Template talk:Wikipedia languages/Archive 4

From Wikipedia, the free encyclopedia
Jump to: navigation, search
Archive 1 Archive 2 Archive 3 Archive 4 Archive 5 Archive 6 Archive 7

Other wikipedias

Hi isn't it about time the other wikipedia count was updated at the bottom. German wikipedia now has 818,834 articles and to place it in a 300,000 column seems to be a bit of an understatement and not giving it credit. in comparison its like saying English wikipedia has over 1 million articles. At the very least an over 500,000 needs to be created. German, French, Italian, Japanese and Polish all have over 500,000 and Spanish, Dutch and Portuguese have over 400,000 Dr. Blofeld 11:58, 28 October 2008 (UTC)

From memory, the problem is the 300k category is already rather small as is. If you increase the minimum, we will end up with an even smaller category leading to greater imbalance in the section. The section isn't really intended to give 'credit' to anything (for starters, article quality may vary even tho we have excluded those with primary bot created stubs) nor to compare them to the English wikipedia (indeed many people have objected to the excessive emphasis we put on the number of articles we have). It's intended to direct readers to other wikipedias they may be interested in and the size was chosen as the simplest, fairest way to sort them that will be most meaningful to the reader. A category with 9 items should be easy to look through for the majority of readers, reducing it further isn't really necessary. At most, I would suggest 400k be the new tier. If people still feel we are being unfair, I suggest we just remove mention of the number of articles we have from that section (we already have it at the top). You may be interested to know that for the main ones, www.wikipedia.org has moved completely away from the article size sorting system and instead started sorting by popularity of the wikipedias. Nil Einne (talk) 10:09, 29 October 2008 (UTC)

Azerbaijani wiki has now 22,787 articles. Please take into account for main page.

— Preceding unsigned comment added by (talk) 20:58, 10 May 2009 (UTC)

Persian Wikipedia now 50000

Persian wikipedia has reached 50000 articles. Please update the template. Dreamfall (talk) 09:37, 30 October 2008 (UTC)

Done. —David Levy 11:06, 30 October 2008 (UTC)

Tagalog Wikipedia at 20,000

Having reached the 20,000-article mark just this evening Philippine Standard Time (and currently standing at such), please add the Tagalog Wikipedia to the template. Thanks! --Sky Harbor (talk) 12:56, 1 November 2008 (UTC)

Done. —David Levy 16:50, 2 November 2008 (UTC)

Macedonian language edition of Wikipedia

The Macedonian language edition of Wikipedia has reached over 20.000 articles. Please add it to the list. Best regards, Bomac (talk) 19:06, 5 November 2008 (UTC)

Done.—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 19:46, 5 November 2008 (UTC)

Set new milestones

I'd propose removing the 50,000 article bracket and adding a 500,000 article bracket (example below). Alternatively, we could leave the 50k and add the 500k making a total of 5 tiers. 500,000 is a big milestone and I think it is worthy of mentioning. Thoughts?

What about cutoff points at powers of two? —Preceding unsigned comment added by (talk) 09:21, 10 December 2008 (UTC)
Extended content

This Wikipedia is written in English. Started in 2001, it currently contains 5,552,725 articles. Many other Wikipedias are available; some of the largest are listed below.

Rjd0060 (talk) 19:08, 6 November 2008 (UTC)

  • Makes sense to me. MBisanz talk 19:14, 6 November 2008 (UTC)
  • I find the 300k mark thorougly obscure, and would propose to follow the 1-2-5-scheme (you know this from a lot of currencies), therefore:
Extended content

This Wikipedia is written in English. Started in 2001, it currently contains 5,552,725 articles. Many other Wikipedias are available; some of the largest are listed below.

  • More than 1,000,000 articles: (as soon as it becomes necessary with german and french wikis in about a year or so)

Call me Monk, but isn't this a lot more orderly? ;-)

-Axel, 9:46, 19 February 2009 (UTC)

Telugu wikipedia over 40K

and yet, it is not mentioned under the 20K+ list in the template. Could one of the admins please add it? Thanks! - Kesava (talk) 02:39, 12 December 2008 (UTC)

The Telugu Wikipedia has a depth of 4. For inclusion in the template, our current minimum depth is 5, and it's been suggested that we increase the cutoff point to 10. We also check for evidence of depth owed primarily to automated edits. —David Levy 02:55, 12 December 2008 (UTC)
Can somebody tell me what is the current depth of telugu wikipedia. Since we have made some substantial contributions after the above conversation. --రవిచంద్ర (talk) 05:03, 13 March 2009 (UTC)
The depth (always listed at m:List of Wikipedias) is now 5. However, I just viewed several dozen random articles, and almost all were bot-created/edited stubs/placeholders. —David Levy 05:15, 13 March 2009 (UTC)

Swedish Wikipedia now over 300 000

{{editrequest}} So, can somebody update this template please? MiCkE 15:52, 13 December 2008 (UTC)

That is sv.wikipedia/svenska btw. MiCkE 15:56, 13 December 2008 (UTC)
Done Rjd0060 (talk) 16:16, 13 December 2008 (UTC)

Danish Wikipedia over 100000

Having reached the 100000 article --Hejsa (talk) 19:04, 28 December 2008 (UTC)

Updated. —David Levy 19:12, 28 December 2008 (UTC)

Simple English now 50k

{{editprotected}} The Simple English Wikipedia has passed the 50 000 article mark; could someone please update this template. (Originally requested by User:Razorflame at Talk:Main Page.) ChrisDHDR 14:42, 10 January 2009 (UTC)

Done. —David Levy 14:52, 10 January 2009 (UTC)

PMS at 20,000

{{editprotected}} Piedmontese Wikipedia arrived at 20,000. Could you update this template please? Thanks. --Piemont (talk) 13:54, 20 January 2009 (UTC)

Sorry, but the Piedmontese Wikipedia has a depth of 2. For inclusion in the template, our current minimum depth is 5, and it's been suggested that we increase the cutoff point to 10. We also check for evidence of depth owed primarily to automated edits. —David Levy 21:24, 29 January 2009 (UTC)

Is this template updated regularly or on request, as it seems? Cause there are some wikis that needs to be included but are absent. -- (talk) 09:26, 14 February 2009 (UTC)

The template is updated on request or when an administrator notices a need.
I've added the Latvian Wikipedia, which just reached 20,000 articles.
The Azeri and Serbo-Croatian Wikipedias' depths appear to have been artificially inflated via minor bot edits (with the former also containing many empty articles).
If you see any other missing Wikipedias, please let us know. —David Levy 16:21, 14 February 2009 (UTC)
It's not really possible to determine to what extent these wikipedias are effected by bot edits or the proportion of empty pages. Even if we find it out, these would reflect only a current snapshot. A cetain depth cutoff is reasonable but not such factors which can not be objecively determined. I think all wikis ove 20 K should be included, except those under certain depth. One another solution to prevent overpopulation in main page would be to increase the minimum number for inclusion, from 20 K to let's say 50 K. -- (talk) 08:46, 22 February 2009 (UTC)
I'm not sure what you mean when you say that "it's not really possible to determine to what extent these wikipedias are effected by bot edits or the proportion of empty pages." Simply visiting a Wikipedia and viewing a reasonable number of random articles (and their revision histories) paints a rather telling picture. For example, it didn't take me long to discover that the Azeri Wikipedia contains over 2,000 pages for the years of the Common Era, most of which contain absolutely no content other than bot-inserted headings and navigation links (which cause them to be counted as articles).
I also don't understand your "only a current snapshot" argument, as all of the inclusion criteria are based on the Wikipedias' current states. If one improves, it can be reevaluated. —David Levy 09:08, 22 February 2009 (UTC)
One can not judge, and usually it is possible to misjudge, such as yourself, from limited data, without taking larger picture into account. We should come up with something that can be determined objectively, be it depth, article count or something else. This template shows wikipedias over cetain number of articles, not those wikipedias satisfying every individual preference and random test. -- (talk) 10:08, 22 February 2009 (UTC)
You keep saying that "one can not judge," but you aren't explaining why you believe this to be so.
The list is described as "some of the largest" Wikipedias, and I don't see what's subjective or random about checking whether thousands of articles contain little or no information. —David Levy 16:49, 22 February 2009 (UTC)

Wikipedia Indonesia now

Now Wikipedia Indonesia has passed 100.000 articles, please read it. Reindra (talk) 16:29, 21 February 2009 (UTC)

Done —David Levy 20:56, 22 February 2009 (UTC)

Telugu wikipedia

Please include Telugu wikipedia (te) in both interlanguage links and in 20000+ article list. Telugu wikipedia has 43,000 + articles and is the largest Indian language wiki. Thanks --వైజాసత్య (talk) 13:21, 15 May 2009 (UTC)

As I noted above in the Telugu wikipedia over 40K section, I viewed several dozen random articles, and almost all were bot-created/edited stubs/placeholders. —David Levy 16:25, 15 May 2009 (UTC)

Arabic Wikipedia

Arabic now has over 100,000. -- [[SRE.K.A.L.|L.A.K.ERS]] 01:32, 26 May 2009 (UTC)

korean wikipedia became over a hundred thousand

korean wikipedia--Youthegreat (talk) 14:08, 4 June 2009 (UTC)

Congrats! Moved accordingly.—Ëzhiki (Igels Hérissonovich Ïzhakoff-Amursky) • (yo?); 14:38, June 4, 2009 (UTC)
Thank you--Youthegreat (talk) 14:38, 4 June 2009 (UTC)


Malaysian has now over 40,000 articles. --Metsavend (talk) 17:22, 9 June 2009 (UTC)

Thanks! Template updated accordingly. —David Levy 02:24, 10 June 2009 (UTC)

Hindi Wikipedia

[copied from User talk:David Levy]

This is regarding inclusion of Hindi wikipedia in the list of wikipedias over 30k articles.One of the condition for inclusion of a wiki in such a list is so called random test. check this link. But When I clicked on 'Random Article' link for 50 times. I saw only 5 articles containing greater than 5 lines, and two of them are English articles copied as it is from English wiki. Can you take a look at it and do the needful? —రవిచంద్ర (talk) 07:08, 26 June 2009 (UTC)

I conducted the same test and arrived at a similar result. Indeed, it appears that most of the Hindi Wikipedia's articles are stubs or placeholders (including some in English). Accordingly, I have removed this Wikipedia from the list. —David Levy 18:09, 26 June 2009 (UTC)
Manipuri wikipedia too does not confirm to the random test. Please have a look at this also. —రవిచంద్ర (talk) 05:59, 30 June 2009 (UTC)
I agree. Accordingly, I've removed that Wikipedia from the list. —David Levy 06:53, 2 July 2009 (UTC)
From the template's revision history: Added Hindi. This has 30K+ articles. Don't agree with David that it has 40% articles in English or not enough to count as articles. Shyam 07:46, 2 July 2009 (UTC)
1. I just again viewed 50 random articles, and only four were not stubs or placeholders. I hope that you aren't relying on the tables of content to judge, as a great many pages consist of multiple headings for empty sections. (These placeholders contain even less content than most stubs do.)
2. I don't know where you got the "40%" assertion (or how you can dispute even that). If the Hindi Wikipedia contained 60% acceptable articles, this wouldn't be an issue. The figure appears closer to 10%.
3. This was not a unilateral action on my part; it was the application of a consensus-backed standard that has been equally applied to numerous Wikipedias. Your reversal was unilateral and selective; you obviously are not impartial. Please participate in the discussion and seek consensus for the Hindi Wikipedia's inclusion. Thank you. —David Levy 17:25, 2 July 2009 (UTC)
Agreed. I've just conducted the 50-article test and saw three articles, at most, that appeared to be of substance and weren't just stubs or nearly-empty skeletons. -CapitalQ (talk) 05:15, 3 July 2009 (UTC)
Follow-up: A glance over the site's Speical:LongPages reveals stubs, lists and skeletal articles appearing as early as the 2,500 mark; concrete evidence that the count of substantial articles is under 30,000, if not under 2,000 (as plenty of the first 2,000 are partially or fully in English[1][2][3], or consist solely of tables[4][5][6] among other things). -CapitalQ (talk) 07:57, 3 July 2009 (UTC)
CapitalQ, your links are helpful. Agreed, if the main page's intention is to show 20K useful self-language pages which should not be stubs. Can you ensure listed tlwp has 20K useful self-language, non-stub articles, which has about 22K articles in total? Thanks, Shyam (T/C) 14:12, 3 July 2009 (UTC)

Bengali/Bangla Wikipedia (bn)

Bangla Wikipedia (bn.wp) has reached to 20,000 articles and the depth of the projects is 61[7]. As I know the condition to add the project at this template. Hope bn.wp will get the space. Thank you.--Bellayet (talk) 17:05, 29 June 2009 (UTC)

I tried hitting a random article link for 50 times. about 95% of the articles displayed are stubs/bot created articles. I think this is not an eligible candidate here. —రవిచంద్ర (talk) 06:07, 30 June 2009 (UTC)
With due respect, your statement about bot generated articles is either a miscalculation on your part, or simply false. Having been deeply involved with bn-WP, I can give you exact stats. About 2000 of the year articles had been created initially using a script, but many of them, especially the recent ones, have been later filled out manually (with birth/date lists, as in enWP). Next, we have around 4000 or so Indian cities article. And 366 date articles (many also filled manually later). Apart from that, the rest of the articles are all created manually. I might partially agree with your comment that some of the rest of the articles are stubs, but other than the articles I mentioned, none of the articles have been created by bots. There is only 1 article bot running on bn WP (bn:User:RagibBot), and that hasn't created an article for a long long time (AFAIK, none in last 2 years, though I'll have to check. Definitely not in last 1 years, in this timespan, this bot only created redirects/iw/etc.).
As for listing in main page, I think we can wait a little bit as we reduce the stub ratio. (Do you have any stats on the stub ratio?). --Ragib (talk) 05:03, 1 July 2009 (UTC)
PS - 95% of 50 random hits is 47.5, say, it's 46. Are you REALLY claiming you saw only 4 non-stub, non-date, non-year article? I am sorry to say that I do not believe this. And also I don't believe that you saw only 4 pages with substantial text ... that's impossible given the distribution of page sizes. No hard feelings, but please be accurate when making such claims. I'll assume good faith now, and investigate the stub ratio and improve them. Thank you. --Ragib (talk) 05:42, 1 July 2009 (UTC)
I have performed only one random test, and the above said results are absolutely true. Let the other admins also run the same test, and we will see the results. —రవిచంద్ర (talk) 05:44, 2 July 2009 (UTC)
For your satisfaction I have run this test again. It resulted in one big article, and four moderate articles (average 7 lines). so there are only 5 reasonable articles. For your information many of the articles are filled with Sections, subsections and infoboxes. Becuase of that some articles look long. But the actual information is very little. —రవిచంద్ర (talk) 06:09, 2 July 2009 (UTC)
I just viewed 50 random articles, and literally all of them were stubs or placeholders. The Siliguri article looked the best of the 50 that I viewed, and this was one of very few that fit the generous definition of "reasonable articles" used by రవిచంద్ర above. —David Levy 06:53, 2 July 2009 (UTC)
Well, I won't argue that the articles are not stubs. I think this request is a few months premature ... let us work on this a bit and re-apply in a few months. However, I hope by now it is clear to User:రవిచంద్ర that only the few articles I mentioned are bot generated ones, not the vast majority of user generated articles. Allow us to improve the article content and we'll reapply in a few months. --Ragib (talk) 05:27, 3 July 2009 (UTC)

Tagalog wikipedia

I noticed that Tagalog wiki (tl) has been included in the template. I just did a random load of 50 articles from it, and found only 6 non-stub or non-bot generated pages. All others were either one liners/2 liners, or tiny stubs consisting of navigation templates. Since the criteria of inclusion into this page is NOT the number of articles, rather the number of articles with reasonable amount of text, I believe Tagalog wikipedia fails the test, and hence should be removed. It will be great if whoever maintains the template can look into it and verify what I saw. Thanks. --Ragib (talk) 21:40, 3 July 2009 (UTC)

I just conducted the same test and also saw only six non-stubs. Accordingly, I've removed the Tagalog Wikipedia from the list. —David Levy 22:54, 3 July 2009 (UTC)
This test with viewing the random articles is stupid and does not make any sense. I saw the first 50 articles on Wikipedia in English and I saw 35 stubs, 2 laceholders and among others there have been no featured articles or articles that satisfy the criterias for high qauality. Also, it's obviously that Wikipedia in English has the same percentage, maybe even greater of stubs or placeholders than the others.--Kiril Simeonovski (talk) 00:57, 6 July 2009 (UTC)
It isn't the percentage of quality articles that's directly relevant; it's the quantity of quality articles that we care about (because the idea is to refer readers to Wikipedias at which they're likely to find decent articles about major subjects). The percentage is gauged purely for the purpose of combining this information with the total quantity of articles to estimate the number of quality articles.
If your findings are statistically representative, the English Wikipedia contains roughly 760,000 articles that are neither stubs nor placeholders. —David Levy 01:13/01:18, 6 July 2009 (UTC)

Removal of 20,000 tier

I strongly oppose the removal of this tier, as was done here, for two reasons:

1. It was not discussed in advance, as requested at the top of this page, and

2. By regularly "moving the goal-posts" of what makes a good Wiki, it removes any incentive for smaller wikis to grow and improve. For example, Welsh Wicipedia has a depth of 22 (well above the threshold of 5), and made a strong effort to reach the 20,000 limit a few months ago (it is now up to nearly 23,500). This effort now seems in vain, and there seems little point pushing for 40,000 if that is likely to change again when it gets there.

I suggest rather than setting a numerical limit, each wiki is judged individually. What do others think? —  Tivedshambo  (t/c) 09:18, 6 July 2009 (UTC)

1. Please see my explanation here. The change reflected a great deal of past discussion (both of periodically adjusting the numerical threshold to limit the section's size and of applying qualitative inclusion criteria). Note that the banner appearing at the top of the page was added in response to a user's continual removal of a tier (which I reverted) purely because it made the section seem too large under his personal display settings.
I actually did evaluate each and every one of the Wikipedias removed (in the usual manner), and most weren't even borderline. I was extremely disappointed, to say the least. (In the past, I've argued very strongly for the inclusion of more Wikipedias, and I successfully proposed that Wikipedia languages be swapped with Wikipedia's sister projects to eliminate the concern that these additional links would push down the latter section.)
2. The problem is that we've been providing more incentive to "grow" than to "improve" (to these Wikipedias' detriment, I'm afraid), as the "depth" calculation is far too easily increased via the use of bots/scripts. (I'm not suggesting that this is done with the deliberate intent of manipulating the data, but it's a side effect of the automated/semi-automated edits by multiple bots/users.)
If editors were to invest as much effort in increasing existing articles' quality as they do in increasing their projects' size, this would result in Wikipedias of tremendously greater value to these languages' readers. That (and not reaching the English Wikipedia's main page) should be their primary goal.
3. While judging Wikipedias individually has clearly emerged as necessary, we must also go by some numerical threshold (per the issue discussed above). Just as we don't want to list a Wikipedia with 20,000 articles composed 90% of stubs/placeholders, we don't want to list one with 100% good articles totaling 2,000. —David Levy 15:30/15:38/15:44, 6 July 2009 (UTC)
David, as per your criteria, how can you make sure mswp qualifies to be listed while euwp not. Just a difference of 2.5K articles is not enough as per your criteria. It's not a good parameter to judge if a particular should be included or not. If you can find a better parameter to judge, then would be appreciated. Else, please revert your changes. Thanks, Shyam (T/C) 20:59, 6 July 2009 (UTC)
The list always has been based on Wikipedias' size (with a non-static, arbitrary cut-off point). Recently, we introduced qualitative criteria that are applied in addition to (not instead of) the quantitative criterion. It then came to my attention that most of the Wikipedias in the lowest tier failed to meet these criteria (and therefore did not qualify for inclusion). I eliminated that tier in its entirety because it could not contain enough items to remain practical.
In other words, no one is saying that there's an enormous difference between 42,355 articles and 39,665 articles. The same is true of the Bengali Wikipedia's 20,004 articles and the Javanese Wikipedia's 19,801 articles, but the line must be drawn somewhere. A 40,000-article cut-off point is no less valid than a 20,000-article cut-off point is (and as I explained, it was necessitated by the lack of qualifying Wikipedias in that numerical range).
It never has been my desire to reduce the number of Wikipedias listed, but the current quantity is approximately the same as it was throughout much of the past (with the numerical threshold adjusted for inflation). —David Levy 21:26, 6 July 2009 (UTC)
Nice to hear from you, David. Could you please redirect me to the discussion happened about this quality check criteria. I believe this has come through the community consensus. Thanks, Shyam (T/C) 14:37, 7 July 2009 (UTC)
I've reverted the change until consensus has been reached. —  Tivedshambo  (t/c) 14:42, 7 July 2009 (UTC)
Consensus for what? The idea of applying qualitative criteria (because empty pages are useless to readers) or the idea of removing the lowest tier? —David Levy 19:10, 7 July 2009 (UTC)
There are several discussions in /Archive 3 and right here on this page. —David Levy 19:10, 7 July 2009 (UTC)

I support David Levy's view above. I just checked out the sq WP, and out of the 50 random articles I checked, only 4 were non-stub / non-bot generated / non-empty. That is, the percentage of real articles with at least a couple of paragraphs of text is less than 10%. This is just an example ... I'm sure there might be some WPs in the 20k category that will have better articles ... but seems like most of the current one's listed there won't pass the 50 article test. --Ragib (talk) 18:03, 7 July 2009 (UTC)

I'm quite happy to agree that there are some projects that don't meet a required standard. However there is no need to tar all projects <40000 with the same brush. —  Tivedshambo  (t/c) 21:15, 7 July 2009 (UTC)
The unfortunate reality is that there simply aren't enough qualifying projects in that range to sustain such a tier. From memory, perhaps two are borderline/debatable. Are you proposing that we retain the "More than 20,000 articles" tier for those two Wikipedias? (If I wanted to make a point, I would do exactly that.) Have we "tarred" Wikipedias below the arbitrary 20,000-article count by not including them? —David Levy 21:34/21:36, 7 July 2009 (UTC)
David, I don't agree with this quality check criteria. It does not signify anything. I checked it on two of the wikipedias. enwp has 16 (32%) stub articles out of 50 random while mswp has 26 (52%) stub articles out of 50 random articles. We expect these numbers to be higher for new and developing wikipedias. So, please don't go with this parameter while listing on the main page. Thanks, Shyam (T/C) 06:17, 8 July 2009 (UTC)
Did you read my reply to the similar post that you made above? Again, it isn't the percentage that's directly relevant. Our intention is to link to Wikipedias at which readers are likely to find quality articles on core topics, not those at which the "random article" feature will deliver quality articles with high frequency. —David Levy 06:25, 8 July 2009 (UTC)

Could someone articulate what the proposed quality criteria are? Dragons flight (talk) 06:31, 8 July 2009 (UTC)

We've been applying the criterion that a Wikipedia of that size not consist overwhelmingly of stubs/placeholders (because this means that the total number of quality articles is very low). In other words, a Wikipedia might contain 40,000 articles, but if 90% of them are [mostly bot-generated] stubs and placeholders, that Wikipedia is less useful than one with 10,000 articles of which 50% are stubs and placeholders (a Wikipedia with too few articles to even be considered for inclusion).
Perhaps unintentionally, many Wikipedias that otherwise wouldn't come close to meeting our minimum number of articles have been artificially increased in apparent size (MediaWiki mainspace page count) to the extent that they surpass it.
We haven't discussed setting a specific cutoff point, as most of these cases have been rather clear-cut and extreme. —David Levy 07:05/07:24, 8 July 2009 (UTC)
This criteria seems very subjective to me and then measurement would be completely based on randomization. Sometimes, stubs could be 60% or 80% for the same language wikipedia, say with 45K articles. Then how one can decide if this qualifies to be listed? Shyam (T/C) 08:40, 8 July 2009 (UTC)
Thus far, the criterion has been applied strictly to extreme cases in which Wikipedias are overwhelmingly composed of stubs and placeholders. But you raise a good question about how we should handle borderline cases. The answer, I believe, is simple.
We can set an arbitrary "minimum number of non-stub/placeholder articles" (perhaps 10,000). In fact, this could even replace the current arbitrary "minimum number of articles" criterion completely, potentially resulting in the inclusion of Wikipedias with fewer than 20,000 articles but relatively low percentages of stubs/placeholders.
It's been noted that our list serves as an incentive for smaller Wikipedias to increase in size (with many setting our numerical inclusion threshold as a specific goal), and I think that we would do a tremendous service to these Wikipedias and their readers by encouraging the creation of respectable articles instead of pages like this and this.
And incidentally, random samples of reasonable size (and we certainly can go much higher than 50 if we decide to analyze borderline cases in this manner) are statistically very accurate (and vastly preferable to blindly assigning the same value to an empty page as we do to a good article). —David Levy 10:03, 8 July 2009 (UTC)

Well I think the depth of a Wikipedia is pretty irrelevant and should be scrapped altogether as a measure of anything as it encourages bad behaviour. Just a look at two encyclopedias. The average article size of B is 50% larger than A, but the depth of A is 2.5 times larger. I did a spot check of random pages, and the content of B was substantially better (obviously not being able to tell POV or whether the material was rubbish) but it was a lot meatier. Some Wikis might go around subdividing their edits into small pieces to inflate stats or simply create empty talk pages for every stub; while en was almost 3 million articles and it is useful to break it into wikiprojects and tags, for some of these small ones dividing into WikiProjects would be useless but people could just do it to create bluelink talk pages to inflate stats. I don't think the depth does anything except to promote stunts by smaller wikipedias. If any raw machinecount should be used, at least average size*#articles = combined article size. Although that would be gameable by creating lots of pages with empty infoboxes with superfluous unfilled fields everywhere and using a script to take the readable prose size of a sample would be better. YellowMonkey (cricket calendar poll!) paid editing=POV 03:14, 9 July 2009 (UTC)

Yes, it's become rather obvious that the "depth" metric is highly unreliable. Whether this stems from deliberate exploitation or is merely a side effect of good-faith actions (such the use of multiple bots intended to improve a project) doesn't really matter.
This is why we've been checking Wikipedias manually, and the method that you suggest above is more or less the same as what I propose. And yes, skeleton framework must be excluded from any measurement. —David Levy 04:02, 9 July 2009 (UTC)

40,000 home page limit

[Copied from User talk:David Levy]

Hi David, I think that this limit should not be the only criteria for putting a link to a local wikipedia, because it only make the "stubs" problem deeper. Take a look for example the Macedonian Wikipedia (about 33,000 articles): [8] - hundreds of "mini stubs" with just one sentence and template appeared, just like this series [9](Aztec cities with just one sentence), [10] (Inca Empire articles with just one sentence), [11] (peaks of Macedonian with just one sentence) and so on. The admins even consider making a bot ([12]) to make stubs from a public available sources. The standard problem of the "year articles" is present as well ([13] - thousands of placeholders). Future raising of the limit will cause the same problem - again rush for more articles (mainly stubs & placeholders). This is only thoughts, you don't need to answer me, just to think about the future criteria :). Regards! --StanProg (talk) 08:24, 9 July 2009 (UTC)

I strongly agree. This is why qualitative criteria are so important. —David Levy 17:45, 9 July 2009 (UTC)

eu:wp on interwikies?

moved from T:MP

Hi! Would it be possible to put eu:wp on interwikies on the Main Page? We have recently reached 40.000 articles and we think it's not yet a small encyclopedia, so maybe it would be possible to do it. Thanks in advance. -Theklan (talk) 19:26, 22 July 2009 (UTC)

Depth is currently 22, moving this to Template talk:Wikipedialang Modest Genius talk 21:40, 22 July 2009 (UTC)
I just performed the 50-article test twice. Both times, six of the 50 random articles (a total of 12 of 100 random articles) were not stubs or placeholders.
However, unlike many Wikipedias, a majority of the articles appeared to be reasonable stubs (often with images), not worthless placeholders. (I encountered only a small number of empty "date" pages.)
This was one of the few borderline Wikipedias from the 20,000-article tier, and I'm inclined to think that it should be added. —David Levy 00:17, 23 July 2009 (UTC)
Why don't you perform the 50-article test on Wikipediaas, that currently exist in the language sidebar on the Main page (more than 40,000 articles, but high number of placeholders and stubs)? I performed the test, and the results were neither better than the test you've done at Wikipedia in Basque. I think that the quality criteria is so important, but according to your test and your criterias, beside of Wikipedia in Volapük, there are other Wikipedias that should be removed of the list. And, please, can you note, which are the requirements for any Wikipedia after reaching the number of 40,000 articles to be added in the list (for instance, requirements are:Depth:25, 8 non-stubs articles of 50 random, feautering an article every week etc.).--Kiril Simeonovski (talk) 00:40, 4 August 2009 (UTC)
We haven't established formal numerical criteria of that nature. Thus far, we've simply removed extreme examples (which wouldn't come close to meeting such criteria) from the lower tiers (Wikipedias with fewer total articles than those in the higher tiers). —David Levy 10:23, 4 August 2009 (UTC)

Query has been re-made at Wikipedia:Village pump (policy)#Basque Wikipedia. -- JHunterJ (talk) 13:27, 26 August 2009 (UTC)

And again at T:MP. Could an admin review this please? Modest Genius talk 18:37, 26 August 2009 (UTC)
Yes, I´m sorry. I have made Wikipedia:Village pump (policy)#Basque Wikipedia JHunterJ mentioned because I did not know that it´was done here by my fellow Theklan from the Basque Wikipedia.
If you read you shall know the reasons and arguments by that I do humbly request that the you add the Basque language in the Main Page of the English language Wikipedia. Thank you.
Goraintziak Euskal Wikipediatik.
Greetings from the Basque Wikipedia.
--Euskalduna (tell me) 15:05, 26 August 2009 (UTC) —Preceding unsigned comment added by (talk)
{{editprotected}} Since there doesn't seem to be any opposition to adding Basque, could an admin do so please? Needs [http://eu.wikipedia.org/wiki/ Euskara] adding to the 40,000 tier on this page and the [[eu:]] interwiki added to Template:MainPageInterwikis. Modest Genius talk 21:03, 26 August 2009 (UTC)
 Done — Martin (MSGJ · talk) 21:23, 26 August 2009 (UTC)
I just saw that you added the link of Basque Wikipedia in the language section of English Wikipedia. It was the best way to celebrate as we come to item 42,000 and give us encouragement to continue our work.
From Basque Wikipedia, sincerely, thank you very much.
--Euskalduna (tell me) 15:05, 26 August 2009 —Preceding unsigned comment added by (talk)
I also asked the same when we reachek 40.000 as some users asked about this in our Village Pump. We try to make the best with the quality, despite it's not always possible to have good up-to-date articles in small languages. We're very happy about your decission and I want to thank you all personally for the work you make. -Theklan (talk) 21:47, 26 August 2009 (UTC)


While we're backslapping over the 3,000,000th article on english wiki I think we need to spread the love. Most of the wiki languages listed under "More than 250,000 articles:" have now passed 500,000. The thresholds need a rethink. Using m:List of Wikipedias I'd suggest 500,000 150,000 and 100,000. Ideally this should be done today, while the spotlight of publicity about 3,000,000 is on the front page, if there's agreement. Bazj (talk) 14:57, 17 August 2009 (UTC)


{{editprotected}} Please sync with the sandbox. This new version improves the markup by using a single list, instead of the three it uses now. It also removes unnecessary nowraplinks classes and turns a switch into a more appropriate ifeq parser function. TIA. —Ms2ger (talk) 16:30, 8 September 2009 (UTC)

Reviewing this right now... isn't there a problem with using noinclude here? Or is that bug fixed? — RockMFR 22:41, 8 September 2009 (UTC)
The issue was that <noinclude> tags were ignored at Special:Statistics (where this template is transcluded). I don't know whether that remains the case. —David Levy 02:51, 9 September 2009 (UTC)
Ok, the noinclude bug was fixed in rev:29292. I'll make that change first. — RockMFR 02:56, 9 September 2009 (UTC)
I've made the full change now. — RockMFR 03:13, 9 September 2009 (UTC)

Vietnamese Wikipedia at 100k


The Vietnamese Wikipedia has reached 100,000 articles. I've updated the sandbox to reflect the change, but could a sysop synchronize the main template? Thanks. – Minh Nguyễn (talk, contribs) 09:55, 13 September 2009 (UTC)

Done Rambo's Revenge (talk) 13:33, 13 September 2009 (UTC)
I was fulfilling the request at the same time. I moved the Vietnamese Wikipedia to the correct alphabetical position. —David Levy 13:35, 13 September 2009 (UTC)

Numerical criterias

I've got a question to the users, who hold out this template. I made a comment to user David Levy about the criterias required for adding in the tier of the Main page. David told me, that any criterias aren't establihed yet. It was, approximately a month and a half befoe. Sionce then, I can't see comments aboutestablishing criterias. Many Wikipedias started their "new pages initiatives" and some of them reached the minimum required. Many Wikipedias which were erased of the tier are based on stubs and placeholders, have Depth bellow 20, the ratio of total articles, and total pages is over 0.8, and there are no more than 3 non-stubs articles in 50 random. I strongly agree that theese Wikipedias should not be added, after reaching the number of 40,000 articles. But, the fact that the link to Wikipedia in Basque was added few weeks after reaching this number criteria is enough to make me think, that the work on this template is irresponsible. In the group of Wikipedias, between 20,000, and 40,000 are projects, with Depth over 50 (Macedonian, and Latvian), the ratio is bellow 0.3 (again Macedonian, and Latvian), and there are 10 random non-stubs articles of 50 random. If it's looks, I suppose that you'll group theese exceptions in the group with others. Therefore, I have a proposal for the holders of this template, to establish other criterias, except minimum articles and minimum Depth required, as soon as possible. This will make the desicion easier. Regards.--Kiril Simeonovski (talk) 22:16, 17 September 2009 (UTC)

Limits 200 000, 100 000 and 50 000

I checked the article amounts for the moment and find that making the limits to 200 000, 100 000 and 50 000; would make more sense and a better distribution of the links per row. Also a link at the bottom of the page to with a table of all languages with the amounts would give better perspective of the Wikipedia. Maybe this table (columns: English name; Local name; Article amount; Founded) can be automated to generate the amount of articles. --Kslotte (talk) 12:29, 22 September 2009 (UTC)

Taking a look on this
  • over 200 000 (15 languages)
  • over 100 000 (13 languages)
  • over 50 000 (16 languages)
and at the bottom also a link to see other languages with less then 50 000 articles. I see this distribution quite even. Is there any chance to also include the English name like "Deutsch (German)". Opinions?
--Kslotte (talk) 13:01, 22 September 2009 (UTC)
Seems like the tooltip show the language in English. I think that is OK. --Kslotte (talk) 13:14, 22 September 2009 (UTC)
I see the "List_of_Wikipedias" on meta being to technical to be listed as "see other languages". Maybe a longer list made in the same way as the front page could be good. --Kslotte (talk) 15:01, 22 September 2009 (UTC)

I'll just point out that m:List of Wikipedias already is linked (as "Many other Wikipedias"). This, of course, has no bearing on the possibilities of linking to a simpler list and adjusting the numerical tiers. —David Levy 17:09, 22 September 2009 (UTC)

I'm thinking of keeping 250,000 and 100,000 but 50,000 makes a lot more sense than 40,000. With the latter the extra wikis spill over to the next line. That being said, the Hebrew and Serbian wikipedias are about to pass the 100K milestone and will have to be moved up making that one look a little more messy. Valley2city 18:51, 12 November 2009 (UTC)

Lang criteria only for India?

Do these criteria apply only to Indian wikipedias? Polish and Dutch are listed in the list although I found nearly half stubs and bot generated articles in 50 random articles. —Preceding unsigned comment added by (talk) 19:47, 10 October 2009 (UTC) Similar results with ukranian, slovak, catalan and norwegian wikipedia. —Preceding unsigned comment added by (talk) 19:58, 10 October 2009 (UTC)

Serbian Wikipedia now 100 000+

{{editprotected}} Serbian-language Wikipedia now has more than 100 000 articles. Please update the list, thanks :) --Dzordzm (talk) 23:22, 20 November 2009 (UTC)

Done. —David Levy 00:11, 21 November 2009 (UTC)

Restructuring proposal on Talk:Main Page

This discussion (and this entire talk page) is about the section Wikipedia languages on the Wikipedia Main Page.

A user at Talk:Main Page#Wikipedias by size update has pointed out that this template is out of date and suggested new number cutoffs. Editors who watch this page may be interested in that discussion. - BanyanTree 12:23, 18 December 2009 (UTC)

Please note that the user's belief that the template is outdated appears to have partially stemmed from the mistaken impression that it's intended to list all Wikipedias with a raw MediaWiki article count of 40,000 or above. —David Levy 14:00, 18 December 2009 (UTC)
The discussion linked above is now in Talk:Main Page/Archive 147#Wikipedias by size update. Also see Talk:Main Page#German and French wikipedia.
It seems several people are worried that the current wording "More than 250,000 articles:" makes it sound like the larger Wikipedias are not so large, so they want to raise the limit for the top group to 500,000 or so. But others think that will make the size of the groups unbalanced. So I instead suggest we change the wording for the top group to something like:
250,000 to ca 1,000,000 articles:
I am not a native English speaker, so I don't know if I am using the proper notation for "ca" (as in "about"). Anyway, writing it something like that should make it clearer how large the largest Wikipedias are. I updated my example over at Template:Wikipedialang/test1 so you guys can see how it looks.
--David Göthberg (talk) 02:22, 26 December 2009 (UTC)
To my knowledge, use of the term "ca." ("circa") in the English language usually is limited to chronology (e.g. "ca. 1900", meaning "about the year 1900").
But I like the idea, and I suggest using the labels "40,000 or more articles", "100,000 or more articles" and "250,000–1,000,000+ articles". I've posted an example at Template:Wikipedialang/test2. —David Levy 02:59, 26 December 2009 (UTC)
Ah, I checked my dictionaries, unfortunately you are right, "ca" is only used with dates. In Swedish we use it for anything that means "about". Ah well, then I suggest "250,000 to about 1,000,000 articles:" instead. And when the largest Wikipedia in that set goes above 1050,000 articles we can change it to "250,000 to about 1,100,000 articles:", and so on.
I feel using "250,000–1,000,000+ articles:" looks a bit strange, like there is a range, and then there might be some items way above that range too. And since the German Wikipedia now has 999,186 articles, does that mean we currently should use "250,000–900,000+ articles:" or perhaps "250,000–950,000+ articles:"? But still, your way is pretty okay and much better than the old way without a range.
For the two lower sets I am fine with both your new way and the old way.
--David Göthberg (talk) 03:51, 26 December 2009 (UTC)
To me, "1,000,000+" comes across as "1,000,000 or more", which is specific enough to approximately convey the high end of the range and non-specific enough to seem detached from any particular Wikipedia. Conversely, "about 1,000,000 articles" conveys the high end of the range, but it also implies that we're singling out the one particular non-English Wikipedia with the highest article count (currently the German Wikipedia) and custom-tailoring our format on its behalf.
If we were to implement such a change before the German Wikipedia reaches 1,000,000 articles, I would recommend the label "250,000–1,000,000 articles" (with the + symbol added when 1,000,000 articles is reached). I'm not sure that the figure should be updated in increments as low as 100,000 articles, again because this would be too specific to the German Wikipedia. (Note that we've drawn relatively few complaints by stating that it contains "more than 250,000 articles," and 1,000,000 seems to be a milestone beyond which the precise count is of far less concern.)
I relabeled the two lower tiers because I felt that it looked strange to begin only one of the three labels with a number. —David Levy 04:36, 26 December 2009 (UTC)
I don't see anything wrong with 'more than 250,000' tbh. Using circa is certainly inappropriate, and putting a + symbol afterwards looks unprofessional to me. Personally I think the status quo is fine, it is entirely accurate and the number will rise as more languages reach larger sizes. There's no need to mess around with upper limits just because one language is quite large. Modest Genius talk 13:45, 26 December 2009 (UTC)
For the record, I'm personally fine with the status quo, but I view this proposal (or some variant thereof) as a sensible means of addressing the criticism that several Wikipedias (not just the German one) are misleadingly (albeit accurately) labeled. —David Levy 13:54, 26 December 2009 (UTC)
This upper bound is not only to serve the German Wikipedia. We have eight Wikipedias over 500,000, and the French Wikipedia is also closing in on the million.
But right, I too am fairly okay with the status quo. But I see that people are complaining on these pages every now and then, and I too feel it would be pretty nice to brag a little about how large the other Wikipedias are. And we can add that "bragging" so easy. It means we are giving them credit for the hard work they are doing, and I think it would also be encouraging for the readers to see that those other Wikipedias are large.
And I have no doubt that the other Wikipedias will come here and remind us when it is time to update the upper bound, if we forget to do it. :)) And if we use increments of 100,000 then the updates won't be that often.
So I think it is merely a question about agreeing on how to best state the upper bound. We should probably wait some days and get some more opinions.
--David Göthberg (talk) 15:01, 26 December 2009 (UTC)

Given that there are 8 wikipedias over 500,000 now I think saying "Over 500,000 is more appropriate, don't you think? It is almost as if we are intentionally trying to belittle the other wikipedias to make english wikipedia look better.... Now if this isn't the case then there should be absolutely no reason why we can't change it to over 500,000... Dr. Blofeld White cat 12:11, 27 December 2009 (UTC)

I agree. There are in the meantime enough wikipedias passing 500.000, such that there won't be a big bias by changing this. Also, dewiki has passed the million today. --PaterMcFly talk contribs 12:27, 27 December 2009 (UTC)

Please add an "over 500.000" and an "over 1.000.000" section. Chaddy (talk) 16:36, 27 December 2009 (UTC)

The former would be reasonable, but the latter would include only one Wikipedia (an inefficient layout with consensus consistently against it). —David Levy 16:58, 27 December 2009 (UTC)

"More than 500,000 articles" tier added

Based upon the above discussion and other recent discussions, it appears that there is consensus for the addition of a "More than 500,000 articles" tier, so I've created one (thereby reverting to the four-tier layout used in the past).
For balance, I've adjusted the second tier's numerical threshold from 250,000 to 150,000. (Otherwise, it would contain only three Wikipedias.) The alternative (retaining the three-tier layout and shifting those three Wikipedias to the middle tier) would have resulted in substantial imbalance.
Of course, we should continue to discuss David Göthberg's proposal (and implement it if consensus is reached). —David Levy 17:52, 27 December 2009 (UTC)

Secure server

When accessing Wikipedia through the secure server the Main Page is kind of broken: The links in the sections "Wikipedia's sister projects" and "Wikipedia languages" always point to the normal insecure servers, even when viewing the Main Page through the secure server. This is bad since the normal way to get to another project is to go to the Main Page and use those links.

Those sections are rendered by the templates {{WikipediaSister}} and {{Wikipedialang}}. I intend to fix this by doing some changes to those templates. For the links in {{Wikipedialang}} I will use {{sec link auto}}. For the links in {{WikipediaSister}} I will use some other solution, since those links currently use full external links to make the URLs visible when printing the Main Page to paper. (I'll have to think a bit more about the best way to fix {{WikipediaSister}}. ) My changes will cause no visible change for users that use the normal servers.

I intend to do the change to {{Wikipedialang}} some days from now, and {{WikipediaSister}} at some later time.

--David Göthberg (talk) 03:34, 19 December 2009 (UTC)

I have coded and tested the new versions, see {{Wikipedialang/test1}} and {{WikipediaSister/test1}}. The WikipediaSister code probably needs some explanation:
I could not use {{sec link auto}} there for several reasons. Among other things since WikipediaSister is set to show the full URLs when printing and WikipediaSister also links the icons to the projects. So instead I hardcoded the secure server links. That meant the tables ended up inside a ParserFunction so as we usually do I changed from wikitable to HTML table markup instead, otherwise all the pipes "|" would have to be escaped with {{!}} which costs more server resources and makes the code harder to read.
To make the code less messy I didn't link the icons when in the secure server view. I think that most users on the secure server are experienced enough to not click the icons when wanting to go somewhere.
I am ready to deploy this. I'd appreciate if some other editors sanity check my code before I deploy it.
--David Göthberg (talk) 16:46, 20 December 2009 (UTC)
They look good, David. Thanks for doing this!
My only suggestion is that we use the "plainlinks" class for the secure versions as well (because the multiple padlock icons are visually overwhelming). I agree that users of the secure server can be assumed to be relatively knowledgeable, so the URLs themselves should serve as sufficient indication (and there's no harm if someone clicks through without noticing that the connection will remain secure). —David Levy 17:01, 20 December 2009 (UTC)
Right, the padlocks Lock icon.gif are a bit overwhelming. But I am planning to update the inter-Wikimedia links in several other system messages. Users of the secure server currently expects any inter-Wikimedia links to be insecure so I want to make it visible which links are secure and which are not during this period. My plan is that when we have updated most of the system messages and some weeks have passed we can remove the padlocks. Meanwhile, one thing we could do to make the padlocks less overwhelming is to change to a blue padlock Lock icon blue.gif in the same hue as the links themselves, thus also making the padlock look more like the external link icon External.png and the other link icons Icon External Link E-Mail.png News icon.png File icon.gif. So I have now suggested the blue padlock as the default for all https links here on the English Wikipedia. (I know how and where to add it in the .css files.)
--David Göthberg (talk) 16:02, 21 December 2009 (UTC)
I like the blue padlock icon and the plan to temporarily display it in these templates during the transition. You have my full support. —David Levy 18:25, 21 December 2009 (UTC)
YesY Done - The Main Page and Special:Statistics now have secure links for most of their inter-Wikimedia links, when using the secure server. (This template is also used on Special:Statistics.)
--David Göthberg (talk) 21:01, 23 December 2009 (UTC)

More than 1 million section needed

{{editprotected}} Now that the German Wikipedia has reached a million articles, could we have a line added to this template saying "More than 1 million: "? -- PhantomSteve/talk|contribs\ 20:43, 27 December 2009 (UTC)

 Done--Jac16888Talk 20:50, 27 December 2009 (UTC)
Phantomsteve: As noted in the template's text, please use {{editprotected}} to request edits that are uncontroversial or consensus-backed.
Jac16888: Please look to existing discussion (or initiate it if it doesn't exist) before fulfilling a unilateral edit request for a substantial change. As noted two sections above this one, there is longstanding consensus against creating a tier for few Wikipedias, let alone one. Only now that the number of Wikipedias with 500,000 or more articles has totaled eight have we reached consensus to create a tier for that threshold.
Also note that you introduced broken code by reusing the "lang-4" list ID. —David Levy 23:28, 27 December 2009 (UTC)
Ok sorry, no need to be so aggresive. It seemed like a harmless request. --Jac16888Talk 00:01, 28 December 2009 (UTC)
My message wasn't intended to come across as aggressive, and I apologize if it did. I merely wanted to explain why I reverted and politely request that you be a bit more cautious in the future. Sorry again if my tone seemed confrontational. —David Levy 00:21, 28 December 2009 (UTC)