Wikipedia talk:Attribution/Archive 4

This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Archive 1

Archive 2

Archive 3

→

A different sort of approach

It seems to me that this discussion is falling into something that was at the heart of the problems with WP:RS - an attempt to make a white line guideline that will on its own allow all of the sources we want to allow and forbid all of the ones we want to forbid. We all have a good sense that an article based soley on press releases is bad, but one based on Outpost Gallifrey probably isn't. We seem to recognize that tagging a sentence that's clearly true with a {{fact}} tag is a bad idea. What we're stumbling over is wordings that can make a perfect general case out of all of these observations.

It is unlikely that any such wording exists. Instead of the current process, in which we try to craft compromise wordings to nudge this policy closer to the mythical white line, why don't we make a move along the lines of admitting that source evaluation is hard in the policy. I'm thinking one bit, towards the top of questionable sources, that notes that "In some areas there may be sources that meet every criteria listed for a reliable source, but still ought not be considered reliable. Such exceptions are rare, and you should be prepared to explain your reasoning when you assert the existence of one. Stringent application of this policy is no substitute for thought and editorial judgment." Then, towards the bottom of the same section, we add one last exception to the effect of "Some articles and topics may require the use of sources that do not meet all of the standards here. These exceptions are rare, and great care should be taken in cases where they exist. You should be prepared to explain in detail why the article would benefit from the use of a non-traditional source, and be aware that these exceptions are highly controversial."

Such wordings do open the door for the clueless and the malicious, but experience has shown that we cannot legislate either of those categories out of existence anyway. Given that both will abuse this policy in any form we present it, we ought do what, in the end, we always do on this project: trust the overall base of editors to reach sensible decisions, and write policies that will support those decisions when referenced. Phil Sandifer 13:38, 16 October 2006 (UTC)

Works for me, although I obviously think my exchange of three words for one is a far better solution. :) But whatever gets the agreement on this issue has my support. I certainly agree that sources are evaluated in context. Hiding Talk 13:47, 16 October 2006 (UTC)

Phil, do you agree that isn't needed if we stick to "unquestionably recognized expert"? WAS 4.250 15:17, 16 October 2006 (UTC)

Not really. I mean, that may address the concerns that we can come up with, but the real test of policy is its flexibility in the real world. There seem to be about a half-dozen of us actively discussing this policy. If, among us, we've read 1% of Wikipedia's articles, I'd be surprised. I mean, I'm happy to help with the fine-tuning of language as well, but I can't really think of a circumstance where some hedging against rules-lawyers on either side would become wholly unnecessary. Phil Sandifer 16:06, 16 October 2006 (UTC)

I understand your point Phil, and honestly I would prefer the approach you outlined above. However the discussion here has shown there is a sincere line of thought that holds to the idea policies should apply equally to articles regardless of the subject matter. I do not think we can achieve consensus with your suggested approach. I do wish that the people holding that line of thought would explictly state what they think of your proposal as well as the current compromise wording. I dislike having to guess what they mean from their silence.--Birgitte§β ʈ Talk 20:33, 16 October 2006 (UTC)

This policy would apply equally to all articles regardless of the subject matter. It just acknowledges that part of the nature of sourcing is that there are exceptions. Phil Sandifer 20:35, 16 October 2006 (UTC)

I do like your idea myself. If no one has spoken against by tommorow you should work it in to the page.--Birgitte§β ʈ Talk 20:38, 16 October 2006 (UTC)

Back from the weekend

I've been off wiki for much of the weekend, and haven't been following the debate as close as I would like--but I see the (for lack of a better word) "pop culture" section remains a sticking point. Both to get myself back into the debate, and perhaps to assist others, let me attempt to summarize the issues.

Many articles on Wikipedia are sourced in such a fashion that would violate WP:ATT without such an exception; further, in many cases these are articles on clearly notable subjects which are not disputed WRT their accuracy.
Granting exceptions to policies to specific subject areas (i.e. pop culture, fiction), is seen as unwise. Other subject areas (quilting, roadgeek pages) probably qualify for looser standards. Furthermore, even within general "unserious" topics; there are subtopics which shouldn't qualify for the exception. A bestiary of Harry Potter, for example, need not limit itself to serious or scholarly works from reliable (not-self-published) sources; OTOH, discussion of critical reaction to the series (books or film), or real-world controversies like whether or not HP is satanic (real-world in the sense that a non-trivial amount of people believe this--I don't intend to promote such lunacy here), should use reliable sources. A review of the books written by a major literary critic is appropriate, a book review posted on someone's fansite is probably not.
Certain categories of subjects shouldn't be exempted in any case. Traditionally-academic subjects (science, liberal arts, business, etc); subjects involving significant active controversies (religion, politics); and other specific topics with a rich literature which meets WP:ATT without the exemption under debate; and/or a large number of established "professional"s who write on the topic.
- OTOH, some debates within fandom, such as whether or not Deckard was a replicant, probably don't rise to the level of "significant active controversies". --EngineerScotty 16:36, 16 October 2006 (UTC)
Some remain unconvinced of the value of self-published, "amateur" sources, even for nonserious topics. (Here, amateur doesn't mean "unpaid" just as professional probably shouldn't be construed to mean "paid"; the term instead refers to a source's amount of dedication and study into the topic. Compensation, or lack thereof, is generally irrelevant; though if demonstrably large numbers of people are willing to pay for an individual's advice on a subject, it's probably a sign that he knows what he's talking about; either that, or he's a highly skilled con artist. Who pays for a professional's advice is also of note; a professional who successfully consults or who is tenured faculty somewhere inspires more confidence than someone who is paid by a tobacco company or other party where a demonstrable conflict of interest exists)
Many are concerned that any exemption which we create which isn't narrowly-tailored, will be open for abuse. Others think that this policy should not attempt to legislate out all forms of abuse--therein lies madness; and that WP:BP, WP:DE, and other policies remain available to deal with abusive behavior.

Have I missed any salient point?

--EngineerScotty 16:11, 16 October 2006 (UTC)

Yeah. There are quite a few people who will argue that a bestiary of Harry Potter is "fancruft" and should not be allowed in without reliable sources. ColourBurst 16:39, 16 October 2006 (UTC)

Good point... which brings up a few questions concerning how this policy interacts with notability and NPOV. A common assertion, which many agree with, is that a "notable" topic is one which is described by multiple reliable sources--(a band or other topic which is only described in self-published promotional literature, for instance, isn't eligible). OTOH, given that Harry Potter is highly popular, the subject is "notable" in the sense that "lots of people can be shown to care about it". Many objections to fancruft don't come about because of sourcing issues--certain fictional universes (Star Wars, Star Trek, Harry Potter, Lord of the Rings are fairly-well sourced; in this way HP is a bad example); many object to fancruft simply because they think that such material is inappropriate for an encyclopedia, regardless of how well-documented the fan material is. Obviously, such arguments are beyond the scope of this policy; though as you point out, some may seek to craft this policy to exclude topics they dislike. I'm probably guilty of the same thing--several suggestions I've made are rather transparent attempts to keep crackpots and pseudoscientists from suggesting that self-published material promoting their theories as equivalent to mainstream science. :) --EngineerScotty 17:11, 16 October 2006 (UTC)

Again, this is not true. The current AfD, Wikipedia:Articles for deletion/Mega Man weapons (and there are many articles under deletion and already deleted that fall under this category) is under attack for not having reliable sources (the sources in the article point to GameFAQs articles which wouldn't be reliable anyhow... okay, they're reliable as game guides to a lot of the population, but game guides fall under WP:NOT.). Much of the source material in a universe is "official" source material which is primary (LotR is an exception because it's a historical text as well as a "geek" text) Consensus is wildly mixed on this, there's a lot of people in the "delete all articles except for the works" camp, a lot of people in the "keep them and clean them up" camp, and a lot of people who are in the middle (which tends to bias them to things they care about). For subjects that are academic, I have no problem insisting on absolute peer review/fact finding, because those are subjects that need that kind of thing. ColourBurst 19:33, 16 October 2006 (UTC)

Slightly off-topic, GameFAQs is unreliable because its content is user-submitted (some but not much editorial oversight), not because it's a game guide. For example, the Ultimania series of guides is invaluble to citing the Final Fantasy series. Nifboy 20:45, 16 October 2006 (UTC)

Not off-topic at all. Would I be right in thinking that the site is one where the username of the submitter is attached to content they submit? If so, a user on the site who built up a reputation for accuracy might be allowable as "an indisputably recognized expert" under the text here. JulesH 21:30, 16 October 2006 (UTC)

I suppose, but I don't know how you would go about showing someone had a "rep" on GameFAQs. There does exist a separate user class for game developers, etc (who have to prove they work there) on the forums, but that's it. Nifboy 01:36, 17 October 2006 (UTC)

The Ultimania guides would be considered primary sources because they're directly published by Square Enix. I think people would discourage having large amounts of information from primary sources in an article. ColourBurst 01:05, 17 October 2006 (UTC)

The most pertinent info in Ultimania is development info, particularly developer interviews, which is basically the scarcest and most in-demand info at GA/FAC. I think the exact same interviews, if they had shown up in secondary sources, would be equally as valid, because the source-source in this case is the developer's own words. There is a difference between relying solely on primary sources and having really really good primary sources. Nifboy 01:36, 17 October 2006 (UTC)

Is it ready?

Is it ready to submit for comment to the general public? I think it is. WAS 4.250 16:30, 16 October 2006 (UTC)

The discussion on this page would suggest not. SlimVirgin ^(talk) 01:20, 17 October 2006 (UTC)

Missing things

Here are some points that this proposal misses or doesn't address properly:

An important factor in how reliable a source is is the reputation that it has to defend. For example: The Washington Post has far more to lose, from the damage to its reputation that accrues, from making an error of fact than someone posting under a pseudonym on a web log does.
Whilst it may be considered by some to be disruptive to "prematurely" remove material for which reliable sources could "reasonably be found", addressing the issue this way is in part telling people to stuff beans up noses, because it suggests that wiping content is the sole way of dealing with material that is not yet sourced. If one sees unsourced material, one productive way of addressing it is to attempt to find sources.
Saying that there is a class of content that "does not need a source", is outright wrong. Everything needs a source. That Paris is the capital of France is in fact verifiable from several of the sources cited in Paris#References and is a bad example if one is trying to give an example of content that "does not need a source", given how copiously sourced that fact already is in our current article. The whole theme of this part of the proposal is backward. Things that are "obvious" and "well-known" should be easy to source as a consequence. Therefore there is no excuse for not having a source cited, against which readers, who may live on the other side of the world, can check this "obvious" and "well-known" fact.
Address the bean-stuffing issue mentioned above, by pointing out that if something "obvious" is lacking a source the best route is to improve the encyclopaedia by finding a source to cite, and there is no need in the first place for creating the notion of "content that does not need a source". Start with something like this:
Any unsourced material may be removed, and in biographies of living persons unsourced contentious material must be removed immediately.

Any material that is challenged or likely to be challenged needs a source, as do quotations; the burden of evidence lies with the editor(s) wishing to retain the material. Material lacking attribution may be removed, but always aim to improve the encyclopaedia: In the case of material that is widely known, and thus easy to source, it improves the encyclopaedia more quickly, and requires less work to be done on the parts of editors overall, to find and to add sources for unsourced material than it does to challenge and to remove the material. Conversely, it damages the encyclopaedia to retain unsourced controversial material about living persons, or even to relocate it to talk pages pending the supply of sources. Therefore such unsourced contentious material must be removed immediately and not be moved to talk pages.
Effectively turning "verifiability, not truth" into "attributability, not truth" rather loses the pithy impact of the original. I suggest losing the maxim entirely and instead starting again from scratch with something like the following:
That something is true or factual is not sufficient by itself for it to be included in Wikipedia.

Something that is true and factual must also have already been recorded outside of Wikipedia, have been checked for veracity by a process of peer review, and gone on to become a part of the corpus of human knowledge. Only material that has been through the processes of fact checking, peer review, publication, and acknowledgement by the world at large (i.e. by people other than its creators/authors/inventors/initial proponents) belongs in Wikipedia. These processes occur outside of Wikipedia.
One major thing that this proposal misses is an explanation of why we have such content policies in the first place. You might start with something like this:
Wikipedia does not require its readers to blindly trust its editors.

Wikipedia does not expect readers to trust its articles solely on the basis of it having selected expert authors for its articles. This is because of the very nature of a wiki, where for one thing it is difficult or even impossible to determine that people are who they claim to be. Wikipedia instead aims to provide readers with the means for them to check all content for themselves, against outside sources that will have done all of the primary research and fact checking.

Uncle G 16:21, 16 October 2006 (UTC)

I agree with what Uncle G is saying here. In addition to the "Paris is the capital of France" example above, I can't resist adding the "the sky is blue" example of something that doesn't need citations, perfectly punctured here by Dpbsmith. Carcharoth 16:59, 16 October 2006 (UTC)
- I think the point is that for certain things, sources can obviously be provided--requiring cites everywhere is a potential opening for disruption. And of course, in practice, this is probably less of an issue than it appears. That Paris is the capital of France is cited in the article Paris; other articles which mention Paris are probably unlikely link to Paris; and not explicitly mention where Paris is located, unless necessary to disambiguate it from other Paris's like the one in Texas or the one in the video. :) --EngineerScotty 17:16, 16 October 2006 (UTC)
  - requiring cites everywhere is a potential opening for disruption — Requiring that everything be sourced opens no avenues for disruption, and results in more accurate content, as Dpbsmith's example illustrates very effectively. Whereas, in contrast, trying to avoid this mythical spectre of disruption results in the creation of a weasel-notion of "content that does not need to be sourced", a dangerous, damaging, and completely unnecessary (See the suggested wording above.) idea. Uncle G 18:39, 16 October 2006 (UTC)
    - Think of it less as content that doesn't have to be sourced and more as content that shouldn't be sourced. Not every sentence of every article should have a reference - it would render them unreadable. Phil Sandifer 18:43, 16 October 2006 (UTC)
      - Actually, the solution there is to be better at finding sources to cite, and citing in a broader fashion. For some articles, you can find two or three standard reference works on the topic that could be used to cite nearly everything in the article. This is preferable to having two or three random sentences sourced to some random webpage someone found on a Google search (this happens on a distressing number of articles). What you then do is add a general qualifier somewhere making clear that unsourced statements can be verified by reading these standard works (and then give details). This is very similar to a "further reading" section, or a "primary references" section. In other words (using an example), the basic details about the life and times of Alexander the Great (those that are agreed by all sources) can all be found in a standard book on Alexander the Great. The assumption should be that the reader knows to look to that book to verify the basic stuff (or the editor points this out). Stuff that is not agreed by all sources, or detailed stuff, can be referenced to page numbers and individual sources. Similarly, a whole paragraph on Alexander's early life could be dealt with by a single cite at the end of the paragraph to a book chapter on his early life. This sort of citing strikes a balance between readability and maintaining a high standard of verifiability. This is precisely the sort of situation where examples help. If you haven't seen this sort of thing done, it is difficult to explain just how effective and readable it is. Carcharoth 20:55, 16 October 2006 (UTC)
      - "content that does not need to be sourced" is a dangerous, damaging, and completely unnecessary idea. The concept of "content that should not be sourced" is worse still and is utterly wrongheaded. That one is thinking along those lines is a big red flag, that one should pay attention to, indicating that one has made a serious error in one's thinking somewhere. Everything needs a source. Discard these weasel-notions about content that does not need to be, or that even should not be, sourced. They aren't necessary at all. The error is thinking that they are in any way necessary, caused by muddy thinking about how to connect sources to articles. See above. Uncle G 22:51, 16 October 2006 (UTC)
- I think the ideas under Wikipedia does not require its readers to blindly trust its editors. should be worked into the introduction before the TOC rather than be presented as a key principle. I do believe it is quite important that we say these things here.--Birgitte§β ʈ Talk 20:52, 16 October 2006 (UTC)

I'm sorry to have mentioned this below without reading this discussion first. My general understanding of W:V has pretty much always been that even common knowledge needs to be sourced if it's knowledge about the subject of the article. It does seem, however, that there would logically be a little leeway if you're merely invoking some very obvious, tangentally related knowledge in passing. For example, if I'm writing a biography of an obscure individual and I have a reference that says "X briefly lived in Albuquerque when he was in his twenties", I would assume it's not out of line to refer to Albuquerque as "Albuquerque, New Mexico", even if my source did not directly state that Albuquerque is in New Mexico. I would also assume I don't need to cite an atlas to do this.

As Engineer Scotty mentioned, I can't imagine this coming up often, but I assume that's what the "Paris, France" clause in this page refers to -- undisputed common knowledge referenced in passing, rather than information about the subject of an article. Does that make sense? -- (Lee)Bailey ^(talk) 01:25, 17 October 2006 (UTC)

Yes it does. The article on Paris should have a reference to a source that confirms that Paris is the capital of France, as should the France article. Other articles mentioning that Paris is the capital of France are either mentioning this unnecessarily, or in passing, and a link to the article should be enough. eg. (1) "Rock star Y's tour went to New York, Sydney and Paris, the capital of France." (mention of Paris being the capital of France is unnecessary); (2) "Concert musician Z's tour included the European capital cities of Paris, London and Berlin." (mentioning they are capital cities here is relevant, but the link is enough, no source needed to say that they are capital cities - what is needed is sources confirming that these tours went to these places); (3) "Paris became the capital city of France in 1657" (source needed here - date picked at random); (4) "The cities of Paris, Marseilles, Lyon and the present-day village of Smallville-sur-Saone have all at some point in the past been the capital city of France." (again, sources needed).

As for your example, you are technically interpreting the information you have. If there are other Albuquerques, you have to be certain that you are correct to assume it is the New Mexico one. A professional encyclopedia would pay a fact-checker to confirm this (probably needs a visit to or correspondence with the Albuquerque, NM, records office). What you have to ask yourself is where your source got their information from. And then follow the trail back and confirm the details. Also, if at the time your source was writing, there was only one Albuquerque in the USA, then that is sufficient to confirm your assumption. Carcharoth 13:56, 18 October 2006 (UTC)

Detailed examples and explanations can help

Seeing that there seems to be a push here to consolidate some policies and present them in a minimum number of words (which is good to a certain extent), can I make a plea for the other side? Sometimes, in the case of complex and subtle points of policy and guidance, examples and detailed explanations really help. I have sometimes tried to suggest some clarifications and small extensions to some guidelines, but often get shouted down with the cry of "instruction creep" (though I think some people's definition of instruction creep is too broad). Can the people editing this proposed policy please keep in mind that a range of users read the policy pages:

(1) Those completely new to the policies and wanting to learn about them, and to then go away and learn from experience how they apply 'out there'.
(2) Those who are experienced in the policies and who want a quick refresher and checklist that they are correctly remembering the details and points of the policy.
(3) Those wanting to read up in detail about policies and to discuss examples and to clarify any questions and issues they might have. Sometimes these will be people in the transition stage between new user and experienced user.
(4) Those wanting detailed step-by-step instructions for a process (eg. AfD).

A well-written policy page will be aimed at (1) and (2), in my opinion, and will indeed attempt to avoid instruction creep. Subset (4) are missing the point of policy pages (which are not easily delimited processes). ie. Detailed instructions for processes can be OK, but are bad for policies which require judgement, more than the ability to follow step-by-step instructions. On the other hand, examples and detailed explanations, for subset (3), really can help people understand what is going on. Of course, in an ideal world, examples and detailed explanations would be found 'out there', but it would be nice to have a set of essay pages as subpages of the policy page, expanding more on the policy. It would have to be made clear that they are only examples, and do not in any way supercede or replace the central policy page on the 'front cover'. An additional bonus is that stuff previously (sometimes unjustly) labelled as 'instruction creep', could be moved to these 'detailed explanation' subpages. Does this sound feasible? Carcharoth 16:47, 16 October 2006 (UTC)

Have you seen Wikipedia:Attribution/FAQ already?--Birgitte§β ʈ Talk 16:54, 16 October 2006 (UTC)

I hadn't. Thanks for pointing this out. This does indeed do exactly what I was asking for. Might I also suggest that this kind of "FAQ subpage" is encouraged for other policies and guidelines, and, crucially, be less subject to the 'instruction creep' enforcers (who should concentrate their efforts on keeping the main, front, policy pages under control)? Does anyone here support that? Carcharoth 20:40, 16 October 2006 (UTC)

Proper sourcing is not a wikipedia invention. Those who wish to learn proper sourcing need to look outside of Wikipedia policy pages. Further, proper sourcing for any specific article depends on the article subject matter at least as much as generic considerstions ... that's a million or so different specific micro-niches each with their own unique considerations. WAS 4.250 17:05, 16 October 2006 (UTC)

I agree entirely. Shouldn't something like this (without the 'million' exaggeration) be part of the wording of the policy? Carcharoth 20:40, 16 October 2006 (UTC)

The FAQ provides enough examples for newbies. The polciy page needs to be kep short and as simple as possible. ≈ jossi ≈ t • @ 17:18, 16 October 2006 (UTC)

Um. If you re-read what I said, I have already agreed with those wanting to keep the principal policy page as short and simple as possible. I was asking for a subpage expansion/explanation, which Birgette was kind enough to point out already exists (apologies for missing that). The first part of your response also seems to indicate that you have misunderstood what I was saying, as I specifically said that new users should in fact read the short and simple version first.

Furthermore, your response seems incomplete. You have failed to say why the policy page needs to be kept short and as simple as possible. I agree it should be kept short and simple, but a prominent note explaining that at the top of the talk page would be a great help. The note could say that people shouldn't expand the main policy document, but should write detailed explanations on the FAQ page, or suggest that they write an essay on it instead. Does this sound like a good idea? Carcharoth 20:40, 16 October 2006 (UTC)

Attribution vs verifiability

The following appeared at User talk:Jimbo Wales; apologies if it is redundant.

I can not, at least now, agree that this is not an overhaul. In the first place, the term attribution is different from verifiability. We already get a lot cleaned out or deleted because it is not attributed (but is verifiable and sometimes attributable if one just does a google search), but today it's done mostly to questionable information. As a concept, attribution implies attribution to anyone, be it a recognized science publication, a cheap ufologist newspaper, or just a myspace celebrity. The concept of verifiability implies more mature and balanced approach of using sources as means to the end, with recoginzing both that information may be verified even if not immediately attributed, and that just attribution is not enough, with many sources having very low publication standarts.

Actually moving from verifiability to attribution as the concept would turn Wikipedia towards plain collection of quotes and links. Renaming policy while keeping old concept would make the title misleading. CP/M^comm |Wikipedia Neutrality Project| 17:45, 16 October 2006 (UTC)

The quoted user believes there exists a significant difference in the terms; and that this proposal has the effect of moving us from a model of having grounds to actually believe something is true; to simply reporting whatever someone else said, regardless of the source.

I'm not sure I agree with the comments--nonwithstanding the present "pop culture" debate; we still require that sources be reliable. Links to "ufologist newspapers" and MySpace are clearly verboten by this policy; he may be giving too much weight to the title and not enough to what the policy says.

Thoughts? --EngineerScotty 17:57, 16 October 2006 (UTC)

--EngineerScotty 17:57, 16 October 2006 (UTC)

I agree. The objections that CP\M is stating are well covered in the formulation of this proposal. I would suggest that CP\M re-reads the text of the proposal. ≈ jossi ≈ t • @ 19:42, 16 October 2006 (UTC)

I've read the text. Objections primarily refer to the title. Sorry for not posting it here first in a complete version. CP/M^comm |Wikipedia Neutrality Project| 19:48, 16 October 2006 (UTC)

I assume thenm, that you do not have major objections to the text in the proposal. If the objection is the title, do you have any suggestions for a better title? ≈ jossi ≈ t • @ 19:51, 16 October 2006 (UTC)

A better name is Wikipedia:Verifiability. I can not comment on the text right now, as I have not yet thoroughly compared it to the existing policies. I'm not a conservatist, but sometimes evolutionary approach actually makes sense. CP/M^comm |Wikipedia Neutrality Project| 20:15, 16 October 2006 (UTC)

Title: Attribution vs. Verifiability

While there are some good ideas in this proposal, I don't find the new name to be a particular improvement.

In the first place, the term attribution is different from verifiability. We already get a lot cleaned out or deleted because it is not attributed (but is verifiable and sometimes attributable if one just does a search), but today it's done mostly to questionable information. Will focusing on attribution instead of verifiability serve the purpose, reliability of Wikipedia?

I'll use here the pure concepts, not implementations. As a concept, attribution implies just attribution to anyone, be it a recognized science publication, a cheap ufologist newspaper, or just a myspace celebrity. The concept of verifiability implies more mature and balanced approach of using sources as means to the end, with recoginzing both that information may be verified even if not immediately attributed, and that just attribution is not enough, with many sources having very low publication standarts. This is the one side: loss of correct, but unattributed information.

There's another side. Let's suppose that some newspaper reports an UFO landing in Newport News. As long as there are no other sources, writing "Newspaper X claims there is an UFO in Newport News" fits both verifiability and attribution. But let's suppose that tomorrow some editors, knowing themselves that there was no UFO at the reported time, notice that, and no other source attests the UFO claims. Under verifiability concept, we can just delete this data as false, marking the newspaper as unreliable source. However, under attributionnothing changed: report of the claim is to be kept as attributable. Essentially this allows false data to be kept until some more credible source disproves it - which it might never bother to do. This is an exaggeration, as editors will apply IAR, but a good rule is one that already recognizes editors' sense and doesn't require ignoring.

Of course, there are probably positive things to attribution as well: a published movie interpretation might be inacceptable as verification, but attributable, so this would allow us to keep review-like information; well, it might really be useful for popular culture. However, the current policy still allows us to post that information, only enforcing form clearly marking this as a claim. Verifiability retains the distinction between verifiable facts and attributable claims: facts can be reported as facts, claims must be described as claims. Attribution concept eliminates the distinction, treating facts and claims equally. If one starts rule-creeping, it means that editors should as well treat facts and claims equally. It can result, depending on approach taken, either in reporting facts as claims, like "Elementary Physics textbook by J.W. claims that water under normal conditions is a fluid", or in reporting claims as facts, like "This movie is interesting just because it discusses the nature of man ^[114]"."

Yes, editors can apply common sense when choosing the form, but they will have no clear policy behind it, which frees the hands of ones wanting to misrepresent data without actually violating policy. Verifiability has no such problem. Rule creep is bad, but a good rule should be as harmless as possible even if creeped. And a really good rule should not have the name implying the opposite.

Actually moving from verifiability to attribution as the main concept would move Wikipedia towards collection of quotes and links rather than an encyclopedia of filtered and verifiable information. Renaming policy while keeping old concept of verifiability would make the title misleading and not self-explanatory, which has no point. CP/M^comm |Wikipedia Neutrality Project| 19:21, 16 October 2006 (UTC)

It is not only about attribution, but "attribution to a reliable source" ≈ jossi ≈ t • @ 19:38, 16 October 2006 (UTC)

Via edit conflict: I was wondering about this myself. Sorry to hop in midstream here, but is there a reason the new policy shouldn't still be called "verifiability?" I would think that verifiability implies "attribution to a reliable source" rather than just attribution -- it says more and offers a better single-word description of what we're shooting for. -- (Lee)Bailey ^(talk) 20:08, 16 October 2006 (UTC)

The word "verifiability" is confusing, as CPM's post shows. CP, can you say more about what this sentence of yours means: "We already get a lot ... deleted because it is not attributed (but is verifiable and sometimes attributable if one just does a search) ..." Can you give an example of something that is verifiable (within the meaning of the policy) but not attributable? SlimVirgin ^(talk) 01:11, 17 October 2006 (UTC)

Yes. What comes to mind immediately is articles about popular culture. Facts like "M1A2 in Operation Flashpoint has only one chaingun and needs only 3 crewmen" are easily verifiable by using the primary source, yet often not attributable, and they sometimes do get deleted already unless someone defends the appropriateness by citing the policy. This is what we'll lose quickly, but there is more to this. It may sound somewhat silly, but the fact that vehicle's ground pressure equals tire pressure is verifiable by taking a look inside elementary physics textbook and reading the third Newton's law chapter, yet it can't be attributed neither to the textbook nor to Isaac Newton. I'm not making things up, the latter case is recent. There's a lot to this. Even a calculation 552*117=64584 can't be attributed, yet is a thing easily verifiable by any reasonable adult. We shouldn't lose this line of the policy; easily verifiable things are allowed to have no sources; changing this would be a major change. CP/M^comm |Wikipedia Neutrality Project| 11:41, 18 October 2006 (UTC)

SlimVirgin, this is just my take, but while I can't think of an example of anything verifiable but not attributable, I can think of many things that are attributable but not verifiable -- the obvious example being things "attributed" to poor sources. Verifiability is stricter, both in the common english sense and in the wiki-sense, because there's a relationship between verifiability and truth, which "attribution" has no connection to. Verifiability implies a higher standard. I'll stict by my previous statement that attribution is the means by which we seek to achieve verifiability in Wikipedia, but verifiability is still the guiding principle, which I'm uncomfortable with losing as a pillar of the project. -- (Lee)Bailey ^(talk) 02:11, 17 October 2006 (UTC)

Lee, our policy on sources has nothing to do with truth, which is why V was a bad name for it, but with whether material has been published by a reliable source. That's the policy no matter what it's called. So verifiability is not stricter; it's just a poor choice of word. SlimVirgin ^(talk) 03:14, 17 October 2006 (UTC)

Respectfully, I disagree. Verifiability is a relative term. In Wikipedia, we don't verify information by direct observation, we verify it by evaluating claims made my others in various reputable sources. That's still a form of verification. It's a limited one, but, without drifting to far into epistimology, all forms of verification are limited to one degree or another. The term is generally understood that way as well; I doubt many editors are deeply worried that in order to 'verify' their Wikipedian claims, they'll first have to prove the universe exists.

"Attribution", on the other hand, has the potential to be taken too lightly, or entirely misunderstood. Without specifying that we mean "attribution to reliable sources", it fails to suggest to the new user or casual passer-through that we have standards regarding what we print -- it sounds like "giving credit", which is not the point at all.

"Verifiability" might be a bit vague, but it's at least evocative of the idea that Wikipedia has standards related to supporting claims with evidence. And that's pretty much what the policy is: our standards regarding supporting claims with evidence. -- (Lee)Bailey ^(talk) 05:43, 17 October 2006 (UTC)

A problem with verification as a title or a concept, using it in the general sense without any Wikipedia spin on the meaning, is that it includes readers doing original research to verify a statement. This is, of course, perfectly acceptable, so long as the readers don't add the original research to the article.

With respect to over-attribution, as far as I know publications that make the greatest use of inline citations are academic journals, and these do not cite common knowledge, nor do they usually have a cite for every sentence. Outside of Wikipedia, there is no perception that this is necessary, so there will be no such perception unless a Wikipedia policy creates the perception. So long as the proposed attribution policy does not encourage people to provide an inline cite for every sentence, there is no outside force making people think they have to. --Gerry Ashton 20:05, 16 October 2006 (UTC)

The problem Gerry mentions with "verification" still applies to "attribution". I can "attribute" original research to myself -- in fact I think "attribution" makes that sound like more of an option, not less. Verifiability still says more, I think, in terms of demanding that the source for a statement be a good source, not just any old source. -- (Lee)Bailey ^(talk) 20:14, 16 October 2006 (UTC)

Things are not as simple as a list of a good sources. The generic concept is covered above, and NOR is already a policy. However, Attribution to a reliable source is still impractical as concept and misleading as title or nutshell description. One of the specific areas is culture.

I think everyone here agrees that interpretations, reviews and acceptance are one the key aspects of every article about works of art, including popular culture. However, it should be noted that most of the times the only available sources are not reliable. They are not reliable by any standart, as they do not even attempt to be so, offering opinions and not facts. Bus still we need them.

Verifiability concept resolves the problem, allowing to use these sources, but explicitly defining them as opinions, since the fact that a reviewer wrote something is verifiable. Limiting attribution to reliable sources does not.

It can be addressed in the policy text, but it would be better to keep the name which reflects the concept and content. CP/M^comm |Wikipedia Neutrality Project| 20:31, 16 October 2006 (UTC)

I don't see the distinction between attribution and verification, in this context, that CP/M does. That false information may be referenced to a source (always a danger) is no more likely under the proposed policy than under the verifiability policy. Verifiability never did mean verification of the information, only verification of the source; and that policy, like the proposed one, makes this explicit in the following:

The threshold for inclusion in Wikipedia is verifiability, not truth. "Verifiable" in this context means that any reader must be able to check that material added to Wikipedia has already been published by a reliable source, because Wikipedia does not publish original thought or original research.

For better or worse, truth doesn't come into it—which is one reason, in my opinion, why the word "verifiability", given its root, is more trouble than it's worth (the catchphrase "verifiability, not truth" amounting to semantic gibberish). qp10qp 21:00, 16 October 2006 (UTC)

Well, maybe it can be a bit subtle; I'll try to repeat it concisely. It's fine as long as we only do report reliable sources. The complication is that Wikipedia is not limited to that. When writing about opinions, we may need to cite inherently unreliable sources. When writing facts covered in any textbook, we do not have to attribute every single phrase. Here verifiability allows more flexibility: unreliable sources can be explicitly attributed, and the fact that [Source] claimed [Claim] is easily verifiable, but not attributable (it can be verified by just clicking the reference link). Verifiability covers it well both by title and content. Attribution concept requires amendments and exceptions inside the policy. That's one of the reasons why verifiability covers the sense better. CP/M^comm |Wikipedia Neutrality Project| 21:17, 16 October 2006 (UTC)

How about combining the words 'attribution' and 'verifiability'? You could even toss 'reliable' and 'sources' in there as well? ie. Verifiability by attribution to reliable sources - hmm, bit of a mouthful? Still, WP:VARS has a nice ring to it! :-) How about a catchy phrase next? Editors attribute! Readers verify! (ie. editors attribute their contributions to the content, so that readers can verify the content). Carcharoth 21:05, 16 October 2006 (UTC)

FWIW, I'm rather agnostic on the title. Assuming the core content is what we seem to agree on; WP:ATT, WP:VARS, or even replacing the old WP:V are all OK with me. --EngineerScotty 21:33, 16 October 2006 (UTC)

QP, of course truth comes into it. Let's not get too epistomological; the first sentence of W:V is information must be verifiable. As a community, we're agreed upon certain means of verification, with the citing of reliable secondary sources being our main one. We're also allowed to include non-controversial observations of primary sources, make straightforward logical deductions (such as age from a birthdate), and even include unsourced general knowledge in very obvious cases, like Paris being a part of France. Point being, verifiability really doesn't refer to the verfiability of sources exclusively. It refers to the information in articles being verfiable.

That is not unsourced. The references in Paris#References actually do cover that. Please don't repeat these canards. They are wrong, and they lead to wrong conclusions. The idea that there is knowledge that is common, and yet somehow difficult to source, is a false one. Common knowledge is easy to source. Uncle G 23:14, 16 October 2006 (UTC)

Ack, I didn't think that over well enough before I started typing. I was pulling the example directly from this page: Material lacking attribution may be removed, but use common sense: that Paris is the capital of France does not need a source. True or false? -- (Lee)Bailey ^(talk) 23:25, 16 October 2006 (UTC)

I believe I agree with CP\M. Verfiability is a core principle of Wikipedia. Attribution is just a means. I don't disagree with the amount of emphasis attribution is given in this document as a means of verifying, almost equating attribution with verification. But let's not conflate them entirely; having everything we publish be verifiable is still our goal. The dictionary definition I've got for verification is a confirmation of truth or authority, or the evidence for such a confirmation. Part of our overall, spiritual goal in Wikipedia is to have everything we print be confirmably true. If it were just the "confirmably" part, we wouldn't care about using the best sources available, which is, and should be, our constant concern. -- (Lee)Bailey ^(talk) 21:47, 16 October 2006 (UTC)

But if you're going to quote dictionary definitions, all the more reason to question the phrase "verifiability, not truth": by the definition you give above, this leads to "a confirmation of truth or authority, or the evidence for such a confirmation, not truth". Show me how that makes sense. qp10qp 22:19, 16 October 2006 (UTC)

"Confirmation of truth, not truth alone." -- (Lee)Bailey ^(talk) 23:11, 16 October 2006 (UTC)

You've lost me there.

Can you point out any specific ways in which the proposed policy is less stringent than the verifiability policy?--qp10qp 11:49, 17 October 2006 (UTC)

Just to be clear, I'm not arguing about changes to the policy, just the name itself. The policy is the same, just reworded for clarity, and that's fine by me. I admit the title doesn't make much of a difference one way or another once you've read the policy, but given the endless stream of newbies and short-term contributors that we'll always be dealing with, I think it behooves us to have the name in and of itself be as clear as is succintly possible. Attribution doesn't cut it in my opinion -- in order to get across the point of the policy, it's dependant on the "... to reliable sources" clause after it. Which is why I believe the term in fine in the document, but weak as a name. The term alone suggests the idea of "giving credit" as much as documenting sources, which is misleading. Granted, everyone who's going to remain a member of community will need to actually read the text of our policies and learn how they work, so one could argue the title is completely irrelevant and that it may as well be called "policy one". I think it's beneficial to have a descriptive name. The faster people get the general concept, the fewer redundant questions we'll be answering from now till time immemorial.

As for your above question, I'll go ahead and admit "verifiability, not truth" is a somewhat unclear phrase. I can think of ways to interpret it that make sense, but they're not obviously clear. I see that as a problem with the little catchphrase, though, rather than with invoking verifiability as a concept. -- (Lee)Bailey ^(talk) 13:29, 17 October 2006 (UTC)

Some examples

section heading retrospectively added as this bit is fairly self-contained Carcharoth 23:37, 18 October 2006 (UTC)

I don't want to jump in that forest of indenting, but I did want to answer Uncle G's question: I believe that there is some commonly verifiable information that may not be attributable. I gave examples above, like that in San Francisco there is an area named "The Avenues" that is well known to residents, but doesn't appear on any of the dozen maps or half-dozen guidebooks I've checked. Similarly, for popular card games, there's a ton of information that I believe to be true but that I would be dumbfounded to find in the kind of reliable sources we're looking for. And as another person describes above, there was a case where someone correctly used personal observation to delete a sourced but now incorrect statement about the state of a bridge. That's not to say we should necessarily make exceptions here for these odd cases, but to me they make it clear that attribution is our primary means, but not the heart of our effort. William Pietri 06:31, 18 October 2006 (UTC)

Your first example seems to be street knowledge, or verbal knowledge, that tends not to be written down. An example of this is slang words, or new words. These circulate for a while before they are documented by dictionaries and, in your example, maps (though if it is restricted to a local area, residents presumably, when talking to non-residents, remember to use the proper name, or if they forget, have to explain what they mean). Also, unofficial terms don't tend to be documented, unlike official terms. This is an important point that needs to be mentioned in the policy document. As to whether to allow such information or not, I'm not sure. The trouble is that this is definitely original research. A dictionary pays people to document the use of new words. Map-makers pay people to find out the names of places. We can't do that, so regretfully, as we can't practically verify this sort of information, this is a good example of something where Wikipedia has to wait for the professional publications to catch up and document the phenomenon. Internet memes might be different though... :-) Carcharoth 12:09, 18 October 2006 (UTC)

Your example of popular card games I can't comment on, as I don't know what you are referring to here. If you mean terminology, please see above. Carcharoth 12:09, 18 October 2006 (UTC)

Consider Hearts (game), Spades, Euchre. Useful, interesting, and personally verified by many editors, but lacking citations and probably not fixable with the kinds of sources we like to see. Much could perhaps be cited through Geocities websites from J. Random Enthusiast, but that strikes me as less reliable than Wikipedia's content, not more. William Pietri 17:13, 18 October 2006 (UTC)

Hmm. Those sort of things remind me of "how to" guides. Most of the content seems to be: (1) How to play the game (beginner level); (2) How to play the game (advanced level); (3) Variants of the game; (4) Terminology. Nothing seems to link the games out to wider society. Some of the history sections are more encyclopedic and fulfil this function, but something on organised tournaments and professional stuff seems needed. Otherwise it is doomed forever to remain an enthusiast's article about something where no non-amateur activity has yet been documented. Compare chess, scrabble, bridge (card game) and poker. I think what I'm trying to get at here is that "how to play", and "tips and strategy" should go in wikibooks on the subject, while the history and cultural impact of the games should go in an encyclopedia article. Ditto for the computer game guides people mention elsewhere. Carcharoth 23:57, 18 October 2006 (UTC)

The example of the bridge is an interesting one. Normally, for this kind of recent news, we attribute a newspaper report. If this is something that isn't recorded in any newspapers, I suggest taking a photograph might help here. This is one of the areas where original observations are allowed. If the photographer uploads the photo to Wikipedia/Commons, and gives a caption with the date, then this can be used to update the article. Unfortunately, you can't photograph someone saying "The Avenues"! :-) Carcharoth 12:09, 18 October 2006 (UTC)

I guess popular names of areas could be found in some king of local media, website, or other written source. There should be at least some website to cite. And if there really isn't, IAR consensus could be applied, just if a few editors can confirm that it's a widely used term. This isn't a normal situation, though, but rather an exception - Wikipedia, after all, is more about written knowledge. In practice mention of a local name for a specified area is most likely to remain uncontested. CP/M^comm |Wikipedia Neutrality Project| 12:59, 18 October 2006 (UTC)

I strongly disagree. That allows anyone to claim anything about names in local areas. If it really is widely used, then something has to be provided in support of that claim. Otherwise you could get situations where old names that have long since fallen out of use remain on a page and are not contested. I would suggest a note on the talk page, with an appeal asking people to find some attributable source that will confirm the name. Carcharoth 14:05, 18 October 2006 (UTC)

By practically being uncontestead I mean that correct info of this kind is unlikely to come under policy-compliance check so not much attention is required - not that it is right to insert claims or leave it uncontested. BTW, false ones would likely be contested and removed by the residents. And I'm pretty sure that at least some website could mention a widely used name. CP/M^comm |Wikipedia Neutrality Project| 17:32, 18 October 2006 (UTC)

Well, Carcharoth, that's exactly why I brought this up. I have a shelf-full of books on San Francisco, and a large stack of maps. I couldn't find a source in any of the ones that I thought might have something. We can find web sites where people use the term, but I'd hate to cite something like that. But any San Francisco resident would know what you meant if you said you were looking for an apartment in "The Avenues". I think our San Francisco article is correct to mention it, even though it seems to violate this proposed policy, and more so than a policy that is rooted in verifiability. William Pietri 17:13, 18 October 2006 (UTC)

Local names of neighborhoods might be found in transcripts of city council meetings or the website of the city street department --Gerry Ashton 19:39, 18 October 2006 (UTC)

Often, but probably not in this case, as "The Avenues" isn't a neighborhood, but more a region. The actual neighborhoods are the Richmond and the Sunset. And this particular example aside, there are enough others that I'm more interested in the principle than the specific examples chosen. William Pietri 20:36, 18 October 2006 (UTC)

Sure. But just sticking briefly with the example, I thought that estate agent descriptions of neighborhoods might be good, or property magazines. I also did a Google search to try and find some online usage of this phrase, and there seem to be some good results here. Sometimes a website is the best that can be found. Just try and find a website that looks reliable... :-) Carcharoth 23:37, 18 October 2006 (UTC)

I particularly liked this one, even though it is 6 years old. Carcharoth 23:40, 18 October 2006 (UTC)

Fact checking and media sources

I think the policy should make a clearer distinction between different parts of a mainstream media source. At the moment the suggestion is that a source (e.g. a well-regarded broadsheet newspaper) in general can be viewed as reputable if it can be assumed good fact-checking processes are in place. But while high profile news stories can expect to receive good fact-checking, other parts won't e.g. whimisical opinion columns about new fads, human interest news stories, semi-humorous pieces etc. Bwithh 21:32, 16 October 2006 (UTC)

Maybe an issue for the FAQ. Certainly, a letter to the editor, a sports column (many sports columnists are shoddy journalists), or a fluff piece in the "living" section should not be treated in the same fashion as a major news item. Likewise for "articles" vs "letters" in many scientific journals. --EngineerScotty 21:35, 16 October 2006 (UTC)

Extremely biased

Someone keeps changing "extremist" organizations to "extremely biased", describing groups that may not be used as sources. Many of the sources we use are "extremely biased." It's groups that are widely acknowledged as extremist that we want to avoid being used. SlimVirgin ^(talk) 01:18, 17 October 2006 (UTC)

Define extremist. Really. Barry Goldwater was widely ackowledged as extremist. You imply "extremely biased" is not extremist. You are not making sense to me. WAS 4.250 03:53, 17 October 2006 (UTC)

Not everyone with an extreme bias is an extremist. The White House has an extreme bias when it comes to the Iraq War. Should they no longer be used as sources? SlimVirgin ^(talk) 03:55, 17 October 2006 (UTC)

So, some extreme ultra-right politician may be quoted, as he isn't an extremist. Why then extremists may not? What's so special about them? CP/M^comm |Wikipedia Neutrality Project| 16:03, 17 October 2006 (UTC)

Current version: "A questionable source is one with no independent editorial oversight or fact-checking process, or with a poor reputation for fact-checking. This includes websites and publications that express views that are widely acknowledged as extremist, and gossip columns and sources that are entirely promotional in nature. Questionable sources should not be used, except in articles about themselves."

The White House is a very questionable source. The problem is with the black and white designation of this group which can only be used in articles about themselves versus those groups which are not "questionable sources" at all. The problem lies with the not-questionable/questionable use/don't-use labeling of sources and presciption for source use. The paragraph is too dogmatic.

Alternative: "A questionable source is one with no independent editorial oversight or fact-checking process, or with a poor reputation for fact-checking. The use of questionable sources should be limited in accord with the degree and nature of its questionableness."

Maybe someone can look at your version and the above alternative and whip up something better than either. WAS 4.250 04:44, 17 October 2006 (UTC)

Questionable sources

The "problematic sources" section has the following in it:

A questionable source is one with no independent editorial oversight or fact-checking process, or with a poor reputation for fact-checking. This includes websites and publications that express views that are widely acknowledged as extremely biased, and gossip columns and sources that are entirely promotional in nature. Questionable sources should not be used, except in articles about themselves.

I'm wondering if maybe we don't need this bullet point, because the first part of the paragraph is mostly redundant to the "self-published sources" bullet (although maybe it might be good to use a different term from "self-published"), and the rest might be better covered at WP:NPOV. Thoughts? JYolkowski // talk 01:22, 17 October 2006 (UTC)

Professional or academic

Hiding, why both? Are you saying academics aren't professionals? SlimVirgin ^(talk) 01:25, 17 October 2006 (UTC)

WAS, this would be a policy change with major implications. That's not what we're here to do. SlimVirgin ^(talk) 01:28, 17 October 2006 (UTC)

Is any "self-published source" an "indisputably recognized expert"? I rather doubt it. They can simply be "experts", when it comes to popular culture articles, but they must be "professional experts" when it comes to the pretty much everything else. The point of this policy is to streamline WP:NOR and WP:V, and allow for reasonable attribution of articles which are hard to attribute from traditional sources, not to make major changes in the policies which open the floodgates to anyone with a website who has figured out a new wrinkle on super-string theory, or knows who is really behind 9-11. Jayjg ^(talk) 03:13, 17 October 2006 (UTC)

I hope no one here thinks "professional" is well-defined. It could be anyone who gets paid, anyone who is good at what they do, or a only a member of one of the traditional professions (if I recall correctly, the clergy, lawyers, and physicians). Then is the whole issue of retired professionals. --Gerry Ashton 03:25, 17 October 2006 (UTC)

"Professional" may not be perfectly defined, but it's far better defined than "indisputably recognized expert", which is a very rare beast. Jayjg ^(talk) 03:30, 17 October 2006 (UTC)

I can't even work out what "indisputably recognized expert" means. SlimVirgin ^(talk) 03:35, 17 October 2006 (UTC)

Gerry, a professional expert is someone who is paid for his/her expertise. SlimVirgin ^(talk) 03:35, 17 October 2006 (UTC)

The third edition of the American Heritage Dictionary (unabridged) includes a definition equivalent to SlimVirgin's, but also includes "3. A skilled practitioner; an expert." --Gerry Ashton 03:42, 17 October 2006 (UTC)

"Professional" connotes a learned practitioner, whose skill and study are comparable that that expected of a physician or lawyer. "Academic" connotes someone who studies the field rather than practicing. A legal academic is concerned more with the history of laws or the underlying philosophy of legal systems than with using the law to benefit specific clients, and the legal professional has the opposite perspective. Of course, the lines blur. Many academics maintain professional practices in their fields, and many practicing professionals perform some research, but when I write or see the two terms counterposed, this is the distinction that forms in my mind. Robert A.West (Talk) 03:55, 17 October 2006 (UTC)

I'm fine with professional since it includes expert. WAS 4.250 04:08, 17 October 2006 (UTC)

When I removed it it said professional or academic. You'd have to check the archives to see where that discussion took place~~, but I have to say, I'm getting tired of this edit war over these words~~. Can someone explain to me why there is constant revision rather than discussion and an attempt at understanding concerns and reaching a compromise? ~~Either side has pretty much equal support, and I'm tired of seeing people attempt to claim primacy for their opinion.~~ I think the neutral point of view policy stops any breach of the use of the word professional as opposed to recognised. I've yet to see a rebuttal of that point~~, which I find disappointing~~. Stricken words not conducive to furthering the discussion. Apologies.Hiding Talk 13:38, 17 October 2006 (UTC)

The professional or academic was in your original creation of the page, Slim~~, so I'm perplexed you are asking me why I replaced it back into the page, and also why you removed it.~~ Stricken words that ain't nice. Sorry Hiding Talk 13:45, 17 October 2006 (UTC) . Hiding Talk 14:23, 17 October 2006 (UTC)

Rewrite of lead section

I found the lead section confusing. So I rewrote it. It will probably get reverted (and it was, before I finished writing this, not a problem, please discuss here), so I'm discussing and explaining the changes here, to see if there is any consensus for them. The edit in question is here.

Before

"Wikipedia is an encyclopedia, not a publisher of original thought: all material published by Wikipedia must be attributable to a reliable published source. This means that all facts, opinions, ideas, and arguments may be included in articles only if they have already been published by a reliable source. The threshold for inclusion in Wikipedia is whether the material can be attributed, not whether it is true. Wikipedia is not the place to insert your own opinions, experiences, or arguments.

Although everything in Wikipedia must be attributable to a reliable source, not all material must actually be attributed. Editors should provide attribution for any material that is challenged or likely to be challenged. For such material, Wikipedia must answer the question: According to whom? The burden of evidence lies with the editor wishing to add or retain the material."

After

"Wikipedia is an encyclopedia, and as such its content is based on existing, published information. It is not a publisher of original material. All facts, opinions, ideas, and arguments published by Wikipedia must be attributable to a reliable published source. The threshold for inclusion in Wikipedia is whether the material can be attributed, not whether it is true -- Wikipedia is not the place to insert your own opinions, experiences, or arguments.

Although everything in Wikipedia must be attributable to a reliable source, the level and detail of attribution varies. Editors should provide detailed attribution for any material that is challenged or likely to be challenged. For such material, Wikipedia must answer the question: According to whom? The burden of evidence lies with the editor wishing to add or retain the material."

Changes

(1) State upfront what an encyclopedia is based on: "existing, published information". This is clearer and sets the right tone for what follows.

(2) Consolidate sentence structures to make flow of paragraph clearer, and placing emphasis on re-packaging and attributing material published by others, as opposed to using your own material.

(3) Change the "not all material must actually be attributed" sentence to say "the level and detail of attribution varies", and use the example of a statement being challenged as one where detailed attributions are required. The implication here being that even unchallenged statements should be verifiable using the less-detailed attributions present in an article (this needs to be expanded upon in the main sections).

Please consider which version of the lead is easier for someone to understand if they have never read the page before. Carcharoth 01:57, 17 October 2006 (UTC)

I went through your analysis, but I honestly think the former is more understandable. We need to mention attribution up front, for example. Jayjg ^(talk) 02:50, 17 October 2006 (UTC)

Carcharoth, as a matter of interest, what's the point of reverting someone who has reverted you, with the words "Please feel free to revert"? :-)

That's easy to explain. I forgot to add "see talk page" to the previous edit summary. Not everyone I interact with uses the talk page, so consider it a dummy edit to update the edit summary. :-) Carcharoth 09:48, 17 October 2006 (UTC)

As for your edits, they introduced problems, such as that WP is "based on existing, published information." (a) "Information" means it is true, and we are not based on material that is true, but material that is reliably sourced. (b) How could WP be based on non-existing material?

So do you want to work with me to come up with something that will satisfy my concerns? I find the current wording repetitive and unclear. I'll explain in more detail below. Carcharoth 09:48, 17 October 2006 (UTC)

What we're trying to do here is merge and streamline the policies to get rid of problematic wording. SlimVirgin ^(talk) 02:54, 17 October 2006 (UTC)

OK. Lots of work still needs doing then, as lots of problematic wording remains. Let's continue this in another section. Carcharoth 09:48, 17 October 2006 (UTC)

Carcharoth wrote "The implication here being that even unchallenged statements should be verifiable using the less-detailed attributions present in an article...." I agree the main points of the article should be verifiable using the references in the article, but it may be impractical and wasteful of effort to provide any reference for peripheral statements in an article. If I were writing an article about heavy objects, and used a battleship as an example of a heavy object, I don't think I should have to provide a reference that lists the masses of battleships (unless the exact mass of a battleship were important in the article). If someone doesn't like it, let them delete the example and provide a different heavy object as an example, one that they have a suitable reference for within reach. --Gerry Ashton 03:07, 17 October 2006 (UTC)

Well, technically, you shouldn't be using examples that you have come up with. That verges on original research. To quote the proposed policy: "A battleship is a heavy object. According to whom?" If it is according to you, you are inserting your "own experience" into the encyclopedia. Admittedly a very mild example, but still something to think about. Carcharoth 09:48, 17 October 2006 (UTC)

This is not a radical rewrite of policy

WAS, please stop reverting on such a fundamental point. This proposal is not intended to change policy, but to streamline it, and perhaps to add an exception for pop culture. Even at that, we're going to have problems steering it through, because there's a lot a opposition to the pop culture exception (just because people aren't commenting here doesn't mean there's no opposition). There's almost no possibility of a general weakening of the policy being accepted. Jimbo writes regularly to the list of the need to insist on the best sources and how it may be better to have nothing than to have unsourced or poorly sourced material, and not only in blps. There's no point in trying to buck that trend. SlimVirgin ^(talk) 03:26, 17 October 2006 (UTC)

WAS, can you give an example of an "indisputably recognized expert" (odd writing), who is not a professional and who is used as a reliable source on Wikipedia in a non-pop culture subject area? SlimVirgin ^(talk) 03:33, 17 October 2006 (UTC)

I've removed the proposed tag for now. We have people trying to change, not just streamline, the content policies and engaging in revert wars without discussion to retain their changes; other people are turning up to change points already agreed on; and the kind of loose writing we were trying to avoid is starting to creep back in, so it appears the page has become unstable. This is a pity because the initial collaborative efforts were very successful. I hope we can return to them at some point, but in the meantime this is no longer a page I'd be willing to propose to the community. SlimVirgin ^(talk) 03:53, 17 October 2006 (UTC)

Let's keep the proposed tag. The fact that there are a few people that oppose this attempt, does not make the proposal less valid. Let us hear what the objections are and see if the proposal can be well defended. It needs to be able to be such to stand a chance. ≈ jossi ≈ t • @ 04:01, 17 October 2006 (UTC)

If others want to propose it, that's fine but I won't be doing it. People are starting to add the kind of writing that made RS such a mess. SlimVirgin ^(talk) 04:05, 17 October 2006 (UTC)

So let us stand firm and challenge the bloat.≈ jossi ≈ t • @ 04:13, 17 October 2006 (UTC)

It's actually pretty good. There is no reason to add a pop culture exception. Now that would be bloat! WAS 4.250 04:17, 17 October 2006 (UTC)

Without a pop culture exception, Phil Sandifer's and others' concerns aren't addressed. If people are fine with that, that's okay, just as long as we understand that. SlimVirgin ^(talk) 04:30, 17 October 2006 (UTC)

Can the pop culture issue be addressed in the FAQ? Would that satisfy Phil's concern? ≈ jossi ≈ t • @ 04:32, 17 October 2006 (UTC)

We can't have something in the FAQ that contradicts the policy. SlimVirgin ^(talk) 04:35, 17 October 2006 (UTC)

I agree with you that this is no longer in shape to be a proposal, but maybe things will settle down shortly. Regarding the indisputably recognized expert outside of pop-culture please read Wikipedia_talk:Verifiability/archive13#How to handle this case?. I really believe not having a way to handle that case in this policy will end up being a deal-breaker. Please encourage the opposition who are not commenting to join in here so we can reach some kind of compromise. I understand the that this is not an attempt to rewrite policy, however there should be room in this new draft to address the perennial problems from the talk pages of the older policies. I really believe the true policy here is not any wording that people revert war over but the wider consensus which is reaffirmed everyday in all kinds of articles on WP. Consensus in this project is clearly in favor of WP being a well-sourced and reliable resource, however it is also clear that consensus is that WP will contain all kinds of articles that will never qualify for being included in the key subjects of Wikipedia 1.0. I do not believe this policy will succeed if it turns its back on either one of those consensus opinions.--Birgitte§β ʈ Talk 04:29, 17 October 2006 (UTC)

Birgitte, some of the opposition that has e-mailed me is too fed up to comment. Can you give the principle of the case you linked to, as it's a lot to read. SlimVirgin ^(talk) 04:34, 17 October 2006 (UTC)

If people are going to seriously refuse to comment this is basically dead. I do not know how to sum up the principle there but I will repeat the original post which describes the problem:

Let's say I want to contribute to the article on the Jefferson Nickel. As a reference, I want to cite one of the leading books on Jefferon Nickels; The Jefferson Nickel Analyst by Bernard Nagengast. This book can be purchased by anyone with $25 to spend, and is offered by multiple vendors on the web. This book is held in extremly high regard by the community of Jefferson Nickel collectors, which is however a very small community.

However, despite this being one of the leading books in the field, it is self-published by Mr. Nagengast himself. Additionally, a search of the catalogs of the five largest libraries in the United States (Library of Congress, Havard, Boston Public, Yale, and Chicago Public) show that none of them have this book in their collections.

So, a strict reading of the Verifiability policy would prohibit using this book as a reference, even though it is the leading reference in the field. How should this be addressed? - O^O 18:49, 25 August 2006 (UTC)

If these sorts of cases are not dealt with in this policy in some manner (and I am completely open to any ideas of how to deal with them), then you will end up back to a policy which people will commonly ignore out of common sense. I completely support WP:IAR, but there is a point where it is used so often that it speaks to the fact the written policies are out of line with the community consensus. --Birgitte§β ʈ Talk 04:44, 17 October 2006 (UTC)

Call Mr. Nagengast a professional and this policy proposal lets the book be used as a source. WAS 4.250 04:57, 17 October 2006 (UTC)

I do not see how you can describe such a person as a professional. He is an expert yes, a professional no. If you are defining the term loosely enough to include Mr. Nagengast, there is no point in having it in the text at all. Who would not be a professional with such a defintion? --Birgitte§β ʈ Talk 05:02, 17 October 2006 (UTC)

Mr. Nagengast is an expert. We do not know if his self-published book makes a profit, but if it's sold by places like Amazon, it seems quite possible that it does, so it is quite possible that he is paid for his expertise. I suppose we could claim that coin collecting is a non-professional field, and thus it is impossible for anyone to be a professional coin collector, but that would eliminate self-published sources in many fields. --Gerry Ashton 05:06, 17 October 2006 (UTC)

Your above reasoning sounds overly complex. I do not imagine anyone who really insists the policy read "professional" would accept your interpretation. Solidifing a wording everyone can agree on only because they are each interpreting it differently is not something I can stand behind. I know I said I didn't care how this case can be covered, but I do care that it is covered in a clear manner that has one common interpretation--Birgitte§β ʈ Talk 05:23, 17 October 2006 (UTC)

Birgitte, EngineerScotty came up with a good exception paragraph that would cover writers like Nagengast without opening the floodgates. See below. SlimVirgin ^(talk) 05:56, 17 October 2006 (UTC)

Some articles concern topics that are notable and encyclopedic, but not well-documented by professional sources. Examples include popular culture and fictional topics. Many articles in these areas rely on self-published primary sources (e.g. posts on bulletin boards, blogs, and Usenet) and secondary sources (e.g. fan-written websites), because few other sources exist for them. Although they are self-published, these sources are often the best available for those topics. In such cases, the source may be used so long as the material was written by a known individual (including those known by a well-known pseudonym) with established expertise in the subject. The individual need not be a professional in a relevant field. Anonymous sources should never be used. Note that these sources are the exception, not the rule. When such sources are used, they should be attributed in prose in addition to having a formal citation; sources of this nature whose reliability or applicability is disputed by editors in the topic should not be used. If in doubt about how to use a source in this area, consult a relevant WikiProject for advice.

This provision applies only to topic areas not addressed by professional sources. Unorthodox or fringe theories in subject areas covered by professional writers and researchers — such as science, history, politics, religion, and current affairs — are not included in this provision.

I've added it, as it seems the best compromise suggested so far. SlimVirgin ^(talk) 06:21, 17 October 2006 (UTC)

Marskell has removed it. SlimVirgin ^(talk) 08:56, 17 October 2006 (UTC)

All challenged material needs an inline citation?

Currently the WP:CITE guideline states: "Attribution is required for direct quotes and for material that is challenged or likely to be challenged". I've taken that to mean that a source needs to be provided on request, but not always as an inline citation within the article. The current wording seems to imply that the footnotes exist within the article to serve the editors, not the readers.EricR 03:59, 17 October 2006 (UTC)

A reader needs to know who the material is attributed to. An inline citation does just that. ≈ jossi ≈ t • @ 04:02, 17 October 2006 (UTC)

I don't understand Eric's point. SlimVirgin ^(talk) 04:03, 17 October 2006 (UTC)

Yes, all. Of course. What else? Jayjg ^(talk) 04:08, 17 October 2006 (UTC)

I don't completely understand Eric's point, but it is normal in all serious writing to provide an inline citation for all quotations. If the challenged material is not a quotation, an inline citation is still a good way to answer the challenge, but occasionally there might be an alternative method to provide a reference, such as stating that all the material in the article without inline citations is based on Dan's Guide to Taxidermy. --Gerry Ashton 04:14, 17 October 2006 (UTC)

In my opinion the policy requirement that every claim of the article should come from a reliable published source, and the guideline as to which claims require an inline citation within the article itself should be treated separately. Take for example the Isaac Newton article. Every statement in the article should be backed by at least one, and hopefully more than one reliable source. But does every statement need a footnote? Should all those citation from the fact-checking page go into the article as footnotes? Wouldn't it look a bit silly to have a footnote or two per sentence, or four or six or more considering the run-on sentences prevalent in articles? If we want to be able to show reader the source of every article claim then separate reference pages or something like m:Wikicite would better serve the reader rather than some mass of footnotes that makes for confusing reading and difficult editing.

If we only provide inline citations for some statements within the article, those which would be most benificial to the reader, then isn't it just an editorial decision which need a footnote? Why should some random editor be able to come along and throw a {{fact}} tag on, say, "Newton wrote Principia" and force me to put a footnote there? Can't i just provide a citation on the talk page, or point them to a separate reference page?EricR 05:06, 17 October 2006 (UTC)

Wikipedia does not dictate the method for providing citations. There are well-regarded books that do not use inline citations, but instead put all the notes at the end of the book, and in that section, give the page number that each note applies to (example: The Code Breakers by David Kahn). Wikipedia does not have page numbers, and because anyone can edit it, there would be a great risk of any such method being disturbed by edits, but if you can find a robust method along those lines, go for it. --Gerry Ashton 05:15, 17 October 2006 (UTC)

While this policy should not replace WP:CITE, and I happily defer the details of citation to that page, Wikipedia is a general-reference work as opposed to specialist literature. As such, it probably benefits from common citation formats rather than multiple topic-specific formats. Wikipedia is also, as you point out, not a printed work. While Wikipedia is the fruit of many people's labors; a common format for presenting citations is a good thing.

There is a lot of consensus behind inline citations; including the fact that their use is a de facto requirement for any article that an editor intendes to become featured. In addition, m:cite.php is a software module which supports efficient citation generation, automatically matching the cross-references embedded in the text to the full cites found in the bibliography at the bottom. As far as details such as Harvard-style vs numeric cross-references.

--EngineerScotty 05:27, 17 October 2006 (UTC)

Which is better?

"Although everything in Wikipedia must be attributable to a reliable source, not all material must actually be attributed."
"Although everything in Wikipedia must be attributable to a reliable source, the level and detail of attribution may vary." WAS 4.250 04:23, 17 October 2006 (UTC)

The first is better. The whole point of the sentence is to make clear the difference between material that CANNOT BE attributed (is not attributable, or OR), and material that simply HAS NOT BEEN attributed. So while everything in Wikipedia must be attributable (must not be OR), not every single sentence actually has to have a source. SlimVirgin ^(talk) 04:27, 17 October 2006 (UTC)
- Can I suggest that the longer sentence you have written here does a better job of explaining what you mean than the shorter sentence, which requires the reader to stop and re-read it several times to make sure they have really understood it. It requires people to stop and understand a subtle point of the English language (the difference between potential future attribution [attributable] and actual present or past attribution [attributed]). This is what I mean by confusing sentences. I understand that you are trying to reduce the amount of verbiage, but there comes a point where a sentence or paragraph becomes edited down so much that it loses all clarity. I do realise that the lead section is trying to summarise the policy, but remember that people will read the lead section first and may not read the rest of the policy. As such, things need to be clearly explained here, and not summarised to death and squeezed into a few information-dense sentences. Carcharoth 10:00, 17 October 2006 (UTC)
The first, as per SM. ≈ jossi ≈ t • @ 04:33, 17 October 2006 (UTC)
The first; the second could be read to read that editors must make sure even the most obvious or peripheral statement is contained in one of the references, which could be absurdly laborious. --Gerry Ashton 04:38, 17 October 2006 (UTC)
- The second wording deliberately leaves vague something that should be left for individual editors to decide (the level of attribution), unless challenged. Carcharoth 10:00, 17 October 2006 (UTC)
- How "obvious" something is depends from who is doing the reading. Not all readers find the same things obvious. Therefore there must be sources against which readers can check everything in an article, obvious or not. Moreover, as Dpbsmith's example makes starkly clear, what is "obvious" or "common knowledge" is often misleading or even downright wrong. Uncle G 11:49, 17 October 2006 (UTC)
The first. Agree with Gerry and Slim. Sarah Ewart (Talk) 05:10, 17 October 2006 (UTC)
- Nothing added to discussion, hence no further comment. Carcharoth 10:00, 17 October 2006 (UTC)
On the contrary, the second is better. Everything must be sourced, without exception. The root cause of the problem here is that SlimVirgin and others are conflating style guidelines (how to cite sources in articles if the relationship between sentences/paragraphs and sources is not a simple 1:1) with content policies. The content policy is, and should remain, that everything must be sourced. That only means that readers must be able to check everything in the article against one or more of the sources cited in the article. It does not imply having to put <ref> tags on every sentence, as several editors think that it does. The style of citation is a separate matter from the content policy.
For an example of employing different styles of citation in a single article, see Vince Foster#References. Uncle G 11:49, 17 October 2006 (UTC)

What's about material which can not be attributed to a reliable source? For instance, statement like "The Flat group website published an article about relation of their faith to others ^{[ref to the article]}" can not, generally, be attributed, unless someone else wrote about it. Yet it is obviously easy to check with absolute reliability by just clicking the article. So something could be done with excessively strong wording to avoid self-contradiction. CP/M^comm |Wikipedia Neutrality Project| 11:54, 17 October 2006 (UTC)

Actually, that probably is a reliable source in the sense of verifiability, but not reliable in the sense of being independent. You can indeed attribute the material, but you can't claim it comes from an independent source. More seriously, there is a reason why Wikipedia asks for multiple sources for things. If you are only using their website as a source, then all you are doing is rewriting what they say about themselves. Looking at this in more detail, there is a difference between:

(a) stating that they have published article X
(b) quoting from the article and referencing it
(c) stating that they hold an opinion and referencing the article

(a) is non-controversial. (b) makes clear to the reader that you are quoting them. (c) is dodgy, because the opinion is presented as an editorial comment, supposedly supported by a reference, but in fact only using their opinion to support your interpretation of their opinion (you should maintain independence by documenting other people's opinion of their opinion). This all strikes to the heart of different ways of quoting and attributing, and needing to be able to distinguish between people talking about other people, and people talking about themselves. Carcharoth 12:46, 17 October 2006 (UTC)

Yes, let's say we are just discussing the group and the article. I admit I'm really rule-creeping here, but a policy must at least have no visible flaws. (c) is really not completely OK even now. But (b) wouldn't work, because the source is not reliable (today it works). And, while (a) is really non-controversial, Attribution policy doesn't allow it! Yes, there's no reliable source stating that they have published the article; it's an obvious fact, yet not allowed by the policy, as this statement isn't attributable to a reliable source.

It means that not everything must be attributed or even attributable. It would be rule-creep to delete such statement, but a good rule must be safe even when creeping it. So some change is needed to facilitate difference between controversial and non-controversial facts; not all of the latter can be attributed. CP/M^comm |Wikipedia Neutrality Project| 16:01, 17 October 2006 (UTC)

I'm sure there is a really simple way to solve this problem, but I can't think of it at the moment! Sorry. I can only default to saying that we should say that everything needs to be attributed (the idealistic viewpoint), while privately acknowledging that this doesn't happen in practice. In other words, if someone asks for an attribution, provide one, or explain why there are problems. In any case, the bit about people talking about themselves or talking about other people, seems fairly important. In other words, if the only source is the person themselves, then it is OK to provide that as a source, provided it is made clear that this is not an independent source. (ie. a qualified attribution). Carcharoth 23:03, 17 October 2006 (UTC)

In my opinion, one of the simple solutions could be saying "All reasonably questionable material must be attributable to a reliable source, while the exact level and detail of attribution may vary". Or something with similar meaning. The key point is reasonably questionable: it allows obvious or immediately verifiable statements to have no problems, while all the rest can be reasonably questioned and therefore require sourcing. Generally, if anyone questions the material, he can require sources. Varying level and detail bit will allow the further part of policy and guidelines to cover different situations without problems. CP/M^comm |Wikipedia Neutrality Project| 23:36, 17 October 2006 (UTC)

This sounds good. I'd support wording like this. Carcharoth 01:23, 18 October 2006 (UTC)

First person facts, fictions, and fantasies

We read: Material from questionable or self-published sources may be used as primary source material in articles about that source's author.

What's the logic for this? -- Hoary 06:34, 17 October 2006 (UTC)

Meanwhile, I'll say what's wrong with it. Clearly people can say just about anything they want about themselves on their own websites. As it's written now, "Attribution" says that they or their boosters or fans are licensed to republish it within WP, which we should perhaps name Puffopedia. I accept that biographical facts about people in reputable books, etc., often come from those people, and this way we're bypassing the middleman. However, as noted elsewhere in the draft page, ~~middleme~~ middlepersons perform a valuable role: probably an insufficient one, but a valuable one all the same. A second objection is that this "direct to WP" policy will further open the barn doors to claims to an article's-worth of fame by squillions of non-entities (remember the mantra: all the talk of a need for "notability" stuff is mere guidelines!), hucksters and charlatans. No: let somebody's autobiographical information first be recycled in a journal or by a reputable publisher, then allow it in here. -- Hoary 07:36, 17 October 2006 (UTC)

You might want to read the next sentence:

The material must be relevant to the person's or organization's notability; should not be contentious or unduly self-serving; and must not be used to support claims about topics not directly related to the source or about third parties.

I.e., we can only use people's own claims about themselves to support uncontroversial facts. This is almost identical to an exception currently in WP:V, and which is generally used for precisely the purposes you outline, i.e. to provide basic biographical information for people or background information on organisations. I really don't see what would be gained by requiring a secondary source for this information. If it's uncontroversial, where's the issue? JulesH 10:51, 17 October 2006 (UTC)

Well, no shinola. I'd never previously read that far down WP:V. There indeed it was. Well, if this is only a streamlining of WP:V and WP:OR, I can hardly object.

Neither would I object at all if this provision were honored as intended. The problem is the "unduly self-serving" bit: If somebody's desperate to let the world know of their notability (as it seems thousands of WP editors are), what to you or me is unduly self-serving will go straight in.

I don't understand the point of this loophole. This isn't a matter of the dead (other perhaps than the very recently dead). If somebody is notable and alive his or her biographical data is likely to be written up somewhere; for what's missing, one can simply add a link to that person's site. Can you think of an example of a WP article benefits from the ability to reproduce biographical information via this loophole? (Today I've been intermittently going through Category:American photographers, and what a depressing amount of self-puffery there is in it.) -- Hoary 11:13, 17 October 2006 (UTC)

It seems to me that these citations are occasionally of use when it is the very nature of a subject's claims (independent of those claims' truth or lack thereof) is among the causes of the subject's notability. Extreme examples of this might include people such as Barbara Schwarz or Kola Boof, in which what they have said about themselves is at the crux of whatever notability they have (in that the claims have generated responses from reliable non-primary sources). Of course, such citations, if allowed at all, need to be very carefully worded (as claims rather than fact). Robertissimo 12:22, 17 October 2006 (UTC)

I hadn't heard of either of these personalities (and now that I look at them, I don't think I've missed much). Well, yes, I see your point, kind of. But in my opinion Kola Boof (for example) isn't notable for what she says but rather for media coverage of what she says: one can and should cite the latter. Meanwhile, WP is subjected to stuff like this, and this seems set to continue. -- Hoary 12:52, 17 October 2006 (UTC)

This would also cover the use of non-autobiographical self-published matierial as a primary source document, would't it? -- (Lee)Bailey ^(talk) 13:56, 17 October 2006 (UTC)

Non-hypothetical situation

A question, we are having a fun discussion on Talk:Healthcare of Cuba regarding sourcing. I'm wondering how this policy would look to resolving the dispute. One user has brought forward a critical source (published on a website) which makes the claim that, "Cuba has a high rate of alcohol consumption. It is estimated that 80% of the adult population in Cuba consumed alcohol regularly on a daily basis or consumes alcohol in binges until the supply available runs out." [1]

Now, while much of the information in the article is footnoted, this part isn't. It makes quite an extraordinary claim and in a very partisan manner. In the old WP:RS, we would have had the "exceptional claims require exceptional evidence" guideline, which under my interpretation would mean that a single source published on a website for a rather extreme claim is insufficient citation. Another more reliable source would need to be bought forward for this to get into the article.

Fortunately this dispute was resolved as several reliable sources (journal articles and WHO documentation) were bought forward which disputed this. But I am interested what should have been done according to this policy if these sources had not been found. - Francis Tyers · 08:15, 17 October 2006 (UTC)

Pop culture removal (again)

I realize the total volume on this talk is probably around 500k, and thus it's hard to consider it in sum—but in sum, there is not consensus for the pop culture addition, which would constitute the most massive change in policy introduced by this page. Most of the rest is paper-shuffling with definitions. There are, as I see it:

People opposed.
People neutral or mildly sympathetic, but who feel exception 2 covers it anyhow.
People in favour.

I worry a) that a radical policy innovation is going to be introduced through attrition (people failing to stay tuned simply because it's so big) and thus that b) many editors are simply going to have this presented to them as fait accompli.

Amongst other issues:

How are we possibly going to police "this provision applies only to topics not addressed by professional sources"? Who decides that?
For the third time, if the source in general is questionable how are we ever going to determine who a "known individual (including those known by a well-known pseudonym) with established expertise in the subject" is?

I could go on but it's all above or in the archives (and not just from me). Marskell 08:55, 17 October 2006 (UTC)

Marskell, the most massive change in policy was not the addition of this section, but the proposed extension of this section to cover the entire policy by removing that self-published experts had to be professional.

On the one hand, we have people who want to see no change whatsoever in the policies; and, on the other, people who want to see the entire policy diluted to say that any non-professional person with a website who is declared an expert by some group can be used as a primary and secondary source.

It seems to me that the reasonable course is in the middle: to leave the policies largely as they are, but to introduce an exception for pop culture and other subjects (quilting, coins, stamp-collecting) not covered by professional sources, as this section allowed. Where does the consensus lie on this issue? SlimVirgin ^(talk) 09:03, 17 October 2006 (UTC)

Debate over "professional" is a proximate concern springing from the addition in general—the "ultimate" innovation here.

I don't understand that sentence? What addition? SlimVirgin ^(talk) 09:17, 17 October 2006 (UTC)

The addition of exception 3 in the first place; once five days ago and again today. Marskell 09:25, 17 October 2006 (UTC)

The irony is that you and the other "side" agree on the exception thing. You don't want it at all; and they don't want it as an exception. SlimVirgin ^(talk) 09:32, 17 October 2006 (UTC)

How are we going to police it? Can you give me any reason not to suspect that this won't be massively misused? I read it and I'm positively frightened by the effect it might have. I don't believe for one minute it will be limited to quilting and I'm still wondering why quilting deserves an exception ITFP. One standard—if you don't meet it, don't add it. There are many things Wikipedia is NOT and there's nothing about Wikipedia that makes it incumbent upon us to forgive information that we cannot properly source.

The same way we police it at the moment, which is not at all. SlimVirgin ^(talk) 09:17, 17 October 2006 (UTC)

Agreed. And apropos of massive misuse, I'm sorry to infer that first-person vanity is similarly being given the green light (see this section above.) -- Hoary 09:39, 17 October 2006 (UTC)

Hoary, this proposal doesn't say anything different in that regard from the current policies. SlimVirgin ^(talk) 10:58, 17 October 2006 (UTC)

Also, exception number 2 does offer something, doesn't it? Do we need to follow it with a more meandering exception? Marskell 09:14, 17 October 2006 (UTC)

Exception number 2 is already in the policies, so adding it here doesn't advance us any. SlimVirgin ^(talk) 09:17, 17 October 2006 (UTC)

"The same way we police it at the moment, which is not at all." Wha? Do you mean we don't police pop culture much, so hey, it doesn't matter? My whole point is it won't be limited to pop culture, which should be clear to anybody after a nanosecond of thought.

Perhaps we don't need to "advance" in this regard. Verifiability --> attribution is an advance; the stream-lining is an advance. Marskell 09:22, 17 October 2006 (UTC)

As you know, I agree with you. I just think there might be a way to accommodate other people's concerns without affecting the policy as a whole. SlimVirgin ^(talk) 09:32, 17 October 2006 (UTC)

I did suggest some compartmentalizing above but it got lost. I'm wondering if the volume here won't allow for a proper discussion of it. A "workshop" somewhere? Marskell 09:37, 17 October 2006 (UTC)

A workshop is a very good idea. Things are getting lost on this page. SlimVirgin ^(talk) 10:58, 17 October 2006 (UTC)

Replying to Marksell's first comment in this section:

There are, as I see it:

People opposed.
People neutral or mildly sympathetic, but who feel exception 2 covers it anyhow.
People in favour.

I think most of the people who feel exception 2 covers it only think so if the word "professional" isn't used. Consensus seems to be that it is required.

How are we possibly going to police "this provision applies only to topics not addressed by professional sources"? Who decides that?

To start with I can see that we would have a lot of long discussions about it. In the end, we'll probably end up with a set of guidelines, somewhat similar to the current notability guidelines, about when such sources could be used. It will take work, yes. But I think it is worth it, because it'll allow wikipedia to cover a broader category of articles (or rather, it'll bring a broader subset of wikipedia's current articles within the scope of policy).

For the third time, if the source in general is questionable how are we ever going to determine who a "known individual (including those known by a well-known pseudonym) with established expertise in the subject" is?

By reputation. By positive reviews. By the number of people who use references to the individual's work elsewhere. By expert recommendations. By any means possible. By taking all of this information and discussing it and making a decision on a case-by-case basis. Do you see anything wrong with this? JulesH 11:00, 17 October 2006 (UTC)

"Reputation"—that tells me nothing in-itself. "Positive reviews", "the number of people who use references to the individual's work", "expert recommendations"—if we have any of these in credible sources then the dubious source isn't needed; if we have these in dubious sources, then it begs the question. This has all been stated before. Marskell 11:05, 17 October 2006 (UTC)

I've set up WT:ATTPC as a workshop page devoted to this issue. SlimVirgin ^(talk) 11:06, 17 October 2006 (UTC)

I would imagine that we are not going to police it. I imagine that the good faith, smart, capable editors in those areas will police it. I mean, in the past, delegating a lot of responsibility and trust to our editors has worked pretty well. We wrote a million article encyclopedia that way, for instance. Phil Sandifer 13:56, 17 October 2006 (UTC)

Critique of wording of lead section (continued)

Following on from the section above about the lead section (where I suggested some changes), I am presenting here my concerns about the readability of the lead section, and the potential for confusion, which may require a rewrite of the section. I am quoting the current version in italics, and interspersing my comments.

Wikipedia is an encyclopedia, not a publisher of original thought: all material published by Wikipedia must be attributable to a reliable published source. This means that all facts, opinions, ideas, and arguments may be included in articles only if they have already been published by a reliable source. The threshold for inclusion in Wikipedia is whether the material can be attributed, not whether it is true. Wikipedia is not the place to insert your own opinions, experiences, or arguments.

First sentence tries to say too much. The second sentence then repeats it, possibly to emphasise the first sentence, possibly to reword the first sentence in case people haven't understood it.
The first sentence uses the word 'encyclopedia' and the phrase 'original thought' without defining or explaining them. This is confusing. Also, the current text does not make clear what this thing called attribution means. The implication seems to be that the text is trying to explain what attribution is, but the text doesn't make this clear.
The logical flow of the paragraph is not clear. The first sentence uses a comma and colon to associate three clauses. The implication here is that the "not a publisher of original thought" is consequent from the "is an encyclopedia" bit. The "attributable" bit then jumps over several steps in the logic to focus on attribution. The first step should be to say that Wikipedia must be written using reliable published sources (otherwise it is just original research). The second step is to say that the nature of Wikipedia (anyone can edit) requires verification methods that the reader can use. The third step is to say that attributing the reliable published sources used allows verification. Simple, clear and easy to understand.
The last sentence seems to follow from the previous one (ie. even if your own experience, opinions and arguments are true, don't put them in Wikipedia because they are almost certainly not attributable to a published source, and will be original research). However, unlike the first sentence, no colon or semi-colon is used to join these related clauses. The separation into two sentences disjoints the point being made here.
The last sentence also uses the word "insert", which is a informal colloquialism, and jars with the more precise language used elsewhere in the paragraph.

Although everything in Wikipedia must be attributable to a reliable source, not all material must actually be attributed. Editors should provide attribution for any material that is challenged or likely to be challenged. For such material, Wikipedia must answer the question: According to whom? The burden of evidence lies with the editor wishing to add or retain the material.

The first sentence is hopelessly confusing. The word 'actually' causes a lot of the confusion. It would be better to spend a few sentences explaining the concept, rather than rely on people to understand the sentence as it stands. The rest of the paragraph is fine.

Wikipedia:Attribution is one of Wikipedia's two core content policies. The other is Wikipedia:Neutral point of view. Jointly, these policies determine the type and quality of material that is acceptable in articles. That is, content on Wikipedia must be both attributable and written from a neutral point of view. Because the policies are complementary, they should not be interpreted in isolation from one another.

For examples and explanations that illustrate key aspects of this policy, see Wikipedia:Attribution/FAQ.

The last two paragrahs are fine.

I hope the above makes clear why I tried to rewrite the lead section. Comments are welcome. I have spotted one small change which I will try out now. Further changes I will suggest here so that people can discuss them. Carcharoth 10:42, 17 October 2006 (UTC)

New lead paragraph

How about this?

"Wikipedia is an encyclopedia, not a publisher of original thought. This means that all material (facts, opinions, ideas, and arguments) published by Wikipedia must be attributable to a reliable published source. In other words, the threshold for inclusion in Wikipedia is whether the material has been published by a reliable source, and hence can be attributed to that source, not whether the material is true. Wikipedia is not the place to publish original thought (such as your own opinions, experiences, or arguments)."

Is that acceptable? Carcharoth 11:00, 17 October 2006 (UTC)

It sounds good to me. It fixes my main problem with the lead paragraph which was that we needed to stress that attribution had to be to a reliable source. JulesH 11:02, 17 October 2006 (UTC)

Just one note. How about (for example) "creation science"? To me, this is nutball stuff, and I'd write off any company that publishes it. But a cursory reading of the newspaper tells me that millions of people have Faith in this, and I suppose they would say the publishers were fine, upstanding, and reliable. (No, I don't have any answers for this, and wonder if it is dealt with elsewhere.) -- Hoary 11:17, 17 October 2006 (UTC)

That should probably be in the FAQ (see Wikipedia:Attribution/FAQ). The short answer is that, in conjunction with the Wikipedia:Neutral point of view policy, the editors of articles need to make a judgment of the reliability of the sources, and the relevance, presenting a balanced and neutral article that gives all side of the argument in the correct proportions, and not using their own arguments. Unfortunately, policies like this, while providing guidance, don't always help with the specific judgments needed on the ground. There will be an article on creation science. Hopefully it is not in too bad a state. But have a look at how things are handled there. Carcharoth 11:25, 17 October 2006 (UTC)

No, that's simple. Problems arise if the proponents of "creation science" want to insert their own "scientific" viewpoint into articles that aren't overtly about "creation science" itself but are on geology, evolution, and the like. The editors will be able to cite books, and the books will be published by what they and hundreds of thousands of people (if not you or me) sincerely consider to be reliable outfits. Of course this can easily shade off into a discussion of WP:POV, but hoping to avoid this, I wonder if "reliable" publishers can be defined in some way to exclude those with a PoV mission. But I can immediately think of problems with that too. (Sorry to have questions but not even tentative answers.) -- Hoary 04:49, 18 October 2006 (UTC)

The solution here is to find a counter-cite that dismantles the argument of the "creation scientists" (see talk.origins for arguments), and also to use editorial judgement as to how much text to devote to these issues within any single article, and which article the information is best placed in. A lengthy documentation of the geological arguments of the "creation scientists" might be best at a geology section of creation science, instead of at geology. A mention and link to creation science is needed at geology and other places, but that mention must not be allowed to get out of control and take over the article. You should probably also find a cite that will back up the claim that geology is mainstream science, while creation science is not accepted by the mainstream scientific community. In other words, just because material is citable and can be attributed, doesn't mean it should go in just any old article, or in any old way. The editorial balance and distribution of information between different articles is a vital part of the process. This is really a question of where information belongs, not one of attribution. ie. WP:NPOV, remembering that not all points of view should get equal amounts of space in any one article, but that all relevant points of view should at least be mentioned. Failure to mention something that is relevant is what NPOV is meant to avoid. Carcharoth 11:48, 18 October 2006 (UTC)

Many thanks for the very thoughtful reply. I'll take a little time to think through its implications. -- Hoary 01:58, 19 October 2006 (UTC)

New second paragraph

How about this?

"Although all material in Wikipedia must be obtained from a reliable published source that can be attributed, not all such material is attributed. Editors should provide attribution for any material that is challenged or likely to be challenged. For such material, Wikipedia must answer the question: According to whom? The burden of evidence lies with the editor wishing to add or retain the material."

This re-emphasises the main points, and simply states the status quo where stuff is often added without attribution, but can be challenged. The new emphasis draws the reader's attention to the fact that pre-emptive provision of attribution is recommended. Carcharoth 11:11, 17 October 2006 (UTC)

Second paragraph is excellent as is. ≈ jossi ≈ t • @ 14:00, 17 October 2006 (UTC)

Carcharoth edited the second paragraph to begin "Although all material in Wikipedia must be obtained from a reliable published source that can be attributed...." I have restored the earlier version because the new version puts an unreasonable burden on editors. While a good editor will consult reliable sources for the main points in an article, the editor might reasonably rely on memory for common knowledge in a field, if the editor is confident that virtually all college textbooks or similar basic sources in the field would confirm the information. The new wording would require to find a reference for even the most trivial or peripheral bit of information, even if the editor does not write a citation for that bit of information. --Gerry Ashton 18:10, 17 October 2006 (UTC)

I was careful not to say that all material must be attributed. I merely said that we require that it can be attributed, and I retained the "must be attributed if challenged" bit (this bit directly contradicts your interpretation). I would also suggest that relying on memory, with all due respect, is a bad way to write encyclopedia articles. It might work for the first draft of the first million articles, but after that, a bit more care needs to be taken. In any case, we know that however this policy is worded, it won't change the way basic, common and trivial information is added to Wikipedia in dribs and drabs, without attributions. The policy is there to emphasise that the goal is for all the material in Wikipedia to be attributable to reliable sources. ie. If someone checked carefully through an article, they would indeed find everything to match with what reliable sources say. Carcharoth 20:43, 17 October 2006 (UTC)

The removed version said, in part, "all material in Wikipedia must be obtained from a reliable published source" (emphasis added). This means that the editor must find every bit of information in a reliable published source. The removed version did not say all information must be attributed, but that just saves the effort of typing the citation; every bit of information must be located in a source. Of course the main points in an article should be checked in reliable sources, but all information is presented against a sea of background information that those able to understand the article can be expected to already know. For example, suppose an editor of the Transistor wanted to mention that silicon is an element. That is common knowledge that anyone can verify with a dictionionary or periodic table. Should the editor really have to stop and find a source that says silicon is an element? --Gerry Ashton 21:08, 17 October 2006 (UTC)

Good example. I think of this sort of thing as peripheral information (or secondary or background information). An article on a transistor would be expected to cover the properties of silicon, but the fact that silicon is a chemical element is not really central to an article on transistors. The correct place for a reference confirming that silicon is an element would be at silicon. This leads into the idea of whether linking to the article on silicon is enough for the purposes of the transistor article (I think it is), and how much of the information in silicon should be recapitulated at transistor (as much as the editor judges is needed, in my opinion). Maybe the idea of only requiring the central information (or primary or foreground information) of an article to be attributed might work? But again, this all comes down to editorial judgement, and the checks-and-balance system where stuff that is challenged must be attributed. But then the merry-go-round of "citation needed" tags starts again. The expert thinks the layperson is asking for excessive citations - the layperson is unsure about some bits of the article, and wants citations so he can verify the information. Ultimately, the editor must respect the layperson or fellow editor asking for attributions, but I agree that in some cases it can be argued that something does not need a detailed cite. That is why I offered the following "the level and detail of attribution may vary" (both in practice and in theory), which has been discussed elsewhere. Carcharoth 23:19, 17 October 2006 (UTC)

I agree with the points made by Carcharoth and Gerry Ashton. See also my comment in the associated page refered to at the top of this page (on pop culture) in the subsection on binary vs continuous (nondiscrete) relationships about the problem with wording things in black/white terminology. WAS 4.250 05:30, 18 October 2006 (UTC)

I tweaked that second paragraph in the lead section with this edit. Does this adequately reflect the consensus reached above? Carcharoth 23:22, 18 October 2006 (UTC)

You have several conflicts in the lead as currently written. The most glaring is:

In other words, the threshold for inclusion in Wikipedia is whether the material has been published by a reliable source, and hence can be attributed to that source, not whether the material is true.

Although all material in Wikipedia must be verifiable using a reliable published source that can be attributed, in practice not all such material is attributed.

These two statements are opposed to each other. Is it one or is it the other? It can not be both at the same time. Brimba 05:01, 19 October 2006 (UTC)

Yes, its the old "verifiable does not equal truth" thing. I wanted the word verifiable to be in there, but I see that it causes more confusion than it is worth. The principle that any editor or reader should be able to use the attributions to check the material, is an important one though. This is what verifiability means. Whether the material is true depends on whether it is a reliable source. A non-reliable source could be lying. A reliable source could be mistaken. This is all something that needs explaining in detail, but not in the main policy document, but in subsidiary pages (like the FAQ). Carcharoth 11:32, 19 October 2006 (UTC)

Operation Clambake

This is a particular source I think that it's important we keep in play, since attempts by POV-pushers to kill it dead was one of the things that broke WP:RS. It's an absolutely massive, major source of anti-scientology information - rightly the top Google ranking for Scientology that's not actually owned by the church.

It is, however, self-published by an expert who is neither professional nor academic. Nor, of course, indisputably an expert, since the CoS denies every one of his allegations.

On the other hand, it would be insane not to use Operation Clambake in Scientology articles.

Is this currently allowed under the policy? If not, how do we fix it? (Not even the proposed popular culture exception addresses this one - it's way past the uncontroversial line.) Phil Sandifer 14:15, 17 October 2006 (UTC)

How did the Operation Clambake issue break RS? SlimVirgin ^(talk) 17:49, 17 October 2006 (UTC)

It's the sole reason Terreyo took interest in the page. Phil Sandifer 17:50, 17 October 2006 (UTC)

User:Phil Sandifer, please feel completely free to discuss any personal issue of any nature whatsoever on my discussion page. Terryeo 23:56, 26 October 2006 (UTC)

I'd never heard of Operation Clambake before. If it has its own article, then I guess that says something about it. Anyway, if we are bringing up specific examples, what about the excellent Talk Origins FAQs centred around the Usenet group of that name (Talk.origins)? How is that viewed by this policy? Carcharoth 15:25, 17 October 2006 (UTC)

The policy should not make distinctions or exceptions. If the Clambake site is a good source, then it will be most probably referred to in scholarly articles, books and other reliable secondary sources. If not, then this source is no better than the Media Matters viewpoints on Rush Limbaugh. A partisan site can only be used as a reliable source for the site itself. ≈ jossi ≈ t • @ 15:30, 17 October 2006 (UTC)

"A partisan site can only be used as a reliable source for the site itself." is the current policy. In other words, you're trying to justify not putting something into a revised policy on the grounds that the current policy prohibits it. Ken Arromdee 05:07, 18 October 2006 (UTC)

As mentioned elsewhere, neutral and reliable sources on Scientology are rare. The CoS itself is obviously biased; as is operation clambake. The same is true for many contentious topics--I can't think of any sources which covers political issues which are unimpeachably unpartisan. In many cases, biases sources need to be used, opinions expressed therein attributed to where they came, and WP:NPOV applied to ensure that the result is fair comment as opposed to a hatchet job. --EngineerScotty 17:45, 17 October 2006 (UTC)

All sources are biased one way or another. But we are not discussing bias, rather, we are addressing reliable sources: Unless a contentious claim is supported by reliable sources, it should not be used. Opinions of detractors and critics can be heard, of course, but through the fact checking/reporting made by reliable sources as described in the proposal. ≈ jossi ≈ t • @ 19:07, 17 October 2006 (UTC)

CSA Illumina has 164 article finds for "Scientology". And that's only scientific journal articles. There are numerous scientific book chapters and entire books on Scientology. You call that wealth of academic, if not somewhat neutral, information rare? Fossa 15:23, 18 October 2006 (UTC)

Needless to say, I agree that the solution lies in attribution and application of WP:NPOV. ≈ jossi ≈ t • @ 19:09, 17 October 2006 (UTC)

I'm with Jossi on Clambake - as far as I can tell, its only indicia of reliability is that it's popular. I also think that the scientology pages' habit of relying on Clambake and similar self-published internet sites has led to a serious lack of rigor. IMHO, although there are several excellent editors on those pages, it would have been strongly preferable for the pages to develop from the reliable sources. (There are reliable sources on scientology, mostly published by religious studies writers, but they take more work than internet sites, and aren't as consistently anti-scientology). TheronJ 13:26, 18 October 2006 (UTC)

Hi, I am a sociologist with a specalization in new religions; I gave up editing on the English Wikipedia, because the articles on Scientology are in firm hands of the anti-cult movement, mainly because (a) hardly anyone knows anything about Scientology except that it has a very bad reputation and (b) the stories of the anti-cult movement sound plausible. Anyone not involved, may want to read Clambakes introduction on Scientology:

"The Church of Scientology is a vicious and dangerous cult that masquerades as a religion. Its purpose is to make money. It practices a variety of mind-control techniques on people lured into its midst to gain control over their money and their lives."

Please. That is an activist statement that bases in no systematic empirical research or credible theoretical approach. "Mind control", for one, is a psychological concept that has been discarded 30 years ago and anyways was only intended to be applied to tatlitarian regimes such as North Korea, who have the power to use coercion and torture on their victims, something that Scientology not even remotely approximates. The question, if Scientology is a religion or not is hotly debated (in courts, in scholarly literature, in the media) and the answer depends on your particular concept of religion. And so on ...

There is, actually, quite a bit academic research on Scientology, but, alas, it is printed on dead trees, so hardly anyone reads it. No respectable scholar would quote Clambake or similar websites (xenu etc.) as a secondary source. If there is some particular primary data bit on Clambake that might be valid to use in an article on Scientology, you will always be able to retrieve more credible information on the very same data via LexisNexis. I believe though, that you'd rather not do your own research, but you rely on scholarly materials, which is what a real encyclopaedia would do. Fossa 14:35, 18 October 2006 (UTC)

Fossa, I am currently reading Roy Wallis' book The road to total freedom: A sociological analysis of Scientology New York Columbia University press 1977 Andries 16:42, 18 October 2006 (UTC)

PS: Scientology versus the Internet is a lemma that is almost exclusively sourced through such dubious sources. Already the first sentence is plain false: This is not a colloquial term at all: It's an invention (the term, not all of the incidents that are referred to in the article) of an anti-cult activist, who sucessfully has broadcast the term via Wikipedia. (see here, basically no-one uses the term outside Wikipedia and it's clones. Fossa 14:49, 18 October 2006 (UTC)

Actually, I count about 100 separate usages on Usenet, plus around 300 websites when you count also "Scientology vs the Internet." Small, to be sure, but for a subtopic, it's a fine number of hits for that specific term - certainly there are thousands of resources on the topic. Phil Sandifer 18:54, 18 October 2006 (UTC)

I should point out that Scientology does present a wealth of problems not afforded to other topics - not necessarily things that require explicit exceptions or, for that matter, subject-based exceptions, but things to be aware of. Publishing a source on Scientology is difficult because of the extremely litigious nature of the church. It's been something that many publishers spent a long time reluctant to touch. It wasn't until the Church had their spectacular failure to combat Internet discussion of them that the information began to drift towards the mainstream. Meanwhile, the Church itself runs a relatively successful disinformation campaign. It's a problem that requires considerably more care than an obvious Scientology operative like Fossa is suggesting here.

In any case, suffice it to say that Xenu, a featured article, relies on Clambake as a source on two occasions, and to self-published material by its maintainer, Andreas Heldal-Lund for a third occasion. Phil Sandifer 18:54, 18 October 2006 (UTC)

On Xenu, some inline external links were recently "promoted" to references, and I did notice a message board link and a Usenet posting that don't make the grade. The next time someone does a reference cleanup of the page, they'll probably go. Other than that, I don't see anything wrong with the links to online copies of published books and the like via Clambake which are well and truely attributed. AndroidCat 01:52, 20 October 2006 (UTC)

Phil: I don't think you should be labelling Fossa as a "Scientology operative". I don't see any evidence for that. As far as I can see; nothing he said is disinformation or intended to deceive. Apart from anything else, he really does seem to be a sociologist specialising in new religious movements (on his blog linked from his wikipedia user page he identifies himself as this person). Alright, he may be an anti-wikipedia activist, but that shouldn't prevent us from treating him civilly. I think he's here to fix what he perceives as being wrong with wikipedia, and which a lot of people have commented on: an anti-expert bias in the editing process. JulesH 07:31, 20 October 2006 (UTC)

Ignore all rules

Okay, rather than try and bash out a set of words that is going to settle the issue of pop culture and self published sources, can we perhaps get an agreement that ignore all rules will allow exceptions to the self-published source part, and that we can accept that in certain, exceptional instances there will be stuff that can be sourced in such a manner. Those exceptions are of course, like anything, open to challenge. I'm not asking that we actually have to state this in the policy itself, I just want to be clear that WP:IAR can apply to this policy. If we can get a consensus on this, that's an acceptable compromise to me. I don't want to open a back door where any ip can claim WP:IAR, but we're all clear that ignore all rules allows us to make exceptions where they improve the encyclopedia. That we have featured articles that may violate this policy shouldn't necessarily imply this page is flawed, just that the supremacy of the neutral point of view encyclopedia takes precedence over any other concern. I hope enough people get the thrust of the point that we don't get tangled up in more lawyering on minor wording. Hiding Talk 14:19, 17 October 2006 (UTC)

I'd be completely OK with this, though I'd like some language like what I suggested yesterday added to the page, since IAR tends to be spectacularly unhelpful when dealing with processbots. If the page actually says "Of course there are some exceptions you'll run into in the real world" it becomes much easier to get a processbot to stop and think. Phil Sandifer 14:25, 17 October 2006 (UTC)

Yeah, that's my worry too. My hope was that if we got a good consensus in the talk page, and also a consensus that we wouldn't implicitly add it to the article, we'd just point any process-bot to that. Kind of keep the backdoor closed but have a security chain to check who's knocking. Maybe we could add it to the FAQ, something like Does WP:IAR apply to Wikipedia:Attribution? Only in exceptional circumstances. Hiding Talk 14:37, 17 October 2006 (UTC)

Or more accurately, "As much as it does to any other policy." Phil Sandifer 14:48, 17 October 2006 (UTC)

I think the point of IAR is defeated if other policies attempt to use it as a reference point. No? Marskell 14:49, 17 October 2006 (UTC)

As a direct reference point, yes. As an indirect one (A note that specific articles may have sourcing needs that are stricter or more lenient than this, but that such exceptions are rare and controversial), no. Phil Sandifer 14:52, 17 October 2006 (UTC)

I agree with this idea very much. But editors who care about a page can already tell administrative interferers to sling their hook, and can remove citation-needed tags, or whatever; and they do.

I think we need to wise up about the limited status of V and NOR; they aren't sets of rules but attempts to ensure the quality of articles, and they have no canonicity as such. I'm always struck when I read anything around foundation issues and principles by the fact that the originators of Wikipedia were at pains that editing spontaneity should never be inhibited by rules and regulations.

If administrators and mechanical bots take these two policies (or this proposed one) as a licence to over-police, they should be reined back. It's "ignore all rules", in my opinion, that makes Wikipedia beautiful. Pages which fail to follow policies (unless real-life laws are being broken or libel courted) should be allowed to survive insofar as they are deemed good articles by editors; Wikipedia won't explode on their account. qp10qp 15:43, 17 October 2006 (UTC)

Except that this means that we need to have a subjective criteria is what is a "good article" (and evidently the WP:GA criteria doesn't count, because I've seen articles that were in the Good Article category get consensus for deletion.). And my opinion is that this currently doesn't happen anyway. ColourBurst 18:47, 17 October 2006 (UTC)

IAR is bad for this purpose. It is for editors in case policy misses something; the reserve permission. Relying on IAR when making policy is like relying on the fire ladder to support balconies when building a house. Generally, relying on IAR when creating policy is elimination of the strength reserve, which was created for a purpose. Let's leave IAR to editors and create a proper policy which will stand without the fire ladder. CP/M^comm |Wikipedia Neutrality Project| 19:53, 17 October 2006 (UTC)

"Ignore all rules" is policy; and it has equal status with anything we may compose here. Policies, other than Wikipedia:Neutral Point of View, aren't binding. We can block as many holes as we like, but the rabbit may shoot from another one.

To Colourburst: of course we use subjective criteria to decide whether we're happy with an article or not. Panels may judge Featured Articles and Good (with a capital G) Articles by stringent criteria, but most articles are judged by the standards of the editors working on them; if those editors are happy, the article stands. qp10qp 20:55, 17 October 2006 (UTC)

Sure. It is a policy, but it's the reserve system. Like the inner hull of a ship. It won't let it sink if the outer hull is breached, but it's the outer hull that must be kept watertight in the first place. We not only can block as many holes (both for inappropriate inclusion and deletion) as we like, but we must. A ship has means to stand a few holes, but it should be built without them. CP/M^comm |Wikipedia Neutrality Project| 21:49, 17 October 2006 (UTC)

Since Wikipedia is not a bureaucracy, any policy (or guideline) can and does have exceptions. In the spirit of education, it makes sense to list some of the common exceptions on the policy page itself. >Radiant< 13:20, 18 October 2006 (UTC)

Someone else just said above that Wikipedia:Neutral Point of View doesn't have exceptions. If so, it isn't true that every policy or guideline has exceptions.

I'm currently involved in a debate at Wikipedia_talk:Ignore all rules where I point out that the current original research policy says it has no exceptions but the Ignore All Rules policy says there are exceptions to everything.

This may be a good time to clear up the confusion once and for all. Either every policy has exceptions, including OR, NPOV, etc.--or else some policies don't have exceptions, and IAR should state that it doesn't apply to everything. Either way, something needs to be fixed; the current rules contradict. Ken Arromdee 14:15, 18 October 2006 (UTC)

Technically, NPOV is not a policy but a Foundation principle, and as such beyond policy. IAR, by definition, contradicts every policy page in Wikipedia, including itself. This paradox is supposed to indicate that the encyclopedia is more important than the rules. >Radiant< 14:28, 18 October 2006 (UTC)

It's only paradoxical because nobody's willing to fix it. If everything has exceptions, then fine. Fix it by taking out the text that says that policies don't have exceptions. It's false information (you're saying the policy has no exceptions, but it does), and seriously confuses people. And we should explicitly say that NPOV, OR, etc. have exceptions, because the misunderstanding there is unusually strong and persistent.

On the other hand, if it really is true that NPOV and OR have absolutely no exceptions at all, then IAR is providing false information and *it* should be fixed instead. (If NPOV has absolutely no exceptions because it's a foundation principle, then say "ignore all rules, except for foundation principles".)

And there's a big difference between "IAR contradicts every policy page" (which may be true in some abstract sense), and "IAR contradicts a policy page that actually says, fairly directly, that it may not be ignored" (which causes real confusion and makes it hard for people to figure out what the policy is). Ken Arromdee 17:45, 18 October 2006 (UTC)

These kind of arguments are what I was hoping to avoid when I asked people to look at the thrust of my argument and not get caught up in lawyering over minor issues. Yes there are contradictions that cut across Wikipedia left right and centre. Why don't we state there are exceptions? Maybe because if you tell kids not to stuff beans up their nose, they will. If anyone really thinks there's a way of writing broad policy pages which will legislate for every and any position, I suggest they think again. Wikipedia is something like the internal combustion engine. Something which has had fixes and hacks grafted on here there and everywhere to make it work. The core of the engine has remained unchanged, but it has grown tremendously. Similarly, the core of Wikipedia, the collaborative NPOV encyclopedia, has had fixes and hacks attached. But let's not throw the baby out with the bathwater here. All people are asking for is that when an exception throws its head up, its taken as exactly that; an exception. All these pages are realistically, are pointers. Once people get experience around here they can see how the old patchwork quilt mats up, and they can see the most important thing about Wikipedia is that the encyclopedia is the goal. Dickering over whether some hypothetical situation affects some other hypothetical situation ain't important. What's important is that everyone understands what the goal is and works towards that goal. Here we are trying to get a consensus on the proposed attribution policy, and in an attempt to get that consensus a compromise is proposed that if'n we need it we can call an exception. Why? Because exceptions exist. I ain't too interested in getting hung up on the minor details of what contradicts what. Let's just try and get our heads together and work out whether we can all take it on the chin that when a good case is made, when a consensus forms, there's a case to be made that exceptions to this proposed policy can, in extremely rare instances, taking Wikipedia as a whole, happen. The discussion about whether the page reflects that or not is a whole different argument, and if there ain't no consensus on that, then people just have to accept that, and let's be honest, it ain't as important as simply agreeing the principle somewhere. As long as it's agreed on Wikipedia somewhere, there's somewhere to point people. You make a good point, Ken, but I don't see it as relevant. I don't see it as relevant because there's not an accepted consensus on when and what WP:IAR applies too. It's a case by case call. Wikipedia is pretty much a case by case encyclopedia. What we decide today, we can ignore tomorrow. Hiding Talk 18:47, 18 October 2006 (UTC)

First of all, it isn't a hypothetical situation. People are actively claiming, right now, that some policies can't be overridden by IAR. I wouldn't be complaining about the contradiction if it didn't have some effect in practice.

Second, while we don't have a consensus on exactly what IAR applies to, we really don't need to in order to resolve the contradiction. We'll never agree on whether IAR applies to a specific example; but we certainly can agree with general statements like "IAR may occasionally apply to OR/NPOV" or "IAR may never apply to OR/NPOV" or even "it is unclear whether IAR could apply to OR/NPOV". I don't think it's terribly hard to come to a consensus on one of those statements, even if it's just the last one.

Third, even by saying "exceptions exist", you're already contradicting the OR policy. The OR policy, after all, says that exceptions don't exist. (And people are interpreting NPOV to mean that no exceptions exist for that, either.) If exceptions exist, shouldn't the policy be changed to say that? Ken Arromdee 22:50, 18 October 2006 (UTC)

If you look, Ken, the point of this debate, the reason I started it, is to settle that IAR does actually apply. So I am unclear what you are arguing with me about. The point here is where a compromise is reached, and what shape that compromise takes. I'm quite happy to edit Wikipedia and when I get challenged discuss what I am doing and build and respect any consensus built. If we all agree that that's the principle behind Wikipedia, it doesn't matter at all what policy pages say. We already contradict ourselves left right and centre and the system hasn't collapsed yet. Hiding Talk 10:59, 19 October 2006 (UTC)

Well, I'm arguing that one of the two contradicting articles should be fixed to remove the contradiction. If, as you say, IAR applies, then the OR article should be fixed. Just remove the line which says the policy can't be overruled by editor's consensus. (Of course, if the attribution policy gets anywhere, the line may be gone anyway.)

It'd also help to say "when we say ignore all rules, we really mean all rules, yes, even those rules", but even not doing that and just removing the line from WP:NOR would do quite a bit of good.

And I would argue that while contradicting ourselves doesn't collapse the system all by itself, it does hurt it. The fact that contradictions don't cause so much damage that everything lies in ruins is no reason to keep them. Ken Arromdee 19:15, 19 October 2006 (UTC)

Personally, I think IAR should be in the singular: "If a rule prevents you from improving or maintaining Wikipedia, ignore it." IAR should be a context specific tool, not a general mindset. As the former, I don't think it particularly contradictory. Marskell 19:20, 19 October 2006 (UTC)

No good. WP:NOR says that *nothing* overrides original research. That implies that it cannot be ignored even in a specific context. You can't resolve it by saying that rules are only ignored in context any more than you can resolve it by saying rules are ignored with consensus, or in exceptional instances, or anything like that. The statement in WP:NOR is so strong that it doesn't allow context, or consensus, or exceptional instances, or any other loopholes. It's *absolute*. Nobody's going to look at a statement which says that it may not be overridden under any circumstances whatsoever, and conclude that overriding is okay as long as it's limited to one context that isn't part of a general mindset. Ken Arromdee 05:27, 20 October 2006 (UTC)

Fine. Argue til you are blue in the face. Most of us will keep on making changes to make this a better encyclopedia. Sometimes that means IAR. No matter what NOR says. But argue away, if you wish. The best contributors (not me) don't read these pages anyway. WAS 4.250 06:34, 20 October 2006 (UTC)

IAR is a policy. And it is does not just permit to ignore all rules. It tells to ignore them when they prevent you to improve encyclopedia. It means that you can't just go and do anything you want - it rather tells that when you surely know what is better, but rules resist, you should ignore them. This also implies editors consent that it improves encyclopedia. One of examples of fully proper use of IAR to skip NOR presented here is correcting information about presence of the bridge, even without printed source. CP/M^comm |Wikipedia Neutrality Project| 07:32, 20 October 2006 (UTC)

Actually, I don't see the bridge example as overriding NOR exactly. Using OR to justify excluding inessential information that appears to be wrong is a far lesser matter than including the results of OR: Attributability is a necessary, not a sufficient condition for inclusion. I would not, in this example, support including the fact that the bridge still carries traffic based on a personal observation, but I agree with the reported exclusion of an apparently inaccurate statement about disuse. Things would be different if the matter of current use were central to the article. Robert A.West (Talk) 08:38, 20 October 2006 (UTC)

So effectively the article could report that the bridge was closed on a certain date (if that is what happened, I'm not entirely clear whether the report of disuse was just wrong, or outdated), but would have to fall silent on the question of whether it is has since been reopened. In fact, looking back at the details, the salient point is that the bridge is now private (direct quote from further up the page). The important bit to document there is the date the bridge went private (that is surely sourcable), and fall silent on the question of susequent use of the road (storage, ocassional cars passing over it, whatever). If people really want to go and look at the bridge to see how the private owner is using it, that is their business. But it is not our business to report that, unless someone, somewhere, deems it important enough to report. Carcharoth 12:12, 20 October 2006 (UTC)

A case which bears upon this

At this time there is now an arbitration case up for Bowling for Columbine which bears upon this. The essence of the matter is that one editor is taking an extremely aggressive stance on sourcing in order to prevent the article from saying much about controversy over the film's content. My feeling is that we should not enable this kind of misrepresentation. Please visit the talk page for the article and Wikipedia:Requests for arbitration#Bowling for Columbine. Mangoe 03:39, 18 October 2006 (UTC)

This is a case of editors overstepping their duties. Wikipedia editors should not engage in describing/analyzing/commenting on primary sources, in particular when the subject is controversial. Wikipedia is not a place for original research. See also the OR mess at Voyage_Au_Pays_Des_Nouveaux_Gourous. ≈ jossi ≈ t • @ 14:37, 18 October 2006 (UTC)

I have commented here Talk:Bowling_for_Columbine#Concern_over_original_research ≈ jossi ≈ t • @ 14:42, 18 October 2006 (UTC)

Principles, not bright lines

What is needed in a one page top level policy are principles that can be used as a basis for reasoned discussion and not bright lines used as a basis for ending discussions. The former leads to sentences on the several desireable qualities of sources and how each of these qualities ranges from worst to best among the sources available and how seleting good sources often involve tradeoffs; example choosing a known biased source over a completely unverifiable source in one case and a less available source (rare book) over a biased source in another case. The later leads to WP:RS. WAS 4.250 18:41, 18 October 2006 (UTC)

Well said. Unfortunately human nature is such that it prefers rulings to thinking. It is easier to apply a rule without thinking, than to take responsibility and apply principles. The former is used to blame any problems on the system and abdicate responsibility, the latter calls for being accountable. Big difference. ≈ jossi ≈ t • @ 19:12, 18 October 2006 (UTC)

I think this is true, but not all that could be said. Are you familiar with the Shu-Ha-Ri model of learning? The basic notion I'm referring to is that novices like and benefit from rules. As you advance thorugh practice and study, you move away from rules and toward principles.

I do a lot of training of groups, and at the beginning people really want some simple rules that let them get things done, even when they understand they will eventually cast the rules aside. For example, when I started learning to ski, I just wanted some simple rules to keep my head out of the snowbanks and let me get down the mountain. Had the instructor only talked of finding my center and listening to the mountain, I would have given up, even though that's how some expert skiiers certainly talk. William Pietri 20:30, 18 October 2006 (UTC)

That is agood metaphor, but note that a page such as WP:RS is certainly not good material for a novice. This page, on the other hand, can give enough structure to a newbie, while not tying the hands of experienced editors. ≈ jossi ≈ t • @ 21:30, 18 October 2006 (UTC)

Off topic, I was a very odd skier. I learned from a friend (which isn't odd), but I didn't really learn from rules (no "bend the knees like this, etc etc", just bend left, bend right.) However, I also applied what I learned from skating into skiing (stopping in skiing is very similar, at least to me). But back on topic, I certainly think that there are plenty of people who like structure and need rules, and the FAQ should suffice for them 90% of the time, and the 10% where it doesn't apply, they can get help. ColourBurst 01:48, 20 October 2006 (UTC)

I guess if the point of a policy is to simply encourage discussion, then what you say sounds most reasonable. Blah blah blah. Talk talk talk. Are we having fun yet? However, the job of making stuff means that somewhere, somehow, decisions have to be made: discussions must end, disputes must be resolved, and work (finally!) begin. A policy that is vague or incomplete because of a desire to engender discussion will, of necessity, manufacture more conversation that article content. I'll also note that any decisions extracted from such a policy will appear to most outside viewers as random if not capricious -- it is likely that many will entertain the possibility that the policy was deliberately engineered to permit such nonsense "when the powers that be wish it", the sort of dishonesty called plausible deniability in the real world. The desire to see a list of formal "rules and regulations" follows not because people are "newbies", or lack formal training, or are just latent control freaks, but simply because people like predictable behaviour more than anything else. It is, in the end, the basis of all forms of "justice" they see around them every day in their lives. mdf 12:33, 19 October 2006 (UTC)

I agree. There's a difference between rules creep and having rules. I'd like a bright line or two. Marskell 12:44, 19 October 2006 (UTC)

There is nothing wrong with two bright lines: one marking the limits of what is conceivably acceptable and the second marking the realm of safe harbor. Statements that are accurately attributed to a third-party independent source with widely accepted credentials are safe-harbor: any editor removing such material must justify the removal under some other theory, such as relevance, NPOV, style, etc. Statements that appear unlikely to be attributed, or that are attributed to sources of doubtful accuracy are never acceptable and may be freely removed. Between these lies the realm of debate, which we have to address, and there are a host of considerations, but the majority of cases fall either within the safe harbor or beyond the pale, and we should draw those lines as close to one another as we can, for clarity and simplicity. Robert A.West (Talk) 08:57, 20 October 2006 (UTC)

Perfect. I couldn't agree more. Carcharoth 12:03, 20 October 2006 (UTC)

Another example

See User talk:DTDelaney for an example of attribution issues. It seems quite common for children of recently deceased people to offer information on that person to Wikipedia. Usually, it is nicely packaged in the form of an obituary (which I tracked down and added as a link in the article), but just as an example, what would this proposed policy say in cases that are not so clear-cut? Carcharoth 22:53, 18 October 2006 (UTC)

Well, keep Wikipedia:Wikipedia is not therapy in mind for one thing. WAS 4.250 04:32, 19 October 2006 (UTC)

Well, I mean notable people whose children might be able to provide or confirm extra information not found in published sources. The question is, do we want that extra information? Carcharoth 11:24, 19 October 2006 (UTC)

What about Waterfalls of the Havasupai Indian Reservation? I was led to this by the picture of the day. Rigorous enforcement of this proposed policy would send this up for deletion. Is that desirable? --Ancheta Wis 09:01, 19 October 2006 (UTC)

It would send it for clean-up, not for deletion. Rigorous enforcement of this policy would justify other editors asking for sources, as per the "sources needed" tag at the top of the page. It is up to the editor wanting to keep the material to explain any problems with obtaining sources. Carcharoth 11:24, 19 October 2006 (UTC)

Most of this is article is hiking "beta", which is spread mainly by word-of-mouth by those who have "been there, done that". While there are printed descriptions of this place (reviews of trips there, maybe a few paragraphs (if that) in a technical canyoneering guidebook, etc), you probably won't find as much detail as given here, and certainly the people who created this article have either moved on or have no idea where to find it, beyond their own eye witness testimony. As Ancheta Wis says, strict enforcement will gut this article to the point of deletion, and if not this article, more that are like it. mdf 12:07, 19 October 2006 (UTC)

But if people have moved on, then you are stuck with choosing between assuming good faith (which shouldn't be applied to article content, but rather applies to editor behaviour), and allowing potentially inaccurate information to remain. In any case, this is another example of something that strictly speaking, because of the way it was added, is not encyclopedic. It belongs more in a wikibook or wikitravel guide. I would say that anything that can't be verified is not encyclopedic. The stuff that can't be verified is mostly "I went here and saw this" stuff, which is clearly original research and contributing "your own experience". This is not what Wikipedia is about. It is undoubtedly interesting, probably accurate, and useful to some people, but it is not really encyclopedic. If people add stuff and move on without telling people where the stuff came from, they really should expect to return in 10 years time and find that the material has not survived (unless someone buttressed the material with supporting references). If you want your contributions to be stable and remain in the encyclopedia, then tell us (editors and readers) where you got your information from. It is as simple as that really. Even if it is "eyewitness testimony", then put that, rather than nothing. Carcharoth 13:13, 19 October 2006 (UTC)

Near as I can tell, the current proposal would deep-six any and all material based on previously unpublished eye witness testimony. Under such circumstances, I would have to delete basically every single image I've uploaded here over the last year. Note that current policy specifically exempts such from the rigors of WP:NOR; this proposal is utterly silent on this matter. Like the beta issue above, I think this would be a ludicrous result, and likely an unintended consequence of these efforts. But if strict observance of policy is to be expected -- and from your last response, I conclude it will be -- what other choice do I have? I eagerly await your detailed analysis of this proposal, to the effect that people like me would be permitted to contribute to this project under its regime. mdf 17:18, 19 October 2006 (UTC)

What? The bit about being less stringent on pictures got dropped? That's just silly. I agree that if this policy is really going to merge and replace WP:V and WP:NOR, then it mustn't miss out stuff like that. As for strict observance of the policy, I much prefer to repair and improve articles, rather than delete or gut them. What I want is for people to stop adding obscure things without sources. It just makes it harder in the long run to check things and add sources later. Carcharoth 21:33, 19 October 2006 (UTC)

If I understand this, this is a big problem for images in the public domain (I asked a similar question earlier in the Village Pump.) By definition most public domain images were created by the users, which is probably not professional (otherwise they'd be copyrighted 99% of the time), and if you don't allow them, you're basically excluding all user-created images, which means we'd have to rely on copyrighted images, which is also bad for the project. And as for people adding things without sources, that already happens 90% of the time because 90% of the Wikipedia population don't interact with policy, and no amount of policy will prevent this, due to the very nature of Wikipedia. ColourBurst 01:43, 20 October 2006 (UTC)

What? 90% of Wikipedia editors intuitively follow WP:IAR? Why did no-one tell me? :-) Carcharoth 08:53, 20 October 2006 (UTC)

Current NOR policy includes an exception for images. I proposed a paragraph (WT:ATT#What is not original research) covering images and diagrams. The touchstone is that if the image can be readily compared to copyrighted images (or the thing itself) then it is not OR. I got reverted and the discussion died. Robert A.West (Talk) 09:13, 20 October 2006 (UTC)

I would suggest trying again. There is discussion happening on this issue now, and some support for adding such a clause. Carcharoth 12:29, 20 October 2006 (UTC)

Relevant to articles written based on hiking beta, if we are discussing the hiking blog of a well-known expert climber, then I would say that the material is already covered by the self-published-expert exception (whatever we are calling it this hour). If the material has been collated, the accounts of different hikers cross-checked for accuracy and published by someone with a reputation to protect, then we have a reliable source. If neither, we have someone's unverified first-person account, and, no Wikipedia should not include that material. If that means 86ing a batch of articles people care about, I am sorry, but those articles belong in a different project. Perhaps we need a Wikia project devoted to things like hiking beta, internet memes and other things, where the intent is similar to Wikipedia, but the nature of the material requires editors to take responsibility for fact-checking. Robert A.West (Talk) 09:25, 20 October 2006 (UTC)

Simple way to assess material for inclusion, removal or moving

Following on from the above, is it not simpler to present the policy this way:

When assessing the suitability of material for an article, answer the following:

(1) Does the material in question supply relevant information?
- If not, can it be moved somewhere more relevant? Another article? Another wiki project?
(2) What is the best way to ensure future readers and editors will not question or remove the material?
- One way to stabilise information is to link it to other articles. Another way is to attribute a source for the information.
(3) Has the original editor or anyone else said where the information came from?
- If yes, investigate to verify the information and assess the reliability of the source.
(4) Can you give a source for the information?
(5) Do you think someone else might be able to give a source for the information?

If the information is not relevant, not stabilised, and the last three questions are answered with a "no", then the material should be removed, or tagged for removal. Does this way of presenting the policy look helpful? Carcharoth 14:08, 19 October 2006 (UTC)

That could be a nice addition to WP:ATT/FAQ ≈ jossi ≈ t • @ 14:24, 19 October 2006 (UTC)

Primary sources

The current version lacks material about primary sources. Specifically, these are, according to existing WP:NOR:

Primary sources present information or data, such as

archeological artifacts
photographs (but see below)
historical documents such as a diary, census, video or transcript of surveillance, a public hearing, trial, or interviews
tabulated results of surveys or questionnaires
written or recorded records of laboratory assays or observations
written or recorded records of field observations
artistic and fictional works such as poems, scripts, screenplays, novels, motion pictures, videos, and television programs (whether recorded in digital or analogue formats).

Secondary sources present a generalization, analysis, synthesis, interpretation, explanation or evaluation of information or data from other sources.

Research that creates primary sources is not allowed. All articles in Wikipedia should be based on information collected from published primary and secondary sources. This is not "original research"; it is "source-based research", and it is fundamental to writing an encyclopedia. Articles which draw predominantly on primary sources are generally discouraged, in favor of articles based predominantly on secondary sources.

The current attribution proposal completely omits this. However, unless we are doing a major overhaul of the principles, these sources must be mentioned in the policy. In current state, it is impossible to use primary sources, and such rule would be very different from the current policy. CP/M^comm |Wikipedia Neutrality Project| 14:25, 20 October 2006 (UTC)

Could you point to the section on the proposal where it states that it is impossible to use primary sources? ≈ jossi ≈ t • @ 14:37, 20 October 2006 (UTC)

The lack of mention of these sources among allowed ones. For instance, movie may be used as a source for describing a movie, game to describe a game, etc. New policy shouldn't change that. CP/M^comm |Wikipedia Neutrality Project| 18:37, 20 October 2006 (UTC)

The current proposed Attribution policy says "Reliable sources are credible published materials with a reliable publication process; and their authors are generally regarded as trustworthy or authoritative in relation to the subject at hand." I don't see why a movie or game would not fit this description. --Gerry Ashton 18:48, 20 October 2006 (UTC)

If you think well, they could. If you don't, it's easy to throw them out using the policy. It already is frequently attempted, or, for articles lacking policy-experienced editors to defend them, done. CP/M^comm |Wikipedia Neutrality Project| 19:07, 20 October 2006 (UTC)

It's clearly OK to use reliable, published sources—so any primary sources may be used if they are reliably published, though any comment on them must also be from a reliable, published source. The present text wisely, in my opinion, avoids conflating published and unpublished primary sources as if they're the same animal and marks an advance on the wording in the existing policy texts. There are reams of discussion about this higher up the thread.

It's not too late to reinsert definitions of primary, secondary, and tertiary sources, but it needs to be well worded if it's not to reintroduce contradiction and muddiness. This is only a proposed policy, so propose away: what wording would you suggest? qp10qp 14:47, 20 October 2006 (UTC)

Trouble is, it's not an advance if the wording, or, dare I say it, sourcing of various sentences of the policy, are not stable enough, or explained enough, to prevent well-meaning people coming along and changing things later. It is not reasonable to expect people to find things on a massive talk page. A case maybe for inserting lots of footnotes to demonstrate to people how to show where the information is coming from in a document stating how people should show where they get information from? Oh, the irony! :-) Carcharoth 15:34, 20 October 2006 (UTC)

PS. Is this talk page a "reliable,published source" for what goes in the policy document? :-) Carcharoth 15:37, 20 October 2006 (UTC)

I do not think you should worry about well-meaning people coming along and changing things later unnoticed. There is always more details in the FAQ.--Birgitte§β ʈ Talk 16:26, 20 October 2006 (UTC)

I'm splitting off your initial comment, to make it easier to reply to it. My point works both ways. People might make valid points later, and get reverted without discussion, in a worthwhile attempt to keep the main policy statements under control. It is extremely hard for people to come along later and make any sort of substansive change after a big discussion like this is over, so I think there should be some sort of attempt to make clearer exactly how consensus was reached for certain crucial bits of wording, to avoid repeating the same endless discussions. This is what I meant by stability of the document. The FAQ is good as far as it stands, but currently if someone doesn't quite understand the reasoning behind the phrasing used in a particular sentence or paragraph of this policy, they have the choice to either: (a) change it and hope the change is acceptable; (b) ask on the talk page and hope that the right people are around to answer the questions; (c) try to find the relevant bit in page after page of talk page discussions; (d) see if the FAQ answers their questions. Many people here know how easy it is for a policy page to get out of control as it evolves over time, built up and added to bit-by-bit. If slightly more organisation was applied to presenting the reasoning behind the consensus bits of text, then a lot of future discussions and edit skirmishes could be avoided. I seriously think judicious use of footnotes might help. I'm sure I've seen some policy or guideline pages using footnotes, and I think it really can help, both to stabilise the document, and to shape and guide future discussion, as well as reducing repetition of certain discussions. Maybe not right now, but maybe when the document is more stable, some of the crucial bits of wording can be explained in footnotes without disrupting the flow of the document? Carcharoth 00:57, 21 October 2006 (UTC)

Still I think we could a paragragh to What is not original research along the lines of: All articles in Wikipedia should be predominately based on information collected from sources which present a generalization, analysis, synthesis, interpretation, explanation or evaluation of information or data from other sources. This is not "original research"; it is "source-based research", and it is fundamental to writing an encyclopedia. Additional information may gathered from sources which are used as primary sources to merely present information or data. Primary source material should only make up a small portion of the article and it requires extra care to avoid crossing the line into original research. More details about primary sources can be given in the FAQ. I would rather not use the term secondary source. According to the Questionable Sources section, questionable [secondary] sources may only be permitted when used as primary sources. So I believe that while we can adequately cover how to use something as a primary source in the FAQ, bringing in the term secondary source will lead to a great deal of confusion.--Birgitte§β ʈ Talk 16:26, 20 October 2006 (UTC)

FAQ is good, but never a substitute for policy - it has no status by itself. The paragraph would fit, however, I'd remove the "small portion" bit: the original WP:NOR states that some articles may be based entirely on primary sources. The problem with primary sources is not that they are worse or less reliable. They are as reliable or more. The problem is that they often don't present conclusions, evaluations, interpretations, etc., and these are required for a good article; so they are not recommended as editors might tend to make their own synthesis when using primary sources. Properly used, they'll usually produce a sourced and reliable, but not very informative or well readable article. CP/M^comm |Wikipedia Neutrality Project| 18:37, 20 October 2006 (UTC)

I was just about to raise the exact same concerns as CP/M; NOR explictly mentions that some articles may rely entirely on primary sources. It also mentions that only published primary sources are acceptable, which probably deserves a nod here, even the if FAQ contains more detail. Attempting to consolidate from NOR and some of description of primary sources in RS, I came up with: Most articles in Wikipedia should be predominately based on information collected from sources which present a generalization, analysis, synthesis, interpretation, explanation or evaluation of information or data from other sources. This is not "original research"; it is "source-based research", and it is fundamental to writing an encyclopedia. Additional information may be gathered from sources which are used as primary sources to present information or data. Primary sources may be used only if they have been published by a reliable publisher, and only to make purely descriptive claims easily verifiable by any reasonable adult without specialist knowledge. Contributors drawing on entirely primary sources should be exceptionally careful to comply with both conditions. -- (Lee)Bailey ^(talk) 18:56, 20 October 2006 (UTC)

I think less distinction could be made, rather just stating that no interpretations should be derived basing on primary sources, and other sources should be used for them; if there are no such sources, only descriptive claims should be made. I don't find primary/secondary terms too confusing, but a better wording could help. One specific important thing is to mention that source works falls into primary sources. There is a deal of misunderstanding about that today. Some people make excessive interpretations, some others (and not too few) are running deleting descriptive claims (mainly about fiction) stating they are not published. So a few example words would help as well. CP/M^comm |Wikipedia Neutrality Project| 19:11, 20 October 2006 (UTC)

Perhaps primary sources could be more explictly mentioned under the description of reliable sources under "Key Principles --> Wikipedia articles must be based on reliable sources" instead? That way it would be clear that things like "no unpublished synthesis of published material" apply without spending the extra ink singling primary source documents out. -- (Lee)Bailey ^(talk) 20:58, 20 October 2006 (UTC)

Probably. Yes, I guess it would work, and explanation could be presented in FAQ. CP/M^comm |Wikipedia Neutrality Project| 21:02, 20 October 2006 (UTC)

So, say the reliable sources text becomes:

Reliable sources are credible published materials with a reliable publication process; and their authors are generally regarded as trustworthy or authoritative in relation to the subject at hand. Published means the material has been disseminated in print or digital format. How reliable a source is depends on context. 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. What these have in common is process and approval between document creation and publication. As a rule of thumb, the more people engaged in checking facts, analysing legal issues, and scrutinizing the writing, the more reliable the publication. Material that is self-published is generally not regarded as reliable, but see below for exceptions.

And say we add this at the end of "What is not original research":

In cases where the subject of an article qualifies as published material, the subject of an article may always be used as a source for purely descriptive claims about itself, provided no generalization, analysis, synthesis, interpretation, explanation or evaluation of information is made.

...we could follow up with examples, either directly after the above, or in the FAQ. Would that cover it? -- (Lee)Bailey ^(talk) 20:51, 22 October 2006 (UTC)

I think it will work. Has anyone else any objections or suggestions? CP/M^comm |Wikipedia Neutrality Project| 21:19, 22 October 2006 (UTC)

How is it possible to make a summary of published material (be it a literary work or a piece of scholarly research) without producing a "generalisation/synthesis"? As far as I can see, the only alternative is direct quotations. Maybe I am missing something, but this would make a guideline that is impossible to follow. --Anonymous44 16:07, 26 October 2006 (UTC)

To me, the distinction is clear. Original research expands the boundaries of knowledge, however slightly. A generalization derives a new rule from specific cases, and accordingly introduces a new idea that may be of use in new cases not previously considered. A synthesis takes two or more ideas and produces a new idea that contains aspects of each, yet differs from both. In contrast, a summary merely reduces a larger amount of material to its essentials. There is research and original effort involved in the selection and presentation, but no new ideas are presented. A survey, of necessity, contains research into what the significant positions are and how much support each has, but that information should be solidly based on published data, and that is not what is being prohibited. If you insist on a formal analysis, deciding whether a summary is fair or a survey reasonably complete takes no special expertise. That answers the fundamental question: "Why should anyone believe this?" Since 75% or more of all Wikipedians are capable of performing such analyses, the consensus process provides a form of review. So far we trust ourselves, and no farther. Robert A.West (Talk) 23:01, 26 October 2006 (UTC)

The boundary isn't so clearcut to me. In a sense, anything is an original interpretation, every single new sentence is, no matter how descriptive it could seem - that's how the human mind works. Now, to "reduce material to its essentials", you have to analyse that material and decide what its essentials are: you describe "specific cases" with a "general" term, which in turn is also "a new idea that contains aspects of each". For example, Jesus Christ Superstar begins with the statement that the rock opera "highlights the political and interpersonal struggles of Judas Iscariot and Jesus." This involves the designation of a number of specific scenes involving disputes etc. with a couple of general terms ("political and interpersonal struggles") - which requires an analysis and classification which may be correct - or not. In vampire fiction, Lord Byron's specific lines depicting a vampire in The Giaour were described by one editor in general terms as showing "the combination of horror and lust that the vampire feels and the concept of the undead passing its inheritance to the living". Apparently, the author felt that this is an obviously correct, descriptive interpretation of the lines in question, but I didn't feel that way and only left a mention of "the inward agony of the vampire", another general expression, while a third editor preferred to expand this and wrote that the work "alludes to the traditional folkloric conception of the vampire as a being damned to suck the blood and destroy the life of its nearest relations". In all cases, it's all about interpretations; in many cases, the correctness of the interpretation is obvious, and in others, it is not - hence in the second case, we may speak of OR. --Anonymous44 11:40, 27 October 2006 (UTC)

(unindent)Defining published as "material has been disseminated in print or digital format" is not acceptable for several reasons. This definition excludes LP records, cassette tapes, VHS vidio tapes, information available from automated telephone information services, among others. Also, "disseminated" is to vague; it could include secret material distributed to selected persons. In short, attempting to define "published" is a big deal and it should not be tucked into the middle of a paragraph about something else. --Gerry Ashton 21:21, 22 October 2006 (UTC)

That's fair, but I still think an explict definition of "publication" needs to established, so it can't be interpreted in such a way that excludes everything other than print. I admit, I was not thinking of analog recordings. Perhaps we can come up with a more inclusive definition and move it to its own paragraph? Incidentally, I assumed the "reliable sources" section was an approprite place for such a definition, because the present text includes "published" as part of the definition of "reliable source". -- (Lee)Bailey ^(talk) 22:40, 22 October 2006 (UTC)

Actually, since I'm at a loss as to how to go about defining "publishing materials" at the moment, would it be acceptable to include at least the latter part? Ie,

In cases where the subject of an article qualifies as published material, the subject of an article may always be used as a source for purely descriptive claims about itself, provided no generalization, analysis, synthesis, interpretation, explanation or evaluation of information is made.

or even:

"In cases where the subject of an article fits Wikipedia's definition of published material as it relates to the statement that "reliable sources are credible published materials with a reliable publication process", the subject of an article may always be used as a source for purely descriptive claims about itself, provided no generalization, analysis, synthesis, interpretation, explanation or evaluation of information is made.

...? I'm really just trying to get the ball rolling here. The second option might not be clear as day, but at least it covers our bases, and perhaps it could be clarified in the FAQ. -- (Lee)Bailey ^(talk) 23:14, 23 October 2006 (UTC)

I put some ideas together regarding primary sources in a proposal, Wikipedia:Use of primary sources in Wikipedia - don't know if that would help out in the discussion above. --Francis Schonken 12:39, 24 October 2006 (UTC)

I have another question: I understand the desire to avoid the use of the words primary and secondary, I just want to get a clarification on something that I asked on WP:V and WP:RS and I want to make sure it still holds with this new proposal. I was told it was okay to use primary documents housed at a public institution, such as the National Archives or Duke University, as long as I was not interpreting the contents. Does that still hold? The reason I ask is, it's verifiable, but not easily verifiable. I was told previously that ease of verification doesn't matter. I just want to know if this still holds true under this new proposal. --plange 03:04, 29 October 2006 (UTC)

You say "I want to make sure it still holds with this new proposal." Please, everyone, be assured this policy proposal seeks to reword only and not to change existing best practices. Policy as practiced does not change with this proposal, only its wording. WAS 4.250 05:04, 29 October 2006 (UTC)

Gotcha, then does the wording as it currently stand allow the practice of this "best practice"? I guess I was both making sure it still stood, and if so, that the baby wasn't thrown out with the bathwater in the re-wording. --plange 05:47, 29 October 2006 (UTC)

Sources requiring expert interpretation

I have boldly taken a crack at this subject, avoiding the primary/secondary distinction, because it has a different meaning in the sciences than in historiography. The real issue is whether a source needs explanation or not. That will, of course, come down to a judgment of the editors in question, but that is unavoidable. Robert A.West (Talk) 07:26, 25 October 2006 (UTC)

This doesn't quite cover everything. Quotations and extracts make sense as acceptable material, but NOR also allows for "purely descriptive claims" about primary sources; we should not have to either quote a specific passage or cite a critic for every single fact in the plot summary of a work of fiction, for example. This should be appropriate even if you don't have "attributable interpretations" present in the article, as long as you're not doing any interpretation yourself -- having no interpretation is a valid option, in some cases. -- (Lee)Bailey ^(talk) 08:31, 25 October 2006 (UTC)

I quote from WP:NOT#Wikipedia is not an indiscriminate collection of information.

Plot summaries. Wikipedia articles on works of fiction should contain real-world context and sourced analysis, offering detail on a work's achievements, impact or historical significance, not solely a summary of that work's plot. A plot summary may be appropriate as an aspect of a larger topic.

So far, most of our articles on works of fiction don't come up to this standard, but many of those can be improved. An article that is merely a plot summary is underwhelming at best, and original research at worst. Robert A.West (Talk) 08:55, 25 October 2006 (UTC)

Agreed. See also WP:WAF, which tries to delimit when where and how to handle plot summaries. Often the people laying the groundwork in this area do just write out plot summaries. What is needed is a follow-up team to prune the results, add in sources to secondary literature, and add in verifiable and encyclopedic information, like sales figures, publication histories, editions in different languages, that sort of thing (for books), and broadcasting history, production details, and so forth (for film and TV media). But to back all this up, there does need to be a ground layer of basic information about the fictional work, to give the reader the context. Carcharoth 09:52, 25 October 2006 (UTC)

I didn't mean to imply that certain Wikipedia articles should include only plot synopsis -- sorry if I was unclear. I'm just uncomfortable with the way you've worded it -- it seems to potentially imply that having critical commentary already in an article is a sourcing requirement for adding basic primary source material -- problematic, given that Wikipedia is a work in progress. Practically speaking, if an editor wants to add a plot synopsis to a stub and critical commentary isn't added until later, that shouldn't be an attribution problem for that editor. I'd hate to have content removed simply because it wasn't added in the right order. I'd be happier if "where appropriate to illustrate the attributable interpretations" were simply "where appropriate".

Regardless, my larger concern was that "quotations, extracts, etc." doesn't adequately cover basic, non-interpretative summary, which is extremely relevant in some cases. Perhaps that language can be expanded. -- Bailey ^(talk) 01:10, 27 October 2006 (UTC)

The spectum of informational reliablity (very poor to very reliable is evolving as the internet evolves. Whether we call it WP:WAF or call it WP:RS, I think we are going to need a guideline page that editors can go to in order to discuss whether a particular source is good enough to reference, or not. We would like to keep such discussion (and policy changes) off of Policy Pages, do we not? Presently WP:RS's discussion page fills this role. So some guideline discussion will need to fulfill the role that WP:RS's discussion page fills, now. Terryeo 00:09, 27 October 2006 (UTC)

Oh! I now see Bailey's concern. It hadn't occurred to me that the language might be seen as restrictive in that way. In practice, we probably have a hundred articles that contain unsourced commentary for every article that avoids a plot summary because of concern about OR. Nevertheless, I am sure the language can be improved, and the FAQ also more explicitly deals with primary, secondary and tertiary sources. Robert A.West (Talk) 06:12, 27 October 2006 (UTC)

Citations for common knowledge or easily verifiable facts

I strongly feel that this should not become a policy unless there are clear guidelines how to implement it, with well-chosen illustrations for a variety of issues that an editor may run into. One thing that particularly puzzles me is how to provide citations for things one would consider common (or easily verifiable) knowledge. Please do not just repeat the mantra: "If it is common knowledge, it should be easy to provide citations". I invite everyone who strongly feels that every last factoid should be cited, to provide reasonable verifiable citations from reliable sources for the statements below occurring in various existing articles. By "reasonable", I mean they should be reasonable for use in inline citations for these facts. Alternatively, you may of course argue that some of these unsourced statements should be removed because they detract from the encyclopedic value of Wikipedia. Hopefully this exercise will result in some useful material or ideas for presenting examples in a section of these future guidelines. --Lambiam ^Talk 21:29, 21 October 2006 (UTC)

It doesn't say that "every last factoid should be cited." It says that material should be attributed if it's challenged or likely to be challenged. SlimVirgin ^(talk) 08:46, 22 October 2006 (UTC)

Addition: "For example, 3 + 2 = 5, since 5 is 2 more than 3."
- [2]
  This covers the first half of the statement, but not the second half. --Lambiam ^Talk 05:33, 22 October 2006 (UTC)
- Logical consequence of other details already explained and sourced on the page and a knowledge of the meaning of the specific numbers involved. No reference needed. JulesH 12:35, 22 October 2006 (UTC)
Apple pie: "The fruit for the pie can be fresh, canned, or reconstituted from dried apples."
- This is not an encyclopedic statement. It belongs in a cookery book. Carcharoth 23:53, 21 October 2006 (UTC)
  So I take it you prefer this (and the following sentences that depend on it) to be deleted. --Lambiam ^Talk 05:33, 22 October 2006 (UTC)
- I'd suggest that a source should be found that describes apple pies made with non-fresh fruit, and simplifying to a statement that the fruit not be fresh. Should be a simple matter. JulesH 12:35, 22 October 2006 (UTC)
  "Should be a simple matter." Perhaps, but don't forget we also need to establish reliability of the source. --Lambiam ^Talk 20:15, 22 October 2006 (UTC)
Egg (biology): "Most bird eggs have an oval shape, with one end rounded and the other more pointy."
- [3]
  The source would support "Some" but not "Most". --Lambiam ^Talk 05:33, 22 October 2006 (UTC)
  I'd definitely want to see a source for "most". I have personal experience of a small minority of bird species' eggs, certainly not enough to be sure of a generalisation from the "some" that I (and most people) would accept without a source to "most". JulesH 12:35, 22 October 2006 (UTC)
  
  I think "most people" should be "some people". Most people would happily assume that there is no observational bias such as that birds laying oddly shaped eggs do so in a more secretive way, and would not believe a claim that several thousand virtually unknown species of birds lay eggs in the form of cubes, spoons and bananas. --Lambiam ^Talk 20:15, 22 October 2006 (UTC)
Google: "Google's core business model revolves around its internet search engine, which also includes a tool to search for images, news stories, and peer-reviewed, academic publications."
- Find sources describing Google's "core business model". Carcharoth 23:53, 21 October 2006 (UTC)
  Well, yes, find such sources. --Lambiam ^Talk 05:33, 22 October 2006 (UTC)
- "Google is a public and profitable company focused on search services."[4] "Users can search for images, multiple file types, pages in Czech or Turkish, posts to Usenet, phone numbers, airline flight info or categories in the Open Directory." [5] Also refer to list of google services at [6]. All of these are non-contentious claims, so self-published corporate descriptions are fine. JulesH 12:35, 22 October 2006 (UTC)
Language: "Human languages are usually referred to as natural languages, and the science of studying them is linguistics."
- The "usually referred to" - who is doing the referring? And the second clause of the sentence can be linked to a definition of linguistics. Carcharoth 23:53, 21 October 2006 (UTC)
  The intended meaning is: ""Human languages are usually referred to by human beings when referring to human languages as opposed to languages-in-general as natural languages". --Lambiam ^Talk 05:33, 22 October 2006 (UTC)
- A non-contentious claim about terminology. Could be linked to a dictionary definition if necessary, but is "unlikely to be challenged" so doesn't need an explicit attribution. JulesH 12:35, 22 October 2006 (UTC)
Mouse (computing): "When mice have more than one button, software may assign different functions to each button."
- Research and rephrase. "The first computer mouse to have more than one button was <give details>. The functions of these buttons could be assigned by software, the first such example being <give details>. This led to an established standard <give details> that has been a universal feature of all subsequent computer mouse devices." Carcharoth 23:53, 21 October 2006 (UTC)
  This only requires more references, including one that would do for the original text! --Lambiam ^Talk 05:33, 22 October 2006 (UTC)
- A logical consequence of having more than one button. Doesn't need a source. JulesH 12:35, 22 October 2006 (UTC)
Quebec, County Durham: "Road access: Minor roads off B6301 and B6302"
- Not encyclopedic. Wikipedia is not a travel guide. Carcharoth 23:53, 21 October 2006 (UTC)
  Not convinced. Sometimes what makes a town important is the quality its transport links to other places. A citation to an appropriate map would suffice for this. this one will do. JulesH 12:35, 22 October 2006 (UTC)
Spain: "To the west and to the south of Galicia, Spain borders Portugal."
- Rewrite to give history of the establishment of the border. Reference treaties and suchlike. Carcharoth 23:53, 21 October 2006 (UTC)
- Again, cite an appropriate map. JulesH 12:35, 22 October 2006 (UTC)
Sun: "The Sun is the star of our solar system.
- [7]
- Other than the horrendous phrasing, having established that the sun is a star in the solar system, which is a unary system, this is a logical consequence of things that are already referenced. JulesH 12:35, 22 October 2006 (UTC)
Universe: "Both popular and professional research articles in cosmology often use the term "universe" when they really mean "observable universe"."
- Quote examples. Carcharoth 23:53, 21 October 2006 (UTC)
  Who is to be the judge that occurrences of "universe" in the examples really mean "observable universe"? --Lambiam ^Talk 05:33, 22 October 2006 (UTC)
- I'd definitely want to see a source for this. It's an interesting and non-obvious claim. JulesH 12:35, 22 October 2006 (UTC)
Week: "A week is a unit of time longer than a day and shorter than a month."
- Go into more detail about the history of the week. At the moment, this is a pointless oversimplification of what a week is. Carcharoth 23:53, 21 October 2006 (UTC)
  This is just one sentence from a whole article. The history of the week, by the way, goes like this: "It all started on a Sunday. Then it became Monday. ... The week finally came to an end when Saturday expired." --Lambiam ^Talk 05:33, 22 October 2006 (UTC)
- Sentence should be removed as non-encyclopaedic. JulesH 12:35, 22 October 2006 (UTC)
  Really? What makes it non-encyclopedic? How else would you formulate this? Or do you mean this information has to be removed from the article all together? --Lambiam ^Talk 20:15, 22 October 2006 (UTC)
  It's misleading in it's generality. For over 99.9% of usages, a week is 7 days. The description should lead off with this information. If there are exceptions, they should be mentioned afterwards. I'd prefer a formulation more like "A week is a unit of time that is usually seven days long[reference to description of a standard week, e.g. in a dictionary] but which may in some disciplines refer to other units of similar length[reference to a discussion of pre-modern era 'weeks' that may be shorter or longer]".
Yellow pages: "In many countries, the Yellow Pages refers to a telephone directory for businesses organized by the category of product or service. As the name suggests, they are usually printed on yellow paper, as opposed to white pages with non-commerical listings, printed on white paper."
- "In many countries" is a vague bit of hand-waving. Be precise. Give several examples of countries and link to their official websites. Carcharoth 23:53, 21 October 2006 (UTC)
  Why should the official websites of the countries have anything to say about this statement? --Lambiam ^Talk 05:33, 22 October 2006 (UTC)
  I think Carcharoth meant the official website of the disitributor of the yellow pages in those countries (e.g. [8]). JulesH 12:35, 22 October 2006 (UTC)

I don't know that I believe all these things need i

nline citations per se, but I do think the information for ten of these can be easliy found. I found three of them in a single reference work. Honestly these are things that my not be easily found online but are certainly available with a trip to the reference section of a library.--Birgitte§β ʈ Talk 22:54, 21 October 2006 (UTC)

Requesting citations for very obvious statements strikes me as unconstructive, because it yields unnecessary work. I think such requests will not make one popular. Andries 23:19, 21 October 2006 (UTC)

What exactlty are you replying to? I don't think the proposal encourages such requests. I certainly don't support making such requests. Who do you think is suggesting citation needed tags should be put on obvious statements?--Birgitte§β ʈ Talk 23:26, 21 October 2006 (UTC)

Carcharoth, in the future would you avoid commenting in the middle of somebody's post? It makes it hard to follow the original, and like interruptions in conversation, I feel it robs the speaker's words of some of their power. Thanks, William Pietri 00:24, 22 October 2006 (UTC)

I don't mind the replies being together with the examples. I see however that Carcharoth has not always considered the issue in the context of the statement; for example, for item 8 the suggestion offered does not work. I'm also not sure that all references are to what I consider a reliable source, and I don't know if the reference [8] was offered in jest or meant to be serious. I should perhaps have also given this option: "This does not need a citation", but from reading the discussion on this page I somehow got the impression that the overwhelmingly prevailing opinion is: No exceptions, every fact must be sourced, including basic common knowledge. Inline citations appear to be a requirement from Good Article reviews, not necessarily from this proposal. I assume we all want to write good articles. --Lambiam ^Talk 05:03, 22 October 2006 (UTC)

I don't know where you got that idea from. The proposed text clearly states that it isn't necessary to provide a citation for every fact. It does, however, say that it must be possible to find such a citation on request, and that information that doesn't provide a citation may be deleted (with the implication that to restore it, one would need to provide a citation).

I find the attituted of GA reviewers regarding inline citations quite problematic in some cases. If something is covered in a general reference provided for an article, it shouldn't need a specific citation at the sentence in question, IMO. JulesH 12:35, 22 October 2006 (UTC)

I agree with JulesH. I do not believe this policy is asking such things to have inline citations, however it is silly to claim that reference cannot easily be found. I did mean my references in earnest, and I think they are valid even if require a slight change in wording (most --> some). I don't know what you are trying to provoke here as what you are against is not supported by this proposal.--Birgitte§β ʈ Talk 15:13, 22 October 2006 (UTC)

I agree also with JulesH. We have any articles with facts that are not cited directly. Any given datum of knowledge has a huge hierarchical tree of facts behind it. There are a wide variety of facts that people take for granted as common knowledge. At some level, there is a point where something is less commonly known. Facts that are disputed, or dubious should have citations. As an article progresses, it will gather more citations. It is clearly not practical to point every single fact to another wikipedia article, or to external citation in every single article. Atom 15:46, 22 October 2006 (UTC)

OK, if you all say so, I may have been seeing ghosts. These days I'm getting a bit jittery on Wikipedia. --Lambiam ^Talk 17:22, 22 October 2006 (UTC)

Lambiam, you're not seeing ghosts -- the requirement for in-line citations on the Good Article criteria page is poorly phrased. The "good article" standard was originally based the "featured article" standard, only with the criteria somewhat reduced; however, not even the featured article criteria demand an in-line citation for every individual fact. All they demand is "the provision of a "References" section in which sources are set out and, where appropriate, complemented by inline citations". That's how the GA requirement has been practically interpreted, too -- Google, for example, is a GA, and it's one of the examples on your list.

Also, let's not forget that there's a difference between "cited", "sourced" and "sourceable" (or "attributable", according to the new re-write). All information in Wikipedia needs to be attributable to a reliable source -- that's the bare minimum for inclusion and the point of this policy. All information should ultimately be attributed -- meaning that there's a source to cover every single fact in the references section at the bottom of the article. That's what the Good and Featured articles criteria demand, but not a policy, because removing sourceable information from an article based on the fact that it is not explictly sourced is usually a bad idea. Depending on how obviously sourceable the information is, it might even be a violation of WP:POINT.

Having every single fact be explictly connected to its specific source -- "cited" -- is an even higher standard, which we may want to adopt someday, particularly if we can come up with a technical implimentation that makes this unobtrusive to the reader. But that's not really relevant at the moment. -- (Lee)Bailey ^(talk) 18:05, 22 October 2006 (UTC)

This is a good summary of the position. But what I find worrying is that these discussions go round and round in circles. I agree the main policy document needs to be short and to the point, but the number of editors turning up and needing clarification indicates that a more detailed document (based on the talk page archives, maybe) needs to be available to sort through and index the many objections, misunderstandings, questions and so forth. This could be handled by a massive expansion of the FAQ, with a prominent notice on the talk page to add something to the FAQ if a new question keeps being asked. Carcharoth 21:28, 22 October 2006 (UTC)

I hope so. It doesn't help to find a requirement of "the mandatory use of some sort of in-line citation" on the talk page of your favourite article.[9] I find the inline citations in an increasing number of articles (e.g. Lancaster County, Pennsylvania or Leonhard Euler) obtrusive and interfering with the reading. I once tried to propose a non-intrusive way of providing attributive information – basically, putting it in a comment in the source text[10] – but this suggestion met with unanimous disapproval. --Lambiam ^Talk 19:46, 22 October 2006 (UTC)

Taking your points one by one: (a) the note from WikiProject Good Articles on the Mathematics article was, in my opinion, poorly phrased, in that it just asked for more citations without being specific. It appears to be a warning note, but the tone makes it appear more than that, and doesn't leave much room for discussion. It is best for critics to be specific and start with the most contentious stuff or obscure stuff, and ask for citations for those bits first. Only then should the "obvious" stuff be tackled, if at all. That tends to require a rewrite to make the implicit sources more obvious. (b) I agree that excessive citations can be intrusive for the reader, but the reader still needs to be able to verify the article somehow. If you can come up with a less intrusive system that is visible to the reader, please do. (c) The link you give to your proposal doesn't link to the responses (which I agree with). The full debate can be found here. Carcharoth 21:16, 22 October 2006 (UTC)

What does this page add?

This page appears to be much longer than Wikipedia:Verifiability and to add nothing of much use that is not already in that page. (Or am I missing something?) WP:NOR follows automatically from WP:V anyway, and can safely be ditched. The worst thing of all is that the policy, as opposed to discussion, section of WP:V is just 57 words long. Is brevity really a curse to be thrown away so easily? jguk 18:22, 22 October 2006 (UTC)

NOR doesn't follow automatically from V to the point where it can be ditched. The point of this is to merge the two. SlimVirgin ^(talk) 18:25, 22 October 2006 (UTC)

WP:V already states (at the end under "Other comments") that "Another effect of this policy is that as original research will not be supported by reputable sources, it cannot be included." This seems to cover the point, jguk 18:27, 22 October 2006 (UTC)

We have to explain what we mean by OR though, because it's not always obvious to editors. SlimVirgin ^(talk) 18:41, 22 October 2006 (UTC)

Why?

I'm not trying to be difficult, but why do we need to explain a concept we no longer need to rely on. Any "original research" (by WP:NOR definition or by any other reasonable definition) will not be supported by reputable published sources. So why do we even need a discussion on original research to begin with? jguk 11:19, 24 October 2006 (UTC)

Material can be apparently well sourced and still be OR because of the way it's been put together. NOR and V are very closely connected, I grant you, and could be described on the same policy page, but they're nevertheless distinct ideas. We can't just abandon one of them. SlimVirgin ^(talk) 17:32, 24 October 2006 (UTC)

If material has been put together in a novel way, you need to consider whether in doing so you formulate a new idea or concept of itself. If it does, that idea or concept needs to be verified. We could add those two sentences to WP:V if that would help.

Are there any other cases where you believe something might be compliant with WP:V but fall foul of WP:NOR? jguk 12:04, 27 October 2006 (UTC)

Professional

I think we should add professional researcher again, as V states, rather than expert or just professional. Just because someone who works in a field sets up their own website shouldn't mean they can be used as reliable sources. As currently worded, any teacher with her own website can be used in articles about teaching; any engineer in articles about engineering. The point of that section in V is to stress that the person must be a professional researcher (academic or otherwise) into that topic; that is, they are being paid by others to be a reliable source and therefore may be regarded by us as one, even if self-published. SlimVirgin ^(talk) 19:37, 22 October 2006 (UTC)

See, I still don't buy the being paid as the defining part of a good source. I think if other people have utilised the research of the person in question, or the person in question has presented at conventions, then that gives them weight, But that's my two cents. I'm not sure we should disregard engineer's opinions in articles about engineering. As engineers, they are paid for their engineering expertise, therefore they are professional experts. Don't know how you are going to close this circle, and even if it is necessary. Consensus should eventually work out what the good articles are and what the bad ones are. Many times magazines like New Scientist and newspapers will seek the opinions of people from within a field; at times teachers opinions on teaching do become valuable. I think it's got to be contextual. What is the source being used to convey, that should guide the reliability. Hiding Talk 19:53, 22 October 2006 (UTC)
I've worked with electrical engineers. Some of them were very good, I would treat anything they said as of the highest repute. But there were exceptions, too. In any sufficiently large group of people, there will be persons who are, at times, not worthy of trust. Terryeo 00:17, 27 October 2006 (UTC)

The issue is credibility. Being paid does not establish credibility. In fact if you are being paid by a someone or an organization with a conflict of interest regarding the subject matter being paid can diminish credibility. Being a research also does not necessarily establish credibility. Expert researchers are notorious for being narrowly focused, as keeping up in their area with the latest advances precludes time spent in a broader study that could lend perspective. Further professional researchers need to be continue to be funded, meaning they are always advancing their careers by overestimating the importance of their latest finding. WAS 4.250 21:11, 22 October 2006 (UTC)

Being paid is almost always the way credibility is established, WAS. If no one is prepared to pay for a person's expertise, why should Wikipedia trust it? SlimVirgin ^(talk) 10:07, 23 October 2006 (UTC)
- On the issue of being paid, it is certainly an indicator. However, there are people who are paid to present (anything) in the best possible light. Bush has his publicity staff who do so. Certainly drug companies too. side effects include propensity to suicide. Such professionals' public statement of who they serve should be considered, also (I think). Terryeo 00:17, 27 October 2006 (UTC)

Professional researchers reference what they do, so we can rely on them to that extent and check their facts, if necessary; experts may just be pundits who talk rather a lot about the same thing. Since you don't trust professional researchers, I'm not clear whether you wish to widen or narrow the definition of reliable sources. If you approve "experts", you widen the pool of sources (and the goal posts) and approve professional researchers anyway (because they are experts). qp10qp 22:52, 22 October 2006 (UTC)

- Credibility depends on other issues - not of being paid; not of being research or non research. "Expert" is the minimum qualification and not the sufficient qualification. WAS 4.250 02:27, 23 October 2006 (UTC)

Aren't we pushing things too hard and expecting too much? Let's remember that most articles aren't sourced at all. It's not like policy where source being expert or professional is not enough is going to make everyone find a widely known, professionally working, cited expert researcher for every claim. A more likely outcome is that we'll just lose references to a lot of sources, which are, in practice, fully correct on over 95% cases. Do we really expect professionals and experts to write such nonsense that they may never be referred to? And let's remember that such source with mistakes can be easily countered by a more reliable one, if it can be found for the subject. CP/M^comm |Wikipedia Neutrality Project| 03:23, 23 October 2006 (UTC)

My problem with "professional researcher" is that it implies a bias towards academic subjects. I still favour "professional expert", i.e. somebody who is employed for their unusually advanced knowledge about some subject or other, whether they actively perform research in that area or not. I agree, however, that merely being paid for something doesn't qualify them as a good source. Perhaps "notable professional expert"? This would imply only accepting sources who were well-enough known in their subject area that their opinion carries at least some weight. JulesH 09:35, 23 October 2006 (UTC)

I don't think researcher implies a bias toward academia, although in reality that's mostly what we'll be dealing with: academics who have set up their own websites. Making it clear the person should be paid to research the area means we don't have to use every professional engineer with a website as a source on engineering. Adding "notable" doesn't really get round that. How would we establish notability? What sources would we use? etc SlimVirgin ^(talk) 10:07, 23 October 2006 (UTC)

WP:V also says "so long as their work has been previously published by reliable third-party publications," which we should probably add here, as this isn't an attempt to change policy, just streamline it. SlimVirgin ^(talk) 10:21, 23 October 2006 (UTC)

How about: "if and only if their work has been previously published by reliable thrid-party publications"? This would allow what we want, exclude what we don't want and would be terribly worded. heh. Terryeo 19:39, 27 October 2006 (UTC)

The latter I agree with. The former... I still disagree. There are many disciplines that you wouldn't describe anybody as researching. To come back to previous examples, the Jefferson Nickel book that was mentioned comes under this category; I don't suspect the author researched any of the information, at least not in the sense of the word that's likely to be read. He probably collated information from a variety of sources to produce his book. While technically that is research, it's research of a kind that any student performs, so is clearly not what we mean here. Lawyers also come under this category. If a lawyer publishes legal advice about some topic, I'd hope we could use that as a reference. But a lawyer isn't a researcher. A professional expert, yes. JulesH 06:43, 24 October 2006 (UTC)

Anyone who is paid to research a topic is a professional researcher, regardless of whether an academic or not. We don't use any old lawyer as a source on legal issues. We do use people with particular legal expertise who may have published on a specific topic they have specialized in (researched). The key difference is that our self-published source must have gone meta. We don't want self-published school teachers to be used as sources on school teaching. But we can make an exception for a self-published school teacher who has been paid to research issues on school teaching and whose research in that area has previously been published by a reliable source. In other words, our self-published sources must be the very best in their field. SlimVirgin ^(talk) 07:24, 24 October 2006 (UTC)

If we use self-published sources about which there is consensus of other sources to be best in their field, we will be safe. ≈ jossi ≈ t • @ 15:25, 24 October 2006 (UTC)

Researcher means whatever people reading the policy think it means. If a significant number of readers will be chilled from using appropriate sources through an overly narrow reading of the word researcher the word should be changed. There are some areas where some of the experts would seldom be described as researchers, for example, sports. --Gerry Ashton 17:09, 24 October 2006 (UTC)

At the beginning of this thread, SlimVirgin expresses concern about using the website of any engineer about any engineering topic, and proposes to plug the whole by adding the word researcher. That won't help a bit; the typical daily activities of an engineer are research.

I agree there is some risk that an engineer in the casual environment of a personal website might comment on some engineering field outside her area of expertise, but adding the word researcher will not solve this potential problem, nor will retaining the word professional, since many engineers are licensed by their state or province as Professional Engineers. I think the greater difficulty would be verifying that a website is actually under the control of a real engineer. --Gerry Ashton 17:25, 24 October 2006 (UTC)

Rewrite of section

I want to focus on these two comments from above which I think everyone agrees on and try to rewirte the this section around them. I am also taking into account other disscusions that have been going on here and at the popculture workshop.

WP:V also says "so long as their work has been previously published by reliable third-party publications," which we should probably add here, as this isn't an attempt to change policy, just streamline it. SlimVirgin ^(talk) 10:21, 23 October 2006 (UTC)
If we use self-published sources about which there is consensus of other sources to be best in their field, we will be safe. ≈ jossi ≈ t • @ 15:25, 24 October 2006 (UTC)

2. Superior self-published sources

In-depth sources sometimes exist which are self-published by [an author] whose other work, in the same field, has been previously published by reliable third-party publications. Such sources may be acceptable only when there is a consensus of other sources that it represents the best work in the field. If there is reasonable doubt about the reliability of the source or the relevance of the material to the subject matter, err on the side of caution and don't use the self-published material.

(Adjusted per intial replies)

In-depth sources sometimes exist which are self-published by [an author] whose other work, in the same field, has been previously published by reliable third-party publications. Such sources may be acceptable only when other sources regard them as reliable. Self-published sources should only be used if they offer better coverage or quality than other publications, and there is no reasonable doubt about their reliability. If there is reasonable doubt about the reliability of the source or the relevance of the material to the subject matter, err on the side of caution and don't use the self-published material.

I believe this rewrite generally represents what we have consensus for, but I understand that people wish to more specifically restrict the type of authors we accept. This is where the current stumbling block really is. The attempt to resrtict certain authors by deciding on a word that will define acceptable authors is not getting very far. To try something different can we list types of authors that would fit the above rewrite and that we still wish to exclude from the policy? After we can better understand the goal of who to exclude maybe we can come a with a better word to distinguish between these types of people and otherwise acceptable ones. Please think of this list as being what you are worried about becoming acceptable with the proposed rewrite. As such no one should be combative about what others choose to list, remember consensus is about trying to mitigate the concerns of objectors and if this comes to arguing about whether someone's concerns are valid or not we doing well. If anyone is concerned about Foo; Foo is valid and we need to try and find how to exclude them. Of course if anyone believes my rewrite is completely off-base please say so and ignore all the ideas about listing concerns etc.--Birgitte§β ʈ Talk 18:30, 24 October 2006 (UTC)

BirgitteSB's rewrite has a problem in that it requires both that the author have published other work in reputable publications, and that the self-published work be considered the best in the field by reliable sources. I believe either condition alone would make the self-published work acceptable. Also, I think requiring it to be the best in the field is excessive; it would suffice for it to be well-regarded by reliable sources. --Gerry Ashton 18:44, 24 October 2006 (UTC)

If there exist equal sources which are not self-published, we should use those. Self-published sources should only be used when they are better either in quality or detail than other sources. That said I am open rewording best into something along the line of better or superior. As to your other point, from reading the opinions here, I think a large number of people would object to simply allowing the self-published work of an author who has published other work in reputable publications. If I am wrong about this and this is acceptable to others, I will not object to it. I was trying to write something I believed the majority of people could agree to rather then my own preference--Birgitte§β ʈ Talk 19:13, 24 October 2006 (UTC)

I agree with Gerry. We aren't that rich of sources to introduce limits beyond necessary. I would remove the word "best" and stay with "...when other sources regard it as reliable", with notice later that "Self-published sources should only be used if they offer better coverage or quality than other publications, and there is no reasonable doubt about their reliability". CP/M^comm |Wikipedia Neutrality Project| 19:20, 24 October 2006 (UTC)

Wikipedia is the encyclopedia anyone can edit. Editors should be allowed to use any source that meets certain minimum standards and which the editor is aware of and has access to. Requiring a person to conduct extensive research to determine the "best" source is a requirement that might be applied to a PhD candidate, but not a Wikipedia editor. Of course, if a subsequent editor is aware of a reliably published source that is equal to or better than a self-published source, it would be proper for the subsequent editor to revise the article accordingly. --Gerry Ashton 19:40, 24 October 2006 (UTC)

Thanks for the input I updated the wording to eliminate the use of "best".--Birgitte§β ʈ Talk 19:43, 24 October 2006 (UTC)

I continue to object to requiring that a self-published source be both from an author who has other publications in third-party reliable publications, and when other sources regard them as reliable. If other reliable sources in a field consider a self-published source reliable, what business does Wikipedia have saying it isn't, even if the author has not published elsewhere.

In the case of an author who has published in third-party reliable publications, the self-publication may not be important enough to warrant mention in reliable sources, but may be useful nevertheless. A case that comes to mind is an author who publishes a work through a well-known publisher, but puts errata on a personal website, or personally controlled university web site. Try searching for errata on Google and see what you get. --Gerry Ashton 20:03, 24 October 2006 (UTC)

I definitely agree. I don't think there's actually any one set of criteria that we can use to judge a reliable self-published source, and constraining it like this eliminates a number of useful sources.

Clearly, what we have at the moment is one category of reliable self-published source. Another is non-contentious biographical or corporate information, which is covered by the other exception. I'd say others include:

An adequately qualified expert, whether previously published by a reliable source or not. For instance, a well-reputed lawyer publishing legal information.
Most people, on the subject of their own opinions (whether those opinions are notable enough to include or not should not be a matter for policy to cover, but a guideline instead, IMO). This could come under the heading of primary sources, for which I think a lower bar should be set. A primary source doesn't need independent editing to be reliable.
An "inventor", on the subject of the details of how they came to invent a device, or how that device functions, unless claims are presented that contradict their own claims.

Comments on these? JulesH 11:52, 25 October 2006 (UTC)

I thought this attempt was broader than the current wording, but it is not something I am going to push.--Birgitte§β ʈ Talk 13:31, 25 October 2006 (UTC)

I think that this goes a bit too far. Consider an inventor of a perpetual motion machine. If for some bizarre reason it's notable enough to get in an article, the inventor should be a legitimate source for "Joe Blow claims that the machine functions by..." but not as a source for "the machine functions by..."

What you might want to say is that someone is generally an expert on things about himself that he can observe without further interpretation. This covers the "their own opinions" case and the "inventor claims that..." case, since you can, without further interpretation, know what your own claims are. It wouldn't cover cases that need interpretation, like saying "I am the greatest person ever". It also would cover observing your own shoe size, which WP:Autobiography currently says isn't allowed. Ken Arromdee 13:47, 25 October 2006 (UTC)

What about: On his own web site, joeblow.com, Joe Blow claims that his shoe size is "the largest ever". ? Look, no interpretation involved. I don't think we need to go as far as this for obviously non-contentious personal data, although people are known to have "juvenated" themselves by adding a few years to their year of birth, or to have mystified their biographical data. --Lambiam ^Talk 15:49, 25 October 2006 (UTC)

On his own web site, joeblow.com, Joe Blow claims that his shoe size is "the largest ever". would pass the test of coming from an expert (since Joe Blow is an expert on what his own claims are) but would fail the test of "Wikipedia is not an indiscriminate collection of information". In other words, it's true that Joe Blow makes this claim about his shoe size--but the fact that Joe Blow made such a claim is of no interest to us. The only reason we'd include such a claim is to imply that the claim is true (which *is* original research, and which I'm not suggesting we allow). This is different from the inventor--"Joe Blow claims to have a perpetual motion machine" is encyclopediac all by itself, regardless of whether he really does. Ken Arromdee 08:06, 26 October 2006 (UTC)

A valid point Ken, I agree. It would be useful to have some direction to editors toward understanding how reputable such a website is. First of course, some attribution must be present. Terryeo 19:45, 27 October 2006 (UTC)

Explicit and implicit attribution

One of the concerns in recent discussions is use of sources essential to coverage of a subject, like own inventor's site, yet possibly biased or unreliable. It's quite certain that the underlying contradiction involves the method of attribution: reliability of claim "party P claims X" does not depend on P, but reliability of "X [footnote:P]" does.

I suggest that instead of attempting to express the use of less reliable sources through more complex wording we rather outline the distinction directly. This will need to involve both policy and FAQ, and uses concepts from existing guidelines and policies. Possible wording, which would need some style correction, expansion for FAQ and compaction for policy, would be like this:

Explicit and implicit attribution

The material can be attributed in more or less explicit manner. The required level of attribution depends on plausibility of a claim and reliability of its source.

Denoting the source in references section without use of footnotes implies that some part of statements in the article can be attributed to the specified source. This is appropriate for accepted facts which are certain for any person with specialist knowledge, when source reliability is not questioned, for instance when referring to a textbook.
Inline citations and footnotes, referencing a specific claim, are the preferred main method of attribution for most cases. This method implies that the statement is most likely correct and is confirmed by the specified source. It is appropriate for referencing reliable sources where there is no contradiction.
Statement can be attributed to a source explicitly, by modifying the text itself to report source making a certain claim. This avoids supporting a statement, clearly denoting the source responsible for it. Explicit attribution is required when using sources of questionable reliability, attributing an opinion or evaluation, or when there is contradiction between sources.

While longer even if compacted, such form would provide more clear guidance to the reader. Other opinions? CP/M^comm |Wikipedia Neutrality Project| 06:25, 26 October 2006 (UTC)

I like this. It makes it clear that you have to be careful while using questionable sources, gives a clear indication of what steps must be taken while you're using them, but doesn't restrict articles from using such sources when necessary. JulesH 08:11, 26 October 2006 (UTC)

Since there are no objections, I'll insert this into the proposal so we may work further on improving it there. To people which are watching it and disagree completely, please consider discussing it here rather than in the summaries; otherwise, just edit as needed. Since it isn't yet a policy, and what I insert only clarifies things already outlined, I guess it's appropriate to be bold.

The new section, merged with the older one, looks this way:

When a large number of non-controversial facts in the article may be attributed to a single source, it may be simply listed in the references section. This is appropriate for accepted facts, which are certain for any person educated in the area, when source reliability is not questioned, for instance when referring to a textbook.
Material challenged or likely to be challenged, and quotations, should be accompanied by an inline citation, written as a footnote or Harvard reference. This method implies that the statement is most likely correct and is confirmed by the specified source. It is appropriate for referencing reliable sources where there is no contradiction.
It is often appropriate to include prose attributions for sources. This is the attribution, within an article's text, of a claim to a specific source — "According to The New York Times, ..." — thus conveying that Wikipedia does not necessarily support the view. Editors should use common sense when attributing positions within the text, taking care not to cast unnecessary doubt on a position; for example, "According to Jane Doe, she was born in 1965", suggests, in the absence of further explanation, that Wikipedia has reason to believe she may not have been born then. Prose attribution is required when using sources of questionable reliability or insufficient scope, attributing an opinion or evaluation, or when there is contradiction between sources.

Please comment on this; I expect it might require changes in the wording, but I hope the intent to make clear guidance is supported. CP/M^comm |Wikipedia Neutrality Project| 22:55, 28 October 2006 (UTC)

Hi CP/M, I reverted the edit because I couldn't quite see what it was getting at. SlimVirgin ^(talk) 22:57, 28 October 2006 (UTC)

I can't understand it either; what does it mean? What is it adding (besides words and length)? Jayjg ^(talk) 02:13, 29 October 2006 (UTC)

It adds pretty much the same as the proposed section above. The current wording is somewhat unclear about when to use which attributions. Specifically, I attempted to add details describing that:

Just inline citations are appropriate only for reliable sources.
Others need to use prose attribution.
Books and papers which support non-contested claims are still desirable.

I guess there are other ways to include this, but please comment if either you disagree with the ideas or suggest other ways. CP/M^comm |Wikipedia Neutrality Project| 12:15, 29 October 2006 (UTC)

Commented out:

I am very sorry to interrupt and interject my views when so many have already talked and discussed the matter: In the name of simplicity and simple English, we should not decide anything which may be used (assume good faith please) to lower our credibility!

Moved from main page:

This looks interesting as many editors enveloped in thick veils of anonymities would have licenses (they anyway have now too) to cite their published work/s while poor people who chose to have a user name disclosing their true identity have lost the pleasure of making such citations! It sounds highly discriminatory. KINDLY REMOVE MY COMMENTS. THANK YOU. Bhadani

I have moved the comment from the main page, and uncommented the other comment. No real reason to have hidden comments here, and no reason to have discussion type comments on the main page. Cheers, Ansell 04:14, 29 October 2006 (UTC)

Is there going to continue to be WP:RS

The idea is to replace WP:V and WP:NOR with this policy, is that right? But when it comes to specific reliable sources, won't we need a guideline page to work with, and its discussion page to confront and handle new questions of reliability as they come up? I think a link to WP:RS (you know, "resolve your RS question here") or its equivalent should appear early in this proposed policy. The internet is evolving, we simply can not make black, white, and stable rules in policy for every source of information. We can, however, 1) make a stable statement of intent, 2) insist that intent be followed and 3) provide a guideline page which evolves to meet an evolving internet. Terryeo 18:18, 27 October 2006 (UTC)

Right now it looks like Wikipedia:Attribution will be presented as a choice to replace WP:V and WP:NOR with WP:RS dealt with as a seperate issue. See Wikipedia:Attribution/FAQ as the WP:RS replacement waiting in the wings. WAS 4.250 19:13, 27 October 2006 (UTC)

But the two are not exclusive: a FAQ could exist (with the status of "essay") in combination with a guideline that details how to evaluate the reliability of sources. Robert A.West (Talk) 20:51, 27 October 2006 (UTC)

All the more reason to deal with these as two separate issues to present to the Wikipedia community. WAS 4.250 21:53, 27 October 2006 (UTC)

We fail to disagree. Robert A.West (Talk) 22:24, 27 October 2006 (UTC)

I think both have to be in order to be presented. Otherwise we'll have RS referring to slightly different wording from WP:ATT, and it would get messy. SlimVirgin ^(talk) 22:58, 28 October 2006 (UTC)

Then there needs to be a top-down rewrite of WP:RS, or else we fail to advance past the problems this proposal was meant to solve. Phil Sandifer 22:59, 28 October 2006 (UTC)

I meant that WP:ATT and WP:ATTFAQ both need to be in order to replace V, NOR, and RS. I was still thinking in terms of ditching RS. Then we also need to decide on a final name for the proposed policy. Are people fine with Wikipedia:Attribution? SlimVirgin ^(talk) 23:02, 28 October 2006 (UTC)

Ah, OK. I still think it's probably a good idea for ATT to be policy, while ATTFAQ be essay, not guideline. At essay, it provides helpful insight, but avoids instruction creep. Phil Sandifer 23:16, 28 October 2006 (UTC)

I agree. SlimVirgin ^(talk) 23:20, 28 October 2006 (UTC)

Yes, that was the plan all along, ATT as a replacement for V, RS and NOR, and the FAQ as a complimentary element as guideline or essay. ≈ jossi ≈ t • @ 23:34, 28 October 2006 (UTC)

That works. Jayjg ^(talk) 02:12, 29 October 2006 (UTC)