Wikipedia talk:Verifiability

From Wikipedia, the free encyclopedia

The project page associated with this talk page is an official policy on Wikipedia. Policies have wide acceptance among editors and are considered a standard for all users to follow. Before editing this page, please make sure that your revision reflects consensus. When in doubt, discuss on the talk page. Always remember to keep cool when editing. Changes to this page do not immediately change policy anyway, so don't panic.
Shortcut:
WT:V
This talk page is automatically archived by MiszaBot II.
Any sections older than 10 days are automatically archived. An archive index is available here.
Archive
Archives
Archive 1 Archive 2 Archive 3
Archive 4 Archive 5 Archive 6
Archive 7 Archive 8 Archive 9
Archive 10 Archive 11 Archive 12
Archive 13 Archive 14 Archive 15
Archive 16 Archive 17 Archive 18
Archive 19 Archive 20 Archive 21
Archive 22 Archive 23 Archive 24
Archive 25 Archive 26 Archive 27
Archive 28
About archives

Contents

[edit] Exceptionaly high-quality reliable sources?

Where is the community consensus to make such a drastic and far reaching change to a policy page? The net effect of all those qualifiers would raise the bar so very high, that very few sources would qualify. To me that would include only the very most prestigious peer reviewed journals, with very well known and unimpeachable authors. The criteria has been: Exceptional claims require exceptional sources. Sorry, but this requires wide and deep community consensus before being changed. — Becksguy (talk) 23:34, 20 May 2008 (UTC)

I changed "exceptional sources" to high quality reliable sources, exactly as it is mentioned further on in the text. The reason is that "exceptional sources" can be misinterpreted as being exceptional in some other respect. E.g. there are "exceptional" journals that exclusively deal report on anomalous phenomena. Needless to say, such journals often have very bad reputation in the scientific community. This clearly is not the intended interpretation of "exceptional".
So, I thought, why not simply say "high-quality reliable sources" as is done further on in the text. But then someone else demanded that the qualification exeptional (in the correct interprepretation, of course) is still needed. If this is changed back I would prefer "high-quality reliable sources" over just "exceptional sources", because pseudoscience POV pushers are guaranteed to misinterpret the meaning of "exceptional" :) Count Iblis (talk) 03:14, 21 May 2008 (UTC)

Fine, but we still need exceptional community consensus before a policy page can be changed. This change will have far reaching effects anywhere this policy is cited, especially in XfDs. It needs to be reverted. — Becksguy (talk) 21:03, 22 May 2008 (UTC)

What is an "exceptionaly high-quality reliable source"? For example, would investigative reports in major newspapers count? Wire service reports that have been widely published? Mainstream scholars writing for large publishers? ·:· Will Beback ·:· 08:37, 1 June 2008 (UTC)

[edit] Good faith efforts to verify

I think we should include in here that if you find something that isn't sourced and you're about to remove it, you should first run a quick Google search and see what comes up. 9 times out of 10 the assertion will have some basis which will come up quickly. If we're really trying to create a good encyclopedia, we shouldn't simply discard the input that people who don't know how to cite have added. Good faith, to me, means that you're trying to create a quality, information-filled encyclopedia rather than pushing your own point of view. With that comes an obligation -- if you're going to cut something, you may have to do a little research into a topic or a position that you don't necessarily believe in. Impin | {talk - contribs} 17:43, 22 May 2008 (UTC)

Generally, I agree... But not always. If I know for sure that an unsourced statement is factually incorrect, I am not going to bother looking for sources. I am going to remove it. If I suspect that it is incorrect (but do not know for sure), then I will tag it, put a note on the talk page, and wait a month or so. If no one bothers to add a citation, then I will remove it. But I am not going to spend time researching something that I don't think is accurate. Remember, the onus to source material lies with the person who wants to add it, not with the person who questions it. Blueboar (talk) 19:52, 22 May 2008 (UTC)
Perhaps more exactly, that one doesn't want to research what one doesn't think i accurate unless there is a reasonable possibility of either replacing it by something which is accurate, or settling definitively that its wrong and certainly must be removed. I've done some negative research from time to time, to give a better reason for removing something than my inability to find it--it's the best way to avoid fighting the issue back and forth. DGG (talk) 02:41, 23 May 2008 (UTC)
I don't think its necessary to explain what a reasonable good-faith effort is. A google search, if editing in a subject you don't know much about about is certainly the minimum. --Relata refero (disp.) 12:17, 24 May 2008 (UTC)

[edit] Speculative text

“and speculative text that cannot be sourced should be removed aggressively”

What is “speculative text” per the definition we are using? As I am unaware of any such definition, it would appear that “speculative text” is in the eye of the beholder. Is it material that is damaging; material that may be damaging, or simply any material that is unsourced? Brimba (talk) 00:33, 23 May 2008 (UTC)

The phrase comes from a Jimmy Wales quotation elsewhere on the page "…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.…"
The phrasing within the quote makes it clear as to what is being inferred, something that I can not say is clear from “and speculative text that cannot be sourced should be removed aggressively”, nor can I see any reason to have that duplicated in two places; once should be sufficient. Brimba (talk) 02:28, 23 May 2008 (UTC)
Well then, isn't there a more professional way to state this than "some sort of random speculative 'I heard it somewhere' pseudo information"? What EXACTLY is the type of info Jimbo wants removed aggressively? I know we've gone on for years applying the obscenity test (you know it when you see it), but don't you think there's a better way to say it? "Unsourceable low-quality text, speculatively inserted"? Or is it just WP style, a la JzG, for policy to be based on abruptly broad brush strokes which permit wide ambiguity? That might be defensible if you could convince me of it. I agree once is sufficient in the long run, but the repetition was a deliberate temporary measure to induce dialogue on just this question. JJB 13:41, 23 May 2008 (UTC) While we're at it, isn't "Any material lacking a reliable source may be removed" contrary unless changed to "Any contentious material or direct quotations lacking ..."? JJB 13:51, 23 May 2008 (UTC)
I don't think the Jimmy Wales quotation resolves Brimba's "What is “speculative text” per the definition we are using? ... in the eye of the beholder. ...". For example would "The sun rises in the east" be “speculative text” if unsourced? If it would be acceptable, then we have grey areas. If unsourced statments such as "The sun rises in the east" should be removed aggressively, IMO the cure would be worse than the disease. Philcha (talk) 14:02, 23 May 2008 (UTC)
Jimbo didn't mean speculative. In fact, it's not very helpful to figure out exactly what he meant by analyzing the wording too deeply. He was simply saying article content should rest on sources, not knowledge (and hearsay, speculation, rumors, etc) retained in the head of the editor. I'm removing the duplication of this section because it's redundant and misleading out of context. Policy does not emanate in pure form from Jimbo's keyboard. Even those who think it does will admit that it takes some interpretation. Wikidemo (talk) 14:19, 23 May 2008 (UTC)
Wikidemo, if Jimbo's form is impure, then you would favor a better paraphrase? Do you think it should remain ambiguous and diversely interpretable, or can it be stated better (er, improved)? That is, in such a way as I don't have to guess what's in the other editor's head? I don't wish to get into a free-for-all, I wish people to reply to my questions, which then can build a consensus. JJB 14:28, 23 May 2008 (UTC)
I'm not a big fan of inserting Jimbo quotes or other marginalia (e.g. historical accounts of the article's genesis, the scientist or philosopher who first pronounced the principle, etc) on policy pages - they confuse more than they illuminate and ought to be in essays where they can be fully explained. But some people are fond of the Jimboquotes and will revert any attempt to replace them. Jimbo often says things for effect, and purposefully exaggerates things to make a point. That's a fine rhetorical style but it does not translate directly to working rules. I believe his meaning is embodied in the article as is, subject to the ongoing process of refining policies over time. The outcome is that (and I paraphrase here): (1) material that is likely to be challenged should be cited and, if uncited should be tagged or removed if there is a good faith doubt as to its accuracy; (2) controversial or derogatory BLP material should be removed if poorly sourced; (3) fringe theories need not all be mentioned; it is all about due weight, and (4) exceptional claims need stronger sourcing. That's all. We need not be concerned with what "speculative" means or what it means to be "aggressive." No need to manhandle our computer when editing articles. Wikidemo (talk) 15:13, 23 May 2008 (UTC)
Wow, I agree with almost all of that. So you also support removing the Jimboquote entirely and add that it need not even be paraphrased? And speculative and aggressive are unnecessary policy ambiguators, to coin a term? I.e., pseudo information is contentious or contended information? JJB 16:44, 23 May 2008 (UTC)
If you can gain a consensus to do so, I too would be for removing the quote, without making any attempt at paraphrasing it. I think that the Burden of evidence section is sufficiently fleshed out so that the need for the quote no longer exists. That is easy to say, the hard part is likely to be removing it without entering into an edit war, and if it does come down to an edit war, then just keep it. Some things are worth edit warring over; but they are few and far between and the quote is not one of them.
One thing that I would not like to see is for the quote to be removed and some new language formulated to replace it. Brimba (talk) 00:51, 24 May 2008 (UTC)
Remove the Jimbo quote? Me and what magic wand? As an experiment I would like to see how many seconds it takes to get reverted, but it would probably be considered a breaching experiment. No, I think Jimbo's words are here to say. And I think there's a lot of wisdom behind them, and insight about how Wikipedia should work. It's just not something that works if you interpret it literally as you might interpret the more formal parts of policy. Wikidemo (talk) 01:10, 24 May 2008 (UTC)
It lasted some 2800 seconds last time! And there does seem to be developing consensus to delete, in favor of either the remaining text being regarded as accurately paraphrasing everything needed, or adding a bit more accurate paraphrasing. Those who insist on retaining the redundancy don't seem to be demonstrating the same kind of consensus. Here's my summary (which of course doesn't mean that these people are more agreed than favoring deletion to retention), for the reasons in Wikipedia talk:Verifiability/Archive 27#Jim's sentiment and the followup "Jim's sentiment" above: (1) I argue redundancy, and that linking is almost always better than boilerplating. (2) Taken in good faith, an IP editor claims that quoting Jimbo at V has encouraged thoughtless or blind deletion, and believes Jimbo's quote tends against WP:AGF. (3) DGG believes it is overstated. (4) Blueboar believes it is overquoted and misunderstood, but would favor it being paraphrased (ADD: and would retain it as well). (5) Dank55 opens the question of deleting the quote and relying on WP:BLP, implying he favors the move. (6) An essay by the esteemed Raul654 (WP:WHEN) also indicates the quote should be moderated. (7) Brimba would remove without paraphrase. (8) Wikidemo is not a big fan and believes Jimbo's meaning is embodied in the article as is, but doesn't think we have consensus. Well, in good faith, I'm going ahead again. If you disagree, please state below where and how many editors have recently said specifically, yes, retain the Jimboquote both here and at WP:BLP. Thanks! (BTW, and to make clear, my motive for pressing forward is primarily because I think redundancy is a blot on core policy; the original U.S. Constitution, e.g., has hardly any redundant clauses, and that only took a few months of committee work to write! My backup motive is to build my personal credibility as a conscientious policy editor, I'll also grant. But I really think this is a good peg to build my credibility on.) JJB 00:19, 27 May 2008 (UTC)
Removing the quotation is a very poor idea, for several reasons. First, it is not limited to WP:BLP, as is obvious from the quote itself ("This is true of all information, but it is particularly true of negative information about living persons"), and as such logically belongs in both places. Second, it is not redundant, and it is needed, in particular, to add the emphasis that speculative and unsourceable claims "should" be removed. Finally, in spite of your claims of consensus, you've already been reverted and others have already expressed agreement with retaining the quote. Jakew (talk) 00:43, 27 May 2008 (UTC)
Thanks Jake! First, the fact that it applies to "all [speculative] information" does not mean it logically belongs in both; if anything, it would logically belong only here (though I think its thrust is better placed there). Second, I follow Wikidemo in thinking it's redundant with the actual policy text here, such as "All quotations and any material challenged or likely to be challenged should be attributed .... If no reliable, third-party sources can be found for an article topic, Wikipedia should not have an article on it. Any material lacking a reliable source may be removed" (may implies should). Third (three is several?), under WP:BRD and ordinary policy editing procedure, please provide a list of others who would retain, that demonstrates your reversion met some consensus; I did mention the prior reversion, which was perhaps no more consensus-based than yours. I appreciate your help. JJB 01:06, 27 May 2008 (UTC)
In terms of where it logically belongs, you're right: it would belong in this policy, and WP:BLP would then refer to it. However, as a practical matter, it is rather more important to ensure that people actually read it, and on that basis I think it belongs in both policies.
As for the "challenged or likely to be challenged" text, it's related but different. One addresses how to add new material, whereas the other tells us how to handle situations in which this hasn't been done. Both are important aspects of this policy, I think.
Finally, I dislike citing the views of editors from several different past discussions. I think it's a poor way to establish consensus, and is susceptible to taking statements out of context. If a consensus emerges to delete the quote, then I'm sure it will be readily apparent. Jakew (talk) 01:31, 27 May 2008 (UTC)
WP is unable to ensure people read anything; it is unwiki. The question of whether promiscuous redundancy is better than wisely interlinked clarity is a larger question, and while it's clear that redundancy is harmless in mainspace, core policies (like FAs) deserve better.
"May be removed" already tells how to handle situations where adding new material hasn't been [properly] done.
"Several" again, referring to two discussions? The first discussion lapsed because no objection to deletion; I reopened it above (and Brimba opened this section on the same question). At least 8 7 editors favor deletion; the silent majority may not be presumed to favor insertion.
Under WP:BRD, how would you address the stated concerns of all these editors: that redundancy is unbecoming to core policy, that linking should deprecate boilerplating, that the quote is immoderate and ambiguous and favors hyperdeletion, etc.? JJB 14:24, 27 May 2008 (UTC)
"May be removed" is perfectly reasonable with the Jimbo quote, but to remove the Jimbo quote without incorporating its meaning would have the effect of severely weakening this policy with respect to removal.
I think it is simply erroneous to claim that 8 editors favour deletion of the quote, and I note that one of the editors you name has already corrected you on this ("John, you are misinterpreting my views... yes, I do think that the Jimbo qoute is overquoted and frequently misunderstood... but the solution to that is to include a better explanation of what that statement means and how to apply it. I think the quote should be included in this policy. [...] " — Blueboar 19:29, 22 May 2008). I think it's probably fair to say that there are a range of different ideas about the quote, including: i) removing it; ii) incorporating the meaning into the text; iii) leaving it alone; and iv) including it along with a better explanation. That does not equate to a consensus to remove it. Jakew (talk) 14:47, 27 May 2008 (UTC)
I apologize to Blueboar for missing his key sentence, some of the times I read that graf. Refactored, and thank you for catching it. My point for now is that there are at least 7 editors on the record in categories i-ii, and only about 3 in categories iii-iv, which is why I asked for a list of the latter. I also asked for a means to address the concerns stated by the former. Thanks. JJB 14:31, 28 May 2008 (UTC)
I'm afraid that I don't find your categorisation very convincing, John. In some cases an awful lot of interpretation is needed to get a statement to "fit in a box", and it is far from obvious which box people fall into. Although polls are no substitute for discussion, I'd suggest that you ask people to categorise themselves if you think it would help to get a rough measure of where consensus lies.
To address the issues you raise in turn, firstly "redundancy is unbecoming to core policy": I disagree with this - policies are not intended to, and cannot, be read in isolation. They are intrinsically linked, and moderate overlaps between them are not only inevitable but are actually desirable, since this overlap actually helps readers to understand how they fit together. Secondly, "linking should deprecate boilerplating" - not sure about this. I think I need to better understand your reasoning before I can comment. Third, "the quote is immoderate and ambiguous and favors hyperdeletion" - that's a valid point of view (even though I don't share it), but fundamentally, this is an argument to change the semantic meaning of this policy, not to remove redundancy. It echoes the comments from myself and User:High on a tree that the removal of the quote would weaken policy. If there really is a consensus to change (not just modify to remove redundancies) in policy, then so be it, but I'd like to see stronger evidence of consensus before we do so. Jakew (talk) 15:07, 28 May 2008 (UTC)
Great! Then we can agree to disagree on the various POVs and get to the meat. I have no intent to weaken or change the semantic meaning of the policy in this case (when I intend to change meaning, I say so)-- unless removing redundancy can be said to change meaning by reweighting. I have no intent to narrow "all information" to "BLP only" or to narrow "should be removed" to "might get removed": [1]. So if you think the quote is a proper statement of policy, then you would not object to a paraphrase which more professionally and yet faithfully preserves its meaning. Correct? So let's address the concerns of these editors (whatever POV they may hold) by getting that professional paraphrase. Again, I hold out a proposal: "and unsourceable low-quality text, questionably inserted, should be removed unhesitatingly,". JJB 15:29, 28 May 2008 (UTC)
Sorry, JJB, I don't think "unsourceable low-quality text, questionably inserted, should be removed unhesitatingly" would help. That phrasing sets 3 logically criteria for removal: "unsourceable", which itself is vague (e.g. within what expenditure of time and money; and by whom, since another editor might easily find a source, either from his / her bookshelf or because different editors phrase Google search strings differently); "low-quality", which comes perilously close to a personal value judgement by the next editor who reads the text; and I simply don't understand "questionably inserted", unless it's equivalent to some combination of "unsourced" and "low-quality". The phrasing also does not make it clear whether just one or both of the 2 independent criteria are required in order to justify removal.
The problem is that editors will differ on what is uncontroverial. "The sun rises in the East" probably qualifies, but scientific propositions that are not much more complex than that run into objections from adherents of various religious beliefs (e.g. Hubble's law implies the Big bang). The only solution I can see is to say that statements that remain unsourced for e.g. 3 months after being been given a "fact" tag should be removed. It would help if someone could produce a bot that could scan for "fact" tags in articles by category or in a watchlist. Philcha (talk) 12:34, 2 June 2008 (UTC)
You've just demonstrated my point: if my summary is vague, so is Jimbo's. That phrasing sets 3 logical criteria for removal: "unless it can be sourced", which itself is vague (e.g., within what expenditure of time and money; and by whom, since another editor might easily find a source, either from his/her bookshelf or because different editors phrase Google search strings differently); "speculative 'I heard it somewhere' pseudo information", which comes perilously close to a personal value judgement by the next editor who reads the text; and I simply don't understand "random", unless it's equivalent to some combination of the other two.
We can certainly say the criteria must be as interpreted by the deleter, because assuming good faith means the deleter believes they are met, and if (someone else believes) the deleter is wrong that is to be corrected later outside the scope of this clause. But now let's figure out what Jimbo really means. How can we clearly say what type of material, exactly, should be removed unhesistatingly or aggressively, in the potential deleter's good-faith judgment? JJB 15:04, 2 June 2008 (UTC)
This whole debate demonstrates that the quote should not be included. No one can agree what it means. It plainly fails to explain. Either get Wales to draft something clearer or chuck it and get on with writing policy. I don't know whether wikidemo is acting as Metatron, but I agree with the analysis that the quote's context indicates it was not intended to be taken as literally as some here are arguing. This is theology being argued here guys. I respect mystic oracles even less than incarnate gods. Sandpiper (talk) 21:35, 6 June 2008 (UTC)

[edit] Any material?

My question above went unanswered:

While we're at it, isn't "Any material lacking a reliable source may be removed" contrary unless changed to "Any contentious material or direct quotations lacking ..."? JJB 13:51, 23 May 2008 (UTC)

However, my hairline change to "any such material", as one resolution to this apparent contradiction, was reverted. Isn't this policy sentence patently false as stated, and needing some kind of moderation? It is the (contentious) material discussed in context that may be removed, not "any material". JJB 14:30, 27 May 2008 (UTC)

Yes, it is contradictory, and in any case we don't remove such material, at east not without trying to source it. The phrase has been defended best by emphasing the may', and interpreting it as "is liable to be removed by someone" which of course rather destroys the effect and the meaning of it as policy. Yes, there are people who will remove particular parts of unsourced material for whatever reasons, and then it generally gets discussed, sometimes rather vociferously--it doesn't usually stay out unless there is some reason to think it contentious. If anyone started in systematically doing this, I wonder how long it would take till the first hundred people showed up at AN/I to ask for an emergency block. if policy has any relation to reality, this is not policy. DGG (talk) 22:54, 27 May 2008 (UTC)
The point is... any material that lacks a source may indeed be removed. Are there additional steps that in many situations (even in most situations) should probably be taken before actually doing so? Sure... but that does not change the fact that unsourced material may be, and often is, removed. Blueboar (talk) 03:38, 28 May 2008 (UTC)

Thank you, now I understand how the discussion went which I only skimmed over! But given that, (1) wouldn't Blueboar and other supporters be agreeable to an unambiguous text (perhaps "may potentially be removed")? And (2) wouldn't that removal of ambiguity also indicate that (against Wikidemo, alas) the Jimboquote, stating that some text should be removed, has not fully been summarized in policy, is not unambiguous, and should be paraphrased to reflect what a consensus of editors believes policy to be? Simply, "may potentially" does not have the liability of implying "should". (My leaning would also be that if the "may" sentence merely intends to restate the obvious facts of vandalism and open-source risks, it is not really part of the WP:V concept.) Again, while ambiguity of text is useful for compromise elsewhere on WP, it should not be used in core policies; the disagreement hidden by the ambiguity should be aired and resolved unambiguously. (More towards my ultimate point, the additional steps that should probably be taken in most situations, as Blueboar identifies, have not yet been fleshed-out here as well as possible, but I will remake my proposal on that issue later.) JJB 14:22, 28 May 2008 (UTC)

FWIW, I'm fine with "may potentially be removed" which I think is clearer overall, without disturbing the basic meaning or prejudging the points at issue in either direction. DGG (talk) 20:49, 28 May 2008 (UTC)

Thanks! If Blueboar remains silent I'll do that too. JJB 15:26, 29 May 2008 (UTC)

I think the intent here goes beyond "may potentially be removed." The point of Jimbo's quote is that unsourced material should be removed... under certain circumstances. So much depends on the situation... different approaches are needed for different articles and how you deal with unsourced statements often depends on nature of the statement being made. Sometimes the right thing to do is try to source the statement yourself, sometimes it is to flag the statement and discuss the issue on the talk page... but at other times the right thing to do is to bluntly remove the statement. The Jimbo quote allows us to do the latter when needed. Blueboar (talk) 17:56, 29 May 2008 (UTC)

Yes, the intent of this section goes beyond "potentially". That harmonizes with this clause saying "may potentially be removed", and Jimbo's quote (or a paraphrase thereof) providing some form of "should be removed". Could you please comment at "Speculative text" above? I appreciate it. JJB 18:55, 29 May 2008 (UTC)

See also Wikipedia talk:Verifiability/Archive 27#May object?.
I reverted the last change by JJB from "may potentially" back to "may" because the burden of proof lies with the editor who adds or re-adds the material because of lack of a citation, putting in the word potentially may lead some to think that the person doing the deletion has to some get sort of other permission to make the delete. --Philip Baird Shearer (talk) 15:27, 30 May 2008 (UTC)
I'm not sure about burden or permission, but the person doing the deletion ought to have some good faith concern that the information is not reliable, rather than wikigaming content the person disapproves of or pointy deletion over the simple fact it is uncited. "May" sounds a little permissive, if not attached to a standard or couched as advice to the person inserting the content rather than the one potentially deleting it. Wikidemo (talk) 17:42, 30 May 2008 (UTC)

Thank you, Philip. Wikidemo says it well. I discover that WP's structure already provides that almost any editor can, may, and might remove almost any material in good faith, not just "material lacking a reliable source". This clause has no way to broaden the extant permission to edit boldly (WP:IAR, WP:BOLD); it also really doesn't address the separate Jimboquote duty to remove "pseudo information". This clause largely describes an obvious "is" and fails to provide a real "should", as Jimbo does. Maybe the whole thing can be scrapped and the paragraph begin "Editors may object"; or it can be changed to a "should" as in "Be bold when updating, and demonstrate good faith, but"; and/or "Unsourceable low-quality text, questionably inserted, should be removed unhesitatingly, but". JJB 18:28, 30 May 2008 (UTC)

One person's "pseudo information" is another person's obvious fact, if it is obvious then they can either restore the information with a citation or they may complain about the deletion on the talk page and be directed to WP:PROVEIT. I would not object too strongly to the first "may" in the sentence being changed to "can", but last time I looked at the OED it seemed to me that "may" is as good or a more suitable word in this context. --Philip Baird Shearer (talk) 11:20, 31 May 2008 (UTC)

Perhaps I'm being too subtle again Philip. The reason people are objecting to you is that both "can" and "may" are ambiguous and that some of the meanings permitted by the ambiguity are misleading or not policy. By that argument neither is more suitable and some other wording should be favored. And I wasn't talking about whether another person finds material to be an obvious fact, I was saying that the clause itself is obvious, because WP's structure is such that "any material may, can, might be removed" already. That means our emphasis on "lacking a reliable source" is superfluous. Policy has two ways of guiding, known as "is" and "should". My point is that this clause has no "is" value because the fact that anyone can delete you is reminded to you every time you edit; and the clause attempts to have "should" value as well, but fails because it is a milquetoast version of Jimbo's stronger statement. Therefore the clause can be deleted.

From another angle, we should look at the related clause in graf 1 (second sentence of policy), which I've failed to do. There a different, clear "should" appears so you can see both aspects better. All material IS always at risk of removal, so you SHOULD cite it if controversial. The "should" is then developed well in the first three grafs of "Burden". Then the "is" takes over, but less thoughtfully, and relying on the overworked word "may". How about "material lacking a reliable source is at high risk for removal"? I don't mind that it takes lots of drafts to succeed in communicating, but I do get impatient also. JJB 13:03, 31 May 2008 (UTC)

JJB you use "should" a lot and like "may" it has more than one meaning, do you mean "should" as in "ought to"? --Philip Baird Shearer (talk) 10:57, 2 June 2008 (UTC)

Umm, yes, why do you ask? JJB 14:55, 2 June 2008 (UTC)

[edit] Better section names

At Wikipedia talk:Verifiability/Archive 28#Shortcuts I said:

I think the section titles to which two shortcuts point should be shortened losslessly to "2.3 Self-published sources" and "2.4 Articles about self-published and questionable sources", and the "SELF" anchor deleted as orphaned. I will wait to see if there is any objection. JJB 16:24, 8 May 2008 (UTC)

This seemed not to raise any eyebrows in itself, though one or two other shortcomings with my related shortcut proposal were raised and resolved. My later research shows that "SELF" is not orphaned as I thought, but has some 100-200 links, so I have restored it. I also discovered several former section names which have sometimes hundreds of links but which had reverted to the top of the article due to section name change. This indicates it should become a point of honor, when changing a section name in a policy, either to ensure the old section name is orphaned in all namespaces, or to add a "span" identifier which preserves the old name as an anchor. I added four of these; there are more. Anyway, all this indicates that we are now ready to proceed with improving these section names, primarily because undue length both clutters the contents box (also creating overlap in some browsers) and looks less professional. I am now favoring "2.3 Self-published sources" and "2.4 Articles about otherwise unacceptable sources", because that is how they are described in the article. The in-line text (and thus the policy) will not be changed, indicating that "otherwise unacceptable" is shorthand for "either questionable or self-published" (also dodging the and/or issue). When these are changed the old names will be again added as spans. JJB 15:26, 29 May 2008 (UTC)

Crum, since you favored just "Self-published and questionable sources", can I get you to comment here on how that compares to the above, or to just "Use of otherwise unacceptable sources", or perhaps "Exception for ..."? I'd like to have better consensus on a single shorter name first, due to not wanting to encourage excess links that would need handling. JJB 20:46, 30 May 2008 (UTC)
I found the titles "Self-published sources" and "Self-published and questionable sources in articles about themselves" better. When I read the idea of "Articles about otherwise unacceptable sources" I had to read it three times and I still don't think it's very clear. After the text change which happened, there sounded to be a contradiction because the section "Self-published sources" says there are some circumstancies where self-published sources may be used (other than on pages about themselves), and then the section "Self-published and questionable sources in articles about themselves" said that they can be used only in articles about themselves. Best regards Rhanyeia♥♫ 07:56, 31 May 2008 (UTC)
Thanks, that was actually a good catch: SPS may be used in articles about themselves and sometimes if a RS regards them as expert. Maybe we need a different tack on the title, like "Self-coverage exception" (that's really what it is!), or else something like "Use of less reliable [or acceptable] sources". A section name need not be too specific and should not be too contentious; honestly, "Self-coverage exception" might fill the bill. JJB 13:12, 31 May 2008 (UTC)

[edit] More on SPS

As already pointed out, I initially liked Crum375's change but then realized Rhanyeia's countervailing logic was correct. With Crum's latest edits there are (at least) three questions outstanding:

  1. What should the title of this section be? No title has supplanted the prior consensus on the overlong title. My suggestions are above, Crum's are in edit history.
  2. In addition to the limited use in WP:SELFPUB for articles about the source, are SPS also usable when a recognized expert in the field as described in WP:SPS? I think obviously so, but Crum hasn't yet abandoned the position that the first use is the only use. Example: Suppose Stephen Hawking says something contentious but sincere about time travel on a blog, but it appears in no scientific journal or news report. Hawking is an established expert, he has reputation for fact-checking and accuracy, but he's not third-party. Doesn't that make him excluded under WP:SELFPUB because contentious, yet includable under WP:SPS due to expertise?
  3. In addition to the limited use in WP:SELFPUB for articles about the source, are SPS also usable when speaking about themselves in any other article, assuming proper weight? I that that would be too loose a rule for Crum to advocate, and two IPs were just reverted trying out another version of it. But I would also hold for a hairsbreadth of expansion, because Crum's concern goes a little ways. Currently the SPS is usable for articles about the author, the book, or the organization (three defs of source), but I think it ought also be usable for articles about the POV or movement or spouse or invention, etc., namely, very closely allied topics. An article about any minority philosophy would be poor if it excluded all SPS's simply because the philosophy is not the source. The SPS ought to be permitted to speak a bit more than about self. That would suggest the wording be "in articles about themselves or very closely allied topics", presuming that the community can tell the difference between closely allied topics (often minority POV articles) and loosely related topics (often overviews of all POVs). JJB 18:04, 2 June 2008 (UTC)

Discussion continues at points 5, 8, and 9 of #Recent changes. JJB 20:37, 3 June 2008 (UTC)

[edit] "Verifiability INSTEAD OF truth" or "verifiability IN ADDITION TO truth"?

I have run into an editor who appears to believe that the WP philosophy of "verifiability, not truth" means that anything that can be attributed to somebody else is sufficient for including it on a page. That is, she seems to believe that "verifiability" is a lower bar than is "truth."

By contrast, I interpret "verifiability, not truth" to mean that for a claim to be included, it must be both true and verifiable. That is, WP has a higher bar than mere truth.

Any input would be appreciated.

—MarionTheLibrarian (talk) 01:38, 30 May 2008 (UTC)

I'd say that it depends on the subject of the page, the subject of the quote, and the notable expertise of the attributee in relation to both. As a top-of-the-head example, this page on Shirley MacLaine's official website says, that astrology "... can provide a road map that will permit us to avoid obstacles, enhance our natural abilities, help us chart the course of our life and reach a deeper, more thorough understanding of ourselves, our past and our future." That page can be cited as a supporting source for an assertion that Shirley MacLaine has expressed such beliefs, perhaps in the WP Shirley MacLaine article. That page might possibly be used as a supporting source regarding e.g., notable astrology-believers in the WP Astrology article. That page probably cannot be used as a supporting source regarding assertions regarding how planetary positions influence our daily lives in the WP Solar System article.
We need to be very careful with diddling core policies, but perhaps this policy ought to say "One threshold for inclusion ..." rather than "The threshold ...". It could be argued that the implies the sole—that once the inclusion threshold of verifiability is reached (e.g., it is verifiable that Shirley has expressed ...) that a statement about Shirley having expressed those beliefs can be included in any WP article. -- Boracay Bill (talk) 04:18, 30 May 2008 (UTC)
  • Permit my pontification? Actually, it says "the threshold for inclusion in Wikipedia", not in any given article. If it's reliably attributed, and text attribution is given for controversies, then it's suitable for inclusion where WP:DUE, but maybe not in the article the person is pressing for.
  • Marion's wording change was discussed last month and the present compromise text has been successful instead. IMHO, the real "higher bar" is "not just verifiability, but also what we believe to be true", and consensus favors only a hint of the second part of that: it actually appears as WP:DGF, "demonstrate good faith". Each part is necessary, not sufficient, which is already covered by the presence of the different policies and guidelines.
  • If "MacLaine says Controversial X", that may be included appropriately. "Controversial X<ref>MacLaine, p. 1.</ref>" may not be included except subject to WP:SELFPUB.
  • "One threshold" can wrongly imply "this threshold OR that" instead of what we mean, "this threshold AND that". True, the presence of other simultaneous thresholds is described by other policies like WP:BLP. That may permit a different modifier like "The quality threshold", e.g. But I don't think the potential implication of "the threshold for inclusion" is so problematic as to fight over.
  • As to objective truth, WP:OR also applies. Say a reliable source says something is 500, but a little OR suggests to me, and most folks I discuss my OR with, that it really is 50. A reliable-source correction may not be published for years. In this case someone unknowingly inserting 500 into WP does properly meet the bar here: it was verifiable and believed to be true, though it may not have been objectively true. The proper solution for me, the doubter, is not for WP to state what I believe to be true here (OR), but simply to delete the potentially erratic statement from mainspace and retain the discussion at talk. This (truly) happened to me. It turned out that 3 or 4 reliable sources stated that the appropriate measurement should be 50 and thus indicated the first source's statement as error, but if they hadn't, deletion would be the proper outcome, assuming that consensus validated the doubts arising from OR. This protects the editors' good faith when they follow sources accurately but mistakenly. Making "objective truth" the actual bar for entry would technically be too high, because no one on earth can know "objective truth", only what they believe to be true. (Other philosophies of truth differ.) JJB 14:35, 30 May 2008 (UTC)

Thanks for the points thus far.
Perhaps the specifics might also be helpful: The editor I have in mind included on a list of external links a link to a lawyer who presented that lawyer's argument to a judge. (The judge's decision on the lawyer's argument was not included in the link.) When I said that those were not sufficient, the editor said, "Wikipedia's policy is verifiability, not truth. As such, a legal memorandum presented in court regarding the scientific unreliability of PPG is both verifiable and a reliable source." This struck me as an improper interpretation of the WP policy. That is, if 'I found somebody who said so' were sufficient, then WP would be just a big rumor mill.
Am I off base?
—MarionTheLibrarian (talk) 15:23, 30 May 2008 (UTC)

It's verifiable all right! But notice how there is nothing to prove the minor premise of reliability? A lawyer's memo in court is easily not WP:RELIABLE unless the lawyer has a good reputation for fact-checking and accuracy (insert lawyer joke here). If in doubt it could be ruled on at WP:RSN. If it's found suitable as an accurate WP:PSTS for scientific claims, you must still reach WP:CONSENSUS as to WP:UNDUE weight. If there is consensus to include, assuming we don't know the judge's decision, you would still need a disclaimer at least as long as "In a legal argument in X Court, attorney Y stated:". And if you know the decision, state it as overruling. Plenty to discuss without WP:V being the policy misinterpreted. FWIW, I am declining to find out what PPG is. JJB 16:11, 30 May 2008 (UTC)
Marion, short answer for you: the editors simply forgot to establish Wikipedia:Neutral point of view, didn't they? --Kubanczyk (talk) 22:56, 2 June 2008 (UTC)
I suspect the editor bothering MarionTheLibrarian will want the answer to be longer than Kubanczyk's sensible, direct response.
Is there any way to make sure JJB's legal pyrotechnics remain visible when this page is archived?
Re PPG, after much Googling I concluded it was more likely to mean "Pollution Prevention Guidelines" rather than Babylon 5's "Phased Plasma Gun". Any other guesses, anyone? Philcha (talk) 23:09, 2 June 2008 (UTC)
Thanks Philcha! The policy changes might be: (1) change "the threshold for inclusion" to "the first threshold for inclusion", which resolves the other concerns about the reality being plural thresholds; (2) Add to WP:RS something like, "When a consensus of editors believes a reliable source to have made a simple error, but the error cannot be demonstrated from another reliable source, the solution of completely omitting the claim may be considered." (probably needs strengthening); (3) Add to WP:DGF something like, "For instance, do not add material you do not believe to be true;" and other examples of good faith. I believe the rest is obvious from current policy. Comments? JJB 13:47, 3 June 2008 (UTC)

Discussion continues at #Recent changes point 11, as well as at WT:RS#Suspected source errors. JJB 14:30, 4 June 2008 (UTC)

I don't think it is the place of a neutral editor to decide what is and what is not truth. Truth is a subjective concept and cannot be demonstrated objectively and thus is incompatible with our goal of neutrality. That is why we use verifiability instead, not in addition to truth.
Just because something can be attributed to somebody else does not make it sufficient for including on a page. We have WP:RS, and WP:UNDUE to address these very concerns. 1 != 2 14:34, 4 June 2008 (UTC)

[edit] Attribution again

Brimba reverted my adding two words as "readers are able via attribution to check"; the edit summary was Rm, Policy is every source must be attributable; not that every source be directly attributed. “Although everything in Wikipedia must be attributable, in practice not all material is attributed.". I agree with the summary except for two problems: it quotes WP:A which is not policy (yet), and I don't think it supports "rm". But since my goal is to build resolution of the WP:A dilemma, let's do it.

  • Since you agree that sentence represents best practice, let's add it right here right now. I'd love to.
  • "Everything must be attributable" leads to "readers must be able to attribute anything" which leads to "policy is that readers are able to check anything via attribution". "Able via attribution" is precisely consonant with "everything must be attributable, not everything is attributed".
  • If you think my phrasing permits the impression that readers must be able to check via someone else's attribution, i.e., that everything must be attributed, then what properly describes the role of attribution? I have no problem with the bolder "whether readers are able to attribute material added to Wikipedia as already published by a reliable source"; how's that? There "attribute [it] as" is a synonym of "check that [it] has been". Thank you. JJB 03:32, 31 May 2008 (UTC)
Crum? JJB 04:18, 2 June 2008 (UTC)
I modified slightly to make clear this only applies to factual claims, and only article space. Obviously, attribution does not apply to images, talk pages, templates (except to the extent that, by transclusion, they introduce a factual claim), opinions (which don't belong, verifiable or not), word choice (as opposed to factual claims) in headings, phraseology, punctuation, etc., except to the extent they make factual claims. Wikidemo (talk) 05:45, 2 June 2008 (UTC)
Based on this reversion[2] I've reverted farther back to a stable version.[3] We obviously don't have a consensus for the latest additions, so we should talk about this further before making changes to a fundamental policy page. The other change I had made, that was rolled back, was adding a statement about "quality" which was wikilinked to an inapt article about data quality that has more to do with signal theory than encyclopedias. We don't have low, medium, and high signal strength in Wikipedia, with an arbitrary cutoff - we have a bright line rule that verifiable information is included and non-verifiable informaqtion is not. Wikidemo (talk) 05:58, 2 June 2008 (UTC)
We're editing in different sections. Give me a moment while you read the below section. JJB 06:18, 2 June 2008 (UTC)
First, remember the attribution clause is long-standing text from WP:NOR. I agree it need apply here only to article space (WP:BLP is an exception, but handled separately). I'm really not too opposed to "factual claims" for "material"-- but anything which an image intends to communicate must be attributable; word choice and such must be supportable by reference to attributable sources; and any opinion (of which there are many) must be attributed (making it a verifiable and true "Y says Z" statement even if "Z" is false). You could well argue I don't need to attribute templates like "citation needed" by telling who opines that citation is needed; but I don't think template would be understood to be material in this sentence. Given that all that may be characterized as factual claims, I don't think "factual claims" is too much of a problem, except that you need to verify more consensus for it than me.
I'm not going to speculate about "stable version" or "don't have a consensus" other than to remind you that I have always cited the consensus (or occasionally the common sense) of my edits, unless I am specifically putting them out speculatively for comment and potential reversion.
Whether info is verifiable or attributable is a bright line. The quality of the info isn't. Quality is determined by whether info is reliable or acceptable, which as I demonstrated below admit of many degrees of difference between, say, "most reliable" and "largely not acceptable". Again, SlimVirgin and others have emphasized the information quality link, and I don't know what article you were reading about signal theory, because that link seems right-on for detailing a key definition in this policy.
So how about we compromise by my restoring one of your two edits in full (though there may be potential disagreement from others), and we continue to discuss the other one? JJB 06:35, 2 June 2008 (UTC)
Many factual claims about images are not supposed to be attributed. In an article about dress, if you have an image of a girl in a red dress, you can say "photo of a girl wearing a red dress" and none of it needs to be verifiable - or could possibly be verifiable, given the non-free policy. We only use free (generally meaning unpublished) images. Could we continue the discussion in one place, below?
Actually, such a caption is verfiable... by the photo itself. On the other hand, if the caption were to say "photo of a girl wearing a red Versachi dress" we might need a citation to the fact that the dress is indeed by Versachi. Blueboar (talk) 14:54, 2 June 2008 (UTC)
I agree, verifiable...but not in the policy sense because the photo itself is not a third party independent published reliable source. It's verifiable by direct inspection, i.e. original research. Hence my comment that we allow original research in photo captions. You could describe this in a few ways. I don't think there's any way you can get a citation that a particular dress in a photo is a red versace because if it's a free photo there is no published version of it, hence no reliable source. You would just hash it out in the talk page if anyone disagrees, reach a consensus, and move on. Or accept the uploader / image inserter's claim on good faith that they took a picture of something known to them to be a versace. This exact question comes up every once in a while in various places. Wikidemo (talk) 19:07, 2 June 2008 (UTC)
Red dress - This also depends on the actual article in which the photo is used. In an article on 'Tourism' a photo of a tourist would be captioned, "Woman in a red dress at Niagra Falls" — that the dress is by versace, even if true, is pretty much irrelevant. In an article on the 'Academy Awards', where apparently who designed whom's dress is considered important, the caption would read "Jane Doe wearing a fabulous red versace gown' which would easily be supported by a reliable reference to newspaper reports etc. --
If some reader wishes to dispute the evidence of their own eyes, that the dress is 'red', or dispute that 'dress' and 'gown' are equivalent in this context then that is discussed in the normal manner; one or two editors venture their opinions on the talk page and a 'consensus version' is implemented. (I don't think it would take long to sort out, lol.) --NewbyG (talk) 23:50, 2 June 2008 (UTC)
Restricting it to article removes "portals" which is not a good idea. I think it is better not to name the space(s). --Philip Baird Shearer (talk) 07:41, 5 June 2008 (UTC)

[edit] Reasonable TIme

What is considered a reasonable time from when a FACT or CITE tag is added to the time the material can be removed from the article? Paulrach (talk) 09:26, 31 May 2008 (UTC)

This is left to editorial judgment... we do not set a time frame because so much depends on the nature of the uncited meterial and the article it is in. Some statements should be removed immediately (for example, see: WP:BLP) and not even tagged. Others can be left with a tag indefinitely. But if you feel that the material in question is objectionable, and should be removed, then I would suggest waiting at least a month. I assume you have explained what the problem is on the talk page? Blueboar (talk) 12:35, 31 May 2008 (UTC)

[edit] Non-English sources

The text of WP:RSUE currently says: "Because this is the English Wikipedia, for the convenience of our readers, editors should use English-language sources in preference to sources in other languages, assuming the availability of an English-language source of equal quality".

I find this wording unclear and potentially harmfull to the universal character of WP. There are probably a lot of subjects on, for instance, American or English historians disagree with French, German or Italian historians who were not necessarilly translated. And as the sources are of equal quality, French, German and Italian historians will then not be used as per WP:RSUE.

The wording should be clarified I think, for instance by something as follows: "Because this is the English Wikipedia, for the convenience of our readers, editors should use English-language sources in preference to sources in other languages, assuming the availability of an English-language source of equal quality providing the same information" (or something to the same effect). Bradipus (talk) 09:46, 1 June 2008 (UTC)

[edit] Which material, factual claims

I undid two of Wikidemo's suggestions on the same grounds: in both cases text that has long standing as policy was replaced with text which has not been vetted for consensus. The edits may well lead to improvements after discussion, but let's discuss them. Though Wikidemo did edit text I had just worked on, I am considering Wikidemo as the "bold" and myself as the "revert" because Wikidemo's text was wholly new, while mine was previously reviewed at other pages, and Wikidemo did not use "undo" while I did.

  • Wikidemo did not like either the canonical "the type and quality of material" acceptable (WP:NPOV, WP:NOR), or my recent version (based on SlimVirgin) which includes information quality (enjoying a successful airing at WP:A and WP:NPOV, V and OR), and Wikidemo favors instead "which material" is acceptable. I don't believe that concern will be sustained. The first is well-respected, clearer, and tighter, and the second is being supported because it reinforces the concept that reliability and acceptability are spectra and not booleans. We don't use "unreliable" and "unacceptable", but we use "largely", "in some circumstances", "high-quality", "more reliable", "most reliable".
  • Wikidemo also changed the statement from WP:NOR, "All material in Wikipedia must be attributable to a reliable, published source", to "All factual claims in Wikipedia articles must be attributable to reliable, published sources". I don't mind the narrowing of Wikipedia to Wikipedia articles, or the pluralizing of source, as they are logically defensible; but I don't believe narrowing "material" to factual claims is useful. The topic is, simply, all material in Wikipedia articles. Subheads, images, bibliographies, all characters must be attributable (and not necessarily attributed).

If Wikidemo or others agree, I have nothing against a reedit on those lines. JJB 06:05, 2 June 2008 (UTC)

  • I did ask Wikidemo here in edit summary, but Wikidemo has instead now also deleted the full statement from WP:NOR and WP:A mentioned in the second point above, and Shoemaker's Holiday has jumped in too with a different edit. As I already mentioned, the first clause has clear consensus at WP:NOR, and the second clause is an obvious statement, regularly discussed on this talk, that not all material is attributed in practice. I don't understand Wikidemo's concern with why that statement should be deleted, so will await instruction. JJB 06:12, 2 June 2008 (UTC)
  • (EC) I reverted back before seeing any question, following normal BRD procedure. The version I reverted back to appears to be the stable version, before of the recent changes were made for which I'm asking for consensus. Its language, though not perfect, is identical to WP:NOR so it is at least consistent, whereas the version proposed, "type of material and information quality level of material" is not the exact same as other policy pages. Shoemaker's Holiday simply re-added what looks like a minor improvement[4] that got deleted when I made the reversion - no problem there.
  • Regarding the "the type and quality of material that is acceptable in articles" language from NOR, it is fine and I think we all agree on the substance of what this means. However, as a matter of wording we have to be careful, especially when making changes, to make clear that we are referring to the type and quality of source material, not the type and quality of material that we actually put in the article. There is indeed a spectrum of reliability among sources, and of quality of information within any given source. However, we don't distinguish between quality levels once the material is in a Wikipedia article. That has one standard, verifiability, and it is the decision of whether or not something is verifiable that rests on a sliding scale of quality of source materia. Saying "...and quality level of material" tends to confuse things, by making it more likely people will mistakenly think WP:V references varying levels of material within articles.
  • The wikilink and phrasing, information quality, isn't direclty related to encyclopedias. The link is to an article that in the lede says it is about Information systems. We ought to be careful when we link words used in policies to articles, because that seems to endorse it as a specially defined term. That is a short, start-class article, and whatever we mean by quality of information it is not exactly what is in that article. The analogy between knowledge and computers / signal theory is useful sometimes but we shouldn't take it too far or stretch the metaphor. If it has recently been added elsewhere it should be changed back there too. However, one of those pages is fully protected and the other is a mere proposal.
  • The phrase "all material in Wikipedia must be attributable to a reliable, published source" is not strictly true, as I explained. It doesn't apply to things that are not facts, and it does not apply outside of article space. It's not part of NOR either - it's part of the nutshell, which is a simplification. That's fine for nutshells, but if we're going to add it here we should state the full version. This is a real issue because people frequently make the mistake of claiming that arguments made in AfDs or talk pages are invalid unless verifiable, that images have to be verifiable, that image captions need published sources, etc.
- Wikidemo (talk) 06:30, 2 June 2008 (UTC)
Thanks for your consideration. Recall that IS is not IT; IS includes the cognitive sciences. My thinking was "consensus" = "type and quality of material" = "type of material and quality of information" = "... and information quality level". How about "type of material and quality of information"? Then it doesn't look like an endorsed term, but it still lets people know that we are in fact an information system that has, dare I say, reputable standards? Let me also look in the archives. I take it you're OK with my restoring your other edit (which, really, is just as new and presumably commonsense as my reintroduction of IQ). JJB 06:51, 2 June 2008 (UTC) Second, this clause applies to all 3 policies jointly, and neutrality is certainly a spectrum, and nonoriginality has some gray areas too. And per the other discussion here, "demonstrably findable" is also a spectrum. So the argument from verify/attribute being a boolean process does not seem to require that we "make clear that we are [not] referring to ... the type and quality of material that we actually put in the article" or that we are "making it more likely people will mistakenly think WP:V references varying levels of material within articles". JJB 07:02, 2 June 2008 (UTC) Third, in actual practice, using a photo of a dress to illustrate an article on dresses does indeed intend to claim factually that the picture accurately depicts something in the class of dresses, and if that should be disputed (perhaps someone thinks it's a gown and not a dress), you'd ultimately need to source the challenged statement that objects such as that depicted are reliably defined as dresses. This would not necessarily happen in this case, but "dress" is not a boolean either, and in more complicated cases such disputes certainly happen. JJB 07:06, 2 June 2008 (UTC) Anyway, I'm ready to restore this version, so long as we understand it's an agreed temporary compromise and not a revert of anyone, and that the question of IQ is open. JJB 07:13, 2 June 2008 (UTC)
I'm fine with your proposal so the following is discussion, not consensus-building. Looking at the information quality article and those linked to it, one definitely gets a lot more about data quality than knowledge quality. So I'm not sure what linking to a start-quality article does as far as clarifying policy. We should make sure we're clear that the goal is high quality content and that we're not importing a different concept. Could we do a footnote "see also" kind of link? The second point is an explanatory issue and I'm not sure of the effect. Maybe someone can just go ahead and improve that article. OR, verifiability, and NPOV together don't combine to produce an emergent trait called information quality do they? It's just OR, verifiability, and NPOV. I do disagree on #3, images. Images verify themselves. If there's an image with a caption that it is Sigmund Freud, and people disagree as to whether the image truly is one of Sigmund Freud, they have a discussion in the talk space somewhere based on a wholly different set of considerations than verifiability. I think you've teased out of the dress example a possible factual claim embedded in calling a thing a dress, but the question of exactly what it is (as opposed to whether that thing qualifies as a dress) is sourced from the image itself, not a reliable third party published sources. Wikidemo (talk) 07:24, 2 June 2008 (UTC)
OK thanks for a BRD success! Archives reveal links which soundly place IQ as a WP discipline, e.g. Wikipedia:Wikipedia in academic studies, Wikipedia:Wikipedia Signpost/2005-08-01/Featured content; here is where SlimVirgin makes the direct application. The three policies do severally and jointly contribute to some emergent trait which many have supported being described as "quality of material" (information), which (to me) is equivalent enough to information quality to use the latter because of its linkability and because of the other links in support for it. Yes improvement of the IQ article is desired.
The image issue is one from a deep philosophy, though it takes a bit to wrap your mind around. A picture of a red dress makes no claims in itself about it being red or a dress, and a picture of Freud makes no claims about it being Freud. When placed in the context of an article they do make those claims. And the discussion does then turn on the factual claim. A photo and a sentence are both data; the data's uploader must demonstrate that it accurately matches a reliable source. Some reliable source had better verify that that is a "dress" and this is Freud (hmmm), just like a reliable source is (theoretically) needed for attributing sentences like "dresses are clothing" and "Freud was bearded". And if it turns out that the uploader heard some rumor that this free image was really a dress and that free image was Freud, why, it's failed verifiability, whether or not the rumor is true. Anyway, "factual claims" subsumes it all.
Let's put the question up for wider discussion as "Is it appropriate for this page to link to the topic of information quality, from the lead?" Anyone? JJB 07:53, 2 June 2008 (UTC)
It's not a very good link here, it's not about the same thing. I think "type and quality of material" was ok, but "type of material and quality of information" could be ok too. Best regards Rhanyeia♥♫ 17:13, 2 June 2008 (UTC)

Discussion continues at point 3 of #Recent changes. JJB 20:41, 3 June 2008 (UTC)

[edit] Carrying this to the point of absurdity

Recently, I've been running into people who seem to be carrying the requirement for third-party sources to what seems to me to be the point of absurdity. For example, it seems to me that it should be entirely legitimate to cite a newspaper's masthead for the names of its editors, or the existence of an article with a certain byline for the fact that a particular person has written for that paper. It is cumbersome to have to find an independent source for this sort of thing, and extremely rare that the third-party source would be more reliable. Yes, there are exceptions (like Private Eye or another satirical publication running an obviously false byline, or the need to cite a third party for "Boz" having been the same person as Charles Dickens), but the normal case is that these things can be taken at face value. For example, I should not need to go anywhere other than the masthead (or the online masthead) of the Village Voice to assert that Tony Ortega is its current editor. - Jmabel | Talk 17:25, 2 June 2008 (UTC)

I completely agree. I have run across an editor who is insisting that every sentence of many spirtuality organization articles be referenced by a completely neutral 3rd party academic source, and fights against citing those organizations' own webpages regarding the history of the organization, what they believe in, etc. What can be done in these situations? Madman (talk) 16:12, 3 June 2008 (UTC)
No, Madman. You restored a wholly uncited article from redirection, threw in four citations, three of them bad (one to the wrong URL, one that would only verify it with some heavy WP:SYNTH thrown in -- leaving me to correct the former and rewrite the latter, and one was sufficiently vague that it let you also throw in a piece of false information without realising it). Thereafter, you've added one source, I've added two and the article is still half unsourced. So what do you do? Edit war to keep the unsourced half in there and bitch at me for not adding more sources (which per WP:V is your "burden" for restoring the material). I did not "fight[] against citing those organizations' own webpages regarding the history of the organization" -- I fought against your slap-happiness in not citing the correct URL in one instance & not rewriting to match the cited source in the other. "What can be done in these situations?" HrafnTalkStalk 17:26, 3 June 2008 (UTC)
Guys, I'd appreciate it if you'd skim pages about how this is not the page for edit war discussion, how there are no personal attacks, and how disputes are resolved. This policy is very clear about when self-published sources are permitted. Thank you for your forbearance. JJB 17:44, 3 June 2008 (UTC)
Ignoring Madman & Hrafn's remarks, but replying to JJB: I don't think it is "very clear". The section starts "Anyone can create a website or pay to have a book published" and then goes on to talk about "experts", but it really says little or nothing about citing institutions when they write about themselves. It seems to me that, for example, barring unusual circumstances, a corporation should be the preferred source for information about its own current officers. Anyone else probably got their information from that same source; it is quite possible that they garbled it or aren't up to date. If there is a source that says explicitly that the information from the organization is misleading, outdated, etc., then we may have good reason to look elsewhere, but (for example) it seems to me the height of foolishness not to go to a newspaper's masthead to find out the names of its editors, or an NGO's web site to find out the names of its officers and board members, or a list of people that NGO may have honored with awards in the past. - Jmabel | Talk 19:54, 3 June 2008 (UTC)
Actually it's "very clear" for their purposes (an organization may be cited about its history or beliefs subject to clear WP:SELFPUB limits) though some details irrelevant to their concerns are not yet very clear because we're still hammering them out. Please comment at 5, 8, and 9 below in #Recent changes. JJB 20:08, 3 June 2008 (UTC)

[edit] Guideline stuff

I have moved this text from this edit to here for further discussion:

ISBN or OCLC indexing, or notability of source). If a reliable source is not readily available (e.g., freely accessible online), the editor inserting or restoring the challenged material should be able to provide page numbers and/or direct quotes from the original text supporting the material, upon request. Direct quotes may be posted on the article's talk page for evaluation—they need not be added to the article.

I think this type of detail ought to be in a guideline. Not in the policy. --Philip Baird Shearer (talk) 18:04, 2 June 2008 (UTC)

  • An ISBN number is not a substitute for notability -- this is a major change.
  • If someone enters an edit and references the information, there is no reason why, a year or so down the line that, the editor who added the citation, will still have access to the original text. To date we have only insisted that the person provides enough information in a citation that a reader reader can check the citation. This is another major change.
  • As for "Direct quotes may be posted on the article's talk page for evaluation" what about copyright issues? --Philip Baird Shearer (talk) 18:17, 2 June 2008 (UTC)
This came from consensus in archive 27 or 28 that "demonstrably findable" was good but needed expansion. I added the parenthetical and I think Crum did the rest, which I and Martinphi (at my talk) immediately applauded. Your first point seems to mistake topic notability for source notability: if source is notable, it's verifiable; if source has an ISBN, it's also verifiable (whether source is notable is then irrelevant). Secondly, regardless of who first cited the material, either no one argues against the deletion, or someone argues against it and thereby takes up the burden of dealing with its inaccessability (the "restoring" editor): if it's inaccessible to all current editors, that's a new consensus of unverifiability. Thirdly, why would copyright issues apply to three sentences quoted in talk if they don't apply to one sentence quoted in the article, as routinely? People transcribe long stuff into talk all the time. This whole piece is a natural description of how to meet the burden of evidence and, while it may easily be tweaked, the archive suggests it is a needed outgrowth. JJB 18:36, 2 June 2008 (UTC)
Not all notable sources have an ISBN, not all publications with an ISBN are notable. It is notable not ISBN that make it a reliable source. As I said if this point needs expanding then do it in a guideline.
"If it's inaccessible to all current editors, that's a new consensus of unverifiability" -- is clearly not true. If the current editors do not have access to the library where the information comes from it does not mean that the information is unverifiable. For example suppose that a there is a fully cited reference for an edition of a book published in the UK but the current editors are American and they do not have access to that UK edition. I have worked on a page were it was expanded using a Romanian edition of I book I have, I believe (as I must) that the person who added the citation did so in good faith. I have no right to delete his additions just because his page numbers and the ones in my book do not correspond. I have started to alter the pages to an English language edition, but it is not a high priority of mine. But the by the implications of what you are suggesting this sentence could be used to removed all good faith editions simply because an editor does not have access to a specific edition of a book.
"Thirdly, why would copyright issues apply to three sentences quoted in talk" Because the addition is not talking about quotes in the article, but adding possibly very large sections of a book to the talk page to provide proof of a point.
"This whole piece is a natural description of how to meet the burden of evidence" Then it should be in a guide line not cluttering up a policy which should be a succinct as possible. --Philip Baird Shearer (talk) 08:55, 3 June 2008 (UTC)
Excuse me, PBS, I think you will be open to the concept that "source notability" is not what makes something a reliable source. Reliability requires at least third-party, published, well-reputed, and findable; this sentence is not discussing "reliable" but "demonstrably findable". "Demonstrably findable" is established by criteria like weblink, library or OCLC existence, ISBN, or proof that source is independently notable. (That is, source notability is one way of demonstrating findability, which is a sufficient but not necessary criterion for reliability.) If the statement could be misleading enough to make it appear that I'm saying ISBN equals notable or ISBN equals reliabile, then rephrasing and/or moving it elsewhere might well be indicated, perhaps even to WP:CITE.
For "inaccessible" I probably should have said "unfindable", meaning that proof of existence is inaccessible. If no current editors can find either an ISBN, OCLC, or WP article about book, publisher, or author, or any archived talk about why the source is so inaccessible, or any other indication that the source is real, then yes it can be deleted in good faith. If it's a matter of pagination in a book everyone has, that is not this. I don't believe that my haphazard talk explanations imply potential bad faith, as I would trust your talk wouldn't either.
If you fear others adding whole book sections, just add "brief" to the edit.
A couple editors found this a necessary clarification of how accessible sources should be. It may well be a guideline-style clarification but it certainly met the initial consensus test. How about you propose what it should say and where it should go? Both my clause and Crum's sentences. Thank you. JJB 14:13, 3 June 2008 (UTC)
Not in this policy document, because there are too many exceptions that need to be added. For example I added information to the Burma Campaign using Bill Slim's book Defeat into victory. I used a 1956 second edition (No ISBN), which is available under OCLC. 1 in Sweden, 1 in England, 1 in Scotland, and Lots in Australia including the library in Wagga Wagga (What a great name). Now if someone deletes a passage that relies on this edition and that edit is reversed. Are we really saying that the person who reverses that edit must have access to this edition of 1956 edition of the book? If so the person doing the reversal better live in Australia! What if I had used a first edition, the OCLC has no record of it. I'll give you another example. What about journals like Journal of Railway and Canal Historical Society which is used to provide for information on commercial navigation River Teme. Tricky to get hold of in most parts of the world, but essential for that debate as it is the only source. As it happens the editor who added the information is still around. But if he were not then should the the information be deleted if current editors do not have access to the Journal of Railway and Canal Historical Society? These sort of details need to be discussed in a guideline and qualified with exceptions, it is not the sort of details that should be in a content policy page. --Philip Baird Shearer (talk) 19:10, 3 June 2008 (UTC)
We have discussed it at Wikipedia talk:Verifiability/Archive 27#Verifiability when you have hard-to-access sources, I'm sorry my hint towards the archive was unclear. The person who reinserts need only state that OCLC proves the existence of the book and thus that the original inserter (you) can be taken to have summarized it in good faith. This is not about deletion of material not quotable on demand. Quotability on demand should only be used in this application when nobody can demonstrate that the source exists. (It should also be used in the application where another editor has evidence for the suspicion that the source may be incorrectly summarized, but that's the NOR policy again.) Perhaps there is an improper ambiguity and it should say instead "If a demonstration of the source's findability is not readily available". JJB 19:24, 3 June 2008 (UTC)

Discussion continues at point 4 of #Recent changes.

[edit] factual claims in Wikipedia

I have moved the following recent addition to here:

All factual claims in Wikipedia articles must be attributable to reliable, published sources; in practice not all material is attributed.

This is a massive change to this policy, and there needs to be far more editors involved in the decision making process before such an innovative sentence is added to this policy document. I personally am not against it, but after the WP:ATT fiasco, such a change needs wide consultation and consensus before it is added, as it is a deletest's charter. --Philip Baird Shearer (talk) 17:54, 2 June 2008 (UTC)

Thanks for watching PBS! I like your fix to the lead quite a bit. However, you may not have noticed that the first clause of the sentence above (except as minorly changed by Wikidemo and vetted above), has been policy at WP:NOR for half a year. And the second half is rather indisputable and I don't think it implies anything erroneous either. See the wide consultation and consensus. This does not seem a massive change to policy, let alone to this policy. Would someone else mind reinserting please? JJB 18:24, 2 June 2008 (UTC)

Not quite the same thing, because it does not carry the implication that everything must be cited and "new analysis or synthesis of published facts", is not the same thing as facts. Besides if you think it is covered there then it does not need to be repeated here.--Philip Baird Shearer (talk)

Well then, I'm always a reliable "support" !vote on a redundancy argument. I'll take that as acceptance for this version, including Wikidemo's tweak and my clause from WP:A (regarded as policy already by Brimba), being brought back into WP:NOR as the suitable location. JJB 14:19, 3 June 2008 (UTC)
Lets get a firm consensus prior to trying to reinstate any version, as that is something we sorely lack atm. Brimba (talk) 14:58, 3 June 2008 (UTC)

Discussion continues at point 1 of #Recent changes. JJB 20:40, 3 June 2008 (UTC)

[edit] Example of the use of questionable sources

If this example is erroneous, then the policy really needs to be rewritten to show that such a use of questionable sources is legitimate. 77.235.40.107 (talk) 18:46, 2 June 2008 (UTC)

See #More on SPS. JJB 18:49, 2 June 2008 (UTC)

[edit] Recent changes

A user has requested comment on Wikipedia policy or guidelines for this section.
This tag will automatically place the page on the {{RFCpolicy list}}.

When discussion has ended, remove this tag and it will be removed from the list.

Could we please slow down on all the updates before we end up getting this page edit protected? —This is part of a comment by Wikidemo , which was interrupted by the following:

Too late. JJB 02:48, 6 June 2008 (UTC)

Let's take them one at a time, and talk about them to see if there is consensus. I've removed three that we were talking about[5] and don't seem to have an agreement on. The page has had so many changes lately it's hard to tell what the status quo stable version is, but I think, this approximates it. If anyone sees any other recent changes they wish to challenge we should take them out too and discuss. I'll start a discussion momentarily on all three. Wikidemo (talk) 07:04, 3 June 2008 (UTC)

Please comment below point by point to show where current consensus stands. RFC by JJB 02:48, 6 June 2008 (UTC)

[edit] Not all is attributed?

1. Although everything in Wikipedia must be attributable, in practice not all material is attributed.

  • Comment: First half of sentence is clearly not true per discussions on this page. It is not policy elsewhere. It comes from a nutshell on another policy page, WP:NOR, which is a simplification of that policy. In any event this page is the seat of that part of policy, not WP:NOR. Wikidemo (talk) 07:11, 3 June 2008 (UTC)
See above #factual claims in Wikipedia -Philip Baird Shearer (talk) 13:00, 3 June 2008 (UTC)
NOR requires that everything ultimately has a source independent of Wikipedia; see “Wikipedia does not publish original research or original thought. This includes unpublished facts, arguments, speculation, and ideas” “"Original research" is material for which no reliable source can be found.” V establishes the quality of the sources that are acceptable, and in so doing requires that some sources be directly cited. The important part of the phrasing per this page is “in practice not all material is attributed.”
It is true that nutshells are not in and of themselves statements of policy. It is also true that “Although everything in Wikipedia must be attributable” is a summarization of NOR and not directly part of V. However, the two overlap and rest against each other; NOR establishes the need for a source, and V the need to cite some of those sources. It is very hard to introduce “in practice not all material is attributed” i.e., that not every source needs a direct citation, without (a) mentioning policy better articulated on NOR, or (B) without having a very high noise to info ratio in the policy; or (C) lacking any sense of clarity.
There have been several recent edits that have either directly of indirectly implied that everything source needs to be cited; “in practice not all material is attributed” would be welcome for the sake of clarity. But its not a live or die item either. Brimba (talk) 13:38, 3 June 2008 (UTC)
Then proposal in prior section will work. JJB 14:25, 3 June 2008 (UTC) If you agree, please mark this one done. JJB 17:50, 3 June 2008 (UTC)
Since Brimba at WT:OR suggested that the second clause stay here but the first not, and I don't see PBS's comments above as barring that compromise, I'll mark done for now per Wikidemo. JJB 18:43, 3 June 2008 (UTC)
This is not done see my comments about guideline and the policy remaining succinct. This is instruction creep and not need in the policy. --Philip Baird Shearer (talk) 11:29, 4 June 2008 (UTC)
Uh huh, well I yielded up the first clause; so where should the statement "in practice not all material is attributed" go? If omitted, doesn't it scare up those deletionists who hack at the uncited? JJB 20:00, 4 June 2008 (UTC)

[edit] Removed by anyone?

2. Editors should provide a reliable source for quotations and for any material that is challenged or is likely to be challenged, or the material may be removed by anyone (addition in italics)

  • Comment: Seems to change sentence from descriptive to prescriptive. However, the focus of this section is on editing, not challenging sources. Moreover, "by anyone" seems to be a meaningless emphasis. Wikidemo (talk) 07:11, 3 June 2008 (UTC)
I agree it: seems to be a meaningless emphasis. --Philip Baird Shearer (talk) 13:00, 3 June 2008 (UTC)
Agree, already removed. JJB 14:25, 3 June 2008 (UTC)
Y Done JJB 17:50, 3 June 2008 (UTC)

[edit] Information quality?

3. Jointly, these policies determine the type and quality of material and [[information quality]] level that is acceptable in Wikipedia articles.

  • Comment: in discussion on this page, no consensus seems to be reached yet. The Information quality article is incomplete and the concept and use of "level" is not a close fit with what the policies get at. Wikidemo (talk) 07:11, 3 June 2008 (UTC)
I put in in as a link as determine the type and quality of material becasue I did not like the new wording, but I have not read the linked article. Now that I have, I do not think this policy should be linked to it until it is more mature than currently. For example "One man's information is another man's data", as the old saying goes, and the essay does not address that issue. --Philip Baird Shearer (talk) 13:00, 3 June 2008 (UTC)
I affirmed PBS's change, but I don't like the delink. I think it should either be linked or should be mentioned as "information quality" or both. Not linking due to maturity seems a weak argument, because if the topic is properly related to this article (isn't it?), then it shouldn't matter how stubby the article is. And, sofixit. JJB 14:25, 3 June 2008 (UTC)
The default (and status quo, if no consensus) is to not wikilink in policy pages as we might in an article. What does "information quality" mean (something not answered by the article), and what does it have to do with our policy on NOR, Verifiability, and NPOV? Wikidemo (talk) 16:02, 3 June 2008 (UTC)
Um, IQ = quality of information = quality of content of information systems = quality of content of Wikipedia. If we can compromise on unlinked "type and information quality of material", then please mark this one done. JJB 17:50, 3 June 2008 (UTC)
Y Done per Wikidemo. JJB 18:43, 3 June 2008 (UTC)
Further discussion

Seems to be some edit warring over nonconsensus changes without participating in the ongoing discussion[6] I have half a mind to simply restore what looks like the stable version. However, that seems unlikely to quiet things. At this point, unless people can agree to discuss here before making changes, I think it may be best to protect the page even at the "wrong version" to enforce stability.Wikidemo (talk) 16:54, 3 June 2008 (UTC)

Believe unnecessary at this time. JJB 17:50, 3 June 2008 (UTC)
I think a disinterested admin should be asked to protect it, so that we can slow down the changes and agree if they are necessary because as Wikidemo has shown there are people reverting without discussing changes. --Philip Baird Shearer (talk) 11:35, 4 June 2008 (UTC)

[edit] Demonstrably findable?

Following comment refers to proposals under 1-3 above (JJB 02:48, 6 June 2008 (UTC)):

I'm okay with all of the proposals in the above section, but they would involve reverting the new edit at least in part. It also re-adds the word "exceptional" a couple places (I have no opinion at all on that) and reintroduces a lot of material in this paragraph that I don't quite get (added material in italics):

4. It is not necessary that the source be findable instantly by any reader, merely that it be demonstrably findable (for instance, by library or archive request, ISBN or OCLC indexing, or notability of source). If a reliable source is not readily available (e.g., freely accessible online), the editor inserting or restoring the challenged material should be able to provide page numbers and/or direct quotes from the original text supporting the material, upon request. Direct quotes may be posted on the article's talk page for evaluation—they need not be added to the article.

- Wikidemo (talk) 18:16, 3 June 2008 (UTC) Numbered 4 by JJB 18:43, 3 June 2008 (UTC)
I respect that you may find this less comprehensible and a rephrase is fine, but I think consensus is building for it and I'm not taking time to write it now. Others are invited. JJB 18:57, 3 June 2008 (UTC) To resolve a potential ambiguity, I believe Crum's meaning is actually "If a demonstration of the source's findability is not readily available". Though this is a hair clunkier, I believe it's accurate summary of the best practices described at the archive linked above, and it need only be streamlined. I also agreed with making it "Brief direct quotes". JJB 19:28, 3 June 2008 (UTC)
Does item 4 mean that if I cite a non-web source (about half the sources I use), I'm expected to be perpetually available (even after my lifetime, and I'm no spring chicken) to provide quotations on demand for anyone who may challenge it? This is not a reasonable requirement. - Jmabel | Talk 19:45, 3 June 2008 (UTC)
No, quotations are only needed if an editor can actually question the book's existence. If there is no ISBN, OCLC, relevant WP article, or other proof that some editor could reasonably find the book, then it's rare enough that it should be quoted-- if it's otherwise reliable. I appreciate these questions because they represent the working out of this (IMHO eminently sensible) practice by a larger and growing circle. JJB 19:53, 3 June 2008 (UTC)
“No, quotations are only needed if an editor can actually question the book's existence.” No, there are many reasons for challenging a source. The sentence “It is not necessary that the source be findable instantly by any reader, merely that it be demonstrably findable” referees to the paragraph that precedes the sentence. To be truthful I am not sure what exactly is being said by the new paragraph as a whole. Brimba (talk) 01:43, 4 June 2008 (UTC)
Please remember that occasionally one may omit a word when one repeats a lot. At #Guideline stuff I said (emphasis added), "Quotability on demand ... should also be used in the application where another editor has evidence for the suspicion that the source may be incorrectly summarized, but that's the NOR policy again." Here I said "quotations are only needed if ...", meaning when demonstrating findability. If findability is obvious but the problem is that the editor's statement is inattributable to the source because it says something different, we call that OR; backup quotations are also needed there, but that application is in NOR's scope rather than in V's scope. I believe you observed the very useful distinction that arguments over whether something is attributable should remain at NOR, while this page can safely discuss when something should be attributed. I think I'm ready to rewrite and defluff; consider replacing the whole paragraph, and its preceding sentence, with:
The citation should clearly, fully, and precisely demonstrate that the source text is reasonably findable, such as by external link to the source website. Providing an ISBN or OCLC number, linking to an established Wikipedia article about the source, and directly quoting brief context on the talk page also each assert findability sufficiently.
I believe that addresses everything present in the other proposals, except for page numbers, which I must abandon because, if the source is otherwise unfindable, page numbers without direct context quotes will be too nonspecific to be sufficient evidence of findability. "Fully" and "established" are new. JJB 06:48, 4 June 2008 (UTC)
This is a non starter "linking to an established Wikipedia article about the source", links to Wikipedia articles can not be used to establish the notability of a source. I still say that this suggestion as to what is a reliable source, should be in a guideline and not in the policy document. As can be seen by the comments made by various editors including yourself JJB, that this new proposal is too wide open to misinterpretation and will be used to remove content that is properly cited and meets the current requirements. When one starts to dig there are too many exceptions for inclusion here. KISS --Philip Baird Shearer (talk) 11:45, 4 June 2008 (UTC)
(1) An established WP article is good-faith evidence of notability; I routinely link small-newspaper citations to their WP articles. (2) But that doesn't matter, because I removed notability from the criterion as unimportant. (3) The criterion is about findability, and a WP article demonstrates reasonable findability. (4) The criterion is not about reliability either. (5) Yes I thought Crum's text was somewhat loose, which is why I tightened it. (6) I believe the criterion is very clear now. (7) Clear statements are not wide open to misinterpretation; of course any statement is wide open to misrepresentation. (8) This criterion proposes loose standards for the untrodden ground of findability; it does not tighten any standards and thereby tempt overdeletion. (9) You are not evidencing that there are any harmful exceptions. (10) In short, we are failing to communicate: you are not taking my meaning and I am not taking yours. What shall we do about it? (11) I do see one omission, that the other cases are only for when the article is not "findable instantly" because of external link to the source website; I am adding that. JJB 13:53, 4 June 2008 (UTC)
It has been discussed many times and Wikipedia articles can not be used as a source. There are lots of other problems with the addition "The citation should clearly, fully, and precisely demonstrate that the source text is reasonably findable, such as by external link to the source website." There is no definition of "reasonably findable" To date it has been accepted that the citation must be verifiable, not "reasonably findable". The second sentence "Providing an ISBN or OCLC number, linking to an established Wikipedia article about the source, and directly quoting brief context on the talk page also each assert findability sufficiently." may be good practice but it should not be policy, this sort of thing should go in a guideline. --Philip Baird Shearer (talk) 18:46, 4 June 2008 (UTC)
(1) I am not using articles as sources. (2) I am using them as wikilinks from citations per standard procedure. (3) The definition of "reasonably findable" is the four examples: web, catalog, source data, quote. (4) The citation must be verifiable and findable. (5) So findability does not confer verifiability; it's only one of several factors. (6) Policy is about good practice. (7) This clause answers the question: What is the minimum you must do to have made cites sufficiently findable for others? (8) What is your answer to that question Philip? JJB 19:48, 4 June 2008 (UTC)
It is covered in WP:CITE and if it is not sufficiently covered there for your liking then, I think, that it the place to add this information. At this level this is covered by a generality and not detail. I would leave it at "The source should be cited clearly and precisely to enable readers to find the text that supports the article content in question." but I have no objection to "It is not necessary that the source be findable instantly by any reader, merely that it be demonstrably findable (for instance, by library or archive request)." --Philip Baird Shearer (talk) 10:20, 5 June 2008 (UTC)
I can try to see if it fits at CITE, but that doesn't mean we have consensus on either the wording or the placement. JJB 02:48, 6 June 2008 (UTC)

[edit] SPS

[edit] As sources anywhere?

5a. Use of self-published sources as sources about themselves, instead of in articles about themselves. See #Better section names. I favor the latter because the former is too broad and suggests fringe self-publication may be used anywhere if limited to claims about self. Other concerns about this section exist above below. JJB 18:57, 3 June 2008 (UTC)

[edit] Very closely allied topics?

5b. Above I said: In addition to the limited use in WP:SELFPUB for articles about the source, are SPS also usable when speaking about themselves in any other article, assuming proper weight? I that that would be too loose a rule for Crum to advocate, and two IPs were just reverted trying out another version of it. But I would also hold for a hairsbreadth of expansion, because Crum's concern goes a little ways. Currently the SPS is usable for articles about the author, the book, or the organization (three defs of source), but I think it ought also be usable for articles about the POV or movement or spouse or invention, etc., namely, very closely allied topics. An article about any minority philosophy would be poor if it excluded all SPS's simply because the philosophy is not the source. The SPS ought to be permitted to speak a bit more than about self. That would suggest the wording be "in articles about themselves or very closely allied topics", presuming that the community can tell the difference between closely allied topics (often minority POV articles) and loosely related topics (often overviews of all POVs).

Jmabel would add: or its employees, or its awards, etc. I note that the Ron Paul newsletters were used as a primary source for their editors, and not a soul blinked an eye about including the mastheads' testimony, though the editorship was highly controversial. JJB 20:16, 3 June 2008 (UTC)

[edit] Deleted by anyone?

6. Deleted by anyone. I think we already have consensus against this one as superfluous. Other concerns about this section exist above. JJB 18:57, 3 June 2008 (UTC)

Marking Y Done because enough people have reverted this text out to indicate consensus to me. JJB 18:19, 4 June 2008 (UTC)
I don't believe it, but all this time I didn't realize this was identical to #2. JJB 02:48, 6 June 2008 (UTC)

[edit] Exceptionally high-quality?

7. Exceptionally high-quality reliable sources. I think we have consensus favoring this one in that it's been on the page without a fight for IIRC a week and a half. JJB 18:57, 3 June 2008 (UTC)

Please see Wikipedia talk:Verifiability#Exceptionaly high-quality reliable sources? as clearly this change lacks consensus. The idea that certain claims can only be reported when vented academically (which is clearly what is actually being said here as all other sources fail by default) is contrary to NPOV and contrary to common sense. It takes 2 or 3 years for anything to be academically vented and so WP would lag behind other sources of knowledge in terms of being up-to-date on the grounds that we are being extra supper double cautious. Brimba (talk) 01:27, 4 June 2008 (UTC)

Actually, your link has one editor who thought "exceptionally high-quality" should be reverted but didn't revert, one editor who doesn't comment on that phrasing, and one who asked questions about it which, frankly, I didn't answer because I thought they should remain on a case-by-case basis. However, the change was added May 20 in one place and May 29 in the other, and I don't see that either were ever reverted, which is a pretty good silent consensus with so much other reversion going on. Now Brimba, where do you get "the idea that certain claims can only be reported when vented academically (which is clearly what is actually being said here as all other sources fail by default)" when looking at the words "exceptionally high-quality"? "Exceptional and high-quality" has been consensus for months, uncombined; putting them together is logically no different and is in no way a sudden invocation of 3-year peer review. Honestly, I am not succeeding in finding any other characterization for your comment than it being an unintentional strawman argument. What is it you really mean here? JJB 06:11, 4 June 2008 (UTC)

PBS dropped this one in a mass-revert which that did not state any reasons for objecting to the word "exceptionally". As I said, there is no essential difference between exceptional sources that are also high-quality sources (the prior consensus), and exceptionally high-quality sources, which is why that text lasted more than two weeks. If no reason is supplied for objecting, I will restore this as a two-week new consensus. JJB 14:11, 4 June 2008 (UTC)
exceptional adds nothing to the sentence and is bad style as the word exceptional has already been used. --Philip Baird Shearer (talk) 18:46, 4 June 2008 (UTC)
I am demurring on this question because there is no logical difference. I merely thought that a two-week silent consensus was sufficient to tip the scales toward verbal consensus. JJB 19:51, 4 June 2008 (UTC)
You guys, this is very hard to follow since people link an example change which seems to be the opposite of some comments. An exceptional claim (ie a most remarkable one) should not require an exceptional source. An entirely mainstream, conventional and respected one would be appropriate. It is entirely right to reference mainstream sources which have nonetheless reported exceptional claims. Anything else would be suppression of facts accepted by mainstream publications. The intent of this statement was (I presume) to distinguish it from its corrolarly: that unlike ordinary claims which may be supported by average or even poor sources, an exceptional claim can only be supported by a high quality source. Sandpiper (talk) 07:37, 6 June 2008 (UTC)
Another option is "extraordinarily high-quality source", which gets rid of the ambiguity of "exceptionally". JJB 18:58, 6 June 2008 (UTC)

[edit] Self-coverage exception?

Discussion break 2

8. (From above:) What should the title of section 2.4 be? No title has supplanted the prior consensus on the overlong title. My suggestions are above at #Better section names, Crum's are in edit history. Right now I'm favoring "Self-coverage exception". JJB 20:16, 3 June 2008 (UTC)

[edit] Expert SPS?

9. (From above:) In addition to the limited use in WP:SELFPUB for articles about the source, are SPS also usable when a recognized expert in the field as described in WP:SPS? I think obviously so (this is the former consensus), but Crum hasn't yet abandoned the position that the first use is the only use. Example: Suppose Stephen Hawking says something contentious but sincere about time travel on a blog, but it appears in no scientific journal or news report. Hawking is an established expert, he has reputation for fact-checking and accuracy, but he's not third-party. Doesn't that make him excluded under WP:SELFPUB because contentious, yet includable under WP:SPS due to expertise? JJB 20:16, 3 June 2008 (UTC)

I have a comment: be careful about reasoning through this based on analogy and example. This is going to apply to several million articles, in contexts we will inevitably neglect to consider. Trying to derive the general rule from a limited set of examples is risky. Will the same rule work when we cover philosophy? Ballet? Train routes? Motorcycle parts? Shapes of pasta? Another observation is that by "source" we mean a thee-fold issue of author, publisher, and specific work. The answer may be that given that there is no reliable publisher here, we should see if this particular work has the indicia of reliability with respect to the specific article content for which it is a source given its tone, context, who the author is? I would tend to think that if it's a truth about astronomy it would inevitably be in a published reliable source somewhere. If it appears only in Hawking's blog it is speculation or hypothesis, and even Hawking is not infallible. Or else it's an anecdote, personal history, recollection, reflection, a statement about himself or his colleagues, for which he is no more an expert than anyone else. Wikidemo (talk) 20:30, 3 June 2008 (UTC)
That was my attempt to show Crum that the "established expert" SPS paragraph is meaningful. Is it an accurate example, if the inserted material is, say, "Hawking has independently proposed Otherwise Unsourceable Bizarre Theory X"? If so, is that paragraph a description of the best practice? The issue is that Hawking is not third-party, so presumably is not reliable by definition even if his self-publication has a reliable tone to it. In this hypothetical his statement is surely notable, as a fallible speculation or hypothesis of an expert. I think the rule also works for self-published pasta experts et al. JJB 20:51, 3 June 2008 (UTC)

[edit] Must be attributable?

10. (and counting - the same editor keeps adding stuff without discussing here) but must still be attributable — i.e. a reliable source must be provided when challenged[7], to which I added the important qualification, a factual claim in article space is challenged reasonably and in good faith.[8] We seem to be going in circles on this, but it is not a moot point. Some people (erroneously in my opinion) challenge word choices, images, arguments at AfD, and other stuff that's not factual or not in mainspace for lack of citations. Further, some occasionally claim that obvious known matters of common knowledge (the sky is blue, the earth is round) need to be cited, hence my reasonableness limitation. Finally, some people wiki-game by removing vast amounts of material claiming lack of citations (e.g. the trivia wars), hence the good faith limitation. We already have a verifiability requirement. Making the language more absolutist would be a change in policy. I would rather keep it subtle, but if we're going to state an absolute requirement we need to be clearer on the boundaries and exceptions. Wikidemo (talk) 20:22, 3 June 2008 (UTC)

Thanks for discussing, I almost reverted the whole clause myself! "Must still be attributable" is redundant with WP:NOR and I would cut it to appease Brimba; it speaks only theoretically, and verifiability is supposed to be about practical proof that something has been attributed. The remainder of the clause is about "attributed" and is in both versions redundant with what came just before: All material challenged should be attributed. End of story. Unless you want to change "should" to "must", or add "in Wikipedia articles". There is no aegis to delete images or AFD arguments or other nonmainspace content because "unverified"-- with the BLP exception of course. Deleting vast amounts of mainspace material as uncited or poorly sourced must be taken in good faith if there is no evidence of bad faith. If there is such evidence, the current provision will not solve the problem, because the bad-faith editor is free to continue indefinitely to argue that all sourcing supplied is poor. Because even policy provisions must assume good faith, a different provision would be needed to deal with bad faith, such as my "burden of proof shifting" proposal from the archive. JJB 20:36, 3 June 2008 (UTC)
Can I close this one, since there hasn't been activity to reinstate this basically redundant text? Y Done JJB 18:28, 4 June 2008 (UTC)

[edit] The first threshold?

11. The first threshold for inclusion in Wikipedia is verifiability. I like this change and proposed it to resolve concerns of Philcha and MarionTheLibrarian. PBS dropped it in a mass-revert that did not state any reasons for objecting to the added word. (It is not a good time to be reverting numerous edits without discussing them individually.) There are technically many thresholds for inclusion in Wikipedia, but we don't want to say "a threshold", so "the first" works quite well. Rather than get in the complex foray, I'll just mention it here and let its obvious merit attract its own attention. :D JJB 14:11, 4 June 2008 (UTC)

I agree with this proposed change ... this may well solve the constent carping about "Truth"... our first priority is verification, truth is secondary(although I still feel that what we want to discuss is "factual accuracy" as opposed to "Truth"... "Truth" can be subjective, while "factual accuracy" isn't. But that is another debate for another time.) Blueboar (talk) 18:01, 4 June 2008 (UTC)

First "threshold" is "The piece of timber or stone which lies below the bottom of a door, and has to be crossed in entering a house; the sill of a doorway; hence, the entrance to a house or building."(OED) How can one have more than one threshold (especially when it is being used metaphorically)? Second what are meant to be the second, third, etc "thresholds"? Third who decides on the order? --Philip Baird Shearer (talk) 18:22, 4 June 2008 (UTC)

Multiple doors, multiple thresholds. First we walk through the door called Verifiability (the front hall of Wikipedia). From there we can walk through other doors, such as the one called "truth" or "factual accuracy" or whatever. But if a different word conveys the same concept, I am open to alternatives. Blueboar (talk) 18:32, 4 June 2008 (UTC)
Concur with PBS. At best, it's question begging. Marskell (talk) 18:37, 4 June 2008 (UTC)

Many buildings have multiple thresholds, such as at both ends of a foyer. It is not necessary to state the order of the other thresholds, but NPOV is probably an important one. The point of the word "first" is that just because something crosses the threshold of verifiability, it's not automatically allowed to be included; it must also be neutral, nonlibelous, and so on. The reason V can be called "first" and the others not designated is that V is usually a clear yes-no test, while NPOV is much more gray-shaded. Another solution which nobody responded to was "The quality threshold". Well, at least talking is better than edit warring, but I've had to put down severe temptations to crack jokes about this behavior. Please continue to build consensus. JJB 18:43, 4 June 2008 (UTC)

When used a metaphor one usually (always?) has only one threshold eg a bride is carried across the threshold, never across a threshold even if the building has multiple entrances, similarly visitors cross the threshold in to a house rarely a threshold into a house. --Philip Baird Shearer (talk) 19:02, 4 June 2008 (UTC)
(actually, carrying the bride across the theshold isn't a metaphor... it refers to an actual threshold... but I get the point.) What I think is being attempted here is to ease the concerns that truth has no place in Wikipedia. I think we all agree that it does... but that we consider Verifiability to be the primary criteria for inclusion. Is there another way to express this. Perhaps if we toss out the entire "threshold" analogy we could move forward on this to everyone's satisfaction. Blueboar (talk) 19:20, 4 June 2008 (UTC)
Was thinking that myself Blueboar. The first criterion. There are plenty more where that came from, but I am having strong challenges remaining civil lately for some reason so I will stay brief. JJB 19:36, 4 June 2008 (UTC)
Per the comments of 1!=2 above, I don't think anyone need argue that truth is necessarily one of the thresholds or criteria. But because we recommend demonstrating good faith, believing something is true might well be. But it doesn't matter, because I think we agree NPOV and BLP are also key "thresholds" or criteria for inclusion, and the only issue with adding a word like "first" is to affirm that you're not all the way in just because you're verified. Which is what I think 1!=2 said. JJB 19:40, 4 June 2008 (UTC)
Perhaps "An important criteria", or "A fundamental requirement" or something similar, but not phrase with a number implying an order, or text that implies more than one threshold. --Philip Baird Shearer (talk) 10:32, 5 June 2008 (UTC)
Either "An important criterion" or "A fundamental requirement" is fine, as many others would be. Took awhile to get there, didn't it? Who will agree to close this one? Everyone knows that anyone can add or remove a {{done}} tag, right? JJB 02:48, 6 June 2008 (UTC)

Are people getting a bit literal here about building design? If the niceties of the metaphor are becoming an issue, then maybe it is a bad metaphor? If you say 'the first threshold for', then it means that what is coming up is pre-eminent above any other consideration. I do not believe this is true. An article about electricty starts off, 'George bush is president of the united states'. No? obviously not, because it is wholly irrelevant to the subject. The first criterion is whether something is relevant, though I wouldnt care to say which acronym this is. Sandpiper (talk) 07:51, 6 June 2008 (UTC)

I support "a fundamental criterion" it is not merely important but it is one of the foundational principles right up there with anyone can edit. Criterion rather than requirement because more precise meaning. First implies an order (Sandpiper's comment) My opinion: there is no first criterion, all criteria must be met. Relevance, verifiability, reliable source, neutral POV, encyclopedic tone/subject matter, the entire cascade of policies applies simultaneously. User:Pedant (talk) 08:54, 6 June 2008 (UTC)

[edit] Partial rewrite of Questionable Sources

In the discussion WP:AN/I#Editor deleting reliable sources, the issue of what happens when an unreliable source copies or claims to copy a reliable source verbatim. For example, if a Pastafarian web site reprinted a New York Times article from 1987 on the nutritional value of pasta. The Pastafarian web site is not a reliable source and should not be used in articles about food, but the New York Times article is not otherwise widely available: You either have to pay to see it or visit your local library.

The consensus in that discussion is still evolving but it seems to be: The citation should be the original, reliable source, but it is OK to supplement it with a link to the web page hosted at the unreliable source. However, there was concern that some unreliable sources may deliberately or carelessly mis-copy 3rd-party materials.

There is also the issue of fringe source != unreliable source: You can be unreliable and not fringe, and you can be fringe and, with respect to reprinted material from reliable sources, reliable.

With that in mind, I would like to change

WP:V#Questionable sources from

Questionable sources are those with a poor reputation for fact-checking. Such sources include websites and publications that express views that are widely acknowledged as extremist, are promotional in nature, or rely heavily on rumors and personal opinions. Questionable sources should only be used as sources about themselves as described below. Articles about such sources should not repeat any contentious claims the source has made about third parties, unless those claims have also been published by reliable sources.

To:

Questionable sources are those with a poor reputation for fact-checking. Such sources typically include websites and publications that express views that are widely acknowledged as extremist, are promotional in nature, or rely heavily on rumors and personal opinions, but may also include blogs and other sources which do not do a good job of fact-checking. While such sites may contain some reliable content, particularly if that content pulled from reliable sources, anyone quoting from such sites should be prepared to defend the accuracy and reliability of the material. Except for content shown to be reliable, questionable sources should only be used as sources about themselves as described below. Articles about such sources should not repeat any contentious claims the source has made about third parties, unless those claims have also been published by reliable sources.
When an otherwise-unreliable source quotes material from a reliable source, and that material is not otherwise readily accessible, it is better to cite the original source and provide a web-link to the questionable source. For example, after verifying that the book and author are not themselves fringe elements:
is preferred over
  • Wheelan, Charles, Naked Economics: Undressing the Dismal Science, 2003, W. W. Norton & Co., ISBN 0393324869, p. 29
because other editors can fact-check it without going to their local library or bookstore.

What do you guys think? davidwr/(talk)/(contribs)/(e-mail) 20:49, 3 June 2008 (UTC)

Speaking of pasta! I think you want to start with WP:CITE with that draft. It's OK as a start but the language needs some tightening (e.g. delete "typically"). Aside from self-coverage, QS content is only usable if it actually quotes and cites RS content, and becomes unusable upon simple proof that the RS content was misrepresented; and I don't think the draft reflects that fully yet. Otherwise it's already covered here, including the blog footnote we already have (which I think should be promoted into the main text as was done in another policy). The general idea that RS may be provided with a link to QS is, I believe, well-supported for inclusion at WP:CITE, and might merit a clause here. JJB 21:02, 3 June 2008 (UTC)
The draft could be a little shorter. I support the intention, though, and would like to see it clarified in policy. forestPIG 21:04, 3 June 2008 (UTC)
I forget where (probably here or in RS), but there used to be a clause somewhere about convenience links. The essay to which I link is worth reading. Jakew (talk) 21:10, 3 June 2008 (UTC)
We should not, however, be knowingly linking to copyright violations (for example, a copyvio of a documentary uploaded to Youtube, or a copyvio of a New York Times article on a personal website). BuddingJournalist 01:29, 4 June 2008 (UTC)

er, the proposed draft says Questionable sources are those with a poor reputation for fact-checking. Such sources ..... may also include blogs and other sources which do not do a good job of fact-checking. So sources which do a bad job of fact checking may include sources which do a bad job of fact checking? Umm. Sandpiper (talk) 08:33, 6 June 2008 (UTC)

[edit] Second draft

Try this:

Questionable sources are those with a poor reputation for fact-checking. Such sources include most websites and publications that express views that are widely acknowledged as extremist, are promotional in nature, or rely heavily on rumors and personal opinions, and may also include blogs and other sources which do not do an adequate job of fact-checking. In general, questionable sources should only be used as sources about themselves as described below. Articles about such sources should not repeat any contentious claims the source has made about third parties, unless those claims have also been published by reliable sources. An broad exception is the convenience link (essay), where an questionable source holds an authentic or at least an apparently-authentic copy of a reliable source. In such a case the reference should include information pointing back to the original source as well as to the convenience link. Use caution when using convenience links: Sites with reputations for hosting inauthentic but apparently-authentic copies of reliable material should not be used, nor should links to obvious cases where copyrighted material is used without permission.

davidwr/(talk)/(contribs)/(e-mail) 21:15, 3 June 2008 (UTC)

If we are going to discuss convinence linking in this policy (and I am not sure we should) we need to point the reader to WP:SAYWHEREYOUGOTIT. The key is that we can not act as if we have read Wheelan's book, if all we have seen is a quote at kookyeconomictheoriesrus.com. Unless we have seen both the book and the website, we have to cite the web page. The important thing is that the material is cited to a reliable source. If that means people have to go to the library or book store to check it, fine. Adding the convenience link is an optional courtesy.Blueboar (talk) 21:42, 3 June 2008 (UTC)
Oppose. Unclear, both in terms of wording and necessity. Marskell (talk) 21:47, 3 June 2008 (UTC)
Do you oppose Blueboar's comments or my draft? If mine, please re-indent.[resolved and correctly indented] I'm open to rewording. I'm also open to reasons why this might not be as necessary as it looks. For example, if the point I'm trying to get across: using an unreliable source to quote a reliable source is sometimes okay - is already covered elsewhere in WP:V and I just missed it - then open my eyes and I'll withdraw the proposal. davidwr/(talk)/(contribs)/(e-mail) 21:50, 3 June 2008 (UTC)
I don't know; I think I'm opposed to this on the grounds that the convenience of a convenience link can be far outweighed by how objectionable an objectionable fringe site is. Nasty fringe sites can boost their Google traffic by hosting RS material. Pastafarian is kooky; a pedophile or neo-nazi website is something else. Also, someone brought up copyvio, which is a good point. If the only thing the site has going for it is a purported full text convenience link of something not available elsewhere, shouldn't the lack of availability elsewhere on the web indicate a copyvio issue? (Perhaps the wording could be adjusted to 1) the site does not promote an extremely anti-social fringe view/Wikipedia should not be abused to boost traffic to these sites 2) the site gives copyright info.) -PetraSchelm (talk) 22:20, 3 June 2008 (UTC)
@davidwr. I was opposing your wording and (lack of) necessity. Your initial example is a split hair. No, obviously, we cannot cite an unreliable secondary source to quote some primary source that we assume is reliable. Marskell (talk) 22:22, 3 June 2008 (UTC)
It's not at all obvious. In fact, I would say that absent evidence that the unreliable source is misrepresenting the quoted source or has a history of misrepresenting similar material, it is obvious that we can use it. By the way, by "reliable" I mean the quoted source is known to be reliable, such as a pre-1923 article from The New York Times, and the copy hosted by the blog or other unreliable source is presumed to be authentic based on lack of a reason - such as someone checking and saying "they mis-copied it" or a history of such mis-copies from the unreliable source - to think it's not. If either fails, then buh-bye. By the way, sources cited by blogs and fringe sites are not necessarily primary sources, they may be secondary or tertiary sources, but that's not the point. davidwr/(talk)/(contribs)/(e-mail) 22:48, 3 June 2008 (UTC)
PetraSchelm, that's Google's problem. Google's algorthims or the games 3rd party sites may or may not use to manipulate them should not affect Wikipedia editorial policy. If Google is an issue, that should be addressed at a technical level, perhaps by either Google or Wikipedia taking steps so including those links does not affect the rankings. To put it another way: We should edit as if Google did not count outbound links from Wikipedia articles. To put it a 3rd way: We should not be held hostage to Google's algorithms. WP:CITE strongly recommends that if material is courtesy-linked from fringe sites, and the same material is available from a less-fringe site, the reference should be changed to the less-fringe site. However, if the material is only readily available from fringe sites, or less-readily available by a trip to a courthouse or library, then I think we should allow it, and WP:CITE seems to agree. The issue of copyvio is already addressed above and by WP:CITE. davidwr/(talk)/(contribs)/(e-mail) 22:48, 3 June 2008 (UTC)
Ah yes, I see here: "When offering convenience links, it is important to be reasonably certain that the convenience copy is a true copy of the original, without any changes or inappropriate commentary, and that it does not infringe the original publisher's copyright. Accuracy can be assumed when the hosting website appears reliable, but editors should always exercise caution, and ideally find and verify multiple copies of the material for contentious items." This means we already address this, and that the burden of proof is squarely on the editor wishing to include such a convenience link, not the other way around. That does still leave the issue of extremely objectionable/anti-social fringe sites. (Re reliability and the site, International Pedophile and Child Emancipation which sparked this discusssion, if there's no factchecking staff/the site is operated by one anonymous person, it can be presumed to be unreliable on those grounds, I think--it certainly can't be presumed to be reliable). Also, I don't think it is "Google's problem"--bringing the project into disrepute is an issue that Arbcom and Jimbo take seriously, and there is no reason to gratuitously linke to pedophile websites for dubious convenience links; there's not a net benefit when harm to the project comes into the equation. -PetraSchelm (talk) 23:02, 3 June 2008 (UTC)
If a source is not reliable, it's not reliable to use as a convenience link either. A source with no positive reputation for accuracy or fact-checking, especially if it's a fringe source, does not instantly gain reliability because it posts a document that it claims is an accurate copy of a document published elsewhere. On the other hand, for websites that are known to be reliable based on reputation of the site and the specific individuals who run it or oversee its editorial content, documents posted on that reliable source website can dependably be used as convenience links.
Another consideration is that when reliable source websites post documents, they do so with permission from the copyright owner. If a website posts what it describes as a copy of a published document, but does not have permission to post it, that introduces significant doubt about the reliability of the website as a source and the accuracy of the copy they present. --Jack-A-Roe (talk) 23:49, 3 June 2008 (UTC)
A scenario to think about:
I run a blog that espouses a fringe theory of world creation. I include verbatim a copy of a 1910 newspaper article from the New York Times. On my blog I comment "see, in 1910 a New York State Senator believes that schools should teach all theories of creation no matter how unpopular" Now, a Wikipedia editor who happens to work at a library edits the article on education and cites the New York Times article and includes the link to my blog as a convenience link. She makes a comment on the article talk page saying "Sorry about the bizzarro link but it's the only online copy I could find. I read the article on microfilm and the copy online is the same." What then? Should the link be allowed?
If so, should the link be allowed if she didn't bother to make the comment on the talk page, leaving editors to wonder why she linked to a blog full of garbage theory for a New York Times article?
davidwr/(talk)/(contribs)/(e-mail) 02:58, 4 June 2008 (UTC)


I think we need this in some form, to avoid both the use of links to apparently unreliable sources that do not say that the material is actually from a good source originally, and the omission of the source from which the material can actually be found by most users. This middle road is the logical extension of what we mostly do already--I see it as a restatement of current practice. I prefer the first version as being much clearer than the second; the example given there illustrates it perfectly. Among the reason for giving the convenience source is that we can upgrade it--in the example just above, the 1920 NYT is now available free on the NYT site, so we could now remove the previous link. As for google, I agree we shouldn't give undue prominence, but sources are where you find them. Its enough to give a non-partisan source when such is available in preference. But some advocacy websites host really remarkably extensive collection of documents, often from both sides of the question, & unless we know they are faked or copyvio, they are in fact valuable resources for the issue. Sure, it's their purpose in putting them there--the hope that you'll also see the advocacy, but if they do provide the material without messing with it, they are making in at least that respect a positive contribution to the discussion. (I think CITE is the better place by the way, not here--this is a guideline, not policy). DGG (talk) 13:06, 4 June 2008 (UTC)

You mean this is policy, not a guideline, I assume. My own position is, no, the link shouldn't be allowed. To paraphrase a Wiki maxim, if some document or other is notable enough for description here a reliable source ought to have it in an archive. Marskell (talk) 13:19, 4 June 2008 (UTC)

er, the proposed draft says Questionable sources are those with a poor reputation for fact-checking. Such sources ..... may also include blogs and other sources which do not do a good job of fact-checking. So sources which do a bad job of fact checking may include sources which do a bad job of fact checking? Umm. It can be argued that a wiki terms 'unreliable source', which in real terms is a source of unknown quality (it might in reality be entirely reliable) ought to be allowed for convenience when referring to a document (or a quote from one) which people can read there online. This is a benefit to readers. There should be a formally acknowledged distinction between 'unreliability' with regard to signposting material from an accepted wiki-reliable source elsewhere and their own original material. Sandpiper (talk) 08:33, 6 June 2008 (UTC)

[edit] WP:ATT

There are discussions and edits being made at WP:ATT that people here need to be awair of (example: changing the line about "The threshold for inclusion..." to "The first threshold for inclusion...".) I personally feel that some of the proposed changes are good, and others are not so good ... but my main concern is that they are being discused and changed there instead of here. WP:ATT is supposed to be (in part) a summary of this page... so I think key sentences should essentially follow what is stated on this page. If a major change is needed or wanted, it should be discussed and made here first, and then imported over to the summary. Blueboar (talk) 12:12, 4 June 2008 (UTC)

I think, when WP:ATT has its backed turned, someone should drop a little cyanide in the kool-aid and pass it over. Marskell (talk) 13:21, 4 June 2008 (UTC)
Discussion continues at point 11 of #Recent changes. Marskell, something like that was reverted 14 times over there this past week. JJB 14:33, 4 June 2008 (UTC)

[edit] Please slow down

Too many chages are happening too fast, and I have lost track of what is being added, changed, reverted, etc. ... I don't mind people being bold and making an initial edit without in-depth discussion... but once a potential change has been reverted please stop, discuss, and reach a firm consensus before continuing. Blueboar (talk) 15:05, 4 June 2008 (UTC)

I agree entirely! I hope I'm not going overboard to direct virtually every conversation back to #Recent changes where consensus can be built point by point, but Blueboar, doesn't it seem to you that there are a few people running around a bit too liberally here? Is it just me, or is this a case of everyone talks about consensus but nobody does anything about it? Okay, frustration over, don't answer that. JJB 17:40, 4 June 2008 (UTC)

[edit] Protected

Given the spate of medium level revert wars over the last week, I have f-protected this for three days. I have been involved in wording disputes previously here but am not involved in this one and take no side in the various competing versions for the moment. (I have made no direct edits myself, recently.)

My main concern is that this generally very stable policy not be hastily changed. There is on-going discussion and I don't think a protection break will hurt. Marskell (talk) 19:04, 4 June 2008 (UTC)

Thanks Marskell... now, can we please sit down and go through the proposed changes one by one... and reach a consensus on one change (or agree to postpone) before we move on to the next? Blueboar (talk) 19:11, 4 June 2008 (UTC)
Thanks from me too, Marskell.
I have a couple of objections to the recent changes:
1. This was added: "It is not necessary that the source be findable instantly by any reader, merely that it be demonstrably findable (for instance, by library or archive request)."
First, the writing is a little odd, and secondly it's stating the obvious, given that we allow offline material, which is never going to be "findable instantly" unless it's on your bookshelves. —This is part of a comment by SlimVirgin , which was interrupted by the following:
This came as a result of Wikipedia talk:Verifiability/Archive 27#Verifiability when you have hard-to-access sources. This and its expanded version attempt to convey a criterion several editors desired for how accessible or findable sources need to be. There is a more recent distilled draft at #Recent changes point 4 that addresses those concerns and I'd appreciate your consultation there. JJB 19:32, 4 June 2008 (UTC)
2. This was removed: "It is important to strike a balance between being quick to removed unsourced material that is clearly wrong or in some way damaging, while at the same time making sure that challenges are reasonable, and that editors are given enough time to find supporting sources. Before you challenge unsourced material, ask yourself whether you really do doubt that the material is accurate. Unsourced material should not be removed simply because of a difference of opinion, or in an effort to make life difficult for other editors. As with all policies, this one must be applied with common sense."
This was added to the policy after long discussions on talk started by Phil Sandifer, who wanted something to be added to prevent editors from removing stuff willy nilly and citing this policy as justification. I think it needs to be restored. SlimVirgin talk|edits 19:14, 4 June 2008 (UTC)
I'm not familiar with this one, although it appears a bit unrehearsed yet. It would probably pass muster with tightening. I too want something added for that reason, and I have a proposal in the archive. We might be able to harmonize. JJB 19:32, 4 June 2008 (UTC)
Regarding SV's 1, my first impression is that it's non-colloquial. I understand what "demonstrably findable" means but it just...sounds weird. However, even if the point is a little obvious and the wording weird, I don't mind the intent. A lot of people show up on this talk wondering about the "reader can check" idea. I'd just suggest better wording.
Regarding 2, I agree that this is potentially useful but we need to discuss its positioning and relation to other wording (such as the Jimbo quote, which creates an opposing mood). (Could you link back previous discussions, Slim?) What constitutes a "reasonable challenge" is an extremely sore point—perhaps more important than the other things being reverted over. We should give it good discussion. —This is part of a comment by Marskell , which was interrupted by the following:
See Edit 18:34, 6 April 2008 by Phil Sandifer and Arhive 26: Do not leave unsourced information in articles for too long? Archive 26: Changes to "burden of evidence" section, Archive 26: This is not wrapped up and Archive 26: Changes to "burden of evidence" section 2 --Philip Baird Shearer (talk) 11:20, 5 June 2008 (UTC)
My position has not changed, I do not think it should be restored. I have already given details of why in the archived sections listed above, but the major problem is the third sentence could be used to reverse the burden of proof and that I think is a retrograde step. --Philip Baird Shearer (talk) 11:33, 5 June 2008 (UTC)
Lastly, I'm sort of opposed to JJB's numberings and Done! checkmarks. (Sorry—but it might just confuse other people.) Marskell (talk) 19:42, 4 June 2008 (UTC)
Agree on #2, that's an important discussion to hammer out. Believe that your concerns on #1 are addressed by the draft above I already mentioned. And what do you propose in lieu of Wikidemo's and my numberings, in order to achieve consensus on several different point-by-point discussions? I apologize if I forgot to say that all can demote the "done" tags or add "not done" tags if they disagree. Thanks. JJB 19:56, 4 June 2008 (UTC)

[edit] Move the WP:PSTS section from the WP:OR page here ?

It is odd that the primary, secondary and tertiary sources are defined in WP:OR instead of in this page. Any issue with moving it here ? Pcarbonn (talk) 11:31, 5 June 2008 (UTC)

You might want to look at the history of PSTS for clues as to why it is where it is, and how it is. There was considerable activity, debate, and a little heat, over PSTS several months ago. I'm not being coy here, I don't really remember. I just recall that it was the subject of some considerable debate. Wikidemo (talk) 19:16, 6 June 2008 (UTC)

[edit] Link to historical page

The link to Wikipedia:NPOV, V and OR under See Also links to a historical page. Should this be changed to link directly to Wikipedia:NPOV and Wikipedia:OR or another current page? WeisheitSuchen (talk) 14:49, 5 June 2008 (UTC)

I don't think that page was ever intended to be an active policy or guideline page... It was created during the formation of WP:ATT... and was designed to give historical background as to how the various core policies developed and grew out of each other. It is useful for understanding the original intent of these polices, so I think a link to is is appropriate. Perhaps it should be renamed to something like "History of WP:NPOV, WP:V and WP:NOR" and marked as an essay (as opposed to "historical", which is really for failed policy proposals). What do people think?

True meaning of the word "The Blister" : The Blis-ster \'blist-r\ 1. A paranoid misguided individual who calls herself a mother, which chooses to harass the Father of her children. 2. Any person resembling this, as a unreasonable erroneous human being. 3. A person who hurts others and says that she is only doing it for the well-being of her children. Her swelling lack of self-confidence is transparent to everyone concerned. -v.; blis-tered; blis-ter-ing \-r(-)ing\ 1. To cause blisters. New wives may cause the blister to fester. 2. To get a rise from the blister. The blister is usually under stress and striking out against the people who love her children but do not care for her. -blis-tery\-r(-)e\ adj. —Preceding unsigned comment added by 209.183.55.46 (talk) 22:23, 5 June 2008 (UTC)

[edit] Reorganization of the discussion

Could we please have a reorganization of the various proposals... What is settled and what is still outstanding? It is getting very difficult to follow the conversations... and there is some overlap (for example: we now have overlaping numbering in different threads: Wikidemo has a proposals #1 and #2... and Slim Virgin has points #1 and #2). When someone posts a reply to something, I am finding it hard to figure out which proposal or point is being referred to.

I would suggest seperate headers for each outstanding proposal, so that if someone wishes to comment on a proposed change, we all know which change they are commenting on. Thanks. Blueboar (talk) 15:22, 5 June 2008 (UTC)

Rather than having twelve bullets at once we could have a section per point. Start with the instantly findable business? Marskell (talk) 17:13, 5 June 2008 (UTC)

[edit] "Wikify" flag v. "Articles and posts on Wikipedia may not be used as sources"

Wondering about an apparent contradiction.

WP:V states, and many talk page discussions endorse, "Articles and posts on Wikipedia may not be used as sources".

In contrast, the "Wikify" flag renders as, "This article or section needs to be wikified to meet Wikipedia's quality standards. Please help improve this article with relevant internal links."

In practice, that text has led at least one editor to remove an external link and replace it with a link to a Wikipedia article. Looks like a good-faith edit by an editor who's simply taking the flag's text at face value -- Wikipedia's quality standard is internal links (I'm not agreeing; I'm just trying to point out that the editor's actions reflect a reasonable interpretation of the flag's text). So, a couple questions for more experienced editors:

1. The flag's text, in its most straightforward reading, leads the reader to think that Wikifying = adding internal links; internal links are the only component of Wikification, and are Wikipedia's quality standard. Am I correct in thinking that's not the intended message?

2. If not, where should I go to discuss getting the language of the flag more in line with the intent?

I think it's likely that readers of this page are almost all experienced editors who read the Wikify flag text and instantly understand its underlying intent. However, as a relative newcomer, the value I'm trying to add is to point out the difference between the intent, and the words that actually appear. I think Wikipedia's goals would be achieved better and faster with a little word-smithing to get the intent and the text into alignment. Thanks, Thirdbeach (talk) 19:17, 5 June 2008 (UTC)

Template talk:Wikify would be the place to discuss any proposed changes in wording. This is an interesting suggestion, though I'm not certain that replacing external links with internal ones constitutes the "most straightforward reading" of the text. =) Powers T 19:22, 5 June 2008 (UTC)
It's a common-sense judgment call. If I say "Earl Campbell is an NFL football player" I need an external reference to say that. If in the NFL article I say "Notable players include Earl Campbell" I don't, since there will be obvious references in the E. C. article that say so. In general, the more obvious the references in the target article are, the less likely you need a reference in the source article. The two things you must avoid: Circular references with no external link, and no highly obvious reference, either in the source or target article. If I have to spend more than a few seconds at the target article looking for the relevant reference, then I should put a reference at the source article in addition to the wikilink. You are allowed to have both. davidwr/(talk)/(contribs)/(e-mail) 23:41, 5 June 2008 (UTC)
A similar difficulty applies where there exist sub-articles. The main article is supposed to contain a summary of the sub article. Any summary of one article has to be original research done by wiki editors, that is what a summary is. Obviously, it cannot have external references because it is original. The only effective way to reference it is to refer someone to the longer version which presumably will reference points as they occur. The only logical answer to the question, therefore, is to presume that because something is wikilinked, it is satisfactorily referenced. If this is doubted, then the matter should be argued out on the linked page. Sandpiper (talk) 06:44, 6 June 2008 (UTC)
Thinking about it more, my impression is that the purpose of WP:V and external links is to make an individual article robust; the purpose of the "wikify" flag and internal links is to make Wikipedia as an integrated system robust. Does that sound right? If so, it seems like the next step is to make the purposes and usage of internal vs. external links clearer, especially in the "wikify" flag where it currently (IMO) could be clearer. Please comment on my proposed rewrite of the flag, which is on Template talk:Wikify. My thanks to LtPowers for the link. Thirdbeach (talk) 19:49, 6 June 2008 (UTC)

[edit] Extended protection

I just protected ATT for three days given edit-warring, and decided to extend protection for the same period to here. Some means of centralized discussion needs to be agreed upon. We have more than enough fora--it needs to be reduced to one forum, in addressing current arguments. One place to discuss the current rash of edit warring. Marskell (talk) 22:11, 5 June 2008 (UTC)

It's interesting that you protected that page and left the transcluded subpage (intended to prevent such protection by separating the header dispute from the main content) unprotected. —David Levy 00:06, 6 June 2008 (UTC)
Whatever policy discussions are going on should take place here, not at ATT, which is just a summary. Can someone summarize what the issues are with V that people want to change? SlimVirgin talk|edits 01:22, 6 June 2008 (UTC)
Agree, and this is the eleventh thread on this page which I am repointing to #Recent changes. Someone objected to my organizing and centralizing and clerking the discussion in one location, so I held off, but nobody else wanted to flag it so I'll go ahead. There are a couple other open questions in some other sections but they don't appear to be edit-war level. I would appreciate all users commenting to demonstrate where consensus lies on each point. JJB 02:24, 6 June 2008 (UTC)
Sorry JJB, for the wheel-spinning. Let's try to solve one thing at a time thing to at least make some use of this protection. How about findability? At present:
"The source should be cited clearly and precisely to enable readers to find the text that supports the article content in question. It is not necessary that the source be findable instantly by any reader, merely that it be demonstrably findable (for instance, by library or archive request). If a reliable source is not readily available (e.g., freely accessible online), the editor inserting or restoring the challenged material should be able to provide page numbers and/or direct quotes from the original text supporting the material, upon request. Direct quotes may be posted on the article's talk page for evaluation—they need not be added to the article."
Some have objected to a bit of odd wording and it's also longer than it needs to be. How about:
"Readers should be able to check source material either immediately through a weblink or through a library, a journal or newspaper archive, or other standard, professional collections. Editors should cite sources precisely, providing as much publication information as possible, including page numbers for book sources."
This drops the direct quote business, which I think obscures the main point and should be handled separately. Marskell (talk) 16:55, 6 June 2008 (UTC)
Thanks, I actually agree that the draft you quote was odd and long (it wasn't mine). Assuming you don't want to comment at #Demonstrably findable? or at WP:CITE#Findability, the question is: what standard do I, the inserter, need to meet in good faith in my own judgment to ensure this checkability reasonably for others? That is, if it's not online, how can I evidence that the material can be found at a library, archive, collection, etc.? My answer (per my draft at those links) has been that the cite should provide something that evidences that, either by way of an ISBN, OCLC, wikilink, or direct good-faith quote of really rare material; if it doesn't, and nobody else can find it, then it's not verifiable. (Also, which page does this info go on, V or CITE?) Your draft is fine but it doesn't answer the question, which was raised in Archive 27 as "hard-to-verify sources". JJB 17:51, 6 June 2008 (UTC)
First, JJB, I think part of your frustration stems from trying to solve everything at once in the recent changes thread.
On your specific Q, I am strongly opposed to foisting specific demands on people adding material beyond the general demand that they provide proper publisher and author information for an RS source. I don't have to mail you a copy of the book I cited this morning. Asking someone to quote a book or journal on article talk is problematic—it amounts to saying "we assume you are lying until we get a direct quote." (And, of course, if I was some sneaky vandal I might just invent a quote for article talk.)
So, we ask people to provide primary publication info without throwing up weird hoops to jump through. I would avoid getting into the specifics of ISBN etc. on a generalized policy. Marskell (talk) 18:17, 6 June 2008 (UTC)
OK, that's an interesting take. Testing the limits of that theory, does that mean that, if I merely post plausible-sounding author and publisher names (and perhaps page number), I have done all I need do to enable readers to find the supporting source text (my WP:V duty)? The reader may take my insertion in good faith but still claim that it is insufficient for meeting the bar of findability because (assuming good faith) the inserter might be mistaken. A full quote, unlike a name and number or two, forces a conclusion of either unmistaken good faith or (if proven false) bad faith.
Or are you perhaps saying that any well-formed citation always passes the bar of findability due to good faith, and that the question is moved to the court of reliability instead? (I.e., if inserter claims author A says X, and I find no evidence of existence of author A, may I delete, or must I go to RSN with the claim "author A is unreliable because not provably existent beyond inserter's good-faith assertion"?) That doesn't sound right either. I think the minimum standard we developed is a good definition of the point at which the inserter should no longer be questioned about being mistaken, and at which good or bad faith can be judged. I'm not too much against putting it at WP:CITE if it will stay there. JJB 18:48, 6 June 2008 (UTC)
Yes, I am all for a more complete specification of how to source and cite things, but that should be guideline material. This policy and NOR say that it should be verifiable. How it is verified is a more subtle matter, potentially with lots of special cases, exceptions, IAR possibilities... in other words, guideline stuff. Too much of that will actually weaken this policy page by getting lost in the details, not strengthen it. As to the substance of the argument I don't have any real preference but as in other areas I urge you to think of all the far corners of the encyclopedia where this might apply...in music articles you have covers of out-of-print rare albums; in other places things may be cited to speeches or conference papers, what have you. A preference, even a requirement, that something be readily and speedily findable makes some intuitive sense, but how much will that really help the encyclopedia if in practice nobody is going to actually go through the exercise? I like JJB's point that the citation should be specific and believable enough that accuracy and good faith would not reasonably be questioned. It's not so critical that someone does actually check the source, but that they could. Wikidemo (talk) 19:13, 6 June 2008 (UTC)
(ec. @JJB) "If I merely post plausible-sounding author and publisher names (and perhaps page number), I have done all I need do to enable readers to find the supporting source text." Yes, in a sense. Title, publisher, author, date—these are basics. Of course, you could lie—nothing in this policy can change that. "A full quote, unlike a name and number or two, forces a conclusion of either unmistaken good faith or (if proven false) bad faith." Wrong. Totally. Assuming the questioner doesn't have access to the inserted source, they could be lied to about the content of a quote as easily as about an author or publisher name. If they do have access to the source, the point is moot: they can simply check for themselves. Demanding a quote makes no concrete difference in determining reliability. Marskell (talk) 19:19, 6 June 2008 (UTC)
In theory, I am a great fan of the idea that citations should contain as much information as possible... ideally, they should contain Title, Author, Publisher, date of Publication, page refs, ISBN or other Index Number...etc, etc, etc. However, I think all that should be discussed at WP:CITE and not spelled out here. Here we should just say something brief like: "when citing to a print publication, provide as much publication info as possible (see WP:CITE)". As for quotes... I actually dislike quotes in citations. It is far two easy for a POV editor to take a sentence or two from a source out of context, and end up with a quote that seems to say something very different than it would if read in context. I would prefer that if someone doubted a print citation, they actually go to the library, obtain the book, and see the text being cited in context. Blueboar (talk) 19:32, 6 June 2008 (UTC)
I don't disagree. As above, I think the policy should demand as much publication info as possible, while leaving specifics to guidelines. Marskell (talk) 19:37, 6 June 2008 (UTC)
OK that sounds like consensus for returning to a short sentence and letting the rest thrash out in discussion like WT:CITE#Findability. For instance, The citation should state, as clearly, fully, and precisely as possible, how a reader can find the source text. We already have the link to WP:CITE atop this section anyway, so if anyone wants to go down the rabbit trail of "how easy must it be to find" we direct them there.
Before Marskell concludes so quickly that I'm totally wrong, it would be good to consider my point: that if an inserter only provides an unfindable author/publisher, it is still possible to overrule the insertion in good faith as unfindable, if the inserter might be mistaken; but if an inserter provides an otherwise unfindable 50-word quote, the insertion can no longer be overruled in good faith as unfindable, and we move to discuss other topics like its reliability. Marskell's valid observations about lying might relate to reliability, but not to findability, which can be audited straightforwardly. JJB 19:59, 6 June 2008 (UTC)

[edit] instruction creep, quote from Wales

Someone elsewhere started using this quote from wales, and I see it has now popped onto this page. Although accurate referencing, never mind truth, is not required of policy pages, this quote is misrepresented. Wales might really believe that wikipedians should adopt an aggresive policy of deleting on sight any information which does not have a direct attribution, but I hope to goodness he does not. The actual context of the quote seems to imply he is talking about biography, where special rules may reasonably apply because of the sensitivity of information about living people.

This encyclopedia was created by people merrily typing away about stuff they know but almost certainly could not reference without further work, or maybe not at all because they don't have the resources. Policies like this are being interpreted that nothing may remain on wiki without a cite right next to it. This is a fundamental abuse of the system by editors, and is being encouraged by statements like this. This quote is exactly the sort of source material which screams for further research and explanation before allowing it anywhere near a page. Placed like this it is begging to be interpreted as: history professor comes and write a learned treatise on his lifelong pet subject. Joe bloggs comes along and sees no attribution and wipes the lot. Joe may read it, think how interesting it is, probably dead on the mark, but he doesn't approve of unreferenced text so he wipes the lot because it is just random speculation and Wales supports his deleting anything without a ref.

This diametrically opposes how this encyclopedia grew to this size, and indeed came to be widely respected. By having unsourced but nonetheless authoritative texts created by anonymous editors who happen to know something about a subject. It only remains authoritative because such people continue to contribute. Attempts at stiffening policies like this, which are going on all the time, are driving this encyclopedia away from the mechanism which has created it. Please remove this quote, it is not appropriate here. Sandpiper (talk) 06:33, 6 June 2008 (UTC)

I generally agree but we have not gotten consensus for it. First it is observed that he does speak of "all information" so it is suggested it might be better if removed from BLP and left here. Second it is observed that he excepts when "it can be sourced" so it is suggested he does not argue for overdeletion, but for deletion when the material can be expected not to be sourceable. Thus status quo has reigned. Actually, the quote is both overstated and ambiguous. I have attempted to replace it with a better paraphrase of what people here think the true policy to be, but no version has jumped out as a clear leader. My current proposal at #Speculative text is, at the "may be removed" clause of WP:BURDEN, to add or replace the policy text as "Unsourceable low-quality text, questionably inserted, should be removed unhesitatingly". While this gets closer, it is still a bit ambiguous and people object to it on that ground without realizing that the familiar quote is even more so. What would you say the practice of "aggressive" deletion ought to apply to, if anything? It certainly applies to vandalism and attack BLPs. JJB 14:57, 6 June 2008 (UTC)
I would not limit it to any particular category of article or type of statement ... If there is an unsourced statement that strikes someone who has a solid understanding of the subject matter as being "speculative, I heard it somewhere" information it can and should be deleted. Admittedly, this is always going to involve a judgement call... and yes, every editor is going to draw the line a bit differently. Sometimes this will result in a statement being deleted that others think should not be deleted... but that is what discussion pages are for. If it turns out that the statement is not "speculative, I heard it somewhere" information it can always be reinserted (hopefully with a citation, so it will not be deleted again by someone else). To be honest, I like the fact that the Jimbo quote is a bit ambiguous... it gives us wiggle room in all directions. Deletionists can delete... inclusionists can include. Eventually a consensus is reached. Blueboar (talk) 17:22, 6 June 2008 (UTC)
Um, if Jimbo's quote is a wax nose, we don't need it. Both sides already have license (WP:GFDL in fact) to insert and delete as they judge fit. What we're looking for is a clearer, more professional standard for making that judgment call than calling it "some sort of random speculative 'I heard it somewhere' pseudo information". If I judge that it's unsourceable (or unsourced), low-quality (speculative), and questionably inserted (heard-somewhere), then (as you seem to admit) that's the category and I should delete unhesitatingly (aggressively), unrelated to what other editors judge. JJB 18:56, 6 June 2008 (UTC)
Partly correct... When faced with unsourced information, you have to ask several questions. One of which is: "Do I think it likely that this is accurate information that simply needs to be sourced, or do I think it is random speculative 'I heard it somewhere' pseudo information?" If, in your best judgement, your answer is the former, then you should leave it (with a citation request). If your answer is the latter... then indeed you should delete. That said, if that deletion meets with opposition, if another editor comes along and says, "wait... this isn't random speculative pseudo-information", then you should go and discuss it further on the talk page. What is wrong with that? Blueboar (talk) 19:49, 6 June 2008 (UTC)
Its nonprofessionalism. What does it mean in grammatical English? JJB 20:02, 6 June 2008 (UTC)
As I said above, my reaction to it is immediately to seek further information, because it is not clear what it means. It really isn't acceptable if half the editors here think it means one thing, and the other half think something else. This is not supposed to be a page for fudge but a definitive statement. It says 'unless it can be sourced'. Did Wales mean 'unless it could be sourced', or did he mean 'unless it is sourced within a week'. The inclusion of the word 'aggressively ' implies something should be done at once. Blueboar seems to feel that a random unattributed sentence should either be given a {fact} tag, or be deleted. I think it most likely should be left absolutely alone, like all the other unattributed sentences in wiki which are maybe 90% of the total. No action should be taken unless there is doubt about it.
If something is 'particularly' true of biography, how much less true is it in other circumstances, and how much should we therefore disregard it in other contexts? What did Wales mean by 'Random speculative I heard it somewhere pseudo information'? Did he mean the history professor in his lunch hour, or the guy who saw a telly program yesterday while having a few beers? How do you distinguish the cases? Under what circumstances did Wales write this? did he think what he was writing carefully, or was he somewhat annoyed, as seems to be suggested by the first line of the original document, so perhaps he did not think before writing. The 'particularly' really doesnt make any sense if action is to be taken in all cases, which again smacks of hyperbole and a not thought out statement, which should not be taken too seriously. If he is not inclined to explain rather better, why is everyone quoting him as a delphic oracle, with just as much comprehensability? If he wants to play god and ruler of wiki, then he needs to explain what he means. If he doesn't, then why is anyone paying attention to this. Wiki is moving away from the encyclopedia anyone can edit because of statements like this. Seems a case of killing the goose which laid the golden eggs because it has fleas. Sandpiper (talk) 21:06, 6 June 2008 (UTC)
All I have to say is welcome to Jimbo's world. The answers to those rhetorical questions are a bit obscure but boil down to this: Jimbo Wales and his words of wisdom do have a unique, and special role in the formulation of Wikipedia policy. Why? That's not hard to figure out given that he created this endeavor. Should it be? That's a matter of opinion but enough people find the quote useful that it is unlikely to be removed anytime soon. I struggle with all of the questions you raise, and have concluded that you simply can't take this quote literally. On the other hand, it is useful as a forceful and cogent reminder of what Wikipedia is all about. Fortunately, we have few fundamentalists around here who actually think that every word in the quote is a specific requirement. Most follow it in spirit, not in the letter. Wikidemo (talk) 21:25, 6 June 2008 (UTC)
That may be true of the sort of people who wrangle here, but I assure you any unclarity in policy will be noted down and used in evidence in a dispute. This similarly applies to points arrising in more than one policy document, where minutiae of difference will be seized upon to prove a point and justify another 1000 lines of argument. Do people find the quote usefull because it confuses? Depending on the meaning attributed to the quote I would pat the guy on the back, or argue fundamentally about how he could go so wrong. Sandpiper (talk) 23:20, 6 June 2008 (UTC)

[edit] Definition of "challenged" ?

I've got a question: Can we get (here of somewhere else in policy) a better understanding of what is meant by "any material that is challenged or is likely to be challenged" ? The problem as i see it is that any wp user can "challenge" any (and EVERY) uncited sentence not because the content is disputable but simply and solely because there is no citation. This ambiguity means that if i say "the united states is a country in north america" that can be challenged unless I provide some kind of citation to support it. Any help or comments on this ? 208.43.120.114 (talk) 23:53, 6 June 2008 (UTC)

We have had people challenge that the Earth is round, and they are sincere in their belief that the world is flat. Yes, we should cite that the world is round, but such obvious facts can be all summed up with the first reference that has any information on the subject. Most sources on the history of the US will mention its geographic location, an inline citation is not needed, but one of the references should establish this. Everyone knows where the US is, but it is entirely possible someone may mistakenly remember a different smaller country being east of a place, when it is in fact west and putting this mistake into an article. These things do need to be checked when challenged. Of course at some point we need to use common sense. 1 != 2 00:12, 7 June 2008 (UTC)

In the example you give the people challenging have a reason ("flat earthers") other than the lack of citation. What im trying to say is why bother including the "challenged or likely to be challenged" language at all ? There are 3 places in the policy that use this phrase and two other mentions of challenged material so i think there must be some special meaning or else we could change all those sentences like this...

  • "This page in a nutshell: Material challenged or likely to be challenged, and all quotations, must be attributed to a reliable, published source."
  • "Editors should provide a reliable source for quotations and for any material that is challenged or is likely to be challenged, or the material may be removed."
  • "All quotations and any material challenged or likely to be challenged should be attributed to a reliable, published source using an inline citation."

I wanna know if merely lacking citations is reason enough to remove material or if there is some kinda criteria for "challenging" material above that threshhold. If the latter then we oughta spell it out for users that lack of citations is not enough to challenge material. If the former then we oughta remove the word "challenge" and related phrases from this policy since it is meaningless redundancy. 208.43.120.114 (talk) 01:34, 7 June 2008 (UTC)

OK that settles it. I just got curious and did a google for this phrase and google came back with

Results 1 - 100 of about 2,270 from wikipedia.org for "likely to be challenged". (0.07 seconds)

Any phrase that is used over 2000 times at wikipedia ought to be defined clearly and it is not. We gotta fix this. 208.43.120.114 (talk) 01:42, 7 June 2008 (UTC)

The rules of Wikipedia already try too hard to define everything, which leads to problems. It's plain and simple English. Ask yourself: is it likely that Joe Average would ask for proof or believe a contrary fact? If the answer is yes, cite the information straight from the start. If someone challenges the fact, provide a reference. It's really not complicated or anything that should require explanation. Vassyana (talk) 04:00, 7 June 2008 (UTC)

I think that this debate is driven by Wikipedia:100,000 feature-quality articles. Yes before 2006 the tendency was to write an article, and maybe include at the bottom of the article a reference or two. I certainly used to do that, but over the last couple of years if I write a new article I usually put citations on every paragraph (unless like the translation of a foreign Wikipedia article like my re-write of the Battle of Ligny). When rereading articles to which I made large contributions in earlier years, I usually look for the references I used and retrofit citations, but this turns out to be time consuming (and not very interesting), because as I did not include page numbers (obviously) it involves lots of re-reading, often of more than one source.

There is an old Scottish joke that goes "When a Scot emigrates to England the average I.Q. of both countries goes up". I think that it is only through the increased use of citations that the average quality of articles will improve. --Philip Baird Shearer (talk) 08:55, 7 June 2008 (UTC)

The difficulty though, is the growing band of people obsessively insisting on citation for ever simpler facts, and the consequent risk of deletion of perfectly good material. Ok, everyone here likes writing for wiki, so if we scrap articles, we get the pleasure of writing them again. But this is in reality a foolish waste of time if there is some objective of expanding content. Sandpiper (talk) 21:37, 7 June 2008 (UTC)
It is only through the increased checking of citations that accuracy will be improved. We've gotten good at asking for citations... but I think there's this terrible belief that the mere presence of an inline citation means a claim is accurate. The citation still needs to actually be checked... until then it's just decor. --Rividian (talk) 18:48, 8 June 2008 (UTC)

[edit] What do we do when there are no non-English sources in an article & an English source contradicts it??

See Siege of Trsat. No sources I can verify, an creator who likes to call battles sieges when they weren't and who has been blockec for copyvios in the past and who tells me on his talk page he has read a lot but can't remember what he read where. His last 'siege' turned out to be a legend. All I can find is [9] "It was last mentioned as a small coastal settlement in 799 when its inhabitants killed the Friulian army leader, margrave Eric, in an ambush on the road near the church of St. Lawrence. The following year Tarsatika was burned down in a raid of revenge, and the surviving inhabitants moved to a more protected hill where they established a new settlement called Trsat." which is contrary to the Wikipedia article. Doug Weller (talk) 10:26, 7 June 2008 (UTC)

I would tag the statement as {{dubious}} and put a note on the talk page saying what you said here. I would also try to find people who might be able to read the original source material and have them assess the use of the source for accuracy and the source itself for accuracy and reliability, and post the results on the article talk page. If the citation is bogus or does not support the use in the article, I would strip the questionable content. You can probably find speakers of those languages on wikiprojects about related countries. Good luck. davidwr/(talk)/(contribs)/(e-mail) 13:36, 7 June 2008 (UTC)

[edit] Self-published and questionable sources about themselves

The section Self-published and questionable sources about themselves says among other things that such sources can only be used when "the material used is relevant to their notability". Shouldn't that be when the material used is not relevant to their notability?? We should not rely on a company's publications, for example, to establish that company's notability, but a company's website and publications can provide all kinds of information about the company and its products. As long as the statements used in the article are uncontroversial and comply with the other restrictions on self-published sources, such references should be acceptable. For example, one might cite a notable company's website as the source for the company's current products. That is often done, but appears not to comply with the policy as written. Am I missing something?--Srleffler (talk) 03:20, 9 June 2008 (UTC)