Japanese (ja) translation group

English

This thread is meant for all those interested in translating Indico to Japanese.

In order to facilitate consultation by other translation teams, I suggest we use English whenever possible. Especially, the subject of every thread shall be written in English or, at least, Japanese with its English translation alongside.

However, messages deeply connected to grammar or the meaning of words can be written in Japanese. In those cases, please consider adding a short summary in English to help non-Japanese speakers to understand the context.

日本語

このスレッドはIndicoを日本語へ翻訳することに関心のある皆様に向けたものです。

他の翻訳チームとの協議を容易とするため、可能な限り英語を使ってください。特に、スレッドのタイトルは英語で書くか、少なくとも英訳を伴わなければなりません。

ただし、日本語の文法や語意と深く結びついた事項は日本語で書いてもかまいません。その場合、非日本語話者が文脈を理解するのを助けるため、英語による要約を添えることを検討してください。

1 Like

So, let me introduce myself. I am one of the administrators of https://indico.iist.nias.ac.jp/ Indico instance. I have participated in open source software projects for some years but I have only a few experience of software translation.

1 Like

There is a known issue related to Japanese:

Guidelines for Japanese translation group (draft)

General information

  • We use Transifex as the translation platform.
  • If you find any problems in translated strings, please add a comment on Transifex as an issue. If you find difficulties in translation due to the structure of source code, please fill an issue on GitHub.
  • If you find a term translated inconsistently, please let us know. After we agree with a canonical translation, a translator with Transifex account will add it to glossary on Transifex.
  • Do not use automatic translation service like Google Translation.
    EULA of these services are sometimes incompatible with open source licenses. If you are 100% sure that any specific service is open source compatible so we can use it for Indico translation, please let us know.
  • Currently, Japanese translation data are not released as part of the official Indico distribution. To test Japanese user interface on your Indico instance, please read Translation workflow? thread.

Translation guidelines

There are many Japanese translation guidelines. Until we find the optimal one for user interface, we use Japanese Standard Style Guide by Japan Translation Federation (version 2.3).

For the guideline of Katakana transcription, we use Loanwords (Katakana) Transcription Guideline by Japan Technical Communicators Association (third edition) as it is referred by JTF Japanese Standard Style Guide v2.3.

In addition to the style guide, there are some remarks.

  • 説明文では敬体を、ボタンやメニュー項目のラベルでは常体を使います。
    例:このイベントを管理する権限がありません。
    例:ファイルを開く(メニュー項目のラベル)
  • ボタンやメニュー項目のラベルが動詞で終わり、なおかつその動詞が「する」で終わるとき、その「する」を省略します。
    例:名前を付けて保存(メニュー項目のラベル)
  • ボタンやメニュー項目のラベルが分詞で終わるとき、「している」「される」に類する接尾辞を省略する場合があります。
    例:protected → 保護(メニュー項目のラベル)
1 Like

Isn’t the quality of such translations usually bad enough to make them a no-go? Even if they produce grammatically correct output, I could imagine that they lack the domain-specific knowledge to use the right terms.

2 Likes

@ThiefMaster Yes, I hope so, but let me share a case of Japanese OSS community.

In the nearly-end of 17.04 development cycle, Ubuntu Japanese Team found that some translations had been contaminated with machine translation. They backed out changes by whom used machine translation and audited the translation process (Japanese) albeit the tight schedule of 17.04 development cycle.

This incident had led to long discussions in Japanese OSS community and they agreed with the principle: no machine translation (unless you are 100% sure). However, not all Japanese developers know this incident, so I wrote it explicitly.

1 Like

Now, NiAS Indico instance started providing Japanese translation! :tada:
(There’re only a few entries translated for now, though.)

2 Likes

Now the NiAS instance synchronizes translation data with Transifex weekly.
I will post step-by-step how-to to Translation workflow? after the cleanup of scripts.

Can this be more often? It is nice to see result without waiting maximum 7 days.

Yes, of cource. It is matter of configuration of the server. However, if we request resources to Transifex so often, it is possible that our server will be banned.
According to the description of Transifex API rate limit, it will be safe to update hourly. Are you happy with this?

Thanks. I think it can be less often than every hour, like, once a day is enough.

OK, I updated the NiAS instance to update translations daily.

Actually, the number of resource files we need is less than 100, so even updating every minute would be accepted if the performance of server and network permits. :slight_smile:

2 Likes

Hello Masanori,
I visited your website, and it looks very nice. Apparently you spotted the most exposed text elements and translated them to Japanese. (I cannot judge the quality by myself, but I am sure you did your best.) Is your translation up to date with transifex? Statistically there are only 11.32% of all strings translated. Do you plan to continue the work?

I have made a change to the Latex template that will enable use of CJK fonts in PDF exports.

I’m hoping to make a pull request soon. Any comments before I do?

*Note: noto fonts also need to be installed to be available.

1 Like

We are nearing the launch of our first Indico event, the 2024 MoodleMoot Japan. We’ve got the abstract book showing properly with Japanese characters, but now I’m noticing that the PDF view of the schedule is not showing the Japanese characters properly. Is there any way we can move to incorporate Adam’s changes to the PDF schedule?

Our event FYI
https://indico.cern.ch/event/1257438/timetable/#20240217

We don’t use Latex to generate the PDF timetable so we will need a different approach here. I think adding a font fallback as described here: python - ReportLab: working with Chinese/Unicode characters - Stack Overflow could work.

I wonder if we should at some point just use weasyprint to generate these PDFs. Would probably look much better and hopefully avoid this kind of bugs as well (assuming it handles languages with a “nonstandard” alphabet nicely)