49

I recently visited area51 for the first time and was disturbed to see proposals for Stack Overflow variants in other languages dominating the site.

I'll admit that seeing these worries me greatly. What is the reasoning behind allowing international variants of Stack Overflow? I understand it might be easier for non-native speakers, but it risks fragmenting our knowledgebase over multiple languages and segregating our users. Is this really the proper approach when English is, essentially, programming's lingua franca?

There has been discussion in the past relating to non-English Stack Exchange sites:

And many outstanding points have been made on the subject:

But I'm still rather curious. Why are we suddenly forgoing Stack Exchange's Non-English Policy in favor of fracturing our userbase?

As an afterthought: A great comment was posted by Emrakul (Actually that entire comment thread is lovely. Read it.) on the answer found just above.

We need to see charts of activity by country to gauge how many people this would bleed away from the main site. My guess is, it's a lot more than you'd expect.

Does this information actually exist anywhere? It would be amazingly useful/interesting if it has been made available to Stack Exchange users.

A Note: I've tried to cite as many great discussion items as I can on the subject. Quite a bit of information exists on alternate language variants, but, to my knowledge, this question has never been directly asked.

In summary, the question asks "Why should language variants be allowed?" and "Does the reward really outweigh the risk of dividing our user and knowledge bases?". NOT "Is it okay to be proposing these sites?", "Should these questions be closed?", or "Does Stack Exchange support sites in languages other than English?" (all things which have already been asked).

17
  • 1
    It's worth stating that just because they're allowed to propose the site doesn't mean they'll actually be created as main sites.
    – Servy
    Aug 28, 2013 at 14:36
  • @Servy Most have made it to beta. The new multilingual interface hasn't been implemented yet so they're on hold. Aug 28, 2013 at 14:37
  • 10
    Upvoted because I'm also not a friend of splitting communities (even though SE will probably/hopefully do it in a proper way so they aren't actually separate sites but a single site with various "views" depending on the languages someone wants to see). In my opinion especially programmers are supposed to know English anyway. Aug 28, 2013 at 14:39
  • 14
    Beta still isn't a main site. Putting the sites in beta will help create a lot of data to help verify or contradict the guesses and assumptions made in the discussions you linked. If they end up creating more good than the problems, we have strong data to contradict our guesses of the problems. If the fears are realized, we now have confirmation that they shouldn't be created in the future.
    – Servy
    Aug 28, 2013 at 14:40
  • 1
    Stack Overflow (maybe the whole Stack Exchange network) is going through major localization these days so I guess we'll soon have something like http://stackoverflow.com/en/questions/... vs. http://stackoverflow.com/pt/questions/... - dunno if it will replace those proposals though. Aug 28, 2013 at 14:43
  • 5
    I think a better idea would be to let the community translate each question and answer--think of it as a kind of editing capability. I agree that SO in multiple languages is a bad idea if implemented as multiple, separated sites. Aug 28, 2013 at 15:17
  • 2
    @ɹǝʞɐʇıɥʍʍǝɹpuɐ How do you see that? Wouldn't that be an utter maintenance nightmare?
    – Bart
    Aug 28, 2013 at 15:34
  • @Bart Any more than repeated questions and answers across the network? If you solve the localization problem the latter way you end up hiding the maintainence problem instead of putting it right in front of users Aug 28, 2013 at 15:35
  • Repeated questions are not a maintenance nightmare @ɹǝʞɐʇıɥʍʍǝɹpuɐ. And given the question coming into SO at the moment, I'm hard pressed to believe we'd be missing out on much original material on either side. But having to translate each and every question and answer to the various languages does not seem to be a feasible endeavour. Unless I'm completely missing the point of what you propose.
    – Bart
    Aug 28, 2013 at 15:38
  • @Bart: I don't see another way to avoid the duplication of knowledge, which seems like it outweighs the benefits of multiple versions of SO. Maybe I'm putting too much weight on downsides of having the same question asked multiple times and losing a single canonical reference for a particular question though... Aug 28, 2013 at 15:42
  • @ɹǝʞɐʇıɥʍʍǝɹpuɐ I think it's an idealistic and noble view of a single set of content, but not a practical or feasible one. I get where you're coming from, but I don't think it's realistic.
    – Bart
    Aug 28, 2013 at 15:43
  • @ɹǝʞɐʇıɥʍʍǝɹpuɐ Just imagine trying to track duplicates across multiple languages while awaiting a translation. Nightmarish stuff. Aug 28, 2013 at 15:46
  • 7
    @ZachSmith You don't need a formal system in place to do that. If someone finds a really good question on a site, with a good answer, they can translate it to other sites, post it on those sites (cite the original of course) and then, if other users find it valuable, they can then vote on the translated version. If the content is not useful, they don't get rep, if it is, they do, based on votes. This would encourage the translation of the quality content, and discourage the translation of the crap.
    – Servy
    Aug 28, 2013 at 16:23
  • 1
    @Zach, there are an estimated 750m to 1,500m speakers of English. Less than 400m are native. There are over 7 billion people on the planet -- even with the most generous estimate, English tops out around 20% of the world population. Compare this map with this map. There is a huge audience for other languages.
    – jmac
    Sep 2, 2013 at 2:27
  • 1

8 Answers 8

41

"Why should language variants be allowed?"

Because there is an audience for them. It's really as simple as that. If a site graduates, there is an audience sizeable enough that a site like that is sustainable.

"Does the reward really outweigh the risk of dividing our user and knowledge bases".

In my opinion that greatly exaggerates the risk. I've already commented on that in one of the questions you link (yeah, should have really made that an answer at some point):

I don't even think we'd lose users (to a significant extent). Those who do speak English will still come to SO. Don't underestimate the power of that site. And perhaps they will also participate on the site whose language they speak as well. I don't see those as mutually exclusive activities.

and

If they successfully participate on Stack Overflow, which is a huge quality resource with more activity than any proposed site is likely to have, why would a site in a different language cause an exodus? If they have a difficult time participating on an English language site, maybe, but then you'd have to wonder if it's so bad for them to find their support elsewhere. I don't see all Dutch users move to an SO in Dutch. I know I wouldn't.

as well as

If a large section of those Chinese users has difficulty with the language, you can fret all you want about the number of visitors going down, but it might well increase the overall quality of SO, require less community support such as editing, and additionally provide that group of users with an excellent set of resources as well. I see only wins here...

In the end I think it will be beneficial to some extent to both the users whose grasp of the English language is not all that great, as well as Stack Overflow who might see some users leave, but overall quality increase.

5
  • 12
    The worry here is that people are, can be, a bit lazy. It would easier for a native Chinese speaker to ask a question on the Chinese Stack Overflow than it would be to ask it here on the English Stack Overflow. Ultimately, in my opinion, any meaningful question with pertinent, useful knowledge that bleeds to a language variant is a greater loss to us than any gain had from losing poor quality questions that would just be moderated out of existence anyway. Aug 28, 2013 at 14:51
  • 6
    There are two things to consider here. For SO/SE it might make a lot of sense to have an SO in Chinese, should that become a seriously large site. Not because that's where all the current Chinese users would go, but it's where all those who currently don't participate at all can ask their questions. Secondly, I have much more faith in the strength of SO than you seem to have. This is where the questions are. This is where the people are. This is where users get their answers faster than anywhere else. Laziness or not, you'd have to be pretty damn good as a site to pull people away from SO.
    – Bart
    Aug 28, 2013 at 14:55
  • Being Dutch myself, I don't think the Dutch are a good example. Dutch people tend to be comfortable with the English language. Aug 28, 2013 at 15:04
  • @S.L.Barth Which I was I tried to mention both ends of the spectrum in my comments/answer. My argument is that SO is strong enough to still pull those in who have a decent level of English. And that we perhaps should not worry too much about those who don't have it, if there is an alternate quality resource for them within the network. (P.s. I mentioned the Dutch because I'm Dutch myself, and I've seen the SO in Dutch proposal on Area51. Don't see that one happening. )
    – Bart
    Aug 28, 2013 at 15:05
  • 9
    @ZachSmith Remember that ESO has a lot more utility than any future language SO. People know that they will get faster (arguably better) answers on SO. The askers who will go to P.SO/C.SO are those who would have had a hard time writing it in English. And usually, such users find it hard to clearly express their problem on the site (even when the problem is actually a very nice one!), getting closed as unclear -- so SO isn't losing much if they go to P.SO, P.SO gains, the ESL Internet gains, and the user himself gains. Win-win-win-win. Aug 28, 2013 at 15:40
24

What risks?

Since you posit that there is a risk, it's up to you to say what you think the risk is. Do the benefits of allowing users to post images outweigh the risks involved? Do the benefits of allowing claims that Java is a real programming language outweigh the risks involved? Do the benefits of allowing teen moderators outweigh the benefits involved? Do the benefits of allowing users whose name begins with Z to ask questions outweigh the risks involved?

it risks fragmenting our knowledgebase over multiple languages and segregating our users.

In your whole question, this is the one line that explains your position. That's not much to go on.

Why do you think that having SO-like sites in other languages would fragments the knowledge base of SO? Roughly speaking, there are three kinds of programmers in the world¹:

  1. Those who are comfortable participating on a site in English. They won't be affected by the existence of SO in other languages.
  2. Those who can read English but not express themselves. They can benefit from SO, but they can't contribute.
  3. Those who can't speak English at all. They can't even benefit from SO.

Creating SO sites in other languages will drive programmers of type 2 away from SO. But they weren't contributing to SO in the first place, so SO isn't losing anything by it.

A slew of unwarranted assumptions

A lot of the claims in your question are plain wrong.

English is, essentially, the world's lingua franca

That is true (for programmers — programmers tend to be better at English than the average profession), but only to some extent. There are many type 3 programmers, and even more type 2 programmers. Of course, you don't tend to meet them on SO or other English language venues: the language excludes them.

If you search the web a little, you will find forums about programming in Chinese and French and Russian and Portuguese and other languages. There are clearly people who do want to converse about programming in languages other than English.

Why are we suddenly forgoing Stack Exchange's Non-English Policy in favor of fracturing our userbase?

Stack Exchange has never had a non-English policy. It is rather bizarre that you should make this claim, considering that you cited an official post by Stack Exchange founder Joel Spolsky stating that there is no such policy:

it is OK to propose new sites in other languages. (…) For example: (…)

  • Programming (in Russian)

The non-English question policy is for the existing sites. Stack Overflow is English-only. As Stack Overflow founder Jeff Atwood put it:

It is not, nor has it ever been, our goal to be the one place in the world for all programming information in every possible human language.

All of this applies to Stack Overflow and the other sites of the original trilogy. SO was and is about programming, in English. Stack Exchange is a different thing altogether: SE sites can be about any subject, in any language.

We need to see charts of activity by country to gauge how many people this would bleed away from the main site. My guess is, it's a lot more than you'd expect.

I have no idea how many people Emrakul expects to find, but that number is irrelevant anyway. How would charts of activity by country tell you anything? The people who are participating on SO now aren't going to stop if there are sites in other languages. The ones who are on SO now are the ones who are comfortable enough with participating in English. Most of them are going to stay. SO in other languages is largely intended for the people who are not participating on SO now, because they can't.

Your follow-up questions

Why should language variants be allowed?

Why not? You've yet to explain why they shouldn't be. And as you acknowledge, there's an official policy that they are allowed.

Does the reward really outweigh the risk of dividing our user and knowledge bases?

Seeing that you don't seem to be able to articulate why there would be a risk, that doesn't put much of a burden on the reward. For the current participants in Stack Overflow, there are several rewards that may be small but still nice to have. There will fewer incomprehensible posts by people who just can't make themselves understood in English but currently have nowhere else to go. There will be a place where answers from non-English-speaking programmers can be found, when such answers are currently buried in the webforums that SO sought to replace if they exist at all on the web. For the people who cannot participate in Stack Overflow due to the language barrier, the reward is considerable: they will gain access to a questions and answers site about programming.

A glaring omission

Nowhere in your rant question do you consider the benefits for the programmers who don't speak English enough to participate on SO. They would gain the same benefits that you and I gain from SO. That is a considerable gain that would outweigh a significant cost to SO — and that cost has yet to be demonstrated.

If you can have nice things, why not let others have them too?

¹ Plus those who don't understand that you can count to 3 even in binary.

6
  • Does the risk of not knowing how to spell 'involved' ...
    – Rosinante
    Sep 2, 2013 at 1:39
  • 3
    In the end, for Stack Overflow specifically ... I don't think it matters. Programmers know that the best chance they have of finding and getting a great answer will be on the English site due to the nature of programming -- I mean, your keywords are in English, aren't they? Not to mention the umpteen million existing questions there in English. So the alt-language Stack Overflows, if they ever launch, will be at best a minor sideshow attraction compared to the main tent. Language is much more interesting for non-programming topics on the rest of the network. Sep 4, 2013 at 7:27
  • The risk is that those programmers on the edge of learning English will get lazy and stick to their pet language SO variants, without any incentive to learn English and thus have access to vastly more information in general. Oct 7, 2014 at 2:59
  • @DanDascalescu Please be nice and do not denigrate people who don't speak English. Oct 7, 2014 at 7:28
  • @JeffAtwood: well, those sites did launch, and they did fracture the community and discourage users from learning English. Feb 12, 2019 at 10:28
  • 1
    @DanDascalescu [citation needed] Feb 12, 2019 at 15:33
23

I don't think those proposals will significantly drain away users that are comfortable with the english language. I'm not a native english speaker, but I tend to visit sites in english and use english terms when searching for technical topics. In many areas there are just more and better english resources than in my native language, so using english has an advantage if your language skills are up to the task.

Stack Overflow in english will also remain by far the largest of the SO sites, through sheer inertia and the prevalence of the english language in programming. So I think that SO in english will continue to be the resource of choice for programmers that can read and write english reasonably well.

But the sites that are proposed are not primarily meant for users that speak english well. They are meant for users that struggle with english or don't speak it at all. And I don't see this as a disadvantage. Someone who is truly struggling with english will have a better experience on a site in their native language.

I'm convinced that learning english is a very important skill for a programmer, but I don't think it is our place to try and force that view onto everyone. Not that we can, we just drive those users to crappier resources in their own language.

The SE team also stated at some point (I don't remember exactly where) that they only want to create those sites for languages where users tend to avoid english resources. There won't be a dutch SO because they pretty much all speak english well enough.

I think creating duplicates of SE sites in different languages makes sense when they can reach a significant user base that does not use english sites.

2
  • 4
    You have totally equal. Dutch shall normal SO stay using, because their English good enough is. (Sorry, just had to. This will only possibly be humorous to the Dutch) But yeah, couldn't agree more.
    – Bart
    Aug 28, 2013 at 15:28
  • Brilliant answer! Couldn't be more perfect.
    – Maniero
    Nov 7, 2013 at 14:17
7

These sites don't belong to 'us'. They belong to a commercial enterprise that is trying to create economic value. If that commercial enterprise thinks that they can create more value with sites in Portuguese or Klingon, more power to them! And to whatever extent they belong to 'us', you are not speaking for me in attempting to throw tomatoes at these proposals.

'Fragmentation' is a herring of the finest redness here. There's room in the world for knowledge replicated in many languages; just look at Wikipedia. What about experts? Well, those experts who are capable of answering in Lower Middle Frisian will answer questions there. Those that can (only) answer in English will stay put. The only possible problem is if SO.com is suddenly fresh out of SQL expertise because all of the experts have run over to the Sumerian site to answer questions in Middle Babylonian.

If they happens, the team can hire Maxwell's Daemon to shoo them back.

7

When one seeks for a piece of specialized technical information, it is typically not done by "browsing one site". It typically starts with an Internet search, and that search is conducted in a particular language. If the asker has a real choice of the language to use, they will use the language of people who are likely to know the answer. This motivation applies not just to the search, but also to creation of a new question on a Q&A site.

Lot of computer technology has been created in English-speaking countries, and large expert bases live there, but not all the technology or all the experts. I expect that the non-English sites will naturally specialize on a relatively limited, but still sustainable (for some languages) width of topics that will have some regional, linguistic or cultural dimension, doubtlessly accompanied by very basic general support to programmers who aren't good at navigating the English speaking Internet.

For me, the former is about more efficiently reaching into new information niches, and the latter is mostly about simple, beginner's questions whose reduction will be no loss to the English speaking SO.

Can a programming question even have a regional, linguistic or cultural dimension? Yes, it happens.

  • Language processing technology.
  • Character encodings, input methods, data exchange formats and conventions.
  • Globalization and localization techniques.
  • Integration to software that survives regionally and that isn't normally even spoken of in English.
  • A non-English, locally popular resource makes somebody thinking and asking.
  • An unexpected error message came out, oh no, localized.
2
  • It would be lovely if the types of programming questions you've listed were asked on regional sites, but that's not the case. As I feared 5 years ago, creating localized version of SO has fractured the community and siloed knowledge. Take a look at pt.stackoverflow.com for example. Perfectly generic questions are asked in a language that has only ~15% of the number of speakers of English. If a PT.SO user answers a question in Portuguese, that answer is lost to the rest of the world. Feb 12, 2019 at 10:27
  • 1
    @DanDascalescu - All posts on that page seem to be in Portuguese. Those posters were under no obligation to hire a translator to get their questions posted on Stack Overflow in English, nor to post in Portuguese on the English site and see them promptly closed. I'm not sure that I care personally. My Portuguese is very basic, I would probably choose to skip every single one of those questions for that reason. Feb 12, 2019 at 11:36
4

To address the question in the title: we'll see how it goes, in practice.

"StackOverflow in Portuguese" should reach private beta soon – and I'm not talking about 6-8 weeks, the team just announced (via email, in Portuguese!) that private beta should start somewhere during the next few weeks, and public beta early next year.


UPDATE

Private beta started last week, https://pt.stackoverflow.com/

-2

Current SO experts that are not popstars (but do contribute with quality) and are not native english speakers may feel tempted to easily become popstars in their native SO site, thus reducing or simply stopping contributions to english SO.

1
  • 1
    You are partially right here: it's much more valuable to me to be a "popstar" in my local community (where I do real-life business, no less) than to "compete" with all the other experts in a global community I'll never truly be part of. The english-speaking community will indeed lose with this (not the SE company, though: ad revenue is ad revenue). The extent of this loss is what I find debatable (besides, there will bet a lot less n00bs flooding the site with "poor quality" questions), especially when compared to the "win" for the other communities (it's not a zero-sum game).
    – mgibsonbr
    Feb 1, 2014 at 10:43
-6

No the benefits don't outweigh the risks involved. Since StackOverflow is the only site I really use regularly this answer only pertains to that site.

Think about the huge amount of information available on StackOverflow. This took years to build up and now we'll have duplicate questions and answers popping up all over the place. As an advocate of DRY, this sounds absolutely terrible to me. You could possibly end up with the same question repeated n times throughout the network, where n is the number of languages. Awful.

For example, just take a look at the sample questions that the Portugese SO proposal says are "Top Example Questions":

There are, of course, already canonical answers to these questions:

All of the proposals have similar language. The point is that they aren't "new" questions. The entire site would be a duplicate of StackOverflow.

If those proposals go through, we'll no longer have a canonical place to find a programming question and answer. How do you know a better answer does not exist on Spanish SO? You don't.

We should keep SO English-only for now.

I know this might sound cold or unaccommodating, but the fact is that the same problem I describe above would exist for every single language-specific SO. You'd never know if you were looking at the best possible answer to your question because it could live in multiple places.


A crazy idea that probably won't work

How about allowing the community to translate SO into multiple languages? There could be a tab on each question where a bilingual speaker of the other language could translate the English question. The non-English version would get marked "dirty" in case of edits. If nobody has come along to translate the question and answers, then you would obviously just see the English version.

I know this is way harder than it sounds, which is why my first recommendation is to keep SO English-only. Just throwing it out there.

This wouldn't be a mass translation effort so much as an editing-like capability that's community driven.

12
  • 10
    We have multiple sites that overlap with SO in topic (look up questions with the agile tag to see similar questions being on topic on Stack Overflow, Software Engineering and Project Management). How/why is having a language-specific site different? Why are you not considering all the people that can't access SO right now because of the language barrier? What would constitute the "best" answer for them?
    – Sklivvz
    Aug 28, 2013 at 16:08
  • 2
    Why would having similar questions in different languages be an actual problem? Aug 28, 2013 at 16:09
  • 5
    ¿pǝʌloʌuı sʞsıɹ ǝɥʇ ɥƃıǝʍʇno ʇuoɟ pǝddılɟ uı os ƃuıʌɐɥ ɟo sʇıɟǝuǝq ǝɥʇ op 'ɹǝʞɐʇıɥʍ ʍǝɹpuɐ
    – gnat
    Aug 28, 2013 at 16:11
  • 8
    Having the same question ask/answered in multiple languages does not violate DRY. It's not duplicate content. Having the same technical information written out in a different language is adding new information/value, just not new technical information. If someone doesn't speak English then the content on SO wouldn't help them. If they speak another language that has a duplicate, but well translated, then it does help them. That's adding value. The concept of DRY only applies when the added work doesn't add value.
    – Servy
    Aug 28, 2013 at 16:26
  • @Servy: Valid point-- I suppose I'm most concerned about the fact that it'd be hard to know where to find the best answer to a question when that question exists in multiple places across multiple languages Aug 28, 2013 at 16:35
  • 1
    @ɹǝʞɐʇıɥʍʍǝɹpuɐ Well, if you don't speak those other languages, you don't need to care. Just stick to the language site that you know. If you're someone who speaks quite a few languages then yes, you may want to search several sites for content you're interested in. Too bad there's no place that we could go to search for content across multiple sites, or even better yet, somewhere that could search basically every site on the web for content that's related to our problem. Google already handles the problem you're describing, and after all, SE was built around traffic coming from Google.
    – Servy
    Aug 28, 2013 at 16:37
  • 1
    @Servy: Not so sure I "don't need to care" about other languages. Why does it matter what language a coding Q&A is in? Does the fact that a coding question is in Spanish vs English really change the essence of what the best answer is? Maybe for other SE sites this logic applies but I don't see language mattering that much for programming questions Aug 28, 2013 at 16:41
  • 2
    @ɹǝʞɐʇıɥʍʍǝɹpuɐ It doesn't matter in the sense that you're unlikely to be able to find a question that matches your problem, and wouldn't be able to understand an answer even if you did. You wouldn't be able to use the site if everything was in a different language, so you'd be better off just not using it.
    – Servy
    Aug 28, 2013 at 16:43
  • 1
    @Servy It does violate DRY. Having translations of original ESO questions is much easier as the variation in information is marginal at best (essentially expect some differentiations due to translation difficulties). Multilingual duplicates raise the issue of a much wider margin of variation. What do we do when duplicates are posted on multiple language variants and each contains valuable information that we'd like to translate accross the variants? Aug 28, 2013 at 16:52
  • 1
    @Everyone The point here is to have a civil discussion, examine all sides, and raise whatever points we possibly can. ɹǝʞɐʇıɥʍ ʍǝɹpuɐ posted a quality answer that raised some valid points. He shouldn't be attacked/downvoted because of a difference in opinion. Aug 28, 2013 at 16:55
  • 2
    @ZachSmith So it's better to just not generate that useful content to begin with by prohibiting the use of other lanaguages, then to generate that useful content and fail to translate it into English. That seems like an, "If it doesn't help me, it's not worth doing." attitude. If the content is particularly valuable, it likely will be translated by someone, although I freely admit there will be some content in each of the sites that's at least somewhat valuable and doesn't get translated elsewhere, but some will, so all other languages will still benefit to a degree.
    – Servy
    Aug 28, 2013 at 16:56
  • 1
    @ZachSmith Downvotes on meta are frequently used to indicate agreement/disagreement with an opinion (when that makes sense in context). See the help center.
    – Servy
    Aug 28, 2013 at 16:58

You must log in to answer this question.

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