> --------------------
> Subject:

> Hello Sven,

> Thank you for your email!
> My name is Akira Matsuda, one of the target of your indiscriminate  
> email attack against Japanese I18n watchers :)
> I was so much surprized(and felt so happy) with the fact that you  
> noticed me(though I had not yet published any information about Ruby/
> Rails in English).

> Firstly, let me introduce myself a little bit.
> I'm a freelance Rails programmer in Tokyo/Japan, I'm a regular  
> member of Rails Benkyo-kai@Tokyo(Rails study meeting held every  
> month in tokyo. I'm gonna talk about it later), I'm a founder of  
> Asakusa.rb(a Ruby hackers group including ko1 Sasada --the developer  
> of YARV-- launched last month with a lot of respect for Seattle.rb.  
> I'm sorry, there's only Japanese website now: http://asakusa.rubyist.net/)
> , one of my hobby is Rails edge watcing, and my another hobby is  
> Music. My favourite Musician is L. v. Beethoven(truly).

> Well, yes, as you expected, I'm watching your Rails I18n project  
> with a big big interest and anticipation. Because we also always  
> have been annoyed and frustrated by Japanizing every Rails  
> application we make.

> So, of course I knew that your I18n library has been merged into  
> Rails edge, I read the announcement on your blog post in almost real-
> time(twittered like this at that time http://twitter.com/a_matsuda/statuses/862447811)
> .
> After reading your announcement, I tried it by myself and found it  
> very useful. It worked pretty well on Japanese!

> Then I thought I have to let other Japanese Railers know about it,  
> so I held a learning session about the Rails I18n on the 31st Rails  
> Benkyo-kai@Tokyo held on yesterday(http://wiki.fdiary.net/rails/?RailsMeetingTokyo-0031
> ) with a very simple material like this(http://blog.dio.jp/2008/7/27/rails-2-2-i18n
> ).

> What I did through the session was, I described the history of the  
> project, introduced some materials on the web, talked about how and  
> where to follow the commits on Github, grabbed the edge Rails from  
> Github, read some points in the source code, did some basic  
> translation live demo on irb, and made a very simple Rails app with  
> a Japanese translated AR validation error message.
> I believe I succeeded to inform more than 50 people there about the  
> Rails I18n!
> And I feel it's so funny that the session was accidentally held on  
> just the same day as I received your email.

> So, I showed them your message there as well, and we discussed about  
> using YAML for that purpose. Everybody there, of course, agreed that  
> that format is preferable than the current hash one, but no one  
> could clearly explain the problem you were told by someone(though  
> there were certain number of famous Ruby/Rails library developers  
> and book writers there).

> One of my colleague at Asakusa.rb, Kakutani-san (I think you've sent  
> the same email to him, and he's gonna send the reply to you sooner  
> or later), has poited out that storing Japanese string into YAML  
> returns binary output like this,
> % ruby -ryaml -e 'puts "あ".to_yaml'
> --- !binary |44GC
> Thus we may face this problem when we trying to generate the ja-
> JP.yml by Ruby script. That's the only problem we counld notice at  
> this moment, and you know, it's qwite a slight problem.

> So, I personally think it's OK to use YAML in this case. I'm sure we  
> Japanese can also be happy with it, and It's not going to cause much  
> problem.
> In fact, I made a little adhoc test implementation with YAML ->  
> I18n::Backend::Simple => AR validation error message, and it worked  
> perfectly again.
> Here's my testing code<<EOF

> lib/locale.ja-JP.rb
> I18n.backend.store_translations :'ja-JP',  
> YAML.load_file("#{Rails.root}/lib/locale/ja-JP.yml")

> lib/locale/ja-JP.yml
> active_record:
>  error:
>    header_message: '{{object_name}}に{{count}}つのエラーが発
> 生しました。'
>    message: 次の項目を確認してください。
>  error_messages:
>    inclusion: は一覧にありません。
>    exclusion: は予約されています。
>    invalid: は不正な値です。
>    confirmation: は確認されませんでした。
>    accepted: を受諾してください。
>    empty: を入力してください。
>    blank: を入力してください。
>    too_long: is too long (maximum is {{count}} characters)
>    too_short: is too short (minimum is {{count}} characters)
>    wrong_length: is the wrong length (should be {{count}} characters)
>    taken: はすでに存在します。
>    not_a_number: は数値で入力してください。
>    greater_than: must be greater than {{count}}
>    greater_than_or_equal_to: must be greater than or equal to  
> {{count}}
>    equal_to: must be equal to {{count}}
>    less_than: must be less than {{count}}
>    less_than_or_equal_to: must be less than or equal to {{count}}
>    odd: must be odd
>    even: must be even
>  human_attribute_names:
>    user:
>      name: なまえ
> user: 'ユーザー'

> EOF

> Finally, I'm sorry that I joined the I18n Google groups the last  
> day, but not yet followed all of the discussions there.I'll try to  
> catch them up as qwickly as I can.
> And, of course I promise that I'm willing to do any help and  
> cooperation for the project if you need.
> Please ask me any question or problem or task around Japanese. I'll  
> do my best :)
> Thank you.

> Akira

> --------------------

> Reply on GitHub: http://github.com/inbox/13655#reply

>>> returns binary output like this,
>>> % ruby -ryaml -e 'puts "あ".to_yaml'
>>> --- !binary |44GC

> It doesn't return "binary" output, rather a base64-encoded byte form
> of the Unicode string. This is something that has been on the Syck
> todo list for years and years but _why never got to fixing it. So the
> problem is that as soon as you try to _save_ a Unicode string into
> YAML programmatically (as opposed to typing it) it becomes human-
> unreadable (it will load fine when unserializing YAML).

> So YAML is perfectly suitable for translations as long as you only and
> exclusively edit it by hand.

> On 30 jul 2008, at 20:15, Sven Fuchs wrote:

>>> returns binary output like this,
>>> % ruby -ryaml -e 'puts "あ".to_yaml'
>>> --- !binary |44GC

> It doesn't return "binary" output, rather a base64-encoded byte form
> of the Unicode string. This is something that has been on the Syck
> todo list for years and years but _why never got to fixing it. So the
> problem is that as soon as you try to _save_ a Unicode string into
> YAML programmatically (as opposed to typing it) it becomes human-
> unreadable (it will load fine when unserializing YAML).

> So YAML is perfectly suitable for translations as long as you only and
> exclusively edit it by hand.