Wikipedia talk:Attribution/Archive 17

Archive 10

←

Archive 15

Archive 16

Archive 17

Archive 18

Editors, please note:

After four months of discussion at Wikipedia:Attribution, a number of editors at Wikipedia talk:Attribution agreed on a means of merging Wikipedia:Verifiability with Wikipedia:No original research, while also streamlining Wikipedia:Reliable sources into a simpler FAQ at WP:ATT/FAQ. This would in no way deprecate those longstanding policies, which are conceptually distinct and handy to refer to as such. In some cases, this page may be easier to refer to, but when attempting to refer to individual elements of this page, it will best to refer to those.

WP:ATT is intended be a more cohesive version of the core content policies with which the Wikipedia community is already familiar.

There is a community discussion on the merits of WP:ATT, at Wikipedia talk:Attribution/Community discussion.

A poll at Wikipedia:Attribution/Poll was held to gauge the community's thoughts about the Wikipedia:Attribution merger from March 30, 2007 at 00:00 UTC to April 07, 2007 at 01:00 UTC. The poll is now closed.

There is a community discussion on the Role of truth that concerns the role of "truth" in Wikipedia policies.

"reliable publication process"

I think this need some clarification, this can be seen as excluding self-published material by organizations such as Greenpeace, Amnesty, political parties, labor unions, and so on, even if they are only issuing a press release, or stating something on their website, and stating their view regarding something.Ultramarine

You are correct. It certainly can exclude self-published material by such organizations. Not in every situation, but often. In general, if you want to include statements made by such, you need to find a reliable secondary source, such as a news paper, that has discussed the claims the organization make and quote that. Blueboar 17:36, 7 May 2007 (UTC)

This is unclear. Can the Party Platform of the Democratic Party be quoted for anything since it is self-published? Or a press release by Amnesty regarding Press Freedom which is self-published? Ultramarine 17:43, 7 May 2007 (UTC)

The Democratic Party Platform can certainly be quoted in an article on the Democratic Party... but probably not in an article on, say, Haliburton... and he press release from Amnesty can be included in an article on Amnesty... but not in the article on Journalism. They might be usable as a relevant statement of oppinion (ie "according to the Democratic Party... " or "according to a press release issued by Amnesty...") but they definitely should not be used to back a statement of fact. Blueboar 18:16, 7 May 2007 (UTC)

Well, there are editors who use the current text as an excuse for excluding all such sources from Wikipedia, even if they are only quoted as above, since they are self-published, so some sort of clarification is probably needed.Ultramarine 18:20, 7 May 2007 (UTC)

Yup... people do that. Sometimes their reasoning is correct (and the source should be removed) and sometimes it is incorrect (and the source is appropriate). It often depends on the article, the context of how the source is being used, and several other factors. Can you give me an example of a situation where the quotation or citation was removed... then I would have a better idea of the context and whether removal was appropriate. Blueboar 18:29, 7 May 2007 (UTC)

Well, the above is not what the current text states, this is your view regarding this, others interpret the current text as blank prohibition. For example, arguing that Freedom in the World should not be allowed as a source for any statement or view in Wikipedia, demanding a higher standard than peer-reviewed academic journals who often use the scores from this source. Presumably we cannot quote das Capital or the Bible since they have not gone through a reliable publication process.Ultramarine

I think we need something similar to Wikipedia:Notability. That is, the views of notable persons and organizations should be allowed in the fields they are notable, obviously mentioning that they are views. So the views of Amnesty could be quoted in article regarding the Thailand coup, the views of das Capital in article about capitalism, or the view of Catholic church in an article about abortion.Ultramarine 19:07, 7 May 2007 (UTC)

I guess I am the one being accused of a 'blank prohibition'. I deny this, and point out that using the best reliable secondary sources is not an undue burden on editors, worthwhile because an encyclopedia based on the best reliable secondary sources is a better encyclopedia. SaltyBoatr 20:16, 7 May 2007 (UTC)

Your novel interpretation would require that Wikipedia removes many of the sources and material now in the encyclopedia. Also, note that current text is not a prohibition of other sources than university publishers, mainstream newspapers, and known publishing houses. It only states that these are in general the most reliable, not that other sources are not allowed.Ultramarine 21:52, 7 May 2007 (UTC)

My interpretation is hardly novel, and incidentally, using the most reliable sources is a desirable thing which will result in a better encyclopedia. SaltyBoatr 22:19, 7 May 2007 (UTC)

So are you actually arguing that Wikipedia should never be allowed to cite the Bible or das Capital or the Democratic Party Platform since they have not had a a reliable publication process? Ultramarine 22:23, 7 May 2007 (UTC)

No. I don't take it on myself to re-write the guidelines. I see wisdom in a strict guideline for attribution, because we all (should) want the best possible encyclopedia. Using the most reliable sources is a good thing! (And, if the material belongs in an encyclopedia, should be easy enough to find and use the most reliable sources. SaltyBoatr 22:41, 7 May 2007 (UTC)

But these sources have not gone through a "reliable publication proces". So according to your interpretation Wikipedia should not be allowed to cite them. Are you really arguing that this will lead to a better encyclopedia? Ultramarine 22:46, 7 May 2007 (UTC)

Furthermore, here you are arguing that only university publishers should be allowed as sources, even denying that a non-university publisher of acadmic books is a reliable. [1]. So Wikipedia cannot even use cite newspapers and such publishers according to you? Ultramarine 22:49, 7 May 2007 (UTC)

If you are questioning the policy, do not address your question to me. I am only trying to follow policy, not re-write it. SaltyBoatr 23:06, 7 May 2007 (UTC)

This is the place to discuss policy and interpretations. If you have an inerest in discussing the best policy for Wikipedia (which I assume you do), and not in Wikipedia:WikiLawyering, then please answer and dicuss in good faith the questions and issues above.Ultramarine 23:34, 7 May 2007 (UTC)

I think that a strict policy is not an undue burden on editors, worth the extra effort because it makes a more credible encyclopedia. Far too often, editors are lazy and choose to Google search the blogs and personal web pages, instead of genuine editing based on the most reliable sources. Lazy editors make Wikipedia worse. SaltyBoatr 15:43, 8 May 2007 (UTC)

As can be seen below, you are alone in your personal interpretation. Arguing on other pages that only university publishers are allowed as sources in Wikipedia, excluding even non-university acadmic publishers, is only destructive. Misinterpreting policy or relying on technicalities to justify inappropriate actions is not constructive.Ultramarine 16:30, 8 May 2007 (UTC)

Personal interpretation? I am just reading what WP:ATT literally says. You have not addressed my point that this strict policy is actually not an undue burden on editors, worthwhile for the improved credibility for the encyclopedia. SaltyBoatr 17:24, 8 May 2007 (UTC)

See below, no one agrees with your interpretation. You refuse to dicuss that your personal interpretation that even non-university academic publishers are not allowed would require deleting much of Wikipedia.Ultramarine 17:28, 8 May 2007 (UTC)

I get the impression that you would rather argue the policy than simply taking the time to use 'most reliable sources'. How hard can it be? This is not an undue burden. SaltyBoatr 17:44, 8 May 2007 (UTC)

So you continue to insist that a given non-university academic publisher is not acceptable in Wikipedia. Again, your personal interpretation requires the deletion of much of Wikipedia.Ultramarine 17:47, 8 May 2007 (UTC)

Read the policy: A self-published source is material that has been published by the author, or whose publisher is a vanity press, a web-hosting service, or other organization that provides little or no editorial oversight. Personal websites and messages either on USENET or on Internet bulletin boards are considered self-published. With self-published sources, no one stands between the author and publication. It's true that "web-hosting service" is an unfortunately vague way of putting things - any document on the Web is on a web-hosting service. But the intent here is very clear - we're talking about the situation where one person writes something and has nobody to second their motion. The Democratic Party Platform is obviously looked at by many different people, and the Democratic Party provides editorial oversight. Likewise with Amnesty International, a union's social commentary, etc. Bibles aren't even in this one because the apostles aren't paying a vanity press to print them. Looking at this policy, the one flaw I see is that while the last sentence I quote is clear, there are circumstances where the first could be misconstrued - for instance, you could argue that a pharmaceutical company's press release or MSDS is a "vanity press" because the company pays to have a Web site, even though these documents are written by many people, checked by editors, and subject to regulation by federal agencies. But I think it's clear that's not the intent. I think that there's a disturbing bias by some commentators to view what a media company pays to put on its web site as somehow better than what Amnesty International pays to put on their web site, which is not grounded in reality. Mike Serfas 13:17, 8 May 2007 (UTC)

This is why I have been saying that such sources can be reliable for statements of opinion, but not for statements of fact. It is appropriate to say (for example): "According to the Democratic Party's platform, Rudy Giuliani showed incompetence when dealing with air quality standards following 9/11.<ref>[http://www.democrats.org/a/2007/05/more_of_the_sam_1.php]</ref>" because this is attributable to the party platform website. We can verify that this is their opinion. However, it would not be appropriate to simply say: "Rudy Giuliani showed incompetence when dealing with air quality standards following 9/11.<ref>[http://www.democrats.org/a/2007/05/more_of_the_sam_1.php]</ref>". Blueboar 14:57, 8 May 2007 (UTC)

Agree with this but I think this should be stated more explicitly. How about "Notable opinions or views, not stating a fact, can use as a source the publication expressing this view. Such views must pass Wikipedia:Undue weight."Ultramarine 00:18, 11 May 2007 (UTC)

Yes, this is very important. The best sources for the statement "Some common arguments against foo are...." are places where these arguments are made. If this were stated explicitly, in the words you express, it would clear up one of the major problems with the official Verification policy: it becomes very hard to source the existence of arguments in disputes if you can't point to the publications of the people making the argument.24.59.100.145 19:10, 25 May 2007 (UTC)

I think we finally have a clear explanation of this. I would add that the safest way of expressing the opinion is a properly selected quotation. There seems to be some bias against doing it. Quotation can be biased, but it is less likely to be wrong than paraphrase. 05:07, 28 May 2007 (UTC)

Publishing houses

I see this section thread in Archive 7. Did anything come of this suggestion to ease this strict guideline? I guess not because I still see the strict guideline in WP:ATT, giving preference for books published by universities over those published by "known publishing houses". My personal opinion is that Wikipedia benefits from high quality standards, policies and guidelines. SaltyBoatr 20:48, 7 May 2007 (UTC)

Per WP:ATT In general, the most reliable sources are books and journals published by universities; mainstream newspapers; and magazines and journals that are published by known publishing houses.. This literally excludes school textbooks, those which are published by a publishing house not affiliated with a university. It seems to me that school textbooks are thoroughly fact checked, and should be included on this list of 'most reliable sources'. Can school textbooks be added to the list of 'most reliable sources'? SaltyBoatr 17:20, 10 May 2007 (UTC)

School textbooks are hotly disputed in many nations and often subject to various kinds of bias and censorship.Ultramarine 17:23, 10 May 2007 (UTC)

Bias and censorship? Not that I see. What actually happens is that textbooks are vetted by committee. Not unlike the fact checking process that occurs at a university publishing house. SaltyBoatr 17:37, 10 May 2007 (UTC)

See for example the Japanese history textbook controversies or this regarding Russia: [2]Ultramarine 17:41, 10 May 2007 (UTC)

You seem to have the premise that Wikipedia is about truth. Indeed Wikipedia is not about truth, see WP:V. Wikipedia instead has the standard of verifiability. The key issue is that 'most reliable sources' have many people involved in the checking, and textbooks are some of the most thoroughly checked books in existence. SaltyBoatr 17:49, 10 May 2007 (UTC)

You are arguing that the textbooks mentioned above, or Palestinian textbooks and Saudi textbooks, are as reliable as peer-reviwed articles? Ultramarine 17:52, 10 May 2007 (UTC)

A Saudi University publishing house now qualifies as 'most reliable', what is the difference? SaltyBoatr 18:04, 10 May 2007 (UTC)

Interesting, so Wikipedia should be able state that the Protocols of the Elders of Zion is a genuine document, citing one of the 'most reliable sources' available?Ultramarine 18:07, 10 May 2007 (UTC)

So back to my question. Textbooks from well known publishing houses are thoroughly vetted by committee. It seems that they meet the standard of 'most reliable' because so many people are involved in the checking, regardless of whether the publishing house is affiliated with a university. SaltyBoatr 18:14, 10 May 2007 (UTC)

American university level textbooks can certainly be considered a reliable source, but they are necessarily less reliable than the sources they themselves quote, that is the scholarly articles and books.Ultramarine 18:18, 10 May 2007 (UTC)

I disagree. As long as fact checking has been used, they should be more reliable in most cases. I can point to scholarly articles that come to very incorrect conclusions, for instance this paper where the authors suggest that with some common programs, an appropriately designed computer processor might execute hundreds of instructions in parallel. This is known to be hopelessly optimistic (and this was pointed out by numerous studies released shortly after the one I link to) and it is unusual for the actual value to exceed 4. Authors of textbooks are usually more conservative in their claims than authors of scientific papers. JulesH 20:35, 10 May 2007 (UTC)

Obviously there are lot articles with wrong conclusions and errors, corrected by later articles, otherwise science would be quite easy if everyone got it right the first time. But do peer-reviewed articles quote, say, introductory textbooks? Generally, no. They go back to the original research articles or maybe review articles. So the journals and researchers do not consider introductory textbooks to be the best sources.Ultramarine 20:42, 10 May 2007 (UTC)

No, but the fact that there are so many incorrect articles means that such articles are not generally considered authoritative by themselves. You have to perform a review of multiple articles to be sure you're not getting one person's erroneous opinion. Such reviews are (theoretically) performed by the authors of textbooks, and therefore the textbooks are (in cases where the reviews are performed adequately) superior to most articles. Seriously, I don't know where this theory that "if A quotes from B and C is of the same class as B, then C must be a better source than A" comes from. A might be (and more often than not is) a better source because its authors did the work to know that B (and not C) was the best source to quote from. JulesH 21:25, 10 May 2007 (UTC)

Say an introductory textbook describes a study. You examine the study itself and finds an error in the textbook description. Obviously the original article is more reliable, not the textbook. But I think you are talking about broader literature reviews. Here again a review article is generally more reliable, an introductory textbook may not even have inline citations, making the statements impossible to check. Only if there are no literature reviews available, for example in very basic uncontroversial subjects, or there are no recent review articles, is an introductory textbook really the best source.Ultramarine 00:25, 11 May 2007 (UTC)

Online

Yukichigai, it makes no difference that it's online. The point is that it's an encyclopedia and shouldn't contain nonsense. :-) SlimVirgin ^(talk) 21:12, 7 May 2007 (UTC)

POVness of the View of majority of scholars

We have a quote from Mark Cohen, a Professor of Near Eastern Studies at Princeton University, whose research specialty is Jews in the Muslim world. This professor has made an statement in the The Oxford Handbook of Jewish Studies about what most scholars hold about Arab anti-semitism. I would like to write this sentence in the intro as "According to Mark Cohen, most scholars hold that"... Jayjg however says that: "Cohen's statement is his own POV, and it contradicts the views of other scholars. We don't insert one specific POV in the lead, that's a violation of NPOV."

I think Cohen is an specialist in his field and knows the view of majority of scholars. He is not saying the view of his, but the view of majority of scholars. Unless someone presents another sourced sentence on the view of majority of scholars, there is no reason to think that Cohen's statement is wrong. Since it is the view of majority of scholars, yes, it should be in the intro. It is not "inserting one specific POV in the lead"

Jayjg says that it is Cohen's POV that: "most scholar hold that view". I say historians relate the facts accurately (and also make professional interpretations of those facts). Those interpretation parts are their POV but to say what most scholars think is not an interpretation or POV. This is mentioned by an expert in the field in The Oxford Handbook of Jewish Studies . --Aminz 23:43, 9 May 2007 (UTC)

Frankly, I think it almost always better to avoid controversy in the lead. The lead should be as descriptive as possible: it's meant for orientation One can't summarize a position accurately in a single sentence before people have even found out what the article is about. Yes, one gives his opinion and all responsible opinion in proportion, but later in the article where there's room for it. Cohen can & does express his POV in a published work we can & should quote him--it's only our own that we cannot quote. But not in the lead. DGG 05:48, 10 May 2007 (UTC)

The point is something else. Cohen says: Most scholars believe in X. He doesn't say "I believe in X". The question here is whether Cohen is relating a fact or his opinion. --Aminz 07:09, 10 May 2007 (UTC)

Everyone is relating their POV when they make a statement, the only real concern is the weight of that POV. Being a reliable source it is acceptable to use his position as a reference as long as the citation format is proper. If there is an opposing view point in direct contradiction from an additional reliable source then the article should reflect that as well. It is not the job of an editor to decide which relaible source is correct, only present the position of all applicable realiable sources in order to make sure that the article is attributed in a NPOV manner to multiple independant reliabe sources. NeoFreak 12:23, 10 May 2007 (UTC)

A completely different topic regarding WP:ATT

on WP:V, it states, under Burden of Proof: Quotations should also be attributed. Is it outlandish to suggest that the WP:ATT page be reserved for a policy detailing that quotations should be attributed? The current iteration of WP:ATT would just get moved to some other name.

Please note that I'm not even touching the merits of the WP:ATT policy. What I'm arguing is that the name of WP:ATT might be better used for policy regarding the attribution of quotes. Thoughts, criticisms, comments, gripes etc. ⇒ SWATJester ^{Denny Crane.} 12:33, 10 May 2007 (UTC)

Im not clear about what you are asking. Are you saying that you think ATT should only apply to quotations and V should apply to everything else? Blueboar 12:58, 10 May 2007 (UTC)

Close. I'm saying that WP:ATT should be quotation specific, and all the current discussion on this page should be moved to some other page, with a different policy name. ⇒ SWATJester ^{Denny Crane.} 00:32, 11 May 2007 (UTC)

Suggest that claims -- at least controversial ones -- whether or not direct quotations, should be attributed as well. See my comment under Wikipedia talk:Verifiability#A proposed change to the nutshell:Specific claims must be checkable by non-experts --Shirahadasha 02:18, 11 May 2007 (UTC)

Does it matter what we call them? The point of ATT was to combine V and RS. That quotations should be attributed ought to be a basic principle of honest writing. Unfortunately, many noncontroversial articles presently in WP do not follow it, and we have no real way of enforcing it. That even copied text should be attributed in addition to being PD is similar ignored--is not even policy. For the pages from PD sources, almost none state explicitly what portions are copied from the source. Personally, I think both are important enough to go into any combined ATT page. DGG 12:20, 11 May 2007 (UTC)

No, I can't go with ATT being quotation specific... it should definitely apply to quotes, sure, but it also should apply to other statements as well. ATT is intended to be a combination of specific concepts in V and NOR - Policies which have a scope well beyond just quotes. The scope of where ATT applies should be no different. Blueboar 12:30, 11 May 2007 (UTC)

Notable views

Moving here for discussion: SaltyBoatr 15:25, 12 May 2007 (UTC) Notable opinions or views, not stating a fact, can use as a source the publication expressing this view. Such views must pass Wikipedia:Undue weight.

See the discussion above at "reliable publication process". You have not replied to the arguments there.Ultramarine 15:26, 12 May 2007 (UTC)

Ultramarine added this paragraph above to the article, claiming 'consensus'. I may have missed it, but I see no history of prior discussion or consensus regarding this wording. I am moving it here so it can now be discussed. SaltyBoatr 15:36, 12 May 2007 (UTC)

I see several problems with the proposal, the biggest is that it unnecessarily validates the use of self published material. Assuming the premise that the 'view' is notable, per Wikipedia definition, to be notable the view needs to be noted and published by credible third parties. Then, why not use the credible third party as the source? All notable views have a credible third party source, or they would not be notable. This proposed new section is unnecessary. SaltyBoatr 15:36, 12 May 2007 (UTC)

See again the section ""reliable publication process" above. You misunderstand "self-published", it is primarily meant to mean single individuals publishing on their own webpages or through vanity press. Not organizations like Greenpeace, the Democratic party, and so on, where many persons examine the material before publication and whose views are notable. See what other think here: [3][4] Ultramarine 15:44, 12 May 2007 (UTC)

Would you address my question, that is: If the 'view' is notable, it has been 'noted' by a credible third party and published. That being true, editors can use then the third party published material. Your proposed new section is unnecessary because notable views always have third party references. SaltyBoatr 15:55, 12 May 2007 (UTC)

That is not correct, the views of the Democratic Party Platform is notable, even if a newspaper has not reprinted it in whole. The Platform is a notable source in itself. Similarly, if a notable presidential candidate expresses an opinion on his website regarding an issue, this is a notable view. Wikipedia has many articles using such sources.Ultramarine 16:01, 12 May 2007 (UTC)

No. All of the topics in the Democratic Party Platform are not notable per Wikipedia's definition of notable which is: A topic is notable if it has received significant coverage in multiple reliable secondary sources that are independent of the topic. Your personal definition of 'notable' may differ from Wikipedia's definition. SaltyBoatr 16:06, 12 May 2007 (UTC)

Note that is states topic, not source. Wikipedia:Notability is about if there should be an article about this topic. I quote "Notability guidelines determine whether a topic is sufficiently notable to be included as a separate article in Wikipedia. These guidelines do not specifically regulate the content of articles, which is governed by other Wikipedia's guidelines, such as those on the reliability of sources and trivia." You are arguing that Wikipedia should delete all the text in the articles about Presidential candidates that quote the platforms of the candidates? Ultramarine 16:16, 12 May 2007 (UTC)

Also, not all of the views in the Democratic Party Platform are notable, per even the dictionary definition of notable. To be notable, it needs to have been noticed by multiple credible secondary sources. It appears that your proposed new 'notable views' section creates a loophole for including of the non-notable views of organizations or individuals in the encyclopedia with no independent attribution, with the threshold being only 'personally notable' to the editor. That allows editors to personally editorialize in the encyclopedia. That would lower our standards and harm the credibility of the encyclopedia. SaltyBoatr 16:26, 12 May 2007 (UTC)

As stated, all views must follow Wikipedia:Undue weight. Again, Wikipedia:Notability is about if there should be an article about this topic. I quote "Notability guidelines determine whether a topic is sufficiently notable to be included as a separate article in Wikipedia. These guidelines do not specifically regulate the content of articles, which is governed by other Wikipedia's guidelines, such as those on the reliability of sources and trivia." You are arguing that Wikipedia should delete all the text in the articles about Presidential candidates that quote the platforms of the candidates? Or that Wikipedia should delete all the criticisms citing Amensty from the Guantanamo Bay detention camp article? Or all citations from the Bible in the Armageddon article? Ultramarine 16:44, 12 May 2007 (UTC)

If, as Ultramarine argues, WP:Notability does not apply; then, we would still need to define 'notable view'. Ultramarine appears to be arguing that the standard should be personally notable in the opinion of the editor. Instead, I argue that 'notable view' should mean noticed by multiple secondary sources, independent of the Wikipedia editor. And, that being the case, then the secondary sources can be used making your proprosal unnecessary. Your proposal appears to allow a loophole that encourages original research by using selective quotations of primary material. SaltyBoatr 17:04, 12 May 2007 (UTC)

WP:Notability explicitly states that it does not apply. Wikipedia:Undue weight is what applies. You have not answered my questions.Ultramarine 17:11, 12 May 2007 (UTC)

Your questions appear to be straw men. My question remains, if WP:Notability does not apply for the definition of 'notable', then how do we quantify 'notable view'? SaltyBoatr 17:41, 12 May 2007 (UTC)

If they are, then you can easily answer them. They are very important questions, so please state your view regarding whether wikipedia should delete this material. Regarding what views should be allowed and how much space for each, see Wikipedia:Undue weight. To anwer my own question, yes, the views of the Presidential candidates, Amnesty, and the Bible pass in the articles described.Ultramarine 17:38, 12 May 2007 (UTC)

My question remains, if WP:Notability does not apply for the definition of 'notable', then how do we quantify 'notable view'? SaltyBoatr 17:41, 12 May 2007 (UTC)

Sources from individuals like vanity press or personal websites are excluded per policy; the views of important organizations like Amnesty are certainly part of a neutral point of view. When is a minority to small to even have a view? That depends on the context and in practice this have rarely been a problem, in most cases it is easy to identify a significant minority. Now answer the very pratical questions above.Ultramarine 17:48, 12 May 2007 (UTC)

And, how shall we quantify what is a 'notable view'? Your answer, 'it is easy to identify' is not quantifiable, but rather is subjective. SaltyBoatr 18:03, 12 May 2007 (UTC)

In practice this has not been a problem as stated above. Now answer the very practical and real questions above.Ultramarine 18:07, 12 May 2007 (UTC)

I prefer not to stray from the topic, which is your new proposed wording. Your proposal starts with Notable opinions or views.... I wanted to know how to define 'notable' and looked to WP:Notability, which you then argued did not apply. Now, the best you offer is that quantifying 'notability' is "not a problem". I disagree, subjectivity is bad because that creates a loophole for editors to subjectively create selective quotes from primary source material, a form of original research, which creates an unnecessary risk to credibility. SaltyBoatr 18:29, 12 May 2007 (UTC)

I have edited Wikipedia for years and have never had any serious problems with Wikipedia:Undue weight. This is the place the discuss policy and implications. Those interested in discussing good policy that helps Wikipedia should discuss the real and practical impications of their proposals. Thus, answer the very real questions above which have important implications.Ultramarine 18:39, 12 May 2007 (UTC)

From Wikipedia:Undue weight: "Jimbo Wales, paraphrased from this post from September 2003 on the mailing list:

• If a viewpoint is in the majority, then it should be easy to substantiate it with reference to commonly accepted reference texts;
• If a viewpoint is held by a significant minority, then it should be easy to name prominent adherents;
• If a viewpoint is held by an extremely small (or vastly limited) minority, it does not belong in Wikipedia (except perhaps in some ancillary article) regardless of whether it is true or not; and regardless of whether you can prove it or not."Ultramarine 10:40, 13 May 2007 (UTC)

I don't think this is a particularly huge issue. If the view is notable, what does it matter what source it was published in? As long as it is reliable enough the we know whose opinion it is, at least. How to determine whether it is notable is a different matter: clearly a republished opinion is notable, but I think there is room for other things. Whenever anyone is clearly qualified to give an opinion on a topic, it seems reasonable to include it if we have an article on the topic. This is the present second exception on self-published sources, phrased in a less-strict way, and only applying to opinions. I don't have a problem with this.

I'm not sure the addition is necessary, but if it does stay I would suggest it would be a good place to point out that when using somebody's opinion in this way it should always have a prose attribution. JulesH 16:59, 12 May 2007 (UTC)

Unfortunately the above discussion shows that it is. How about Notable opinions or views, not stating a fact, can use as a source the publication expressing this view. Such views must pass Wikipedia:Undue weight. Prose attribution should always be used in such cases. Ultramarine 17:09, 12 May 2007 (UTC)

I disagree, the existing policy covers just fine. The new proposed text adds a bit of ambiguity that creates a loophole which encourages original research through selective quotation of primary material, which creates unnecessary risk of loss of credibility for the encyclopedia. SaltyBoatr 17:35, 12 May 2007 (UTC)

This is in effect the long standing policy, your novel interpretation requires deletion of much of Wikipedia. Primary sources should obviously always be avoided. Also, they are rarely opinions or views, that is a secondary source.Ultramarine 17:42, 12 May 2007 (UTC)

Pardon me, I am advocating for keeping existing policy unchanged. This does not mean I require deletion of anything that meets present policy. SaltyBoatr 18:03, 12 May 2007 (UTC)

No, as you are arguing for the deletion of much of the sourced and accepted content in Wikipedia you are arguing for a new policy.Ultramarine 18:07, 12 May 2007 (UTC)

If what you claim is true, it is magical. Consider that I advocate to change not a single word. SaltyBoatr 18:29, 12 May 2007 (UTC)

I have edited Wikipedia for many years, you a few weeks. I have never enountered your interpretations of policy. The policy is not being changed, only clarified to avoid misunderstandings by new editors.Ultramarine 18:39, 12 May 2007 (UTC)

Your ad hominem attack on me does nothing to support your proposed addition of wording. Neither have you presented any evidence that the existing policy lacks clarity. All I see is a proposed wording that adds new ambiguity with a loophole, a subjective definition of the term 'notable opinions and views', which could encourage editors to selectively quote from primary material, as original research, causing a risk of loss of credibility. SaltyBoatr 19:11, 12 May 2007 (UTC)

I count three other editors on this page who supports me, [5][6][7] No one supports you. Thus, unless new arguments, I will shortly restore the text.Ultramarine 19:16, 12 May 2007 (UTC)

I am not seeking support, I am asking you questions. You, who is making this proposal, should explain how we can quantify the ambiguous term 'notable opinions and views', and you dismiss with a reply: "not been a problem". I asked you to show that there is an existing problem with existing policy that needs to be clarified, and again, no answer. I ask you to say why existing policy is not good enough. I ask you to explain how the risk of loss of credibility to Wikipedia makes your proposal worthwhile. Would you answer? Do you have consensus to ignore my questions? I might be able to support your proposal if you answered my questions about it. Your counting of 'three supporters' is odd. Did you count [8]this opponent? SaltyBoatr 20:31, 12 May 2007 (UTC)

Already explained earlier. On the other hand, you refuse to even discuss the destructive implications of your intepretation which is not very constructive for bettering Wikipedia.Ultramarine 21:35, 12 May 2007 (UTC)

It is a rhetorical trick throw up a lot of chaff, hoping people will lose patience, then claiming 'already explained earlier'. I've looked. You have not actually answered my questions. SaltyBoatr 23:02, 12 May 2007 (UTC)

I have cited Jimbo Wales from the official policy above. Now, you are arguing that Wikipedia should delete all the text in the articles about Presidential candidates that quote the platforms of the candidates? Or that Wikipedia should delete all the criticisms citing Amensty from the Guantanamo Bay detention camp article? Or all citations from the Bible in the Armageddon article?Ultramarine 10:43, 13 May 2007 (UTC)

Vote

Shall the following section be added to the article?

Notable opinions or views, not stating a fact, can use as a source the publication expressing this view. Such views must pass Wikipedia:Undue weight

• Oppose - At least I oppose at present, pending answers to my questions. We need some method to quantify 'notable'. If we agree to use WP:Notability, that is: multiple credible secondary sources, then those secondary sources can be used instead of the primary material. Also, no need to encourage more use of self published and primary material. And further, the existing policy already covers this question adequately, and this new sentence fixes nothing while adding ambiguity. SaltyBoatr 21:21, 12 May 2007 (UTC)

Already discussed for several days now and you are the only one opposing.Ultramarine 21:36, 12 May 2007 (UTC)

Ultramarine just rushed and put[9] his proposed text into the article prematurely, before allowing time for a vote. I would revert, pending results of the vote, if not for his 3RR threat[10] against me. SaltyBoatr 23:10, 12 May 2007 (UTC)

What does it mean? SlimVirgin ^(talk) 03:22, 13 May 2007 (UTC)

• Um... why would we need to have a vote, instead of just the discussion up above? -Amarkov moo! 04:02, 13 May 2007 (UTC)
  • Perhaps because the discussion above is predominantly two editors who seem intent on disagreeing with each other despite what they might say. I don't see the discussion in its present form resolving anything. I don't think a vote is the solution, though. JulesH 07:45, 13 May 2007 (UTC)

Since Slimvirgin asks what it means, it obviously needs to be clarified. How about Statements describing opinions or views, not stating a fact, can use as a source the publication expressing this view. Examples include a party platform as a source for the view of the party or a report by Amnesty when describing criticisms by this organization. Such views must pass Wikipedia:Undue weight. Prose attribution should always be used in such cases.Ultramarine 09:35, 13 May 2007 (UTC)

Legislate, procrastinate or delineate?

I feel sure that the people spending time on this are doing a combination of the three, with a heavy if accidental focus on the first. I think you'd make more progress if you worked out what is already the dividing line between yes/no, and take it from there. This rather than trying to divine new such dividing lines, carving up things that were previously not carved up and generally resulting in noone agrees with anyone else. No? Splash - tk 22:01, 14 May 2007 (UTC)

Explicit mention of examples (source code, math, etc.)

This topic is likely to be controversial, so please remember to be civil and assume good faith.

Neither Wikipedia:Attribution nor Wikipedia:No original research make any explicit mention of examples. Many editors (including myself) have thought it appropriate to include examples in articles - many of which (in my experience reading articles on computing and mathematical topics) are not referenced to a reliable source. However, on a plain reading of both policies, examples not published in reliable sources, and statements (aside from obvious logical deductions) about examples that are not published in reliable sources, are both against Wikipedia policy.

Examples of types of examples that editors might wish to include in an article are:

• Computer source code.
  • For example, an editor might wish to state that a piece of source code is not syntactically valid (which is sometimes simple and uncontroversial deduction from the published, reliable source).
  • However, they might want to state that one piece of code is "more efficient" than another, which is a potentially complicated and controversial subject (efficiency can depend on the nature of the input data), and is certainly not a trivial deduction from a reliable source, typically.
  • In the most common case, they might want to state, without evidence, that a piece of source code does a particular thing, correctly - which requires access to a compiler or interpreter to verify it yourself, which might require payment. Also, as is well known, just testing it on some input data does not prove that it works correctly for all input data. Also, in the process of verifying the code, you could risk executing obfuscated malware, and thereby compromising your personal security and privacy!
  • Also particularly problematic could be programming languages for "ancient" computers, for which no modern implementation exists - it is then harder to check if an example actually works, or does what the article says it does.
  • An even worse example: they might want to state that a particular piece of source code is "easier to read" or "easier to understand", which can be extremely controversial and subjective. The latter, at least, I think, should be totally unacceptable, unless it is correctly cited to reliable sources.

• A specimen of text written in a computer language which is not a programming language - for example, JSON.

• Mathematical formulas, equations, or proofs. Proofs, for example, are often not a trivial matter to manually check (that is to say, to go through them and see if they are valid arguments, rather than to check them against sources) - relying as they often do in practice on unstated assumptions, and non-trivial technical knowledge.

I suggest that, because many editors seem to be unaware that attribution and NOR presently apply to examples (or believe that it is pedantry to apply them to examples), we should either (a) make explicit mention of examples, particularly computer source code, without actually changing policy, or (b) change our policies to allow at least some use of examples that have not been published in a reliable source, and/or arguments about those examples.

Opinions?—greenrd 10:07, 18 May 2007 (UTC)

Wikipedia:Scientific citation guidelines makes allowances for just this point. --ScienceApologist 11:16, 18 May 2007 (UTC)

Examples should be cited if they are taken from external sources. However, examples and comments which are obvious can be included by editors without violating WP:NOR: the question as to what is obvious is subjective, but all the natural sciences have well-functioning WikiProjects where advice can be obtained from experienced, non-involved editors. I do not think that this is a general problem, although disputes will always occur and will (hopefully) be resolved by the normal mechanisms. Physchim62 (talk) 11:34, 18 May 2007 (UTC)

To add to this obvious discussion, we just removed most code examples from the Null (SQL) article because of such objections. The examples were included to specifically demonstrate some of the more non-obvious aspects of Null-handling in SQL. We were also told by reviewers that the article needed to be dumbed-down (yes, that was their words) so that laymen could understand it, which included getting rid of *all* code samples. I think a very good test of whether or not code samples work in an article is to test it against the Featured Article criteria, get it peer reviewed, and see what they say. Based on experience, they will tell you that the articles need to be dumbed-down and all source code removed completely; after all, not everyone knows how to program. SqlPac 14:57, 18 May 2007 (UTC)

I didn't see any mention of dumbing it down at Wikipedia:Peer review/Null (SQL)/archive1. Are you referring to some other place? Also, I think some of the code examples are needed here. (Particularly the CASE example, which I especially liked.) Silly rabbit 15:37, 18 May 2007 (UTC)

See FAC Null(SQL). Direct quote from a reviewer: "This reads a bit too much like a programming guide for my taste. You need to dumb it down so that laymen like me can understand it." We thought the code examples added considerable value as well, but there are apparently conflicting opinions here. On the one hand, there are those of us who believe the code examples add value to the article(s); on the other hand, the Featured Article reviewers think these topics need to be dumbed-down and source code stripped. So WTF are you supposed to do in this situation? Keep the examples because they add value (with the understanding you will never be able to achieve Featured Article status under current guidelines), or dumb it down and strip code examples to make it more marketable to the Featured Article reviewers? It seems to me that there needs to be separate paths to Featured Article status. The current guidelines strongly favor biographies, historical events, and places (just look at the Featured Article list of articles), but the rules are set firmly against promotion of technical articles. SqlPac 16:03, 18 May 2007 (UTC)

I see what you mean about a double-standard, and it is unfortunate. This isn't really the place to talk about the de facto realities of FA versus the published criteria. To bring this back into line with the main discussion, Null (SQL) is a perfect example of an article where code is completely appropriate (and even necessary). The snippets of code weren't terribly sophisticated, and they illustrated the main features of Null handling. No citation should even be required for such things, above and beyond a general reference on the topic. These are the sorts of examples that are "Known to those who know", and should be known to those who want to find out about Null handling. Code snippets are more problematic, for instance, when you have ten different dubious implementations of currying written in everything from C# to Io. Silly rabbit 16:30, 18 May 2007 (UTC)

Null is a hard concept for non-programmers (heck, it's a hard concept for some programmers!) I think that some examples, showing correct and incorrect assumptions and uses, would be a good thing, but perhaps put them in a separate section (perhaps a subpage?) with a warning about the technical details therein. It is important that a non-technical reader understand both the usefullness of NULL and danger in misuse of NULL, even if they are not taught explicitly the right and wrong ways. Not the first time I've wanted to use Knuth's Dangerous Bend in a document. htom 17:30, 18 May 2007 (UTC)

The code examples were not overly complex, mostly one-liners. SQL Null itself, while not a ridiculously hard topic, is often misunderstood and the source of a lot of frustration (particularly the three-valued logic). It's all so closely intertwined with implementation that it is difficult to dumb-down the topic, and still simultaneously eliminate the single-line code samples that demonstrate the concepts in question. Harder still would be to try to find published code for 3 or 4 samples and maintain any type of coherency or consistency, since each author for these single statements would have their own "style", their own naming conventions, etc. We haven't even started on the SQL DML article yet. That will be interesting to try to describe everything that each DML statement does while providing no code examples. As for the currying article, if the code samples make up 99.9% of the article, it might be a good idea to eliminate a few. SqlPac 17:42, 18 May 2007 (UTC)

As for the comment below, about Wikipedia:Scientific citation guidelines ("I read it as saying that the statements made about the example need to be obvious to anyone who reads and understands the references"), that apparently is not good enough to reach Featured Article status. Something needs to be changed somewhere; whether it's the Wikipedia:Scientific citation guidelines being changed to state that "the statements made about the example need to be obvious to anyone, period", or the Featured Article process/standards being changed to be more inclusive of technical material. SqlPac 17:48, 18 May 2007 (UTC)

Your suggested "anyone, period" standard would preclude any advanced mathematical content on Wikipedia. There are plenty of people who do not understand even simple arithmetic, let alone anything more complicated. But if Wikipedia is not to be merely a repository of celebrity trivia, it needs technically-nontrivial content. If the FA guidelines are broken, the solution is not to break the other guidelines to match. —David Eppstein 18:05, 18 May 2007 (UTC)

The "anyone, period" standard is what the standard that the WP:FA reviewers are currently using. All I'm saying is that the standards should reflect how business is conducted around here; or something should be changed in the way business is conducted to reflect the lofty ideals in the standards. Either way, a little honesty in the standards would be nice. Currently the WP:FA review process is thoroughly stacked against technically-nontrivial content, and the WP:FA standards might be said to more accurately reflect the "best non-technical content in Wikipedia". Just look at the list of WP:FA articles for verification on this. You have a much better chance of having your work recognized as the "best content in Wikipedia" if you work on a "Brangelina" article or something. The reason being that an article about celebrity trivia can be "dumbed down" to the level where the layman can understand it. With technical content you can only make it "as simple as possible, but no simpler." (apologies to Einstein) SqlPac 18:46, 18 May 2007 (UTC)

So it sounds to me as though you have a gripe with the featured-article process; I don't think the answer is to make the rest of Wikipedia just as bad. I basically came to the conclusion some time ago that the FA process should simply be ignored; it has little relevance to the sorts of WP content that I care about. --Trovatore 19:28, 18 May 2007 (UTC)

You could put it that way ("gripe with the featured-article process"), although I prefer to think of it as a problem with the dishonesty. I think the FA process could be very valuable, and motivate people to work hard to improve the articles they care about. I assume (though I could be wrong), that this was one of the motivations behind adopting the FA process in the first place. The way things are run around here now, though, it's mostly just a big holding Category for peer reviewed Biographies and articles about Geography and Politics -- basically a bunch of "soft sell crap". Anyway, where do you go around here to recommend that they set up a parallel to the FA process for articles of substance? SqlPac 20:19, 18 May 2007 (UTC)

An example of a parallel process: Wikipedia:WikiProject Mathematics/A-class rating. You need a sufficiently active project for this, though; e.g., it might be more problematic in CS than in Math. And it's relatively new, so there isn't a lot of experience with it yet, though. —David Eppstein 21:12, 18 May 2007 (UTC)

Many technical articles simply require examples - period. NULL (SQL) is surely a very good candidate here. Removing the examples or source code snippets makes this particular article completely worthless. (At that point, I wouldn't see any compelling reason to continue working on SQL-related articles, for example.) One cannot easily explain all the details in a verbal way w/o showing some code. As for the references and verification, we have the SQL:2003 standard (for which no full implementation (product) exists. Related to that, there are a lot of CS students and professionals who look at Wikipedia to get an introduction and basic understanding of certain concepts. The only effect that the "dumbing down" would have is that you turn those folks away. I see the following solutions:

1. Ignore FA and concentrate on the quality of the articles - no matter what certain guidelines or policies try to accomplish.
2. Write two versions: dumbed-down (without any helpful content in the end), and one exhaustive with a lot of technical detail and examples)
3. Find another place for the information.

--Stolze 09:14, 19 May 2007 (UTC)

Thanks for the link to Wikipedia:Scientific citation guidelines - I wasn't aware of that. It could be read as covering source code as well, I suppose, but I wouldn't have thought to read that page to find out the policy on source code. Basically, I read it as saying that the statements made about the example need to be obvious to anyone who reads and understands the references. That sounds sensible. So effectively the example can be verified by reading the cited works, even though it doesn't actually appear there in so many words. It's a special case of the subjective boundary between obviousness, which is OK, and synthesis, which is original research.

I am not sure how or whether this should be applied to source code, though, because it is well known that source code of more than a few lines in length often contains one or more bugs, despite looking "obviously correct" at first. I'm not sure that "it's obvious" is enough here.—greenrd 11:58, 18 May 2007 (UTC)

I have replied to you over at Wikipedia talk:WikiProject Mathematics. My recommendation is to leave policy matters like this to the wikipedia project editors. Every subject has its own internal logic; its own notion of when an example is needed and when it isn't. Lumping together source code and math examples is apples and elephants. If you want to make policy decisions of this sort, you should do it project-by-project rather than try to implement a broad "totalitarian" policy of summary deletion. Silly rabbit 12:24, 18 May 2007 (UTC)

I agree with Rabbit. Furthermore, we shouldn't elevate an issues to a policy level without actual cases of contentious situations that have arisen which could not be resolved under exiting policies. I've yet to encounter one concerning examples.--agr 12:41, 18 May 2007 (UTC)

The actual case that prompted me to post this - sorry, I should have mentioned this at the start - concerned source code: an editor went through Boolean datatype and put {{fact}} all over the place - including on examples. I thought this was over the top, but from reading WP:NOR, it looked to me like he was actually right. So I was struck by how strange this conclusion was, and I eventually started this discussion. I appreciate your point though about not unnecessarily starting global policy discussions and I'll try to bear that in mind in future.—greenrd 13:35, 18 May 2007 (UTC)

It really does help to have specific cases. I think Boolean datatype can be worked out under existing policies by the CS community. One problem with an article that compare numerous programming languages is the need for a references for each language. It might help to create a standard reference list or bibliography of major computer languages, with an entry for each, perhaps containing the official spec and a couple of major texts and/or on-line resources. This could be included by reference for such articles. I don't think simple examples that follow directly from the cited works present a problem. When you start making claims like example A "does the exact same thing as" example B, a citation may be needed if it is challenged. I think project editors can come up with reasonable approaches for situations like this. --agr 14:23, 18 May 2007 (UTC)

Regarding "summary deletions", WP:V actually does say "Any edit lacking a source may be removed" - and goes on to quote Jimbo as saying "I heard it somewhere" (i.e. unsourced) types of info should be "aggressively" removed. Maybe this is over-the-top and should be modified. This is kind of one of the issues I'm raising here. WP:V and WP:OR don't currently reflect or even acknowledge these nuances about how Wikipedia actually works - and pointing to WP:IAR is unsatisfactory. It's a little hypocritical if we go around criticising newbies for ignoring WP:OR etc. - and then don't follow those policies ourselves because they haven't been written with science/math/technology examples in mind.—greenrd 13:47, 18 May 2007 (UTC)

Well I agree with the spirit of Jimbo's words, and I would argue against any changes to the stated policies. Even for the purely pragmatic reason of eliminating frivolous or odd-ball edits. Things should be sourced. Period. But there is a grey area involving things like elementary examples. For instance, choosing a particular cubic polynomial to illustrate the cubic formula. This polynomial may or may not be out there in the literature, but it is of such a completely uncontroversial nature that it would be mean-spirited and pedantic to demand a reference. This is all covered (somewhere) in the Wikipedia science article guidelines. I think the blanket "Everything must be sourced" dogma gives editors something to point at when they revert strange OR-type edits (physics is awash with this sort of phenomenon). The finer points are (perhaps deliberately) buried in the Wikipedia documentation, so that more established editors can see what is actually expected of them. Silly rabbit 14:08, 18 May 2007 (UTC)

Wikipedia must not require that examples come from published sources. I can't state this boldly enough. Examples are, in many or even most cases, complex enough and creative enough that copyright applies to them. Wikipedia therefore cannot use examples that are taken from published sources because doing so would be a copyright violation. This applies particularly to example source code (and, yes, I'm aware of the problems of this -- one of the articles I've worked on in the past is spinlock which, when I first started working on it, contained an example that not only didn't work but gave a completely incorrect impression about a significant aspect of the subject). JulesH 13:10, 18 May 2007 (UTC)

I don't think that applies to all examples, not even all complex and creative ones. Just as we can explain a published theory in our own words without violating copyright, so we should be able to reword a published example in our own words, at least in some cases, without violating copyright. Having said that, source code is a difficult one in terms of "rewording", of course, because modifying it can easily introduce bugs. —greenrd 13:35, 18 May 2007 (UTC)

It's not an issue in the way you have framed it, see Copyright#Idea-expression divide. An example isn't copyright-able, but the expression of it as it occurs in a printed work is. (See also, of course, Wikipedia:Citing sources which gives more stringent guidelines on attribution for inclusion into an encyclopedia.) What would be an issue is if non-free source code were published on a Wikipedia page, but that isn't what you are talking about. Silly rabbit 13:45, 18 May 2007 (UTC)

WP:IAR is indeed relevant, translate as use your head if preference to relying on rules. This could get ridiculous, consider that every mathematical diagram is actually an example. So in fact all the mathematical diagrams in Wikipedia:Featured pictures/Sciences, break NOR. Yet due to copyright restrictions we can't use other peoples images. Ruthlessly applied this would leave mathematical content devoid of images, a very much poorer place. --Salix alba (talk) 14:18, 18 May 2007 (UTC)

I think we're in agreement about that. I was just pointing out that the legal flaw in the idea that such-and-such is copyrighted, and therefore cannot be included in an encyclopedia. To use your example, commutative diagrams aren't copyrighted, although particular images of them may be. My indication of citing sources was to show that Wikipedia's policies in fact go beyond the mere requirement of posting no copyrighted material. Attribution is a stronger constraint. Of course, most diagrams (that I have seen) are either completely standard, or already well referenced. Silly rabbit 14:55, 18 May 2007 (UTC)

Even core policies need to be applied with common sense. Well-chosen examples illustrating and exemplifying the topics and definitions in an article enhance the value of the articles. Imagine, for example, removing the examples from the article named Dependent clause. It would become utterly incomprehensible, except to those readers who already know quite well what a dependent clause is – and therefore don't need the article anyway. Requiring the examples, in all cases, to be taken from published reliable sources, overshoots the mark. Mathematical examples are not essentially different in this respect from linguistic examples. Many cases I've seen of source code in articles, however, can in my opinion not be considered to be examples in that sense. In some cases it is not an illustration but a full implementation of a fairly complicated algorithm, and often wrong at that. In my opinion this then does qualify as "original research" if unsourced, and should be removed. In other cases, we have an indiscriminate repository of implementations. See, for instance, the ridiculous list of "examples" in the article on Currying, many of which, moreover, miss the point, as they show partial parameterization or whatever, but not currying.  --Lambiam^Talk 14:22, 18 May 2007 (UTC)

Source code is just a form of text. Source code is a way of communicating algorithms to both people and computers; it is, therefore, qualitatively little different from a paragraph of English text. It is no more reasonable to require that source code be attributable to some other source than it is to require that we lift paragraphs wholesale from our sources. The algorithm described by the source code must be sourcable, though, of course. —David Eppstein 16:11, 18 May 2007 (UTC)

I totally agree with David Eppstein's comment: "It is no more reasonable to require that source code be attributable to some other source than it is to require that we lift paragraphs wholesale from our sources." There are cases in which the exact original wording is useful and should be kept, such as the first two examples in Hello world program. However, quoting source code must follow the same rules as quoting other textual sources, and suffers from the same problems — too much quoting is plagiarism, and the incorrect attribution of quoted material is dangerous. This is particularly bad for quoted source code, because (IMHO) if an English quotation has a typo, someone will add [sic] and move on, but if an example program has a bug, it's very tempting to fix it, thus getting the program out of sync with its attribution.

For example, if in Tiny Encryption Algorithm we gave the original authors' original implementation, it would be buggy. But if someone innocently added parentheses to correct and clarify the order of operations, it would no longer be "the original authors' original implementation", and we'd have a factual error in the accompanying text. However, by coming up with our own implementation, we avoid these sticky issues, and, besides, we're free to make the example as clear and pedagogical as needed, rather than being forced to stick with an inferior example just because someone with a book deal thought it was better. (Note that published source code is not exempt from errors.) --Quuxplusone 17:02, 18 May 2007 (UTC)

For astronomy articles, this has been a problem when people calculate their own ephemereses for objects and then put them in Wikipedia articles. In my opinion, this is dangerous, especially for some of the more creative derivations, because the derivations rely on the individual contributors' derivations and calculations, which could be faulty. I would support a policy that explicitly states that such calculations should not be included in Wikipedia. However, I would suggest allowing for exceptions that are relatively simplistic (such as for one-step calculations that can be performed by secondary school students). Dr. Submillimeter 17:43, 18 May 2007 (UTC)

Your point about ephemerides is well taken. But I think, in the context of articles describing algorithms, it is often useful to present worked-out examples of the algorithm's steps on a simple input. Such examples may not be present in the source material (and if they were present, copying them could be considered plagiarism) but they go well beyond "one-step calculations". And I think "a one-step calculation that can be performed by a secondary school student" may again be too restrictive — it precludes situations in which the algorithms being described are built from simpler algorithms described elsewhere. However, each step should clearly follow the algorithm as described. For instance, in Størmer's theorem I included an example that involves the solution of 8 Pell equations; it would be inappropriate in that article to describe how to solve Pell equations but their solution is necessary as part of the example. Maybe the difference between the algorithm examples and the ephemerides is that, in the algorithm example case, the example is presented only to show how the algorithm works, while in the ephemeris case the result of the calculations is itself an important informational component of the article, and therefore needs to be sourced more carefully? —David Eppstein 18:16, 18 May 2007 (UTC) P.S. I should add that in the Størmer's theorem example, the results of the calculation are sourcable. It is only their presentation as a worked-out example of that algorithm that is not. —David Eppstein 18:32, 18 May 2007 (UTC)

I agree that source code is a form of text. The problem is not in the source code being or not being text, but in its being used in a claim such as "the following program implements the Bellman–Ford algorithm in C". (See Bellman-Ford algorithm#Implementation.) Apart from the fact that (in my opinion) presenting the C program does not add any encyclopedic value to the article, it also raises the question how we can verify the claim that this specific program implements the algorithm. (And if you take the description of what the Bellman-Ford algorithm is strictly literally, it actually doesn't, since the given C function does not compute paths.) This goes beyond being an example. The last sentence of this edit shows that this is not a sudden qualm I have.  --Lambiam^Talk 19:37, 18 May 2007 (UTC)

We had some extensive debates about the use of source code examples over at WikiProject Computer science. The general consensus seemed to be that including a source code example in an article about a language made sense. But wholesale inclusion of (for example) algorithm implementations was pointless:

• How do you choose the language(s) in which to present implementations (and if you don't make such a choice, how do you prevent the implementation examples from overwhelming the content of the article)?
• How do you verify that the implementation is correct? As you've pointed out, there's a factual claim that "the following program implements the Bellman–Ford algorithm in C" - but is that claim actually true?
• Wikipedia is not a source code repository. There are other venues for implementations and source code.

Our guideline for code samples is that actual source code should only be used where it is really necessary (and preferably with a reference), and that in all other places we should favor pseudocode over actual code (it's language neutral, much more flexible, and tends ot be more accessible to lay readers).

--Allan McInnes (talk) 04:59, 19 May 2007 (UTC)

Reply to Allan's points: (1) You choose programming languages the same way you'd choose whether to write an article in British English or American English — the first editor gets to pick, and any editor who cares enough can change it later, which may be a good thing (if it makes the article clearer) or a bad thing (if it's a pointless change). You keep lists under control by pruning them regularly, the same way dedicated editors already do in hundreds of non-CS-related articles.

(2) "How do you verify that the implementation is correct?" The same way you verify that the algorithm description is correct; i.e., generally speaking, you can't. Remember, please, that source code is not magic; it's just a highly regimented form of communication. If someone actually wants to see what the algorithm is, then source code is the fastest and easiest way to do that. You could write out the algorithm in compilable pseudocode, or vague pseudocode, or English, or pig latin (in descending order of usefulness), but you'd be introducing ambiguity and uncertainty to the article, and that's bad. Re your example, consider how you'd verify a claim that "The following pseudocode expresses the Bellman–Ford algorithm".

(3) We agree; WP:NOT a source code repository. Likewise, WP:NOT a repository for English text. However, it is an encyclopedia, and therefore it ought to contain the sorts of things that you'd find in an encyclopedia — things that will help readers gain a better understanding of the topics of its articles. WP:NOT a repository for English text, but if that text is useful in our mission, we should keep it.

A general remark: The situations for math, CS, and physics articles are not remotely similar, so it's unfortunate that this discussion is attracting all three camps. In my comments, I'm talking only about source code in CS articles, not formulas in math articles or crackpot theories in physics articles. Really, IMO, these discussions should be taking place at the appropriate WikiProjects, or on the talk pages of real articles where some kind of problem (not purely hypothetical) has been identified. --Quuxplusone 09:51, 19 May 2007 (UTC)

Allan was only reporting on the consensus that emerged from the discussion at WikiProject Computer science about wholesale inclusion of algorithm implementations, and you yourself state that these discussions should be taking place at the appropriate WikiProjects, and yet you react to Allan's report as if he stated something that urgently needs to be contradicted or amended.

With regard to examples, I don't see how or why the situations at the different projects are so radically different that a joint discussion is necessarily without value. Whether the topic is in the realm of maths, linguistics, or ethics, the same issues with examples arise, and why couldn't the principles that apply to one (such as "use common sense") work as well for the others?  --Lambiam^Talk 10:38, 19 May 2007 (UTC)

As Lambiam already pointed out, my comments were merely a description of the prevailing consensus at WPCS, and the rationale for it. However, in response to your specific comments:

1. Choosing a single language inevitably results in complaints about favoritism, and endless arguments about why language A would be a better choice than language B (or C ;-) ) - I have seen this happen many times. If you don't pick a single language, then you end up with everybody and their dog adding implementations (sometimes multiple, different implementations) in their favorite language. The resulting list of implementations overwhelms the rest of the article. Pruning the lsit is possible, but is (a) a never-ending task, and (b) difficult to justify since there is no established list inclusion criteria (cue complaints about favoritism, POV, etc.)
2. You verify pseudocode by checking the citation provided as a source for the pseudocode in question. Texts and journal articles on algorithms typically include a pseudocode definition of the algorithm, which is citable. The implementations being added to the article were generally original implementations (certainly, I never saw a reference given). If a particular algorithm was originally provided as source in a particular language, then it might be reasonable to include the (cited) source code.
3. By the same token, the criteria for the inclusion of text and source code should be the same. Uncited original implementations do not strike me as particularly verifiable.

--Allan McInnes (talk) 07:06, 24 May 2007 (UTC)

Original source code is only as unverifiable as original text -- you can read it and check that the ideas in it are verifiable. Yes, it is a type of writing that requires an expert to read, which is unfortunate as it limits the number of people who can verify it, but there is nothing about source code that is inherently unverifiable. As the source code should be (essentially) restating the existing content of the article in a different form, I do not see why it should require any additional citations. As an example, the source code in spinlock can be verified by anyone with an understanding of x86 assembly language and access to the textbook that is used as a reference for other claims on the page.

Non-original source code, like non-original text, would be a copyright infringement, so original is the only kind we can use.

Regarding language choice, I see no reason the rule currently used for national variants of English shouldn't work: whoever writes an example first on a specific page gets to pick which one is used, unless there is a good reason to choose a different one for that particular page. I would support a style guideline that contained a suggestion that only one language be used for any example code and used this method to pick the language. JulesH 10:02, 24 May 2007 (UTC)

To add to the language choice discussion, it makes sense to me that source code examples should attempt to be platform-neutral when possible; preferably with the capability to run on an accepted standards-compliant implementation of the language of choice (for instance, an ISO standards compliant C++ or SQL platform). Of course some articles might need to address platform-specific features, and exceptions should be made for those, but on the whole platform-neutrality/standards-compliance of code examples should be a goal, regardless of language choice. SqlPac 17:24, 24 May 2007 (UTC)

Again, I'm not looking to have a debate here. I'm merely reporting the guidelines that WikiProject Computer science ended up agreeing upon, and the rationale for those guidelines. Please note that I'm not saying that source code should not be included at all (as my original comment indicates, there are certainly places where it makes sense). The concern was with algorithm articles that became filled with sample implementations (take a look at earlier iterations of the Quicksort article, for example), in many cases not a (as you put it) "restatement of the text in another form", but implementations containing all sorts of optimizations, or intended to show off features of a particular language.

Just to avoid further confusion, the specific guidelines in question are:

Samples of actual source get included in articles for a variety of reasons, although the most typical reasons are to demonstrate the "look" of a particular language, to provide examples of language-specific constructs or features, and to provide examples of algorithms not easily expressed in pseudocode. While there's nothing inherently wrong with including sample code, excessive amounts of it can detract from the content of the article itself; avoid writing sample code unless it contributes significantly to a fundamental understanding of the encyclopedic content.

and

* Sample implementations of algorithms are fine, but every algorithm article should include a pseudocode description of the core algorithm when possible, so that anyone can understand how the algorithm works.

* The sample should use a language that clearly illustrates the algorithm to a reader who is relatively unfamiliar with the language — even if you believe that the language is well-known. To remain language-neutral, choose languages based on clarity, not popularity. Avoid esoteric or language-specific operations.

* Source code implementations must be compatible with the GFDL (which is incompatible with the GPL).

* Multiple source code implementations are not appropriate unless they contrast specific aspects of the code and that contrast is important to the encyclopedic content of the article. If possible, accentuate differences by providing the alternate implementation in the same language as the original.

If you wish to discuss these guidelines further, I suggest we move the conversation to either the project talk page, or the talk page for the guidelines themselves. --Allan McInnes (talk) 20:57, 24 May 2007 (UTC)

Recap

Summary findings: This discussion thread covers some fairly subtle, yet significant distinctions. The following is an attempt to summarize and condense the core issues here:

Summary removal of all unsourced "examples":

• 1.1) Under a good-faith reading of existing WP policy, is it correct for any contributor to summarily remove any and all illustrative examples, diagrams or derivations that have not been properly attributed verbatim from a reliable source?;
• 1.2) If "yes" to 1.1) above, does this warrant a clarification of existing general WP policy on the grounds that such an interpretation leads to absurd outcomes;
• 1.3) If "yes" to 1.1) above, is the status quo nonetheless adequate for addressing abuses of this good-faith interpretation, or over-application by new contributors;
• 1.4) If "yes" to 1.1) above, is this a matter best resolved by adherence to the established norms of each discipline and wiki sub-project individually;
• 1.5) If "no" to 1.1) above, ... then what?

Peer-review of unsourced "examples":

• 2.1) Assuming summary removal of any and all "examples" is not justifiable under a good-faith reading of WP policy, what are the standards for: a) inclusion and b) review of examples?
• 2.2) Is there is a "threshold test" for distinguishing an "example" from "original research" and "novel claims"? on what basis should contributors determine if that threshold has been passed?
• 2.3) Is there an "accuracy test" for distinguishing "correct" examples from those that are questionable on the grounds of: a) accuracy; b) relevance; or c) superfluous content?
• 2.4) Is there a "distinction test" that contributors must first apply in order to differentiate "examples" from "clearly oddball contributions from crackpots", or the ordinary sort of content that are obvious candidates for immediate removal? (e.g. vandalism, WP:LIVING violations, obvious nonsense);

Intellectual property:

• 3.1) What limitations or policy considerations arise when we consider copyright restrictions?

Since this discussion merges these distinct and arguably separate considerations, the preceeding is intended strictly to clarify, and not re-formulate, the views stated previously. Consequently, feel free to modify or extend any of what's been written here so far, if there are any mistakes or omissions.

Additionally, it would seem a "yes" answer to 1.1) clearly does not justify a summary removal of all "verbatim-unsourced" examples from WP. As far as I can, tell no contributor to this discussion has yet supported such a conclusion. Thus, if there *is* a good-faith reading of policy that seems to lead to this (so far unsupported conclusion) then that issue should be addressed (or skirted, as some have implied) separately. My personal view: handle it on a project-by-project basis, and indicate this general approach somewhere in the core policy documents. dr.ef.tymac 22:35, 18 May 2007 (UTC)

Regarding Summary removal of all unsourced "examples", I agree with 1.2 - there needs to be a clarification to existing WP policy, and quite possibly to WikiProject policies to prevent (a) the loss of examples that add considerable value and (b) the overuse of examples as a substitute for good, quality text. I also sort of agree with 1.4 - WikiProjects and disciplines should also clarify their policies on this matter. A clarification on the amount of code that should normally be included in an article, and the size of code samples that require sourcing, would be phenomenal. For instance, it doesn't make a lot of sense (in most cases) to try to source a single code fragment like: "i = i + 1". Maybe something like code fragments longer than 4 lines should be sourced and source code/examples should normally not make up more than 20% of an article, or something similar. I strongly disagree that the status quo is acceptable.

Regarding Peer-review of unsourced "examples", I would propose that a possible threshold test might include aspects such as length of the code sample, and the amount of space the code sample takes up in the article as a whole. As for the accuracy test, perhaps the relevant WikiProjects could set up a validation system that would include validations by other contributors. For instance, if I said the expression "COALESCE(NULL, 0)" would return "0" in SQL, other contributors could validate (or invalidate) my claim as a sort of assurance for non-technical readers and others who might not know whether the example is correct or not. You could rate the article "All Samples Cleared" (apologies to Biz Markie) or something.

As for Intellectual property, I think you need to ask the lawyers. Although I doubt that something short of a complete algorithm (not in the public domain, of course) would be copyrightable. We could try to copyright/patent "i = i + 1;" but it's doubtful it will succeed :) SqlPac 23:23, 18 May 2007 (UTC)

I'm still neutral as to (1.1). There are things like Wikipedia:Scientific_citation_guidelines#Examples, derivations and restatements which clearly indicate that there is some flexibility in when something needs to be sourced (verbatim or otherwise). I suppose it is possible for an editor to miss this on a good-faith reading. But I would probably not interpret any summary actions based on such a reading to be acting in good faith. It's probably happened to most of us at one time or another: getting tagged {{OR}}, and then the ensuing edit war over: The reference you provide doesn't say it precisely that way. It's all usually because of some agenda or another, and so not in good faith (in my opinion).

As for (1.2), I'm not sure. You pointed out Jimbo's reason for the NOR policy, but you didn't make it explicit. Here is his quote in full:

"The phrase 'original research' originated primarily as a practical means to deal with physics cranks, of which of course there are a number on the Web. The basic concept is as follows: It can be quite difficult for us to make any valid judgment as to whether a particular thing is true or not. It isn't appropriate for us to try to determine whether someone's novel theory of physics is valid; we aren't really equipped to do that. But what we can do is check whether or not it actually has been published in reputable journals or by reputable publishers. So it's quite convenient to avoid judging the credibility of things by simply sticking to things that have been judged credible by people much better equipped to decide. The exact same principle will hold true for history." (from WP:NOR)

Having a non-nuanced WP:Citing sources and WP:NOR policy in plain sight does make it easier to deal with some of these issues. But the nuances are there, and more experienced editors need to be aware of them before implementing any sort of policy decisions from on-high.

(1.3) The status-quo seems fine to me. It's not a perfect system, but it seems to work pretty well for most issues. A recent example of this happened at High-Frequency Gravitational Waves. This was a case-study in what could go wrong: It was a well-thought-out, apparently well-referenced, exceptionally well-organized article. Unfortunately, it was a complete load of pseudoscientific nonsense (with portions plagiarized verbatim from elsewhere). Luckily, the system worked: with enough editorial pressure, the author decided not to fight the AfD. (IMO it was the plagiarism that made him cave.)

(1.4) I can't speak for all subprojects, but it seems that the editors over at Wikipedia:WikiProject Mathematics are pretty good at keeping up standards. We have a ways to go as far as bringing articles into decent shape, but I think that you'll find that with any other solution the negative consequences are going to outweigh the positive. Qualified people need to be the ones making the editorial decisions on whether an example is needed or not. User:SqlPac's recent troubles are an excellent illustration of this. The "dummies" at the FAC review couldn't understand the article, so they told him to remove the contentful parts of it. Not good. Let the people who are qualified make the editorial decisions within a particular subject-matter.

(2.*) If only there were a good way to address these universally. Sadly, it seems that it must be done on a case-by-case basis. Also, I think there is a breakdown of some of the conventional mechanisms for fixing problems like these. {{fact}},{{OR}},{{dubious}}, and {{cleanup}} tags don't seem to generate much interest from other editors. I've stopped using them and resorted to fixing problems when I see them, if possible. Anyway, I think the way to do it is to try to preserve the spirit of the author's original intent (insofar as can be done without advancing a POV). So far it hasn't led to any serious complaints. Silly rabbit 03:57, 19 May 2007 (UTC)

I think that the {{fact}},{{OR}},{{dubious}}, and {{cleanup}} tags help the reader clue in that fact that the source is questionable. Carl 14:42, 24 May 2007 (UTC)

Images/Diagrams. WP:OR has the following to say:

Pictures have enjoyed a broad exception from this policy, in that Wikipedia editors are encouraged to take photographs or draw pictures or diagrams and upload them, releasing them under the GFDL or another free license, to illustrate articles. This is welcomed because images generally do not propose unpublished ideas or arguments, the core reason behind the NOR policy. Also, because of copyright law in a number of countries and its relationship to the work of building a free encyclopedia, there are relatively few publicly available images we can take and use. Wikipedia editors' pictures fill a needed role.

-- Jheald 12:31, 19 May 2007 (UTC)

Just a point, but one could argue that if a code sample in an article had to be removed, then the accompanying flowchart that some contributor creates should also be removed. I'd personally prefer to see strong guidelines concerning what is and is not acceptable, while leaving open the possibility of the occasional article-by-article exceptions rather than weak guidelines which make almost all articles an exception. It would definitely make it easier for contributors, as well as non-technical editors, to make better quality contributions. SqlPac 21:32, 19 May 2007 (UTC)

There appears to be nearly universal, violent agreement that the answer to 1.1 above is an emphatic NO, based on my reading of the commentary. The answer to question 2 can't be determined, since there have not been enough example cases presented to talk about. Every mathematics example I've seen has been appropriate to the article. Most programming language examples that I've seen seem appropriate, if a bit tedious and over-long; my eyes glaze over. The only bad example I've ever seen was an article on some obscure fractal that had 23 different programming examples; worse, what made the fractal important was not the algoriithm that generated it. linas 15:25, 20 May 2007 (UTC)

My feelings about this are:

1.1) Under a good-faith reading of existing WP policy, is it correct for any contributor to summarily remove any and all illustrative examples, diagrams or derivations that have not been properly attributed verbatim from a reliable source?

No. The policy does not encourage the wholesale deletion of material, whether sourced or otherwise. A fair argument could also be made that the above-mentioned exclusion for images also applies to examples. Both exist to make clearer text that is already in the article, both have similar copyright issues if copied wholesale from another source, and both are difficult to "rephrase" in a manner that renders them an original expression of an idea (because they both are already an expression of an idea).

1.5) If "no" to 1.1) above, ... then what?

A contributor who believes an example is incorrect should challenge any editor who wishes to retain the example to show that it is correct by using techniques that are described in a reliable published source. Requiring the use of an unattributable technique would, effectively, mean the example is original research.

2.1) Assuming summary removal of any and all "examples" is not justifiable under a good-faith reading of WP policy, what are the standards for: a) inclusion and b) review of examples?

An example should not be included if it makes a point that is not attributable (this would be original research); it should be the case that anybody who has thoroughly understood the article text could understand the example. If any editor believes an example to be incorrect, he should discuss the issue on the talk page and correct the example if nobody disputes his assertion. A formal review process might be a good idea; perhaps a {{unreviewed example|field of study}} tag could be added below the example (which would add small text reading "the above example has not been reviewed for accuracy")

2.2) Is there is a "threshold test" for distinguishing an "example" from "original research" and "novel claims"? on what basis should contributors determine if that threshold has been passed?

If the example does not rely on any knowledge that is not in the article, and all information in the article is correctly attributed, I don't believe the example can be original research. If the example uses a technique that isn't described in the article (and isn't common knowledge among those reasonably well-read in the domain), however, it should include a citation to a reliable source that uses the technique.

2.3) Is there an "accuracy test" for distinguishing "correct" examples from those that are questionable on the grounds of: a) accuracy; b) relevance; or c) superfluous content?

I think this is basically down to the individual editors of a page, or perhaps members of a relevant wikiproject to decide. I certainly can't think of anything that would be globally applicable.

2.4) Is there a "distinction test" that contributors must first apply in order to differentiate "examples" from "clearly oddball contributions from crackpots", or the ordinary sort of content that are obvious candidates for immediate removal? (e.g. vandalism, WP:LIVING violations, obvious nonsense)

I think this is covered by my answer to 2.2: if the example relies on some technique or other information, then that information must either be common knowledge or be in the article and attributed somewhere reliable. This should rule out anything that's obviously crackpot.

3.1) What limitations or policy considerations arise when we consider copyright restrictions?

We can't lift examples from somebody else wholesale, because they are an expression of an idea rather than an idea in itself. Therefore, the idea of attributing an example to a reliable source seems to me to be totally wrongheaded. The ideas inside the example, however, must be attributed. JulesH 17:42, 20 May 2007 (UTC)

In regards to CS articles would it be so hard to hold that all source code come from a publish source. Given the number of books that are out there on the market, including just code books and not text book, I don’t think this would be that hard of a burden. I do think there is a separate case to be made for math and physics articles, but here too is a need for more documentation. One last thing that came to mind, because proofs and code can be implemented differently, any claim of superiority of code should be look on with extreme sceptsisum. Carl 20:56, 22 May 2007 (UTC)

I think, in many instances, it places an unecessary burden on contributors. There are plenty of single sentences throughout Wikipedia that are unsourced, even in WP:FA articles. I don't see how forcing people to source "i = i + 1", or other extremely short code snippets would add any value. An example I think parallels this one is the line "1 + 1 = 2", which is not sourced in Addition. Neither is "2 + 2 = 4". Of course those would be considerably easier to source than most code snippets if it were really necessary. I think there has to be a reasonable limit set. As I said, I could see having a rule to require sourcing code of a certain length or complexity, precisely because you get into the issues of WP:OR, etc. SqlPac 02:34, 23 May 2007 (UTC)

Please reply there, not here, if you wish to contribute.—greenrd 11:20, 18 May 2007 (UTC)

Source code and mathematical explanations are somewhat different; I agree that complete segments of code should be copied to a wikibook and linked to by the article rather than being dumped into an article. Mathematical examples are, however somewhat different -- a well placed explanatory equation can be very useful in illustrating topics to people who aren't implicitly mathematics graduates or others. However, if those become unsightly -- i.e matrices, integration and others that are huge.. then by all means, remove them. ♥♥ ΜÏΠЄSΓRΘΠ€ ♥♥ ^{slurp me!} 17:00, 1 June 2007 (UTC)

That we should include a hello world example is bordering on blatantly obvious. As for other code examples, perhaps in most cases depending on the particular language and example. Mathmo ^Talk 01:25, 20 June 2007 (UTC)

FAQ page

There has been a bit of discussion at WP:RS about the duplication that exists between Wikipedia:Reliable sources/examples and the Wikipedia:Attribution/FAQ. It has been pointed out that we could get into a situation where the examples and advice given about them could (once again) conflict; and the suggestion has been made that we combine them into one single examples/FAQ page. I think this is a good idea, but we need to discuss this further and agree on details (For instance, do we merge it all into the ATT/FAQ? do we merge it all into RS/Examples? do we move and merge both into a new page? etc.)

For the record, the FAQ page was originally created when we were expecting this policy to replace WP:RS (along with V and NOR). It primarily consisted of bits from WP:RS and WP:RS/Examples that did not really belong in a Policy Statement... stuff that was more advice than policy. The idea was that we would work on it more, and eventually promote it to a new (purely advice) guideline that would help people answer those commonly asked grey area questions about sources and their reliability.

What do people think about re-combining these advice/examples pages... and if it is considered a good idea, how should we go about it? Blueboar 12:48, 31 May 2007 (UTC)

I would merge into the subpage Wikipedia:Reliable sources/examples. It is better to save the material than to reignite the question of ATTFAQ's status. I would also consider merging the result into WP:RS, which is presently a handful of similar examples and a list of links. Septentrionalis PMAnderson 18:05, 31 May 2007 (UTC)

Too many footnotes

As an example that anything can be done to excess, I present the current text of Eyes of the Insane, in which five consecutive sentences, in the same paragraph, are sourced to the same footnote with the same source (an on-line text of an interview). This may well be the best available source, given the nature of the article; but it only needs to be cited once. The proud author has nominated this for FA at Wikipedia:Featured article candidates/Eyes of the Insane and some editors appear to believe that this sort of thing looks academic. They have also been insisting on absurdities at GA, and thereby contributing to its endemic problems. If you think I'm wrong, do tell me; if you agree with me, could you tell them, with more patience than I have managed to muster? Septentrionalis PMAnderson 18:15, 31 May 2007 (UTC)

I'm afraid the problem has gone too far to be easily fixed by just talking to a few people. I've worked on articles where editors have insisted that something is "unverifiable" if you need to actually read the only and only source given for the entire article to determine whether some particular fact is in it. I'm sure that in that case, some editors will only be happy if every sentence ends "^[1]". JulesH 21:07, 31 May 2007 (UTC)

I include a version link; the article is being improved. I'm not sure if GA can be, although there are promising discussions at Wikipedia talk:Good articles and Template talk:Grading scheme#Good articles.

I am infavor of adding citations in general (too many articles don't have any) but you can push this to absurdity. To some extent the same thing is going on at Freemasonry, which is undergoing a GA review. There are reviewers that want citations for every third word... even where it is clear that the section is a summary of some other article that is fully cited. I thought there something in one of the guidelines that says you don't have to repeat citations when summarizing what is more fully explained in another article... especially when the section in question clearly starts with a (see Main article: XXX) tag. Blueboar 18:29, 1 June 2007 (UTC)

Can newspaper articles be used to support a political analysis?

I'm currently in a debate at [Talk:May_2007_RCTV_protests] about the suitability of newspaper articles for supporting a political claim. An editor provided various articles from different newspapers that state that the members of the Supreme Court of Venezuela are Chávez political allies. My issue with the articles is that although the articles state that, they never quote anyone as being a source for that, they simply state it to be fact. I find that these statements are not verifiable, because the articles never state from where are they coming to that conclusion. Carlos5678 21:52, 1 June 2007 (UTC)

Cost of completeness?

Is the goal of Wikipedia still to become a complete, accurate encyclopedia? What happened to the mention of editing at the cost of completeness? J. D. Redding 21:29, 2 June 2007 (UTC)

I hope so. Re the "accurate" part, some of us have been working to prevent the second sentence from being changed to an unqualified "not ... true"; see the Role of truth page and the section "Second sentence" below. Re completeness, perhaps you could look through the history pages to find older versions of policies and quote parts of them here for discussion. See also Verifiability. --Coppertwig 20:32, 4 June 2007 (UTC)

The second sentence

I suggest reverting the second sentence back to the compromise form that it was in a few weeks ago. Currently it says " The threshold for inclusion in Wikipedia is whether material is attributable to a reliable published source, not whether we think it is true. Wikipedia is not the place to publish your opinions, experiences, or arguments.". In the compromise form, these two sentences were joined with a colon and "we think it is true" was a link to the section of the FAQ about whether Wikipedia cares about truth. The current form stray s too far from the original meaning of the policy, in my opinion. --Coppertwig 22:07, 2 June 2007 (UTC)

User SlimVirgin has changed the presentation of the policy to "The threshold for inclusion in Wikipedia is whether material is attributable to a reliable published source, not whether we think it is true. " without a colon joining it to the next sentence and without a link to the FAQ. I object to this change because it strays too far, in my opinion, from the meaning of the original policy "Verifiability, not truth". If you wish to make this change, one criterion you must meet is to discuss it thoroughly first. Thorough discussion will include: discussing each of the alternative wordings I suggested at the "Role of truth" page; and answering these questions: "Are you really suggesting that an editor who knowingly includes cited, but false, information, is helping write the encyclopedia?" and "I don't understand the phrase "do truth". Perhaps you would be willing to explain what you mean in different terms." (See "do truth" in archives 12 and 13 of this talk page.) I'm also still waiting for a single person who supports the proposed new wording to volunteer to discuss back-and-forth with me long enough to have a reasonable chance of arriving at an agreement on a mutual statement of clarification of positions. Without such discussion you can't make this change from the original "Verifiability, not truth" (or proposed alternatives, including the compromise version you just reverted away from) over my objections. With discussion, we will see. Please carry on discussion on the talk page, not in edit summaries. If you wish to remove the link to the FAQ, you can change to one of the other suggested versions from the role of truth page. --Coppertwig 23:56, 3 June 2007 (UTC)

As another detail, I believe that this is usually intended as the bare minimum, not that everything which is ATT can be used--in which case, just to remove doubt, it should begin "the minimum threshold..." (The preceding unsigned comment was added by user DGG on June 4 at 1:39 UT.)

I agree with DGG that that is the intent of the policy. I support the insertion of the word "minimum" as suggested and consider it an improvement but do not consider it sufficient as a qualification of "not ... true". "A minumum threshold" rather than "the" would be even better but still not sufficient in itself. I've restored the compromise version with the link and the colon. In addition to not addressing the points raised above I note that no reason for changing the colon to a period has been given. --Coppertwig 20:24, 4 June 2007 (UTC)

I object to user Jossi's revert on the following grounds. I don't see any discussion by the user on this talk page. I had specifically requested discussion before any such revert. The various points raised above have not been addressed. No reason for changing the colon to a period has been given. Please restore the compromise version. --Coppertwig 21:28, 4 June 2007 (UTC)

The issue is whether the proposed policy is to contain an unqualified "not ... true" which could colour the whole meaning and purpose of Wikipedia. (In the actual current policy, "not truth" is qualified by the presence of the word "verifiability.") The compromise which was worked out was "The threshold for inclusion in Wikipedia is whether material is attributable to a reliable published source, not whether we think it is true: ..." Note the colon joining two sentences, which clarifies the meaning of "not ... true", as well as the link to "Doesn't Wikipedia care about truth?" Please comment here on the talk page before reverting. --Coppertwig 17:31, 18 June 2007 (UTC)

I object to this revert by Crum375 on the following grounds:

• No talk page explanation in spite of requests for it.
• No explanation why the colon from the compromise version was changed back to a period.
• Previous discussion on the talk page supports the compromise version; I haven't seen discussion on the talk page supporting the version changed to by this edit.
• Objection or strong objection by a number of people has been expressed to the unqualified "not ... true".

Please restore the compromise version. --Coppertwig 18:04, 18 June 2007 (UTC)

I strongly agree with Coppertwig's concerns about material being reverted without discussion. Discussion is essential to establish consensus.

As for the wording in the introduction, I think it's very important that we get this right. Newcomers to Wikipedia (including some journalists), have misinterpreted this in the past to think that Wikipedia doesn't care about truth or normal academic standards.

I prefer the colon to the period - but only slightly, as I still don't think it is getting the right message across. The wording talks about "The threshold for inclusion..." where we are really talking one among many thresholds, or a minimum standard, and I agree with Coppertwig and DGG that this should be addressed.

I would suggest something along the lines of "Material should not be included in Wikipedia if it cannot be attributed to a reliable published source, even if we think it is true. This is because Wikipedia is not the place to judge the truth of new ideas."Enchanter 12:27, 24 June 2007 (UTC)

Note: editors are using this page to threaten stubs with deletion

For everyone involved in this project to note, people such as User:Mikkalai, are using the “Wikipedia:Attribution” page to threatening new page stubs and short articles, having no references, with deletion. Mikkalia states, in reference to pasted-article Questioning (sexual orientation) (from disambig questioning), “find references or else I’ll delete you”.

Is this what Wikipedia has become? If this user would have taken the time to check the comment history for the article or my edit history for June 03, you would see that I did not write this article it is a paste from questioning disambig page because I found the term linked there on many pages but found no main article, but have now found it (Questioning (sexuality)) after being threatened on my talk page. Please do something so that other well-intentioned editors don’t have to be threatened with deletion on their talk page. If not, please instigate a rule, something to the effect that “on all new articles, three or more references are needed, or else they will get deleted”. If this is the new rule at Wikipedia, I think someone should spread the news. --Sadi Carnot 01:21, 4 June 2007 (UTC)

Sadi Carnot appears to be referring to this edit, which looks unobjectionable to me. Mikkelai found an article needing references, tagged it appropriately, and notified interested parties appropriately. Note that "this article may be deleted" is not the same as "I will delete you". I would read that message, not as a threat, but as a suggestion: since you obviously care about the subject of the article, this is what you need to do to make the stub good enough to survive the editorial process here. All moot now, of course, since it ended up being merged with Questioning (sexuality). Even stubs can and should have some kind of reference, especially one such as this which appears to the uninitiated to be simply an odd title for an article about the process of discovering one's sexuality, rather than (as with proper references it clearly is) a term of art used with a very specific meaning in gender studies. —David Eppstein 01:34, 4 June 2007 (UTC)

The article is not the focus of the concern, the concern is that in the Wikipedia I know, a user short on time, can start a legitimate stub, and let others build on it. The basic premise of what makes Wikipedia good is that we all chip in to each article and the combined synergy is what makes the encyclopedia robust. Yet, I see Wikipedia:Attribution turning into a goad used justify deletion on anything without references, thus stifling potential synergy. Is this the correct point of view? --Sadi Carnot 01:46, 4 June 2007 (UTC)

The correct point of view is (IMO) that you have stepped into one of the paramount philosophical debates ever to take place on-wiki: inclusionism vs. deletionism. --Iamunknown 02:03, 4 June 2007 (UTC)

Of course, if Mikkelai were truly a deletionist he would have {{prod}}ed the article rather than merely tagging it. —David Eppstein 02:30, 4 June 2007 (UTC)

Yes, well the top of WP:A it states “the burden of evidence lies with the editor wishing to add or retain the material. If an article topic has no reliable sources, Wikipedia should not have an article on it.” This is all great, but what if an editor (such as myself) wishes to contribute a little by starting a stubby on a topic that could easily have multiple references or is referenced on multiple pages already, etc., but is short on time, i.e. the time needed to dig around for references? If this is the case than at the top of WP:A it should say “in a nutshell, all articles need to have 1, 2, 3, etc., or more references, or it will be deleted.” This is the point I’m getting at. --Sadi Carnot 02:58, 4 June 2007 (UTC)

One option is to create a user subpage (or personal "sandbox") and develop the article there for a while before moving it to the article namespace.

It is indeed true that one result, for better or for worse, of the demand that all information on Wikipedia be attributable is that stubs may be deleted. They, however, may not; there is a WikiProject dedicated to patrolling proposed deletions and the general masses of editors who contribute to deletion discussions. If an article that should not be deleted is proposed in one of these venues, it will likely not be deleted. Similarly, if an article that should be deleted is proposed in one of these venues, it will likely be deleted.

I understand your concern, but it is unfortunate consequence of Wikipedia's size and growth, of certain persons (vandals) wishing to comprimise Wikipedia by inserting false information and other factors that information on Wikipedia is required to be attributable. I hope this helps, Iamunknown 03:11, 4 June 2007 (UTC)

A immediately possible way of dealing with this is for more people to participate in Afd discussions, and in those cases where something is deleted which should not have been, to request at Deletion Review that it be undeleted. In non-controversial cases, there is usually no difficulty in this.

Other steps have been taken already: there is now an automatic method in place to notify the authors of articles that their articles have been proposed for speedy deletion. There is more to do--though an automatic method is not possible when pages have been extensively edited, various proposals are always being discussed at the talk pages for deletion policy, and proposed deletion, and speedy deletion--it would be generally helpful if more people were involved in these discussions also, for we might get some new and better ideas. There are also some people engaged in looking for inappropriately placed tags, and reminding those placing them of the advantages of following the rules.

As Iamnunknow says, the root of the problem is that there is generally several hundred pages deleted each day, and even a 1% error will produce an unfortunate number of removals of worthy articles. However, reducing an error rate in a process that requires human judgement to below 1% is exceedingly difficult. That said, there should be some way of evaluating a stub to say that it is an appropriate but short article. One way that might help is an active project to place at least one reference on each stub for which a reference can be found. DGG 03:50, 6 June 2007 (UTC)

Intent to clarify "direct" verses "indirect"

[I am planning to revises this section of the article to read as below, with new material noted for this discussion]

Wikipedia does not publish original research or original thought

Original research refers to material that is not attributable to a reliable, published source. This includes unpublished facts, arguments, ideas, statements, and neologisms; and any unpublished analysis or synthesis of published material that appears to advance a position. Material added to articles must be directly and explicitly supported by the cited sources.

---

[New material begins here]

---

Note the difference between direct and indirect relationship to an article's topic:

• Directly related indicates the reference is supplying details centered on the subject itself of the article.
• Indirectly related indicates the reference is supplying details centered on something the article happens to mention.

The line between direct and indirect can be difficult to see when a detail is very compelling. A film provides a less fuzzy example. Details about character's personalities in a film are directly related to the film. Details about the actors who play the characters are indirectly related. To relate to details about an actor's life, one steps outside of the experience/topic of the film, seeing the actor for who he/she is. Seeing the personality traits of the actor—although interesting—would be details mostly about the actor, not the film. However, such personality details would be must-have information for an article about the actor.

---

[New material ends here]

---

Note the difference between unsourced material and original research:

• Unsourced material is material not yet attributed to a reliable source. It is unattributed.
• Original research is material that cannot be attributed to a reliable source. It is unattributable.

The only way to demonstrate that material is not original research is to cite reliable sources that provide information directly related to the topic of the article, and to adhere to what those sources say.

---

Proposed by WikiLen 19:46, 5 June 2007 (UTC)

The proposed distinction between direct and indirect is important to help write a good article, but does not really relate to how directly used in the sentence "material added to articles must be directly and explicitly supported by the cited sources." That sentence is about whether a source adequately supports a statement in an article or not; whether the statement is directly related to the topic of the article is another matter. If a source had to be analyzed in light of the reader's extensive knowledge of quantum physics in order to see that it really does support the statement after all, that would be indirect support.--Gerry Ashton 15:03, 7 June 2007 (UTC)

I see that the proposal has been inserted into the proposed policy. I disagree with this addition, because, for purposes of attribution, it does not matter whether the fact being supported is directly related to the topic of the article or not; either the source adequately supports the fact or it doesn't. If the fact is too indirectly related to the topic of the article, that should be covered in something like Wikipedia:Article development, not here. --Gerry Ashton 15:17, 7 June 2007 (UTC)

Hummm, a good debatable point. This article does have a section, "Unpublished synthesis of published material". It addressess issues of "facts" in an article. As noted at the top of this article, "attribution" policy is kind of an attempt to have a one-stop-tells-it-all article. Specifically, this article is combining the policies on "Verifiability", "No original research" and "Reliable sources". My contribution does seem to falls under the rubric of "No original research". My intent is to make it easier for new editors to get that content must be relevant and what it means for it to be relevant. Too often editors argue material is important and therefore must be in the article.

In my opinion, directly and explicitly are not meant to be synonymous or to modify the meaning of each other. Two completely different issues are being addressed. (1) Directly addresses: when content belongs at another article and (2) Explicitly addresses: when claimed content is really a phantom or a synthesis. I base the opinion on this policy:

[avoid] sites that are only indirectly related to the article's subject: the link should be directly related to the subject of the article. A general site that has information about a variety of subjects should usually not be linked to from an article on a more specific subject. Similarly, a website on a specific subject should usually not be linked to an article about a general subject. If a section of a general website is devoted to the subject of the article, and meets the other criteria for linking, then that part of the site could be deep-linked.

A policy that appears in Links normally to be avoided which I take to be a direct parallel to content to be avoided. —WikiLen 16:09, 7 June 2007 (UTC)

---

---

The third paragraph of WP:External links begins "the subject of this guideline is external links that are not citations of article sources." If a fact is in an article, it is good to provide a citation, and if the fact is challenged, the citation must be provided. It is always acceptable to support any fact with a citation from a reliable source, notwithstanding anyting in the External links guideline. If the fact is too indirectly related to an article to belong in the article, it should be removed for that reason; leaving the fact and removing the citation would just be wrong. On that basis, I am removing the change. --Gerry Ashton 16:27, 7 June 2007 (UTC)

I don't follow your reasoning. Don't forget this article is more than just being about citations. It also covers issues of "no original research". —WikiLen 17:07, 7 June 2007 (UTC)

My confusion: Your reasoning seems to suggest that the whole section on "No original reseach" does not belong in this article. You can't be saying that, so I must not be getting what you are trying to say. —WikiLen 17:20, 7 June 2007 (UTC)

Forget I was attempting to draw parellels with the External Links policy. I agree "leaving the fact and removing the citation would just be wrong". The issue is content, not the link. I submit that facts that are "too indirectly related to an article [to be in an article]" is a policy issue that should be addressed in this article. My basis is:

• I find no other policy that explicitly addresses this.
• This is a good place to address it since creating content-that-is-original-research is already addressed here.
• Relevance of facts is a common problem and as such should be addressed in a high-profile article such as this one.

I would like to hear from others on this. thanks, —WikiLen 18:47, 7 June 2007 (UTC)

(Unindent & edit conflict) Let me give you an example. Suppose I decided to add a section to the Mathematical logic article about the mathematician Charles Lutwidge Dodgson, and I wanted to mention Dodgeson's date of birth, 1832. So, I cite a University of Toronto web site, even though that site is about poetry.

Now, you could argue that poetry has little to do with mathematical logic, but that does not matter, the site is still a good source for Dodgeson's date of birth. You could argue that Dodgeson was not an important enough mathematician to be included in the article, but that is an argument about how to write the article, not an argument about attribution. If the argument that Dodgeson was not an important mathematician prevailed, the whole Dodgeson section could be removed, but there would never be any reason to leave his birth date in but remove the citation to the source of that information.

Perhaps you could provide an example of when your "clarification" would apply? --Gerry Ashton 19:01, 7 June 2007 (UTC)

The proposed modification was placed in a section titled "Wikipedia does not publish original research or original thought". Whether something is original research has nothing to do with whether it is related to the topic of an article. This section is certainly the wrong place to address the issue of keeping content relevant to the topic of an article. Indeed, I would say that issue does not belong in the Attribution proposed policy att all, because irrelevant facts can be well attributed, and unattributable vile lies can be directly related to the topic of an article. --Gerry Ashton 19:07, 7 June 2007 (UTC)

Seems correct, that I found the wrong place to put my "direct" verses "indirect" piece, and at the moment the real issue is whether of not it belongs here at all. —WikiLen 05:06, 8 June 2007 (UTC)

The issue I am raising is an issue of "how to write the article" — i.e.: policy on what facts are OK to have in an article. The title of this policy—"Attribution"—is misleading. It should include a subtitle such as, "Attribution centered revision process". I say this because this policy page also includes direction on avoiding "original research". Why is that part of this policy? This policy is more than just about citations. It is about the whole constellations of issues around attribution-centered editing. The first line of this page—see top box—says as much: "This page was created to serve as a combination of Wikipedia:Verifiability, Wikipedia:No original research and Wikipedia:Reliable sources." —WikiLen 05:06, 8 June 2007 (UTC)

---

A real example follows of a struggle over what facts are OK to be in an article: See this edit: 20:03, 6 June 2007.

---

I am on the edge of being in an edit war with someone over whether or not scandalous financial details about "investment guru", David Schirmer belong in the article The Secret or just in the article for David Schirmer. He is one of the so-called "secret teachers" in the film. This cited edit has both issues of synthisis (original research) and relevance (indirect relationship to film). I can find Wikipedia policy addressing the "original research" issue, but can't find anything overtly addressing the relevance issue—because it does not exist, I think. —WikiLen 05:06, 8 June 2007 (UTC)

The only way to demonstrate that a statement is not original research is to attribute the statement to a verifiable, reliable source. Original research is covered in the Attribution policy because the desire to avoid original research, and demonstrate that we have avoided original research, is one of the main reasons attribution is necessary. --Gerry Ashton 12:52, 8 June 2007 (UTC)

In my opinion, it's the other way around. Attribution is required to attempt to assure that the information is true, and original research is not allowed because it can't be attributed. --Coppertwig 22:57, 8 June 2007 (UTC)

Gerry Ashton, OK — I see why OR is in this article. I am becoming convinced issues of relevancy of content do not fit in this article. Unfortunately, I am getting no guidance on where it would fit in or in what policy it is already addressed. I put in a question at the Help Desk regarding this and got no answer — I take that to mean there is not a stated relevancy-of-content policy. It may be this "Attribution" policy should be renamed and refactored so that relevancy of content can be included. Alternately, a policy article on "Relevancy of content" could be created. —WikiLen 03:12, 9 June 2007 (UTC)

There is no policy because there are no bright lines--things are more or less relevant. Family life is generally irrelevant to a company president, but relevant to a national president. The nearest we have as policy is undue weight, which is very clearly a matter of judgment. You'll need to find some consensus over the actual article,if necessary by a third party or Rfc. DGG 06:07, 9 June 2007 (UTC)

Thanks for the info that there is no policy. I have an eye to flesh this out. I think there could be a policy if only to express there are "no bright lines" — need some place to go to for questions on relevancy. I see a policy that more or less has a defined boundary on relevancy. It would have the usual qualification that there are exceptions and then advise that this policy might have more exceptions than usual. It would seem that directly related verses indirectly related would be the normal boundary and that generally, topics of greater interest or importance might have relevancy expand into indirectly related matters. (unsigned post by WikiLen 21:29, 9 June 2007)

There is currently discussion aimed at developing policy on "relevance" at Wikipedia talk:Avoid trivia sections in articles#What is relevance?. --Coppertwig 13:13, 11 June 2007 (UTC)

publishing?

{{Soul characters}}

All the 'Character Analysis' sections were deleted from each character's article per this guideline. I mean come on. The game (like all games) is directed toward youths. What are they going to do. Publish their papers on these characters? Seems to me that a fatal flaw in Wikipedia is that it relies primarily on published work (be it right or wrong) for it's content. --89.180.150.49 15:18, 12 June 2007 (UTC)

The game may be directed towards youths, but not this encyclopedia. OUR standards are higher. Blueboar 15:38, 12 June 2007 (UTC)

"OUR standards are higher"?! Only under the metric of requiring published works. Which arguably isn't the best one to have chosen, which makes it very understandable why [User:89.180.150.49|89.180.150.49]] is perceiving a problem with the content. No need to be quite so dismissive towards him/her. Mathmo ^Talk 02:06, 20 June 2007 (UTC)

"challenged or likely to be challenged"

Recent discussions in FA-land have revolved around the phrase "challenged or likely to be challenged". Many expressed concern that this is too broad, because it's not specific what exactly a "challenge" is. Current practice seems to be that any editor can lodge a "challenge" against any uncited material for any reasons, and thus demand its sourcing. Is this not too broad? For example, someone suggested "...likely to be challenged by people who are reasonably familiar with the topic." Or that challengers must provide a "demonstration of how the facts are likely to be challenged and/or are inherently insupportable". Or that a challenge must consist of "a demonstration that the facts are in play in the professional field" itself. I feel we must do something here, because the phrase is basically a tautology as it stands now: "It's likely to be challenged because I'm challenging it right now", as one editor put it. - Merzbow 18:27, 12 June 2007 (UTC)

As far as I know, this is indeed intended to be a broad statement... indicating that an unsourced statement can be challenged (and potentially removed) simply because it is uncited. This is to encourage people to source their material. It is better to oversource than undersource.

A "challenge" means someone has expressed concern about the statement. It does not really matter why it was challenged. A "challenge" can be expressed in broad spectrum of ways, ranging from a mild comment raising some concern on the talk page - through placing a {{fact}} tag on the statement - to, in some situations, summarily deleting the material. Can this be abused?... sure. All of our guidelines and policies can be abused.

The whole point is to encourage people to do their homework before they write something in Wikipedia, and to provide sources for anything that is even remotely likely to need it. Blueboar 19:19, 12 June 2007 (UTC)

"and to provide sources for anything that is even remotely likely to need it" - that is a much clearer and stronger formulation than "challenged or likely to be challenged". One can reasonably interpret the latter to mean "challenged by informed editors" or the like, which is the source of much of the confusion in the thread I linked to. Thus I would still put forth that no matter what one's position is on sourcing, the phrase as it stands now is just too vague. If consensus is closer to what you're saying, we should just go out and say in the policy that "all non-trivial statements should be sourced, and must be sourced if any editor asks it to be". I still think that a challenge should be something more than a drive-by editor who knows nothing about the subject in question demanding that some fact undisputed in the field be sourced. - Merzbow 20:34, 12 June 2007 (UTC)

"indicating that an unsourced statement can be challenged (and potentially removed) simply because it is uncited" - uh, no. If an article says that George W. Bush is the president of the united states, and some damned fool comes along and {{fact}} bombs the article, putting a tag at the end of that sentence, the tag can be summarily removed. We do not have to cite common sense.

More generally, User:Raul654/When to cite is a page I created to try to address this concern. It's still incomplete and could use work, but I think it's a good start. I eventually hope to move it into the main wikipedia namespace as policy. Input and help from others is desired. Raul654 20:43, 12 June 2007 (UTC)

It does look like a good start, something like this is sorely needed. - Merzbow 21:29, 12 June 2007 (UTC)

It is best if every number is sourced, because people cleaning up after a number vandal will not necessarily know what the correct number is. --Gerry Ashton 23:13, 12 June 2007 (UTC)

OK... we are getting into "silly example land" here... of course you do not have to pay attention to rediculous "challenges" of things that are obviously common knowledge (such as the statement that George W. Bush is the President of the US). In fact the sentence right before the one we are talking about admits as much: everything in Wikipedia must be attributable, in practice not all material is attributed. The point is, we could attribute the statement, but your average editor would agree that it is rediculous to have to do so in this case. (and, seriously, would anyone actually make such a "challenge"?)

However, let us go a half step deaper... Editor A might think a given statement is common knowledge... he doesn't think is needs a citation... but Editor B disagrees and asks for one. Editor A should provide it - simple. An example of this would be... "Bush owns a ranch in Crawford, Texas"... it could stand without a citation (most people do know this fact), but a request for a citation is perfectly reasonable. If challenged, a citation should be provided. However this is probably not contentious enough for anything other than a "citation request" tag.

Now let us take it another half step deaper... Editor C wants to add a statement to the effect that "Many of Bush's policies have proven to be controvercial". This is likely to be challenged, and so Editor C should definitely include a citation at the same time that he puts the statement in the article.

And finally Editor D adds "Bush's policies are bad"... this is not only likely to be challenged, and thus should have a citation from the get go (in fact, I would say it should have an in-text attribution as to who says it and should be stated as as the author's opinion not as fact)... but, if it is added to an article uncited, it should be removed summarily.

Now... all of this is a very wordy way of saying ""look, if there is any doubt, add a citation to a reliable source with the statement", and while we could put all sorts of examples in the Policy, they tend to encourage instruction creep that defeats the point of the policy - as editors try to tweek examples to fit what they want to include in their articles. And by the way... the language here was taken from WP:V, it has been part of Wikipedia Policy for a very long time. Blueboar 00:36, 13 June 2007 (UTC)

I endorse this view. It has been part of policy, but widely ignored in both directions. Some science articles take the opportunity to give a research level density of citing; some give an offhand reference to a textbook. Some contentious articles give a reference per word, because each one may have been challenged. Common sense is a good enough guide, and we should direct our efforts to actually adding sources. DGG 03:21, 15 June 2007 (UTC)

The problem with this idea is that it makes the rule prone to intentional abuse. It's quite possible for someone to pick an article and then start insincerely "challenging" common knowledge or uncontroversial statements. Typically this happens either because 1) the user is on a power-trip and likes forcing people in random articles to jump through hoops, or 2) because the user doesn't like the article and is making insincere challenges in the hope that nobody will respond and he can kill the article. Ken Arromdee 04:23, 15 June 2007 (UTC)

I have to ask... how common is this "abuse"? I don't doubt that it takes place, but wonder if it is a localized issue (to a specific type of article, or a specific kind of statement) or a general one that takes place throughout Wikipedia. The idea behind the phrasing: "challenged or likely to be challenged" is to avoid the very common problem of editors adding material that violates WP:V, WP:NOR and WP:NPOV. It gives serious editors a much needed tool to demand that such statements be attributed, and the authority to remove such material if no citation is provided. In other words, the wording is there to prevent another (and I think far more common) form of intentional "abuse".

I can see that this discussion needs to take place... but let's not throw the baby out with the bath water. We do need something that says citations should be given if they are legitimately requested. To my mind the language may need a tweeking, but something like it should remain ... The first question that needs to be addressed is: When is a "challenge" legitimate, and when is it just "abuse" of the rule? Blueboar 14:45, 15 June 2007 (UTC)

I think it's okay to say that challenges must be legitimately requested, but the language used is much too strong. This happens a lot in Wikipedia policies: someone writes a policy in such strong terms as to imply that it's absolute, in order to prevent others from trying to find loopholes in the policy or to casually violate it; but then, the fact that we don't *want* the policy to be absolute pops up. You could at least say "sincerely challenged", or add another sentence "Challenging material that you do not genuinely doubt is considered abuse and such challenges may be ignored". It would weaken the force of the statement that anything challenged can be deleted, but we need to weaken it; it's gone too far in one direction, and has reduced NPOV abuse only at the cost of greatly increasing spurious-challenge abuse. Ken Arromdee 19:30, 15 June 2007 (UTC)

Ken, I encourage you to come to Wikipedia:When_to_cite. We should probably move this discussion there. - Merzbow 19:57, 15 June 2007 (UTC)

On contentious articles, it's pretty common, actually. See my comment here Raul654 14:51, 15 June 2007 (UTC)

This does not surprise me... an article on a contentious topic is bound to attract people who disagree with every little bit of wording that disagrees with their POV. And it is exactly that sort of article that prompted the need for the citation rule in the first place... with POV editors claiming that their POV edits were "common knowledge" and did not need citations. In contentious articles, where it is likely that editors can not even agree on what is "common knowledge, I side fully with the "if you want to say it, cite it" view of things. Yes, it can be frustrating to have to cite things that truly are common knowledge (Bush is President... The Catholic Church is against abortion... etc)... but I see the alternative as being worse.

Out of curiosity... I know part of this discussion came out of discussions over Featured Article criteria... how many contentious articles go up for FA status? (I would assume that with all the debates, vandalism, etc. that typically go on at such articles, few reach even GA status... but I could be wrong). Blueboar 15:24, 15 June 2007 (UTC)

Plenty of contentious articles go up for GA/FA and succeed (just look over the list). However, there are several articles on completely uncontentious subjects over which epic battles have broken out related to citations. Hence the need for guidelines. - Merzbow 17:36, 15 June 2007 (UTC)

Specific reliability

What I don't see mentioned in the documents to date on reliable sources is that many sources are reliable for specific topics or specific subtopics. People Magazine, for example, is reliable for celebrity information, but not for legal information. Likewise the Christian Science Monitor is reliable for world events information, but not for medical information. The point is, just because a source is reliable in one area doesn't mean it's reliable in another area -- and vice versa. - Keith D. Tyler ¶ 21:02, 19 June 2007 (UTC)

Wholesale deletion

I have a question: Suppose that an article has entire sections which, though correct, are unreferenced. Is it reasonable to delete those sections entirely without discussion, or should some discussion take place beforehand? Silly rabbit 23:12, 20 June 2007 (UTC)

If it is about a living person, it should be removed immediately as per WP:BLP. If the article has been tagged with {{fact}} and {{unreferenced}} templates for a matter of months with no response, and the validity of the material has been challenged it can also be removed under those circumstances. There isn't much discussion to be had, just restore it with a valid citation and move along. Burntsauce 23:41, 20 June 2007 (UTC)

If an editor knows that entire sections of the article are correct but unreferenced, the issue becomes how does the editor know it? If the editor knows of references that show the information is correct, the editor should add those references, rather than delete the information. But if the editor does not know the information is correct, the editor is free to delete it. --Gerry Ashton 23:44, 20 June 2007 (UTC)

Thank you for both for the clarification. My concern is that User:Burntsauce has been apparently making it a policy to delete massive amounts of material from articles on the grounds that it is unreferenced. It seems to me that a more productive route to achieving the same ends is to find references for the questioned material, rather than engaging in a deleting spree. Articles can be tagged, and references provided. But usually this process is slow, particularly for low-traffic articles. Deleting 90% of an unreferenced article usually has the side-effect of burying the material to be referenced deep in the edit history. Can I suggest that perhaps moving it to the talk page might be a more constructive approach? Silly rabbit 23:53, 20 June 2007 (UTC)

I will add a third option... if an editor knows the uncited information is incorrect, it is his duty to delete it. As for grey areas, all I can comment on is my personal approach ... I tend to be a strict "have your citations ready before you write it in Wikipedia" person... even when I think the information is correct, I demand that it be cited. However, I also understand that it can take time to gather such citations. So, what I do (and what I think everyone should do) is this: 1) I tag the article, section or sentence. 2) If there is no response after one week I leave a comment about my concerns on the talk page. 3) If there is still no response after an additional week, I post a stronger comment on the talk page stating that I may remove the material if no citation is provided. 3) I then wait a few more weeks and then, if there is still no response, I assume that it can not be cited and I start deleting. However... if there is a response (say a note on the talk page saying that someone is working on it) I back off... usually for a few months. If the problem is not resolved, then I remind the user of my "threat"... and wait a while longer.

In other words... I feel editors should delete unsourced material wholesale... but only after giving the creator due courtesy and notice. Blueboar 00:57, 21 June 2007 (UTC)

Yes, I agree with this approach. The issue I'm objecting to is whether it is acceptable to delete summarily without any due-process. I think that other editors should be given every opportunity to shore up an article before it gets the axe: and {{unreferenced}} tags are quite ineffective. Talk is better, since it is more engaging and inclusive. It would be nice if there were a stated policy about this somewhere, as if it didn't follow naturally from WP:CIVIL. Silly rabbit 01:29, 21 June 2007 (UTC)

Just to state the obvious: uncited information that violates WP:BLP should be immediately deleted and not copied to the talk page or anywhere else. —David Eppstein 01:01, 21 June 2007 (UTC)

That is one interpretation of the WP:BLP policy. But attempts to clarify it have been thwarted recently. There is some debate over whether unreferenced material needs to be contentious or not. I think the phrasing at that project page needs to be clearer on this point. It should say something along the lines of: "Any material which is not suitably cited should be deleted." Instead it says that any contentious material should be deleted. I am not happy about this at all: does contentious mean libelous? controversial? or just unreferenced? There is an ongoing debate at the project page about it now. I think the statement is deliberately vague, perhaps to avoid endorsing a unilateral course of action. For example, to say that so-and-so was born at such-and-such a place, clearly needs a reference, but should not be deleted without some discussion and due process. On the other hand, so-and-so was a vicious dictator who liked to eat babies should probably be speedily deleted. Anyway, I feel that certain editors are abusing the BLP rules. A fairly extreme case is http://en.wikipedia.org/w/index.php?title=Somy_Ali&diff=prev&oldid=139545223]. Some other diffs: [11], [12], [13], [14]. In all cases, a truly enormous amount of information was deleted. Most of it was factual biographical information, after the fashion of "so-and-so was born at such-and-such place." Clearly, it needs referencing, but I find the summary wholesale deletion to be a rather extreme measure. Silly rabbit 01:29, 21 June 2007 (UTC)

See Jerry Springer, under which WP:A was cited for significant deletion of the article. (Guess who performed the deletion.) The only place in WP:A in which I see the word "deletion" is: "Absurd unsourced claims and original research should be deleted rather than tagged or moved to a talk page." None of the deleted text appears to be "absurd" nor is it original research because no original ideas or interpretations of facts or any of the things that constitute original research appear in the article. And if the rationale for deletion is changed to a violation of WP:BLP, deletion is in regard only to "unsourced or poorly sourced contentious material" or "contentious material about living persons that is unsourced or poorly sourced", which is said explicitly and which is used as context for the rest of that page. Again, nothing about the deleted text appears contentious to me. What remains in the Jerry Springer article is indeed absurd:

• The lead paragraph and infobox were unchanged, but no sources are provided. By the logic of deleting any unsourced material, those would also have to be deleted because no sources are provided there!
• The "Other appearances" section remains, but in the span of his career, Springer on the Radio, The Springer Hustle, and even America's Got Talent is negligible compared to The Jerry Springer Show, his obvious namesake. The Jerry Springer Show only appears in the lead, and any reasonable biography of Springer would at least expound on that. His political career is notable as well.
• The "Scandals" section remains, which is about Springer having sex with a porn star and the aftermath. The way the article is now, it gives undue weight to the scandal compared to his career. Has the scandal hurt his career? Not that I know of. I never even heard of the scandal until now.
• Many categories are left behind. If the remaining text cannot support them, why should they remain?

The article was rated as B-class (it's probably a good start-class in my opinion due to the need for referencing), but now it's effectively a stub. The point is, the deleted text can be attributed properly, and users should assume good faith that someone will come along and find sources. One month I think is too short a time to expect responses, perhaps 3-4 months at minimum. Deletions in the name of referencing marginalize the work of many contributors and lead to very bad morale. I did a web search for "Jerry Springer East Finchley", and after a few seconds, I found a reliable source that can support Springer being born in the East Finchley tube station during World War II. Simple actions like that are constructive, deletions after a month of putting up the {{unreferenced}} tag are not. Notice that this applies only when attribution is possible.

Furthermore, deletions should not prompt people to reinsert (reference-improved) text from article histories, since although that is desirable in this case (to acknowledge previous contributors), there is nothing to direct users to do that, and not all users would do that. (God knows if someone adds in referenced but poorer quality prose than before.) For all I know, that is the recommendation from the GFDL: net deletions (overall deletions and additions that result in deletion at the end) should not be performed without a good reason. I am now happy to find references for the Springer article (give me at least a month), of which I have not had a significant hand in, but I'm also pissed off that it took the deletion of text to prompt my reaction. (And I apologize if this is a rant, but this is a good rant.) Tinlinkin 05:32, 21 June 2007 (UTC)

I feel your pain. I've been browsing through you-know-who's contributions, and found that he similarly gutted Lucky Charms. He certainly had a case that it was not suitably referenced, but it took me just an hour or two to reference nearly everything that had been removed. (Despite my handle, I'm not at all an expert on breakfast cereals produced by General Mills.) I think that efforts to delete material, rather than attempt to properly source them in the first place, is horribly misguided. If an editor finds that a source is lacking, he or she should try to provide the reference. Failing that, go to the talk page. Only after demonstrating significant good faith in trying to tease the references out of the more reticent editors should material (of an uncontroversial nature, at least) be deleted. Silly rabbit 05:56, 21 June 2007 (UTC)

An article's regular editor (or group of editors) who doesn't question other users' edits to the article would most likely believe the edit is factual, based on the regular editor's knowledge. But another editor who does not see any inline references may be skeptical of the information that is disseminated. This skepticism may be equivalent to "contention" in their mind. (This is certainly not my view.) Randomly going through articles to find unreferenced sections is not my cup of tea, but once a nominator posts an unreferenced or similar tag, he or she should either leave it alone or make every effort to help find sources, not just order someone to find references and say "time's up" (delete) when references are not found in some determined span of time. There is no deadline, not that I know of, to find references if the information is uncontroversial. Tinlinkin 09:31, 21 June 2007 (UTC)

Smith/Jones example

There's a discussion at Wikipedia talk:No original research#Chicago manual of style in which I have proposed removing the Smith/Jones example. The discussion is long, but to summarise the key points as I see them:

• The example claims that the Chicago Manual of Style defines plagiarism as "using a source's information, ideas, words, or structure without citing them". In fact, the Chicago Manual of Style contains no such definition, or anything that resembles it. Noone has contested this.
• The example is based on a real case on Wikipedia. However, it misrepresents the original case, at which different issues were at stake. The original case it was based on is very convoluted and difficult to understand - however, there was no misquoting of the Chicago manual, which was added later.
• Quite apart from the way that the example misrepresents the source and the original case, it is unclear and does not follow logically. It frequently confuses new users, and I have asked for other users to explain it several times, with no satisfactory reply.
• The case it is based on is from a highly controversial dispute (vaguely related to Israel/Palestine). In essence, it is saying "here is an example of where an editor violated the policy". In this kind of situation, I think it is essential be careful not to misrepresent the editor who made the edits, not to misrepresent the sources used, and not to misrepresent the original dispute. Failure to do so risks damaging Wikipedia's reputation for impartiality and objectivity. Honesty and objectivity are essential to Wikipedia, and bending the facts is not justified on the basis that this is "just an example".

I think the problems above are too fundamental to be fixed with modifications to the example, and in my view this complicated and controversial case is a bad case study to use. I would support exploring ways of improving the policy to explain it more clearly - perhaps including alternative examples. In the meantime, I propose that the example is removed immediately. Enchanter 00:05, 24 June 2007 (UTC)

I would replace it with another example that makes the same point about synthesis. Blueboar 12:18, 24 June 2007 (UTC)

Wikipedia has a reputation for "honesty", "impartiality and objectivity"?? This is a real question, BTW. SqlPac 19:29, 24 June 2007 (UTC)

Idea for a tag re: reliability

There is an interesting proposal at WP:RS to creating an inline tag for situations where a source is reliable, but hardly the best of sources available. The proposed wording would be something polite and non-confrontational, like [more reliable source requested]. To give you an idea of where such a tag might be used... take an article on an historical topic, containing a statment that cites to a Histroy Channel documentary as a fact source. While the History Channel is not completely unreliable, I hope you would agree that there are probably sources that are more reliable for the same information. However, more community input is needed. (Please join the discussion at Wikipedia Talk:Reliable sources#Inline tag?) Blueboar 22:42, 24 June 2007 (UTC)

Status?

So apparently this is no longer a policy proposal. Is it an essay? —Ashley Y 22:35, 25 June 2007 (UTC)

I think it's a good idea to leave it without status for the time being. It's not a personal essay. There's a sense in which it's policy, in that it's a summary of two other policies, but strictly speaking it doesn't have that status (though it did for a while). It's not a proposal anymore. Yet it's a page that lots of people still use and link to in order to understand the policies. Therefore, it has a kind of unique status. SlimVirgin ^(talk) 22:38, 25 June 2007 (UTC)

But it's definitely not a policy or a guideline or anything official, is it? That should probably be mentioned at that top. —Ashley Y 22:43, 25 June 2007 (UTC)

Like was said, it has a sort of unique status -- it's sort of half-official and half-unofficial. Best to leave it this way unless and until it can be pinned down. mike4ty4 09:15, 4 July 2007 (UTC)

References in wikipedia lists

Please comment in Wikipedia_talk:Verifiability#Verifiability in lists. `'Miikka 23:18, 5 July 2007 (UTC)