to start adding this data. I’ll have to confirm with the rest of the project memebers.

[edit]: some related information in https://tickets.metabrainz.org/browse/MBS-7830 which looks like the kind of direction that I was looking towards.

Thanks!

[jesus2099]

The language(s) of an MB work is the language(s) of its lyrics.

We don’t store the language of its main title, only the ones of its aliases.

[yvanzo]

@jesus2099

The language(s) of an MB work is the language(s) of its lyrics.

This topic is not about lyrics, it is specifically about work’s title, on purpose.

@jesus2099

We don’t store the language of its main title, only the ones of its aliases.

We can store the main title in aliases, so we can store its language. This is what he described.
Here is a recent example (with a recording): “Maighdeannan Diùraigh” a.k.a. “The Maids of Jura”.

[loujin]

@alastairp

• In the case that a title is written in a non-english language, should we also add that exact title as an alias with a primary locale in order to be able to identify this language?
• In the case that a title in a non-english language can be translated to english, should that translation be added as an alias with a locale (with or without primary)?

I would say yes and yes.

Unfortunately adding aliases probably takes too many clicks at the moment, so people don’t add them on works. I thought about two improvements to add aliases faster:

• mass import from wikidata localized page names (if they are different from the main work title)
• convert related recording names to work aliases with an existing userscript. People still need to add/check the locale by hand though.

If you have a way to import names from a curated source, great

[jesus2099]

@alastairp

In the case that a title in a non-english language can be translated to english, should that translation be added as an alias with a locale (with or without primary)?

If it’s self translated or if no releases are published with this translated title, then, it’s not primary.