The current state of ArchWiki internationalization

pointone · 2009-12-21 04:26:03

Rather than duplicate effort, I am posting to draw attention to this discussion I began on my user page. Please comment. Any ideas or opinions on the matter are welcome!

Gen2ly · 2009-12-21 12:31:54

Categorization is definitely possible. The i18n could be standardized with the most popular languanges.

Pierre · 2009-12-21 13:30:53

I didn't have time to look into this, but we should use interwiki links to connect the international wikis. E.g. there are some orphaned german articels in this wiki which should be moved or linke to archlinux.de.

thayer · 2009-12-21 15:59:58

FWIW the naming scheme, "Title in English (Language)" is the standardized form. It was standardized some time ago (circa 2007/08). I believe this conclusion was reached because of maintenance and administrative issues. English-only admins wouldn't know one article from the next if they were titled in foreign tongue.

pointone · 2009-12-21 17:21:06

Thanks for the quick responses thus far.

thayer wrote:

FWIW the naming scheme, "Title in English (Language)" is the standardized form. It was standardized some time ago (circa 2007/08). I believe this conclusion was reached because of maintenance and administrative issues. English-only admins wouldn't know one article from the next if they were titled in foreign tongue.

This was my thinking exactly. Unfortunately, I was unable to find any clear statement of the standard on the wiki; is this still the general consensus? pwd raised an important point on the discussion page about alienating foreign-language users: just as non-English titles are confusing to English-only admins and other readers, English-only titles would be equally confusing to non-English readers. However, as Arch is primarily an English distro, perhaps this is a moot point?

Perhaps "Title in English (Language)" redirect pages could be established to facilitate maintenance and inter-language links, whilst keeping non-English users happy.

Pierre wrote:

I didn't have time to look into this, but we should use interwiki links to connect the international wikis. E.g. there are some orphaned german articels in this wiki which should be moved or linke to archlinux.de.

As I understand it, Dusty and other developers wanted to avoid multiple sub-domains or otherwise segregated communities in past discussions. Nonetheless, enabling such interwiki links (e.g. [[de:Main Page]]) requires back-end changes in the database, as far as I know. Whether or not we move toward or away from separate international wikis, having the option of interwiki linking would certainly be appreciated.

Pierre · 2009-12-21 18:05:35

That's new to me. We have been through this several times and have tried different solutions. The best is to have independent wikis for each language. This might be different for others but at least this was decided by the german community. (I could put a few arguments here but I hope you just trust me here for now ;-))

For the interwiki stuff: if someone would look into this and open a bug report about how to implement this I could go for it (I am the wiki maintainer).

karol · 2009-12-21 19:38:58

> The best is to have independent wikis for each language.
+1 This is how it's done in Wikipedia.

Which one do you prefer:
Title in English (German)
Title in English (Deutsch)
Title in English (de)
Title in English (deu)
Title in English (ger)

or some other?
http://en.wikipedia.org/wiki/ISO_639

pointone · 2009-12-22 19:06:31

I opened http://bugs.archlinux.org/task/17580 on the bugtracker. I am not completely sure how to implement; I simply linked to Manual:Interwiki for instructions (someone with more MediaWiki experience should comment/clarify). There seems to be some important distinction between interlanguage links and generic interwiki links... But perhaps it can be ignored.

Gen2ly · 2009-12-26 17:52:27

thayer wrote:

FWIW the naming scheme, "Title in English (Language)" is the standardized form. It was standardized some time ago (circa 2007/08). I believe this conclusion was reached because of maintenance and administrative issues. English-only admins wouldn't know one article from the next if they were titled in foreign tongue.

I like this idea too. I'd like to begin implementing this. Any thoughts, considerations?

faelar · 2009-12-26 23:46:40

I missed this thread before, but you should be interested by our solution on ArchServer : http://bbs.archserver.org/viewtopic.php?pid=60#p60

Svenstaro · 2009-12-27 04:58:38

Good thing this discussion has been kicked off. I was meaning discuss this for some time now. The current Arch Linux language handling is a mess. I agree that there should be a consistent way for handling non-English languages. We could keep our wiki strictly English-only and refer to the language-specific Arch communities. This would probably lead to a split-off community but those international communities exist anyway. Should we encourage them to contribute mainly to the English wiki? It is a rather hard situation to resolve here.

Personally, I think those other languages in the original Arch wiki mostly do no good because they are usually too old to be considered helpful. Some are even harmful due to their old contents. The international Arch communities care about their language and keep their wikis maintained. Perhaps we should refer strictly to them and remove international languages from the original wiki? Or should we encourage all other communities to contribute all their articles to the main wiki and remove theirs? It's complicated.

pointone · 2009-12-27 06:29:16

Gen2ly wrote:

I like this idea too. I'd like to begin implementing this. Any thoughts, considerations?

I agree; all that remains is to decide on the standard (English language identifier (e.g. German), localized language identifier (e.g. Deutsch) or some other short form as karol has listed above?) I favor the full localized version.

faelar wrote:

I missed this thread before, but you should be interested by our solution on ArchServer : http://bbs.archserver.org/viewtopic.php?pid=60#p60

I made a similar suggestion above, and am glad to see a working example. I think having both localized names and accompanying English redirects will greatly improve navigation for non-English users.

Svenstaro wrote:

Perhaps we should refer strictly to them and remove international languages from the original wiki?

Whilst possibly outdated, these translations may still be useful. I agree that independent wikis should be encouraged, but articles should not be removed until transferred or otherwise replaced on another wiki. There's no sense in removing articles unless proven to be irrelevant or harmful.

Last edited by pointone (2009-12-27 06:30:04)

faelar · 2009-12-27 11:50:44

Independent wikis have some downsides :
- If an article is missing, I must go back to the main wiki anyway.
- If I know intersting tips about, say Xorg, and I post on my wiki, nothing guaranty it will be translated and added to ArchWiki. So maybe I will find half of what I need on archlinux.de wiki, the other half part on archlinux.fr, and nothing on the main wiki...

However, you end with better content (Doing some translation myself, sometimes it sounds rather odd). I don't really know what is the best solution... maybe clear guidelines for the organisation of all the wikis ? and some bridges between them...

Svenstaro · 2009-12-27 11:59:48

A rather crazy idea of mine that just came up in my twisted mind was to provide machine translations of articles that have no properly translated international pendants yet. The moment a user would write the actual article, the machine translation would no longer appear and everybody would get to see the real article. This way, we could always provide updated information derived directly from the original English sources. The Google web translation service has gotten to acceptable levels even for technical documentation. As long as a big red fat marquee blinking disclaimer is provided, might this work?

pwd · 2009-12-27 13:21:46

Svenstaro wrote:

Good thing this discussion has been kicked off. I was meaning discuss this for some time now. The current Arch Linux language handling is a mess. I agree that there should be a consistent way for handling non-English languages. We could keep our wiki strictly English-only and refer to the language-specific Arch communities. This would probably lead to a split-off community but those international communities exist anyway. Should we encourage them to contribute mainly to the English wiki? It is a rather hard situation to resolve here.

The only sane solution is to have different namespaces for each language, but that's just not happening. Proposed before/turned down.

English speaking admins just can't do a good job with content they can't understand, and vice versa.

Svenstaro wrote:

The international Arch communities care about their language and keep their wikis maintained. Perhaps we should refer strictly to them and remove international languages from the original wiki? Or should we encourage all other communities to contribute all their articles to the main wiki and remove theirs? It's complicated.

That's not the case. Even if you don't understand the language, a quick look at some of the recently updated pages shows that they haven't been touched in months or weeks. A lot of good content is spread between the main wiki and the intl. ones.

Edit: Well, not all of them are outdated; specially the German wiki seems to go against that perception.

Last edited by pwd (2009-12-27 13:26:36)

karol · 2009-12-27 15:44:59

> Well, not all of them are outdated; specially the German wiki seems to go against that perception.
Do you mean https://wiki.archlinux.de? That's one great wiki :-) I've tried to use http://translate.google.com on it - the text made sense, so I don't see a need for an automatic translation: if there's no / too little info on <you mother tongue> wiki, try other one. We just need a list of wikis a la Wikipedia http://wikipedia.org. Interwiki links would be nice, because I may not be able to find the correct wording for wiki.de or wiki.fr search - they're too terse, too short for translation services to work correctly.
Of course, there may be no article on xyz in English, so interwiki links won't help - hopefully Google will.

The problem is that it's imho impossible to keep the wiki up to date. It should be updated after every mayor change which in many cases occurs about twice a year. That's the cost of progress. Until we can clone ourselves or somehow spawn maintainers, I don't think that this'll change.

Gen2ly · 2009-12-27 19:15:30

pointone wrote:

I agree; all that remains is to decide on the standard (English language identifier (e.g. German), localized language identifier (e.g. Deutsch) or some other short form as karol has listed above?) I favor the full localized version.

Since the title name is only important to the relevant locality, I think the full localized is the way to go.

pointone wrote:

faelar wrote:
I missed this thread before, but you should be interested by our solution on ArchServer : http://bbs.archserver.org/viewtopic.php?pid=60#p60
I made a similar suggestion above, and am glad to see a working example. I think having both localized names and accompanying English redirects will greatly improve navigation for non-English users.

I like this too. Been looking at the language categories and a number of people from different locales will be needed. Currently we got 18 language categories with a good number of them already having 'English Title (Language)'. To do this they will need to be renamed to 'Título Inglés (Idioma)' and then redirected back again. Will begin moving (redirecting) pages that already have a local title to the English title, unless there are any major objections!!?

Last edited by Gen2ly (2009-12-27 19:27:18)

totalwormage · 2009-12-31 15:38:05

Gen2ly wrote:

pointone wrote:
I agree; all that remains is to decide on the standard (English language identifier (e.g. German), localized language identifier (e.g. Deutsch) or some other short form as karol has listed above?) I favor the full localized version.
Since the title name is only important to the relevant locality, I think the full localized is the way to go.
pointone wrote:
faelar wrote:
I missed this thread before, but you should be interested by our solution on ArchServer : http://bbs.archserver.org/viewtopic.php?pid=60#p60
I made a similar suggestion above, and am glad to see a working example. I think having both localized names and accompanying English redirects will greatly improve navigation for non-English users.
I like this too. Been looking at the language categories and a number of people from different locales will be needed. Currently we got 18 language categories with a good number of them already having 'English Title (Language)'. To do this they will need to be renamed to 'Título Inglés (Idioma)' and then redirected back again. Will begin moving (redirecting) pages that already have a local title to the English title, unless there are any major objections!!?

good idea, it works both ways, improves non-english navigation and wiki maintainers would still know what the article is about.

Gen2ly · 2010-01-01 05:22:52

Went through all the categories and renamed any non-English titles to 'Title in English (Lingua)' so things are up-to-date now for locale titles (at least as long as they were categorized). I had planned to rename non-English titles of the type 'Los Títulos en Inglés (Inglés)' to ''Los Títulos en Inglés' to 'Title in English (Inglés)' then deleting the original but this created broken links. I did this for the first few Espanol titles before realizing this - so look for some broken links there. As for renaming titles in the native-tongue (Bonjour) and then re-directing them to the English title (Welcome) they can still be done for those that are interested. It's nice feature that some may find useful.

I noticed that most titles where already in in the English title, however, a good deal of the categories are not. I'm guessing the categories are going to fall under the same category?? (can't see how they wouldn't). If categories are going to be renamed then there is to consider that creating a new Language category each page that contains that category will need to be edited. Perhaps there is a better way to do this? Manual doing this seems like a lot of work.

A couple other things I noticed as I went through:

Category:Nederlandstalig is wrong, should be Category:Nederlands.

Some translations marked the 'Official Arch Linux Install Guide' as 'Official Arch Linux Install Guide (Lingua) (Deprecated)'. I removed (Depracated) and marked articles 'Out of Date' as it seem the proper way to do this.

Last edited by Gen2ly (2010-01-01 05:33:13)

pikiweb · 2010-01-01 17:23:05

Gen2ly wrote:

As for renaming titles in the native-tongue (Bonjour) and then re-directing them to the English title (Welcome) they can still be done for those that are interested. It's nice feature that some may find useful.

I personally like more the idea of redirecting from the "english title (language)" to the fully localized title "Title in language" than the opposite way around, I think it makes it more clear to non english-speaking people like me, even the difference is quite small. Why "through away" internationalized titles in an international wiki?

Gen2ly · 2010-01-01 22:34:51

I'm thinking this standardizing for titles ('Title in English (Espanol)') not only helps Admins find pages for maintenance but also for tasks like bot maintenance for cleanup tasks. The titles aren't technically 'thrown away', they are still there but are now exist as redirects. I've seen it in a good number of wikis use this structure and believe it's important to keep this here as well. Perhaps to help non-proficient English readers a fully localized ToC could be used (i.e. all links there should point to fully localized titles). This can help with any ambiguities!?

Last edited by Gen2ly (2010-01-01 22:40:23)

pikiweb · 2010-01-04 16:21:40

Gen2ly wrote:

I'm thinking this standardizing for titles ('Title in English (Espanol)') not only helps Admins find pages for maintenance but also for tasks like bot maintenance for cleanup tasks. The titles aren't technically 'thrown away', they are still there but are now exist as redirects. I've seen it in a good number of wikis use this structure and believe it's important to keep this here as well. Perhaps to help non-proficient English readers a fully localized ToC could be used (i.e. all links there should point to fully localized titles). This can help with any ambiguities!?

I think it should work quite good that way. Would it then be possible and useful to have a 'Title in Language' page redirecting to 'Title in English (Language)', for each of the last ones? I'm thinking of people using "Search" function in particular

pointone · 2010-01-05 03:09:51

A few issues remain to be discussed; please see my updated user page: http://wiki.archlinux.org/index.php/User:Pointone/i18n

Specifically:

1) What should be done with categories? i.e. should categories follow the same scheme (Title in English (Language)) or something different (keeping in mind that category redirects will not behave ideally)?
2) How should i18n templates/linking be implemented? See http://wiki.archlinux.org/index.php/Use … te_Sandbox for some early ideas.

I'd prefer to use English category names simply to facilitate administration. i18n links could also be added to category pages. I also prefer the "Main Page"-style i18n links.

pointone · 2010-01-09 04:09:47

Another concern: Is there really a need for separate "Português" and "Português do Brasil" categories, or could the latter be merged into the former? As I understand it, the difference is comparable to that of "British" versus "American" English. Are these dialects not mutually understandable?

tlaloc · 2010-01-09 10:07:33

I don't speak either variant, but it should be pretty much the same. We have only the Brazilian variant in pacman translations, and no one from Portugal ever complained. We do have a en_uk version, but I considered that always to be kind of an internal joke (no offense to guy who writes that)

Arch Linux

#1 2009-12-21 04:26:03

The current state of ArchWiki internationalization

#2 2009-12-21 12:31:54

Re: The current state of ArchWiki internationalization

#3 2009-12-21 13:30:53

Re: The current state of ArchWiki internationalization

#4 2009-12-21 15:59:58

Re: The current state of ArchWiki internationalization

#5 2009-12-21 17:21:06

Re: The current state of ArchWiki internationalization

#6 2009-12-21 18:05:35

Re: The current state of ArchWiki internationalization

#7 2009-12-21 19:38:58

Re: The current state of ArchWiki internationalization

#8 2009-12-22 19:06:31

Re: The current state of ArchWiki internationalization

#9 2009-12-26 17:52:27

Re: The current state of ArchWiki internationalization

#10 2009-12-26 23:46:40

Re: The current state of ArchWiki internationalization

#11 2009-12-27 04:58:38

Re: The current state of ArchWiki internationalization

#12 2009-12-27 06:29:16

Re: The current state of ArchWiki internationalization

#13 2009-12-27 11:50:44

Re: The current state of ArchWiki internationalization

#14 2009-12-27 11:59:48

Re: The current state of ArchWiki internationalization

#15 2009-12-27 13:21:46

Re: The current state of ArchWiki internationalization

#16 2009-12-27 15:44:59

Re: The current state of ArchWiki internationalization

#17 2009-12-27 19:15:30

Re: The current state of ArchWiki internationalization

#18 2009-12-31 15:38:05

Re: The current state of ArchWiki internationalization

#19 2010-01-01 05:22:52

Re: The current state of ArchWiki internationalization

#20 2010-01-01 17:23:05

Re: The current state of ArchWiki internationalization

#21 2010-01-01 22:34:51

Re: The current state of ArchWiki internationalization

#22 2010-01-04 16:21:40

Re: The current state of ArchWiki internationalization

#23 2010-01-05 03:09:51

Re: The current state of ArchWiki internationalization

#24 2010-01-09 04:09:47

Re: The current state of ArchWiki internationalization

#25 2010-01-09 10:07:33

Re: The current state of ArchWiki internationalization

Board footer