Talk:Top Ten Wikipedias/poll

From Meta, a Wikimedia project coordination wiki
Jump to navigation Jump to search


Where are the translations? It's a important Poll - for such things is a translation in the most used languags a must. Not all people speak english very well. Have they to be quiet? It's an international project. Wherever a translation an one of the main languages is missing, the result can't be acceptable. Marcus Cyron 14:11, 8 July 2008 (UTC)

There's an edit button: be bold and add them. --::Slomox:: >< 14:27, 8 July 2008 (UTC)
Most wikipedias have added a translation under the announcement that was made in their village pump. I'll post the links later today, right now I am out of time :P --Waldir 14:49, 8 July 2008 (UTC)

Ok, here are the links and current status:

Current top 10
remaining 100K+ wikipedias

Anyone who is able to translate these please do so, and update the pages here :) --Waldir 19:05, 8 July 2008 (UTC)

Determined by the language of the browser, OS, and IP of the visitor?[edit]

If the result of Question 3 is NOT Option A (do nothing), how is the language determined? Referring to browser, OS, or IP? (These setting may not be the same, such as a Chinese OS using English browser though a IP in Germany.) Another problem is that: is it technically possible to change the layout according to users' language? I especially worry about the possibility of Sub-option D.2 & Sub-option D.3 (replacement options). Orangemoon 04:40, 9 July 2008 (UTC)

It is definitely possible to replace the languages on the list, but your question does make me wonder if it's possible to detect a user's language. —Kal (talk) 04:49, 9 July 2008 (UTC)
The specific details of each question's implementation weren't discussed in depth during the discussion phase. But this is not a controversial issue. A little discussion among web-savvy editors, and the devs, might be enough to determine the best method to choose a user's language. For example, it has been proposed to only make the choice of a local language if the browser language settings, the OS and the IP all agree, but the criteria can be a little less restrictive -- it is up to us to decide, and anyway, it can always be changed later if it is proved non-effective. :) --Waldir 09:53, 9 July 2008 (UTC)

Question 3[edit]

Hello! Regarding Question 3, I think that before we set upon a specific technical solution, this question should be removed from the present poll and asked at a later date. Top Ten Wikipedias#By IP has a few discussions concerning this but aren't very extensive at the moment. I think that this question will require much more discussions delving into the specifics of the determination procedure for a user's language. I think that it would be best to have the results of the first two questions implemented first before adding the new functionality of a language-specific portal display. I also think that this could later be extended to Wikimedia sister-projects' portals. Although I am not entirely against the idea, I've voted "No" to question 3 because I am unsure as to how successful the language determination function will be. The poll page vaguely states that it'll be "determined by the language of the browser, OS, and IP of the visitor". Could it be a combination of these three? Will a user be able to set one's own preferred language, specific to that user's IP address? Will the system be able to "learn" and correct mistakes if a user from a specific IP address consistently selects a different language Wikipedia link than the one determined from the browser-language/OS-language/IP-origin information gathered? What options would a user have in case the preferred language determined by the system consistently fails to match that of the user's actual preference? (For example, an active editor for the Russian Wikipedia located in Israel using Hebrew OS and internet browsers, consistently having the system automatically place the Hebrew link at the top, while pushing the Russian link out of the top 10). How will dynamic IP addresses be accounted? Each of these questions a nd many others will affect my answer to question 3. Unlike the first two questions, I think question 3 needs a working model first before a decision could be made. Therefore I propose the postponement of this question for a later date. Thanks for reading! --Shibo77 06:21, 12 July 2008 (UTC)

When a user wants to have a specific language, he changes his user preferences. When we are to guess what language goes with a specific PC, there are several ways in which this can be determined. Only when a user is known, we can give him his preferred language.
I fail to see why you vote against something because you "do not see how something is to be implemented". If you want to know technically how something is to be done, develop it. If you are in favour of an idea but vote against just because you are do not understand the technicalities, you vote against what you want... I do not understand why. GerardM 07:00, 13 July 2008 (UTC)
Hi! I would not rashly put a vote on something that I do not understand, I feel that that would be quite irresponsible on the part of the voter. At the moment, Question 3 states "Should a visual clue be used to highlight the language version determined by the language of the browser, OS, and IP of the visitor?". The example of the hapless .RU Wikipedia editor with a Hebrew system and browser, and an Israeli IP address that I've given above, now would that fit the statement of Question 3? I tend to think so, and that is something to which I am against and would vote accordingly. If the developers are able to bring about a solution and demonstrate it to us voters, only then would I be able to grasp the intricacies of this new functionality, voice my opinion and place a vote.
In your reply, you talked of a user's specific language, does that mean the user has to be logged on, as in a "User:"? I believe that many visitors to the main Wikipedia portal concerned here in this poll, will not be logged in, and possibly not even registered. If you are talking about the general visitor to the Wikipedia portal, then I have to ask, how will you enable the general visitor to set user preferences for a preferred language? Would this warrant a new functionality to be given on the portal for anonymous visitors to set their perferences? Would the preferred language chosen by the visitor be linked to the originating IP address? What about shared computers and dynamic addresses? I suppose in the case a guess fails, nothing is to be changed, no language links highlighted or pushed to the top? Or I suppose the English link would become the default? Disregarding the portal's anonymous visitors, at present, unified login still has many problems and I should think being given a Wikipedia wide user preferences setting based on their unified user account would be highly problematic. As a non-developer, I can only see the option is to guess the user's preferred language, and as put forth in Question 3, the guess being made on the basis of the origins of the visitor's IP address, the language of the visitor's internet browser, and the language of the visitor's operating system. I've already voiced my concerns regarding this.
Overall, I regard this as a new functionality quite distinct from the first two questions, moreover, the third question's statement is quite ambiguous on how exactly this new function will be implemented and how successful it will be. How will it solve the RU Wikipedia editor example I've described, and the many other concerns I've voiced in my original question above? I think this question should be taken out of this poll and a separate poll regarding this question could be set at a later date when the mechanisms for the implementation of this functionality matures. Or perhaps a developer could rephrase Question 3 to help the voters better understand the process involved in "guessing" a visitor's preferred language link. Votes signed previous to the change-in-statement would probably have to be discounted. As far as I'm concerned, personalisation begins with registration, and the user can then set one's own preferences. While I agree that quick and easy navigation for the logged off or unregistered is of great benefit, I am skeptical about its successful implementation. I hope this has helped you understand the "why" of my vote against, and thanks again for reading! --Shibo77 05:33, 28 July 2008 (UTC)

Rephrasing the end-of-poll result[edit]

I want to point out that in point 3 of "How to vote", we should just eliminate the part that says "(considering a majority of 2/3 over each of the other options)". If the most voted opinion does not reach two-thirds, essentially the poll becomes moot. OhanaUnitedTalk page 04:28, 12 July 2008 (UTC)

Although a majority of 2/3 is not reasonable, a majority of 1/2 may be considered. If so, when no options gets 1/2 or more votes, a second round of vote can be held and the options are the 2 most voted options in the first round. Otherwise, if just "the most voted option" (in this round) is chosen, such option may get less than 50% support (especially Question 3). --Hargau 06:34, 12 July 2008 (UTC)
Just for the record: the poll description was (by omission) assuming a simple majority, but I changed that, after the following comment in Talk:Top Ten Wikipedias:

Well, I think this is pretty important and major issue on very exposed page thus I'd expect there should be very major consensus reached for the change. Absolute minimum of 2/3 but rather 3/4 of votes. So yes, I am strongly oppose to simple majority because it wouldn't show the consensus, and the result could be very fragile (simple majority is also n : n+1).
Danny B. 04:18, 8 July 2008 (UTC)

But it does make more sense to me that this be discussed, as I agree that this poll shouldn't end up in a stall for too restrictive passing criteria. By the way, I am in favor of 1/2, with a second round if that ratio is not reached (or if it is too close according to the perception of who closes the vote). That would actually be good, since more thoughts and proposals have been added to Top Ten Wikipedias that perhaps could be taken in consideration (I know, those people should have commented while we were in the discussion phase, but I guess it always happens this way... polls always receive more attention than "mere" discussions). --Waldir 10:56, 12 July 2008 (UTC)
Just go for simple majority, because it's... simple... OhanaUnitedTalk page 17:38, 13 July 2008 (UTC)
But I think a majority of 1/2 can ensure the consensus. If just using simple majority, the chosen option may just have a little more than 1/3 support, which means that there may be over 1/2 opposing that option. -- Kevinhksouth 02:34, 14 July 2008 (UTC)
You're right, but note though that the people who vote for other options don't explicitly state their disapproval of the one with most votes. A concrete example: people who are voting for keeping the current article count criterion probably disapprove the traffic option, but those who are voting for the most spoken languages, for example, are probably not (just guessing).
This is also unfair for questions with more options, since the voters might get split up among several choices, diminishing the possibility of one of them to reach 1/2 of the total votes.
What about using a 2/3 majority, not against the total votes for each option of a question, but against each of the other options individually? --Waldir 13:35, 15 July 2008 (UTC)

Any more opinions? Or this will be the consensus? OhanaUnitedTalk page 17:53, 22 July 2008 (UTC)

We're approaching the closing date. I believe that since there are no objections, yes, we can go forward with the last proposal made in this thread (to avoid any ambiguity, it's "2/3 majority, against each one of the other options for the same question, individually"). I seriously hope people not to show up afterwards complaining about this, the same way as people showed up for the discussion after the vote has started... :( --Waldir 12:55, 28 July 2008 (UTC)
I don't understand clearly. For example, if option X is 60%, option Y is 31%, option Z is 9%, do you mean that option X still cannot pass, as comparing option X (60%) with with the total of option X & Y (91%), option X is not 2/3 majority? If no option passes, what should we do? -- 09:37, 29 July 2008 (UTC)
You are right, and this is the original concern expressed by OhanaUnited. I have it also, but I can also see the reasoning of Danny B. and Kevinhksouth. The problem is, we have to find a compromise between not changing important stuff with weak community support and not having such a mass community effort end up in a useless stall... We basically only have to choose the threshold. Honestly if I had to choose, my personal opinion is let's use the 50%+1 rule (simple majority).
I find it easier to look at this issue through a graphical example: in the chart on the left, even a 1/2 rule (50%+1) would mean we'd choose the option "SPD" (red). A 2/3 rule would prevent that, since the option "CDU" is that close. But if CDU had, say, about 15%, then yes "SPD" would win even though the remaining options combined would overweight it.
That obviously isn't the case in politics: a country would have no government in such a case, would a 2/3 majority be required. Thus, simple majority suffices. Indeed, why should we be more picky about such a change?
So, what do we prefer, if we had to choose: having a weak consensus, or declaring no consensus? I prefer weak consensus — at least we'd have an outcome, and wouldn't disregard all the effort that the community has put in a month-long, interwiki-wide vote, with discussions on the village pumps of dozens of the largest wikipedias, etc. --Waldir 23:30, 30 July 2008 (UTC)
Or, we could have a second round with only the two most voted options. but I wonder if the people would bother showing up again... then we'd be forced to declare no consensus by lack of community participation, which would be even awfuller. --Waldir 23:34, 30 July 2008 (UTC)
I would agree on lowering the majority to 1/2. Hargau and Kevinhksouth agreed on 1/2 majority. That's 4 out of 5 people agreeing to amend it. I echo Waldir's concern. A month long poll with messages to village pumps is pretty much what we could do the most, unless someone can send out an email to all members in all projects and ask them to vote (but this would be impractical). Having over 50% of support when presented with a large amount of opinions ensure consensus. Then if some opinions do not reach the 50% threshold, such as the current situation for Question 3, then we can re-run a smaller poll just for that option without voting everything else again. OhanaUnitedTalk page 14:28, 31 July 2008 (UTC)
Agreed. Well, it looks like the poll is over. I am going to have little time online the next few days, so someone please make the summary of the results, and ask an admin to protect the page. --Waldir 03:11, 1 August 2008 (UTC)

Voting Instructions[edit]

Ok, maybe I'm just being dim, but the voting instructions seem very unclear to me. From they way I read thing, each editor can vote for one of the options for each of the questions. A sample of my interpretation is shown below (with my comments in italics). In addition, having the questions' and options' text spread out all over the page makes navigation very difficult, regardless of what that text may read. I am concerned that without clarification and improvement of the voting page and process, results will be inaccurate. — MrDolomite • Talk 22:46, 24 July 2008 (UTC)

  • Q1 - Which criterion should be used in the main wikipedia portal ( for sorting the various language editions of wikipedia?
  • choose one option from the list below to support.
    • A = Number of articles
    • B = Most visited wikipedias
    • C = Size of compressed database
    • D.1 = Number of articles with a given minimum size of 1 kB
    • D.2 = Number of articles with a given minimum size of 2 kB
    • D.3 = Number of articles with a given minimum size of 3 kB
    • E = Most spoken languages

  • Q2 - Should only one of the above criteria be used (for Q1), or a combination of them?
  • choose one option from the list below to support.
    • A = One — 1 (the most voted one, on question 1.)
    • B = Two — 2 (the two most voted)
    • C = Three — 3 (the three most voted)
  • — MrDolomite • Talk's comment for Q2 - remember that this choice for criteria is independent of how you voted for Q1. Example: just because you voted D1 (articles over 2 kB) for Q1 does not mean that when you vote A (use one criteria) for Q2 that the one criteria to be used will be D1 (articles over 2 kB).

  • Q3 - Should a visual clue be used to highlight the language version determined by the language of the browser, OS, and IP of the visitor?
  • choose one option from the list below to support.
  • — MrDolomite • Talk's comment for Q3 - I admit, I am very confused about Q3 C and Q3 D.1-.3. Does this mean that for Q3, I should really be voting one option from A-B-C-E and also voting for one option from D.1-D.2-D.3?
    • A = No.
      • — MrDolomite • Talk's comment - No, a visual clue should not be used. This is almost a double negative and may confuse voters
    • B = Reverse video.
    • C = If the [visitor's] "local" language version is one of the top ten: Reordering of the ranking (dynamic portal approach: move the local version to the top position and the others which were above it will drop one position each)
    • D.1 = If the [visitor's] "local" language version is not in the top ten: Do nothing (leave it as it is)
    • D.2 = If the [visitor's] "local" language version is not in the top ten: Replace one of the top ten the editions (remove the last one of the top ten and add the "local" version in its place)
    • D.3 = If the [visitor's] "local" language version is not in the top ten: Replace the editions (remove the last one of the top ten and add the "local" version in the first place, moving all the others down one position)
    • E = Different font color
      • — MrDolomite • Talk's comment - Yes, a visual clue should be used and it should be a change in font color (e.g. like this)

  1. First of all, yes, your interpretation is correct: "each editor can vote for one of the options for each of the questions"
  2. Secondly, replying to your comment to Q2, you're right, like it is exposed right now it is impossible to choose, for example, to use one criterion only if the option you voted for is the chosen one. But it is not the purpose of this question to decide that. Q2 only tries to balance between a simple, easy to calculate, understandable ratio, and a more complex (combined), less cheatable one, regardless of which criteria are used.
  3. > "I admit, I am very confused about Q3 C and Q3 D.1-.3. Does this mean that for Q3, I should really be voting one option from A-B-C-E and also voting for one option from D.1-D.2-D.3?"
    Not exactly: it means that it works as normally, only one option can be chosen from A to E. But IF you vote C, you can also vote for D(1,2,3), and the same thing in reverse.
Feel free to improve the descriptions of the vote if you feel the result can be biased by the confusingness of the current descriptions. But also bear in ming that nobody has complained so far, so it might be wise to wait for more people to reply in this thead before. --Waldir 09:05, 25 July 2008 (UTC)
Just some comments: Although I cannot vote, I would still like to point out a serious problem in Q3. It is confusing in Option C & D, as it violates "each editor can vote for one of the options for each of the questions" and make the situation a lot more complex. If someone can vote Option C & D at the same time, it creates trouble when counting the votes. For example, if one sub-option in option D gets the most votes, how to decide whether option C should be chosen or not? (As what I understand, Option C & D are non-exclusive so they can be chosen at the same time.) -- 09:53, 25 July 2008 (UTC)
Both C and D are measured against the other options, not against each other. It is like having A-B-C-E and A-B-D-E (and this double counting would only occur if C or D are the most voted option), it's pretty simple actually. I mean, if you can think of a simpler way of making that choice (given the content of the options C and D), please suggest it! --Waldir 10:41, 25 July 2008 (UTC)

Final Result[edit]

Here's the final result for the poll, % rounded to one decimal place (bold text indicates a majority and established consensus, that is, the option receives over 50% of total votes):

Question 1 (385 votes)[edit]

  • Option A: 33.5%
  • Option B: 57.4%
  • Option C: 0.8%
  • Option D
    • Sub-option D1: 0%
    • Sub-option D2: 3.2%
    • Sub-option D3: 0.3%
  • Option E: 4.9%

Question 2 (248 votes)[edit]

  • Option A: 64.5%
  • Option B: 31.9%
  • Option C: 3.6%

Question 3 (249 votes)[edit]

No option exceeded the 50% threshold, action required

  • Option A: 27.7%
  • Option B: 0%
  • Option C: 13.8% (non-exclusive with Option D)
  • Option D (non-exclusive with Option C)
    • Sub-option D1: 0.9%
    • Sub-option D2: 3.9%
    • Sub-option D3: 37.7%
  • Option E: 27.3%

So... languages are reordered already and use only 1 criteria. What should we do with the visual clue to highlight users' local language? OhanaUnitedTalk page 18:41, 2 August 2008 (UTC)

Strong disagreement[edit]

I strongly disagree with this presented result.

  1. The poll in Question 1 ended up with 59:41 ratio which is way so close to each other to be considered as a consensus. The difference is only 18 %! How come only 18 % of users can set such important thing which main Wikipedia site is?
  2. The conditions have been changed during the poll, which is totally unacceptable procedure - you can't let part of people vote in some conditions and other part in other conditions.
  3. It was never declared in poll, which method is supposed to be considered for Question 1 Option B. Also other options lack closer specifications.
  4. As I already commented in Talk:Top Ten Wikipedias#Necessity of Top 10 vs. full table listing this entire poll has been manipulated from the very beginning towards certain option.

So my words about it's totally unfair from the very beginning and my apprehensions that this poll is just a comedy to push some opinion were both right.

I am strongly disappointed and feel very uncomfortable by the entire procedure because of lack of fair conditions, same start line for all options, lack of specification of criteria and many other defects.

I would strongly suggest to annul this poll (or maybe better say its result presented by VasilievVV) and start from scratch with correct and equal conditions for all options again since this definitely does not represent community consensus and was manipulated from the very beginning.

Danny B. 01:44, 3 August 2008 (UTC)

I will answer your concerns.
  1. You have to look at the old consensus (use # of articles as criteria) vs. then-purposed consensus (changes to the consensus). Viewed this way, 66.5% (just 0.1% shy of 2/3 majority) wanted changes to be made.
  2. As in what? Can you give us examples?
  3. I, as well as almost everyone, thought that the description in each option is clear enough. Alexa is the household name for measuring traffic and is up-to-date. Others are outdated from half a month to 3 years. Which option would you choose then?
  4. You are free to view the poll in your own views. But it seems like not a lot share similar thoughts as yours. From the link you provided, only you and Waldir exchanged ideas and nobody else. If people think there's a problem then they would have voiced their concerns.
We're not lawyers. We don't use law terms in Wikipedia polls. We don't argue the difference between the interpretation of majority or simple majority. It's not fair to us when you disappear for half the time in July while the poll is happening and then come back after the poll ended to share your displeasure. And my final words is that I'm disgusted to see you saying "(the poll) was manipulated from the very beginning". What you said was not assuming good faith and offended a lot of people. OhanaUnitedTalk page 15:53, 3 August 2008 (UTC)


As far as I can see Option B of Question 1 received 221 out of 385 votes. That's 57,4%, not 59% as stated above. Wikijens 22:52, 4 August 2008 (UTC)

Thanks, corrected now. OhanaUnitedTalk page 00:20, 5 August 2008 (UTC)


I would say that we have at least some consensus that marking the user's local language (somehow) is welcome. Over 70% voted for one of the Options B, C, D or E. Which version to use is probably a matter of personal style preferences. Maybe we need to discuss whether somebody strongly dissagrees to the variants that most users supported (D3 and E combined). -- 10:32, 11 August 2008 (UTC)

You cannot combine those two. The vote instructions stated clearly that all options were mutually exclusive, except for 3C and 3D. It is feasible to both move the local version to the globe and highlight it with a different color, but that was not what people voted for. As the wuestion was posed, people were choosing whether to highlight it on its place, or move it to a more visible place, without highlighting. What we can do is make a second round for question 3 with only these two options (D3 and E). What do you think? --Waldir 16:58, 14 August 2008 (UTC)
(Was me above) I don't really see why it wouldn't be possible to do both D3 and E. Marking the matching wiki in some kind of special color is quite unrelated to it's position on the screen. But I agree that it needs another poll, maybe with options D3 only, E only or both. --PaterMcFly 21:42, 14 August 2008 (UTC)
I said it myself: it is possible to do both (in my own words above, "It is feasible to both..."), but it just was not what people voted for. In the second-round vote we can of course have an option to choose one, the other, or both. Anyone willing to set that up? Waldir 01:45, 15 August 2008 (UTC)