Maybe just a carbon-copy of the HTTP Accept-Language header so that users can specify weightings? No need to normalise the data into fields for querying; this is for making user-specific rendering choices at run-time.
Default to empty; probably hide this preference on mono-lingual wikis (i.e. non-Wikidata/Wikifunctions/etc.).
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Open | None | T290537 Use a user-defined list of languages to select the languages to display and prompt for | |||
| Open | None | T268675 Decide on how to get the user language settings | |||
| Open | None | T290595 Let logged-in users define a *content* language preference array (as well as the existing interface language setting) |
Which problem does this solve? I can more or less guess some, but it needs a clearer definition.
On monolingual sites like Wikipedia, it's not so relevant because there is just one language. There are people who choose an interface language that is different from the content language on these sites, but I'm not entirely sure why. I'm not aware of any research about this. I do it sometimes because I want to understand how something works there and I don't know the interface language, but I don't think that it's a really common scenario, especially for logged-out people. And there may be some people, possibly even the majority, who just did it by mistake. There may be more reasons, but I'm not aware of any.
On multilingual sites like Wikidata, is there a good reason for interface and content language to be different? Maybe; I've never felt this need strongly, but perhaps other people did. I can imagine a situation in which a user wants to keep using Wikidata with German UI, but to see item and property labels in French at least temporarily. But is changing preferences the right solution for this scenario? Maybe, but as I wrote in the beginning, it requires clearer design research and problem definition.
For other multilingual sites like Wikifunctions, Commons, and Meta, the scenarios and the problem definitions can be different, too.
In my brain, the interface and content are quite separate things. Why would my mental model of how the interface works vary according to which language I happen to be working in? As has often been observed, native speakers of under-served languages tend to be bilingual or multilingual. If they are flipping between content languages, having to change the interface language as well is an additional obstacle.
Please note the experience of another user in the Telegram interaction preserved here:
https://www.wikifunctions.org/wiki/User:GrounderUK/page#cvictorovich_interlude. It’s not easy to understand how they would work in an ideal world, but they can speak for themself.