Last modified: 2014-03-07 11:44:06 UTC
As far as I can determine all fonts are connected to languages right now. But shouldn't they be connected to language-script combinations ? And shouldn't we 'expand' a language to it's default language-script combination before matching ? I imagine for fonts you would have: A lang attribute: "zh", would be expanded to "zh-Hant". A lookup occurs for a zh-Hant match. A lang attribute: "zh-Latn" lookup though would match a defined "*-Latn" setting and avoid usage of a zh font for a latin script. This is one particular area where the browsers are still FAR FAR behind. In 2010, I even disabled the lang attribute on our (en.wp) transliteration template, simply because it was causing more trouble than it was worth and it would be nice to reintroduce this. It seems that FF support for it at least hasn't made much progress: https://bugzilla.mozilla.org/show_bug.cgi?id=192636 This would also make it easier to support IPA, Fraktur fontstyle (*-Latf) and similar differentiations and transliteration forms.
Nice idea, Derk-Jan. If you'd like to take a stab at it, please do. At the moment, we cannot prioritize it high in the Wikimedia LangEng team priorities.
