Misplaced Pages

Revision as of 23:06, 24 May 2007

Template:Moveoptions

Korean:  Error: {{Lang}}: Latn text/non-Latn script subtag mismatch (help); Hanja:  Error: {{Lang}}: Latn text/non-Latn script subtag mismatch (help)

Japan NA‑class

This page is within the scope of WikiProject Japan, a collaborative effort to improve the coverage of Japan-related articles on Misplaced Pages. If you would like to participate, please visit the project page, where you can join the project, participate in relevant discussions, and see lists of open tasks. Current time in Japan: 02:25, December 25, 2024 (JST, Reiwa 6) (Refresh)JapanWikipedia:WikiProject JapanTemplate:WikiProject JapanJapan-related NA This page does not require a rating on Misplaced Pages's content assessment scale. WikiProject Japan to do list: edit history watch purge Peer review: None A-class review: None Featured content candidates –  Articles: None Pictures: None Lists: None Good article nominations: Vinland Saga (TV series), Godzilla Minus One Add requested images to articles that need them. Pages for Deletion: Participate in Japan-related deletion discussions. Improve and expand Japan-related stubs. Create some requested articles. Help translate an article from the Japanese Misplaced Pages into English. Assess unassessed articles

Archives

1 2 3 4 5 6 7 8 9

Requested Move May 2007

Survey to find out which name editors think is the best one under which to place this article. --Philip Baird Shearer 16:43, 21 May 2007 (UTC)

This survey lays out several proposals. Please only express # Support and an explanation followed by a signature for any of the proposals. Please place any comments like Oppose opinions in the Discussion section. You may change your opinions at any time up to the close of the survey. Changing you opinion and supporting more than one option in the Survey is encouraged as it helps to build a consensus. Please add any new proposals at the bottom of the list in chronological order. --Philip Baird Shearer 16:43, 21 May 2007 (UTC)

Can you note the rationale for your move request here? Consensus can change, but we don't need to have additional move requests unless someone believes it actually has changed. Stability has benefits, and this is a controversial topic. Dekimasuよ! 00:23, 22 May 2007 (UTC)

(A) There is no agreement over what to name this page. This is obvious from the talk page archives. (B) Any contriversial move can be requested at WP:RM, as it opens up the discussion to disinterested editors. (C) It is customary that if a move is requested at WP:RM, then another move is not requested for at least six months, so at the end of this process hopfully the discussions on the talk page over the naming issue will be kept to a minumum for at least the next six months.

I would suggest that this survey is widely advertised, as an RfC and on village pump to get as many disinterested editors as possible to express an opinion. As I said above changing ones opinion to try to reach a consensus is desirable, please don't just express an opinion and walk away, but come back after a day or so and see if the discussion has changed you opinion and also consider adding an opinions under other options to try to build a compromise and a lasting consensus. --Philip Baird Shearer 09:40, 22 May 2007 (UTC)

I was hoping that you would refer to a reason - guideline or policy - why you are proposing a move; I know the purpose of WP:RM. Anyway, I'll have to disagree with point C. I don't know of such a custom, and there's no evidence of it in practice. It still sounds like you've opened a discussion with the knowledge that it won't lead to a consensus decision, so I'm still confused as to the purpose of the debate. Dekimasuよ! 10:30, 22 May 2007 (UTC)

See Talk:Orange (colour) revision 18 October 2005 as an example. I can dig out more that I know about, but I would prefer that you take it on good faith. --Philip Baird Shearer 13:08, 22 May 2007 (UTC)

Survey

For previous surveys see:

1. Talk:Liancourt Rocks/Archive 3#Requested move Dokdo → Liancourt Rocks, result of the debate was move, 2 May 2005
2. Talk:Liancourt Rocks/Archive 4#Requested move Liancourt Rocks → Dokdo, result of the debate was move, 1 June 2006

--Philip Baird Shearer 15:57, 15 September 2006 (UTC)

Keep the article at Dokdo

Hot77 22:58, 24 May 2007 (UTC) support. Dokdo has been part of Korea's territory for centuries. It does not seem to make sense to change the name now.

1. Support local name used by entity in actual possession, arguably most common in English.17:53, 24 May 2007 (CDT)
2. Support local name used by entity in actual possession, arguably most common in English. (No strong preference between this option and move to Liancourt Rocks). --Reuben 16:46, 21 May 2007 (UTC)
3. Support b/c it has the most hits on Google web, Google book, the major newspapers, major websites (i.e. UN.org), etc., as shown in the archived discussion. (Wikimachine 18:03, 21 May 2007 (UTC))
4. Support same as above Janviermichelle 18:32, 21 May 2007 (UTC)
5. Support Per above. John Smith's 18:33, 21 May 2007 (UTC)
6. Support Per above. John Smith's 18:33, 21 May 2007 (UTC)
7. Support per above, especially because it's most common. — AKADriver ☎ 19:50, 21 May 2007 (UTC)
8. Support local name used by entity in actual possession, arguably most common in English. --Kingj123 21:33, 21 May 2007 (UTC)
9. Support I do not like the name "Liancourt Rocks" to be the title of this article because there are already established historical names while Liancourt Rocks do not have any cultural or historical meaning. BombAFT 21:59, 21 May 2007 (UTC)
10. Support. Actual possession and reasonably similar name usage provide adequate support to the status quo. -- Cjensen 23:26, 21 May 2007 (UTC)

    Support per most common naming. Sorry, didn't read first, I misunderstood above comment which is wrong, since it is equal used, I withdraw this vote. It appears UN, etc. supports a different name, but I'll read up the huge page first, so sorry. It seems per policy local use doesn't matter, and usage is the same. Dokdo and Takeshima is both officially local name, so we cannot use. Seoyoon 02:22, 22 May 2007 (UTC)

11. Support Yo peoples, I'm back after a refreshing 24 hour break from Misplaced Pages. Support using Dokdo as per above. Good friend100 03:12, 22 May 2007 (UTC)
12. Support Per King123. Endless Google arguments here and in archives just show Dokdo and Takeshima are widely used, Liancourt Rocks is not. Dokdo is the official locally used name to break the tie. Wikiment 05:21, 22 May 2007 (UTC)
13. Support Liancourt Rocks is outdated name as Pinnacle for Senkaku islands. Some encyclopedia may refer the islets to Liancourt rocks, but it is largely a legacy from old days (before WWII) without any meaningful contents with it. Therefore it is questionable that it is comtemporary and the most common name and these days it is actually being used less and less compared to Dokdo or even to Takeshima. Between Dokdo and Takeshima, Dokdo should be the title per many reasons above. Ginnre 07:23, 22 May 2007 (UTC)
14. Support Goggle hits shows more Dokdo than Liancourt and it seems that the Japanese is closing one eye on their claim---unlike the Koizumi tenure whereby this was an issue of controversy...Mr Tan 10:51, 22 May 2007 (UTC)
15. SupportLiancourt Rocks is outdated name as Pinnacle for Senkaku islands.John95 19:23, 22 May 2007 (UTC)
16. SupportLiancourt Rocks is outdated and does not account for the fact that the island is controlled by ROK. The name of the article has gone back and forth enough with in the last few years and it really needs to stop. The only reason this poll is up is because of the stink some editors have made about the article being under the name "Dokdo", mainly Japanese editors. It's about time this stopped. Davidpdx 00:52, 23 May 2007 (UTC)
17. SupportChanging its name to Liancourt Rocks seems unreasonable cause local name of this island has been kept using for a long time while Liancourt had claimed to found this island in 1849, The naming of any geographical reason should be relevant to its history and just buried long tracking histroy background using one French ships encounter with remoted island of republic of Korea User:Henry 01:06, 25 May 2007 (UTC)
18. SupportUsage of Liancourt Rocks is relatively scarce, and the island is currently under Korean administration and control whether or not Japan disputes it. Cydevil38 03:10, 23 May 2007 (UTC)
19. SupportClaims of dispute is POV. Islands are under South Korean sovergnty and until this changes, title of this article should reflect current status of islands.melonbarmonster 04:03, 23 May 2007 (UTC)
20. Supportper above. Oyo321 19:37, 23 May 2007 (UTC)
21. Support per Melonbarmonster's comment.--210.241.95.245 06:23, 24 May 2007 (UTC)
22. Supportper above. — Preceding unsigned comment added by Lions3639 (talk • contribs) 02:09, 24 May 2007 (UTC)
23. Support per Reuben's comment
24. Support Both Liancourt Rocks and Takeshima names are more recent names, keeping Korean name for Korea's occupied islet is more sensible.

Note: This is the above user's first edit on English Misplaced Pages.

1. Support per comments above
2. Support per comments above
3. Support per above, especially because it's most common.
4. Support b/c it has the most hits on Google web, Google book, the major newspapers, major websites (i.e. UN.org), etc., as shown in the archived discussion.Merongmerong 22:56, 25 May 2007 (UTC)

Move the article back to Liancourt Rocks

1. Support This is my first option as it complies with WP:NPOV --Philip Baird Shearer 16:43, 21 May 2007 (UTC)
2. Support Reasonably neutral name. (No strong preference between this option and Keep at Dokdo). --Reuben 16:47, 21 May 2007 (UTC)
3. Support Disputed territories should never be referred to by names used by any of the claimants --H27kim 17:36, 21 May 2007 (UTC)
4. Support If the slashed names (and their equivalents) are excluded, this is probably the most common English name.--Dwy 18:59, 21 May 2007 (UTC)
5. Support As shown recently, neither Dokdo nor Takeshima are more common in English pages. In additional, nearly all English speakers will encounter this item in the media where both will be mentioned, making siding with one name odd. Liancourt Rocks is a modern, accepted English name for this place used by all major encyclopedias and official naming conventions of at least the American government. I might also add that besides Dokdo not being more common, there is a sticky NPOV issue in newspapers where pundits are claiming, "In consideration of the popularity and influence of Misplaced Pages, the preference given to the word Dokdo demonstrates global Internet users are tilting toward us." here. --Cheers, Komdori 20:10, 21 May 2007 (UTC)
6. Support As per Misplaced Pages:Naming conflict and Misplaced Pages:Naming conventions (geographic names). Liancourt Rocks is the most common name among encyclopedias. Other external references suggested by recommended external references for identifying common names that supports Liancourt are "geographic name servers" (according to the NGIA GNS server, "Liancourt Rocks" is the BGN Standard and both Japanese and Korean names are variants) and "international organisations" (the United Nations Cartographic Section uses "Liancourt Rock" in List of Territories). --Kusunose 22:38, 21 May 2007 (UTC)
7. Support As per above. Parsecboy 23:46, 21 May 2007 (UTC)
8. Support as per above Masterhatch 01:38, 22 May 2007 (UTC)
9. Support More NPOV name. Hermeneus (user/talk) 02:39, 22 May 2007 (UTC)
10. Support pending resolution of the dispute between Japan and South Korea. This name sucks, but at least it's neutral, which makes it better than the alternatives. - Sekicho 03:16, 22 May 2007 (UTC)
11. Support This is the title used in English in Encyclopedias and trusted sources. This is the title as used in Colombia encyclopedia, Britannica refers to it as a name, and Encarta has a map. None of these three major references uses either Dokdo or Takeshima except to explain what the island is called in Korea and Japan respectively. It is also stated as the name in the CIA world factbook and the UN. When a name is actually used in media Liancourt Rocks is sometimes used and once again Takeshima or Dokdo as names in Japan and Korea simply reported but never actually used. In recent scholarly publications, Liancourt is actually the most common name used in English as the English name. Seokwoo Lee, from Incheon’s Inha University in South Korea and an authority on territorial disputes and this one in particular says "The two tiny rocky islets are called 'Tokdo' in Korean, 'Takeshima' in Japanese, and internationally recognised as 'Liancourt Rocks'." Dokdo and Takeshima are essentially never used in trusted sources as the name of the island in English. (Although Google hits gives Dokdo and Takeshima about equal hits and more than Liancourt Rocks this is misleading, because most of those hits are simple reporting what the island is called; or are on webpages set up specifically to promote one of the names). Macgruder 03:29, 22 May 2007 (UTC)
12. Support NPOV.--Watermint 08:50, 22 May 2007 (UTC)
13. Support sadly, per Kusunose and Macgruder Seoyoon 11:27, 22 May 2007 (UTC)
14. Support per WP:NPOV, elegance, and the collective judgment of earlier Wikipedians. WP was a much lower-profile website a few years ago, and it was possible to approach issues such as this in a much calmer and more sensible way. This article was originally created at Liancourt Rocks (in 2004) and remained there until being moved unilaterally last year. -- Visviva 12:33, 22 May 2007 (UTC)
15. Support NPOV, and per Macgruder's summary. Neier 14:26, 22 May 2007 (UTC)
16. Support per Kusunose, Macgruder, and Komdori's excellent Fixed-winged aircraft vs. Airplane and Aeroplane example (in the huge Archive 9). —LactoseTI 14:31, 22 May 2007 (UTC)
17. Support per Komdori, Kusunose, and Mcgruder. Not that anything we do here will keep people from complaining about it... ···日本穣 23:15, 22 May 2007 (UTC)
18. Support per WP:UE and, as an added bonus, it forces the Dokdo and Takeshima factions to fight it out by making article improvements instead of title changes. —  AjaxSmack  02:43, 23 May 2007 (UTC)
19. Support Whether we keep the article at Dokdo, or we move it to Takeshima, Dokdo/Takeshima or Takeshima/Dokdo, we can never reach agreement and the controversy will continue for ever. In order to avoid any POV, I believe we should choose Liancourt Rocks even if it is not populary used.--Michael Friedrich 05:07, 23 May 2007 (UTC)
20. Support NPOV --Isorhiza 16:48, 23 May 2007 (UTC)
21. Support It's the English name for the islands, and this is the English wikipedia. It's also the most neutral choice. Bnynms 02:53, 24 May 2007 (UTC)
22. Support Until Japan and Korea solve the issue, I prefer this neutral English name. Poisonotter 08:23, 24 May 2007 (UTC)
23. Support per Philip Baird Shearer and WP:NPOV. As shown below, in the "Google Book search" section, we were asked to believe that "Takeshima" has only 74 hits, out of which we can only use 12. However, if we manually inspect the results of a Google-Book search for Takeshima Island, we have much more than that. Because of this unfortunate result, and seeing how people have manipulated the Google results, I believe the search engine counts are totally useless, and have decided to completely discard them. Also, regarding jurisdiction: Jurisdiction has nothing to do with naming in English. For example, Japan has jurisdiction over Bonin Islands, but we don't use "Ogasawara Islands", the English name used in Japan. Macgruder has sufficiently shown "Liancourt Rocks" to be used in major English publications. And per Philip Baird Shearer, WP:NPOV must be adhered to.--Endroit 17:53, 24 May 2007 (UTC)
24. Support It is a very encyclopedic, popular, and neutral English name. Jjok 19:04, 24 May 2007 (UTC)

Dokdo is part of South Korea and North Korea and because it's their property, i think they are allowed to call their land 'Dokdo'. It is impossible to call Dokdo as 'dakeshima' or 'liancourt rocks'. If you are calling dokdo as 'dakeshima' or 'liancourt rocks', you are saying that 'dakeshima' is property for Japan or it's prorperty for the French.

Move to Takeshima

1. Support I do not like the name "Liancourt Rocks" to be the title of this article because there are already established historical names while Liancourt Rocks do not have any cultural or historical meaning. I also voted for Dokdo!BombAFT 21:59, 21 May 2007 (UTC)

Move to Takeshima/Dokdo

1. Support This is a second option as I don't like a slash in a name (means something in a URL) and which of the two names should come first--Philip Baird Shearer 16:43, 21 May 2007 (UTC)

Move to Dokdo/Takeshima

1. Support This is a second option as I don't like a slash in a name (means something in a URL) and which of the two names should come first--Philip Baird Shearer 16:43, 21 May 2007 (UTC)
2. Support By far the most common English name. I don't foresee much dispute as to which name should come first. Most Japanese will be happy to have the Korean name first on the basis of the current control, and this appears to be a rather established practice.--Dwy 19:01, 21 May 2007 (UTC)
3. Support English speakers will invariably hit this combo. I see no reason to violate alphabetical order. I would prefer to avoid the slash, but it's far better than siding with one side. --Cheers, Komdori 20:11, 21 May 2007 (UTC)
4. Support A second option. This is by far the most common name encountered (but usually as a reference) outside encyclopedias and scholarly publications. (the slash could be replaced by a | ). When an English user encounters one they usually will encounter the other especially in News, and sources that are not regarded as respected. Macgruder 03:54, 22 May 2007 (UTC)
5. Support This is my second choice since both names are concurrently used in almost all cases. Jjok 19:06, 24 May 2007 (UTC)

Discussion

I think enough time has passed since the last WP:RM (in 1 June 2006) for another to be held. So I am going to post this to WP:RM --Philip Baird Shearer 16:47, 21 May 2007 (UTC)

See Archive 9: Data on reliable sources for a survey of names in carried out on "Encyclopedias", "Newspapers", "Reliable websites" and "Google Books" about a week ago. Also another earlier survey Archive 9: Data from archive --Philip Baird Shearer 17:28, 21 May 2007 (UTC)

The "most common name" from the Misplaced Pages:Naming conventions guidlines does not trump Misplaced Pages:Neutral point of view policy. As Liancourt Rocks is a neutral name in the ongoing dispute over the ownership of Dokdo/Takeshima, I think that this article should be placed under that name. The guideline Misplaced Pages:Words to avoid makes a valid point under the section "Extremist, terrorist and freedom fighter" to paraphrase "Use of the name Dokdo or Takeshima implies a moral cliam to the island, if one party can successfully attach the label to the island, then it has indirectly persuaded others to adopt its viewpoint. --Philip Baird Shearer 16:47, 21 May 2007 (UTC)

The problem that I see with the article names of "Takeshima/Dokdo" or "Dokdo/Takeshima" is which name comes first. I can see that being as much a debate as whether it resides at Dokdo or Takeshima. Further because a "/" implies a URL hierarchy I think that names with a "/" in them should be depreciated. --Philip Baird Shearer 17:00, 21 May 2007 (UTC)

I request that certain discussions (i.e. the ones containing the data of the web results, etc.) be restored from archive. That way, people who are not familiar with this dispute may have more factors to consider in making votes. (Wikimachine 17:31, 21 May 2007 (UTC))

The trouble is that there are 100 of kilobytes of discussion over this. It is not at all clear why we need to pull up information from the archives instead of just providing a link to the informaiton like this: Archive 9: Data from archive --Philip Baird Shearer 19:14, 21 May 2007 (UTC)

What are the parameters of the survey? How long is it to run? What breakdown of opinion is required to consider a consensus to have been reached? Are there to be qualification requirements for valid participants, such as number of edits or length of time editing? The last incarnation of this survey elicited large numbers of sock puppets and single-purpose accounts, as well as outright vandalism, and considerable confusion about the criteria for closing and evaluating the survey. Who will have the discretion to remove disruptive edits from the survey area of the talk page? Although I already registered my opinions above, I suggest that this survey should be put on hold until these questions are decided. --Reuben 17:56, 21 May 2007 (UTC)

See WP:RM, but the normal length for a survey is 5 days, what is a consensus etc. As to the question of sockpupets etc, I have found when I have closed a contentious WP:RM, that at the end of a survey it is usually best to ask the participants if they think there are any irregularities and then to investigate those bought to my attention. However this is really a question for the administrator who closes this request. As I have opened it I will not close it. --Philip Baird Shearer 19:14, 21 May 2007 (UTC)

All legitimate voters from the last poll (~ 30 voters) have been notified of the new poll. (Wikimachine 19:06, 21 May 2007 (UTC))

This survey is not a vote. It is a survey of opinions to try to reach a Misplaced Pages:Consensus, the administrator who closes the poll will determin what the consensu (if any) is --Philip Baird Shearer 19:24, 21 May 2007 (UTC)

1. Comment: Let the readers decide for themselves after reading the archive. You cannot deny that Dokdo has the most hits in English pages with the data I've put up. You also agreed that encyclopedias tend to lag on disputed issues. And I explained that the article you've linked does not say as you claim. Repeat, it's only an observational statement. It's unfair that you quote only that sentence. Readers will know by reading the rest of the article that the author doesn't mean anything POV. (Wikimachine 20:32, 21 May 2007 (UTC))
2. Comment: Your data is flawed, with things like this listing all the transliterations of Dokdo and Takeshima once meaning you get 5 counts for Dokdo and 1 for Takeshima. I hope editors do read that article, it will show how the title of this article is interpretted in a POV way. --Cheers, Komdori 21:30, 21 May 2007 (UTC)

Fine there is a risk that my search result will be seriously erroneous due to multiple variation usage (which info, by the way, the site was specifically aiming to provide... unlike most other sites). But so what? Does that help Liancourt Rocks? Never. Even if I don't include the different variations, Liancourt Rocks is outnumbered by both Dokdo & Takeshima. The fact that 80% of Takeshima results are names completely justifies that minor rounding error. Every Google page (10 results), about 8 sites use Takeshima as a name for Japanese individuals. What, maybe about 1 site per Google page could be accounted for the multiple variation use. Fine, I'll do the searches again. (Wikimachine 23:11, 21 May 2007 (UTC))

In response to Wikimachine's comment:

It is not true that Dokdo is more common in Google. A simple test as required by Misplaced Pages shows the names are used equally. To do a simple test, you follow the guidelines that say to include some English words, and examine the first few pages for false hits. We need to check for variations, and to ensure Takeshima is not referring to names. We can achieve this easily by phrasing our single test carefully. The following test meets all those requirements and give equal results:

• takeshima island Japan Korea disputed -wikipedia 18200
• Dokdo OR Tokdo island Japan Korea disputed -wikipedia 16200

A careful look at the Takeshima results shows few if any uses outside this dispute. This is a tie.

Macgruder 04:14, 22 May 2007 (UTC)

Huh. Seems that you forgot about the archive. I already told you that WP naming convention does not limit google test to simple tests & that would lead to an especially strong bias. I also showed you this link on why simple tests shouldn't be used because they contain similar pages that have nothing to do with the terms. Also, what's so bad about excluding similar pages? There's nothing bad about it, but there's something good about it. Why not use it? And does it matter? Liancourt Rocks have the most votes currently, but even if I were to use your faulty search method, Liancourt Rocks has only 39,800 sites while Dokdo has 466,000. (Wikimachine 04:23, 22 May 2007 (UTC))

The archive is a flawed test as shown in this discussion below. Stop talking in generalities and find the flaw in my test. Yes, some of those pages might be referring to another Takeshima but it is very small and not enough to change the essential result. . Macgruder 04:31, 22 May 2007 (UTC)

That's a complete lie! As you can see, if you exclude similar pages, Dokdo comes out on top, but if you don't Takeshima comes out on top. All of you liars! I just can' believe people lying straight at my face like that. (Wikimachine 04:38, 22 May 2007 (UTC))

Once again, shouting at people and calling them liars when your 'method' of excluding similar pages is wrong is not productive. Google stops storing at around the 1000 page mark, so EVERY test with more than 1000 pages gives a similar result. Or are you telling me that of the 153,000,000 Smiths in Google only 930 of them count because of 'exclude pages'. This kind of reasoning is why I call you results flawed. I think when you make a mistake and call people liars it's time for you to consider your place in Misplaced Pages. Macgruder 05:04, 22 May 2007 (UTC)

"Change your attitude, and get a new suit & tie before coming back to this discussion", Macgruder. If this was the case, why not tell me? Fine. I'll base my statistical inference on your web searches & I'll prove to you that Takeshima is mostly used as names. (Wikimachine 05:39, 22 May 2007 (UTC))

We did. Twice. Both Komdori and I pointed this out to you before. Once again you're resorting to rude comments. It's you who is going around calling people liars. That proof needs to apply to the search that I have made btw because they have been framed in such a way already to exclude most Takeshima's as names. Macgruder 05:53, 22 May 2007 (UTC)

The burden is not on me to catch all errors. Macgruder, you failed to point this out for a long time. I already acknowledged that I didn't know that Google searches limited their results to 10000 sites maximum. However, I already moved onto your method of search, so it doesn't matter any more. Dokdo still contains larger count that Takeshima. (Wikimachine 16:28, 22 May 2007 (UTC))

The burden of proof is on you if you decide call people liars. We did tell you. Twice. And this was in the original discussion, not not for a long time. So you call people liars and the complain it's their fault when they point it out. The problem is you regularly call people liars. This is not a one-time case. Rather than trying to shift the blame, apologizing tends to be accepted method. Macgruder 19:13, 24 May 2007 (UTC)

Ha. And I did give evidence. Burden of proof means that you have obligation to provide evidence, and that's what I did. Now, burden of accuracy is not on me but you. (Wikimachine 19:56, 24 May 2007 (UTC))

The burden of proof to catch all errors is not only on you, but despite the fact that I have shown several times now that your results for Dokdo/alliterations are terribly flawed, you still present them as perfectly fine and usable. This is highly intellectually dishonest. This type of behavior makes clear that you are a POV pusher and do not care one whit about NPOV naming conventions. Parsecboy 16:40, 22 May 2007 (UTC)

Umm... I'm writing English, so read again. I already moved onto your method of search. Someone who accuses another of being POV pusher sure sounds like a hypocrite. (Wikimachine 16:45, 22 May 2007 (UTC))

The archive is not a flawed test. Spamming is part of Google test. Also, the new statistics I'm going to build will exclude spams, unlike yours. And spam applies to your searches as well. I see no net benefit. This is ridiculous.

• Dokdo: 747
• Takeshima: 648

That is your search without similar pages. (Wikimachine 04:36, 22 May 2007 (UTC))

No it's isn't. Sorry, but I think you are confused about how to use Google. Looking at the last page is meaningless for more than about 1000 results. Google doesn't store more than about a thousand pages. If you do the same search with Smith, the most common name in English, then you get the same result. It stops at around 900 even though there are 153,000,000 Smith results. This is an example of flawed reasoning. (Your 466,000 is also meaningless because it is not limited to English language as is required.). Macgruder 04:56, 22 May 2007 (UTC)

Actually, we just got finished showing this is not true unless you include Korean results. Dokdo is actually in the minority if you don't use the variations (there are variations because it's not an English name, it's a transliteration--besides, there is a standard English name, see my comments undre my vote). Of course, nearly all users will see the combo together, or see the established English name. --Cheers, Komdori 20:13, 21 May 2007 (UTC)

all right let's stay right here. Wait.

• First, define what you mean by Korean results.
• Second, Dokdo is minority only if you don't use variations. Its search size is far greater if you use the different variations. (i.e. tokto, dokto)

• Now... You haven't given reasons why not to use variations. I'll explain:

• Because Japanese words use Latin-based pronunciation (all letters have only 1 sound designated to them) while Korean words do not, in English, Dokdo happens to have many variations. They include Dokdo, Dokto, Tokdo, Tokto, Dok-do, Dok-to, Tok-to, Tok-do, Tok Islet, and Dok Islet... while the Japanese counterpart contains only one, Takeshima.
• Web search under Misplaced Pages policy seeks to measure the degree of usage of the term.
• I already explained in the archived page that authors will not switch between different variations. That means that there should not be double, triple, quadruple, etc. tags if I were to search for all variations. Any addition can be ignored as a rounding error.

• Third, I already proved that 80% of Takeshima is used as Japanese name, not for the name of the island.
• Fourth, people keep saying NPOV NPOV, but they're wrong. S. Korea has 100% administrative rights over the island. Japan may dispute the claim, but the mere act of dispute doesn't put Japan on equal level of ownership of Dokdo with S. Korea. If S. Kore was to make territorial dispute on Alaska with U.S., does having Alaska as the title become POV?

• This spills over to another WP policy.

The title: When a widely accepted English name, in a modern context, exists for a place, we should use it. This often will be a local name, or one of them; but not always. If the place does not exist anymore, or the article deals only with a place in a period when it held a different name, the widely accepted historical English name should be used. If neither of these apply, the modern official name, in articles dealing with the present, or the modern local historical name, in articles dealing with a specific period, should be used. All applicable names can be used in the titles of redirects.

• Since S. Korea is the official country controlling Dokdo, Korean local name should be used.
• Even if Liancourt Rocks is widely accepted name, Dokdo in itself has made much larger presence in the English language & can be considered as a English title as well.

• Fifth, Misplaced Pages concerns only with the commonality of the title in its naming convention.
• Sixth, final analysis: only candidate that matches Dokdo in commonality is Takeshima. Even if you were to not add the different variations for Dokdo, Dokdo still offers Liancourt Rocks no chance. I'm surprised that so many people are advocating Liancourt Rocks over Takeshima. Everything comes down to the commonality between Takeshima & Dokdo. Even if you say searching multiple variations may lead to buildup of same sites, remember that 80% of Takeshima is just names. Plus, I'm planning to do another data search with the method Komdori suggested (with all the search variations in "-"). That will clearly show that Dokdo trumps over the 2 other titles completely. And I already said in the archive that

• Misplaced Pages does not recommend double or triple names.
• Misplaced Pages does not recommend "/", or any other symbols that might suggest a hierarchy of articles.

(Wikimachine 20:55, 21 May 2007 (UTC))

We can use the variations if we agree it is not an English name but a transliteration of a Korean one (Tokyo is an English name, Toukyou is a transliteration). There is an established, modern, English name (Liancourt Rocks). Even if we were to ignore the English name for this place, and include the variations, it's still a statistical dead-heat. --Cheers, Komdori 21:06, 21 May 2007 (UTC)

Ah, nice double bind. Nope. An established English title can still have different transliterations. Plus, it doesn't matter really because we advocate Dokdo, the most common of all variations of the Korean name for the island, just as Misplaced Pages uses Tokyo instead of Toukyou. That there can be only one single established form of a name or title is non-argument. (Wikimachine 23:15, 21 May 2007 (UTC))

Those of you supporting Dokdo based on the Google hits should probably take a look at the results listed below. Wikimachine's method (the only google test that demonstrated a dominance of Dokdo over Takeshima) is terribly flawed. Most of the results from his/her google test are unrelated to Dokdo. Parsecboy 00:15, 22 May 2007 (UTC)

Still better than the previous tests, and in much better condition than Takeshima search results. Like I said below, I'm going to do manual count & statistical procedure so wait couple of days. (Wikimachine 02:47, 22 May 2007 (UTC))

If his edits are so flawed, they why don't you try a google test? Its not easy providing statistical data and it can get very tedious (as I have done it before). Don't simply reject his edit list simply because you don't agree with the results. Good friend100 03:15, 22 May 2007 (UTC)

Are you talking to me?

Philip, I'm really frustrated with the fact that you archived the previous discussions because all these arguments that users like Komdori & Macgruder are bringing up I completely answered & annihilated in the previous discussions. The editors participating in the poll seem to have neglected reading the previous discussions, and, as I predicted, have voted based on their personal dogmas & experience from the previous poll. (Wikimachine 03:39, 22 May 2007 (UTC))

For example, even if Liancourt Rocks is used by encyclopedias, most books use Dokdo. And I already explained that encyclopedias tend to lag on current issues & territory disputes.

Second, commonality of the title outweighs NPOV, per WP naming convention. People seem to have forgotten that.

Third, Google results show clearly that Dokdo is clearly more common than Liancourt Rocks & Takeshima. I just hate these people who constantly make obvious lies (as if to confuse outside readers & win support). I'd really love it if Komdori would explain how in the world Liancourt Rocks has higher number of hits on the search results.

Fourth, I also see people who make links to Misplaced Pages naming conventions, etc., but tell obvious lies about what those policies actually say (or make leeway interpretations).

(Wikimachine 03:47, 22 May 2007 (UTC))

Answers to opinions expressed in favour of Liancourt Rocks

• Philip Baird Shearer: Common usage of the title in English outweighs NPOV policy.

No it does not "common usage" comes from Misplaced Pages:Naming conventions (WP:NC) and it is only a guideline, Misplaced Pages:Neutral point of view (WP:NPOV) is a Misplaced Pages content policy and to quote from that policy "Neutral point of view is a fundamental Misplaced Pages principle. According to Jimmy Wales, NPOV is "absolute and non-negotiable."". Now that I have explained that to you Wikimachine, I hope that you reconsider the opinion you have expressed in the survey. --Philip Baird Shearer 09:50, 22 May 2007 (UTC)

I already know that the naming convention is derived from the NPOV policy. I was talking about NPOV within the context of how the Liancourt Rocks supporters were making the claim that Dokdo is disputed & therefore we must use a non-disputed title. However, the naming convention specifically states that common usage of the title in English outweighs all other NPOV considerations. (Wikimachine 16:34, 22 May 2007 (UTC))

Wikimachine please read Misplaced Pages:Policies and guidelines. Naming convention do not specifically state that common usage outweighs all other NPOV considerations. It can not because WP:NC is only a guidline while WP:NPOV is a policy. Common usage has to be weighed against WP:NPOV and if the more common name does not carry a neutral point of view when there is a very strong case to use the less common name. Further the policy WP:V means that when deciding on a name the reliablitly of the sources used should also be taken into consideration when making the decision. --Philip Baird Shearer 18:00, 23 May 2007 (UTC)

Huh, I thought Misplaced Pages naming conventions had it all. I'll see it later, thanks. Remember that reliable sources also include reliable websites & Google Books. In both Dokdo showed larger hits than Takeshima & Liancourt Rocks in the archive data. (Wikimachine 19:05, 23 May 2007 (UTC))

• H27kim: Misplaced Pages naming convention does not limit usage of names advocated by any of the claimants in the title of the article as long as it is the most comon title.
• Dwy: This is not the most common name. See archive & current data.

We are all rather familiar with Wikimachine's opinions so he does not have to try to refute the other side every time we say something. I only hope he will accept that people may have different opinions.--Dwy 06:46, 22 May 2007 (UTC)

Construction of "we" and the other, isn't it? Yes, we know how Dwy tries to avoid the fact that Liancourt Rocks is not the most common name. (Wikimachine 16:35, 22 May 2007 (UTC))

• Komdori: As shown recently, Liancourt Rocks can compete neither with Dokdo nor Takeshima, even if you exclude all transliteration variations. The fact that Liancourt Rocks is an accepted name is not a reason to choose the title when Dokdo is also an accepted name.
• Kusunose: encyclopedias are not the only places to search. Google Book search results favor Dokdo more than Liancourt Rocks. See Talk:Dokdo/Archive_9#New_Data_Using_New_Method. UN uses Dokdo more than any other terms.
• Macgruder: You should, of all people, know best that only encyclopedias refer to the island as Liancourt Rocks, but most of the academic sources refer to the island either as Takeshima or Dokdo, not Liancourt Rocks. See Talk:Dokdo/Archive_9#New_Data_Using_New_Method. Out of 10 major reliable news press, Liancourt Rocks receives 23 total while Dokdo receives more than 100.

Not so. My search of Google books for publications since 1980 threw up Liancourt Rocks being around 80% higher than Dokdo. On top of which many of these states clearly and unambiguously that Liancourt Rocks was the English name as opposed to just explaining the names in Korea and Japan. Macgruder 16:51, 22 May 2007 (UTC)

And why should we use results since 1980? Who defines these dates? Umm no. You failed to include all romanizations for Dokdo. Seems like Parsecboy needs to come her and accuse Macgruder of being a POVpusher. Something I don't like is how I am the one constructed as POV when I constantly make adaptations to your criticisms... but you guys still try to make lee ways out yourself. If you include all romanizations, Dokdo comes out on top. (Wikimachine 16:59, 22 May 2007 (UTC))

Well, I think 1980 defines recent rather better than 1800, or are you suggesting that a choice like 1980 is going to be bias against one rather than the other. The problem with your assertion that Dokdo is on top is that you don't take time to read what the books actually say. For example, the UN text that talk about Aquatic Sciences doesn't actually use the name at all but the name simply appears in a bibliography reference to a book by a Korean author. Then almost every time Dokdo (or whatever) is used there is the qualifier 'which is the Korean name'. This is not a use that can be counted as English because the author is simply reporting what the islets are called in Korea. Remove all these uses and such publications produced by say the "Korean Overseas Information Service" and Liancourt Rocks which is actually stated as the English name in a number of scholarly works very probably comes out on top if we are talking about actual English name usage. Misplaced Pages Naming policy warns about false positives, and this is particularly a problem here. For example, BBC never uses Dokdo or Takeshima as the name but simply reports what they are called in Korea and Japan. A simple count gives Dokdo or Takeshima about a 27 - 14 advantage over Liancourt Rocks. But if we read the BBC we see that NONE of these count. So in reality in the BBC Liancourt Rocks is a 14 - 0 'winner' over Dokdo or Takeshima. Same with most trusted resources. This obsession with counts (when comparing with Liancourt) doesn't mean much because of the overwhelming number of false positives. On top of which all you really end up promoting is Dokdo/Takeshima anyway.

As for the POVpusher thing, when you call people liars after they point out a mistake, and state that when your first count of Takeshima vs. Dokdo in media gave an advantage to Takeshima meant that the news was biased, but when your second gave an advantage to Dokdo it became I take my previous comment about the media bias back, you'll find that people may not come running to your defense. Macgruder 18:41, 22 May 2007 (UTC)

I have no idea what you're talking about. It'd be helpful if you specify in which discussion I said that. I'm also confident that even if my fellow agreeing editors are lazy & unresponsive, I'm skilled enough to take on all of you at once. (Wikimachine 19:31, 23 May 2007 (UTC))

Umm... those two events are not mutually exclusive.(Wikimachine 19:59, 24 May 2007 (UTC))

The media has a tie b/w Takeshima & Dokdo (I take my previous comment about the media bias back).. So far your 500,000 vs about 80,000 has been reduced to 21,000 vs 19000, so please continue taking us on. Macgruder 09:00, 24 May 2007 (UTC)

2,000 lead. You are the one who argued in 100,000s, not me. I argued in 100's. Remember our different search methods w/ different results? I think you're taking that too far when you're trying to make a hypothetical argument that I made as my main point. (Wikimachine 19:59, 24 May 2007 (UTC))

Wikimachine's Data based on Komdori's proposals

(Since it's using my name, I'd like to make notes of the suggestions I had made here until you follow what I suggested. The results below are not according to my proposal, I will remove this note if/when they are) Please include "-wikipedia", and the other fixes described in the lower section. Could you please redo all of these below with English pages please? Including the Korean and Japanese language pages is meaningless. Go to the preferences of Google and select "in English" from the dropdown. It would also be nice to see them with the variations in Takeshima rather than just the variations in Dokdo (getting pages like this) It might be nice to include some words to make sure its about the islands. I just randomly hit the "Dok to" on .net, and 3 of the 4 English pages are about things like the Ukranian Arlines destinations "DOK to IEV", "DOK to KBP", "redhat-dok" to the build", etc. Also note that your "similar pages" idea is flawed since it double counts all similar pages. --Cheers, Komdori 18:45, 22 May 2007 (UTC)

I will do this only on web search b/c Google Book is too small to be affected. Also see Talk:Dokdo/Archive_9#New_Data_Using_New_Method.

— Preceding unsigned comment added by Wikimachine (talk • contribs)

Interesting: you claim many of the hits for Takeshima are irrelevant because they're last names. How about hits for Tok-to that aren't relevant either. Hit the 40 button on the .com domain, and you get "tok to da hand", "doan tok to me about", dun tok to me abt", "AND FUCKING TOK TO THEM", and finally, "・・・ＴＯＫ－ＴＯ (TAKE-SIMA) とか". So we're talking 20% relevance in that one google page. So much for Dokdo and transliterations dominating other results, eh? Not to mention your objectivity in the matter, where it's apparently ok to make sure Takeshima hits actually apply to the issue at hand, but Dokdo + transliterations hits are accepted at face value. Can we say POV pusher? Parsecboy 00:08, 22 May 2007 (UTC)

Also, lets look at #10 from .edu domain. We get "Kohji Suzuki, Department of Engineering, University of Tokto, Tokyo", "Tahashi' Tokto", apparently a last name, (here's an interesting return) "Liancourt Rocks ("Tokto" in Korea, "Takeshima" in", " Tokto Ilimbezova", a first name, " Tokto, Japan", and only two other articles that actually discuss the disputed islands. Wikimachine, your method is horribly flawed, and the only thing we can take from your test is that you're a blind POV pusher. Parsecboy 00:13, 22 May 2007 (UTC)

Interpret as you wish, but I doubt what you say would apply to most of the searches. In my opinion, that type of spamming should apply to both searches - meaning that there is no net benefit of factoring such into the equation. Furthermore, I'm just doing as Komdori asked --> I'm open to more suggestions. If you want, I could manually count all for Takeshima, Dokdo, etc. because there aren't that much. But you'd have to wait maybe a week or two before the counting is finished. (Wikimachine 00:14, 22 May 2007 (UTC))

Ha, read the archive before you say that. I'll do something about it. (Wikimachine 00:15, 22 May 2007 (UTC))

What do you mean, interpret as I wish? A slice of your Google test results show that only 20% of the .com sites examined relate in any way to the disputed islands, and only 30% on the .edu page examined. And yet you claim only Takeshima is affected by false returns. What exactly am I interpreting besides cold hard facts? Perhaps you should explain your comment a little better. By the way, I never said the Takeshima returns are 100% perfect either; but it's intellectually dishonest of you to claim the Dokdo returns are 100% acceptable, while "80% of Takeshima is just names". Parsecboy 00:20, 22 May 2007 (UTC)

I'm just commenting on my cool stance with people who accuse me of being POV. It's just not my responsibility to catch all the errors. I noticed that Takeshima was used as names when I was searching newspaper sites, but I never saw Dokdo being used as names in those sites. So I got the impression that Takeshima was having a major problem while Dokdo wasn't. And I also think that spamming would apply to both Takeshima and Dokdo. I'm confident that I can find as many spams as you describe for Takeshima & Liancourt Rocks results as the amount for Dokdo variations. That is part of the Google search procedure. You make the assumption that all similar searches have the same technical problems, or Google search shouldn't even be used because there would be arguments over which is more accurate, which as more spams, etc. Spams are technical problems. However, people using Takeshima as names for individual is not a technical problem but one that has much larger implications (cultural) & necessitates appropriate response. (Wikimachine 00:35, 22 May 2007 (UTC))

I don't think the erroneous results returned from the Dokdo (or more accurately, from what I've seen, Tok-to) are any less a "technical problem" than Takeshima as a last name. I saw a couple Tokto's used as a last name or first name. I don't think it's properly classified as spam, as it doesn't appear to be an attempt to raise one side's name higher than the other. Most of the erroneous results I saw for Tokto in the .com domain were misspellings of "talk to", while many in .edu and .gov were sheer gibberish that happened to randomly have totko in it. I personally don't believe Google tests should ever be used, especially as a core portion for the justification for one side over another. Parsecboy 00:46, 22 May 2007 (UTC)

I know that Koreans don't use Dokdo as a last name. That's impossible b/c Koreans have last names of only 1 syllable (i.e. Bak --> Park in English, Oh, Yang, etc.) Like I said, I'll do manual count so it doesn't matter, does it? Just wait peacefully. (Wikimachine 01:29, 22 May 2007 (UTC))

When did I ever say it was a Korean last name? The last name I saw as Tokto looked more Central Asian to me, but I could be wrong. The point here is, the google search results are horribly flawed. Only about a quarter of the results I checked for Dokdo's returns were actually about Dokdo/Takeshima/Liancourt Rocks. And I'm sure the results for Takeshima are just as skewed as well. Those editors basing their support on maintaining Dokdo as the article's name are basing it on erroneous data. Parsecboy 15:56, 22 May 2007 (UTC)

However, if you want a suggestion, start with removing Tok-to from the search criteria. From the results I've looked at from .edu and .gov domains, it more often than not refers to Tokto, Japan, or the Tokto National Musuem, also in Japan. Tokto county in China may also skew results. Parsecboy 00:29, 22 May 2007 (UTC)

Maybe, I'll just count manually. (Wikimachine 00:36, 22 May 2007 (UTC))

I'll perform statistical procedure for large search results & manual count for small search results. I already did a statistical procedure at Talk:Heaven_Lake(Wikimachine 01:29, 22 May 2007 (UTC))

No need for all this complication. Here is the simple statistically unbiased test that makes it clear

• takeshima island Japan Korea disputed -wikipedia 18200
• Dokdo OR Tokdo island Japan Korea disputed -wikipedia 16200 Macgruder 04:23, 22 May 2007 (UTC)

But that is still not clear. It will still contain all the spams that Parsecboy complained about. It also contains similar pages. Back to here, what does Saint Thomas University have in common with MOFA? None. You know it. But why do you insist on using similar pages?

It also fails to exclude individual name usage for Takeshima, as my statistics will exclude. It also fails to bring in all variations of the transliteration Dokdo. (Wikimachine 04:30, 22 May 2007 (UTC))

Including Japan Korea disputed in the search terms also excludes pages that are about the rocks themselves, but not about the dispute over their ownership. This fairly well forces the results to be similar for the two terms, since documents about the territorial dispute will usually mention both names. For some examples of what gets excluded with this search: , , . --Reuben 04:40, 22 May 2007 (UTC)

That's non-argument. If a site contains the term Liancourt Rocks, it will show up regardless of whether the similar pages are excluded or not. Show me the Google pages that show this differnece. (Wikimachine 04:42, 22 May 2007 (UTC))

Is this comment addressed to me? I don't understand what you're referring to. --Reuben 18:17, 22 May 2007 (UTC)

Good point Reuben. You can try different search terms and they all pretty much give a similar result. Such as removing dispute. The point is that all such searches give similar results, and it is not clear which is in the majority. As I have pointed out a random check of name usage for Takeshima shows very few examples and would reduce the number by only a few percent. I have included the two most common transliterations Tokdo and Dokdo. A simple Google search shows that Tok do Tok-do and other variations are used very rarely (about 1%) in comparison. This 1 or 2 percent issues don't change the basic result. As Komdori and I have pointed out before looking at the last pages of a search with more than 1000 pages is meaningless as Google never stores more than that, and anyway looking at the last page has nothing to do with similar pages. Google always has that same message at the end of their searches. This is just Wikimachine's misunderstanding of that message and ways to exclude similar pages.

Reuben, you can also do the search for both Dokdo and Takeshima which only gives 11200 results. This shows that about 7000 pages use only Takeshima and about 6000 use only Dokdo. Once again this is a statistical tie. (I for one and not going to take a close result and say it means anything beyond a tie).

And yes it does effectively remove the name Takeshima. Because the words are chosen in such a way to make a Takeshima far more likely as the island than the name. You may find 1 or 2% Takeshimas as names but it doesn't change the statistical result. This is why Zogby and other polling professionals only choose a 1000 people to get a countrywide opinion. Because a cross-section gives a more accurate and sufficient statistical result (e.g. within 5% margin of error). More parameters simply introduce more errors. Macgruder 05:47, 22 May 2007 (UTC)

I'd like to also comment that the data above is still reliable overall (except for the .COM domain) because the searches do not exceed 1000. (Wikimachine 16:38, 22 May 2007 (UTC))

Please include "-wikipedia", and the other fixes described in the lower section. Could you please redo all of these below with English pages please? Including the Korean and Japanese language pages is meaningless. Go to the preferences of Google and select "in English" from the dropdown. It would also be nice to see them with the variations in Takeshima rather than just the variations in Dokdo (getting pages like this) It might be nice to include some words to make sure its about the islands. I just randomly hit the "Dok to" on .net, and 3 of the 4 English pages are about things like the Ukranian Arlines destinations "DOK to IEV", "DOK to KBP", "redhat-dok" to the build", etc. Also note that your "similar pages" idea is flawed since it double counts all similar pages. --Cheers, Komdori 18:51, 22 May 2007 (UTC)

Wikimachine's Statistical Inference based on Macgruder's web searches

Let me note that Dokdo results do not contain all transliterations (i.e. Tok Islet); however, I'm confident that Dokdo has much higher web usage than Takeshima even without the 8 other variations.

• randInt(1, 18100, 300)--> L1
• randInt(1, 16200, 300)--> L2

• SortA(L1)
• SortA(L2)

• Repeat searches with omitted results included.

• Takeshima

• Accept: 4, 122, 263, 277, 376, 392, 405, 487, 513, 521, 529, 560, 631, 910,

• Reject procedure. Google does not serve more than 1,000 results per query. Use results without similar pages to identify the percentage of the acceptable terms. This time, include all 10 transliterations for Dokdo.

• Takeshima (fixed): 19,500
• Dokdo: 21,500

• Since excluding similar pages show the most relevant pages, we can infer that the slightest inclusion of non-relevant searches (i.e. spams) makes a much larger presence in the larger web.

• randInt(1, 694, 50)--> L1
• randInt(1, 734, 50)--> L2

• SortA(L1)
• SortA(L2)

• Takeshima

• Accept: 681, 654, 590, 582, 576, 571, 551, 550, 542, 529, 525, 522, 501, 499, 493, 492, 476, 473, 473, 429, 341, 330, 303, 298, 291, 291, 229, 222, 218, 207, 207, 197, 184, 179, 179, 160, 157, 149, 141, 141, 56, 48, 45, 31, 31, 23, 18, 10, 7
• Reject: None.
• Accept: 7, 10, 11, 25, 55, 92, 93, 114, 115, 115, 126, 153, 179, 217, 224, 230, 256, 259, 275, 280, 296, 329, 330, 331, 332, 356, 361, 361, 365, 368, 376, 379, 399, 439, 507, 514, 517, 517, 522, 524, 529, 556, 564, 577, 579, 595,
• Reject: 640 , 642

• Dokdo

• Accept: 28, 47, 102, 115, 129, 147, 167, 172, 194, 223, 236, 281, 283, 288, 293, 299, 305, 311, 332, 347, 373, 387, 387, 388, 389, 422, 425, 432, 463, 481, 482, 502, 506, 513, 514, 525, 528, 533, 552, 556, 578, 588, 591, 591, 623, 637, 646, 649, 683, 734
• Reject: None.

• Assumptions:

• The samples are SRS's/completely randomized.
• n<.1N
• np>5, nq>5

• Umm... No rejected sites for Takeshima. This confirms that the current search is flawless. (Wikimachine 18:53, 22 May 2007 (UTC))
• 2-PropZTest: H0= p1=p2; Ha= p1=/=p2. z= -1.428571429. p-value: .1531275427.
• At the usual 95% confidence, the null hypothesis that there is no significant difference between the two proportion of acceptable terms for Takeshima and Dokdo. However, at 80% confidence, the null hypothesis may be rejected.

(06:31, 22 May 2007 (UTC))— Preceding unsigned comment added by Wikimachine (talk • contribs)

Unfortunately you don't seem to understand statistics very well. You have shown that in a single test of the web that Takeshima and Dokdo are within 10% of each other. That means for the whole of the web they are very close but we can't say exactly how. So all you have proved is the point I made in the first place. The two names are about equal in use. A 10% difference is not sufficient for your claim 'I'm confident that Dokdo has much higher web usage than Takeshima even without the 8 other variations. You have essentially proved our point. It's up to the admins to decide whether 21500 vs 19500 represents a substantial difference as you claim. Macgruder 19:38, 22 May 2007 (UTC)

Your null hypothesis is that they are equally common, and you have to accept it at the 95% confidence level, which basically supports the assertion that they are about equally common. But the assertion is that Dokdo is substantially more common as Misplaced Pages Rules require or as you say I'm confident that Dokdo has much higher web usage. Well, your own statistics have just rejected this assertion. A statistically test of say Dokdo being just 10% more common than would be totally rejected by the statistics (perhaps you'd like to add this one!). So the conclusion of the above test is from a statistical viewpoint is "The difference between the numbers is not statistically significant, and any hypothesis that one name is substantially more common than the other is totally rejected". (and this would apply with a very very very generous application of the word substantial to be 10%. If you took substantial to mean 30% or double then that hypothesis is totally totally rejected. (This is a pretty good statistical analysis; a small quibble is that Takeshima has variants like Take Islet, Takesima although these don't add much).

In short, the above statistics confirm that one is not used more than the other to any significant degree. Thanks. Macgruder 10:15, 22 May 2007 (UTC)

Statistics can be interpreted openly. Remember the disclaimer I put at the top? Without similar pages, only the searches that Google deems as most relevant will show up. That may explain why only two sites at the end of the search results happened to be rejectable ones. Therefore, any trace of irrelevant results must be weighed on heavily. 10% is not lenient. I see 20% as fitting in this case. (Wikimachine 16:22, 22 May 2007 (UTC))

Plus, even if we were to say equal percentage of sites per each search result are relevant sites, Dokdo counts outnumber Takeshima counts. Remember that most of the votes go for Liancourt Rocks. However, Liancourt Rocks has very little to offer. (Wikimachine 16:24, 22 May 2007 (UTC))

Well, essentially you are arguing for Dokdo/Takeshima because the difference you show is not significant. (Your 20% wouldn't even be statistically significant at the 50% level let alone the 95% level on this test, let alone cyberspace.). The very fact that you have to have ten different variations on Dokdo to make a slight difference tells us that Dokdo itself is hardly standard. Liancourt Rocks does have a lot to offer. It is used in all three encyclopedias per Misplaced Pages policy and is the only one that BBC, UN, London Times, CIA factbook, Korean legal scholar per above STATE to be a/the English name. If you remove all the Tokdo and Takeshima results from the test above that say "Tokdo is the name in Korea, and Takeshima is the name in Japan" then the number of hits for both will diminish significantly. Then remove the results of pages that unlike the BBC and Columbia encyclopedia were simply created for the purpose of promoting one name over the other then you would find that these numbers would be much lower. A statistically insignificant result requiring ten variations hardly to scrape a slight advantage doesn't come close to trumping respected encyclopedias and news sources.

Your argument against Liancourt Rocks only points to a second alternative of Dokdo/Takeshima. Macgruder 16:44, 22 May 2007 (UTC)

Wikimachine's whole procedure is flawed. This is a classic example of using samples with bias. For example, Wikimachine includes unusual Korean romanizations such as "Tok-to" while excluding similar Japanese romanizations such as "Take-shima". "Takejima" and "Take-jima" are excluded as well, while all sorts of Korean romanizations are included.--Endroit 16:50, 22 May 2007 (UTC)

Fine. I'll redo Takeshima with the romanizations you've suggested. Are there any other? (Wikimachine 16:59, 22 May 2007 (UTC))

I cannot help but complain that some of the users here are POV pushers who accuse me of being POV. No, my whole procedure is not flawed. Only a part of it. This is a classic example of using power tags. Like I said somewhere above, it seems almost as if these users are leading a desperate battle to win over outsider readers by using these power tags, such as "completely flawed!" "That's POV!" "Wikimachine's a POV pusher"!. Let's deal with situations more intelligently. (Wikimachine 17:02, 22 May 2007 (UTC))

When you compare multiple "Korean names" vs. multiple "Japanese names" you're opening yourself to criticism. Where in Misplaced Pages:Search engine test does it suggest it's OK to do that? It's inherently flawed (Wikimachine's Google tests).--Endroit 17:15, 22 May 2007 (UTC)

I'll change what I'm saying. It's OK to use such Google tests if you explain in plain English (where everybody can see), that your test results include "Dokdo", "Tokdo", "Tokto", etc. (in full detail)...but the Japanese test results include only "Takeshima". Also, you need to specify in plain English, how many irrelevant data you threw out from a random sample, based on your usage of a random number generator.... and explain this procedure in plain English. I still think you're opening yourself to criticism, but hey, the usefulness of your Google tests, like beauty, is in the eye of the beholder.--Endroit 18:12, 22 May 2007 (UTC)

I stated somewhere above that Japanese romanizations are based on Latin pronunciations. That shows you that I assumed that Takeshima was the only alternative. And yes I already clarified many times the need to include all variations for Dokdo. And I thought I made it specific by showing 2 sites that happened to be irrelevant. (Wikimachine 18:21, 22 May 2007 (UTC))

You need to say each time, it's "Dokdo, Tokdo, Tokto,..." etc. (give the full list), and hope everybody agrees with you to count like that. You only mention "Dokdo" most of the time, and that's very misleading.--Endroit 18:45, 22 May 2007 (UTC)

Adding "Take Island" as in here gives some more as well. As for similar pages, remember that if you ever do that your tolerance is +/-50% as the similar pages are counted twice for the non-similar's once. --Cheers, Komdori 18:59, 22 May 2007 (UTC)

In English, T is not D whatever way you cut it. Count the D's but not the T's. An article can mention alternative spellings but it doesn't mean they should be added up because they originate from the same language.

I'm also a little bit confused by your decision to use a null hypothesis to make your point that Dokdo is more common than Takeshima. You then get a statistically insignificant result thus in fact showing your assertion Dokdo has much higher web usage to be false (who was complaining about power tags just now). If you are doing such a statistical test to test your assertion then that should be the null hypothesis. If your null hypothesis is 'Dokdo has much higher web usage', then we can see right away that it has totally failed. If your null hypothesis is 'Dokdo has a 10% higher web usage' this too is going to fail and pretty badly. Even your 'Dokdo has a higher web usage' (and this would include the case where Dokdo has 1 more use than Takeshima) statistically fails, and this by your own analysis.

You continue to complain about people attacking your results but only a few days ago you were pushing the Dokdo having around 500,000 hits versus a much smaller amount for Takeshima. Now you are doing a variation of a test that I suggested days ago, and finding the similar counts that I also pointed out to you at that time. And suddenly from the 1000% or 2000% difference, suddenly it's down to 10-20% and that is only if we allow the Ds and the Ts which I dispute. Ironically, suddenly a few percent has become the new benchmark about what is 'much higher usage'. Well, if I discovered that my results had dropped from 1000% to 10% I might consider revision my opinion. Macgruder 19:28, 22 May 2007 (UTC)

If his results are so flawed, why don't you do the search yourself? I am curious as to what your response would be if Wikimachine's statistics showed that Takeshima was used more often. Good friend100 20:14, 22 May 2007 (UTC)

Funny you should say that. You don't need to be curious. Because this is my test, that I suggested with the addition of Tok variations. And do you know what? When it showed a 10-20% advantage to Takeshima, I called it a dead heat. What do you have to say to that? Macgruder 05:48, 23 May 2007 (UTC)

It is quite unfortunate both of you are just completely rejecting everything and telling him to do everything. What level of detail are you going down to? Good friend100 20:16, 22 May 2007 (UTC)

We are not rejecting everything. We are agreeing with the tie. As I did when Takeshima had a 10-20% advantage. Macgruder 05:48, 23 May 2007 (UTC)

It's funny how Parsecboy missed Macgruder as a POV pusher... He turns 10% advantage for Dokdo into 10~20% advantage for Takeshima by saying that only one of the two "Tok" and "Dok" variations should be used. Who gives you the right to specify that? You know you're just talking garbage. (Wikimachine 17:01, 23 May 2007 (UTC))

Wrong again. Right at the beginning I said usually the words are used together. When I had a 10/20% advantage for Takeshima what did I say? It's a draw. When you have a 10% advantage for Dokto, I say exactly the same thing. It's a draw. How is this POV? The search has the following 95% both, 3% Dokto alone, 2% Takeshima alone. Macgruder 07:08, 24 May 2007 (UTC)

Um... No, you just gave a POVish & faulty search (by not including all of the Dokdo variations), so your conscience hooked you with "tie". Also when all of my searches since the last poll favored Dokdo, you probably were going carefully & also feared that I'd point out that your search is flawed b/c it excludes some of the Dokdo variations. (Wikimachine 20:01, 24 May 2007 (UTC))

Macgruder isn't POV pushing, because he/she's not saying the 10% difference between Dokdo and Takeshima is any advantage between the two. It's a statistical tie. He/she is also not deliberately presenting data in a misleading manner to support his/her own position. Recall that Macgruder supports Liancourt Rocks, not Dokdo/Takeshima. Clearly not arguing in support of his/her own position. Parsecboy 19:35, 23 May 2007 (UTC)

Basic conclusion.

The two counts are within 10% of each other. (Didn't I say this about 3 days ago?) When it was in favour of Takeshima I called this a statistical dead head, now that it is in favour of Dokdo (plus the Tok variations which I dispute should be included), I'll draw the same conclusion.

Statistical dead heat.Macgruder 19:44, 22 May 2007 (UTC)

Goodfriend--Takeshima was ahead by the same margin before by some search methods. I and others said it was close enough to be tied. Everyone knows they are basically tied, but really just don't want to admit it because it would have implications for where this article wound up. --Cheers, Komdori 20:16, 22 May 2007 (UTC)

Don't call it a tie. 10% lead for Dokdo is significant. Everyone don't know that they are basically tied, and you guys are the ones who claimed that Takeshima has the lead, not in a tie, to begin with. Cheers! (Wikimachine 17:03, 23 May 2007 (UTC))

While the results had Takeshima slightly ahead (10% or so) we always claimed that was statistically a tie. I never advocated moving this to Takeshima during the discussion. Perhaps you misread something? --Cheers, Komdori 17:09, 23 May 2007 (UTC)

I haven't read all of the new edits to talk pages. I'll respond accordingly. (Wikimachine 19:07, 23 May 2007 (UTC))

Basicier Conclusion

All of you seem to have forgotten some stuffs here.

1. Just because I did the tests does not mean that any possible flaws should rest on me. I think I deserve a little praise for doing all the data work here.
2. Second, I reserve all right to define all possible interpretations for the statistics or any other data to test the neutrality of my fellow disputing editors.
3. Third, all of you seem to misunderstand how I ran the test. I ran the test to see which search term had higher proportion of acceptable terms. That way, we can see which search term, even if it has higher search outcome (or lower), may in fact have significant number of spams or irrelevant sites. In other words, the statistics was performed to test the accuracy of the search.
4. Fourth, I'm really confused why you guys would go after the test done last night, and not the latest test. It's almost as if you want to emphasize anything faulty in my test.
5. Fifth, Dokdo still has significantly higher number of usage (2,000 sites more) than Takeshima. I can run additional tests to exclude less relevant sites such as forums, government sites, etc.
6. Sixth, most voters went for Liancourt Rocks, not Takeshima. That means that it doesn't matter if the two results are close. Since most people advocate Dokdo & Liancourt Rocks, & Liancourt Rocks offers very little competition to Dokdo, I am performing statistics on Takeshima to make buffer zone in case voice opinions change.
7. Finally, let me add, Macgruder's "our point" was that Takeshima was in greater use than Dokdo, not in equal use. And to deliver the fact that "I am confident" that Dokdo has much ider use with some POV implication is being POV pusher b/c I had based my assumptions on previous tests & previous tests had shown that Dokdo is more commonly used than Takeshima.
8. Finallierly, Dokdo has higher usage in Google Books, Google Web, major news papers, and presumably reliable websites (drawn from Google search, therefore flaws must be assumed as part of Google procedure).

• Also, we cannot just select reliable web sites manually because that would result in voluntary bias by the site gatherers. Therefore, it is crucial that when we perform reliable site tests, we just run Google search & go straight down the list. (Wikimachine 20:42, 22 May 2007 (UTC))

You should also note that almost no votes went for "Dokdo/Takeshima". I don't know if this happened because everybody was paying attention to "Dokdo" and "Liancourt Rocks", but "Dokdo/Takeshima" should not be used because of several other points other editors made. Good friend100 20:39, 22 May 2007 (UTC)

Well, as for number two, you don't have a right to define all possible interpretations. We and other editors are welcome to draw our own conclusions, and no matter how much you would like Dokdo to be more often used, it's clear to anyone who looks at it objectively that it's not.

You are wondering why we pick on a specific test; the reason is because that one happened to be near the bottom. Nearly all of your tests are chock full of flaws, for the same old reasons.

As for number six on your list--try moving the article to Takeshima, then do the RFM. You'll see it the same kind of split between Takeshima and Liancourt Rocks, with few for Dokdo. The status quo tends to get (in this case very) unearned bias in general. --Cheers, Komdori 20:44, 22 May 2007 (UTC)

Again, why don't you make your own test? If his tests are chock full of flaws, why don't you make your own test and edit out all the flaws that you have pointed out in his tests? A number of flaws you describe are minute or you have gone into intense scrutiny and detail. No test will be perfect, especially on google. There is a margin of error, but that doesn't mean Wikimachine's tests are null. Good friend100 20:48, 22 May 2007 (UTC)

Again, for me, because there is nothing to prove. Even by his faulty tests, it still shows a statistical tie. --Cheers, Komdori 20:51, 22 May 2007 (UTC)

Fifth, Dokdo still has significantly higher number of usage. No it doesn't 10% in this one test is not statistically significant for the web as a whole, (and that statement doesn't mention the Tok variations being included which I dispute can be). The fact is you have come down from 1000% to 10% but refuse to draw the obvious conclusion. A tie.Macgruder 04:10, 23 May 2007 (UTC)

What? I conclude with 100% confidence, 10% lead for Dokdo. The statistics test was performed to show which results had considerable faultiness. None. So, the statistics proved that the tests were completley dependable (to practical degree of plain usage). (Wikimachine 19:08, 23 May 2007 (UTC))

No you still don't understand statistics. Your count is dependable. That is not disputed. 10% for this one test does not mean 10% for the web as a whole. You would need to do a number of test find the standard deviation of those tests and then test your null hypothesis about the whole web. All you showed was the counting was not flawed. But we don't need a test do we? The 10% is not significant even it is were true. As I stated when I found the opposite. Macgruder 03:57, 24 May 2007 (UTC)

Umm... No, 2 tests is sufficient when dealing with something as simple as Google search. Multiple tries fit better with medical & scientific data gathering. If you want, you can do them. I understand statistics well enough. (Wikimachine 20:03, 24 May 2007 (UTC))

The other tests are not relevant to this discussion. This test people agree is at least an accurate count of the search in question, but your other tests don't even meet that basic criteria.

Others will decide that for themselves. I thought I already said that some of the data above are still reliable. Do not falsely discredit everything. (Wikimachine 19:10, 23 May 2007 (UTC))

Good friend, we basically agree with his result. It's his conclusion that is faulty. He has kindly shown there are few false positives (as I said days ago but it's good to get confirmation). Unfortunately for him he has simply shown a draw. As for 'why don't you make your own test'. This was my own test. Right from the beginning I stated that the name Takeshima wouldn't show up and Wikimachine confirmed that. I got a small advantage for Takeshima and concluded a tie. Wikimachine got a small advantage for Dokdo (only after adding the Tok variations which I dispute). So I'm going to conclude the same thing, and so should anyone. It's a tie. Macgruder 04:10, 23 May 2007 (UTC)

Ah, did you do another test on your own, I'll check it out. A nice Memorial day present. (Wikimachine 19:11, 23 May 2007 (UTC))

Sixth, most voters went for Liancourt Rocks, not Takeshima. That means that it doesn't matter if the two results are close. Sorry, this is a circular argument. The issue is not just voting but, as is stated, a way, to get people arguments. If people in the pro-Dokdo section are arguing based on Google hits being much higher and they are shown not to be then effectively their votes are meaningless towards the overall result. Macgruder 05:13, 23 May 2007 (UTC)

I for one am not blaming Wikimachine for the flaws in his google test. Google tests are inherently flawed. I am, however, blaming him for his biased presentation of the results. He states that most of the Takeshima returns are last names. Fine. But then he turns and presents the Dokdo/alliterations results as 100% acceptable for use. This is clearly false, as the results I examined myself clearly show only about 25% relevance. It's not about the results themselves, it's how they're being presented that indicate bias. If anything, the supposed drubbing of Liancourt Rocks in the google test is more or less irrelevant, because Liancourt Rocks will 99% only really turn up relevant articles. I doubt there are many "Liancourt Rocks" last names, or other places with that name. Dokdo and Takeshima results are both skewed by those factors (and several others, I might add). Parsecboy 01:14, 23 May 2007 (UTC)

This is a classic example of everyone taking turn to attack me with "misunderstanding" that after I explain thoroughly, gets dismissed... but through the process, my image stained. I'm fighting a swarm of wasp, and I don't have a bug spray.

Parsecboy, thanks for taking advantage of confusion in chronology - I stated that most of the Takeshima returns are last names based on previous tests, which of course did contain names. However, on the new statistics, I confirmed that Macgruder's method of search is flawless. There's no skew, no 25% relevance, 100% acceptable whatever junk you're talking about. (audience claps) (Wikimachine 01:21, 23 May 2007 (UTC))

Thank-you for showing that my original assertion that there wouldn't be Takeshima names was true. But your misunderstanding of statistics is that even though this one test shows a slight advantage, it does not mean that it is true of all the web. This is the nature of Statistics. If I did a telephone survey where the counting was flawless, and it showed 210 people liked strawberries and 190 liked raspberries, the fact that the counting was flawless doesn't not mean that we could extrapolate to the population as a whole which one was more popular. We would conclude that they were about the same. This is called random variation. 210-190 would not be statistically significant. And to say strawberries are much more popular than raspberries would be absurd. Macgruder 08:18, 23 May 2007 (UTC)

You welcome. Get a new suit and tie & change your attitude. I think I understand statistics as much as you do, sir. You don't need to tell me anything like that. And I'm sure that everybody else in this talk page are smart enough to know what we are talking about also. Assuming such risks is part of statistics procedure. (Wikimachine 19:14, 23 May 2007 (UTC))

You clearly don't understand statistics. A fundamental statistics concept is variance. The more tests you do the less the variance. All tests make assumptions about how well the sample correlates to the the whole population. After you do many tests you can get an idea of what this variance is. For example, Zogby know that after doing many surveys that there is an approximately 2.5% standard deviation in their results. Therefore when they interview a thousand people they state something like within a 5% confidence range. No statistician who knows what he's talking about says: 'My result has no variance when applied to the whole population'. Gallup and other organisation whose business depends on the accuracy of their results have done 1000's of tests and they still state that their results have a 4 or 5% margin of error. This is one test. Your margin of error is going to be significantly higher. Now if you had a substantial advantage then we would probably say, 'OK, fair enough'. When we had Takeshima with 10/20% advantage before we said the same thing it's a tie. This is no different. Macgruder 16:44, 24 May 2007 (UTC)

Macgruder, get out, buy a new suit & tie, and change your attitude. You make these arrogant assumptions that you know something while the others don't. The whole point of doing statistics was to avoid manual count. Doing multiple statistics fails that purpose. (Wikimachine 20:05, 24 May 2007 (UTC))

Ah... Macgruder. You still don't understand how I ran the test. Google search already affirms that Takeshima is 2,000 behind Dokdo. 2-Proportions Z test only aims for which of the 2 results is more accurate than the other. Now... only benefit for you to argue on this topic is if Dokdo's Google results were so inaccurate in case of Dokdo that instead of 21,000, the relevant sites are 17,000--> Takeshima would have lead. However, 2 tests already proved Dokdo is 100% accurate while 1 test for Takeshima proved that Takeshiima might have inaccurate result at 0.2 alpha. It's common procedure in statistics to assume that there is no possible way for Dokdo to have some 3,000 irrelevant sites all of a sudden & Takeshima gets the lead... that's impossible. (Wikimachine 20:14, 24 May 2007 (UTC))

Puh-leeze. What, are you trying to win an Oscar for best performance? I took no advantage from this alleged "confusion in chronology". In your test results from the "komdori method", you argued that the takeshima results are badly skewed because of last names, etc. You made absolutely no mention of the Dokdo results being similarly skewed for other reasons, instead implying they were 100% acceptable. This is clearly not the case, from the slices of results I examined, which, as I demonstrated in the section above, were only about 25% relevant to the disputed islands. This is what I'm talking about. Thanks for playing the victim, it really helps the debate. Parsecboy 01:40, 23 May 2007 (UTC)

No, I'm not a movie actress. Look up what preterition is. You'll learn that in Latin III. It's a literary device. I wonder why you'd want to focus on the komdori method when the newest test is on macgruder method. Check the newest security update, or you'll get chronology virus. (Wikimachine 19:16, 23 May 2007 (UTC))

I was referring to it because you mentioned it in your initial "basicier conclusions" post. You said other users were attacking you for the flaws in your test. I replied that I wasn't attack you, I was attacking your presentation of the results of said test. Get your facts straight. Parsecboy 19:28, 23 May 2007 (UTC)

Hah, it certainly feels like a personal attack if all you talk about are flaws on his tests while praising Macgruder's tests. Good friend100 20:44, 24 May 2007 (UTC)

Here you go:

• Wikimachine's Dokdo test without Takeshima: 787

This basically proves most of Wikimachines data above (approx. 20,000) are dual usages of Dokdo/Takeshima.--Endroit 20:50, 22 May 2007 (UTC)

Here you go:

• Takeshima without all variations for Dokdo: 472

This basically proves that Endroit is playing with words & trying to confuse outside readers as to shift votes toward Takeshima/Liancourt Rocks. (Wikimachine 20:57, 22 May 2007 (UTC))

So? You haven't proven that Dokdo is used more than "Dokdo/Takeshima".--Endroit 21:01, 22 May 2007 (UTC)

"Dokdo/Takeshima" bests "Dokdo" by a factor of 25 to 1, according to your data, Wikimachine.--Endroit 21:06, 22 May 2007 (UTC)

Umm, I'm not even going there. That because Dokdo & Takeshima are used in pairs & therefore must not be counted as usage is....childplay. Misplaced Pages policy requires usage in itself, nothing more. I suggest that you look in the archive on this debate because I crushed it by placing 2-level double bind on Macgruder. This proves that Takeshima is extremely dependent on the Dokdo-Takeshima pair while Dokdo is less dependent on it. But, I really don't care about it. This is non-argument. (Wikimachine 01:26, 23 May 2007 (UTC))

As in your 500,000 Dokdo hits that we no longer talk about, or your multisearch analysis that other users on this page showed was completely flawed(this is the archive you keep insisting we refer to), or the test that I constructed days ago that I said days ago that would not throw up wrong Takeshima and you said it would, but now we see it doesn't. Or the result where I said the counts would be similar and after adding Tok(which I dispute) it still has a similar count as I said days ago. Is this the kind of 'crushing' we are talking about? The fact is we are essentially using my test and it is essentially throwing up the results I stated days ago (a tie). I suggest you (1) look up crush, (2)stop referring to past discussions which were in the middle of the now 'crushed' 500,000 Dokdo count that you were so keen on. Besides you are wrong about the pair argument. If Dokdo and Takeshima are almost always used in a pair it means that most sites are reporting foreign usage which therefore leads back to Liancourt Rocks. Macgruder 08:18, 23 May 2007 (UTC)

Umm.. Please don't use hyperbole & preterition. And I only claimed some 6,000 total hits across the web (.GOV, .EDU, .COM, etc.) using my previous method, not 500,000. However, You're the one who claimed some 500,000 for Takeshima.

If you can't look it up in the archive, I'll copy and paste for you. Further embellishment on this issue is below, so check the newer issue out. (Wikimachine 19:22, 23 May 2007 (UTC))

This is what you said: MacGruder, look at the archives. The archives contain a more accurate Google search data than what you've just posted. You ignore the fact that those numbers account for similar pages more than the individual sites. With no link. Well, in those archives in your discussion there was the number 500,000 and you didn't correct it. You yourself said the archives were more accurate. Ergo you are promoting 500,000. Then we have this quote: 'but even if I were to use your faulty(sic) method, Liancourt Rocks has only 39,800 sites while Dokdo has 466,000.'?Macgruder 17:36, 24 May 2007 (UTC)

Umm... archive is hands-off, and WP already assumes that discussion goes chronologically down the page. There's no need for me to correct anything. However, you can post "comment: this is indeed not true", etc. I don't understand why you're trying to argue with me on every front... even though that brings you no benefit for the entire debate. Is the number "500,000" that important to you? (Wikimachine 20:18, 24 May 2007 (UTC))

This is from the archive

Double bind 2: Either we must consider how the terms are used, & Dokdo & Takeshima have nearly 0% presence in the media while Liancourt Rocks has nearly 100%, and Liancourt Rocks & Takeshima suck in web results & Google Books, etc., and Dokdo dominates (b/c you'd have to count all the variations. Remember, we must consider on how the terms are used.); or we must not consider how the terms are used, & Dokdo & Takeshima have nearly 50-50% presence in the media while Liancourt Rocks has nearly 0%, & Dokdo shouldn't include its variations for web & book searches - still much greater usage than Liancourt Rocks, and a near tie with Takeshima.

(Wikimachine 19:22, 23 May 2007 (UTC))

Simple. Liancourt Rocks doesn't suck in Google Books when we consider how the words are used as is required in the first bind. Dokdo does not dominate in search results. 10% in one sample = dominate??? Unbound. (Unbinding was a lot easier than deciphering the English - do I get a smartie?). Also your statement requires the assumption of Dokdo dominating web results within a discussion about Dokdo's web results. This is known as a circular argument. The irony is delicious as is the smartie.Macgruder 17:49, 24 May 2007 (UTC)

Um... You can't unbind a double bind unless you lose the entire debate. It's either you gain access to half of your arguments while losing half of your arguments & therefore you lose.... or you gain the other half while losing the first half and you lose. Good job. (Wikimachine 20:21, 24 May 2007 (UTC))

Exactly. The fact that everytime I do a stat or data analysis, there are so many POV pushers who try to stain my user name that it's a must for me to test their neutrality by preemptively reserving buffer zone interpretations against their POVish counterinterpretations.

For example, I think Macgruder above said that adding "Take Island" adds significant amount of sites. That's (I'd love to say "lie") false. Look , and .

Umm... I don't know what you mean by your last sentence. RFM?

I understand statistics perfectly well. It is common procedure that you adapt your interpretation of the z test according to various observations you make. And I make the assumption that most of the irrelevant sites do not appear since the similar pages are excluded. That means that any hint that there are more rejectable sites for either candidate means a significantly higher percentage of rejectable sites in the rest of the search. (Wikimachine 20:52, 22 May 2007 (UTC))

Yes, but the flaw is drawing a small margin of difference in one test to make a conclusion about the web as a whole. Your test is not flawed. Your conclusion is flawed. Macgruder 05:05, 23 May 2007 (UTC)

My conclusion is not flawed. (Wikimachine 17:04, 23 May 2007 (UTC))

So you are saying that a single sample has a zero percent S.D. when compared with the population? Well, this is never true. And to quote what you said: Third, all Google searches have flaws no matter what. Macgruder 18:17, 24 May 2007 (UTC)

I didn't run the test to prove that Dokdo is more common than Takeshima. I only proved that both tests are very reliable & accurate.... and that your method does not include many irrelevant sites and wrong usage (i.e. individual names). That is my conclusion, and that is not flawed. It is statistically impossible for Dokdo to suddenly appear with 3,000 irrelevant sites & turn the lead to Takeshima. You should know that yourself. (Wikimachine 20:23, 24 May 2007 (UTC))

• Wikimachine: Macgruder above said that adding "Take Island" adds significant amount of sites is false/lie.
• Macgruder's actual words:a small quibble is that Takeshima has variants like Take Islet, Takesima although these don't add much

So Wikimachine what do you have to say about yourself here. You throw around liar again, and guess what. It's your mistake again. Stop complaining about other users if you have to resort to misquoting and straw to call them liars. Macgruder 03:13, 23 May 2007 (UTC)

Umm... no. That's not what I said. For example, I think Macgruder above said that adding "Take Island" adds significant amount of sites. That's (I'd love to say "lie") false. Look , and .

BUT what Macgruder actually said was: 'Takeshima has variants like Take Islet although these do NOT add much'. How is 'DO NOT ADD MUCH' = 'adds significant amount'. This you said was false, and you'd 'love to say a lie'. Macgruder 16:44, 24 May 2007 (UTC)

Ok. But I was talking about Komdori. Does it really matter who I was talking about? I think that once I clarify that I was talking about Komdori, you're out of my concern and so am I out of yours. ?!* (Wikimachine 20:25, 24 May 2007 (UTC))

In fact, it was Komdori who said Adding "Take Island" as in here gives some more as well. As for similar pages, remember that if you ever do that your tolerance is +/-50% as the similar pages are counted twice for the non-similar's once.

And I showed my Google search, which actually decreased number of total sites from 19500 to 19400 when adding Take variations. (Wikimachine 19:28, 23 May 2007 (UTC))

"Take island" added 100 to the search you had before. Try to assume good faith rather than calling people liars. Read Endroit's proof above in order to heighten your understanding of google searches. --Cheers, Komdori 20:57, 22 May 2007 (UTC)

Guess it still does add 100... and you're arguing about this when the difference is now ~1000. That's 10% of the tiny difference gone. Rather than preach how much you know about z and mu and statistics in general, try reading about them. That tiny of a difference is statistically insignificant. --Cheers, Komdori 20:58, 22 May 2007 (UTC)

Cheers! I better count backwards. 19400 - 19300 = 100! guess what. The search reduced by 100. (Wikimachine 20:59, 22 May 2007 (UTC))

I'm sorry, you aren't being clear... when I add "Take Island OR" to the search that doesn't have it, you get an extra 100 results... In any case, see Endroit's reasoning below. --Cheers, Komdori 22:13, 22 May 2007 (UTC)

Wikimachine's Google counts are misleading, because the usages of "Dokdo" (& variations) DON'T have the highest Google count, when compared to "Dokdo/Takeshima" combinations. When you put it into perspective, Wikimachine's Dokdo test without Takeshima gets only 787 Google hits. Therefore it appears that choosing "Dokdo" (over the other choices) based solely on Google counts seems ill-guided.

See Talk:Dokdo/Archive 9 for further discussions on other criteria, such as NPOV, etc. Since it's a very difficult choice, please be sure to look at all the criteria and past the Google counts.--Endroit 21:51, 22 May 2007 (UTC)

Ok. I'll explain this to all other users. This is a really stupid argument. Endroit, and Macgruder are arguing that most of Dokdo & Takeshima counts should not be considered because they are always used in pair with each other & so the authors are just mentioning them for outside reference, not actually giving favor to one of them.

But, do you really expect reliable sources to voluntarily not mention one or the other in its entire article? That's like requiring Misplaced Pages to either use Dokdo or Takeshima only. That's very unlikely. Also, Misplaced Pages policy requires only usage in itself, without the consideration of how or in what context the terms are used. (Wikimachine 01:30, 23 May 2007 (UTC))

Let me add, I offered a 2-level double bind to Macgruder on this topic, and he couldn't answer it. So, look in the archive. (Wikimachine 01:31, 23 May 2007 (UTC))

No, here is Misplaced Pages Policy:

• False positives
• Search engine tests should be used with care: in testing whether a name is widely accepted English usage, we are interested in hits which are in English, represent English usage, and mean the place in question. Search engine results can fail on all of these.
• "Search engines will find hits when a paper in English is quoting foreign text, which may well include foreign placenames." etc.

You sure do love the word double-bind. I think I was probably too busy telling you that your 500,000 google Dokdo hits was wrong. Macgruder 04:00, 23 May 2007 (UTC)

Um, let's analyze is quoting foreign text, which may well include foreign placenames. What does the word quoting mean? Basically copy-paste/translation of a foreign text (i.e. government site: "according to the Japanese government"...; news: "according to Joon-ang Ilbo newspaper"). Is there any quoting going on in any of these sites? (Well, most) No. These words are written by the authors themselves.

Don't make ridiculous claims. You can't possibly expect newspaper editors to only use one name and voluntarily exclude the other. It's like "know about their culture" thing. You tell the readers what the name of the place is in their own language. (Wikimachine 19:39, 23 May 2007 (UTC))

But you didn't read the next line did you?

• "Hits which are of the form "X " attest to English usage of X, and Korean usage of Y. The latter matters to the Korean Misplaced Pages, not to us." (replaced foolandish for clarity) Macgruder 18:45, 24 May 2007 (UTC)

Is this a Misplaced Pages guideline, I'd really love to know. (Wikimachine 20:27, 24 May 2007 (UTC))

A simple search test for "english sites only" shows that Dokdo has 121,000 hits vs. Dokdo/Takeshima, which has 12,100 / Good friend100 22:12, 22 May 2007 (UTC)

I also think these google tests are flawed. This is because they count statements like "known as Dokdo in Korea" or "the Japanese call it Takeshima". Those hits should not be counted because they just assert local usages and not English usage. --Kusunose 02:41, 23 May 2007 (UTC)

Umm... Ask admin such as Visviva. He'll definitely disagree. WP naming convention specifically aims for usage in itself. Then let me ask you this. Shouldn't the articles mention Liancourt Rocks along with the quotation of foreign text because it would be appropriate for many reasons that I'm not going to mention right now? (Wikimachine 19:42, 23 May 2007 (UTC))

It was you Wikimachine who created a section specifically to deal with my test. That test has a basic tie. But you keep on referring to the other tests when your conclusion is challenged. The basic fact remains, the one test that has had a reasonably amount of support as an effective count is a tie (withstanding the objections to using Google in the first place). I think your checking for one type of false positive was useful (although I disagree that you can use the 'Tok' variations but this doesn't change the conclusion anyway). It's you who is saying the result of this test is flawless, and therefore you need to draw the conclusions that this test shows. You cannot in this section refer to other tests that are not under discussion and as other users have pointed out in those sections are flawing.

Umm... no. I only confirmed that Dokdo has 2,000 site lead over Takeshima in the web. Other data such as Google Books, reliable sites, and the media data still confirm that Dokdo is used more than Takeshima. Other than that, I'm not confirming anything like that. (Wikimachine 19:50, 23 May 2007 (UTC))

So we go back to the conclusion of this test:

• Usage of Takeshima and Dokdo is basically a tie
• This test supports the use of the Dokdo/Takeshima or Takeshima/Dokdo variant

Points to remember

• Google hits don't necessarily measure actual English use
• Test will return results asserting to local usage which probably should not be counted in the context of a Liancourt Rocks comparison
• A single test is a snapshot of overall web usage so will have a margin of error. This confirms the statistical tie more than anything

Final Conclusion

If we accept that Google tests are an issue to be considered in the naming debate then we discover they do not lend support for either the Takeshima choice or the Dokdo choice, but to the slashed choice. Thus the choice can be reduced to Dokdo/Takeshima or Liancourt Rocks depending on the outcome of other arguments and the validity of such tests. Macgruder 04:00, 23 May 2007 (UTC)

So we go back to the conclusion of this test:

• Dokdo has 2,000 sites lead over Takeshima, and that's significant.
• This supports only Dokdo, not the "/" variants because Endroit's tests revealed that Takeshima is more dependent on the pair usage (which I don't agree in the first place, but hypothetically) than Dokdo.

Points to remember

• We have nothing to base our debate on if we can't trust Google. Assuming practical flaws is part of the Google procedure.
• As for the dual usage:

• WP naming convention seeks for usage in it itself, and does not consider how it is used.
• WP naming convention specifies an exception, false positive - but that's only when a foreign text is quoted (none of the sites that were used in the data quote foreign texts)
• You can't possibly expect websites and articles to voluntarily use one name over another. If the authors want, they can mention both names.
• The authors should mention Liancourt Rocks as the main name and then "quote foreign texts", but they don't. They use Dokdo & Takeshima as the primary names. Newspapers have around ~10 total usage for Liancourt Rocks while over 100 usage for Dokdo & Takeshima.

• There were two tests. The first test confirmed that Takeshima brings more irrelevant sites than Dokdo. (Wikimachine 19:50, 23 May 2007 (UTC))

It should also be noted that searching Takeshima on google gives you many unrelated "Takeshimas", including several names and a "Takeshima restaurant". Good friend100 20:58, 23 May 2007 (UTC)

This is already noted. It should also be noted that the alliterations for Dokdo with which Wikimachine is so in love create mostly unrelated hits, as I demonstrated in the section above, relating to Komdori's method. Wikimachine, the reason I am bringing this test up is because these alliterations were not used in Macgruder's test, so don't worry, I have no "chronology virus". Parsecboy 21:04, 23 May 2007 (UTC)

Then you should have no problem. (Wikimachine 20:30, 24 May 2007 (UTC))

Of course indiscriminant searchs for Takeshima might give some unrelated results, but look at Macgruder's results below--they are carefully crafted to avoid this kind of interference. --Cheers, Komdori 21:57, 23 May 2007 (UTC)

You are clearly a POV-pusher. Wikimachine's tests, you say are flawed and changed. Yet Macgruder's tests are accepted and they are the correct tests. It certainly isn't fair. Your attitude is so biased. Oh wow look at Macgruder's test, they are carefully crafted. But Wikimachine's tests are so flawed and full of crap.

Using an incredibly long search "island Japan Korea disputed "Take-shima" OR Takejima OR "Take-jima" OR Takeshima -wikipedia", you can certainly see how the text was manipulated to get the maximum google searches for "Takeshima". Do people searching the dispute type in all that?

Seeing this, I think the google searches are altogether pointless when you get down to that kind of level. It isn't accurate either. You can still see irrevelent links. Again, google searches can't be used like that to justify things. Good friend100 00:17, 24 May 2007 (UTC)

"Using an incredibly long search "island Japan Korea disputed "Take-shima" OR Takejima OR "Take-jima" OR Takeshima -wikipedia". What??? This was because of Wikimachine's insistence he wanted to add all variations of Tok Island, Tok Islet, Tokdo, Dokdo. This was his idea not ours. Of course, if you do for one you must do for the other. But if you wish you can just do Dokdo and Takeshima. Please go ahead.

You are wrong when you keep repeating "Oh wow look at Macgruder's test, they are carefully crafted. But Wikimachine's tests are so flawed and full of crap." We are saying exactly the opposite. Macgruder's test gave a 10/20% advantage to Takeshima and BEFORE we said basically said that means it is a tie. Wikimachine's result gives a 10% advantage to Dokdo, and we are saying the same thing "it's a tie". We are not saying his counting is flawed, we are saying his conclusion is wrong. We've been totally fair. In both cases we said 10-20% means a tie.

You make a very good point about all the variations. This is why we call it a tie. No test can be exact because there is a statistical problems with choices of words, names and counting methods (google vs. yahoo). Because of these unavoidable errors when a result is close it's called a tie. We said that with my test, and we are saying it with his test. The result of the test is a tie. Macgruder 05:44, 24 May 2007 (UTC)

And Macgruder that's exactly where you go wrong. I don't understand how a search with only 2 variations for Dokdo and a search with all variations for Takeshima & Dokdo make a tie. Simply, my search is more accurate because it adds more terms while yours don't. It's called biased sample (eh hem, somebody brought that up above.) (Wikimachine 20:34, 24 May 2007 (UTC))

Umm actually, what I'm really confused about is how when I make use of Macgruder's tests to perform statistics I get criticized but Macgruder doesn't.... when he's original designer of the test. I'm really corn-fused. (Wikimachine 20:32, 24 May 2007 (UTC))

Good Friend, I've been arguing that Google tests are too flawed to be used for something like this the whole time. We're all well aware that Takeshima and Dokdo return erroneous results, probably more than relevant results. This whole process here should be discarded. In doing so, let's establish some facts. The vast majority of media does not use Dokdo or Takeshima by themselves. Therefore, leaving the page at Dokdo or moving to Takeshima is unacceptable. Misplaced Pages guidelines advise against split names such as Dokdo/Takeshima, for problems with url, etc. Misplaced Pages naming rules also state that when an English name is present, it should be the name used. Therefore, Liancourt Rocks, while perhaps not as widely used as Dokdo/Takeshima, is the best choice. It is also the least POV, as the other two names are the local names used by the two disputing powers. This is not to say that the local names cannot be used in the article. By all means, they should be mentioned in the first sentence of the introduction. But the title of the article should be NPOV, and Liancourt Rocks fits that role best. Parsecboy 00:43, 24 May 2007 (UTC)

Parsecboy, then what are we going to argue with? See archive. The discussion has matured much more than you think. We already did reliable sites test, Google book, and major media tests. (Wikimachine 20:37, 24 May 2007 (UTC))

We already went through that one. I pointed out how Liancourt Rocks was not widely used in english newspapers. Good friend100 02:58, 24 May 2007 (UTC)

That may be because the issue never comes up in English newspapers. Parsecboy 11:25, 24 May 2007 (UTC)

Then it's not used. Usage in itself matters. (Wikimachine 20:38, 24 May 2007 (UTC))

But Tokdo and Takeshima are essentially NEVER used in English newspapers. Only to attest what Japan and Korea call themMacgruder 05:44, 24 May 2007 (UTC)

This is why Liancourt Rocks should be used as the title. Parsecboy 11:25, 24 May 2007 (UTC)

Tokdo is NEVER used in english newspapers??? I didn't know that! Good friend100 20:44, 24 May 2007 (UTC)

Discussion about changing name of Dokdo Article

I have some serious concerns about the timing and method of which this poll was put up. While I have not been a regular contributor for the last year or so, I still am around occasionally to see what's going on. The main reason I don't use Misplaced Pages anymore is BECAUSE of the POV fights, which I frankly get sick of because they are counter productive to this project.

As was mentioned above, the poll most likely will have problems with sockpuppets voting multiple times. The person who posted the poll is depending upon the arbitrary decision of whichever admin closes this. Frankly it leaves much to be said that it leaves it open to the possiblity of abuse on both sides.

Also as someone who has participated in other discussions about changing articles (ie deletion) I find people often try to jab at other's comments by putting degrading comments in an attempt to counteract a person's vote. Just some things to think about. Davidpdx 01:04, 23 May 2007 (UTC)

1. Misplaced Pages:Requests for checkuser/Case/Goguryeo
2. Misplaced Pages:Requests for checkuser/Case/Room218
3. Misplaced Pages:Requests for checkuser/Case/NekoNekoTeacher
4. Misplaced Pages:Requests for checkuser/Case/Appleby
5. Misplaced Pages:Requests for checkuser/Case/Kamosuke

I hope Appleby is not operating sock fleet anymore. Jjok 06:08, 23 May 2007 (UTC)

It's my sincere hope that weren't not just targeting Korean users, but also checking Japanese users as well. Certainly someone needs to make sure neither side is pulling crap. Davidpdx 08:46, 23 May 2007 (UTC)

These aren't on-going checkuser cases, as I understand, but ones from over a year ago. The NekoNekoTeacher case is a "Japanese side", Jjok and Komdori are included there, but were determined to be unrelated users. Parsecboy 11:50, 23 May 2007 (UTC)

Google Test Results

Here are the results of the Google test results that I proposed before with slight changes due to Wikimachines analysis. To date this seems to be the only test done that people don't dispute the counts themselves and the methodology of the way they are counted. All choices are counted with variations in spelling and naming (Dokdo, Dok Islets etc.) There are as indicated above disagreement about the validity of doing such a search, the choice of words and the conclusions that can be drawn from the results:

• Takeshima 19400
• Dokdo 13400
• Dokto or Tokto 21500
• Takeshima without T/Dokdo 420
• T/Dokdo without Takeshima 652

• Both the Dokdo/Tokdo and Takeshima on the same page. 95% (Approx 20850 ,21500 - 652)
• Dokdo/Tokdo only: 3% (652 )
• Takeshima only: 2%( 420)

Do not discuss these results here (unless there is a significant problem with the count itself). This is simply a place so people can easily see the results. Go above to do that. I'm aware that the last result seems to contradict the first result, but it is approximately correct. It is probably due to adding up different Google results and missed false positives and negatives giving some slight adding error. The test seems to indicate that the two are usually used together in the vast majority of uses. Macgruder 05:57, 23 May 2007 (UTC)

Macgruder, why did you not include "Tok Islets", "Dok Islets", "Tok Island", "Dok Island" variations? (Wikimachine 19:56, 23 May 2007 (UTC))

Sorry. I thought I just copied and pasted your link. I've redone it to make sure it is correct. It now has 21500 (another 200) Macgruder 16:58, 24 May 2007 (UTC)

When I googled Takeshima, there were many articles that did not have any connections to these islets. From this analysis, I believe thatTakashima can also refer to many other things in Japan while Dokdo has no other meanings. So google counts may bring a false result.

Kingj123 20:08, 23 May 2007 (UTC)

That's why we make sure to not search for Takeshima alone. See the discussions of Google below, where we've made sure to eliminate those issues. --Cheers, Komdori 20:14, 23 May 2007 (UTC)

Alternative Results

Goodfriend 100 voiced a concern that we shouldn't include variations of Takeshima. In order to address this concern I have repeated the results without these Takeshima variations. Of course, this has meant removing the Dokdo/Tokdo variants as well. (Removing Tokdo from the search gives a decisive results towards Takeshima so I have included it).

• Takeshima 18300
• Dokdo/Tokdo 16200
• Takeshima without T/Dokdo 759
• T/Dokdo without Takeshima 568

• Both the Dokdo/Tokdo and Takeshima on the same page. 93% (Approx 17500 ,18300 - 769)
• Takeshima only: 4%( 759)
• Dokdo/Tokdo only: 3% (568 )

This gives a similar result to before. Variations of searches do not always add up totals (i.e. 100%) consistently. This probably indicates small issues with how Google counts results. Macgruder 15:44, 24 May 2007 (UTC)

Comments

I'm still trying to catch up to current discussion but this google search results still doesn't resolve problem raised by Kingj123 above.melonbarmonster 19:29, 24 May 2007 (UTC)

Yes, it seems some of us are obsessed with google counts. You should note that google counts are not very accurate. Good friend100 20:41, 24 May 2007 (UTC)

Google Book search

These are manual counts.

• 2
• 2
• 4
• 0
• 1
• 1
• 2
• 0

• Takeshima total: 12
• Dokdo total: 88

(Wikimachine 01:02, 24 May 2007 (UTC))

• I don't know why but searching with Takeshima islands increases the number to 257 and most of the books seems relevant to the topic. --Kusunose 03:08, 24 May 2007 (UTC)

Clearly we need a closer look. Because Dokdo and Takeshima gives 20 results alone which is more than the above 12. I doubt that any of these Takeshimas would not be the island.

Also we need to remove any non English books. Further we need to check how these are being used. For example, Misplaced Pages policy states to by careful not to include books that by definition give local usage such as maps and guide books.

Then we need to remove from both the books that are simply attesting to each countries usage. Such as 'Takeshima is the name in Japan, and Dokdo is the name in Korea'.

Also, we need to consider modern usage. A book from 1907 would hardly be relevant. I suspect that from this we'll find that Liancourt Rocks is the one that is used most as the actual name.

I'm concerned too about who publishes the books: I wouldn't consider a book published by the Japanese fishing agency to reflect English usage.

Once again counts can be misleading, and need to be considered in context.

Note: Remove non English Uses

• Hits which are of the form "X (Japanese Takeshima)" attest to English usage of X, and Japanese usage of Takeshima. The latter matters to the Japanese Misplaced Pages, not to us.
• Hits which are of the form "X (Korean Dokdo)" attest to English usage of X, and Korean usage of Dokdo. The latter matters to the Korean Misplaced Pages, not to us.

When you remove all these you are going to be left with nothing in Encyclopedias, media, and perhaps books too for both Takeshima and Dokdo. Macgruder 03:49, 24 May 2007 (UTC)

Umm... I don't know what's happening either, but I'll do manual count on both Dokdo & Takeshima.

Using "islands" gives searches with the terms "islands". That makes it irrelevant. (Wikimachine 22:42, 24 May 2007 (UTC))

Actually, it's 17, not 20. Hmm.. what you're saying is quite interesting. I'll do this. I'll search for all combinations of Takeshima & Dokdo. And then divide the total result into Takeshima, Dokdo, and irrelevant. Good? (Wikimachine 22:44, 24 May 2007 (UTC))

That won't work, actually. Search can't show Dokdo & Takeshima simultaneously (they may be on different pages)--> I can't identify which uses both Takeshima and Dokdo. (Wikimachine 22:53, 24 May 2007 (UTC))

Second Google Book Search

• Dokdo: 877
• Takeshima: 495

Dokdo Manual counts

• 7
• 3
• 1
• 2
• 2
• 0
• 4
• 2
• 0
• 0
• 0

• NA-Class Japan-related pages
• NA-importance Japan-related pages
• WikiProject Japan articles