8 years ago
8 months ago


(Reporter: openjck, Unassigned)



(Whiteboard: [type:feature][triaged])



8 years ago
Localized tags, categories, etc. It needs to be possible, for example, to say "the tag 'Foo' in English is equal to the tag 'Le foo' in French". This would allow a list of articles tagged "Le foo" to list both articles in French and in English for articles not yet translated.
Priority: -- → P4


7 years ago
Blocks: 728417
Priority: P4 → P2


7 years ago
Whiteboard: u=localizer c=localization p= → u=localizer c=localization s=2012-06-05 p=
I think this is a new feature, at least with respect to MindTouch
Blocks: 756266
Whiteboard: u=localizer c=localization s=2012-06-05 p= → u=localizer c=localization s= p=


7 years ago
Priority: P2 → P3
Component: Website → Localization
Version: Kuma → unspecified


6 years ago
No longer blocks: 756266


6 years ago
Duplicate of this bug: 818485


6 years ago
Priority: P3 → --
Summary: Localized tags and categories → Make tags localizable
Whiteboard: u=localizer c=localization s= p=
Whiteboard: [type:feature][triaged]
Severity: normal → enhancement
See Also: → bug 776048
Blocks: 1167992
Duplicate of this bug: 1167992
Duplicate of this bug: 1215445
This is still a desired feature.

Tags are not translated. Instead, a new tag is created, where the tag name is a translation of the English tag. There's no link between the "English" tag and the "Translated" tag.

This causes real problems when the tag is used for organizing content:

* Content-collecting KumaScript macros like {{SubpageMenuByCategories}} need to be passed per-locale tag variants, or localizers have to set both the English and "translated" tags. As reported in bug 1215445, this means that a page like is fully populated in English, sparsely populated in French (, and manually copied in Chinese ( Reliable tag identifiers are a requirement to making these macros more effective.
* Search categorizes content by tags. This is a bad way to categorize content, and it also makes translated content even less searchable than English content. A strong tag identifier would improve non-English site search (but not as much as re-writing search)
* Various "Doc Status" tools on MDN rely on tagging, and are broken by default when the localization team doesn't apply English tags to new pages.
* Tags that only differ by an accent character, such as the French "Référence" versus "Reference", are not well supported by MySQL. This was handled with a custom collation in the past, but this was removed in bug 1391084 to allow using unmodified MySQL. It is currently impossible to "Translate" these tags into some languages.

A better system would be:

* Use the English tag (such as "Reference") or the non-translated tag (such as "HTML", "CSS") as the strong identifier.
* Mark a tag as localizable ("Reference") or not localizable ("HTML", "CSS").
* Pass the localizable strings to Pontoon for translation.
* Apply the translation in interfaces, such as rendering or editing.

This system would be much easier to implement if there were a limited set of tags that were controlled in source control, rather than in a database, but if rapid addition of tags is a requirement, then a database-to-Pontoon processing step could be added, similar to search topics (which is also a bad idea).
You need to log in before you can comment on or make changes to this bug.