Game Mechanics Localization Crowdsourcing

Repost after a year with attempted better wording.

Localization suggestions are somehow unpopular, and when I see it being suggested in the forums, most of the responses would be along the lines of saying "cool but this would not work". Here is my take on how it could be worked on.

One of the easiest ways to implement localization is to crowdsource it. Crowdsourcing is basically to diffuse and give a large task to a large group (of volunteers) to complete, which is great for translating a lot of lines. See Minecraft, the base game Wynncraft runs on. All of its important strings are given a key for translation, and the game looks up the key on a map encoded in json format to find the string to display. All translations are crowdsourced using Crowdin during snapshots, which makes it fast to update and making it possible to have the game translated entirely at version release.

Wynncraft could adopt a similar system. Since almost all text like quest dialogue, item names, item lore, ids, mob names, NPC names, boss bars, status messages, sign text, needs to be sent from the server to the client at some point, they could be assigned keys for translation. The keys then could be used to look up translated text from a json/properties file of the client’s desired language before sending the text to the client. If the key hasn’t been translated yet, it will default back to English, just like Minecraft does. This would not require putting additional text files into the client resource pack.

All of the keys and values could be put onto a localization crowdsourcing platform such as Crowdin or Weblate (these are the ones I know) for the community to translate. The translations could be updated regularly (perhaps monthly) after passing a check for validity. As for the concern for secret content, the translation keys attributed to the secret content could be omitted from the crowdsourcing until the content has been verified to be found, such as being posted on the forums or Discord. This can be achieved by putting translation keys in many separate files and each file added to the crowdsourcing platform once it can be translated.

As for actual implementation after the files have been retrieved from the crowdsourcing platform, I do not know how current text works as I have never been in CT, but I suppose that all text except sign text must go through Wynncraft code. Therefore, a server-only implementation is possible, and to do this I suggest adding a single method to handle localization of keys, defaulting to English if there is no translation, and defaulting to key when the English translation is absent as well. This makes it easier to transition as it makes it possible to just pass text into the method when there isn't a key yet. There could also be a client-sided mode where the keys are directly used in localizable text sent to the client. This allows for custom language packs to be made, allowing the testing of unreleased languages in Wynncraft itself (and allowing stuff like badly-translated language packs)

The downsides to this suggestion would be server memory usage (though I did some simple calculations and I assume with 20 languages the size would be under 200 MB so pretty insignificant), possible extra time to get required text (doubt this would be a major lag source), the possible need to have additional personnel to verify that the translations aren't lazily/maliciously edited, and the costs of renting/maintaining a crowdsourcing platform.

TL;DR: Add translations using crowdsourcing platforms.

Might add images later

 

sounds good

 

What even is groundsourcing?

 

What is groundsouring? And how is it relevant to this suggestion?

 

Meant crwordsourcing ..
autocorrect somehow hit in ( im on a phone)

 

o hey we could have poorly translated wynncraft by editing the language systems lets goooo

 

anyways added a simple description of crowdsourcing

thanks for the input, added to suggestion

 

I had to read this like 5 times to understand it

 

Love the Suggestion I think it could honestly work

 

hello THE LMiffy1111

 

bumpbumpbump

 

one year importance bump

 

new year new bump

 

Bump, great idea!

 

one and a half years

 

unfortunately, this can be difficult, especially for item names and brings up questions. i mean, do we keep or change item names? is satsujin gonna stay japanese or change to ‘murder’ in other languages? however this is a great idea for quests and stuff.

 

I'd say that item names should be able to be translated, but whether or not that should actually be done should be up to that language's community.
For "satsujin" specifically, my opinion is that it should be kept similar to "satsujin" in most languages.

 

I believe we've stated a few times why it wouldn't work out. We encourage players to create or use modifications instead.

 

There are some problems with client-based translations:
1. Much of the game's content is one-time only, which causes the translations of these contents to be mostly unverifiable.
2. Client-based translations depend on string matching. Every patch usually contains changes to the game's text, which, even if insignificant, will cause the client translations to fail.

And, therefore, there isn't an active translation mod right now.

 

Perhaps see if you can gather volunteers who'd be able to actively translate stuff, or use a mod that generally translates the idea of stuff (using something like google translate) even if it is inaccurate