Thu 16 Jun 2011 08:00:01 AM UTC, comment #25:
Closing the bug: there is no /chinese in www CVS,
and the file names from /software/chinese follow the convention.
|
Sat 31 Mar 2007 05:47:21 PM UTC, comment #24:
William, I think we're almost done. Pheew...
All files are renamed and symlinks are in place (will take a few hours to update, though). All links are changed accordingly, except in /chinese and /software/chinese -- I left those for you.
It is quite possible that I've broken something, please check carefully. There are a few cosmetic changes to be made (changing the [cn] and [zh] in the templated pages), I'll do that in the next few days.
|
Sat 31 Mar 2007 11:13:30 AM UTC, comment #23:
No, CVS history will be lost for all files (zh, zhs and cn). In other words, if you now have foo.cn.html rev 1.5, you won't be able to see the changes from 1.4 or restore it to an older revision. What we can save are the CVS commit logs (attached).
(file #12353, file #12354, file #12355)
|
Sat 31 Mar 2007 10:11:24 AM UTC, comment #22:
Nice idea! We can move the files to graveyard. (I was thinking that CVS may maintain the history.)
Thanks,
William
|
Sat 31 Mar 2007 09:52:15 AM UTC, comment #21:
That's entirely up to you to decide. As I don't read/speak Chinese, I can't judge whether the files are identical or not. I'm not comfortable in deleting them, but we can move them from "www" to "trans-coord", say, under /graveyard/zh-tw.
|
Sat 31 Mar 2007 09:45:59 AM UTC, comment #20:
I'm not clear about the history that why we named BIG5 files to *.zhs.html before. But if the zhs.html file is a duplicate of an existing zh.html file, I think we can remove it. Your opinion?
|
Sat 31 Mar 2007 09:22:37 AM UTC, comment #19:
Sorry, I see that you probably mean the opposite -- we have to delete the files listed in comment #18. Please confirm.
|
Sat 31 Mar 2007 09:16:19 AM UTC, comment #18:
Here's a list of the duplicated zh and zhs files:
/doc/doc.zhs.html
/home.zhs.html
/chinese/issue-template.zhs.html
/chinese/the-template-page.zhs.html
/licenses/translations.zhs.html
/music/free-software-song.zhs.html
/philosophy/free-software-for-freedom.zhs.html
/philosophy/free-sw.zhs.html
/philosophy/luispo-rms-interview.zhs.html
/philosophy/selling.zhs.html
/philosophy/wsis-2003.zhs.html
/software/chinese/bgw/issue-48.zhs.html
/software/chinese/otcl/otcl.zhs.html
/software/chinese/otcl/topic-1.zhs.html
/software/chinese/gnuctt.zhs.html
/software/chinese/index.zhs.html
/software/chinese/mailman.zhs.html
/software/chinese/otcl.zhs.html
/software/chinese/philosophy.zhs.html
/software/chinese/efforts/hurd-and-linux.zhs.html
/software/chinese/efforts/hurd.zhs.html
/software/chinese/guide/emacs-chinese.zhs.html
/software/chinese/manual/manual.zhs.html
So you want me to rename these files to zh-tw and delete the corresponding zh files, right?
|
Sat 31 Mar 2007 09:01:03 AM UTC, comment #17:
Yavor,
I checked those *.zhs.html files, they are BIG5, not GB2312. Since free-sw.zhs.html is the duplicate of free-sw.zh.html, we may delete free-sw.zhs.html. What do you think?
Thanks,
William
|
Sat 31 Mar 2007 08:12:11 AM UTC, comment #16:
AFAICS zh files are in Big5 and zhs are GB1232. The problem is that there duplicates -- i.e. free-sw.zh.html and free-sw.zhs.html, and I don't know how to deal in these cases. Both cannot be renamed to zh-tw...
|
Sat 31 Mar 2007 02:18:15 AM UTC, comment #15:
Yavor,
I checked those .zhs.html files. They are in BIG5 encoding. So we may rename them to .zh-tw.html.
Thanks,
William
|
Thu 29 Mar 2007 05:17:44 AM UTC, comment #14:
Thanks for the explanation, Alex. I don't see an obstacle adding these symlinks, so I'll do it as part of the migration.
|
Wed 28 Mar 2007 10:44:14 PM UTC, comment #13:
Yes, I was suggesting to go ahead with renanimg but adding symlinks for backwards compatibility. However, I'm afraid that wildcards won't work in .symlinks but you can ask sysadmin to get a final response, since I could be wrong.
About disappearence of current filenames from indexers, it should happen some time after the whole http://www.gnu.org has been updated with newer file names. The .symlinks aren't the best way to achieve that (redirections are much better) but we lack the sysadmin resources needed to get things done the right way :(
Thus, we'll get references to current filenames out of (e.g.) Google as soon as we update our own site content; but stopping people from reference explicitly to those filenames will need more effort, since we'll have to request a manual update (let's hope that there won't be much of these cases).
HTH
|
Wed 28 Mar 2007 08:22:21 PM UTC, comment #12:
If I understand you correctly, you suggest to do the rename but adjust .symlinks for compatibility. Do asterisks work or do we have to list explicitly every file, i.e. would this suffice in /philosophy/.symlinks:
.zh-cn.html .cn.html
.zh-tw.html .zh.html
Second question (might be obvious): Will the .symlinks solution provide natural disappearance of all references? In other words, if /gnu/gnu.cn.html does not exist (but is provided via .symlinks), would that (1) stop indexers indexing the cn page and (2) stop people refering to that URL?
|
Wed 28 Mar 2007 08:10:57 PM UTC, comment #11:
Since I guess that many of those pages are indexed and referenced in many pages, I'd like to suggest that we keep those wrong page names with symlinks. Once we're sure that any of those pages are referenced anywhere, we should get rid of the symlinks. What do you think?
BTW, preserving the CVS logs seems a very good idea to me :)
|
Wed 28 Mar 2007 08:02:09 PM UTC, comment #10:
> Give me sometime to investigate. I'll answer
> your questions by this weekend.
No problem, thanks. I intend to do this migration at once, as I fear that it won't work partially. I mean, if I rename some files (say, under /gnu), then other links to them will be broken.
I'll save the CVS logs prior to the rename and attach them to this bug -- you can then review them when you have time.
I'll also inform the webmasters and web-translators for the forthcoming changes.
|
Tue 27 Mar 2007 03:05:58 PM UTC, comment #9:
Actually, there are 56 *.zhs.html files, most of them under /software. Please tell me what should be TRT.
|
Tue 27 Mar 2007 02:59:02 PM UTC, comment #8:
OK, thanks, but you will lose CVS history. Agreed?
(What about free-sw.zhs.html?)
|
Tue 27 Mar 2007 02:57:29 PM UTC, comment #7:
Yavor,
We can start renaming .cn.html to .zh-cn.html, and .zh.html to .zh-tw.html.
Thanks,
William
|
Tue 27 Mar 2007 02:46:35 PM UTC, comment #6:
Cai,
There is also http://gnu.org/philosophy/free-sw.zhs.html. Any idea what is it?
I thought for a minute that I could start renaming the files (with your permission, of course), but I can't figure out whether to rename foo.zh.html to foo.zh-hk.html or foo.zh-tw.html...
|
Tue 22 Aug 2006 05:49:47 PM UTC, comment #5:
> I sent an email to Debian.org Chinese maintainer a week ago,
> but still don't get response.
Maybe he's on vacation.
> Could we go ahead to use zh-cn/zh-hk/zh-tw distinguishing
> different Chinese encoding?
We don't need to wait for Debian people, so go ahead with
the changes... :)
|
Sun 20 Aug 2006 10:56:02 AM UTC, comment #4:
I sent an email to Debian.org Chinese maintainer a week ago, but still don't get response. Could we go ahead to use zh-cn/zh-hk/zh-tw distinguishing different Chinese encoding?
-William
|
Thu 10 Aug 2006 06:39:35 AM UTC, comment #3:
Very good! :)
|
Thu 10 Aug 2006 01:41:25 AM UTC, comment #2:
Well, I think I get your point now. 'zh-cn' stands for China mainland where Simplified Chinese is used. 'zh-hk' and 'zh-tw' stand for HongKong SAR and Taiwan where Traditional Chinese is used. In a word, 'language code-area code'is enought to distinguish Simplified and traditional Chinese.
I'll get contact with Debian.org Chinese folks and see what happens there.
|
Wed 09 Aug 2006 10:23:48 PM UTC, comment #1:
> I captured the the following information with firefox.
> ************************** HTTP Header *********************************
> Host: [2]www.loc.gov
> User-Agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; zh-CN; rv:1.7.12) Gecko/20050919 Firefox/1.0.7
> Accept: text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,/;q=0.5
> Accept-Language: zh-cn,zh;q=0.5
> Accept-Encoding: gzip,deflate
> Accept-Charset: gb2312,utf-8;q=0.7,*;q=0.7
> Keep-Alive: 300
> Connection: keep-alive
> Cookie: ForeseeLoyalty_MID_MsF0ld0M1Z=2; s_sess=%20s_cc%3Dtrue%3B%20s_sq%3D%3B
> ************************** END *********************************
> It seems the request sent by the browser (at least firefox)
> contains enough information. What we need is a mapping table
> between language and page, say zh-cn <--> index.zh-cn.html,
> etc.
Right, but I'm wondering if it's useful enough to have the
language and country code to differentiate simplified from
tradicional Chinese.
I mean, "zh-cn" stands for Chinese from China and it does
not imply simplified nor traditional symbols. See the point?
> Technically, the page name can be everything. But to be
> consistent with what is defined in browsers, I agree to
> using the format of "language name-area name".
> What do you think?
If the langcode-countrycode is enough to solve the issue,
I'll consider it the right thing. Did you try debian.org
and see if that same solution works for them? Maybe you
could contact their Chinese translation coordinators and
ask them what are they doing.
One last thing, please use the savannah interface to update
this "bug" from now on. If I'm to transcript our conversations
my mail into savannah, it will take a lot of time.
Thanks!
|
Tue 08 Aug 2006 09:58:07 PM UTC, original submission:
There's no such language code on the standard ISO-639-2.
We need to find a way to diff the pages with simplied Chinese from those in the traditional form, but the standard does not make such a difference.
Note that "cn" is the country code for China and "zh" is the
2-letter language code for Chinese.
|