My guess would be that non-English categories will complicate the introduction of such a system significantly. This issue should be considered. –Cwbm (commons) (talk) , (UTC)
Outputs of Google language tools are rather funny than usable still
After returning to Commons after editing at non English wikis the page displays in the language of the site you’ve just been at, with that language used for the header, side bar etc, the page continues to work in that language until you begin to click through the categories. The category and the category tree itself would not be renamed or duplicated it would just be translated into the language that the user set as their default, a linguistic filter through which the unmodified page is viewed, like browsing through a website using the translate url option on Google language tools. KTo288 ( talk ) , (UTC) Just had a go with Google language tools English to Chinese website translation and navigating seems to work okay, but I wonder how much money they had?
KTo288 ( talk ) , (UTC) On the other hand https://paydayloanalabama.com/stockton/ we don’t have money but we do have a lot of public minded volunteers, Google languages has a “suggest a better translation” option, a very wiki attitude to the world, we don’t have to have it perfectly right the first time, as long as we create a framework in which an ever better version could be created. Its more of Meta thing but a Wiki-translation tools would be a benefit to other user communities other then Commons, it could potentially allow all the articles in all the wiki projects to all users. KTo288 ( talk ) , (UTC) Which has all inspired me to createProposal:Wiki language tools at Wikimedia strategic planning, thoughts and suggestions? KTo288 ( talk ) , (UTC) Good thoughts.
If you want only translations on the fly in one way (from English to whatever), solutions are rather simple and straightforward, but the user will always see the categories in his language but he will be unable to use them for entering in the system. If you want to make it work in two ways, then substantial changes will be needed, along with a more formal and strict naming method (and a database with keywords and prepositions). That’s one of the reasons why I need quite some time to formulate a more precise naming rule proposal. The user could edit in his or own language to a layer which would then be translated by a bot, which would create a script to be acted upon.
An analogy would be like placing an transparency sheet over an image and annotating the image by writing on the sheet. Once the user has finished on the transparency sheet the sheet would be passed to a second party who would be use it to create instuctions for amending of the image, and once these instructions have been completed they would be passed to a third party to act upon. My experience of bot translation is that they work best with short and formulaic phrases, so to work best this system would greatly benefit from the strict and predictable naming system you propose. KTo288 ( talk ) , (UTC) : A translating of terms is very sophisticated activity which cannot be made automatically.
How hard would it be to extend this to the categories
Only a transliteration from some non-latin-based writing systems (or back) could be made automatically. As I mentioned above, every category should have its own prie and other variants should be an additional help only. In order to retain a possibility to make understood with one another, the prie should be displayed always and the select localized name should be displayed as an additional legend and a subsidiary tool only. Of course, there are many names which cannot be translated to some language or which haven’t a special English form (see Umgebindehaus as an example). It is not reasonable to anticipate that the prie can be allways in English. However every category should have only one right and prie. The equality of all language forms of the selfsame category name is utterly unrealistic idea.