There are so many proposals entitled: Stack Overflow (in language x).

My question is - are these proposals actually viable?

For example, if these sites reach beta, will the SE team implement a completely separate site for them?

e.g. http://stackoverflow.fr (for french), http://stackoverflow.de (for german).

Or will they have a similar take to this question. - For example, if you post a question with the german filter on (de.stackoverflow.com) - metadata is added to that question to say that it is in german, and then that question will only appear on de.stackoverflow.com - but all the sites use the same users and everything else as stackoverflow.com?

If these sites are going to be implemented, how will they be implemented?

share
2  
I also wonder the answer for your last question a lot. You didn't mention it in the question but I think you separate the other non-English sites from the scope of the question and consider only the code-related proposals. Because it is OK to have non-English sites as discussed in meta.stackoverflow.com/questions/52331/… and discuss.area51.stackexchange.com/questions/2571/… –  petrichor Jan 7 '12 at 13:42
2  
@İsmailArı so I should rename the quesiton "How will non-english sites be implemented"? –  Alex Coplan Jan 7 '12 at 13:52
1  
I think the name is nice. I just wanted to add the links to the related discussions. I don't know any non-English site launched by SX Inc and I really wonder what they think about the format and the process. –  petrichor Jan 7 '12 at 17:27
2  
@İsmailArı German Language and Usage is in public beta phase. –  kiamlaluno Jan 10 '12 at 2:39
    
@kiamlaluno Thanks, I didn't know that. It seems to be a good example for bilingual SX site. –  petrichor Jan 10 '12 at 10:05
1  
It will be viable when: 1-) They make it possible to translate the user interface of the localized version of Stack Overflow and 2-) When they ease on the requirements for getting a proposal to beta stage. Currently it is nearly impossible to get a localized version of any StackExchange site because you can't gather enough people with the required reputation - They have no reputation (and most likely will never earn any) because they don't participate on SO and that is because THEY CAN'T UNDERSTAND ENGLISH, which is why all of these proposals were created in the first place. –  Caio Proiete Jan 22 '12 at 18:38
4  
I'm sure there are foreign language programming languages, but I really hope anyone programming in any major language would be more than capable of navigating the basic nav structure of stack overflow. Programming is so English-centric (and I find it hard to call that a bad thing, since it fosters communication!) I doubt any non-English version would be particularly helpful, it would just segregate other language programmers away from the good answerers on SO. –  Ben Brocka Feb 26 '12 at 17:48
1  
@Ben Brocka I agree with you except that people who don't understand English is already segregated from SO. You probably don't live in a country where a big portion of programmers don't understand English. –  bigown Jul 13 '12 at 14:53
7  
Small point: The stackoverflow.fr pattern doesn't work, as the 2-letter TLDs are named after countries, not languages. And we want a "Stackoverflow in French," not "Stackoverflow for France". The distinction is more obvious for any languages that don't have a corresponding country name (Hindi, for example). fr.stackoverflow.com would make much more sense. (To say nothing of the problems registering stackoverflow.* at a bunch of unrelated registrars.) –  Flimzy Jul 18 '12 at 23:03

10 Answers 10

up vote 102 down vote accepted

First of all, English is not my native language and I'm not a big fan of using it. Even given that I'd say those proposals are doomed - they have no chance of creating a site as useful as original StackOverflow.

There're two reasons why I think so. The first reason is: if the site is not in English you lose most of the current top users. The second reason is those site will dilute effort. I speak English and Russian - I would now have to participate in two sites and answer the same questions twice.

And finally it is not really necessary, because currently you can't be a qualified developer unless you can write and read in English at at least some intermediate level - it is just impossible, because most of the blogs, documentation and new books are in English.

This is why I firmly believe that currently StackOverflow in any language except English is destined to languish in poverty - it will have little traffic and be mostly useless.

share
6  
Actually I agree with you in most points, but you probably not consider that some languages are spread far enough to build a solid base of documents(blogs,tutorials,..). For example there is a german version of the Microsoft Developer Network, so someone from a german speaking region can actually become a developer without knowing much more english than the coding languages english keywords. –  SwissCoder Feb 11 '12 at 15:50
17  
@SwissCoder: I can't evaluate the German version of MSDN, but I've seen the Russian one and it's horrible - the translations are done formally and are often just misleading so I'm much better off just reading the English version than try to deduce what the translator meant. –  sharptooth Feb 17 '12 at 7:48
    
No, at the german MSDN site, they don't do that. Sure some resources get translated automatically, which are not realy readable many times. The difference is, they do let people translate many pages. And they even have German people writing own articles about the technologies, so you get good resources there. –  SwissCoder Feb 17 '12 at 16:24
25  
Very true, if you don't know English, you can't be a software developer. Period. –  markus Feb 26 '12 at 23:49
8  
English is not the center of world. Of course anyone can be a great software developer without know english. –  Rafael Jul 13 '12 at 2:23
14  
It's a lot of arrogance and prejudice stating that you can't be a developer if you don't understand English. Most developers I know don't understand English and they are the main target to a native language SO. They can be poor developers, sure, and they need to improve their knowledge. Why just English speakers can be developers? It's a blind attitude. –  bigown Jul 13 '12 at 15:04
3  
And I'd like to add that just because some country originated a technology, that doesn't mean everyone else must be constrained to working with it in that country's language. A common ground might be necessary (think lawyers and latin), but it's unfeasible to force everyone to learn that language before working in the field. That would raise the barrier of entry a lot, no sane country would accept that, especially if the main "benefit" of doing so would be to contribute better with the country that already has a huge head start... –  mgibsonbr Jul 15 '12 at 3:33
3  
I've worked in software shops where almost none of the devs speak english, but coincidental or not, they don't participate in devs communities also. So I agree that, even when english is not needed to be software dev, these communities don't have much chance. –  Eduardo Molteni Jul 23 '12 at 14:40
    
+1 totally agree. –  Stephane Rolland Nov 9 '12 at 15:11
    
It's probably useful to keep a distinction b/w county and language. They aren't the same thing. True yes in the US we speak English, but Jeff and Joel selected English b/c they speak it, not b/c they live in the US. –  slm Nov 3 '13 at 15:35

They're as viable as any other proposal in Area 51, and just like any other Area 51 proposal, if the proposal has sufficient support from a community, it will get created.

It's a bit premature to discuss the specifics of a hypothetical site, but the thinking right now is that language-specific sites will NOT be created as a filtered version of — in this case — Stack Overflow. They will be created as independent communities with their own subdomain, their own users, their own tags, their own meta, their own moderators — in short, their own community with their own norms.

Beyond that, the actual implementation details will be for the dev team to decide as they work out the localization issues.

share
2  
Did you intentionally use "subdomain" here? i.e, you're intending to create programmation.stackexchange.com rather than stackoverflow.fr as the OP specified? –  Kevin Vermeer Jan 9 '12 at 20:38
1  
@Kevin Yes, "subdomain" is intentional. Currently sites are named as subject.stackexchange.com. Sites do not receive their own domain name but typically are a subdomain under "stackexchange.com." There are no immediate plans to follow a different naming convention for international sites. But that can all change, once a site is launched. It's still a bit too soon to speculate definitively. –  Robert Cartaino Jan 9 '12 at 23:52
4  
@RobertCartaino instead of allowing Secession to happen, what about user-improvable translations of SO? –  Tobias Kienzler Jan 11 '12 at 12:05
    
Stack exchange is a safe & ,reliable sorce in putting it mistakes or human errors Web I've doubled up emails to the same wallets they give I it original details I'd certainly go straight to area 51 stack exchange over any difficulty a peer may AV I wish I'd known it was on offer for the official&original details lost or forgotten very highly recommend dnt waste time get straight to area 51 stack exchange for a question I may AV trouble over –  Andy Farnworth Dec 6 '13 at 17:25

I believe that country and/or language specific sites are indeed viable, not as much because of contents (I agree English is the lingua franca of software development, and refusing to learn it is stupid) but mainly because of culture.

A community is defined not only be its objective properties (ex. what subjects are on/off topic) but by subjective ones as well (ex. how strictly you have to stay on topic, how much duplicate questions are tolatetad, how personal/impersonal the questions can be, etc). Two or more groups of people may have similar goals (ex. develop software) but wildly different means of collaborating with each other to seek those goals.

I believe this Wikipedia article about Low Context Culture vs High Context Culture might be relevant to illustrate my point. Certain behaviors that feel "natural" for a high context individual can be seen as "annoying" or "disruptive" by a low context one. Conversely, normal behaviors by the latter are seen as "cold" and "unhelpful" by the former. A site with similar objectives but different culture may cater more to large niches of people who otherwise would have trouble "fitting in" a different one.

At last, I'd like to point out that just because someone has knowledge of a foreign language it doesn't mean s/he has the fluency and vocabulary to be productive in that language: I often find myself wanting to know about a subject, and while I know precisely the concept I'm interested on I can't translate it to the right English words to google it. Trying to ask other people is equally clumsy, and when I'm finally able to explain myself the retort is usually like: "oh, so you mean X?" ...and I realize that searching for X would have given me what I wanted in the first hit. This situation is embarassing for me, wastes other people's time, and is much more prevalent than a native English speaker might think it'd be.

share
  1. I am indian i know hindi but its pretty tough to write hindi

  2. All good it/computer books and resource are in english than if one can read english where is problem in writing english

  3. I dont know any programming languages which uses hindi.

  4. if we keep stackvoerflow in one languages it will be more helpful since there are more user who know english i think 10x so getting proper help would be easy than a hindi stackoverflow site

also i would say same about all other language x stackoverflow proposal site

share
3  
1. I'm Portuguese and I know Portuguese, and it's pretty easy to write it. 2. There are several good programming books in Portuguese. 3. I know a programming language that uses Portuguese. 4. You're only right in this point. And indeed I'd say the same about many other language x Stack Overflow proposals. –  JMCF125 Nov 3 '13 at 17:24
    
I'm English n was all ready in English termination I'd not relised it was multi langues –  Andy Farnworth Dec 6 '13 at 17:27

================

Analysis

================

Good Side

People who don't have enough English knowledge or just don't prefer English can participate.

I work in a wonderful software company which uses both English and Sinhalese languages to communicate. So when a person feel uncomfortable or stuck when speaking in English they can easily switch to Sinhalese and they will not be frowned upon. Also I have noticed that there are some people who prefer to use Sinhalese for nearly all purposes even if they are fluent in English.

No more misunderstanding occurred by software-language-translations

I have seen some good questions/answers posted on Stackoverflow that were simply lot harder to understand because of the automated translation used by original poster and/or other participants.

Eliminate Language Barriers

Bad Side

Duplicate Questions

  1. Harder to search in languages that you don't know.
  2. Harder to flag as duplicates.

Cannot read

  1. You don't understand the language.
  2. Rare font-rendering or unicode related problems.

Create Knowledge Barriers

================

Solving it

================

What to do ?

First we need to accomplish good and then eliminate bad.

Eliminate language barriers while preventing knowledge barriers.

How to do it ?

  1. Allowing people to ask questions in commonly used languages(or any language that has an online translator API support).
  2. Automatically Providing an To-English-Translated keywords and meta-data or even the whole question/answer so it can be easily Googled.
  3. Give points (or use manual translation up-votes) to people who provide good manual translations.

GUI Example

GUI Sample

share

It is possible to edit questions on stackoverflow. I would believe that if someone who does not know english, if they post the question in russian/spanish or any other language, there would be enough bilingual people there to edit and translate the question to english. Same for answers, someone could answer in english and someone else could take the time to translate the answer back to the OPs language. Although it sounds like more work, it will prevent fragmenting, and since the edits give points, the incentive is always there.

share
4  
That might work if the questions were for instance tagged in the target language and filtered out for everyone who doesn't understand it. Otherwise, it would create much noise and make the site unusable (imagine opening the "newest questions" tab and seeing several questions you don't understand, having to paginate several times to find the ones you're interested on). –  mgibsonbr Jul 15 '12 at 3:15
1  
Support. It's obviously inefficient to separate SE across linguistic borders, but it's kind of draconian to insist that all SE Q&A be in English (to say nothing of all the language learning sites within). Tagging+filtering would solve this nicely, and definitely expand the commmunity. –  Trevor Alexander Sep 30 '13 at 0:30

When the first proposal came up for a German StackOverflow, I supported it. That proposal has failed. Since, I have been granted access to moderator tools etc. on SO, gained a bit more of a big-picture view and have now concluded that such seperate site are not the right way to go about this.

Yet I do think there should be a way to post questions in other languages.

Here's what I'd propose:

  • There is in fact only one SO site, though with multilingual versions of all the labels etc.
  • When users sign up, they specify both their preferred language and a number of secondary languages; expectedly almost everybody will have English as one of them.
  • Questions can be posted in any language where there are users that speak it (and somebody has gone through translating the FAQ etc. at some point). Those questions are only shown to users who understand it, preventing Babel.
  • For users with multiple registered languages, there will appear a translate button on all posts. That allows to generate, most often, an English equivalent of a non-english question, and for some particularly popular questions perhaps the other way around (similar to Wikipedia). When translating, the user may opt to be notified about further post on the question in order to translate these as well / translate back to the original question's language so the OP can understand it (which though, recall, is not the primary purpose of SO questions!). You might of course also just post a question in multiple languages right away.
    I'd expect translation to happen only in the more interesting questions. Rather the majority is of course pretty trivial, in which case an answerer would be more likely to just give a quick reply in the OP's language. But the great thing about SO is the really interesting questions that come up occasionally; in such cases I'd often be happy to translate just to see more useful answers.
  • The translated version[s] of a question will appear to all users speaking any of the available languages. The different languages are not seperate threads, but the same thread with different view parameters. Inside each question, there will for each post be the possibility to view other language versions, and do edits to any of them (but not all languages are shown by default, we don't want the jarring mess of <hr>-seperated subposts). Translations might also be voted upon to give some reputation for the effort. Posts available only in a language the user doesn't understand are by default at the bottom / greyed out, but may still be useful through machine translation.
  • When somebody posts a question that has previously been asked in another language, there will be the option to merge the two versions: possibly with minor changes, the new question text is used as a translation of the old one. Answers already given to the new question would just be added to the ones on the old questions. Especially if there are none, a proposer of the merge should be encouraged to translate at least the top one of the old answers to the new language.
  • Only for search engines / users not logged in, there will be actual disjoint sites with country-specific domain, displaying only one language.

Of course, this would mean major changes for the entire StackExchange framework, but after all these guys have proven they can handle hard tasks with great results, haven't they?

The changes would in fact mean we could instantly have non-english posts not just on StackOverflow (I think the only site big enough so other languages can stand a chance of getting their own version at all) but also on all other StackExchanges. Most often, there just won't be anyone who needs anything apart from English; in this case pretty much everything stays just as it is. But whereever another language is needed for some reason, it can be used, without fragmenting the community, without requiring actual experts on each question's subject in every language (you can just translate to english, where there are experts), and without needing to manually switch from site to site.

share
    
Excellent ideas. I think you intended this, but I wanted to say that users should be allowed to post in several languages with they ask the question. That way, someone can post in their native language, and give English a shot as well. –  nograpes Feb 23 at 2:17

I don't like English. I have to know it, but one shouldn't be forced to know English in order to program; programming should be like Mathematics in this aspect: even if I use an "if" construction, I don't need to know exactly what it means as an English word (of course all programmers do, as this example is trivial), but as a symbol, as a construction for a condition.

We don't have to come up with another solution though. I think this already exists. In Wikipedia, this problem has been solved. And that's what we should follow.

So yes, they are viable. Is this the way to do it? No. It would be like creating completely separate versions of Wikipedia per language and saying it's a different site.

This means the system must be changed. Each question in a language should be linked to its duplicates in other languages' Stack Overflows. A button "translate this to..." would help. We don't need new Stack Overflows in different languages. We need to expand the original into those. This is just a suggestion on how to, there probably are other ways to follow the Wikipedia model.

share

I think I am a huge fan of different Stackoverflows in different languages, because some people are not that great in English and should be able to use their native language to ask their question to get better in programming.

Although I think this is a great opportunity for them to learn programming, I think we should keep all the great questions and answers in one central place, the English Stackoverflow. Why? Because what if a Dutch guy is asking the master of all questions? Only the Dutch people will be able to read the answer. I think our knowledge should be shared with every programmer so everyone can get better, instead of only the Dutch programmers.

FYI. I am a Dutch guy, so I am not saying this because I don't understand the language, I am saying this because I am Dutch and I think we should not get a different SO for every other language we can think off. Just so we can share our knowledge with everybody instead of only the people of our own country, we have forums like that enough already. I have seen some great questions/answers by English people, guys from India, Turkey or Germany. From everywhere. When we all get on separate websites, we will not be able to read/share our awesome knowledge anymore! We won't be able to Google the answer to our question if it is posted in another Stackoverflow, We won't be able to follow the discussion of a topic and we can not reply if we do know the answer!

I agree with every word said by Sharptooth.

Long story short: do NOT get separate Stackoverflows for every language, and thus keeping everything on Stackoverflow (english).

share

I'm a native speaking English person and primarily participate on the Unix and Linux site. I think as a business of providing a centralized resource for everyone in the world that wants to discuss subject X, creating many sites is counter productive, if you're truly trying to provide the best centralized resources.

I realize it's burdensome for people of the world to learn English, and it can be frustrating for these users too, so as is typically done in the computer and software industry, I say it would be more productive to try and create better tooling around translating the site into their native language where appropriate.

My idea

I've used google translate from time to time when I come across a non-English blog or site and it's worked reasonably well. Is it perfect? No, but based on the comments on many of the other answers, much can get lost in translation from one language to another.

By providing tools we could do the same thing with the translations as we do with the site, where members over time could refactor those translations through editing and commenting. Think of it as an additional view on top of the English backed content.

In this way if the viewers don't understand the translated version, they could always "view source" and see the actual content in English. It's a bit of a site within the site. It could be done using a few languages to start.

Creating these translations could be seeded with google translations to start, I would also limit to key questions to start.

Mixing in the data from which sites by top level domain visit could also be mixed in so as to only roll this feature out to questions that they frequent.

Closing thoughts

There are many directions to go with the above idea. But I'd much rather keep the pool of people as large as possible and not create competing silos of information than for no other reason than one group of people speaks X and another speaks Y.

Our power is our numbers and our different ideas we all bring to the table, and I for one would rather see a central solution that we all can participate in, rather than 100's of diluted sites where we all have to sift through looking for information.

share
1  
Machine translation doesn't work, not to the extent that this would be viable. It usually works well enough so that you know what a text is about (and even then only for simple enough texts and not for all language pairs). For a technical text where it's important to tell the difference between “X is Y” and “X is not Y”, forget about machine translation. –  Gilles Nov 5 '13 at 22:18
    
@Gilles - the machine translation was just to seed the initial results. The translations would be editable and modifiable in the same way as the site sites are editable by members. In this case the translations in other languages would be alternate views on the same page that's already in English underneath. This is a rough idea, anyway. –  slm Nov 5 '13 at 22:25
    
It's a terrible idea. If you seed the site with machine translations, you'll seed the site with crap. That's not an auspicious beginning. Instead, seed the site with real, useful content. –  Gilles Nov 5 '13 at 22:27
    
@Gilles - it was just an idea. So take the machine translation bit out of the equation. Does what I'm suggesting seem like a feasible way to go otherwise? Multi-language isn't something I deal with day to day (in the sense of having to work within it) so I'm not sure if I'm completely off the mark with the views idea on top of the English backed site. –  slm Nov 5 '13 at 22:39
    
What is your idea if you take out machine translations? To start with translations of SO questions instead of questions freshly asked in language X? That won't work. Translating stuff is a lot less interesting than producing fresh content. This has been a consistent experience on French Language: in principle we encourage bilingual posts, and in theory all meta posts should be bilingual, but in practice nobody bothers and the few meta posts that are bilingual were a drag. –  Gilles Nov 5 '13 at 22:44
    
@Gilles - yes the general idea was to make translations an attribute that would hang off of a Q or A. But given your feedback, it doesn't sound like that's viable either. –  slm Nov 5 '13 at 23:54
    
The Wikipedia model sounds more viable. Any thoughts on it? (please comment my answer below) –  JMCF125 Nov 7 '13 at 20:17

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .