Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Translations should follow voting approval rather than Crowdin’s machine translation recommendations #1350

Open
guobao2333 opened this issue Nov 29, 2024 · 4 comments

Comments

@guobao2333
Copy link

Thank you for your development! This has helped a lot of players, and what I'm about to say is not a criticism of you, but my opinion as a translator/proofreader about the dissatisfaction.

As the title says, this problem is very difficult in Chinese because there is a standard in the field of Chinese translation, which can be summarized in three words:

Translating them into English is: Faithfulness, expressiveness, elegance. At the same time Chinese is an ideographic language and does not have quite the same expression order as other epigraphic languages, but this can be worked around through the magic of translators.

First of all, you need to know that the Simplified Chinese of this project on Crowdin is almost entirely machine translated, which is a disaster for users, serious translators and proofreaders.

Secondly you need to know that the unofficial translations are even better than the garbage translations on Crowdin, after all, they are the result of a lot of effort by the translators, whereas these pirated versions do nothing more than change the language a little bit, they are called "汉化版(Chinese language versions)"

@sds100
Copy link
Collaborator

sds100 commented Dec 4, 2024

This is a good point, thank you for the info and respectful explanation 😅. I didn't realise most of the Simplified Chinese translations were just the machine translations - I was expecting/hoping people were writing them manually. What should we do? Unapprove the translations and have people do them again?

I'll disable machine translations in the project.

Edit: it seems like machine translations were disabled. But i remember people submitting Simplified Chinese translations many many years ago so perhaps I didn't have it disabled back then.

@guobao2333
Copy link
Author

I'll disable machine translations in the project.

I think one more thing you need to know is that people are going to use pirated versions, is what was Chinese language versions as mentioned above, and no one cares about the official version, because since the official version didn't even have localized translations many years ago, and only merged those blunt machine translations a while ago, many people would rather use the old version with quality translations, than the official new version.

And you don't need to do much, you just need to do what is mentioned in the title, it maybe not only be Chinese is machine translated, other languages may also have this kind of problem, you don't need to turn off the translation suggestions provided by Crowdin, it's just that I think you are overly relying on the suggestions provided by Crowdin, please try to approve the translations that are higher in the number of votes, And it is not necessarily closer to the suggested ones.

@sds100
Copy link
Collaborator

sds100 commented Dec 8, 2024

Would you recommend I delete the Chinese simplified translations? Several people including you submitted translations over the years, is there one person's translations who you would recommend the most?

@guobao2333
Copy link
Author

Would you recommend I delete the Chinese simplified translations?

I think the current localization contribution specification may need to be modified, for example, in many languages other than Chinese, there are different ways of writing KeyMapper, and even different ways of expressing it after translation. For example, many languages other than Chinese have different ways of writing KeyMapper, and even different ways of expressing KeyMapper after translation:

  1. 键映射器: regular direct translation.
  2. 按键映射器: added emphasis on as an action to disambiguate.
  3. 按键映射: same as 2, but removing is equivalent to removing the suffix er, which is allowed in Chinese because it doesn't prevent the expression of the subject being referred to.
  4. 外设映射器: Although 外设 does not mean key, but because the keyboard, mouse, joystick and other tools are collectively called "peripheral", so it can include key.
  5. 外设映射: same as 4, but with the suffix er removed, here the same as 2.

From the above example, there are many variations in Chinese, resulting in translations of varying quality, hence the need for the above mentioned: 信达雅, in Chinese it is permissible to omit a lot of things ...... as long as it doesn't affect the subject that is ultimately being expressed it is permissible, in English there is a similar situation, but the Chinese language can be even more arbitrary 😂 Ancient Chinese (a.k.a. 文言文) is more elliptical in its expression, but today it is simplified and uses clearer expressions for the sake of 扫盲 (elimination of illiteracy) and is thus called Simplified Chinese, the difference between Simplified Chinese and Traditional Chinese is only in the simplification of fonts, and the expressions used in Traditional Chinese have been simplified in the same way.

I remember there were three people including me who submitted translations, the remaining two were machine translations, the difference is only that the quality may be better later on because Google Translate updated ...... (in Chinese it is easy to tell what translation tool was used for the translation) Although the machine translations are a bit raw, they are equivalent to the same after proofreading work Although machine translations are a bit stiff, they are equivalent to high quality translations after proofreading, so please do not delete them. The proofreaders may need to consult a dictionary for some unfamiliar terms to ensure that the translation is correct, and they are sometimes responsible for touching up the translation to ensure that it conforms to the local language and to ensure the quality of the translation.

is there one person's translations who you would recommend the most?

They are all machine translations, and the recommendation in the OP is the latest one. And doing the proofreading would take so much time that I'm only doing it in my free time, so I may follow up by talking to good translators on some platforms and getting them to submit their translations through Crowdin instead of reverse engineering them.


Actually my English is not good, just barely using translation tools and my broken English to communicate 😂 I'm afraid that DeepL is not accurate enough because sometimes it distorts the meaning of the expression... so I use more English expressions and order of expressions, as well as the original text at the end of this reply, so you can try to translate it by yourself.

I'm better at proofreading, I'm very confident in proofreading and touching up Chinese translations, but if I don't have the original translation I'll need a translation tool. Sometimes I also use ai to assist me in translation work, but I will follow my own ideas and understanding to proofreading and touch-up work, so you will see that my submitted translations may be very little, after all, people are not like machines, and there are many factors that lead to this situation...


Say a topic, the Chinese community is not too like 流汗黄豆: 😅, because it most of the time to express the meaning of the yin and yang, or dissatisfaction with something, which involves the history of the development of the Internet in Simplified Chinese... probably understand a little bit of the good, after all, different languages are not the same as the network environment, Simplified Chinese Internet some of the Wu Smoke and Miasma...

comment_zh-CN_original.md
comment_en_deepl.docx
comment_en_conversion.md

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants