Closed Bug 1426808 Opened 7 years ago Closed 7 years ago

[vi] Change rule for Vietnamese from #1 (2 forms) to #0 (1 form)

Categories

(Mozilla Localizations :: vi / Vietnamese, enhancement)

Product:
Mozilla Localizations
Component:
vi / Vietnamese
Type:
enhancement
Priority:
Not set
Severity:
normal

Tracking

(Not tracked)

Status:
RESOLVED FIXED

People

(Reporter: flod, Assigned: pierreneter)

Details

Vietnamese has 1 plural form according to CLDR http://www.unicode.org/cldr/charts/dev/supplemental/language_plural_rules.html#vi But Firefox and Firefox for Android are set to use the plural form 1, which means 2 plural forms (like English). Translate House has 1 form like CLDR http://docs.translatehouse.org/projects/localization-guide/en/latest/l10n/pluralforms.html Which form is correct? Does Vietnamese have plurals (book, 2 books, books)?
(In reply to Francesco Lodolo [:flod] from comment #0) > But Firefox and Firefox for Android are set to use the plural form 1, which > means 2 plural forms (like English). Clarifying, since this might be confusing: is set to use "rule #1", which has 2 plural forms. #1 cookie;#1 cookies 0, >1: #1 cookies 1: #1 cookie
In Vietnamese, no special form of plural. Just like singular. Example: 0: 0 cookie 1: 1 cookie >1: 2 cookie, 100 cookie. For > 1, can't count: English: books Vietnamese: 'các quyển sách', 'những quyển sách', 'mấy quyển sách', etc... many words refer to plural in the prefix of this word in the singular. Depending on the noun.
(In reply to Pierre Neter from comment #2) > For > 1, can't count: What do you mean with "can't count"? > English: books > Vietnamese: 'các quyển sách', 'những quyển sách', 'mấy quyển sách', etc... > many words refer to plural in the prefix of this word in the singular. > Depending on the noun. What's the difference, for example, between 'các quyển sách' and 'những quyển sách'?
To hopefully clarify it, can you translate these? word 0 words 1 word 2 words 10 words words
(In reply to Francesco Lodolo [:flod] from comment #4) > To hopefully clarify it, can you translate these? > > word > 0 words > 1 word > 2 words > 10 words > words từ 0 từ 1 từ 2 từ 10 từ các từ, những từ 'các' and 'những' are the same, in each case, it will be chosen as the more appropriate word. But, the same.
OK, thanks for confirming. This matches CLDR and the setting in Pontoon (one form). I've updated the rule in Firefox https://hg.mozilla.org/l10n-central/vi/rev/99d3b2257060f90580f2049f13c4e0b019b816b1 At this point, several strings needs to be fixed in Pontoon https://pontoon.mozilla.org/vi/firefox/all-resources/?status=translated&search=Localization_and_Plurals https://pontoon.mozilla.org/vi/firefox-for-android/all-resources/?status=translated&search=Localization_and_Plurals For example, the first one #1 tab has arrived from #2;#1 tabs have arrived from #2 In English has 2 forms, each separated from the other by ";". In Vietnamese it should only have 1 form #1 thẻ đã tới từ #2 Currently it has 2 #1 thẻ đã tới từ #2;#1 thẻ đã tới từ #2 Sometimes the string starts with ;, meaning that the form for singular is empty in English ;List All #1 Tabs Translation should not have the starting semicolon, since Vietnamese only needs one form Hiển thị tất cả #1 thẻ Let me know when you have managed to look into the strings, I can run a script to see if there are more to fix.
Summary: [vi] Verify plural rule for Vietnamese → [vi] Change rule for Vietnamese from #1 (2 forms) to #0 (1 form)
Ok, I'm doing with strings.
Hey, what about this: >and one more;and #1 more >và một cái nữa;và #1 cái nữa Should I keep 2 form for more correctly?
(In reply to Pierre Neter from comment #8) > Hey, what about this: > > >and one more;and #1 more > >và một cái nữa;và #1 cái nữa > > Should I keep 2 form for more correctly? You can't have 2 forms at this point, simply translate "and #1 more"
Ok! I got it.
And: https://pontoon.mozilla.org/vi/firefox/all-resources/?status=translated&search=Localization_and_Plurals&string=74462 >#1 has been installed successfully.;#2 add-ons have been installed successfully. The var of the singular is diff with the var of plural? What should I keep? #1 or #2?
Poor old string that has been fixed recently (there's another want very similar), only translate the second part. #2 add-ons have been installed successfully. It's not a plural, it's a string that assumes the first form is singular, which is not true for many languages.
(In reply to Francesco Lodolo [:flod] from comment #12) > Poor old string that has been fixed recently (there's another want very > similar), only translate the second part. * another one very similar https://pontoon.mozilla.org/vi/firefox/all-resources/?status=translated&search=addonsInstalled&string=74463 Just translate: #2 add-ons will be installed after you restart #3.
Assignee: nobody → pierreneter
Thank you, I'm putting that in the Style Guide [1] https://wiki.mozilla.org/L10n:Teams:vi/Style_Guide
We moved style guides here last summer. Can you make sure the content is in sync? https://github.com/mozilla-l10n/styleguides/blob/master/vi/README.md
(In reply to Francesco Lodolo [:flod] from comment #15) > We moved style guides here last summer. Can you make sure the content is in > sync? > https://github.com/mozilla-l10n/styleguides/blob/master/vi/README.md To update it you can follow this https://help.github.com/articles/editing-files-in-another-user-s-repository/ Looks like there are 3 more strings to fix: - pendingCrashReports2.label (browser/chrome/browser/browser.properties) has 2 plural forms (requested: 1) - downloadMessage.deleteAll (mobile/android/chrome/aboutDownloads.properties) has 2 plural forms (requested: 1) - newtabpopup.opened (mobile/android/chrome/browser.properties) has 2 plural forms (requested: 1)
> Looks like there are 3 more strings to fix: > - pendingCrashReports2.label (browser/chrome/browser/browser.properties) has > 2 plural forms (requested: 1) > - downloadMessage.deleteAll > (mobile/android/chrome/aboutDownloads.properties) has 2 plural forms > (requested: 1) > - newtabpopup.opened (mobile/android/chrome/browser.properties) has 2 plural > forms (requested: 1) All done.
I look like page: https://wiki.mozilla.org/L10n:Teams:vi/Style_Guide is out of date, the regulations in this are no longer compatible with modern Vietnamese, need to be updated and need refactor of the content of this guile. Example: > sidebar: thanh lề (no-one understands what it is mean, `thanh bên` is making sense.) > blog: nhật ký (now just blog) > clipboard: bảng tạm ((?) I don't know what's 'bảng tạm' mean) > SMS (sms): Nhắn tin (I think `Tin nhắn` is more make sense) ...
Flagging myself for compare-locales.
Flags: needinfo?(l10n)
(In reply to Pierre Neter from comment #18) > I look like page: https://wiki.mozilla.org/L10n:Teams:vi/Style_Guide is out > of date, the regulations in this are no longer compatible with modern > Vietnamese, need to be updated and need refactor of the content of this > guile. I suggest to work directly on the GitHub one, we want to get rid of the wiki pages for this purpose. No more errors, thanks for fixing them so quickly. Make sure to use Nightly, and double check that everything looks good in tomorrow's build https://mozilla-l10n.github.io/localizer-documentation/products/firefox_desktop/testing.html
Status: NEW → RESOLVED
Closed: 7 years ago
Resolution: --- → FIXED
https://hg.mozilla.org/l10n/compare-locales/rev/8c04181b6733, updated compare-locales.
Flags: needinfo?(l10n)

Vietnamese has 1 plural form according to CLDR
http://www.unicode.org/cldr/charts/dev/supplemental/language_plural_rules.html#vi

That was an error, as described in https://unicode-org.atlassian.net/browse/CLDR-14273 . It sounds like an upcoming version of CLDR will change the Vietnamese rule from #0 (1 form) to #1 (2 forms).

Firefox has gotten away with #0 for the past few years because there are very few pluralizable strings in Firefox that omit the number outright, as in “Delete this item”/“Delete these items”. If there were, we would have to include glosses like “(các)” more often in translations.

Back in 2009, Hùng (CC’d) and I also determined that some messages that include the number could be problematic as well. For example, “Every year”/“Every 2 years” would be translated “Mỗi #1 năm”. That could result in a string like “Mỗi 1 năm”, which means “Each and every year”, even though the original string wasn’t designed to add that extra emphasis.

(In reply to Minh Nguyễn from comment #22)

Vietnamese has 1 plural form according to CLDR
http://www.unicode.org/cldr/charts/dev/supplemental/language_plural_rules.html#vi

That was an error, as described in https://unicode-org.atlassian.net/browse/CLDR-14273 . It sounds like an upcoming version of CLDR will change the Vietnamese rule from #0 (1 form) to #1 (2 forms).

Once the change in CLDR is official, please file a new bug to get the rules changed back to 2 forms.

I'm also sure they'll appreciate all the help in fixing the errors and warnings that will result from this change.

I just recently noticed this issue and was intending to file a bug. Unfortunately, what I and Minh used to think and did thoroughly was easily changed without some consideration or contacting us. Even if there would not be a change in CLDR, this should still be fixed. Following rules strictly is not always a good option, like this situation. As the original strings are in English which has 2 plural forms, strictly force Vietnamese translation to follow the rule of 1 form actually limits the translation options. And actually, the way plural forms used in Vietnamese language allows it to be set to ANY other plural rule. Hence, in my opinion, the translation guideline should even says something like this "For languages that have one plural form (Vietnamese, Chinese, Japanese, Korean etc.), translators may want to set the plural form rule to other rule to enable more translation options."

Thanks to Minh for bringing this up. I abandoned my Yahoo email and rarely check it, must find some way to change my email on all sites of Mozilla ecosystem to my current gmail.

(In reply to Hùng from comment #24)

I just recently noticed this issue and was intending to file a bug. Unfortunately, what I and Minh used to think and did thoroughly was easily changed without some consideration or contacting us.

The current localization team was involved, and there's no way to know why a decision was made years ago, or why, unless it's documented. Was that the case and it was missed?

Having said that, please be respectful of other people's work or commitment, this wasn't done "easily".

Even if there would not be a change in CLDR, this should still be fixed.

I don't intend to reintroduce differences between CLDR and Mozilla:

  • This rule is only used for legacy content (.properties) in Firefox. With Fluent strings, Vietnamese can already add new forms as they see fit.
  • Fluent uses CLDR data to select the correct plural form. That's the main reason for pushing 3 years ago to get consistency, or report back issues upstream to CLDR.

(In reply to Francesco Lodolo [:flod] from comment #25)

(In reply to Hùng from comment #24)

I just recently noticed this issue and was intending to file a bug. Unfortunately, what I and Minh used to think and did thoroughly was easily changed without some consideration or contacting us.

The current localization team was involved, and there's no way to know why a decision was made years ago, or why, unless it's documented. Was that the case and it was missed?

Having said that, please be respectful of other people's work or commitment, this wasn't done "easily".

Hi Francesco, sorry if my last comment bothered you. I didn't mean to mention you or any other people/devs who don't speak Vietnamese here - you were just doing what you thought were right (i.e. applying a standard), but a Vietnamese translator involved here and didn't see "the issue". That's explainable, anyway. thanhdd.it and Vietnamese devs involved in this project may have knowledge in the technical aspect, but may miss things in the language aspect. I mentioned this in our community group (not on Mozilla), but people might have missed it.

This is a special situation, though. After all, I just wanted to point out something: applying a standard may not be right/perfect in some situation.

(In reply to Hùng from comment #26)

Hi Hùng

This is a special situation, though. After all, I just wanted to point out something: applying a standard may not be right/perfect in some situation.

I completely agree with you, in the style guide [1] we even wrote "thận trọng khi sử dụng số nhiều" (be careful when using the plural). We never said that one special case would be right/perfect in all cases.

but a Vietnamese translator involved here and didn't see "the issue".

We are sorry. Maybe we don't understand clearly.

Now here's the question: Does Vietnamese have plurals (book, 2 books, books)? The answer is no.
We all know that the best translation involves complete sentence. In some cases, if we want to split the sentence we must ensure that it makes sense to combine parts of a sentence.
We are understanding that the form is used in counting cases. In the example of Minh Nguyễn "Every year"/"Every 2 years", it's not just a number, it also has an emphasis in a particular context.
I have a question: that form is used in all cases of numbers, not only counting cases?

[1] https://wiki.mozilla.org/L10n:Teams:vi/Style_Guide

Hi Thanh, hope we had more chances to discuss things.

Alright, we can take this example:
English: "One download is in progress, cancel it?" vs "Two downloads are in progress, cancel them?"
Vietnamese: "Một tập tin vẫn đang tải, dừng nó?" vs "Hai tập tin vẫn đang tải, dừng CHÚNG?"

Yes, at first, it looks like there is no "plural" form for counted object(s) as English have two ("download is"/"downloads are") while Vietnamese has only one ("tập tin vẫn đang"), but the true result is shown in the last part: For "it"/"them" in English, we have to use "nó"/"chúng" in Vietnamese - there is NO one word/concept in Vietnamese to mention both "it/them" and Vietnamese has clear distinction between "one thing" and "more than one thing" like English.

So, yes, this should have been noticed when they set up "the standard" in CLDR. Obviously there are many reasons to explain why the rule could stay for years, but a rule can always be improved when we find out the issue, right? And that's why what Minh is doing is on the point - report the issue directly to CLDR.

You need to log in before you can comment on or make changes to this bug.