Template talk:Languagename/db

I'm not a fan of having such parallel/custom databases and language code definitions, they add confusion and risk to make any fix harder in the long term.

Indeed, I also don't think a template should be a correct place where to fix language assignment error, aka why $wgDummyLanguageCodes isn't enough for Norman?

This is not a parallel. But a complement because Mediawiki is far from being complete enough to support all languages supported on projects hosted on this wiki (where translation occurs even before a new project is created in Wikimedia). Not all projects hosted here are even based on MediaWiki or sponsored by Wikimedia.

This template uses "#language:" as much as possible but has been maintained since many years to cover cases not handled at all (and where "#language:" only returns the specified code as a really dummy "fallback" (which helps no one).

Note as well that this wiki still does not have any support for Lua modules in Scribunto (and we have no access to the PHP API of Mediawiki).

It just has the SemanticMediAiwki extension, which is in fact unused, and just causes problems, notably in standard Mediawiki categories, without solving any practical problems). It's most probably because it runs on a server with limited resources (and lack of very active administration: problems are often signaled and ignored for long, so we have to live with that).

I've not made any attempt to do more than what was already in use and fix existing problems with working workarounds that have a minimal cost and does not require a lot of admin tasks.

So may be you don't like this template, but for now no alternative solution is available, that would solve the Wikimedia needs as well as the needs for other projects hosted on this wiki. I've just filled a gap and maintained it with patient efforts since years (and very progressive manual updates). I've made lot of active searches to relevant standards to make sure all was OK in terms oif usability (and non ambiguous!).

So when you speak about "confusion" or "customization", you should first demonstrate them, may be there are some typos, or some needs for updates (if there were changes in relevant standards), but they are solvable. And for now most of these are not covered in CLDR data (the version used by Mediawiki) and CLDR data is also very slow to progress (and ISO 639 is defective by only supporting English, and not fixing a lot of names, and supporting absolutely no native names!)

So where is the duplication ? Nowhere. There's no existing parallel/equivalent. And for the long term, it has been stable since years, and was always easy to maintain incrementally. We may review this later, only if there's a new software support added on this wiki by its very few admins (that also need to check the available resources, that are limited by the generous donator company for the hardware and its hosting).