Home > Plans > Languages for zkanji, maybe with your help

Languages for zkanji, maybe with your help

November 19, 2011

I have received requests several times before, to allow translation of the zkanji user interface to other languages. First of all let me make it clear, this is not an “official” announcement of a new feature, but an explanation of the work needed for the program to be translated into other languages. There might be such an announcement later, but I’d like to hear your opinion and ideas about it now. (I could use a “disclaimer” here, but people usually skip those.)

Nowadays almost every software out there gets translated. It’s not just a dream anymore that could never come true, unless you are a huge company with lots of money for internationalization. Thanks to the internet, there are whole communities helping to translate things into their favorite languages (which is usually their own). But it is also a huge task to attract and keep together such communities, especially if they don’t speak your language.

In a program like zkanji, there are hundreds (sometimes thousands) of phrases to be translated. I can imagine translating them takes months, and it’s not just that, but you would also have to check whether the translations fit in the controls where they will appear. (As you see I was never good at marketing, I can only say the truth…) Of course there will be nothing to translate if I don’t start collecting the lines that should be translated, but I wrote the originals so I deserve it.

The third problem is finding the means to share the translations between me and the translators, and between the translators themselves, who might be able to work together. I only know of the site https://www.transifex.net/ which was made for community translations, and is absolutely free for open-source projects, but I don’t know exactly how it works and how difficult it is to use its output. I admit I have a little distrust with things that were not made by me personally.

This all is probably just a little part of the difficulties anyone has to face who would like to translate a program with even the least complexity. In the case of zkanji, the problems won’t stop here. While I acknowledge the need for the users to see an interface in their own language, the main dictionary that comes with it would and will always be in English. I’m saying main dictionary, because it is possible to create additional ones. (Unfortunately English is needed as some functions (for example adding furigana in vocabulary printing) require a large enough database, or they won’t work.) I think though, that even if there is a possibility to create your own dictionary, this won’t be such a strong pull for most users.

JMdict (the database used in zkanji) has French and German entries, and the documentation says that it has a little Russian and Dutch as well. I have the means to convert those to zkanji format, but it’s already a huge work releasing the program every time (just that usually takes 1 – 1.5 hours), that I would rather not do anything else on top of that.

Though that’s not the only source that could be used in dictionary creation. Funbit, who often posts comments in this blog created a Russian dictionary for the program available from here, with a tutorial on how to use zkanji.

So to summarize the difficulties:

  1. collecting what to translate (work for me only)
  2. finding a community who helps with the translation
  3. finding the means
  4. translation (obviously)
  5. checking the translations (do they fit, are they correct etc.)
  6. sharing the translations
  7. including the translated data so zkanji can work with it (it is only my problem, sorry for bringing it up)
  8. finding some way to include other available dictionaries (I’m hoping I can convince users to update (and maybe host) them)

Please share your thoughts in the comments!

  1. November 19, 2011 at 4:39 pm

    I there was a list of string constants used in zkanji – I think we could translate it into Russian pretty easily within 1-2 months. I hope, you’ll finally manage to make it 🙂

  2. sas
    November 19, 2011 at 8:16 pm

    Speaking of translating the gui… just saw how it was done in good old Wakan. The control caption/hint fields and menu entries have a default English string prepended with a four-digit number. Matching string lists are stored in plaintext files, e.g.en.lng, ge.lng, fr.lng. Entries look like e,g, 0001 Enter in en.lng, 0002 Eingabe in ge.lng, After form load a translate function walks thru the control list and menu and replaces the number+default by the matching entry from x.lng or just default if no match was found. Assuming there are about a thousand mostly very short strings, the translation to a new language can be done in an evening. Text labels and button captions are a bit iffy if the translated string is longer than the English and does not fit the control width. E.g. German words are generally longer than English equivalents, but one could use abbreviations, Eing instead of Eingabe.

    If Ejmar Eistfodd wants to do a Norwegian gui, he just copies en.lng to a new nor.lng and walks thru the string list with notepad. it’s okay to translate only the most important parts for starters, i.e. menus and control captions. Hints and messages can be done later. upload to zkanji site and voila, Norwegian zkanji.

    Translating the help file is obviously more problematic, since it at least requires a chm help builder program and not just notepad.

    Since Japanese learning success more or less presupposes good command of English (because of gairaigo and the infinitely better selection of textbooks) and Japanese learners have good linguistic skills anyway, a translated gui is a lot less important than in a cookbook or bookkeeping app

    Imho the only alternative gui language that would really be cool is JAPANESE. for advanced learners and native language speakers this would make zkanji a lot more enjoyable. as a bonus, the kanji strings are very short, e.g. 選択 versus “Select”.

    • November 19, 2011 at 10:21 pm

      I mentioned transifex.net because it allows more people to work on a single language. If someone doesn’t agree with a translation, they can vote it down and suggest a better one, which in turn can be discussed by others and voted in or out. That site also seems to notify translators if new lines were added, and I wouldn’t have to send emails about it to anyone. The old-fashioned way of translating from a text file might look super easy for translators, but it would be difficult to automate some tasks that must be done, and there would be no way for me to decide whose translation to accept when more people wanted to help. (Or am I dreaming? :))

      My concern about that site is merely technical, but I will test it out. I would like to avoid using a simple text file if possible. I would rather put together my own online translation system if I must.

      I’m not sure about that you need a good command of English (or any at all) for Japanese. It’s true that there are many loan words, but anyone should be able to learn them in my opinion. But that aside, there were requests before for other languages, and if it can make more users see why zkanji is so good (dreaming again?), then hopefully everyone will win. Including Japanese will be possible as well of course.

  3. me
    November 21, 2011 at 11:12 pm

    My idea has nothing to do with the user interface, but it is also about adding a new language.
    Here is my idea: You write it is possible to import dictionaries from Wakan. Wakan can be used for learning Japanese AND Chinese characters. Since it seems that the development of Wakan has been stopped, why not expand zkanji in a way that you can learn and compare characters of the TWO languages? It would be interesting to compare the characters of both languages. (Some characters or character combinations, the “kokkun,” even have different meanings in Japanese and Chinese, for example, 手紙 is “letter” in Japanese, but means “toilet paper” in Chinese. As a beginner in the field of learning kanji / hànzì, I am interested in learning what different meanings the jôyô kanji can have in Chinese. I know, for example, that 娘 means “daughter / girl” in Japanese, but “mother” in Chinese, and I heard of a jôyô kanji meaning “princess” in Japanese, but “concubine” in Chinese.) What do you think?

    • November 22, 2011 at 12:49 am

      I have long resisted the request of someone to add features for Chinese, because I have no interest in that language. I don’t see any practical advantage of including Chinese meanings apart from that it’d be interesting.

  4. November 22, 2011 at 1:13 am

    By the way, after zkanji had started to support Unicode it became possible to create a simple JapaneseChinese dictionary, so “me”‘s idea could be implemented in much easier way 🙂

  5. me
    November 22, 2011 at 3:06 pm

    As mentioned above, one could use the Chinese dictionary of Wakan (or, if there is a more up-to-date one, another one zkanji can import), so it is not necessary to create a completely new dictionary. The only thing is that Wakan can only be set either to the “Japanese mode” or the “Chinese mode,” so it cannot display both languages at the same time. zkanji would need such a feature for the comparison of the characters. And one must note that there are TWO sets of Chinese characters: the traditional ones and the simplified ones. 🙂

  6. November 22, 2011 at 5:10 pm

    In theory it is possible to create a Chinese-Japanese dictionary (if one is available for free online), but a Chinese-English one that Wakan uses cannot be included with zkanji, because it can only use Japanese dictionaries. The data structure and the interface was made with Japanese in mind, but the other language can be anything you want.

    As I wrote above in the post, if you want to, you can create extra dictionaries, and I’ll be happy to help. (But you’ll need a little scripting knowledge.)

  1. No trackbacks yet.
Comments are closed.
%d bloggers like this: