Not release script related but another interesting tidbit to consider what “script” means in MB I guess:
https://musicbrainz.org/edit/112832785
As author of that “modest proposal” you linked to, this discussion reinforces my belief that the Script field of Release is on the whole making the MusicBrainz project worse, not better. It adds little or no value, and it causes people to spend time and energy fretting about questions like this one.
Dropping the Script field makes it impossible for software to differentiate Pseudo-releases from each other, although the lack of a meaningful Language field means that dropping the Script field does little additional harm.
Boy, I sure don’t follow this reasoning. I very rarely encounter pseudo-releases, so maybe I just don’t know enough.
Reading Style / Specific types of releases / Pseudo-Releases, and Relationship types / Release-Release / Transl-tracklisting,I don’t see anything that says the design requires a Script field for software to differentiate Pseudo-releases. Pseudo-releases can still have their Language fields filled out. An editor who chooses not to fill in a Language field is unlikely to fill in a Script field any better. Software can still identify the set of pseudo-releases for a main Release by following the Transl-tracklisting Relationship.
Could you please give examples of software which does differentiate Pseudo-releases now, and how that software uses the Script field? I think it would help me understand.
The pseudo-release use is that you can have, for example, Japanese in the Japanese script(s) vs transliterated Japanese in Latin, which would be the same language, just a different script.
Note,
- even though we can select multiple secondary release group types, Style says you should only enable a type if it represents most of the tracks on the releases.
- even though we can select multiple lyrics languages, Style says you should only enable a language if it’s used for an extremely large percentage of the lyrics
Presumably, the reason for this is because we want these fields to serve as disambiguations between objects at the same level on the same artist, rather than simply indicating that some portion of the object is a thing. So I’d expect that a hypothetical schema change that allows an MB Release to indicate multiple text languages and scripts (rather than the fake and useless existing “multiple” option) should not mean that we select all scripts used in text of the tracklist if that’s not a reasonable representation of the tracklist as a whole and when compared against other tracklists in the same release group.