Jump to content

Wikipedia talk:Attribution/Archive 2

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1Archive 2Archive 3Archive 4Archive 5

Explanation

I've put up a proposal that would involve getting rid of WP:V and WP:NOR as policies, and combining them into one new policy, Wikipedia:Attribution. WP:A is taken, but it would be good if we could claim it; I've also created WP:SOURCE as a shortcut.

I've used the word "attribution" for two reasons. First, "verifiability" confuses people because usually when we verify something, we confirm that it's true, which is not how the word is being used in the policy. Secondly, the word "attribution" allows us to summarize succinctly the difference between the V policy and NOR (as they now stand): we know something is not original research if it is attributable to a reliable source — that is, if it is capable of being attributed — and all edits must be capable of being attributed. But not all edits actually have to be attributed. "The sun will rise tomorrow" does not need to be attributed, but we all know that it could be if necessary.

In conjunction with this proposal, I propose:

  1. that WP:V be deleted;
  2. that WP:NOR become a description of how the idea of NOR developed, with some examples of OR (not a policy or guideline)
  3. that WP:RS be renamed Wikipedia:How to find reliable sources and become just an advisory page (not a policy or guideline), with everything that duplicates material in Wikipedia:Attribution removed; and
  4. that WP:CITE be renamed Wikipedia:How to cite sources, and that everything in it that discusses why we need sources be removed. SlimVirgin (talk) 04:39, 11 October 2006 (UTC)
Rereading the above, it's not crystal clear, so to clarify:
I know it sounds a bit radical, but so long as we don't throw the baby out with the bathwater, we'll end up with more streamlined, rational, easy-to-follow content policies. SlimVirgin (talk) 11:41, 11 October 2006 (UTC)
Small technical problem. There are tens of thousands of talk page posts and edit summaries citing the existing policies by their redirect. These will cease to make sense if major changes are made without care. I'd suggest that WP:NOR WP:V WP:RS and WP:CITE would need to point here, as the content would be here. If new 'how to' essays are needed, let's put them elsewhere to avoid confusion. The top of this poicy would need a note saying it superceeded the previous policies, and pehaps linking to an archives of them.--Doc 11:57, 11 October 2006 (UTC)
Good points. SlimVirgin (talk) 12:04, 11 October 2006 (UTC)
You can't delete WP:V, it would have to point here. Just a procedural comment and not an endorsement of the proposal. --kingboyk 12:01, 11 October 2006 (UTC)
Agree. SlimVirgin (talk) 12:04, 11 October 2006 (UTC)

NOR needs to remain a separate policy. Let's not get too ambitious, please, otherwise this proposal will not accepted by the wider community. Let's merge RS into V, as needed and propose this as the new WP:V policy. I still think that Verifiability is a better name than "Attribution", as an attribution needs to be verifiable to a reliable source. ≈ jossi ≈ t • @ 14:51, 11 October 2006 (UTC)

  • Could we set up a bot to handle updating the links? Maybe run it in the quiet periods? Whenever they are? I mean, depending on the time scale we have Christmas coming up, I assume the server load is reduced on that day so it'd maybe be able to handle that? Not that I'm a dev or anything, but maybe that's an idea? Hiding Talk 15:14, 11 October 2006 (UTC)
    • The easier solution would be to copy/paste this page over WP:V (after we're done discussing, of course) and call it "verifiability" since that's the term that most everybody in Wikipedia refers to these days. Otherwise, we'll just get people who see some kind of artificial difference between attribution and verifiability. I think that in general it is a good idea to merge related policies such as these to discourge divergence and to make Wikispace less convoluted. >Radiant< 15:30, 11 October 2006 (UTC)

Comments

Aren't WP:NOR and WP:V foundation issues? Perhaps they meaningly can be combined, but I was under the impression that these (along with WP:NPOV) weren't negotiable. Perhaps that means that their effects must remain in place, but the names and arrangement of the policy documents which contain them can change. I'm all for reforming WP:RS, which I think is a bit of a mess... but we should tread carefully when replacing WP:V and WP:NOR. Few editors have issues with those.EngineerScotty 05:16, 11 October 2006 (UTC)

BTW, I'm all for removing duplicate material from WP:CITE--it should be a technical/style guide only; not a content policy. One other thing I think sorely should be added--either to WP:CITE or elsewhere--is an article documenting different styles of prose attribution: When to simply state something along with citation, when to attribute it to a source in prose ("So-and-so said..."), etc. This has a great deal of impact on the presentation of a Wikipedia article; many edit wars spring up when someone tries to use the more authorative voice-of-Wikipedia (no prose attribution) instead of an explicit attribution in reference to a controversial or disputed claim.EngineerScotty 05:19, 11 October 2006 (UTC)

I agree with the stated goals. WAS 4.250 05:26, 11 October 2006 (UTC)

This gets it the wrong way round, I think. The aim of Wikipedia is to be a compendium of the world's knowledge, that is why we do not conduct research ourselves but weigh and compile the research of others. The necessity for attribution stems from "No original research". Dr Zak 05:31, 11 October 2006 (UTC)

The need for any content rule is to choose between content choices. The rules/policies only exist to help us make those choices. In particular when two groups both know they are right (cretionism vs evolution; scientology vs sane people; democrats vs republicans; etc) WAS 4.250 05:40, 11 October 2006 (UTC)

Other than the foundation issue above (and the existence of nearly 6 years of precedent concering WP:V and WP:RS)--how this is organized, and which policies are axioms and which are theorems (to borrow a bad analogy) isn't probably important.

Slim--I think the content of Wikipedia:Acceptable sources (which like this, was composed just today) might be a useful addition. Interestingly enough, there already are questions on that article's talk page, as if it were policy... :) Thoughts?

--EngineerScotty 05:35, 11 October 2006 (UTC)

Engineer, the spirit of the policies is a foundation issue, but the spirit is adhered to in this. The wording is allowed to be changed so long as the baby isn't thrown out with the bathwater. SlimVirgin (talk) 05:36, 11 October 2006 (UTC)
I agree with SV on this. Further, we need to prune back the verbage, not add to it. This proposal does just that. WAS 4.250 05:43, 11 October 2006 (UTC)
There is a page m:Foundation issues, which may or may not be correct and up-to-date. That page does not talk about no original research or verifiability. -- Jitse Niesen (talk) 07:31, 11 October 2006 (UTC)

I am concerned that the "reliable" aspect of sourcing will be watered down. Especially since "truth" is not the requirement (and never was). Conspiracy theory cruft is shot down frequently with the "reliable" policy. If it's not a fundamental, stand-alone pillar and is only a part of the "attribution" pillar, it may be harder to enforce. Maybe "reliable attribution" would be better but I am not sure how this policy is better than the ones we already have in place.--Tbeatty 05:48, 11 October 2006 (UTC)

It's better because it's one policy instead of two, for a start (one instead of three if you count getting rid of RS) and there's no unnecessary fluff that I can see. SlimVirgin (talk) 05:55, 11 October 2006 (UTC)
As long as it's written in a way that citing it to removed unsourced and poorly sourced information doesn't make it vague or muddled, I think it's fine. Deleting a comment that is sourced by saying it violates "WP:Attribution" will create edit wars simply because the defintion of "Attributed" means sourced. It will be confusing to newcomers if the "reliable source" aspect of "attribution" is buried here. Imagine User:Smith citing a blog (correctly and completely) but being told that his comment was deleted because it violated "Attribution." It's much easier to explain a blog isn't reliable that therefore violates "Reliable Sources" rule. It will be harder to explain how an attributed comment violates the "Attribution" rule because of the "Reliable Source" sub-clause.--Tbeatty 06:20, 11 October 2006 (UTC)
Note that you could say, for instance "Violates Wikipedia:Attribution#Questionable_sources". It's unfortunate that you can't set up a redirect to a section, because that would be helpful. JulesH 09:53, 11 October 2006 (UTC)

I like this, it reduces the alphabet soup which is always good. Perhaps it would need foundation endorsement, but if we iron out any problems, I'm sure that wouldn't be hard to get.--Doc 10:58, 11 October 2006 (UTC)

I'm quite keen on this, with the usual proviso that it needs to be discussed carefully. In particular the clarification on current practice related to popular culture articles is welcome. --Tony Sidaway 11:27, 11 October 2006 (UTC)
Doc and Tony, thanks for the feedback, and I'm glad you see the potential in it. I think it might iron out quite a few of the problems some good editors have been having with V, NOR, and RS. SlimVirgin (talk) 12:06, 11 October 2006 (UTC)

Just wanted to say that I think this is an excellent idea. It is silly to have three or four policy pages where one will suffice. These issues are closely related, and having four different policy pages to deal with different aspects of it is just confusing. The current draft looks mostly good to me - the things I'd take issue with are, I think, things I already take issue with on the existing pages, and are relatively minor. Further comments probably to follow as I read more of the talk page. john k 16:45, 11 October 2006 (UTC)

This is a very beautiful piece of work. It takes care of so many different problems on all the various pages quite well. Kudos!--Birgitte§β ʈ Talk 17:57, 11 October 2006 (UTC)


"Trivial"

Given agreement over previous ("could be attributed") I don't know why Jossi would revert latest. Let's not even suggest editors can deduce what is "trivial" synthesis. Really, let's not. Marskell 00:24, 13 October 2006 (UTC)

The reason for the revert (sorry, for not explaining before) is that it substantially changed the wording, i.e. "It is legitimate to make extrapolations that are verifiable " vs. "It is legitimate to make trivial, obvious, uncontroversial extrapolations". My highlights. It needs a qualifier, and you emoved the qualifier. ≈ jossi ≈ t • @ 00:27, 13 October 2006 (UTC)
Your last edit is better, Marskell. Although I would want the distinction "uncontroversial" to be present as well. ≈ jossi ≈ t • @ 00:31, 13 October 2006 (UTC)
No, no, no to this qualifier. My concern Jossi, is not that it is right, but how it will be used. I don't want a pseudoscientist telling me in a year "hey, this is a trivial deduction, what's yr problem?". Do you know what I mean? Let's give specific examples (2 + 2 = 4 is allowed) and nothing else in the affirmative. Marskell 00:34, 13 October 2006 (UTC)
I am happy to remove trivial. But I think that adding uncontroversial would be good, as it clearly implies that controversial deductions are not allowed. ≈ jossi ≈ t • @ 00:57, 13 October 2006 (UTC)
I guess I'm looking at it from different coloured glasses. It disallows controversial deductions (shouldn't that already be obvious?) but allows a deduction of what is and is not controversial. As with above re things that are so obvious they don't need a source, editors make commonsensical choices of that sort regularly. But let's not actually suggest it. It'll just be a bullet in the hand of bullshit artists. Marskell 01:03, 13 October 2006 (UTC)
You are right.... ≈ jossi ≈ t • @ 01:04, 13 October 2006 (UTC)

Added meat to (previously empty) "Citing sources" section

...including some which might be novel.

--EngineerScotty 01:04, 13 October 2006 (UTC)

It is pertty good, Scotty. ≈ jossi ≈ t • @ 02:00, 13 October 2006 (UTC)
Somewhat strange wording here, though: "At a minimum, all references used in an article should be cited with an inline citation, such as this[3] appearing after any substantial, surprising or controversial claim."
Is that meant to be: "At a minimum, any substantial, surprising or controversial claim in an article should be cited with an inline citation, such as this.[3]"? There's nothing wrong, as far as I know, with having references that aren't cited directly; but the current wording seems to suggest that it's every reference, rather than every claim, that needs citing. Kirill Lokshin 03:01, 13 October 2006 (UTC)
The latter is correct. Only these "substantial, surprising or controversial" claims require an inline citation for these claims. ≈ jossi ≈ t • @ 03:08, 13 October 2006 (UTC)
Re-reading the entire policy, this section seems especially long. Is there any chance of getting it down to one paragraph? --Birgitte§β ʈ Talk 03:24, 13 October 2006 (UTC)

Sure. I'd be happy to move the bulk of it to WP:CITE. Speaking of which, that page is seriously out of date, not mentioning cite.php at all.

(This is why the section focuses on prose attribution--detailing the other types is already in WP:CITE).

--EngineerScotty 04:32, 13 October 2006 (UTC)

This looks really good now. Thanks!--Birgitte§β ʈ Talk 16:06, 13 October 2006 (UTC)
I agree. This is not a replacement of WP:V and WP:NOR but it should work in conjunction. We have no original research because we need to be verified, and in order to be verified we need to give attribution.--HereToHelp 01:21, 15 October 2006 (UTC)

Better phrase for unoriginal research

I continue to object to the jargon meaning that Wikipedia assigns to original research, but since my objections are falling on deaf ears, what is the opposite of original research? Is there a better phrase than unoriginal research?

Before anyone suggests it, I won't accept source-based research because in publications where the identities and expertise of the authors is known, authors may drawn non-trivial conclusions from published sources, while this is not allowed for Wikipedia editors. In my mind, source-based research means any research that relied only on published sources and did not rely on new surveys, interviews, experiments, field observations, or the like. --Gerry Ashton 03:09, 13 October 2006 (UTC)

summarisation of existing research perhaps? JulesH 10:39, 13 October 2006 (UTC)
Original research works pretty well. ≈ jossi ≈ t • @ 14:34, 13 October 2006 (UTC)
I agree with Gerry that "original research" is problematic—we are stretching the English language, really, by applying that to all forms of original thought—but the term is probably too much of a Wikipedia shibboleth by now for it to be thrown out with the bathwater at this stage. Perhaps we have to keep this baby, for the time being, even if it's the wrong baby. qp10qp 14:53, 13 October 2006 (UTC)
Picking up from Qp10qp, if we are going to use the jargon sense of "original research", it would be handy to have a phrase for all the research that does not fall into the Wikipedia original research pit, but I can't think of anything good. Summarization of existing research is better than unoriginal research, but does not quite cover reorganizing. --Gerry Ashton 18:06, 13 October 2006 (UTC)
"Reporting"? As in "Reporting ongoing discussions or debates on the subject"? -- llywrch 21:50, 13 October 2006 (UTC)

How to format citations

Not sure why this section is called "How to format citations", but anyway, a few things:

  • I'm not sure all the detail on prose attributions is necessary—I say define it, give an example, and move on. Most of it is more related to NPOV anyway (evident from the fact that we link to it twice in a space of less than 100 words).
  • More important than prose attributions, I feel, is citations and what exactly should be cited. Perhaps the three words are sufficient (substantial, surprising, controversial), but I'd much rather see more detail here than in prose attributions. Of course, there's probably more dispute over what should be cited, which might explain the silence.
  • I've made some changes to prose and emphasis (the bold didn't make sense to me); please review.

Oh, by the way, this whole combining and shortening policies thing is brilliant. Let's not let inertia and status quo kill this. --Spangineeres (háblame) 03:47, 13 October 2006 (UTC)

Wikipedia:Attribution/FAQ

I've created the above as a proposed guideline to accompany the proposed policy. The aim is to keep this page short and sweet, giving essential examples only, and sticking mostly to the issues that are mandatory. Everything else — non-essential examples, longer descriptions, issues that go slightly off-track — can be included in WP:ATTFAQ. Needless to say, it's important that the two pages stay consistent. SlimVirgin (talk) 06:57, 13 October 2006 (UTC)

If I've moved too much, please feel free to revert. SlimVirgin (talk) 07:30, 13 October 2006 (UTC)

Organisations as self-published sources?

I've again removed that "well-known and reliable organization[s]" can be examples of self-published sources. A self-published source is one where there is no-one (or hardly anyone) standing between the writer and the act of publication. No fact-checking, no legal checks, no editing. Well-known and reliable organizations don't operate that way. To include those is to miss the point of what a self-published source is and why they're problematic. For example, we can't say that The Washington Post is a self-publisher, although of course it is, in a sense. SlimVirgin (talk) 09:02, 13 October 2006 (UTC)
I'm not yet fully at ease with the current distinctions w.r.t. self-published sources:
Re. "we can't say that The Washington Post is a self-publisher" - Correct, but I'd see it this way: an organisation that (1) has publication as its core business, and that (2) grants at least some minimum (or even more) of journalistic freedom for its reporters, that in other words does not have complete central control over every letter written on its pages, is not usually referred to as a "self-publisher". Common sense, no? The second point implies that the more control an organisation exerces over its published content ("controlling" implies it is thoroughly checked, up to the minutest detail), the more likely its publications might be, in a normal understanding, self-published. For instance, the publications by a state-owned press agency under a centralised regime, would more likely be considered self-published by that regime, than publications by a press agency subsidised by a state, under an agreement journalistic freedom.
Another example: suppose you want to write something on the topic of whether or not Microsoft has open source software:
  • If you use Microsoft as a source, you'd be able to find confirmation that they do. Microsoft has a host of legal advisers, and technical personnel, and nothing would be published by Microsoft without fact-checking and legal checks. So in the current definition not self-published.
  • Then there is Eric S. Raymond, who, using the website of the Open Source Initiative as a platform, disputes Microsoft's claims that they have open source software. Whether content published by Mr. Raymond on the OSI website is usually checked by anyone besides himself before publication, is an open question. Well, sure, OSI would have legal advisors if they need them. But if these play any role in the OSI publications is an unanswered question, and would at least require original research to establish. So, in the current definition, Raymond's publications on the OSI website would be self-published, and rejected for use in Wikipedia on a page discussing Microsoft's policies.
IMHO, this is not how it's supposed to work.
I'd rather see it work like this: Both Microsoft and Eric Raymond self-published on the issue. Even if Raymond had his writings systematically checked before publication: again, it's not up to Wikipedians to do the original research on finding out whether he did or not. Even if, by now, Microsoft maybe already has filled a library's worth of "checked" publications: publishing is not their *core* business. And then, in Wikipedia's policies and guidelines, we'd have to get rid of some of the odium now rather artificially imposed on self-publications: both sides (as well Microsoft, as OSI-founder Raymond) are best placed to characterise their own viewpoint on the matter. Wikipedia shouldn't draw a conclusion who's "right" in this issue, per WP:NPOV. Muting one of the two viewpoints on the matter by an inappropriate definition of self-publication is not a good idea. --Francis Schonken 13:31, 13 October 2006 (UTC)
Various definitions of self-published and problematic source are possible. Speaking as an engineer, if it were not for technical publications by organizations that are not usually thought of publishing houses, Wikipedia would not exist, because there would be no computers to host it, and no leisure time to write it because we would all be too busy hunting and gathering (well, maybe that's a slight exaggeration). So someplace in our policy we must acknowledge that technical publications issued with decent quality control over the editorial process are reliable sources; not necessarily unbiased, but indispensable in spite of any bias.
Also, the prohibition on Wikipedia being the first publisher of original research only applies to the finished article. I have never heard that an editor may not do original research in choosing among various sources. If I have never heard of a source, I should be able to look up how often it has been cited in academic journals. If two reputable electronics sources give contradictory diagrams of a circuit, I'm allowed to build both and cite the source that produced a working circuit (but I can't write about my experiment in the article.) --Gerry Ashton 18:21, 13 October 2006 (UTC)
Should there be links directly to the relevant sections of the FAQ from the sections of this page? JulesH 10:53, 13 October 2006 (UTC)
Probably a good idea, since people are likely to miss the link at the top (especially if they're directed here through links to particular sections). Kirill Lokshin 11:40, 13 October 2006 (UTC)

Reference standards

"Although they are self-published, these sources are often the best available. In such cases, the source may be used as long as the material used was posted by named, or well-known pseudonymous, individuals with known expertise in the subject. The individual need not be a professional in a relevant field" - with that type of exception, why pretend we are a reference work at all. We cannot afford to accept a double standard, where references on only some subjects can be of low quality.

You're confusing the rules with what they're meant to accomplish. The point is that certain self-published sources about popular culture *aren't* low quality. The idea that someone must be a professional in a relevant field is ridiculous for pop culture; knowledgeable people about pop culture are hobbyists, not "professionals". You don't get a college degree in Star Wars. Ken Arromdee 13:50, 13 October 2006 (UTC)
No, you get it in Cultural Studies, Film, New Media, or the like. There are academic disciplines that cover these things. — BrianSmithson 14:19, 13 October 2006 (UTC)

There are plenty of things Wikipedia is not and information we exclude every day. That is ok and the way it should be to produce a high quality reference work. I realize there is a large contingent of people that want to write on subjects there aren't reliable sources for, but there's Wikia or other places for that. - Taxman Talk 12:31, 13 October 2006 (UTC)

I agree completely. I have cut the section until we have consensus on wording.
Some articles about popular culture and fiction rely on self-published primary sources (e.g. posts on bulletin boards, blogs, and Usenet) or secondary sources (e.g. fan-written web sites), because few other sources exist for them. Although they are self-published, these sources are often the best available. In such cases, the source may be used as long as the material used was posted by named, or well-known pseudonymous, individuals with known expertise in the subject. The individual need not be a professional in a relevant field. Anonymous posts should never be used. Note that these sources are the exception, not the rule—few subjects have more than one or two such sites, and many have none at all. If in doubt about how to use a source in this area, consult any relevant WikiProjects for advice.
This is an enormous loophole, which we should not casually create. Marskell 13:10, 13 October 2006 (UTC)
Hell yes. That would open the door to every conceivable kind of fancruft, a loophole through which you could drive an Arcturan megadonkey and cart. Guy 14:44, 13 October 2006 (UTC)
I, for one, am happy with the current wording. It provides a useful way of working on articles for which the general consensus appears to be that they should exist on wikipedia (at least as far as consensus is manifested over at AfD) but for which sources are sparse. Academic courses on media & culture rarely produce the kind of material that it has been generally accepted wikipedia should have about these topics, particularly detailed explanations of elements of popular works of fiction, usually combined with information about how fans of those works of fiction have reacted to them. The example cited some way above of Spoo is a great one -- it relies on a self-published source the authority of which cannot easily be questioned, and it provides useful insight into the culture of Babylon 5 fandom by analysing how fans have reacted to the concept.
Also, I'm a firm believer that policy should describe as closely as possible how wikipedia editors work (at least when they are contributing their best material). It's pretty clear that most editors at popular culture articles already work like this.
I don't see it as a loophole. It's pretty clear from the phrasing that such sources can only be used when there is a lack of non-self-published sources, so all it does is extend the ability to have sources to articles that previously couldn't have any. If such articles aren't wanted, then they should be deleted as unsuitable material anyway; lack of available sources is a poor excuse to delete an article when a small change to policy can permit perfectly acceptable sources for it. And if the source isn't acceptable, one would have to wonder why a "named, or well-known pseudonymous, [individual] with known expertise in the subject" would post it. Inaccuracies in it would call into question the indvidual's expertise in the subject.
As for "fancruft", current rules about deciding whether content is encyclopedic or not should be adequate to remove it, if it is felt necessary. JulesH 14:48, 13 October 2006 (UTC)
  1. First, I disagree with this notion that policy ought to be largely descriptive. The prescriptive element is paramount. There are vast areas of Wikipedia that have no references. If people don't want to do it, why don't we just give up referencing altogether? The descriptivist argument on this is putting the cart before the horse.
  2. Why stop at pop culture? What about the physics cranks, who were the original impetus behind the content policies? "Only George Looney believes The Moon is Made of Green Cheese so it's only natural that we use his Usenet postings to describe the idea." No, take the article to AfD and make sure George Looney has no place here.
  3. Re "It's pretty clear from the phrasing that such sources can only be used when there is a lack of non-self-published sources, so all it does is extend the ability to have sources to articles that previously couldn't have any." This is what's sooo wrong about it. We can't find reliable sources so let's use non-reliable ones? Think about it—it turns the entire basis of the content policies on their head.
  4. As with previous issues on this page, that something already occurs does not mean we should encourage it. A double standard in our main space presentation should not be self-fulfilling. I oppose the addition on that basis alone. Marskell 15:11, 13 October 2006 (UTC)
(edit con)JulesH I agree with what you say above but I am confused about what wording you are supporting. The "current" wording or what was removed? I remember your attempts to find a solution to the Buffalo Nickel discussion at WP:RS and I am glad there is a version that you believe will solve these sorts of problems. However I am not clear on what you think the policy should say.--Birgitte§β ʈ Talk 15:13, 13 October 2006 (UTC)
Sorry, I meant the wording which was removed and placed in the discussion above.
Why stop at pop culture? Pop culture (and fiction, as it is stated in the wording I prefer) is a subject which the general consensus seems to be should be included in wikipedia. That is what differentiates it from the crank science articles User:Marskell is talking about. If consensus were that such articles should be included, we'd probably have to drop the entire self-published source rule, but clearly there is no such consensus.
If other topics emerge that have similar problems associated (i.e., consensus they should be discussed, lack of non self-published sources, availability of high-quality self-published sources, and those sources are not written by professional experts so would be allowed under exemption 1), they should be considered for addition too.
IMO, Wikipedia has too great a bias against using self-published sources. There are some great sources out there we could be using, but we're too frightened of losing grip on the control we currently have. It's time to experiment a little, see what we can achieve if we allow ourselves to. JulesH 15:30, 13 October 2006 (UTC)
I think the excised paragraph should stay out. We should recognise that the identity of authorities will be context-specific; for example a statement by Gene Roddenberry on the origin of Klingons could be accepted as gospel in a Star Trek article whereas a statement by Stephen Hawking on a cosmology topic would need to be attributed because while Hawking is an authority he is not the authority. But we should not go as far as dropping any requirement for authority where none can be found. That goes fundamentally against the principle of no original research, for a start, and it subverts the core value of Wikipedia as an encyclopaedia (i.e. a tertiary source, summary of secondary sources of known authority) and mooves us right in there with MySpace and all the others. If something is not reported and discussed by relevant authorities, we should not be afraid to walk away from that subject. Otherwise we'll have no way of excluding crank theories which are shunned by all but their proponents, for example. Guy 15:36, 13 October 2006 (UTC)
I believe there is a clear difference between subjects which have plenty of published sources on the topic (which we will use) as well as self-published sources which are the only source for a particular fringe theory (which we will not use), and subjects (like Buffalo nickels) where the only comphrehensive sources availble on the topic at all are self-published.--Birgitte§β ʈ Talk 15:48, 13 October 2006 (UTC)
Guy, why do you see this paragraph as "dropping any requirement for authority"? The paragraph in question contains the following requirements:
[The sources being discussed] are often the best available. In such cases, the source may be used as long as the material used was posted by named, or well-known pseudonymous, individuals with known expertise in the subject.
The point is to establish that the source is authoritative, by virtue of being the best available authority on the subject. Known expertise means a good reputation for the source. These are all things that suggest that authority is required.
Whatever you may think, there are authoritative self-published sources. We should be able to use them when we find them. JulesH 16:23, 13 October 2006 (UTC)
"If something is not reported and discussed by relevant authorities, we should not be afraid to walk away from that subject." Sometimes the relevant authorities are self-published. This happens frequently in pop culture and fiction-related fields. That is what the paragraph in question is trying to point out and explicitly allow for. JulesH 16:32, 13 October 2006 (UTC)
And what source tells us the pseudonymous source has well-known expertise? A show of hands amongst fans? The idea that being the only available source on a subject qualifies something as a reliable source, is awfully flawed logic ("if we can't find it, lower the standard"). If everyone but you drops out of your law class are you necessarily a good lawyer? JzG is right: if there are no reliable sources available—with reliable defined by a single wiki-wide standard—then we should walk away from a subject. This strikes me as incredibly self-evident. Marskell 16:40, 13 October 2006 (UTC)
There is no broad white line between reliable and unreliable that I can see. Please define what your single wiki-wide standard for reliable would be.--Birgitte§β ʈ Talk 16:45, 13 October 2006 (UTC)
I meant that in terms of application, not substance (i.e., the content policies should be applied equally across the entire mainspace). Re substance, the line can indeed hard to pin-down but I sure as hell know which side of it I'd place fan written websites on. Marskell 16:54, 13 October 2006 (UTC)
I truly believe part of this policy should be to prefer the most reliable source availble. There is way too much reliance on websites when books exist on the topic. But I think equal application across the mainspace would hinder this. Basically you would end up encouraging editors to claim that all sources above a certain threshold of reliabilty (i.e. organizational website) are to be treated equally because we must apply the policly equally. I strongly oppose such results and I thik this idea will lead there. I think there is something to be said that the application of a policy depends on the subject. We must use judgement here. Unreliable sources should never be used. However a reliable source should not be excluded simply because of it's means of publication. Of course authors of self-published sources need to extra scrunity to ensure they are reliable, but don't make the rule of thumb (means of publication) of judging reliabilty become a line in the sand.--Birgitte§β ʈ Talk 17:11, 13 October 2006 (UTC)
(Written earlier but not posted due to a connectivity failure:)
"And what source tells us the pseudonymous source has well-known expertise?" That is down to editors on a single article to decide. If they can't, mediation would assist. This is exactly the kind of thing we deal with on a day to day basis, and there's little or no difference between a self-published source and a disputed source published by a usually reliable publisher. We judge based on the general reaction of people with knowledge about the area. If most of them seem to think the source is good, we use it. If there is significant dissent, we use the source cautiously, being careful to attribute in the body of the document and not judge as right or wrong. If there is strong dissent, we don't use it. And this is the way it should be.
Nobody's talking about lowering the standard. We're talking about recognizing that there are perfectly good sources available, that are of more than high enough standard for us to use, but they happen to be self-published and in those cases (which are described as rare in the paragraph we're talking about, which indicates a presumption that any individual case is not covered) we should relax the rules and allow the source. This is already covered by WP:IAR. It's not *really* any change to the way things are at the moment. It's just codifying a case where rules are ignored on a day to day basis and where the majority of editors seems to accept that that is the right thing to do. JulesH 17:13, 13 October 2006 (UTC)
"I meant that in terms of application, not substance (i.e., the content policies should be applied equally across the entire mainspace)." But they already aren't. Living-person bios have firmer requirements on the types of source that are acceptable. Consensus at AfD has often been to keep articles that are sourced by self-published documents. We've even had Spoo pointed out as a featured article that has self-published documents. The standard we're talking about is already commonly applied. Let's recognise it. JulesH 17:23, 13 October 2006 (UTC)
"Nobody's talking about lowering the standard"—then I don't know how to read.
Re the continued descriptivist argument, see above: "that something already occurs does not mean we should encourage it. A double standard in our main space presentation should not be self-fulfilling." This is becoming circular, so I'm not going to repeat that further.
Regarding last, there are differences in broad article types (BLPs, non-BLP) and there are differences in subject matter and discipline; you're arguing for codifying differing standards for the latter. Babylon 5 should be treated no differently than Middlemarch and Frank Sinatra.
As for Spoo, if it went to FAR now I have no doubt fans would vote keep and virtually all of the currently regular reviewers would vote remove. See Halloween (film) and Sex Pistols for pop culture articles with sourcing a step above Usenet. Marskell 17:36, 13 October 2006 (UTC)

A B C

It's an excellent idea to condense policy to essentials and avoid duplication and possible conflicts. But this amounts to rebuilding a large portion of the foundation of this project. Hey, that's fine by me -- but as usual, I see things a bit differently. If we're going to do over, let's do right.

This proposal -- and the policies on which it is based -- and the new concerns it attempts to address -- and the lengthy discussion above -- send up dozens of red flags to my eye. They always have. In a nutshell, they are mostly attempts to define a bright-line rule in an area where good, mature judgement is required. Or, as older people used to shout at me when I was young and extremely foolish:

It's not all black and white; it's mostly shades of gray.

Written policy to the contrary, there really are places where a little OR is acceptable. Any experienced editor can drag out a dozen examples. Ideally, every single statement would be sourced but this is not an ideal world. Most statements in most articles are poorly sourced. This should improve. Much content is warred over and the quality of the textual support for various positions is questioned. To a point this is essential; to another, destructive.

The difficulty in trying to draw a neat line around all the permitted things which excludes all the forbidden things is that the world does not break neatly into these two components. At any moment, the borderline is going to run right through somebody's farm. So, this kind of policy -- in any organization, on any project -- becomes the occasion for bitter, pitched battles over the exact position of a tiny segment of the line. And we have seen very clearly how true this is in Wikipedia.

Here we fight it out over what sources are reliable and which not, which topics are susceptible to real sourcing and which always seem to end up sourced to primaries, which passages can be removed on sight, which commented out, which tagged or flagged, which debated on talk. This leads to the misconception that we can draw that bright line. We can't. All we can do is strive for better sourcing.

What makes this all the more difficult is that no one of us has a true grasp of the range of topics in this project. Each of us is an individual human and this project is a reflection of the sum of all human knowledge. It is absurdly arrogant for anyone to say he has a handle on the whole thing. But every time I push a piece of the line to reflect my concerns in this general topic area, you worry that it will affect that area. We're both right.

Written policy has always poorly reflected actual practice. I speak of these core content policies. These have always been written very specifically, yet they have always been applied very loosely. I feel that this is a failure of our community, as a whole, to practice self-honesty. It's a little like a fellow who says, "I only ever drink one beer a day," yet if you go through his grocery bills, you see he manages to go through 2 cases every month. I don't think we need to struggle bitterly over the number of beers drunk but we do ourselves a service (a) by being truly honest about how many beers we really do drink and (b) trying to cut back a little, when that's feasible.

What's wanted is a source grading system. Far better to assign letter grades to sources -- A, B, and C, no need for anything more complex and we have no interest in anything of lesser quality. These standards can be applied across the board, regardless of topic. Hard science, fancruft, religion, history, current events: anything. We can draw up a list of A sources and establish fairly rigid criteria for inclusion; we can even set explicit standards for B and C grade sources. These criteria do not change from field to field.

My first, rough stab:

  • A: Major newspapers, magazines, and peer-reviewed journals with searchable online archives. Books by reputable authors, reviewed in other A sources, highly regarded by experts in the field....
  • B: Books by reputable publishers, regardless of other merit. Websites whose content passes over some bar or review (not blogs). Local newspapers and magazines. Official government or corporate publications....
  • C: Primary sources. Other books and websites, if there is some evidence that more than one person had some control or input....

Okay, that's just a half-assed stab at what needs to be defined in greater detail. Here, hard distinctions can and should be made. Now, on to application:

Each article contains a references section. Here is the place to display the letter grade for each source. I strongly favor a color-coded system; it makes it easy to see at a glance what you have got. There's no need for cute icons, though; simple CSS backgrounds will do nicely. All that's wanted are 3 templates: {{sa}}, {{sb}}, and {{sc}}. These also cat the article into A-sourced, B-sourced, and C-sourced articles. (I also have another CSS trick up my sleeve, which I promise to reveal a little later.)

Now, the furious pressure to defend fair sources is off. Now, we look over the sprawling corpus of this project and see that in some fields, A sources dominate. In Mathematics or Physics articles, even a B source is questionable; a C source is beneath notice. There are a few exceptions and we can wonder why. In some social fields, say, the history of silent movies made by directors of small stature, B sources predominate. Pokemon articles, not surprisingly, get by on a few B sources and a lot of C's.

We now have a true reflection of our inevitable compromise between irresistible principles on one hand and immovable reality on the other. Most articles in most fields rely on the best available sources. When they do not, it's time to ask questions:

  • Can we find better sources?
  • Should we delete poorly-sourced material?
  • Shall we permit status quo?

I'll bet this reads like a novel, radical proposal. I say it's not. This is what we've been doing all along in our heads. This is the reason for all the windy debates, all the attempts to make exceptions for this class of article or that, all the furious battles to hold the line against the hordes of bloggers and viral marketing agents. We all know that in a perfect world, all articles would be A-sourced. We also realize this is unrealistic. Source grading just makes our process transparent.

Our shortcoming, to date, has been that we struggle to do everything all at once, to draw the bright line and hold every article to that standard instanter. This fails us twice: by causing us to delete useful material that really is as well-sourced as possible and by allowing poorly-sourced articles to languish when the need for improvement could be made evident.

We do our readers a service by honestly indicating the strength of our material. I think letter-graded sources go a very long way toward answering our critics who claim that all sorts of whiffle oozes through the cracks around here. It does. Let's be honest about that second beer. John Reid 15:27, 13 October 2006 (UTC)

I really like this proposal over the current mess of categorizing sources as "problematic" or whatnot. It reflects reality, particularly at the FA level (such as the aforementioned Pokemon). At multiple points in time I've suggested to prospective FAs that "We have a NYT article, use it". I am a little concerned over displaying the source quality in the article, and how that will affect GA (where snap-judgement reviewers say "No A-class sources, fail") and to a lesser extent FAC. Nifboy 15:44, 13 October 2006 (UTC)
I agree with John Reid's main point, but I do think that it's covered by "Ignore all rules"; in other words, Wikipedia is saying, here are our guidelines, but "ignore all rules" can trump any of them, if necessary. The letter grades are likely to break down in as many quibbles as any other categories (John already has "not blogs" in brackets; some people say "but some blogs are OK" etc., and so then we disintegrate into B3(ii) etc. and we're off up one of the tubes in Terry Gilliam's Brazil with our backsides on fire. qp10qp 16:00, 13 October 2006 (UTC)
That is a problem with GA and FAC process not a problem with the proposal. I like the basic proposal at first glance. But I do not see it an alternative to a good attribution policy. I actually think that the suggested method of grading is tooo rigid and narrow. I think it would be a good idea to work at grading sources through Wikiprojects and keep the grades subjective to that particular Wikiproject. Obviously if the grades are not absolute they would need to relegated to the talk page, which I think John will dislike. But his main premise about how we need to do less talioring in the policies is an excellent point.--Birgitte§£] ? Talk 16:04, 13 October 2006 (UTC)
I think that "as well-sourced as possible" is not good enough. It could be taken as an invitation for people writing in obscure fields to use unvetted, essentially self-published sources where it's hard to judge the actual expertise of the sources' authors. Simply because the field has been little-examined does not mean we should allow this to take place. Andrew Levine 16:53, 13 October 2006 (UTC)
  • Another person who agrees with the thrust of John's point but not the solution. The grades just give us another area to fight it out in and create division. It's definitely a grey area, and we should definitely state that front and centre. I think it's best we make that clear and try to steer and ultimately address issues through the rfc and wikiproject arenas case by case rather than attempt to legislate the clear line. We can't do it centrally, by bringing it here or WP:RS, then we're just arguing in front of a panel who may not understand the nuances of each specific instance, because these policy pages do gain a following. I think we have to keep all our guidance and policies pretty grey, I think that's the idea that Wikipedia was founded upon and why we had the WP:RFC system, to flag issues. Centralising creates a bureaucracy, and Wikipedia is not a bureaucracy. Wikipedia is not a central entity, it's a community, there's no central power, so let's not cede power centrally.
  • Returning to the issue of sources, there's always been an argument that we allow all sources and let the reader decide, which is something I can sympathise with, up until the point where you hit the grey area of presenting in a WP:NPOV. Hiding Talk 16:47, 13 October 2006 (UTC)
  • I think a source grading system is a good idea. I'm not sure it should be part of this policy; at least not without experience. Rather, I'd like to see a WikiProject to evaluate sources (and the quality of article sources), just as the Wikipedia 1.0 is doing for articles. If that works, perhaps it can be written into policy, but at this point I'm not prepared to support it as part of this policy. --EngineerScotty 16:51, 13 October 2006 (UTC)

Hard to believe I've been so quickly misunderstood. I'll try to clarify:

Letter grades don't break down as easily in quibbles because it's easier to define absolute standards for sources once it is understood that all grades of sources are permitted. It's easier for me to compromise on a point if there is an area of compromise, rather than "winner-take-all". We fight out sources on the level of sources with hard distinctions; this allows us flexibility in evaluating an article later as well sourced or poorly sourced. Current written policy has the fight all in one place: Is-this-article-sourced-to-our-standards-or-must-it-be-deleted-now.

Blogs may be permitted as C-sources; that's for the community to decide. I don't think it's hard to see how a blog might rise to the level of a B-source; I just think it's very difficult.

We don't allow unsourced cruft but some of the Pokemons come pretty close. I don't propose any loosening of restrictions at all. Poorly-sourced quasi-cruft exists today; source grading just makes such pages evident: all C sources and not too many of them, hmm, maybe delete.

We definitely can't set standards of inclusion here. The problem with current written global standards of inclusion is that they are too inflexible and attempt to draw hard lines. While this can be seen as essential to our integrity, the reality is that such global standards are often honored in the breach while local standards prevail. I think it's much more realistic to set global standards for sources -- when a range of qualities is understood. This does not resolve the conflict between global and local standards for articles but it does direct the discussion onto something previously agreed upon.

Source grading needs to be developed in detail. I suspect the best approach is gradual. I've already created the templates; go ahead, be bold, and apply them to sources as you find them. This is the "bottom-up", community-based approach to developing policy. Don't war over source grades; at present no sources have been graded so there's plenty of fresh turf to work. When a large number of sources have some grade attached to them, then will be a good time to try to write standards that reflect actual practice. Even at this point, these grades carry no "weight"; they're just "there".

Over time, perhaps, the overall grade of sources cited in an article will come to figure in discussions of that article's inclusionworthyness. I suspect that some fields will be quick to adopt local policies requiring a certain grade of sources; others will not. At no time is it necessary to write an exhaustive global policy to cover this. The key global principle is simple: We want all articles to have the best sources possible and we'd rather not have very poorly sourced articles at all.

Source grading is only relevant here because it takes so much of the pressure off of this discussion. Instead of wasting energy trying to define what we mean when we say "articles must be based on reliable sources", I suggest that all our policy really needs to say is: Articles must be based on reliable sources. One line of text, more or less. John Reid 17:43, 13 October 2006 (UTC)

This section has now gone entirely. This is a BIG PROBLEM. Popular culture definitely, and I'm sure many other areas I don't edit in NEED an exception for the often very high quality self-published sources that exist online. Remember - not every topic has professional experts. Many hobby and leisure topics only have dedicated fans and enthusiasts. This exception needs to be made possible. Not easy, but possible. This is a case where specific subject areas have different needs - exactly the sort of thing we have WikiProjects for. We should note that. (Or at least note SOMEWHERE on the page that specific subject areas may have particular requirements for sourcing - both more and less restrictive than this - and that you should consult relevant WikiProjects.

I cannot stress how important this is for this page and for the project. Phil Sandifer 16:45, 13 October 2006 (UTC)

  • Do we need to guide that only professional experts are acceptable? And how do we decide what constitutes a professional expert? I think we might have enough greyness here to get certain fans and enthusiasts through as experts. I mean, what is an expert but someone with specialised knowledge. Hiding Talk 16:49, 13 October 2006 (UTC)
Right now the exception is explicitly "professional" self-published sources. Phil Sandifer 16:51, 13 October 2006 (UTC)
Yes, but I'm here thinking how we define professional? I mean, someone with google ads on their blog is getting paid, yes? I agree with you that there's got to be flexibility to the point where we can say, look Tom Spurgeon is a bloody reliable source for comics commentary. We've talked around issues on sources at the comics wikiproject a tiny bit, and we seem to have roughly agreed that it depends on what you source as to what is a reliable source. I mean, Wizard is a publication but we're not holding to the rule that that's wholly reliable. It's all context. So yes, I agree that the concerns about this need to be met, even if a popular culture section is unacceptable people have to look at what constitutes an expert. Many fans are considered experts within the field, so why do we have to stick to a professional tag? Hiding Talk 17:56, 13 October 2006 (UTC)
Above a variety of users have expressed at length the problems with the section (it's becoming a bit repetitive). Given that this is not live, excising it seemed acceptable per lack of consensus. I would re-add something like this:
"While exceptions exist for broad types of articles (biographies versus non-biographies, for example) there are no subject based exceptions. The standards applied to popular culture articles are the same standards applied to articles in general."
I cannot stress how important this is for this page and for the project. Marskell 17:51, 13 October 2006 (UTC)
This isn't about popular culture. It's about the fact that there are over a million articles on Wikipedia, and the people weighing in on this proposal have edited a tiny fraction of them. We are not in a good position to address the sourcing needs of articles on quilting, ultimate frisbee, clockmaking, and other oddball topics. I want to make sure that editors working in those areas have the latitude to write good articles on those topics. The standards are that every article should use the best sources available for a topic. We have a good sense of what the general traits of best sources are. But we should not be in the business of imposing inflexible standards of best sourcing over a million articles. Phil Sandifer 00:27, 14 October 2006 (UTC)
The more I think of this I more out of line I this it is to be to developing this policy in a way that outlaw what is now common practice. Things which are common practice must by their very nature have consensus.--Birgitte§β ʈ Talk 18:04, 13 October 2006 (UTC)
The problem is when common practice runs up against the basic principles of Wikipedia. In this case, the principle of properly sourcing articles is not being followed consistently enough. Andrew Levine 18:21, 13 October 2006 (UTC)
Amen. Unsourced, unwikified start-ups are also "common practice". Marskell 18:25, 13 October 2006 (UTC)
And are codified here: it says (paraphrasing) you don't need to provide a source, but that you should be prepared for your content to be removed if you don't. Also, I'm sure if you ask those editors, you'll find that either (a) once you explain why attribution is important to them they'll understand and accept that the policy is a good idea or (b) that they already think it's a good idea, but feel that other people will add sources later, and that the most important thing to do first is to get the content down so sources can be found. I doubt you'd find a consensus against requiring sources, even if you only considered the editors who never add them. JulesH 18:49, 13 October 2006 (UTC)
(edit conflict)You misunderstand me I am not arguing that common practice shows consensus against properly sourcing articles. I believe common practice is to constantly improve the sourcing of articles, and I have never heard of people going a round removing the sourcing of articles. SoI thik this show common practice supports the sourcing of articles. However I think common practice shows consensus is against the deletion of popular culture articles and I think it is out of line to back-door outlaw them by outlawing the only sources they could possibly use to become properly sourced.--Birgitte§β ʈ Talk 18:55, 13 October 2006 (UTC)
  • I think most of us, if not all, sign up to Reliable sources are published materials or documents whose authors are regarded as trustworthy or authoritative in relation to the subject at hand I'm not sure why we have to expand beyond that, to be honest. Can we perhaps amend When a professional or academic expert writing in his or her area of expertise to When a recognised expert writing in his or her area of expertise? The onus would still be on the person using the source to demonstrate the recognition, so I think it equates to the same as the previous text. Hiding Talk 19:05, 13 October 2006 (UTC)
I'm prepared to take Phil Sandifer's word for it that there need to be exceptions for popular culture, but could he give some specific examples? I'm not usually involved in that sort of article, but I love popular music and films, and it seems to me that published secondary sources other than blogs and fan websites do exist for those areas. Even the obscurest British band tends to be mentioned in NME and other British music papers, for example (the journalism may not be of the highest standard, but it would suffice, I think). There are also many film and television magazines and columns in newspapers, as well as reviews. Stepping outside my areas of knowledge, I also know of magazines and columns that review video and computer games, celebrity events etc. On all these subjects there are guidebooks and various publishing epiphenomena. I actually own a book about Survivor, I'm embarrassed to say, and one about the minor footballer Robin Friday. I'm not doubting Phil, but I'd genuinely like to hear of types of exception. Are they significant enough in number to merit an exception in the main policy, or could they not be treated individually as they crop up, with a touch of "ignore all rules" thrown in?qp10qp 19:18, 13 October 2006 (UTC)
By far the best source for Babylon 5 information is the fan-run Lurker's Guide. Doctor Who information continually relies on the Reference Guide and Outpost Gallifrey. the J!Archive is by far one of the best resources for Jeopardy. And comics, don't get me started. But these are just the examples that I idly know. I assume my knowledge is not comprehensive, and that there are other areas that, like popular culture, have many of their best sources self-published online. Hence my belief that we need to clearly have an exception that defers to the WikiProjects to be allowed to identify "special case" sources. Not an "open the doors to crap" policy, but a policy that lets editors in areas bend the rules as needed for their areas. If areas become a problem, we have dispute resolution mechanisms to rein them in, but I don't anticipate that being a problem often. Phil Sandifer 00:27, 14 October 2006 (UTC)
The places I've seen this exception argued for, Qp10qp, are in articles about individual elements of a larger work of fiction. Someone writing an article on, say, a specific Pokémon or a planet featured in a Star Trek story arc. In these cases, often the only sources available are the source material itself (a primary source) and fan-written webpages. Some folks here are arguing that those fan-written webpages should be allowed when they are written by "acknowledged experts" or are the "best sources available". But this fails to address the point of who determines that these are the best sources available? Phil has mentioned WikiProjects, but that's hardly a solution. The fans of Babylon 5 get to decide that a fanpage is reliable? And this assumes that WikiProjects even exist in the area. What about topics with inactive or non-existent Projects? No one is advocating "backdoor elimination" of these topics based on Reliable Sources policies. If the primary sources are the only non-self-published sources available, source the article from primary sources. It takes time for popular culture to get academic or press attention, and it takes even longer for specific elements of that popular culture to get noticed in many cases. Thirty years ago, it would have been difficult to find much information on Speedy Gonzales. Today, there's enough information to write a Featured Article (from high-quality, academic sources) on that character. It takes time, and I don't think relaxing standards (or sanctioning widespread disregard for existing policy) is a good idea in any way shape or form. — BrianSmithson 02:32, 14 October 2006 (UTC)
As someone who works in comics and animation studies, I am reasonably confident that one would be hard pressed to get a good article on Speedy Gonzales using wholly academic sources. Phil Sandifer 02:58, 14 October 2006 (UTC)
Well, I considered writing just such an article, so I obviously disagree, but I've responded more fully on your talk page. — BrianSmithson 04:41, 14 October 2006 (UTC)
I'm not convinced that special exceptions need to be made in the policy, but, reading the comments above, I believe these articles could be written within the terms of the present wording, though, as Phil says, with perhaps a little bending of the rules here and there. The advantage of keeping the exception out would be to deny personal-opinion mongers a loophole for using certain contentious blogs and websites as sources.
The present wording has, to my delight, more or less removed the previous deprecations of primary sources, which means, for example, that Speedy Gonzalez films or cartoons can be used as sources for themselves, and that a planet in a Star Trek episode could be the basis of an article that merely recorded everything about the planet in the episode—these aren't purely primary sources, anyway, because they have gone through a process of editing and production by a large number of people. And, Phil, I don't believe that "academic" sources are required in particular, just published ones. I believe that the present text's relaxing of opposition to primary sources that come to us through a secondary medium creates far more freedom for popular-culture articles to operate than before and answers many of the needs expressed. qp10qp 12:24, 14 October 2006 (UTC)
It's been consensus at WP:FAC for a while now that an article on a work of fiction can use that fiction as a source for things like plotlines and character descriptions. See the featured articles The Wire (TV series) and Palpatine for two recent examples of this practice. Andrew Levine 02:05, 15 October 2006 (UTC)

We need better wording for what is a source

We are saying that people are sources, but that is incorrect. Only the published opinions or research of people are sources. ≈ jossi ≈ t • @ 17:18, 13 October 2006 (UTC)

I've cleared up a couple of places where this confusion exists in the 'problematic sources' section. There may be others. JulesH 17:27, 13 October 2006 (UTC)
Strictly speaking, Wikipedia only uses published sources, but people are still sources. --Gerry Ashton 18:32, 13 October 2006 (UTC)
That may be, but the terminology gets confusing if you use it like that. JulesH 18:44, 13 October 2006 (UTC)
I think we ought to leave in that people are sources. That their material has to be published before we can use it is clear, but we regard some people as reliable sources no matter where they are published, even if it's on their own blog. It's the reliability of the person as a source that counts in these cases, not the medium that published their work. SlimVirgin (talk) 20:03, 13 October 2006 (UTC)
Thanks for the clarification. I think that the current wording avoids quite elengantly this issue. If not, please add some text that explains this. ≈ jossi ≈ t • @ 20:11, 13 October 2006 (UTC)
I don't quite agree that certain people are reliable sources no matter where they are quoted for two reasons: (1) We must be sure the publication can be relied on not to fabricate or alter quotes, or quote out of context. (2) A person known for accuracy when presenting material in some sources (e.g. academic journals) many not be as accurate when joking, hashing out a problem on an online forum, or answering questions in a place where reference materials are not at hand (i.e. sitting at the front table in a panel discussion). --Gerry Ashton 20:14, 13 October 2006 (UTC)

Avoid "prematurely removing material for which reliable sources could reasonably be found"

Can we square this with Jimbo's well-known nugget:

"I can NOT emphasize this enough. There seems to be a terrible bias among some editors that some sort of random speculative 'I heard it somewhere' pseudo information is to be tagged with a 'needs a cite' tag. Wrong. It should be removed, aggressively, unless it can be sourced. This is true of all information, but it is particularly true of negative information about living persons."

? Marskell 18:06, 13 October 2006 (UTC)

The policy needs to remain brief. That quote is referenced at the bottom of the polilicy. I don't believe "random speculative 'I heard it somewhere' pseudo information " qualifies as "material for which reliable sources could reasonably be found". I do not believe further details are really needed, but if you have a way of emphasing this point briefly please add it.--Birgitte§β ʈ Talk 18:19, 13 October 2006 (UTC)
The difference is "intensitive"; what should we choose to emphasize? One person's "pseudo information" is another person's "material for which reliable sources could reasonably be found". Marskell 18:23, 13 October 2006 (UTC)
I don't think we should be overly emphasize anything. We should aim for good judgement, aim square in the middle of good judgement.--Birgitte§β ʈ Talk 18:42, 13 October 2006 (UTC)

No, agree with Marskell and Jimbo. We need to be hardasses about sourcing. I don't think it hurts to add a {{fact}} tag to a questionable statement and make a note on the talk page that someone needs to provide some references in the next 24 hours/48 hours/week or the information will be removed. But we do not need to hem and haw about this. Information on Wikipedia needs sources. Without sources, the information can be removed with extreme predjudice. — BrianSmithson 02:38, 14 October 2006 (UTC)

Basic Principles

I already know that some people here disagree about basic priniciples. If we cannot come to consensus on such things we will have a major problem here. Please add more principle to this section.--Birgitte§β ʈ Talk 18:14, 13 October 2006 (UTC)

At this moment, while we are engaged in developing the text of this page, polling editors can be waste of time, and detract from the focus we have ad so far. Let's continue with the collaborative work that so far has produced excellent results. ≈ jossi ≈ t • @ 18:25, 13 October 2006 (UTC)

I don't think the poll detracts from anything. We've reached a point where a number of editors are polarised on a particular issue, and I don't see us getting anywhere near resolving it just by continuing editing. Discussions like this have helped this kind of situation in the past. JulesH 18:41, 13 October 2006 (UTC)
I don't think of this as a poll but rather a way to reformulate the discussion on the fundemental issues of disagreement. The idea is everyone explains there views on the principles and hopefully someone can use that disscusion to formulate new principles until we can find one that demonstrates consensus. Of course it may not work, but I think it will be more productive than arguing over popular culture/fancruft.--Birgitte§β ʈ Talk 19:16, 13 October 2006 (UTC)

Good policies support existing practice rather than shape it.

  • I strongly support this principle.--Birgitte§β ʈ Talk 18:14, 13 October 2006 (UTC)
  • I strongly disagree with this; problematic practice that undermines the reliability or legitimacy of the project must be changed. When I first edited Wikipedia (2003) it was very common practice to upload images from anywhere you felt like without tagging them. Now we have reformulated policy to prevent this from being done. Other existing practices that have changed in the past three years: AfDs being decided by vote; Featured articles being written without citations (later expanded to inline citations); stub categories being created without prior approval; biographies of living people being held to the same standards as other articles, instead of higher ones; and userboxes in template-space being employed to express political and religious opinions. These practices have all been amended by policies and the encyclopedia is improving every month for it. Andrew Levine 18:28, 13 October 2006 (UTC)
    • I guess my question is How did these practices change? Was this change shaped by policy or did policy change to adapt to new practices?--Birgitte§β ʈ Talk 18:43, 13 October 2006 (UTC)
      • In the case of image coprights, userboxes, and BLP, policy changed in order to shape practice. The others were contentious and it could be said that policy and practice both exerted pressure on one another. Andrew Levine 19:08, 13 October 2006 (UTC)
        • Copyright and libel issues I believe are clear exceptions to this idea. I think the userboxes are a prime example as to why we should avoid using policy to shape practice. There had to be a better way to achieve that same result.--Birgitte§β ʈ Talk 19:24, 13 October 2006 (UTC)
          • Userboxes were distinctly contentious; and much of the contention arose when some editors decided to change practice without changing stated policy and other editors clung to policy. (If you can tell my position from that sentence, I'll rephrase it.) WP:GUS was a policy change, and if any part of that disaster was peaceable, it was. But I agree with WP:PRO on this. Septentrionalis 19:49, 13 October 2006 (UTC)
  • Support. This is true in most cases. Perhaps "good policies support existing best practice" would be a better statement. For practices that improve the overall quality of wikipedia, I'd say it is usually true. JulesH 18:39, 13 October 2006 (UTC)
    • I dislike the qualifier best because it is too subjective. How about "common practices"?--Birgitte§β ʈ Talk 18:46, 13 October 2006 (UTC)
      • I agree with "good policies support best practice." The policies have to be descriptive and prescriptive. In writing this, we're trying to do three things: (1) distil the essence of what good editors already do; (2) make sure what we write keeps the project safe legally; and (3) include anything we know the Foundation insists on. That's the package we then prescribe to editors as a whole. SlimVirgin (talk) 20:12, 13 October 2006 (UTC)

Good policies must be applied equally to all subject matter

A poll will not solve this

  • As (nearly) always. Marskell 19:18, 13 October 2006 (UTC)
  • I agree with the numbers don't matter only what can be gained from the disscussion.--Birgitte§β ʈ Talk 19:27, 13 October 2006 (UTC)
  • Polls tend to polarize and are not conducive to a posituve outcome at this stage. ≈ jossi ≈ t • @ 19:41, 13 October 2006 (UTC)
    I would also add that this discussion could have been had without resorting to a poll format. ≈ jossi ≈ t • @ 19:44, 13 October 2006 (UTC)
    • Please alter the format however you think would be helpful. I really had no intention of starting a poll.--Birgitte§β ʈ Talk 19:48, 13 October 2006 (UTC)
      • To add further, the conversation in its present form has exhausted itself. We've seen from Slim (way up there now), Phil S., and in Brigette's and Jules' replies, a kind of "descriptivist defence". And we've seen strong arguments from Taxman, Andrew L., Brian S., JzG, and myself, to the contrary. The poll at this point only invites a repetition of "A," "not-A." There's certainly enough dissent that the section as it stood should remain out; I suggest someone draft compromise wording and we go from there. Marskell 19:54, 13 October 2006 (UTC)
        • Marskell, I'm not sure I'd place myself in the descriptivist camp. I feel I'm between descriptivism and prescriptivism, in that what we do when we formulate policies is try to capture the spirit and letter of what the best editors already do, plus anything the Foundation regards as essential (whether good editors do it or not). Bear in mind that this is not a complete reformulation of existing policy. This is an attempt to merge two policies that already exist. We can tweak them here and there, but their essential points have to remain. If they don't, we're unlikely to get community consensus for the new policy. SlimVirgin (talk) 20:17, 13 October 2006 (UTC)

Please reread what we have now

Please reread what we have now. It is no longer dogmatic. I think Phil's problem's are addressed. Are there problems with the current version? WAS 4.250 19:58, 13 October 2006 (UTC)

Thanks, WAS 4.250. Can someone clarify the disagreement in simple words? ≈ jossi ≈ t • @ 20:01, 13 October 2006 (UTC)

Phil says "Spoo is a featured article based almost entirely on BB posts and Usenet, and nobody with any knowledge about the topic would criticize a one of them."[1] Does that help? WAS 4.250 20:42, 13 October 2006 (UTC)
The disagreement was on self-published sources when better sources were nonexistant. I am happy with WAS 4.25's version, which is allowing self-pushished sources when they are written by reliable experts. That leaves enough room for the things I was most concerned about.--Birgitte§β ʈ Talk 20:16, 13 October 2006 (UTC)

I like how the difference between the two sections has basically been distilled down to a three word change. That's good work. JulesH 20:44, 13 October 2006 (UTC)

Must vs should

What are we seeing as the difference between "must be attributable" and "should be attributable"? Everything in Wikipedia must be attributable because if it's not, it's OR. It needn't actually be attributed, but it must be attributable. That is the absolute essence of this policy (of existing policy too). SlimVirgin (talk) 20:20, 13 October 2006 (UTC)

I agree. Policies should probably use must for their main points and guidelines should probaly avoid must entirely.--Birgitte§β ʈ Talk 20:26, 13 October 2006 (UTC)
My dictionary distinguishes between the two as "must" meaning "to be compelled to" and "should" as "expressing obligation or expectation". If we mean to say that we want and expect something "should" should be used. If we mean to say that we will will force it to be so then "must" should be used. I would rather we didn't add to existing bad feelings by deputizing source-nazis to carry out the mandate implied by "must". WAS 4.250 20:33, 13 October 2006 (UTC)
It's not a question of source nazis, but of using the words correctly. Articles must not contain original research, and any OR is removed immediately. They must not contain any defamatory material. We're saying that those are minimum standards. To change it to "should" would suggest these things are sometimes allowed, or that we don't really mind. SlimVirgin (talk) 20:50, 13 October 2006 (UTC)
Our objective is to successfully communicate; not use words correctly. Or to argue over what is correct useage. WAS 4.250 21:10, 13 October 2006 (UTC)
I agree, and I wasn't making a point about vocabulary for its own sake, but because "must" does communicate what we need to say in this case. SlimVirgin (talk) 21:15, 13 October 2006 (UTC)
See also RFC2119. If we can keep our interpretation of must and should in line with the ones introduced there, it would be beneficial, I feel. JulesH 20:35, 13 October 2006 (UTC)
I share SV's concern about the change from must to should. Should implies that it is not a requirement but an expectation. Additionally, "should" is often used for things that we ought to do but for some reason don't - we need to stay far away from that implication. If the must language implies (must right now) then I would have reservations but I read must as an eventualist - it must be attributed and we are working on getting it there, removing spurious and controversial claims that are not attributed as we continuously improve the articles. --Trödel 21:03, 13 October 2006 (UTC)
It sounds like words are getting in the way of communication here. We should (must?) find a third alternative between "must" and "should". Perhaps a different word or a two-word choice maybe with one word in parenthesis. Any suggestions on a third choice? WAS 4.250 21:04, 13 October 2006 (UTC)
I'd say "must" is important for material "must be attributable but need not actually be attributed" to a reliable source. That sentence is the core of the proposal. SlimVirgin (talk) 21:13, 13 October 2006 (UTC)
I proposed some alternative language - not exactly right - but I think it is getting there. --Trödel 21:17, 13 October 2006 (UTC)
Using "will" instead of "should" or "must" works perfectly for me. I love it. WAS 4.250 21:25, 13 October 2006 (UTC)


Sorry, it must have been a misunderstanding on my part. I was under the impression that we are saying that all material in Wikipedia, at a certain point, needs to be attributed to a reliable source. We can leave temporarily with material that is not attributed, if there is a sensible expectation that such source can be found. Only that material that is challenged, need to be attributed within a sensible period of time, otherwise it will be deleted. That is why I chose to use should rather that must. But if this opens loopholes that will be exploited by the usual suspects, I would not object to returning to the must version. ≈ jossi ≈ t • @ 21:27, 13 October 2006 (UTC)

Changed "should"/"must" to "needs to be". I think this captures what we want to say. ≈ jossi ≈ t • @ 21:31, 13 October 2006 (UTC)

I don't see how that helps. What's the difference between "must" and "needs to"? I would prefer just "must". —mjb 03:45, 14 October 2006 (UTC)

Must vs does

In the key principles, there is this:

  • Wikipedia does not publish original research.

I understand that this statement is intended to express an ideal for the purpose of providing guidance, but it is also a refutable observation. I would like to replace "does" with "must" so that it just provides guidance. Support? Objections? Concerns? —mjb 02:04, 14 October 2006 (UTC)

Impetus behind this proposal

In case we lose sight of this, the reason this proposal was started is that, over the last year or couple of years, there have been a lot of complaints on the mailing list and elsewhere about the way V and NOR were developing. Many good editors felt they were too convoluted; the way RS developed was such a mess that it was being mostly ignored; and people were frequently complaining that the policies meant that many articles on pop culture couldn't exist, and that even some that were featured articles (for example, Spoo) were actually in violation of policy. Those of you who patrol the talk pages of V and NOR know how often we were asked what editors working on pop culture were supposed to use as sources, given that we had ruled out the only sources that existed in their areas.

Because of all of the above, some good editors announced that they were going to ignore the policies from now on, and just use their common sense. When you reach the point where good editors are having to ignore policy so they can get their work done, the policies have to be tweaked back in the direction of the best practises of those editors. A purely prescriptivist position just isn't tenable.

At the same time, we can't throw the baby out with the bathwater. We have to retain the essential points (the spirit) of V and NOR, or the community and the Foundation won't support the new policy. So we have a very careful juggling task ahead of us. The new policy has to be good-user-friendly, while not opening up any loopholes for careless editors, bad-faith editors, newbies, and so on. This therefore has to be a cautious and conservative streamlining of V and NOR that, at the same time, takes into account the complaints we've heard over the years and accommodates those editors' concerns. If we can produce that combination, we'll have a winner. SlimVirgin (talk) 20:46, 13 October 2006 (UTC)

Exactly right. And I think we have that now. WAS 4.250 20:58, 13 October 2006 (UTC)
Slimvirgin: I must admit I sometimes wondered if you actually read the talk page at WP:V. I was also originally against the idea of pushing WP:RS into WP:V, as I didn't think it could be done without making things worse. Whenever I first read this policy, it was obvious you had paid close attention to everything I could remember from those talk pages and did your best to address everyone's concerns. The current version here is not only readable, but focused on what is truly important for the enitre project. I am quite impressed by what everyone has accomplished here but I do not think anyone surprised me so much as you did. This is such an improvement over the previous policies, I have trouble finding problems with it at this point. Besides settling the should/must issue what is really still outstanding?--Birgitte§β ʈ Talk 21:27, 13 October 2006 (UTC)
Kudos to SlimVirgin, EngineerScotty and all others that contributed to this so far. The work and process by which we have created this page makes me proud of being a Wikipedian, and shows what is possible and how great (an fun) collaborative editing can be. ≈ jossi ≈ t • @ 21:39, 13 October 2006 (UTC)
Thank you, Jossi. Kudos to you too and to everyone else who has contributed. You're right that it's been great collaborative editing. :-)
I haven't read over the latest version, so I'd like a chance to do that, but when I last looked, it was looking very good. I'm still unsure about having removed the primary/secondary distinction, because a lot of editors considered that a very important part of the policy. (I found it useful, but was in two minds as to whether its usefulness outweighed the confusion it sometimes introduced.) Also, the popular culture exception seems to have been removed or reworded, which I would prefer to see back in.
Birgitte, I think you may have misunderstood my stance on V and NOR. They were under frequent assault by editors trying to change little bits here and there, but without taking into account that there had to be coherence between V, NOR, and RS. I was very conservative when it came to that kind of random tweaking, but I've been wanting to see the pages merged for quite some time. In fact, I tried to merge them and WP:CITE when I was a new editor myself, but was reverted after about three seconds. :-) SlimVirgin (talk) 21:43, 13 October 2006 (UTC)
I won't take it as a personal insult that you once reverted my change on WP:V from 'researcher' to 'expert' then. :)
We're down to less than 2,000 clear and concise words (including links and references) that seem to cover everything necessary to me. If we exclude the sidebar with links to other policies, that's less than 100 words more than WP:V alone. And WP:NOR exceeds 3,000 words by itself. Condensing 5,000 words to 2,000 is a remarkable achievement. JulesH 21:52, 13 October 2006 (UTC)
And we haven't begun to consider the novel which is WP:RS. --EngineerScotty 21:55, 13 October 2006 (UTC)
Which was over 6,000 when I last checked. SlimVirgin (talk) 22:01, 13 October 2006 (UTC)
That's 6,500 words. Combine it with one of the others and we have a novelette. :) JulesH 22:02, 13 October 2006 (UTC)
I do like the page, though I think verifiability/attribution and OR can be distinguished to the point that having them under this title isn't right. I've already said it to Slim, but to repeat here: given that you can create OR while attributing every word, OR should not be considered merely a component of attribution. Merging definitely has its pluses, but maybe not to this title.
I don't want to conflate this with the pop culture issue. Regardless of whether this goes through, that will remain outstanding. Marskell 21:55, 13 October 2006 (UTC)
Marskell, the phrase that everything must be attributable (whether or not actually attributed) takes care of OR, which is why I prefer the word "must." If something is not attributable (that is, no source for it exists) it is OR. SlimVirgin (talk) 22:04, 13 October 2006 (UTC)
Tim, maybe this will help to clarify. It's true that the individual parts of an edit can be reliably sourced and yet the edit itself is OR because it synthesizes those bits in a novel way that advances a position. That is OR, and that is your concern about this proposal: that we don't say enough to make clear that that's not allowed.
But what that means is that the edit as a whole is not attributable. It's various parts may be, but the synthesis is not. And we do make clear that material "must be attributable." We could simply strengthen that with a sentence or two if you feel it's not clear enough. SlimVirgin (talk) 22:10, 13 October 2006 (UTC)
That's an insightful way to look at it. Can we police it through this page only? I encountered this exact problem early today, here (do follow the link to the bottom—it dovetails with my point perfectly! :). I like the idea of a seperate NOR page to deal with the synthesis issue. This isn't an invitation to continue these massive, rambling policy essays (are you going to have a go at NPOV next :), but just an observation that distinct issues should be treated distinctly. I'm still not convinced that NOR shouldn't be distinct from attribution... Marskell 22:31, 13 October 2006 (UTC)
That was why I chose the word "attribution" for the new policy, and why I want to retain the word "must," not "should."
NOR and V are summed up with: "All material must be attributable to a reliable source, but need not actually be attributed. Any material that is challenged or likely to be challenged must be attributed."
I think we can police it through this page in combination with examples and a fuller explanation on WT:ATTFAQ. Although we're back to having two pages, it's clear from the title that WP:ATTFAQ is a subsidiary of the policy and dependent on the policy for its status, so the pages would not have parity or be in competition with each other, as happened with V, NOR, and RS. SlimVirgin (talk) 22:46, 13 October 2006 (UTC)
SM, is the current formulation, using "needs to be attributable", too weak? Should be revert to te must be attributable formulation? ≈ jossi ≈ t • @ 22:51, 13 October 2006 (UTC)
"Needs to be" is just another way of saying "must be," so I think we should have "must be." SlimVirgin (talk) 02:48, 14 October 2006 (UTC)
If you thought they meant the same thing then there would be no point in your advocating one over the other. WAS 4.250 02:56, 14 October 2006 (UTC)
"Must" is shorter and clearer, that's all. SlimVirgin (talk) 03:50, 14 October 2006 (UTC)

Design freeze?

I would propose a "design freeze" at this point (and a page protection, maybe?). We could then "present" this proposal to the Foundation, Jimbo, et all for comments. Would this work? ≈ jossi ≈ t • @ 21:43, 13 October 2006 (UTC)

I agree. Freeze for comments. So everyone can read and comment on the same version. WAS 4.250 21:47, 13 October 2006 (UTC)
Clearly the FAQ isn't finished. Perhaps we should wait until both are ready...? JulesH 21:55, 13 October 2006 (UTC)
We can freeze this and work on the FAQ. --EngineerScotty 21:57, 13 October 2006 (UTC)
Is there general agreement that this is a good version? Jayjg (talk) 22:00, 13 October 2006 (UTC)
One last suggestion: "The threshold for inclusion in Wikipedia is whether the material can be attributed, not whether it is true." This is a lot like a current sentence in WP:V that has been criticised before. Changing it to either "...attributed to reliable sources..." or "...not just whether it is true" would address the criticism there, that taken by itself, the sentence makes wikipedia's reliability sound flawed: "we'll publish anything somebody has said without regard to whether it is true or not" is a valid interpretation of that sentence. JulesH 22:06, 13 October 2006 (UTC)
Jules, "not just whether it is true" misses the point. Truth doesn't come into it — at least not directly. We simply report what reliable sources are saying. End of story. SlimVirgin (talk) 22:12, 13 October 2006 (UTC)
"attributable to reliable sources", then. Emphasise that if the sources we use are reliable, then so are our articles. JulesH 22:16, 13 October 2006 (UTC)
Sorry, my comment wasn't constructive. We could possibly just not mention the issue of truth. The reason I always liked that phrase (verifiability, not truth) is that it immediately told new editors what was going on i.e. that we don't want to know what they think is true, but we do want to know what reliable sources are saying, true or not. We don't know what is true for any given issue anyway, and usually there is no "truth" involved, just opinions. Some reliable opinions, some kooky ones. We report the reliable ones. SlimVirgin (talk) 22:18, 13 October 2006 (UTC)
I've always liked the sentence myself, too. It sums up the point very neatly. But it has generated negative PR in the past. I think emphasising reliable sources is definitely the way to fix that. JulesH 22:22, 13 October 2006 (UTC)
No, don't freeze it. That's not the way to develop policy. It's okay if people are commenting on different versions, comments can always be changed too. JYolkowski // talk 22:21, 13 October 2006 (UTC)
Maybe you are right... if the proposal is really good' then it will not be challenged and changed dramatically. There are enough people involved in the creation of it, that I am sure it any massive challenge can and will be well defended. After all, we have summarized the main aspects of V, NOR and RS, and hopefully captured the spirit of these without diluting them. ≈ jossi ≈ t • @ 22:28, 13 October 2006 (UTC)
Please do not freeze this yet. I only just discovered this proposal discussion, and I have an issue I need to bring up that has only been slightly touched upon, from what I've read so far. I would like to work up and post some notes about it, but it's 3 am here as I write this, so I'll need a few hours tomorrow to bang out my thoughts. Thanks, --Aaron 06:49, 14 October 2006 (UTC)

FAQ

The Wikipedia:Attribution/FAQ still needs lots of content. I've added a section relating the what I call the Wikipedia Paradox (don't create that article, please, as it would be filled with OR)--why we think we're a reliable encyclopedia (or can become one) when we clearly don't meet the requirements of this policy, or WP:RS. See what you all think.

Also, I've put in a placeholder for primary/secondary/tertiary sources. While the distinction is confusing on the policy page, it might be useful on the FAQ. I'd like User:Phil Sandifer to fill it in... :)

--EngineerScotty 22:54, 13 October 2006 (UTC)

Applying this policy

Here is a real life example: A comment made today by an editor at Wikipedia_talk:Biographies_of_living_persons, with this question: "I need to know whether Youtube is reliable source".

How our proposal stands in regard of this question? Does it provide an unambiguous answer to this question? ≈ jossi ≈ t • @ 22:59, 13 October 2006 (UTC)

No, it doesn't, but WP:ATTFAQ might. The answer is that Youtube is not a reliable source because anyone can contribute to it; although if all they're doing is using it as a place that stored a reliable source (e.g. a clip from a documentary by a reliable source), then we can link to it, but Youtube is not the source in that instance; the documentary is. Lots of people get confused on that point. It's perhaps worth reiterating in the policy. SlimVirgin (talk) 23:05, 13 October 2006 (UTC)
(edit conflict) I am a little unclear on what is being asked there. 1) You can view YouTube a a convience link to a CNN report (provided it was officially uploaded by CNN and not a copyvio). In that case YouTube is not the CNN is, this is equivalent to asking if Wikisource is a reliable source. The question is invalid because these two sites are simply a convience link to the actual source. 2) YouTube in regards to it's original content can be seen as any other self published source and evaluated per the rules on self-published sources. Which basically means it would be acceptable only it were autobiographical or a reliable expert commentating on their area of expertise. --Birgitte§β ʈ Talk 23:10, 13 October 2006 (UTC)


[edcon] Youtube clips may be useful as primary sources with caution if there are no doubts about authenticity; youtube seems to be worthless as a secondary source. Sound right? --EngineerScotty 23:10, 13 October 2006 (UTC)
Yes. That is my view as well. Providing that the authenticity of the material can be asserted (i.e. it was uploaded by MSNBC) a YouTube video would be a good convenience link, although you could also link directly to the msnbc.com website. Other reasons why a specific clip can be not used as a convenience link, is if the video clip is a copyvio (a known problem in YouTube). ≈ jossi ≈ t • @ 23:16, 13 October 2006 (UTC)
EngineerScotty, I'm not sure I understand your point. Youtube would be a convenience link, not the source (primary or secondary). Or did I misunderstand? SlimVirgin (talk) 02:45, 14 October 2006 (UTC)
Sorry, I do see now. I was being dense. SlimVirgin (talk) 03:21, 14 October 2006 (UTC)


Another case

According to the current proposed formulation, can Media Matters or National Review be considered reliable sources for anything other than the publications themselves? ≈ jossi ≈ t • @ 23:36, 13 October 2006 (UTC)

I already treat them as advocacy groups whose claims require verification of sources. Mangoe 23:41, 13 October 2006 (UTC)
Should we give such examples in the FAQ? ≈ jossi ≈ t • @ 23:50, 13 October 2006 (UTC)
No clearly questionable sources. Anything worthwhile should be also available from a reliable source. --Birgitte§β ʈ Talk 23:43, 13 October 2006 (UTC)

What about use of such sources when writing about some controversy, to give an example of a notable media source making a particular criticism? Ken Arromdee 13:07, 16 October 2006 (UTC)

Some other cases

Recently there were long discussions at WP:RS regarding material such as:

  1. Court papers, such as affidavits, rulings, etc.
  2. A sales brochure of a certain card dealer
  3. A transcript of a TV/Radio interview that is not available from the station, but posted on a blog

Does the proposal or the FAQ cover these? ≈ jossi ≈ t • @ 00:01, 14 October 2006 (UTC)

  1. This is a no-brainer yes.
  2. I don't see why not, though it is a bit of a tricky source to access. But we don't seem to be particularly set on that.
  3. Probably not, as it's a problematic source, though in practice I'd IAR unless there were actually a belief that there was a problem with the transcript.

Phil Sandifer 03:04, 14 October 2006 (UTC)

Re: #1 - only if they're publically available. Any sealed papers (or what is purported to be the sealed papers, actually) shouldn't be used unless somebody reliable has published them first. (But this is probably obvious from other points.) Kirill Lokshin 03:11, 14 October 2006 (UTC)
For the transcript on a blog, not only is there an issue of accuracy, but also a possible copyright violation. --Gerry Ashton 03:39, 14 October 2006 (UTC)

A few questions about application

Before I go further, I'd like to ask the editors with in-depth knowledge of this proposal how it would apply with the following:

  • Assuming the term can satisfy people on notability grounds, I should think that, assuming the sources are respected in the relevent communities, they ought be usable. Phil Sandifer 01:41, 14 October 2006 (UTC)
  • A list of incidents or occurances built by editors. List of films by gory death scene is the ur-example. All entries are verifiable by watching the movies - but should every one come with a citation of a respected source discussing the specific acts of violence?
  • This is probably a more complicated issue, actually. The films themselves would be suitable sources for a hypothetical list of films by death scene, but the actual list is making a value judgement—that some death scenes are gory while others are not (and, glancing through the list, there certainly seem to be a number of entries for which this is not at all obvious)—and would therefore need to be more thoroughly sourced. Kirill Lokshin 02:54, 14 October 2006 (UTC)
  • An uncontroversial article about an old video game, containing information gained from playing it. Does there have to be a reference to a game review saying the same things? (Some have claimed as much in the past.) Is Cosmo's Cosmic Adventure allowed to have an illustrated description of the game's ending without citing a reputable publication? I do not enjoy the prospect of making emergency backups of several hundred of these, then going through two decades' worth of computing magazines, not to mention contestions of reputability. --Kizor 00:20, 14 October 2006 (UTC)
Phil, could you comment on the #Some_other_cases above? Thanks. ≈ jossi ≈ t • @ 02:18, 14 October 2006 (UTC)

This policy proposal is not designed to substitute for thinking

This policy proposal is not designed to substitute for thinking. It is a misunderstanding of human behavior, of Wikipedia policy, and of this proposal to ask it to substitute for actual application of educated informed judgement on specific questions of proper attribution. WAS 4.250 01:35, 14 October 2006 (UTC)

Absolutely. That is why we have this in the policy: ≈ jossi ≈ t • @ 01:40, 14 October 2006 (UTC)
  • Building an encyclopedia requires the use of good editorial judgment and common sense.
  • Formal rules are not a substitute for good judgment and intellectual honesty. Formal rules cannot determine whether material is notable or relevant to a particular topic. Our goal is to represent significant, published opinion fairly and without bias.
Amen. This policy cannot be black and white, or impose a flawless set of standards that can be applied across all million articles. Phil Sandifer 01:41, 14 October 2006 (UTC)

I've just read through and it's looking good. The only point of real contention that I can see is the popular culture section. This was exception three in the Problematic sources section. It read:

Some articles about popular culture and fiction 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 used was posted by named, or well-known pseudonymous, individuals with known expertise in the subject. The individual need not be a professional in a relevant field. Anonymous posts should never be used. Note that these sources are the exception, not the rule — few subjects have more than one or two such sites, and many have none at all. If in doubt about how to use a source in this area, consult the relevant WikiProject for advice. This provision applies only to articles that fall clearly within the popular culture or fiction categories.

Can people say here whether they support or oppose the inclusion of this passage, and say why? SlimVirgin (talk) 04:37, 14 October 2006 (UTC)

I support it strongly in a wider form - there are no doubt other areas that need this, and we shouldn't set up a paradigm of drawing endless exceptions. In its current, narrow form I think it does more harm than good. Phil Sandifer 04:51, 14 October 2006 (UTC)
Phil, even this is attracting strong opposition. There's almost no point in trying to extend it beyond popular culture. SlimVirgin (talk) 05:48, 14 October 2006 (UTC)
Then I oppose the popular culture section, but I also oppose the entire policy - this is central to repairing our sourcing problems. (That said, the change from "professional" to "expert" goes a long way, and it's possible my concerns could be addressed through some revisions to this section) Phil Sandifer 06:07, 14 October 2006 (UTC)
I support it in principle, but would rather see WP:RS be the place where what is or isn't an acceptable source for certain kinds of claims about certain kinds of topics. —mjb 05:01, 14 October 2006 (UTC)
(1) The proposed text needs copy-editing. (2) I suggest that the rules for using these looser sources be constrained more severely. (3) There's the problem of what happens over time to these sources (a major issue at FAR/C). Tony 05:44, 14 October 2006 (UTC)
Tony, if you have time, could you produce a version of the above, copy edited and with the tighter constraints, and post it here? It would be interesting to see what you come up with; more constraints might satisfy the people who are opposing it. SlimVirgin (talk) 05:50, 14 October 2006 (UTC)


Howabout:

Some articles concern topics which are notable and encyclopedic, but which are not well-documented in scholarly or professional sources or literature; or which contain notable aspects which are not documented in a fashion anticipated by this policy. 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 used was posted 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 — few subjects have more than one or two such sites, and many have none at all. When such sources are used, they should 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 which are not addressed in professional literature. Note that unorthodox or fringe theories in academic or professional subject areas, such as science, are not covered under this provision, even if the theory in question is not directly addressed in the mainstream literature.

I've generalized the wording, but added a few restrictions:

  • If editors in a topic object to a source, out it goes. We may want to loosen this a bit, otherwise we may get cruft-haters objecting to every source on crufty topics they don't like, then demanding deletion because what remains is OR.
  • Explicitly states that such claims should be attributed in prose; stronger evidence is required for claims of the type "X is true" as opposed to "S states that X is true."
  • Made it plain that scientific crackpottery does not qualify. Since science is an acadmic pursuit, all claims which fall under the rubric of science are expected to use high-quality sources; the fact that Aetherometry is largely ignored by the mainstream science literature doesn't entitle the "rise of the latrines" fellows to claim "our topic isn't covered by first-class sources" and thus self-publish their junk here.

One topic area I often write in, just as an example, is roadgeek articles. Few professional secondary sources exist for this; though lots of informal web sites provide seocndary sources. Primary sources are better--many government road departments or DMVs have detailed web databases available--but writing a decent article on these topics requires either using web-based secondary sources or engaging in a level of synthesis which easily qualifies as "original research".

--EngineerScotty 06:35, 14 October 2006 (UTC)

Looks pretty solid to me as is, Scotty. It will resolve a lot of issues with these types of articles. ≈ jossi ≈ t • @ 06:48, 14 October 2006 (UTC)
I'm not convinced that requiring prose attribution is a good idea. Many of these sources are uncontroversial, and frankly in these cases a prose attribution will look a bit silly. Other than that, I like it. JulesH 07:16, 14 October 2006 (UTC)
I think we'd need to add political subject areas and living persons to the last sentence, otherwise we're opening things up to conspiracy websites: "Note that unorthodox or fringe theories in academic, scientific, or political subject areas, as well as allegations about living persons, are not covered under this provision, even if the theories or allegations in question are not addressed in the mainstream literature."
Perhaps "unorthodox or fringe theories in mainstream subject areas", covering all such content without explicitly naming them...? JulesH 07:35, 14 October 2006 (UTC)
I disagree with the exceptions because exceptions always damage the authority of policies, and also because we have "ignore all rules" as a resort for good-faith articles that cannot easily operate within the policy.
As I said higher up the page, I believe that the present text has met the need of the vast majority of popular-culture articles by no longer deprecating primary sources in principle but only unpublished sources. Films, comics, songs, TV shows etc., since they come to us as produced by a large number of people through a secondary medium, can now be used as sources for themselves without contradicting the policy. In other words, an article about a planet from an episode of Star Trek could be constructed by describing everything about that planet in the episode, giving the episode, with its credits and broadcasting history, as the reference; no need at all to quote what people say about it on fansites. Of course, this amounts to writing from within the fictional world, perhaps, rather than through the refraction of critical or historical comment, but writing from within the fictional world has never been banned by Wikipedia, only deprecated against a more critical style where secondary sources do exist.
I don't really agree with the invitation to amateurism provided by the following in either version (this from Tony's):

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 used was posted by known individuals (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.

This exception may apply in a certain number of cases but in my opinion not a large enough number to require an exception clause in the policy; let each good-faith article in that category have its head in accordance with commonsense, critical judgement, and "ignore all rules". The above clause offers a loophole as big as a barn door for political extremists and pseudo-science nutters to justify the quoting of their pet websites and blogs on Wikipedia: and that's the real problem, not rule-oozing in odd little corners of popular culture. qp10qp 13:01, 14 October 2006 (UTC)

Thumbs down, still. Re Slim's initial question ("Can people say here whether they support or oppose the inclusion of this passage, and say why?") I realize that this talk is like 300k, but do read above.

Someone answer me this, because it wasn't answered earlier: if we can't establish that a given site is itself reliable how can we possibly determine which "named, or well-known pseudonymous, individuals with known expertise" are posting to the site?

If compromise wording comes forth, I definitely agree with demanding explicit attribution ("S says...") per Scotty though the suggested addition is bordering on a small essay. The rest of the page is crisp. Marskell 13:25, 14 October 2006 (UTC)

My suggestion would be to remove all of the exceptions and just say to use the most reliable sources that you can wherever possible, but that there may be cases where self-published etc. sources may be acceptable. Then cover specific examples of exceptions elsewhere. JYolkowski // talk 01:55, 16 October 2006 (UTC)

Anti-religious, racist

Jossi, can you say what you have in mind in adding these? "Racist" is covered by "political," and I can't think of an anti-religious website that is widely acknowledged as extremist; but even if such a thing exists, it would be covered by "political" or "any other." SlimVirgin (talk) 06:19, 14 October 2006 (UTC)

I think that if you keep it only in the "political, and religious" realm, we are asserting the POV that only in these two you can find extreme views. We can either add more exmples of extremist websites, or remove the examples. I will try to edit that section so that we do not include examples at all, and see if that works. ≈ jossi ≈ t • @ 06:23, 14 October 2006 (UTC)
That works better, thank you. SlimVirgin (talk) 06:33, 14 October 2006 (UTC)

Questionable sources should usually not be used...

Having edited Stormfront.org and some other highly contentious areas of Wikipedia, I have some reservations about the following sentence.

“Questionable sources should usually not be used as sources except in articles about themselves.”

It is an unfortunate truth, IMHO, that people with extreme viewpoints, especially highly charged ones, will almost always twist any rule to justify either retaining material that supports their viewpoint, or as a means to remove material that contradicts their viewpoint.

If you leave in the word “usually”, you have opened the door wide open for an edit war (having many battle scars myself, I feel entirely confident in that statement.) As Admin and other potentially fair minded editors are often stretched pretty thinly, the usually outcome is for the extremist position to hold for an extended period of time, as the other editors simply get worn down and leave.

I hate black and white statements as much as the next person, and would like to leave some wiggle room for when cold hard logic would indicate that the rule should be set aside in a particular case. However, as currently written, at least for the areas that I frequently edit in, I can see certain people running a tank through the door that has been left cracked open.

I should say that overall I think Wikipedia:Attribution is a better approach than what is currently employed, and looking forward to it becoming official policy. Thanks Brimba 07:03, 14 October 2006 (UTC)

Good point, Brimba. SlimVirgin (talk) 07:26, 14 October 2006 (UTC)
The problem is that "questionable sources" covers a huge range, including any source with no fact-checking process. To say that any such source cannot be used is taking the matter too far. For instance, during the James Frey controversy, his publisher admitted to not performing fact checking on memoirs. Should we therefore ignore all memoirs published by them? That doesn't sound reasonable to me.
Perhaps moving extremist sources into a section of their own and stating that they should never be used except in articles about themselves is better. JulesH 07:33, 14 October 2006 (UTC)
I am not sure that there is a perfect answer, but I think some improvement could still be made. Perhaps tweaking the section “A questionable source”.
As an example of what I mean, look at Illegal immigration to the United States, it is currently a poster-child for such problems. The editor(s) involved will claim that every rule has been followed, and everything is fully cited. And yet, in truth, almost every rule has been twisted and broken, and the citations are mostly worthless and unwarranted. You have an extremist editor who is dug in and (except for Jossi), almost everyone else now avoids the edit wars that go with editing/fixing that page.
People are human beings, and we can never entirely “legislate” away such problems. People with extreme views will always exist, and the smart ones will always find a loop-hole which will allow them to do largely what they wish. The question is, can we build into the new policy enough safeguards so that at the very least, the not so intelligent ones can be rendered harmless? Brimba 08:40, 14 October 2006 (UTC)
If we explicitly stated that a consensus of editors could deem a source unreliable and hence unsuitable for inclusion, then editors insisting on using that source afterwards could be dealt with through standard procedures for persistent policy violation (e.g. warning, then edit block, perhaps followed by a page-specific ban in cases of repeated offense). As long as the consensus were clear, it wouldn't be a "content dispute" of the kind that currently doesn't get dealt with in such ways, but something more like a 3RR violation. As this is a somewhat novel approach, I'm only suggesting it very tentatively, but it seems worth considering to me. JulesH 12:57, 14 October 2006 (UTC)
I would certainly oppose any use of consensus for solving disputes. The problem with consensus is that in fringe topics, is you can easily get a “consensus” that George W. Bush was behind the 9/11 attacks; that IRS has no legal authority to tax anyone and therefore people in the know have no obligation to pay taxes; or that as it is a well established fact that the US Supreme Court error in (fill in the blank) and its holding in that case has been made moot though (fill in blank). Each of these three examples lead directly to fringe areas of Wikipedia that one point had enough editors with that viewpoint to validate them though a simple consensus. Once that occurs, anyone objecting would have to go out and round up editors to take their side, and of course the “consensus” side would do the same…. In theory, someone on a white horse will ride in from above and force through the rules that ultimately would correct the problem; However, the number of people riding white horses and willing to jump into edit wars seems to be somewhat lacking. Thanks Brimba 14:50, 14 October 2006 (UTC)
You can't legislate clue, and you can't close the door to malice. No policy will solve the policies at Stormfront.org. The closest thing to a solution we have is WP:BP. Phil Sandifer 15:22, 14 October 2006 (UTC)
You can't legislate clue, and you can't close the door to malice. Absolutely right. And trying to write instructions that legislate clue and really nail it down so you can't get a truck through some crack in the instructions (the crack is where the instructions suggest actually thinking rather than pretendimg the world is black and white) results in overgrown deformed monsters like WP:RS. WAS 4.250 15:35, 14 October 2006 (UTC)

A hypothetical case

A think tank has a reputation for producing solid research in the field of economics. It is widely known that it employs experts in the field to produce papers for it, which it then publishes without independent fact checking. The papers are published under an organizational banner, without the name of the individual expert author. Is this source acceptable?

A few points of analysis:

  • The papers are probably self-published. There is no independent review; the think tank instructs the author what to write about and then publishes the result without checking it.
  • While the author is an expert, there is no way to verify this fact.
  • Could the organization itself be called an expert? Its reputation may go some way towards this, but it seems a stretch to apply the word.

Ideas? JulesH 08:01, 14 October 2006 (UTC)

The papers aren't self-published if they're being published by a respectable think tank. No respectable think tank would publish what the authors wrote without reading it to make sure they wanted to be associated with it. SlimVirgin (talk) 00:55, 15 October 2006 (UTC)
I'm underwhelmed. A considerable number of "think tanks" are highly partisan, operating as a kind of intellectual engine for this or that political party. They have grand names and intellectual pretensions, but are blinkered. Solid research nurtured (financially and otherwise) within a good think tank can then be published after traversing the same hurdles (no higher or more numerous) as those for anyone else: the authors can submit their works to peer-reviewed journals or to reputable publishers. -- Hoary 07:30, 17 October 2006 (UTC)
Retrieved from "https://en.wikipedia.org/w/index.php?title=Wikipedia_talk:Attribution/Archive_2&oldid=1190098067"