Wikipedia talk:Attribution/Archive 16

Original Research changes (take 2)
The section on OR is a mess. The current phrasing is:


 * Material counts as original research if it:
 * introduces a theory, method of solution, or any other original idea;
 * defines or introduces new terms (neologisms), or provides new definitions of existing terms;
 * introduces an argument without citing a reliable source who has made that argument in relation to the topic of the article; or
 * introduces an analysis, synthesis, explanation or interpretation of published facts, opinions, or arguments that advances a point that cannot be attributed to a reliable source who has published the material in relation to the topic of the article.

This is an overly complicated way of describing a relatively simple concept, and is also wrong in a few ways. Most obviously, a literal reading of it states that "material counts as original research if it introduces a theory" (which General relativity does without issue) or "method of solution" (which Proof by induction achieves without straying into OR). It also says that we can't define new terms without restricting the meaning of "new" to "previously unused outside of wikipedia". Additionally, that last sentence makes me have to go back and reparse every time I read it because I naturally read "arguments that advances a point" as a single gramatically-invalid phrase.

I've worked on my previous proposal and now suggest the following text:


 * Material is original research if it introduces if it promotes a point of view or idea in a way that cannot be attributed to a published reliable source. Original research can include any of the below if they have not been published in such a source:
 * a theory, argument, idea, or definition, or
 * a synthesis of information from multiple sources, or
 * a novel analysis, explanation or interpretation of published facts, opinions or arguments.

I think this is substantially clearer than the current phrasing, and doesn't include the problems of the current phrasing. Does anyone object to this phrasing? JulesH 15:31, 30 April 2007 (UTC)


 * I don't really object... but I do not consider your phrasing any clearer than what we currently have (in fact I find it more convoluted and clunkier). The current language seems very clear to me. I also question some of your statements ... for example: I don't see how the article on General relativity "introduces a theory".  The theoroy of General relativity was published by Albert Einstein in 1915/16 and has been known by scientists for decades ... so how can it be "introduced" by Wikipedia's article?
 * Your other examples confuse me even more... could you elaborate on what you see as the problem? Blueboar 16:56, 30 April 2007 (UTC)


 * I approve of what this proposed edit is trying to do, but IMO a bit more cleanup is needed. Part of the issue here is semantic.  As noted above, the Wikipedia general relativity article is not introducing that theory as its "introducion" was done by Einstein in 1915 and 1916.  Even more important to me is general relativity has since been described in numerous secondary sources.  Without the attention that it has gotten, general relativity would have remained "original research" to this day.  So IMO, the criteria for getting away from the "original research" label should be the topic being described in source that is
 * Secondary,
 * Independent of any primary source, and
 * Reliable.
 * My reason for adding the first two criteria is that there is plenty of scientific material which goes nowhere but still was originally published in journals that make good reliable sources. Independent secondary publication, being in another journal by another author with the journals and authors being independent of each other, is what in my mind represents the change from "original research" to "appreciated research".  Even then, topics which pass this test are usually not notable at first, but notability is a different issue. --EMS | Talk 17:22, 30 April 2007 (UTC)


 * This totally blows away source based research. Original research is only a problem when it introduces novelty. SchmuckyTheCat 17:04, 30 April 2007 (UTC)
 * Schmucky... I am not sure what you mean. To me any Original research would introduce novelty... if it didn't it wouldn't be original. Blueboar 17:26, 2 May 2007 (UTC)
 * Source based research may be original, but it isn't novel, and it is fundamental to the project. The original OR page was relatively stable until 2006, and had good wording about this. It's been diluted on the current page but it is still there. SchmuckyTheCat 17:36, 2 May 2007 (UTC)
 * Um... I'm still not getting what your objection is. Do you mean that you think Jules's proposed edit would ban doing any research (ie going out and looking for sources to back what is said in an article)? Blueboar 20:52, 2 May 2007 (UTC)
 * It bans stating the obvious. Beyond that it is also a much stricter ban on any source based research (ie, using primary sources) than what there is now, and what is there now is a far stretch from the origination of the policy as a way to keep out cranks. SchmuckyTheCat 04:33, 3 May 2007 (UTC)
 * In what way is it stricter than the current definition? It restricts the definition to material that "promotes a point of view or idea in a way that cannot be attributed to a published reliable source" which in my opinion narrows it from the way it is currently written (although clearly not the way it is usually interpreted). JulesH 09:53, 3 May 2007 (UTC)

The threshhold issue again: replacing WP:NOTTRUTH with matterial stolen from WP:NPOV
Go here to see what I have in mind. Basically, material stolen from Neutral Point of View (mostly Neutral point of view), spliced together with a bit more added by me. I believe this would resolve some of the issues - the binary nature of a threshhold, and the truth concerns. It probably still needs tweaking though.

Let us know what you think.

Thanks,

Armed Blowfish (mail) 23:59, 1 May 2007 (UTC)


 * I don't see how this answers the concern that the section in question needs to answer, which is that the reader coming across the statement "the threshold for inclusion in Wikipedia is whether material is attributable to a reliable published source, not whether we think it is true" might be concerned that this means we will publish anything that anybody has said without regard for its truth, even if the statement is known to be wrong. We need to address this concern, and show the reader why it is a good thing that wikipedia demands attributability, and why truth is not necessarily achievable in many circumstances. JulesH 16:46, 2 May 2007 (UTC)
 * .... but, er... we do "publish anything that anybody has said without regard for its truth", with some caveats: (a) the anybody should be notable; (b) It must have been published in a reliable source. See for example, Mahmoud Ahmadinejad: He says that the Holocaust is a myth, which is 100% not true. But we still report what Mr. Ahmadinejad said, even if the statement "is known to be wrong". Again, this issue of "verifiability, not truth" or similar wording is there for a reason, and that reason has not changed. ≈ jossi ≈ (talk) 01:54, 3 May 2007 (UTC)


 * JulesH - I have problems with the statement, "the threshold for inclusion in Wikipedia is whether material is attributable to a reliable published source, not whether it is true". Firstly, it implies to me that reliability is a binary issue - either a source is reliable and meets the threshold for inclusion, or it is unreliable and does not meet the threshold for inclusion.  Several flaws with that implication - reliability is relevant to the statement being made.  A source, using jossi's example, can be reliable for that statement, "According to Mr. Ahmadinejad, the Holocaust is a myth," but not for the statement, "The Holocaust is a myth."  I.e. all sources are reliable for reporting the opinions of their authors, but it takes more to be reliable on other matters, or even for the opinion of the author to be worth reporting.
 * Reputability might be a better word, as it is a bit more independent of the statement being made, although it may still be relative to a broad topic. The BBC has a reputation for being good on most topics they report on.  But a biology journal might have a good reputation on matters of biology, but not on matters of nuclear physics.
 * And then, of course, there are degrees of reliability and reputability. Who should we trust more on nuclear physics - the BBC or the widely respected (within the international scientific community) nuclear physics journal?  So it's really not a binary thing.  There are degrees of inclusion (see WP:UNDUE) and degrees of reliability/reputability.  Neutral Point of View suggests dividing space (determining degree of inclusion) based on the relative reliability/reputability.  (It also suggests other methods, but sources are probably the most straight forward.)
 * Also, "not whether it is true" might be misleading and confusing without surrounding context. We don't care about absolute truth - we do care about the general consensus, or lack thereof, of the experts and people of our time.  We do not report that the Holocaust is a myth, we convert it to a fact and say that Mr. Ahmadinejad said so.
 * Suggested replacement catchphrase - "Assert facts, including facts about opinions — but do not assert the opinions themselves, regardless of whether these facts and opinions are "true" in the absolute sense of the word." (Most of that sentence was stolen from Neutral Point of View.)
 * — Armed Blowfish (mail) 04:07, 3 May 2007 (UTC)
 * But that doesn't say anything like the same thing. It allows users to assert that something is a fact, without requiring them to provide attribution to support it being a fact.  The point made by the sentence you want to replace is that something can only be included if it is said by a reliable source.  Frankly, this is a binary issue: either there is a reliable source that has made such a statement, or there isn't.  If there is, the material can be included.  If there isn't, it can't.  That's the way WP:V works, and it's the way WP:A should work also.  JulesH 09:57, 3 May 2007 (UTC)
 * That's inconsistent with WP:NPOV. And, if what you said was right, dispute resolution would be quite hopeless.  But if it's not a binary issue, if it's a scale like WP:NPOV suggests, then there is room for compromise.  —  Armed Blowfish (mail) 11:09, 3 May 2007 (UTC)

Illustrated replacement text for the lead here. I think it makes more sense in context, but how much can you fit in the lead. (Thanks for reverting that for me Amarkov. I meant to do it myself, but the autoblocker has been fairly persistent today.)  —  Armed Blowfish (mail) 22:10, 6 May 2007 (UTC)

OR on Talk pages
(copied from Crum375's Talk page) Your recent reverts of my edits seem to be non-sense. I don't know at all. You said in your edit summary of WP:ATT, " 'everything in Wikipedia must be attributable' - this edit seems to imply OR on Talk is OK". If you say that Original Research on Talk page is OK, though I didn't edit anything concerning WP:OR, there is obviously no problem in my edits. In addition, YOU evidently published your opinion, experience, or argument in Wikipedia. See and. Please explain the meaning of your recent reverts of WP:ATT. -- PBeaver 23:21, 2 May 2007 (UTC)
 * I am simply saying that the edits you made to the policy are effectively implying that WP:OR is freely acceptable on Talk pages, whereas the ATT policy states that "everything in Wikipedia must be attributable". I think that the change to the policy that is implied by your edits would require a consensus. Crum375 23:29, 2 May 2007 (UTC)
 * Hmm. The edit you link is indeed problematic, because one should not "publish" original theories other places.  On the other hand, "original research" has always been OK on talk pages as part of deciding on the content of an article.  It is not true that claims in talk page comments must be attributable in the same way that article material must be. I don't know if there is an easy way to explain this in the policy itself.  CMummert · talk 23:55, 2 May 2007 (UTC)
 * I agree that we don't have the identical strict rule about OR in Talk as in the main article, but we still can't just freely post arbitrary OR in Talk. And if the material is related to living persons, the criteria are just as strict, per WP:BLP. Overall, I think the edits in question imply that OR is freely acceptable in Talk, and it is clearly not so. Crum375 00:15, 3 May 2007 (UTC)


 * That we can't publish OR on talk pages is a corollary of the fact that we can't publish them in article space and the fact that per WP:TALK talk pages should only be used to discuss changes to the article they're attached to. Users are perfectly free, however, to publish relevant OR in any appropriate non-article namespace; for instance original research about Wikipedia itself is more than welcome in the Wikipedia: namespace.  Similarly, OR that does help write a valid policy-compliant article is appropriate on the talk page of that article.  Restrictions on the use of talk pages do not need to be addressed here.  They're adequately covered already in WP:TALK and in WP:BLP.  JulesH 10:03, 3 May 2007 (UTC)
 * OR on talk pages may address which attributable text is reliable. For example, the many published claims that Group X or Group Y are plotting to take over the world are all attributable; the problem with simply stating and sourcing them is that they are a small minority view - and bunk. Showing either of these problems may take original research. Septentrionalis PMAnderson 14:49, 3 May 2007 (UTC)


 * I see nothing wrong with discussing material that may be OR on a talk page... Indeed it is often necessary. The key is that OR should not be added to the article itself.  Blueboar 15:05, 3 May 2007 (UTC)
 * Can you give us an example where OR material is 'necessary' on the Talk page? Crum375 17:58, 3 May 2007 (UTC)
 * Well, one obvious situation is where OR material was placed in the article, and then removed to the talk page for discussion. It is nescessary to see and discuss the OR material to explain why it is (or isn't) OR.
 * To take another... suppose the subject for a BLP wants to complain about something that is stated in the article on him. He may need to relate his own information on the statement (which is OR, since it is not sourced) to explain why he is upset.
 * Yet another is where IAR has been invoked because there are no sources for an article - editors may need to discuss the information to hammer out the language.
 * Finally, in a dispute over a given source, it may be necessary for the discussion about the source to stray into OR interpretation of what the source is saying... or it's context, in order to understand the different sides of the dispute and reach a resolution.
 * Perhaps 'necessary' is the not the exact word to use in most cases... helpful, desirable, informative, all are good words that convey what I am talking about. As long as we are strict with ourselves about what we add to the article itself, we should be flexible on the talk page and allow people to talk ... that's what it is for. Blueboar 18:46, 3 May 2007 (UTC)
 * I agree, as I noted above, that we can be a little more flexible about OR in Talk pages, but giving a blank check for OR in Talk space is not the way to go. Certainly the BLP example is wrong: we may never use BLP related OR in Talk (or anywhere else) if the material is controversial. I don't think IAR would help if we have no sources - in that case, the related article or section should be deleted. Overall, there is a bad tendency in Wikipedia for people to rely on OR, and we need to root it out. Certainly article space is the most critical, but we should make an effort not to include OR anywhere, and if it's BLP-related, unsourced or poorly sourced controversial material is prohibited everywhere on this site. Crum375 18:57, 3 May 2007 (UTC)

Talk pages need not be perfect. Besides, the encyclopaedia is only in the article namespace, so only the article namespace need abide by encyclopaedic standards. And most talk page comments are probably original research. I think this and I think that is original research, but essential for a discussion to reach consensus. Reserve refactoring for cases of intolerable incivility. — Armed Blowfish (mail) 19:12, 3 May 2007 (UTC)


 * ABF makes a good point... If you took Crum's comments to extremes, you could argue that we should delete this talk page as being completely filled with Original Research ... Indeed, you could do the same for all of the guidlines and policies themselves (I don't see a citation on any of them), not to mention the Village Pump, our User Pages... etc. No... It is the article space that needs to adhere to NOR... not the support pages. Blueboar 19:30, 3 May 2007 (UTC)
 * Again, some flexibility on Talk pages in discussing the sourced material that goes into the article, is obviously necessary. But BLP is a specific case where there is no flexibility at all - if it's controversial and unsourced or poorly sourced, it does not belong anywhere. Overall, we need to focus on sourced material, and discuss whether the source quality is sufficient, whether there is notability, relevance, redundancy, neutrality, proper writing, etc. There is no need for people to publish original articles (unless they relate to the running of Wikipedia itself) anywhere on this site. Crum375 19:39, 3 May 2007 (UTC)
 * No one is saying that placing an entire original article on the talk page is acceptable... but you have to be able to discuss your views, ideas, concepts, questions, comments, etc. on an article talk page. All of this is technically OR.  Obviously such views, ideas, concepts, questions, comments, etc. should relate to the article and it's subject, and the talk space should not be a place for people to rant or disrupt the community, but better there than in an article.  I agree that BLP should be held to a tight standard, but see no problem with discussing an original idea or material that relates on the talk page... It might turn out that it isn't that original after all (and jog someone's memory of a reliable source they saw)... or it might get shot down by editors saying "that sounds like OR... do you have a source to back that up?"  But there isn't really any harm in discussing it. Blueboar 20:09, 3 May 2007 (UTC)
 * I agree there has to be some flexibility and balance, so as not to stifle discussion, but the focus has to be on getting sourced material for the article, and presenting it properly. Given that important focus, and the fact that BLP requirements are the same everywhere, we should not just open the floodgates and say 'OR is OK on Talk pages', as we already have way too much OR (or just unsourced material) everywhere. Crum375 20:33, 3 May 2007 (UTC)
 * Talk page version of Biographies of Living Persons is No personal attacks. There isn't really consensus on removing personal attacks, especially the minor ones, but severe things like outing of personal information and outright libel can be oversighted away.  —  Armed Blowfish (mail) 20:54, 3 May 2007 (UTC)
 * I am not sure what you are referring to. This is the relevant part from the BLP policy, and it applies to Talk pages as well as anywhere else on this site. This is not personal attacks or outing, simply controversial material which must be removed immediately by anyone if unsourced or poorly sourced. Crum375 21:08, 3 May 2007 (UTC)

I'm not sure why we're having this discussion. Nothing we write on this page will stop WP:TALK and WP:BLP from applying to the content of talk pages. Therefore I'm not sure we need to consider the effects that ATT has on talk pages because there won't be any. PBeaver's original request was for a clarification of something that is in my experience commonly misunderstood: that it's OK to use OR on talk pages to guide us in the process of writing the encyclopedia, as long as we don't ask our readers to believe its conclusions. I don't see why any concerns that are addressed by either of the two pages I link above need interfere with us instructing editors that this is fine. JulesH 21:17, 3 May 2007 (UTC)
 * The discussion arose due to this edit by PBeaver. I reverted it because I believe that it implies that OR is fair game outside of article space, when that is clearly not the case. Crum375 21:40, 3 May 2007 (UTC)


 * At which point I disagreed with Crum's contention ... I think OR is "fair game" outside of article space, within limits ... it has to be in order to discuss things that should or should not be put in the article. Discussions may go off on a tangent, a user may come by with a question that can not be answered fully if we stick to purely OR material (for example... If someone were to ask: "Which of Einstein's theories is the most important?" ... I don't think you could answer that without going into OR and each editor's opinion.  The answer might not be something we include in an article, but it might have relevance as to how we structure an artcle).  There are a host of reasons why OR material and discussions might be appropriate on talk page. Blueboar 21:56, 3 May 2007 (UTC)
 * Again, I agree that OR rules, with the exception of BLP, are not as strict in Talk space as they are in article space. But the edit I objected to could easily be taken to imply that OR is only an issue in article space, and that would be wrong. Also, the policy currently says: "All material in Wikipedia must be attributable to a reliable, published source". Crum375 22:03, 3 May 2007 (UTC)

My original title was "Your contradiction" on User talk:Crum375, and I didn't change anything concerning WP:NOR or Original Researches on Talk pages. I just think you should make the meaning of "Wikipedia" clearer, this is, the range that WP:ATT should cover. I think, "Wikipedia" has two meanings at least, one is "all of the articles in article namespace in Wikipedia" and the other is "all of the things related to this Wikipedia project, including all of Talk pages, Templates, and other pages in Wikipedia project namespace(here, too)". See the upper left logo. Apparently you can see this "WIKIPEDIA" logo in all pages in Wikipedia project. Therefore, if you want to say, "All material in Wikipedia must be attributable to a reliable, published source", and "Wikipedia is not the place to publish your opinions, experiences, or arguments", you should make the meaning of "Wikipedia" clearer at the beginning of the page. I'm a beginner and was(am?) very very confused. The current begininng sentence, this is, "Wikipedia is an encyclopedia — that is, a comprehensive compendium of well-established knowledge" is not enough. This configuration is not enough that you are forced to have to regulate, saying, "Wikipedia is not the place to publish your opinions, experiences, or arguments". But, on the other hand, to improve articles and resolve disputes, anybody has to publish his/her opinions and get involved in arguments in Wikipedia. -- PBeaver 03:11, 4 May 2007 (UTC)

What does this WP:ATT cover?
What cases are Attribution applicable to? See this talk page. And see the edit summaries, eg.. Is this policy(WP:ATT) ordering us not to say any opinion everywhere, even Talk namespace? Has this WP:ATT denied all people's fundamental human rights, this is, freedom of speech, which is obviously attributed to and guaranteed by the Constitutions of all of the advanced countries? (I can't believe it.) If so, you can't say anything meaningful here and there in Wikipedia. After all, all Talk pages are no use. Before my editting, and now, WP:ATT says, Wikipedia is not the place to publish your opinions, experiences, or arguments, but this Talk page and others are obviously provided for all kind of editors as places to publish our(their) opinions, experiences, or arguments. In fact, even Crum375 said his opinion and editted without indicating the reliable reason to deny freedom of speech in Wikipedia. Apparently, "Wikipedia" has the two meanings. One is only articles in article namespace in Wikipedia, and the other all pages in Wikipedia project. "Wikipedia" used in the sentrence such as "everything in Wikipedia must be attributable" seems to be identical with one of Wikipedia is not. This "Wikipedia" means all Wikipedia's articles in article namespace, not opinions, experiencers, arguments, etc on other namespaces. Besides, the sentences such as "Wikipedia is an encyclopedia — that is, a comprehensive compendium of well-established knowledge", and, "everything in Wikipedia must be attributable" - these sentences, I think, can never imply "opinions, experiences, arguments, and OR on Talk pages are OK". They seem to imply that opinions, experiences, arguments, and OR on Talk pages are not OK as well as on article pages. Crum375 published his opinion and experience, saying, "I have a problem with implying that we can freely publish OR in Talk pages etc. But I didn't say such a thing. I didn't make such a edit. I only want to make the meaning of "Wikipedia" clear. My edits have been already reverted twice by User:Crum375. So, someone, please improve Wikipedia:Attributions as it doesn't have a mass of contradictions. I only mean, "Be accurate and make it reasonable". -- PBeaver 03:11, 4 May 2007 (UTC)
 * My personal opinion is that other than where this policy explicitly says otherwise, it refers only to articles (and templates that are transcluded into articles, by logical extension). There are other policies that govern the type of material acceptable in other types of page, and we should leave those policies to make their own rules without interfering here.  This policy is about the encyclopedia, and pages that aren't part of the encyclopedia (e.g. talk pages, the wikipedia namespace, user pages, etc.) shouldn't be subject to the same rules.  JulesH 09:09, 4 May 2007 (UTC)
 * Thank you for your answer. I was relieved to see your reasonable opinion. Thanks. -- PBeaver 23:03, 4 May 2007 (UTC)
 * That isn't just JulesH's opinion, that is the only possible answer. If a reader sees it, policies apply.  Everything else is subject to user and talk page guidelines, which are more free-wheeling. WP:NPA, WP:BIO, WP:CIVIL, etc. SchmuckyTheCat 01:40, 5 May 2007 (UTC)
 * Thank you, thank you so much. You are a nice guy, too. It seems that you put an end to this problrm as well as JulesH. Your opinion makes perfect sense to me and probably to everybody. And then, my current problem with WP:ATT is the following one: Everybody's business is nobody's business? I want to do what a smart beginner should do. Please don't bite me, even if my edit is not smart enough. -- PBeaver 17:55, 6 May 2007 (UTC)
 * I'm afraid I don't like this change at all, see WP:NOT, WP:V, and the other policies; they all make the same claims about all of Wikipedia.
 * The threshold for inclusion in Wikipedia is verifiability, not truth. "Verifiable" in this context means that any reader should be able to check that material added to Wikipedia has already been published by a reliable source.
 * Why is that not a problem, and this one is? I'm sorry, since this is no longer tagged official policy, then changes should first be made at the pages that are official policy! --Merzul 17:20, 7 May 2007 (UTC)

Search engines are not original research
I ran into a debate at Talk:AACS_encryption_key_controversy, probably not the only one of its kind, in which some people have claimed that using a Web search engine as a reference violates WP:NOR or is not a reliable source (or now WP:ATT). I believe these searches may sometimes have problems with WP:V but are otherwise good references.


 * A search engine indexes the Web whether you type in a search or not. You do not "create" research by entering search terms.


 * Search engines are reliable sources - well known, consistent, reliable, with fixed URLs to reference. They can be checked by third parties.  They are often cited by media outlets e.g. the BBC.


 * Limitations of search engine results include inaccuracy (the databases are never complete, and the estimated number of results given can be far off the mark) and verifiability (for most search engines, historical searches are not possible so the number of hits will change over time). The cure for this is to limit claims to what you expect to remain verifiable - e.g. "xxxx is well-known", "future scientific publications by XXX can be located with this search", and so forth.  You may be wrong - xxxx might be censored out of the search engine, XXX might die tomorrow - but many cited online sources vanish over time, and these are not the worst offenders.


 * There's a strong temptation to use search engines for original research that has to be watched. There are 517 million hits for "color" and 123 million hits for "colour", but that doesn't prove that more people speak Americanized English, even on the Web.  For all I know the hits come from automated E-mail headers or Web page generators.

If consensus supports me, I'd like to see a bit of this discussion added to the WP:ATT policy to avoid many future debates on talk pages. Mike Serfas 14:38, 6 May 2007 (UTC)


 * The biggest problem with using search results as a reference is that they can change. They're also a primary source that requires interpretation: how many of the results are relevant to the topic at hand (I'll grant that in this situation, the answer is "all of them", but in most it's pretty likely there will be some false positives).  Interpreting them in the way you suggest (e.g. "xxxx is well-known") does constitute original research, I'm afraid.  "Future publications ..." isn't a reference, so is outside of the scope of ATT; I understand links to search engines are discouraged by the WP:LINKS guideline, although I do not understand the reasoning behind this.
 * As for the fact that other media use search engines like this, it must be noted that there is no reason these other media cannot perform original research, and in fact journalists must generally perform original research to do their jobs. This is an important difference between a news publisher and an encyclopedia.
 * That the results existed before you entered the search is irrelevant: have they been published, by somebody who knows something about the subject area and who has checked the interpretation of them makes sense? If so, cite that publication.  If not, they are almost certainly original research. JulesH 15:11, 6 May 2007 (UTC)


 * Search engines are not reliable sources, they keep their methods of searching a trade secret, they allow ad income to influence results, and you can choose what type of results you get by crafting your query. They lack any sort of peer review, they don't reveal their sources or methods, I could go on. Now if a reliable source did research on search engine results, and published a paper, that would be a good source. HighInBC(Need help? Ask me) 15:13, 6 May 2007 (UTC)
 * Even if search engines were a reliable source, what exactly would they be reliable sources for? They make no statement, implied or otherwise, that many hits means a subject is well-known. -Amarkov moo! 16:46, 6 May 2007 (UTC)

I have twice used a google search as an inline reference for something that I wasn't sure really needed a reference for one reason or another (such as I thought it was probably already covered by noninline references) and I was too lazy to adequately documment a fact with mostly unimportant yet actual conditions/limitations (true but ... (ref) except for ... (ref) yet this possibility has never been observed (ref)). A made up example of when I feel it is appropriate would be "the sky is blue http://www.google.com/search?hl=en&q=%22the+sky+is+blue%22+&btnG=Google+Search" in an article where the details of this claim are not important, yet for some reason there is a need to provide a cite (such as someone stuck "fact" on it). WAS 4.250 17:32, 6 May 2007 (UTC)

Thanks - I think it's time to concede the point. WP:LINKS is already the consensus, and that manual's exception - that the search engine link can be properly used when it's the subject of the article - is probably the best answer. I don't like the idea of replacing "citation needed" with "go look it up yourself" as a link. I also don't believe that trade secrecy or ad skewing are any worse for search engines than news media. Writing the article about the search engine rather than using the search engine - that really is a better style. Mike Serfas 22:13, 6 May 2007 (UTC)