Wikipedia talk:Speedy deletion criterion for unsourced articles/Archive 1

Comment by Triona

I think that in general this is both a good idea and a good precident, but I would like to see it require a tag be in place for some period of time, similar to how we do with prod, only removing the tag without correcting the problem would be prohibited. - Stephanie Daugherty (Triona) - Talk - Comment - 22:51, 17 October 2006 (UTC)

Two more minor issues. Unreferenced should be defined better, so that articles we'd have to completely gut of meaningful content under WP:V are deletable. A specific tagging template should be made, which will come into use after the addition of this criteria, and will be similar to orphanbot tagging or prod tagging in that it gives the date for deletion if the problems are unresolved. - Stephanie Daugherty (Triona) - Talk - Comment - 23:03, 17 October 2006 (UTC)

Wikipedia does not state clearly which sources are considered acceptable and which are not. Some people don't accept non-English language sources, although on many subjects I'm interested in English-language sources do not exist in principle. The only consequence of this proposal, if implemented, will be the proliferation of fake and univerifiable "sources" appended to the end of the article. Those who is active in mainspace know that such activity becomes increasingly common. If you want to eliminate unreferenced articles, the easiest and most effective way is to create a bot renaming "external links" to "online references". --Ghirla ^-трёп- 08:43, 18 October 2006 (UTC)

• Well, you're right. I don't think this will be a cure-all. But people can game our other CSDs too, by putting in an untrue assertion of notability, by adding a fake fair use rationale or source or license, but the vast majority of pages these affect are ones with editors that add their work and never edit again. There is potential for gaming, and they'll have to go to AfD, but I expect the benefit to outweigh the bad. Dmcdevit·t 18:58, 18 October 2006 (UTC)
  • I don't deny the merits of this proposal: 90% of newly-created articles is cruft. Nevertheless, we should clearly say which sources are considered reliable. Today, I had a lengthy discussion with a guy who is persuaded (citing WP:RS) that directly quoting chronicles in an article about a 10th-century ruler is not appropriate and that only secondary sources (read, biased interpretations by modern writers) are appropriate for Wikipedia. In my opinion, an 11th-century chronicle is a more reliable source than an early 20th-century interpretation. Another example. An article about Chumbo-Yumbo is translated from unsourced Swahili Wikipedia article. Although the translated material might be first-class, the trans-wiki translation should be deleted, no? Or perhaps the Swahili Wikipedia should be considered a reliable source, in the absence of English-language sources on the subject? In short, the proposal needs to be discussed at length, which I'm sort of uncapable now, having just been called an idiot (twice) and "Russian NeoNazi skinhead" (once). Regards, Ghirla ^-трёп- 20:31, 18 October 2006 (UTC)
    • A historian writing in the 11th century is little different, in terms of the nature of the source as a secondary source, from a historian writing in the 20th century. But a wiki is not a good source, and Wikipedia is not a source at all. See User:Uncle G/On sources and content#Evaluating_sources. Uncle G 16:33, 12 November 2006 (UTC)

Comment by Grafikm

Mmm, I have mixed feelings about this. Referencing is clearly an ultra-major problem (as User:Worldtraveller stated on his user page) and I'm all for dealing with unreferenced articles, deleting them if necessary.

However, what bothers me is a very vague definition of reliable sources. An outstanding amount of work is needed to get these world mean something precise. For instance, I never write an article without adding links to places I got information from, as it is a basic principle of WP:V. However, what is a "reliable" source? How do I know that. And on some things, there are no "reliable" sources in an academic way. BZflag, for instance, used to be an FA, yet by definition, it only has links to online forums and stuff? Should we delete it? Heck no.

In short, a lot of discussion and caution should imho be applied :) -- Grafikm ^(AutoGRAF) 11:14, 18 October 2006 (UTC)

How about, once references are added, their reliability would have to be judged at AfD for wider exposure. (Trying to make this like images. Images with no source or copyright info are CSD, while images with possibly defective copyright claims are IfD'd.) Suggest the same procedure here. No references=delete+14; any reference=keep or AfD. Thatcher131 18:35, 18 October 2006 (UTC)

Also, a lot of new articles are created hit-and-run by editors who never show up again. Editors who stick around, even newbies, should have no problem with this. However, we must be especially careful not to bite newbies; the RCP is already pretty rough on first articles. Thatcher131 18:47, 18 October 2006 (UTC)

The intention is to skirt the whole problem of reliable sources or not. That is a gray area, and a content dcision that should be worked out by the article's editors. However, any CSD-patrolling admin can objectively distinguish between an article with no sources and one with them. Dmcdevit·t 19:02, 18 October 2006 (UTC)

100% agree on that one. -- Grafikm ^(AutoGRAF) 19:17, 18 October 2006 (UTC)

Amazingly good idea, I'd give you a barnstar for it if you'd not already got one. Stifle (talk) 13:49, 22 October 2006 (UTC)

Comment by User:Dragons flight

This is one of the things that WP:PROD and WP:AFD are for. Those processes give the author and other interested parties a chance to correct the problem (which they may not have appreciated was a problem). By contrast speedy is intended to deal with urgent problems, while in most cases unreferenced content is not an urgent problem. Dragons flight 05:42, 12 November 2006 (UTC)

• Actually, I've found this impossible. References are erroneously regarded as a "cleanup" issue at AFD and PROD, and nominations for lack of sourcing are consistently shouted down, since they should be fixed, not deleted. This leaves us in the position of deciding to keep articles for which we idea or proof of their accuracy, neutrality, or existence. This is precisely not the reason that things like poor writing are not reasons for deletion, but our current deletion mechanisms don't handle this problem. Honestly, try nominating a recent article for deletion solely because of sourcing, regardless of its notability, and see what yu get. Note, this criterion is also designed to give the author a chance to fix the problem: AFD can't do that, since it ends in a "keep, but cleanup" decision, with no binding enforcement of that. It remains kept even if it never is cleaned up. Dmcdevit·t 06:17, 12 November 2006 (UTC)
  • Well if people look at it and say: "Yes, Wikipedia should have an article on this", then broadly speaking, I don't think it should be deleted. At the same time, I don't generally believe that always passing the buck is a good thing either. If there is going to be some process that deletes works solely for being unreferenced, then I think it needs to be such that it provides a substantial waiting period from the point at which the content was identified as problematic (not the point at which was created). That would give people, who might not otherwise appreciate that there was a problem, a chance to correct it. It would also give helpful minded Wikipedians a chance to peruse those articles looking for things to save. Like {{nsd}} and {{nld}}, it's not that we want to rush to delete the content, it is that we want to identify problems and have them be fixed without lingering on indefinitely. Also, I think "delete unless sources added" is a behavior we should strongly encourage over "keep and cleanup" when dealing with AFD. Dragons flight 07:16, 12 November 2006 (UTC)
    • The struckout bit is because I misread the proposal to be 14 days after creation. My bad. I'll go away now. Dragons flight 07:20, 12 November 2006 (UTC)

Comment by User:Antandrus

My initial reaction is positive. We have been tightening up our standards here for a while now, with regard to things like WP:BLP, WP:CITE, WP:RS, to the point that most of the articles as they were three years ago would just not be acceptable to most of us now. This CSD addition is another step towards overall reliability. As a side note, if you take something to AFD just because it has no references, the probability of its being either deleted, or repaired with addition of references, is not all that high. You are likely to get lots of "no consensus to delete" unreferenced articles flooding through. I personally trust a CSD on this more than AFD.

As long as there is sufficient time given for referencing, I like this proposal. Antandrus (talk) 05:51, 12 November 2006 (UTC)

Comment by User:KillerChihuahua

I support this, with the addition of the tagging and wait period (currently tag and wait 14 days). It seems to me there may be objections that this is too much like Prod; thus perhaps the difference - that sourcing is required, not merely protest - should be more emphasized. KillerChihuahua^?!? 06:59, 12 November 2006 (UTC)

Comment by Dijxtra

My opinion is that Wikipedia needs to be pragmatic at this point. Sure, if we enact this rule, Wikipedia will lose some of it's important additions because they were unreferenced. So, yes, we loose some potential content and contributors. But, what we get is that in another few months when we reach 2 million articles, we will have half a million new referenced articles. Maybe poorly referenced, but we will have a starting point. Which of the two is more valuable for a encyclopaedia?

I have seen a lot of plain text articles which are probably copyvio just copy/pasted into Wikipedia (I've just checked the newly created pages: of 10 new pages 3 were unwikified, unsourced plain text). Sure, even this articles have potential. But, after having 1.5 million articles, do we desperately need those? Or do we desperately need referencing?

Sure, some of the contributors might try to game the system. But, it's not like the system is not open for gaming at this moment. And, we do not see all that much of gaming. Sure, it happens once in a while. But not all that often (at least from my perspective). I do not see why this rule would be an exception. We'll get some false references. But, again, we gain (1 - some/all)*100% of (poorly or not) referenced stub articles.

Some newbies might decide to leave Wikipedia feeling bitten by this rule. But, again, it's not like we don't have AfD or prod or copyvio to bite newbies... this new rule will be just one more on the list.

So, these are the potential drawbacks of the rule. Potential gain form this rule is a complete shift of perspective. I think that Wikipedians need to start referencing their articles. Not just new ones, but the old ones too. When people create a new articles, they always think of wikilinks, of categories and of wikifying. If a semiexpirienced contributor encounters plain text article, he will wikify it, categorise it and add wikilinks. We have to change a state of mind of Wikipedians to accept referencing as another thing you routinely do to articles. And, I feel that this rule is a an excellent way to start. --Dijxtra 11:23, 12 November 2006 (UTC)

You make a great point about the balance and this being in the right direction for the right reasons. I've said for a long time, we don't need just any material anymore, we need good material with references. I'm not sure we could get this through, but it is well worth a shot. It fits well with Jimbo's desire to improve articles over adding new ones. - Taxman ^Talk 04:48, 13 November 2006 (UTC)

Comment by Dalbury

Yes! We need this. We really need to change the culture on sourcing, and this would be a good tool for doing that. One problem with 'prod' is that anyone can remove the tag and the only choice then is AfD or tagging as needing citations. We all know how AfDs go, and we have editors that are insistent on removing requests for citations without providing any, or who insist on putting back material that was deleted because it was unsourced. This proposal will hopefully be a tool for educating editors on the need for good sourcing. -- Donald Albury 12:29, 12 November 2006 (UTC)

Comment by Wknight94

Fantastic idea. One dumb question: is this going to stop at requiring sources which prove accuracy? Or is it going to require sources which prove notability? One of the most common new article types I see are for local companies, etc. which include a link to the subject's web site. One could make a case that the article is sourced and probably accurate but it doesn't go anywhere to demonstrating notability (i.e., no secondary sources). —Wknight94 (talk) 14:07, 12 November 2006 (UTC)

Comment by Lar

Good idea but the devil is in the details. If I put up a big unsourced article, but add one source for one small part of it, have I dodged this criterion? It's got "a source" even if most of it is not sourced. What about if half the stuff in it is sourced? Judgement will be required, and what I like about speedies is that for the most part they are cut and dried. Good idea though. ++Lar: t/c 14:22, 12 November 2006 (UTC)

This is a good bare minumum, though. Once judgment is required, when a source is questionable or the total sourcing is small, then we'll have to deal with it as we do now. But this at least, is a useful addition to the effort. Dmcdevit·t 20:17, 12 November 2006 (UTC)

Lack of references as a cleanup and deletion issue

An important part of our Wikipedia:Deletion policy

An article is only deletable for being unverifiable if both of the following conditions hold:

• The article cites no (supporting) sources at all.
• Reasonable searches for sources on the parts of several editors turn up no sources.

This is for historical reasons.

Yes, lack of references is viewed as a cleanup issue at AFD. This is in part because although we do have an expanding culture of strong sourcing now, we have several years' worth of articles to deal with that were created when we had no strong sourcing culture at all, and articles being created now by editors of long standing who have become accustomed to never being required to cite sources. If we had had a culture of strong sourcing all along, as Wikinews has, then the fact that an article cited no sources could be used by itself as a deletion criterion. But we have to deal with the legacy of not having that culture, which means that it is incumbent upon editors at AFD to check, using a range of search tools, that there are in fact no sources. "Fails WP:V" cannot be synonymous with "cites no sources".

But that doesn't mean that one cannot get unsourced articles deleted via AFD. It is simply necessary to show that the second condition also holds. It is necessary that editors show that, in addition to just reading the article, they themselves went looking for sources and couldn't find any. In other words: It is necessary that nominators and other contributors do the research.

We in fact came close to this proposed criterion for speedy deletion for one class of article with Wikipedia:Criteria for speedy deletion/Proposal/2. That was based not upon whether a biographical article "asserts notability" but whether it contains any source citations. The basis for that was very much the same argument as put forward above. An administrator can far more easily determine from the article content alone whether an article cites any sources than xe can determine whether an article asserts notability in whatever field the person may be involved.

Ironically, with the later adoption of Wikipedia:Biographies of living people as an official policy, allowing for the speedy deletion of "biographies that are unsourced and controversial in tone, where there is no NPOV version to revert to" then if one substitutes the (ironically more widely encompassing) qualification of "currently alive" for the (ironically narrower) qualification of "not provably born more than 25 years ago" the aforementioned proposal is effectively a de facto speedy deletion criterion (it already being a requirement in all speedy deletions to check the article history for prior good versions).

Although lack of references is not by itself a reason for deletion, it is definitely a reason for not creating articles at Wikipedia:Articles for creation. So we've already introduced strong sourcing into article creation.

On the subject of biting newbies, I suggest that always handing out to newbies advice such as User:Uncle G/On sources and content#Always_work_from_and_cite_sources, which explains how if one cites sources one avoids a range of difficulties, from questions of notability to unstable content, works when handed out to the kind of editors that are going to benefit Wikipedia. I most recently gave this advice to Sullivan.t.j (talk · contribs), for example. Xe went through many of xyr past articles, and they now have references and should remain far more stable against major content fluctuations and proof against nomination for deletion.

Similarly, liberal, and insistent, use of the {{unreferenced}} notice works. See Template talk:Unreferenced#Effect_when_used_by_New_Page_Patrol. Uncle G 16:33, 12 November 2006 (UTC)

Comment by John Broughton

I think this is a good idea. I think that for every article "incorrectly" deleted by this policy, a hundred junk articles will be identified. And if an article is really needed, someone else will again create it, with at least one source.

I'd actually prefer something stronger: an automated block to prevent an article being CREATED without at least THREE sources within the article. I just don't believe that (a) there are a lot of needed articles still missing which (b) only an inexperienced/anon editor is interested in creating and (c) if the system warned the person creating an article that it was unacceptable without sources, he/she wouldn't - if the article was really needed - be able to quickly find them and get the article added. This is 2006 (about to be 2007), not 2004 - wikipedia's problem is now much more with junk articles than it is with missing articles. John Broughton | Talk 17:13, 12 November 2006 (UTC)

Could I suggest you have a look at the WP:MISSING project, in my experience finding three sources isn't always easy. In terms of strengthening this proposal, I agree with other editors that after a trial period, the grandfather clause could be removed. What are your thoughts on allowing the creation of a bot that automatically tags new unsourced articles?Addhoc 11:34, 14 November 2006 (UTC)

Excellent point. Yes, I think all new articles created without any source whatsoever should automatically get the tag, so that they disappear in 14 days if not even a single source is provided.

I'd also like an one-day block of any user who removes the tag (if it is in fact user-removable) but fails to add a source. (I have in mind something similar to the 3RR violation.) It would be even better if the system prevent tag removal until a source was added, of course. John Broughton | Talk 15:17, 14 November 2006 (UTC)

If that becomes a problem we can always create something like the {{drmafd}} series of warnings. When I apply those users normally get the point pretty quickly. --ais523 15:23, 14 November 2006 (UTC)

Also, the existing {{drmspeedy}} tag could be used, this allows editors to be warned prior to being blocked. Addhoc 15:43, 14 November 2006 (UTC)

I strongly disagree with this. Automatically blocking people for "not the right content" (as opposed to just vandalism) and blocking people without warnings (unless they've already been blocked for the same thing) will get people to start accusing Wikipedia of elitism and strongarm tactics. "Wikipedia's problem is not with missing articles" - if you're in the main WP editor demographic, maybe. See WP:CSB. ColourBurst 15:42, 17 November 2006 (UTC)

Practise, principle and extent

When considering a new CSD it is worth wondering how it will be misapplied, for it surely will be. It might go something like this: X writes a shortish article, and does not include any sources for they do not know they should. In short order, the obligatory tag (presumably pink, black, bold, italic, blue, linked, uppy-downy linky, block-for-removing, red-octagonal stop hand tag) is applied. Huh, they think, and copypaste a link from e.g. a Google search on the end of a sentence somewhere, but leave the tag. Eventually, this minor article passes the 14th day guillotine and an admin, largely ignorant in the field of the topic comes along. Sees the link, sees that it is with reasonable probability a link to a forum, blog etc. Decides that is unreliable, and by a misguided application of the new WP:CSD and (when asked for reasoning) WP:IAR, deletes it. Hmmmmmm. It's not long before there is a de facto "...and the source must be reliable" clause in the policy, in a twistedly descriptivist way.

This entirely plausible scenario has to be balanced against the principles of the thing. Wikipedia articles are little use to a reader without their sources (even if the reader is unlikely to check the sources if they just want a quick answer). Is it really true that no article is better than one without a source? Or was that just Jimbo on the mailing list doing what he does on the mailing list? I suppose this relies on the assumption that in some sense Wikipedia probably has most all the articles it ought to have (it really doesn't, though, particularly in specialist areas) and that we can therefore justify the removal, rather than the editing of poor work. I'm personally not sure that I'm persuaded of that balance yet.

Extent: some things don't need references. 1+1=2 doesn't, and neither does the entirely less obvious $\int x\,dx=x^{2}/2+c.$ (WP:V used to be congnisant of the point but has been masticated to pieces.) It is entirely possible that an entire shortish article could be written out of facts that are fundamental to a field and really only need a section for a "bibilography" rather than a set of references for things like "grass is green". Is the argument then that we already have all, or nearly all, the articles like that?

In short, I haven't made up my mind on this yet, but I really want to stress that is vitally important to consider factors relating to the benign ignorance of many editors, old and new, and to consider that the success of this project is due, at a fundamental level, to the very low bar to lending a hand. Every raising of that bar should be done with caution aforethought. And not, in the tone of some comments on this talk page, in the mindset of knowing better already because we benefitted from that same low bar to entry. Splash - tk 23:48, 12 November 2006 (UTC)

• That integral most definitely does need a source. Wikipedia should not aim to be less than ISBN 1584883472, where every single article has references. Furthermore, I can look out of my window right now and see grass that isn't green.

  Once again, the chosen examples in fact prove that there are no exceptions to everything. Uncle G 00:37, 13 November 2006 (UTC)

  • I don't really agree with the commentary on your subpage in its entirety, and I do think that the integral is as much an 'exception' as the addition (both statements are elementary once you know how to either add or integrate), because I hope noone would insist I sourced the fact if I used it in part of a derivation in an article. Green grass is an easy target: the alternative statement that "grass is not green" does not become acceptable if I find a news report of brown grass. It is merely a weak statement, lacking in generality, just as is the "grass is green" reciprocal. I am off topic some distance, now, however. -Splash - tk 17:23, 13 November 2006 (UTC)
  • Just looking in Integral finds a source for the integral, http://www.lightandmatter.com/calc/calc.pdf (it's at the bottom of page 51, stated in a slightly different form). As for 1+1=2, try looking at page 4 of http://www.math.umn.edu/~jodeit/course/Peano.pdf. As you can see, it's not too hard to find sources for trivial statements such as these. --ais523 09:23, 14 November 2006 (UTC)
• To address your first point about the probability admin misapplication to include reliability, I intentionally skirted the issue of reliability of sources (the question is "sources: yes or no?") because that's murky ground. Reliabilty of sources is a content issue. An admin has no more discretion over that kind of judgment call, and shouldn't really be making that decision. We might move in that direction in the future, but I'm more skeptical than you that the first people who tryit won't be skinned alive and made examples of. In any case, as Radiant! suggests we could easily put a clause in to specify that an article with any source at all should be evaluated by the community at PROD/AFD, not speedied.
  • Yes, and I agree that the lowest bar is the only one that should be applied. However, those who are fans of Common Sense, The Right Thing and related euphemisms are not known for particularly caring about such fine details, and nor are they particularly concerned with being skinned alive. (Witness the speedying of the cookie articles under G11 just recently.) Still, misapplication would remain a problem just about however the wording were worded. -Splash - tk 17:23, 13 November 2006 (UTC)
• You have an interesting point regarding the potential that there could be articles that don't need sources. I would point out that no admin must speedy an article that fits any CSD criterion, but that that they must read the article and then they may do so if necessary. I guess I'm still curious whether, if we assume that fundamentally sourceless statements exist, "It is entirely possible that an entire shortish article could be written out of facts that are fundamental to a field". Could you really construct an entire article that way, and not just statements within an article? Surely an assertion of notability still needs a source in every case? Dmcdevit·t 10:12, 13 November 2006 (UTC)
  • The proposed Wikipedia:Attribution, which is envisioned to replace WP:V and WP:NOR contains several notes to "use common sense". It contains explicit exceptions for simple calculations and trivial deductions that do not present a novel viewpoint. There is consensus that we don't need explicit sources for every facts that only a troll would dispute.
  • I prefer to think of the situation discussed above as an instance where potential sources are so numerous and obvious that citation would serve no purpose. I find it hard to imagine the utility of an article that consists only of such material. Robert A.West (Talk) 15:39, 13 November 2006 (UTC)
    • You are probably right. I am worried about articles created today without sources being ok and kept on their value (which is non-zero whatever Jimbo says) compared with the same articles created 14 days from now having somehow zero value. -Splash - tk 17:23, 13 November 2006 (UTC)
  • Well, probably not, no, it was more a thought experiment to clarify the questions of extent and effect. A halfway-house is Scalar-vector-tensor decomposition in which a reasonable number of the textual statements probably don't appear in the source, and that source, whilst fundamental, is more a token gesture. Would we want to delete that in its earliest form? This raises an important procedural point: the reference that is required need not be inlined. -Splash - tk 17:23, 13 November 2006 (UTC)

Comment by Radiant

Given Jimbo's request that we focus on quality rather than quantity, I believe this is a good idea. It shouldn't be too hard for the creator of any article to find a single source somewhere. Personally I would remove the grandfather clause, because it's overly judicial and it's already covered by the fact that a page must be found and tagged, and because I'd prefer not to get people arguing that something was "out of process" when its intent was clear.

However, there has been a recent controversy at WP:RS debating what exactly constitutes a "reliable" source, and there has been discussion at the new WP:ATT about how articles about fiction can be attributed. I think therefore that it's wise to make clear in this criterion that any reference counts, and that AFD must be used if it is disputed whether a reference is reliable - just like for A7, any assertion counts, and disputed assertions are thrown on AFD. (Radiant) 09:48, 13 November 2006 (UTC)

I agree, the issue of reliability should be skirted entirely. Aside from the ambiguity of the concept, I'm also wary of giving admins the authority to make deletions based on content judgments of source reliability. Admins have extra tools, but no more power when it comes to content decisions. Dmcdevit·t 10:18, 13 November 2006 (UTC)

Note: See also WT:CSD#Suggested BLP Criterion. (Radiant) 11:53, 13 November 2006 (UTC)

Agreed with Dmcdevit on this point (and I wholeheartedly endorse the proposal as written). We already have mechanisms for dealing with sources which may or may not be reliable. A lack of sources is never reliable. Mackensen (talk) 12:28, 13 November 2006 (UTC)

Yes. That keeps it simple and clear. This proposal is not a cure-all, but it should make the job of getting Wikipedia properly sourced a little easier. -- Donald Albury 14:29, 13 November 2006 (UTC)

Agree with Radiant, a good idea and in due course the grandfather clause should be removed. Addhoc 10:50, 14 November 2006 (UTC)

Comment by User:Robert A West

Ignorant of this proposal, I proposed something of the sort for the specific case of articles about living persons. While the current BLP phrasing allows obviously defamatory articles to be deleted, any unsourced biographical information can be harmful. In biographies, we use unverifiability as a proxy for determining what information is private. If it has been published in a reliable source, we use it on the grounds that it has been vetted for both truth and privacy concerns.

Consider a few examples. Assume in each case a colorable claim of notability, thereby evading CSD-A7 as a reason to delete. The harmful information in each case has no reliable source, either because it is not true, or because the subject has successfully kept the matter private.

1. A GLBT activist is biographied without source as an "Active volunteer for the Family Values Coalition." The assertion is not prima facie defamatory, as the FVC has many proud volunteers, yet the subject would rightly regard the falsehood as harmful.
2. An attorney, not a public person, worked her way through college as a stripper. In the absence of evidence that she performed illegal acts, this is not a crime, and it can be argued is not defamatory. Nevertheless, she reasonably feels that this information would scare off certain clients.
3. An unmarried man is biographied as married. The assertion is not defamatory, but I think everyone would understand the potential for inconvenience.
4. Person X has a biography that lays claim to the actual accomplishments of person Y. Person Y has been harmed without even being mentioned.

I don't have evidence that Wikipedia is being used for mischief. I know without a doubt that it is being used for self-aggrandizement, which is just the flip side of the same impulse, and in the last example is itself harmful to a living person. If we want to apply this policy gradually, we could quite sensibly apply it to bios first.

That said, I am heartily sick of seeing AFD votes, "Keep. All this needs is a reliable source, which I am sure must exist." The person with the best clue where to find a source -- if one actually exists -- is usually long gone.

Whatever we do in this regard should generate a warning when creating new articles. "Articles with no sources may be deleted. Articles with unsourced defamatory content may be deleted without warning." Robert A.West (Talk) 15:30, 13 November 2006 (UTC)

• If you are "heartily sick" of bad AFD arguments, then read Wikipedia:AfD Patrol and start doing something about them. For every "Keep. All this needs is a reliable source, which I am sure must exist.", point out that that argument does not cut the mustard when it comes to countering the assertion that an article is unverifiable, and that the only counterarguments are sources, sources, sources. For every "Delete. This article cites no sources." point out that the editor needs to actually put in the effort of doing the research, because articles are only deletable for being unverifiable if both they cite no sources and reasonable efforts on the parts of editors to turn up sources come up empty-handed.

  Expanding CSD criteria is not the way to get people to make AFD arguments that conform to our Wikipedia:Policies and guidelines. Uncle G 12:05, 14 November 2006 (UTC)

• • I already try to do this in my own way, but I was unaware of the AFD Patrol. Thank you for pointing it out to me. Robert A.West (Talk) 17:41, 14 November 2006 (UTC)
• Comment Robert, I know exactly where you're coming from. Right now, we have a bio article written by the subject. We deleted it due to COI and now she has reposted it through a friend. In the article it asserts, by name, that a person sexually assaulted her. Further, she mentions other names of prominent elected officials. See the debate Wikipedia:Articles for deletion/Barbara Biggs. I ask myself, if we allow people to pop up and write this kind of stuff confirmed only by their own statements in an interview, what kind of litigation future does the wikipedia project face? How about the family of the people she accuses?Alan.ca 08:41, 10 December 2006 (UTC)

Comments by Mangojuice

I like the idea. A couple of notes: I wouldn't want to see people trying to get articles deleted by using this criterion along with removing sources (even ones they feel are unreliable). This should be made clear somehow, but I'm not sure of the best way. Also, why 14 days as opposed to any other amount? I would figure 7 would be enough to find one source. Finally, this should come with an automatic undelete clause, sort of the way WP:PROD does. If any user wishes an article deleted under this criterion to be undeleted, and has a source, any admin should be willing to undelete it. Though work without sources isn't valuable enough to keep around anymore, it should be made available afterwards if people start working on it. Anyone who deletes a page under this criterion should watchlist it, and if a new version appears with sources, they should make the deleted version available. Mangojuice^talk 15:57, 13 November 2006 (UTC)

Be wary of adding that much instruction creep to the proposal. Keep the proposal short, or it won't be sweet. CSD already comes with an auto-undelete (after a sanity check), it's just that fewer CSDs actually pass said sanity check. Consider your new version of the proposal:

1. Check if the article is 14 days old.
2. Check it hasn't been vandalised.
3. Check it has no sources, nor any that can be reverted to.
4. Delete the article.
5. Watchlist it.
6. Check every edit to it subsequently.
7. Delete if no sources.
8. If sources, undelete history.
9. (implicitly)Leave a talk page note somewhere, probably on both the article and user pages.

It approximately doubles the load. -Splash - tk 17:27, 13 November 2006 (UTC)

Discussion on WT:CSD

Just highlighting what Radiant noted above, discussion of this is now basically live on the criteria for speedy deltion talk page. Might as well take care of it now and keep it in one place there. - Taxman ^Talk 21:18, 13 November 2006 (UTC)

Perhaps this page should be moved to a Wikipedia space page, where there can be more directed comment from the entire community. The CSD talk page, at any rate, seems to be contemplating the idea of such a proposal, without ddressing an actual proposal. Dmcdevit·t 23:41, 13 November 2006 (UTC)

Yes, I'd think that would be appropriate unless you think more discussion is needed before a formal proposal is launched. How were the other CSD expansion proposals named? Just do similar after updating it for any improvements suggested so far and perhaps incorporating options so that any individual drawbacks don't sink the ship. - Taxman ^Talk 23:53, 13 November 2006 (UTC)

For lack of a better idea: Wikipedia:Speedy deletion criterion for unsourced articles. Dmcdevit·t 08:38, 14 November 2006 (UTC)

Comment by Pmanderson

Jimbo's quotes relate specifically to unsourced articles on living persons; one of them explicitly. We should indicate this.

I agree that this is a well-meant proposal; but I foresee, if it is implemented, a repeat of certain recent unpleasantnesses. All it takes is a handful of admins to construe this as "Delete all unsourced articles now; Jimbo said so!" to produce vast deletions. If one of them belongs to the "The only real citations are in-line citations" movement, it will be much worse. We should take steps to avoid this misunderstanding before doing anything to make this discussion more official. The present notes are a good start, and confirm that the intention of this proposal is entirely sensible. Septentrionalis 00:16, 14 November 2006 (UTC)

The first Jimbo quote may be less on-topic, but he states "This is true of all information, but it is particularly true of negative information aboutliving persons" so it's not as if he's misquoted, I think. You're free to edit the page, by the way. It seems there is already developing support on this talk page for the notion that admin overextensions are not welcome. Ambiguity in the form of any source at all that an admin questions should be taken to AFD or PROD. Dmcdevit·t 08:43, 14 November 2006 (UTC)

I believe the second one is part of the same discussion, isn't it? In any case, this is what lawyers call obiter dictum, not a decree.

As for editing this: thank you. I have on the quotes; I would have edited more sweepingly if I saw how to clarify this further. Septentrionalis 19:19, 15 November 2006 (UTC)

Comment by Shreshth91

I fully support this idea. With increase in the size and popularity of Wikipedia, we'll be overrun with poorly referenced articles, which may be hoaxes, and sorting through them would take mind-numbing amounts of time. I also agree with the specifics spelled out in the proposal, and the 14 days grace period. --May the Force be with you! Shreshth91 08:57, 14 November 2006 (UTC)

Comment by ais523

I proposed basically this on WT:CSD after reading the discussion (link permlink to current version), without even realising the benefits to AfD closure. As Robert A West says on the CSD talk page, the person who writes an article is usually in the best position to source it, and if there is a deletion system in place we can put warnings in interface messages telling people to provide sources if they don't want the article they created deleted.

By the way, I've listed this page on policy RFC to get more feedback about the proposal (in a case like this, it's pretty important that the wording's reasonable to begin with).

We still need an article tag and a usertalk warning template; I'd be willing to have a start at creating these so that there can be a proposal with all the details filled out. --ais523 09:12, 14 November 2006 (UTC)

Implementation?

How is it going to to be implemented? If it's "14 days after tagging", not "14 days after creation", we need a mechanism of keeping track of the waiting time. That means: Either a system of dated tag categories, like the dated prod system, or a list page, like the copyvio system. Either way, it will need rules: Who decides when it's legitimate to delist an article when sources have been provided? What if the creator adds a source but it's blatantly inadequate? Basically, this seems to be introducing not just a new criterion to the CSD system, but a new (fourth? fifth?) deletion process. From the perspective of the tagger, I'm a bit afraid it's not going to be "speedy" at all: learn yet another new reporting system with its own tag templates and rules and everything, tag the thing, keep a watch on it for 14 days to see if the response is adequate; if not, back to square one. If I wanted to get rid of a bad article "speedily", I might still go for an AfD right away. Fut.Perf. ☼ 09:34, 14 November 2006 (UTC)

• None of your quesions really seem new: the process already esits. It will likely work as we'd naturally imagine it to work on a wiki. Anyone may tag an offending article they notice. Anyone may add a reference and remove the tag, or remove the tag if a reference is already there. If anyone wrongly removes a valid tag, it should be reverted, just as removing any valid, say, {{db-a1}} without fixing it would be reverted. Perhaps a more apt comparison would be to I4: anyone can remove the tag when an image source is added, but if they remove it without doing so, they should be reverted. A "blatantly inadequate" source is a content judgment, and should be dealt with using the current means (AFD/PROD/cleanup), but this CSD would nevertheless catch many with no source at all that would languish in the current system without anyone watching them. Even tagging them without watching is better than what happens now (typically, nothing). Dmcdevit·t 09:55, 14 November 2006 (UTC)

I've had a go at an implementation: {{nosourcedel}} (which presumably needs a spiffy abbreviation (nsd is taken) if this becomes accepted); it's based on prod. As always, feedback is welcome. --ais523 10:39, 14 November 2006 (UTC)

• Thanks. Maybe the naming could be in line with the other {db-} templates? I'd go for {subst:db-nosource}. And perhaps make the template a little bit smaller and less obtrusive? After all, we might be seeing them all over the place in a short while... Fut.Perf. ☼ 11:04, 14 November 2006 (UTC)
  • I thought about db-naming the template, but all the db-tags seem to be for immediate deletion (I5 isn't immediate; compare {{Orphaned fairuse not replaced}} (the 'timeout tag') to {{db-i5}} (which causes deletion after the timeout)). As for being obtrusive, it's quite important this isn't mistaken for a cleanup tag, and the obtrusion is what's wanted (WARNING! WARNING! THE ARTICLE WILL BE DELETED IF YOU DON'T SOURCE IT!), in the hope that a user who's created an article will see it if they miss the other warnings (it could be applied very quickly as part of newpage patrol). --ais523 11:10, 14 November 2006 (UTC)
    • Template looks ok to me, would prefer a name such as {PROD-NS} or {db-ns}, but that isn't very important. Addhoc 11:25, 14 November 2006 (UTC)
      • I'll put a redirect in from {{nsdel}}. --ais523 12:25, 14 November 2006 (UTC)

If the problem to be addressed is articles without any cited sources that "languish in the current system without anyone watching them", then the solution is to make Category:Articles lacking sources more usable by those who would like to patrol unreferenced articles and watch them, by having a 'bot do the same categorization by date with {{unreferenced}} that Pearle (talk · contribs) does with {{cleanup}} and MarshBot (talk · contribs) does with {{linkless}}. Uncle G 12:25, 14 November 2006 (UTC)

• No, the problem is that the person best positioned to provide sources (the person who wrote the article) often doesn't, and it's hard to find the sources later. --ais523 12:27, 14 November 2006 (UTC)
  • No, that's not what Dmcdevit wrote above. Please read it again. The solution to that problem is use of {{unreferenced}} by New Pages Patrol, as demonstrated by Template talk:Unreferenced#Effect_when_used_by_New_Page_Patrol. Uncle G 13:28, 14 November 2006 (UTC)
    • The problem I'm dealing with is different from Dmcdevit's, but we're both supporting the same solution. In response to your NP point, it would look a bit silly if MediaWiki:Newarticletext (or some other appropriate interface message) said 'Please give sources for the article you create, or you risk having an {{unref}} tag placed on it by New Page patrollers'. The problem I'm addressing was flagged up on WT:CSD. --ais523 13:36, 14 November 2006 (UTC)
      • You're creating a straw man. MediaWiki:Newarticletext doesn't need to say that. Moreover, as I have already pointed out, the solution to your problem is use of {{unreferenced}} by New Pages Patrol. It works in practice. I linked to several examples examples of the process in action. Uncle G 01:40, 16 November 2006 (UTC)

Undeletion

The proposal is a good one. I'd like to add that I hope admins will be extremely sympathetic to requests (particularly from newbies) to undelete articles deleted under this criterion if the editor requesting undeletion undertakes to add their references shortly after undeletion (and/or says what sources they used in the article that they'd like to add to it as references). jguk 10:58, 14 November 2006 (UTC)

Seconded. In fact, that ought to be added to the proposal; I'll go and do it now. --ais523 11:01, 14 November 2006 (UTC)

Thoughts from badlydrawnjeff

Completely unnecessary. The eventual creep will then move to sources that may have questionable reliability, then to sources that may not be readily verifiable, and judging by how poorly newer CSD criteria such as A7 and G11 have been handled, there's no way we can trust that this can be handled properly.

We have prod and AfD for a reason. Let's stick with them. --badlydrawnjeff talk 12:16, 14 November 2006 (UTC)

I disagree that this is 'completely unnecessary'; as for the creep problem, careful watching of CSD shows that creep (in criteria such as G11) tends to go in the other direction, towards lenience. --ais523 12:24, 14 November 2006 (UTC)

If that's the case, I haven't seen evidence of it. --badlydrawnjeff talk 14:18, 14 November 2006 (UTC)

I also disagree that this is 'completely unnecessary'. I'd argue the opposite, that it is completely necessary if Wikipedia is ever going to become a reliable source itself (it certainly isn't at present!). Nor will there be a creep to speedily delete articles with source with questionable reliability. Where articles have questionable sources, they will be discussed - if they are vindicated, or if better sources are found, the article will stay. It will never be a CSD issue, but one for sensible discussion. If an article is unable to be justified except by sources found to be dubious then it really has no place in Wikipedia, jguk 13:47, 14 November 2006 (UTC)

No single source is inherently "reliable." We can't expect Wikipedia to surpass any other encyclopedic-style source, which should never be used as a singular source regardless. There has always been creep with speedies and this would be no different. More eyes, less deleiton, more fixing - that's the answer. --badlydrawnjeff talk 14:18, 14 November 2006 (UTC)

At present university undergraduates are often warned not to use Wikipedia as it is not reliable (Wikipedia being specifically singled out). There's no reason why we shouldn't aim to make these warnings a thing of the past. Neither is there any reason why Wikipedia shouldn't become the most reliable encyclopedia - we have enough people working on it, and we're not too far behind Encyclopaedia Britannica already. Wikipedia won't get there with one big bang, but positive moves such as this proposal will get us there in the end, jguk 14:32, 14 November 2006 (UTC)

I don't think this move is "positive," or will bring us any closer. That's the point. Hell, as an undergraduate history student, I was told to stay away from all encylcopedias - they're not meant for any sort of significant reference. Just because we hear Wikipedia singled out in the news doesn't mean that it's not across the board in reality for encyclopedias. The amount of lost content combined with the fact that we simply have no evidence that such a CSD would be handled properly by adminsitrators given the way previious CSDs have been held makes this a huge, huge mistake. --badlydrawnjeff talk 14:50, 14 November 2006 (UTC)

The flip side of this proposal is that everyone creating articles will soon get to know that they always add their references. I imagine this will happen very quickly. We have to allow for undeletions to allow sources to be added later, as from time to time even the best editors will forget to add them, but that's now provided for in the policy.

I'm not really sure what the rationale for keeping entirely unsourced (and therefore unbelievable) articles in Wikipedia. If they can be sourced, source them - it will make a tenfold improvement. If they can't be sourced, why do you (or anyone else) want to read them? jguk 14:54, 14 November 2006 (UTC)

Comment by BlankVerse

Learning how to do proper referencing on the Wikipedia is, quite frankly, a pain in the tochis. The minimum requirement should not be a single reference (by that criteria, if applied retroactively, a big chuck of the Wikipedia would get deleted), but should be, instead, a single External link to a reliable source (not directly connected to the topic if it is a company, person, or school).

The second change should be 14 days after a message is left on the article creator's talk page. ^BlankVerse 16:01, 14 November 2006 (UTC)

I'd agree with IfD-style mandatory Talk-page warnings; as for the 'external link' condition, the problem is that too much discretion would be needed (speedy criteria are meant to be objective and uncontestable). The requirement's not necessarily for a reference in the {{cite book}} sense, but for a statement saying what the source of the information is (just like should be provided with images at the moment). --ais523 16:18, 14 November 2006 (UTC)

Agreed: A link to a relevant webpage is, perforce, a source. I am not sure what distinction is being drawn here. Robert A.West (Talk) 17:10, 14 November 2006 (UTC)

Sure, a webpage is (or may be) a source, but not every source need to be a webpage. There are still people out here who use those old-fashioned things, you know, what were they called, books. A minimum requirement of external links would be terribly wrong. Fut.Perf. ☼ 13:37, 15 November 2006 (UTC)

No, not an external link. This CSD doesn't care about formatting or citation type, it just cares whether the References section is blank or not. If you got it from a webpage, put that in the references section. Night Gyr (talk/Oy) 18:07, 15 November 2006 (UTC)

The policy now requires a reference section with at least one entry. It seems pretty clear that an external link without "ref" tags around it is not a reference/footnote, even though it's a source.

I think that is a mistake. I think the policy should require at least one external link OR one reference/footnote, so that new editors aren't forced to learn WP:CITE. (I think footnotes are important to reduce the impact of link rot, but that's another matter.)

We should start small with a new policy like this - err on the conservative side. If the first implementation works well, then an expansion can be considered. If the policy is too broad, it risks not being adopted at all, or being reversed. Starting small would mean NOT tagging articles that have an external link, even if there is no reference/notes section. John Broughton | Talk 19:27, 15 November 2006 (UTC)

You don't need to use CITE to put a link in the references section. External Links are a "See also," not a reference. Just a URL in the references section would be sufficient. Night Gyr (talk/Oy) 19:38, 15 November 2006 (UTC)

Negative Comment by GRBerry

Keep in mind the basic standards for a criteria for speedy deletion, from Wikipedia talk:Criteria for speedy deletion#Read this before proposing new criteria. The criterion should be objective, uncontestable, arise frequently, and nonredundant. "Consider explaining how it meets these criteria when you propose it."

I think the proposal fails the uncontestable point, which is more fully explained as "it should be the case that almost all articles that can be deleted using the rule, should be deleted, according to general consensus. If a rule paves the path for deletions that will cause controversy, it probably needs to be restricted." I can see a significant number of articles being created that don't have reference, but readily could have them, and probably should not be deleted because the subject meets WP:BIO, WP:WEB, WP:CORP, etc... This is even more of a problem in areas like schools, where we don't even had a standard because there is no consensus on what the standard should be.

I know the current intent is to restrict the criteria to future articles, but sooner or later it will get applied to long standing ones also. My personal example of a historical article that is obviously a subject we should cover and for which references exist, but which lacked references, is Geology. I added the first citation style reference on 1 September 2006, but the article was first created on 2 August 2001. It went just under 5 years before getting a citation style reference, and none of the external links were really references for the article as a whole (though undoubtedly some of them backed up some portions of the article).

Although I believe that sourcing is important (see User:GRBerry#Quality), I don't believe we have moved the culture enough for lack of sourcing to be a basis for speedy deletion. Our policy is that content must be verifiable, not that it must be verified. First move the policy, then establish a CSD. For now, PROD and AFD are adequate. I would start my moving the AFD culture, to establish that if sourcing is challenged during an AFD, sourcing needs to be demonstrated during that AFD for the article to be kept. GRBerry 18:11, 14 November 2006 (UTC)

This appears to me to be a first step in the right direction. Of course, in time, I'd like to see every article properly referenced. That will not happen overnight. Instilling a culture of referencing will, of itself, encourage people to add references to old articles that do not happen. We're a fair way off (years and years) looking at removing all old unreferenced articles. When it happens (as ideally it should) it will need to be done in a way that means WP does not lose lots of content and that concerns such as yours can be allayed. If they can't be allayed at that stage, then I can't see such a proposal succeeding. But realistically we are years away from being in a position to do this. The current proposal stops the current situation from worsening. It should be welcomed as such. jguk 18:40, 14 November 2006 (UTC)

This is a critical misunderstanding of the Verifiability policy: "Our policy is that content must be verifiable, not that it must be verified." . When we use the term "verifiable", we do not mean information that could be verified with a source, we mean information that is able to be verified by readers because it has a source. This is the central point of the verifiability policy: "Editors adding new material to an article should cite a reliable source, or it may be challenged or removed by any editor." So our policy is certainly not, in the sense you're proposing, to require verifiable, but not require verified information. Using this more correct interpretation of verifiability, it becomes clear why the criterion would be uncontestable: it is already our policy that unsourced material be be removed if a source is not provided when challenged. Dmcdevit·t 21:26, 14 November 2006 (UTC)

"The threshold for inclusion in Wikipedia is verifiability, not truth. 'Verifiable' in this context means that any reader should be able to check that material added to Wikipedia has already been published by a reliable source. Editors should provide a reliable source for material that is challenged or likely to be challenged, or it may be removed." (from the current version of WP:V, although this may not be authoritative as it was protected due to an edit war). To me this means that the current policy requires sources to be given for anything that has ever been disputed (which must be quite a portion of Wikipedia by now), and technically speaking the placing of the tag would constitute a challenging of the material IMO (but see WP:IAR; I think it may be justified to discount this argument using it). The AfD point is a good one which was already brought up on WT:CSD; it's possibly worth changing policy to prevent a 'keep pending sources' result on AfD. --ais523 09:02, 15 November 2006 (UTC)

Comments from Pengo: This is the internet.

There are a large number of false or misleading assertions made in the text of this proposal:

False underlying assumptions

Main problem we face has long since shifted from coverage to reliability, accuracy, and neutrality.

The main problem who faces? Where is the evidence of this? Which articles? Which readers? Which editors face this problem? I focus mainly on animal-related articles and the main issue is still coverage, and will remain coverage unless every one of the next 1.5 million articles is about an individual plant or animal species.

Have you read any news coverage of Wikipedia? A very large proportion of it focuses on the fact/perception that on the whole Wikipedia is not reliable. Add to that Carnildo's point that on the whole existing articles don't get referenced and we have a problem. Having more unreferenced content is not helpful. More referenced content is extraordinarily so. - Taxman ^Talk 15:14, 15 November 2006 (UTC)

Sourcing will not solve this problem. The perception is only partially due lack of sourcing, but it also has to do with who edits the encyclopedia. If you think people are lazy about sourcing now, they'll be even lazier to check up on the sourcing. In addition, WP:CSB has some major areas that haven't been filled in yet. ColourBurst 15:42, 16 November 2006 (UTC)

Inherent contradiction

One of the most important efforts in this regard is referencing all the articles we already have. Unfortunately, this is an impossible task, because we are inundated every day with more and more unreferenced new articles that will languish in that state, while more are created, faster than we are referencing, or likely possibly can reference, existing articles. The fact is that an unreferenced article is not helpful ...

There are a number of assertions and hyperbole here, but I'd just like to point out the obvious contradiction made here:

1. You are saying people are creating articles.
2. You are saying a different set of people are adding references to those articles.
3. You are saying that unreferenced articles are basically worthless and should be deleted.

Obviously step 2 above cannot happen without step 1, and step 2 is happening without access to the original source material that the articles have come from. So, it is fair to say, that not only is the unsourced article helpful for creating a sourced article, but, as the referencing can occur without the help of the original author, it seems perfectly clear to me, that other sources can also be found by anyone READING the article. And, therefore, an article is still helpful even when it is unsourced.

People accessing Wikipedia are on the Internet and connected to the World Wide Web. Searching Google is probably how most people come up with their references anyway. We can safely assume readers of Wikipedia can do the same.

Misses the point that the article isn't very helpful until it is sourced, and the person that comes along and does the sourcing could easily have created the article with sources in the first place. - Taxman ^Talk 15:14, 15 November 2006 (UTC)

Scaring away newbies

Requiring referencing is like holding up a big sign to any and all potential new Wikipedia editors saying "FUCK OFF". Editing an article should be easy, writing your first article should not be difficult. Adding references and learning to reference properly is not easy, and should not be a requirement for contributing to Wikipedia. Especially, as explained above, unsourced articles are still helpful.

Perhaps some people think that deleting one good article for every 100 nonsense ones is okay. I don't. Especially when the person who sees his or her genuine article deleted for trivial reasons decides not to waste his or her time on Wikipedia any longer, and doesn't bother with the next 99 articles he or she may have written. —Pengo ^{talk · contribs} 00:27, 15 November 2006 (UTC)

• 'Writing a your first article should not be difficult' - I agree. But it isn't that simple. In the case of most problems (lack of wikification for example), it's trivial for another editor to sort the problem out. In the case of sourcing, it's often only possible to get an accurate result if the sources are given by the same writer who wrote the article; the original writer can provide sources trivially (I don't care whether it's a {{cite web}} masterpiece or a little note saying 'I got this from (name of some book)' at the bottom of the page, and neither does the CSD). If you don't believe me, just look through the WP:AFC archives for anything with a green background, and see how anons are sourcing things. --ais523 08:47, 15 November 2006 (UTC)
• It's not a trivial reason. It's critical to a reference work to have references, and the Verifiability policy is one of our most important core policies. And it's not saying f-off, anyone that knows what they are talking about will know adding references is important for a reference work built anonymously over the internet. - Taxman ^Talk 15:14, 15 November 2006 (UTC)

Another fugly message

Now besides having:

• Do not copy text from other websites without permission. It will be deleted.

Appear on every edit screen,

You will also have:

• You must include the source of your statements, or your text will be deleted.

Doesn't make Wikipedia very fun.

"Encyclopedic content must be verifiable." can be replaced with "Encyclopedic content must provide its sources" at a minimal extra character count; providing sources is a strictly stronger requirement than verifiability, so the old requirement can be dropped from the line. --ais523 08:43, 15 November 2006 (UTC)

We're not here to be a party. We're here to write an encyclopedia, and citing sources isn't exactly hard. Night Gyr (talk/Oy) 18:09, 15 November 2006 (UTC)

No, you see. Actually, we're each here because we enjoy it. Easy to forget this in the sort of overseer-with-whip mode some people operate in. We're here writing an encyclopedia because we find it fun, each for our own meaning of fun. Make Wikipedia un-fun wouldn't achieve a great deal, after all. Splash - tk 22:11, 15 November 2006 (UTC)

Ok, how about "Source what you write :)"? Night Gyr (talk/Oy) 22:13, 15 November 2006 (UTC)

Carrot, not Stick

Rather than deleting whole articles because you don't like the lack of references, how about another solution that instead encourages referencing another way.

How about some more creative solutions:

• Write some docs on how to find sources, on a topic per topic basis
• Replace ugly stub notices and "THIS ARTICLE IS UNREFERENCED" noticed with "Help us find references for this article" with some links that are actually helpful.
• Brainstorm other solutions

—Pengo ^{talk · contribs} 00:27, 15 November 2006 (UTC)

• Excellent idea. I would certainly support changing the stub and unsourced message to something like that. That way we can come at it from both sides, drastically improving the sourcing of new articles through this CSD and aiding the sourcing of existing articles through a more useful and friendly message. Not to say the latter will be a revolutionary change in effectiveness, but we need to try everything. - Taxman ^Talk 15:14, 15 November 2006 (UTC)
  • I suspect from what you have both written that neither of you have actually read {{unreferenced}}. It already says "You can help Wikipedia" and it already links to helpful places. Uncle G 11:55, 16 November 2006 (UTC)

On the quotes given

"If an article topic has no reputable, reliable, third-party sources, Wikipedia should not have an article on that topic."

This does not apply to the vast vast majority of unsourced articles. Most articles do have reputable, reliable, third-party sources, they simply don't mention them.

"Any edit lacking a source may be removed".

The operative word here is MAY, not MUST.

Jimmy Wales has stated about articles on living persons, which are especially in need of accuracy and sourcing, that "Zero information is preferred to misleading or false information".

This is only about one topic (living persons), and Jimmy Wales is just trying to cover his arse because he's tired of seeing potential law suits. Moray eels being unreferenced is hardly going to cause a law suit.

—Pengo ^{talk · contribs} 00:38, 15 November 2006 (UTC)

Wonderful closing. How about substitute "Jimmy Wales is just trying to cover his arse" with Jimbo is rightfully protecting the Wikimedia foundation (ie the entity that pays for the servers and is responsible for making sure it keeps running) from potential lawsuits. But lawsuits aren't the only issue, information quality is. Whatever we can do to increase information quality accross the board is the most important thing we can do. Also you're misunderstanding WP:V, if the article doesn't list a source, then for all practical purposes it doesn't have one. - Taxman ^Talk 15:14, 15 November 2006 (UTC)

Yes, but be careful. The quality of the information isn't improved if it is deleted. Perhaps the net total quality of the entire encyclopedia is increased, by virtue of dividing the static total quality between fewer articles, but that's not the same thing. It is not unreasonable to be pretty certain that some perfectly correct information will be removed by this CSD; again that's not likely to increase information quality, so much as result in higher average reliability per piece of remaining information. Splash - tk 22:14, 15 November 2006 (UTC)

Is unreliable information good information? Night Gyr (talk/Oy) 22:15, 15 November 2006 (UTC)

Uh, I didn't mention "good information". I talked about correct information. Should we delete correct information? You're going to ask something like "what legitimate reason is there for not sourcing correct information", but that question expires at the 14 day guillotine: the questio is, at that point, should we delete correct information? Splash - tk 22:32, 15 November 2006 (UTC)

We don't know if it's correct unless we have a source, though. Anyway, I originally proposed three months at template talk:unreferenced. I'm not a huge fan of 14 days, but the line has to be drawn somewhere short of forever. Night Gyr (talk/Oy) 22:40, 15 November 2006 (UTC)

What legitimate reason is there for not mentioning the sources? Night Gyr (talk/Oy) 22:20, 15 November 2006 (UTC)

Comment from Mallanox

I'm not comfortable with this proposal. We have patrols on new pages and on recent edits that should stop anything too disastrous from appearing. I cannot see the need for more things to be deleted and it just makes creating new pages that little bit more arduous. New users will be put off as their work disappears, ok there's a tag for 14 days. Meanwhile they've been scared off and won't look at Wikipedia again. I'm sure it's stated somewhere in wikilore that one should not look to delete where one can improve. Instead of tagging articles, why not use the resource to find sources and citations? Because it's quicker not to? Time to quote J. K. Rowling (sort of): sometimes one must choose between what is right, and what it easy. Mallanox 00:58, 15 November 2006 (UTC)

Reaction from the field

See Wikipedia talk:WikiProject Tree of Life#Wikipedia:Speedy deletion criterion for unsourced articles. -- Donald Albury 01:05, 15 November 2006 (UTC)

Looking positive. Glad to see that editors understand the need for sources and aren't taking it as harshly as some people in this discussion seem to think they will. Night Gyr (talk/Oy) 19:13, 15 November 2006 (UTC)

Two points

Two points to consider:

1. Approximately 80% of all Wikipedia articles are unsourced.
2. Articles have only a minor tendancy to become sourced over time.

--Carnildo 05:07, 15 November 2006 (UTC)

Interesting. I just repeated your experiment (albeit with a smaller sample size -- 30 articles), and found more like 60% as unsourced. Well within the bounds of experimental error for such a small sample, but perhaps influenced by an increasing tendency for new articles to be well sourced? Have you considered rerunning with a new sample of articles, rather than examining the same articles again? JulesH 22:25, 19 November 2006 (UTC)

Killing a flea with a bomb.

That's what this proposal is. Killing a flea with a bomb.

The problem can already be addressed far less offensively by merely editing the articles in question. If you don't feel you have time to source the claims in the article yourself, edit them out. The advantages of this are that it works for ALL Wikipedia articles, no matter the age, gets the unsourced edits off the current revision, yet allows people who aren't administrators (remember us, anyone???) to see what edits may have been perfectly good, but needed sources. If you doubt it, copyedit.

All this requires is a gradual change in the culture of editors rather than adding more bureaucratic hell and rules to memorize. Unfocused 05:26, 15 November 2006 (UTC)

It can be solved by editing the articles. That's exactly what this criterion is about. An unsourced article is tagged and the creator alerted, and it is given 14 days to be edited to fix it. It's no bomb (and unsourced articles are no flea, either). If it can't be sourced in 14 days, than he article itself is "edited out," which is deletion, for an article. It's teh same principle. Of course, we need a deletion criterion to be able to enact that culture change that will get people to find the sources. At the moment, there's absolutely nothing failing an AFD (and even that is unlikely) that can compel sources to be added to an article. The problem with saying "just edit it" is that obviously that is a red herring for most of the articles this will affect: almost all articles on Wikipedia have one or fewer maintainers, and with the amount of unsourced articles, it simply will never be fixed without movement in this direction. What's the "bureaucratic hell," by the way? I can summarize this in a succinct sentence. Dmcdevit·t 05:42, 15 November 2006 (UTC)

You're talking about destroying articles. The rest of us are talking about removing the small bits of unverifiable information. General Sherman shouldn't be an administrator here. --badlydrawnjeff talk 14:25, 15 November 2006 (UTC)

If an article has no sources, then surely all of it is unsourced? The CSD is to delete articles which have no sources (and therefore within which all information is unsourced). --ais523 14:38, 15 November 2006 (UTC)

Another way to define a stub is an article so incomplete that an editor who knows little or nothing about the topic could improve its content after a superficial Web search or a few minutes in a reference library. You'd rather delete that. Flea, meet bomb. I love that analogy, it's perfect. --badlydrawnjeff talk 14:48, 15 November 2006 (UTC)

Either it can already be deleted under CSD A1, or there's some information there. If there's enough information there to do the research, it must come from somewhere (possibly suggesting where to look). If the stub's just half-remembered, possibly incorrect information, a post on Wikipedia:Requested articles would probably achieve a better request than placing it in Wikipedia. To try to get some evidence, I just hit Special:Random until I got a stub and came up with Cowie Hill, Halifax (first stub I came across); and it had a reference which at least proves the information's correct and acts as a starting point. Second stub (although it isn't tagged) is Mirko Derenčin, which has no references. I know nothing about Croatian history, but if this policy was in place the reference that the article was created from would have been given. (I'll just add {{unref}} and {{stub}} tags to it, but it would get sourced a lot faster if this rule had been in place when it was written.) Next page: John Powers (poet). BLP, with no references. All the information here is unsourced, and again knowing the source would be useful. If any of the information here were false, Wikipedia might get into trouble with the media/person concerned, and I have no way of checking it from the information provided. So I can place {{unref}} on it, or {{afd}} on it, under present policy... (/me thinks a bit; /me tries prod). I'd call this situation more than a flea; the only good way to stop this happening is by preventing it at source by requiring sources when the article is made. How would you edit it to correct the sourcelessness only by removing information? --ais523 15:05, 15 November 2006 (UTC)

So now you're advocating misuse of A1? This is why I'm afraid of the misapplication of this possible CSD. To answer your question, I'd do what I'm already doing - either stubbing the article, or finding some sources. I don't even bother prodding or AfDing, I'll assume good faith that it can be cleaned up by someone more knowledgeable. We can all do this in the meantime, and we don't need to lose a ton of articles doing it. A stub is much, much better than nothing at all, and stubbing an article causes little to no strife, as opposed to pissing off most of the editors here by just deleting things. --badlydrawnjeff talk 16:56, 15 November 2006 (UTC)

If you edit out the non-conforming information, you have compliant articles. If you edit out almost everything in an article, you have a compliant stub. Compliant articles are good, compliant stubs are less good, but are still good. It is that simple.

If you question the value of having an article on the topic in question, you have a topic issue. That is what AfD, prod and CSD are for. Topic issues, answering the question "should this topic have an article in Wikipedia?"

Otherwise you have a content issue. Speedy deletion is very seldom appropriate for content issues. Editing is. Lack of sources is a content issue. Speedy deletion's role in content issues as proposed on the project page should not be expanded.

Bureaucratic hell is having interwoven rules and guidelines for everything instead of just being courteous and cooperative, yet persistent in pursuit of the project's broad goals. This proposal weaves CSD deeply into questions of content. Unfocused 07:01, 15 November 2006 (UTC)

• So, what if editing out the non-conforming information would leave you with not a stub, but a blank article? That is the case this CSD is intended to cover. Robert A.West (Talk) 07:15, 15 November 2006 (UTC)
• As Robert says. This isn't intended to be a wide-ranging proposal on articles and verifiability. We're only talking about articles with no source at all. "edit out the non-conforming information, you have compliant articles" is obfuscation. The only issue here is entirely non-conforming articles. Editing out the non-conforming information then is deletion. We have the same premise, and I can't make out why you disagree. Dmcdevit·t 08:13, 15 November 2006 (UTC)
• I suspect Unfocused's fears are similar to mine. This is a licence for block deletion of masses of articles irrespective of verifiability. If someone can't be arsed to find a source, BANG, it's gone. There are already far too many deletionist rules within Wikipedia and far too improvementalist ones. Mallanox 14:09, 15 November 2006 (UTC)
  • That comment seems a bit strange; this isn't 'irrespective of verifiability', it's about enforcing verifiability. WP:V says "'Verifiable' in this context means that any reader should be able to check that material added to Wikipedia has already been published by a reliable source."; it's a bit misnamed, really, because it implies that things have to be verified. The point is that anyone reading an article must be able to find its source; that's the only thing that Wikipedia's credibility is based on. Remember, this is only for new articles, and the existence of this criterion is what will cause the sources to be added in the first place. Would you really support the retention of articles which were created without references to sources? --ais523 14:21, 15 November 2006 (UTC)
    • If I may clarify, I meant that there is a distinction between verifiability and verified. What we are talking about is deletion of what is not verified, not what is unverifiable. Mallanox 23:54, 15 November 2006 (UTC)
    • The BANG part is misleading. BANG is a tag sitting on the article and the article sitting in a category for 14 days asking someone to legitimize it - and I'll bet plenty of people would relish the opportunity to hang out in that category legitimizing articles. That's not so much a BANG as a polite tap. —Wknight94 (talk) 15:07, 15 November 2006 (UTC)
    • Perhaps BANG is hyperbole but the point is valid. What happens when the category gets backlogged, there aren't enough people to check all of the articles and they start being deleted for no fault of their own? We've all found articles with tags that have been there for months. 14 days tick by pretty fast especially if its a page no-one is watching. Mallanox 23:58, 15 November 2006 (UTC)
      • Agree, there would have to be an equivalent of WP:PRODSUM set up and a new project smilar to Wikipedia:WikiProject Proposed Deletion Patrolling, but that would be possible. Addhoc 15:21, 15 November 2006 (UTC)

This entire line of worry is unfounded. this would only apply to articles with zero sources whatsoever. Any article containing any sourced information would be exempt. "Stubbing" an article means reducing it to the small amount of sourced information, but these articles have no sources, so there would be nothing left to include in a stub! Night Gyr (talk/Oy) 18:13, 15 November 2006 (UTC)

Comment by nae'blis

It doesn't often happen, but I have to agree with badlydrawnjeff here; this is massive overkill. We're already deleting too many people's work because people find it easier to argue in XFD, hit the delete button, or whistle up a tree on talk pages than find sources. Our backlog for sourcing is HUGE; I'll be the first person to admit that. But the solution isn't to delete unsourced articles, it's to source them! If half the effort in AFD went into finding sources rather than deleting the merits of whether or not to keep someone's well-intentioned stub, we'd have more than 20% sourcing already. It's funny, but when I hit Special:Randompage, I end up with one stub about a radio station with an external link, one article about a river in France (with source), a stadium stub w/o sources, and a television station with some sources. This is too subjective and will be misapplied, much like A7 is now. -- nae'blis 16:21, 15 November 2006 (UTC)

But the fact is they aren't getting sourced. We need to do something, the current system is not enough. Since sourcing the articles we have is not working, the only way to improve the percentage of sourced articles is to stop the flood of unsourced additions. The only acceptable alternative is one that can make a drastic impact on the sourcing of articles. - Taxman ^Talk 17:10, 15 November 2006 (UTC)

So work harder to change the culture without making crazy moves that are guaranteed to be misapplied and cause massive wikidrama. --badlydrawnjeff talk 17:15, 15 November 2006 (UTC)

Not good enough. I've been doing that, but we need to do that and more. After years of promoting the need for sources, the percentage of sourced articles is unacceptably low. - Taxman ^Talk 17:46, 15 November 2006 (UTC)

Then can we start with a dated {{unreferenced}} tag? If we need to do more, and we do, let's leverage the wiki system rather than using it to delete hide information from casual and new editors. I'd love to see some statistics on what gets improved faster: redlinks or stubs, but I don't know how to get it. -- nae'blis 15:37, 16 November 2006 (UTC)

What's subjective about zero sources whatsoever? Either there's an entry in the References section or there isn't. Night Gyr (talk/Oy) 18:04, 15 November 2006 (UTC)

Comment by Necrothesp

While I agree that it's very desirable for articles to be sourced, I completely oppose this proposal. Why? Because it's a deletionist's charter and will be employed gleefully by some people to kill as many articles as possible. I have no problems with deleting rubbish. As an admin, I do it myself. However, I have noticed in recent months a disturbing increase in the number of people who seem to come to Wikipedia solely to delete and not to create, to criticise and deride other people's work and not to contribute their own. I am certainly not in any way denigrating the many people who do good housekeeping work on Wikipedia, but our main aim here is to create an encyclopaedia, not to delete what others have written, unless it really is unsalvageable rubbish or fan drooling about utterly non-notable subjects. I have seen some people say that their ultimate aim is to delete articles at a faster rate than they are created. I cannot agree with that sentiment and this proposal will merely make it easier to achieve, while deleting a lot of good work and alienating a lot of good editors. Not being sourced does not automatically mean a bad article! -- Necrothesp 17:14, 15 November 2006 (UTC)

Not being sourced means not being able to tell if it's a bad article, because we have nowhere to look to check! Night Gyr (talk/Oy) 17:45, 15 November 2006 (UTC)

Funny. I've always been able to get a sense of whether an article is complete rubbish or not! Usually the bad writing and ludicrous claims give it away. For any controversial or ridiculous claims then I'd entirely agree you need a source. For anything else? Sorry, but no. Desirable, yes; speedily deletable without, no. -- Necrothesp 17:55, 15 November 2006 (UTC)

It's not the ridiculous ones you need to worry about, it's the ones that sound perfectly reasonable but in fact contain completely wrong and misleading information. Someone creates an article on some item, like say nihilartikel that sounds perfectly reasonable on its surface, but in fact is a totally made up word that never existed before wikipedia, and you'll have no source to check it against. Night Gyr (talk/Oy) 18:02, 15 November 2006 (UTC)

Indeed, but they are rare enough for me to be bothered far more about the carving up of Wikipedia by the deletionists than by the handful of plausible-sounding hoax articles. -- Necrothesp 19:26, 15 November 2006 (UTC)

I doubt this will lead to carving up. I nominated Patience Dabany to DYK; someone pointed out that it needed sources, I found a couple and added them to the article. No pain, and now we have a sourced article, and it reminded me to include sources when I wrote USS Firebolt (PC-10). Would you trust those articles if they didn't have sources available? Night Gyr (talk/Oy) 19:35, 15 November 2006 (UTC)

My experience differs. Walk away from an article, and in a month it won't be trustworthy, sourced or unsourced. Editors will change text, up to reversing its meaning, without removing the source. Sourced articles are somewhat easier to fix, but many of them offer the appearance of reliability without the existence. Sourcing is a good thing, but it's not the magic bullet. Septentrionalis 21:42, 15 November 2006 (UTC)

(back over) It's a start, though. SS-N-15 had a wrong yield on the warhead (20 kt, instead of 200), and I had to go digging to find the correct answer. With a source, I can just go and see "oh hey, that's a typo" and know what's right. Night Gyr (talk/Oy) 21:55, 15 November 2006 (UTC)

Try doing the same on Alexander Hamilton, which is exhaustively sourced, to off-line sources; and is still invaded by cranks, who change the text without dislodging the sources. Sourcing is, I repeat, a good thing; but it will not solve our problems. Septentrionalis 23:21, 15 November 2006 (UTC)

I'm not saying sources aren't useful. I'm saying that I don't trust the deletionists to be sensible about applying a new policy. For every person who will go off and try to find a source there are five more who will say "it's unsourced, get rid of it" even if it's obviously a perfectly legitimate article, and who won't even try to find out about the subject first. It's them I worry about. In addition, it's easy enough to make up a published source - and nowhere does it say (or should it say) that sources have to be electronic. Your apparently brilliantly sourced perfect article might in actuality be a load of rubbish and its extensive sources all made up. Sourcing, as Septentrionalis says, is not the answer to all ills. -- Necrothesp 23:30, 15 November 2006 (UTC)

I don't think anyone claims it is, but it's a start. Night Gyr (talk/Oy) 03:38, 16 November 2006 (UTC)

Being a start is not a reason to widen the speedy deletion criteria! -- Necrothesp 02:03, 17 November 2006 (UTC)

NO one thing solves all of our problems, should we not do any of them? Night Gyr (talk/Oy) 03:26, 17 November 2006 (UTC)

Actually, if expanding the speedy deletion criteria would be a start to improving wikipedia, I would vote for it in a heart beat. I have been reading this argument thoroughly and I have come to appreciate that there are good points on both sides. It seems like the debate is over an age old security problem, do we ban everything and only permit what we know to be good or vice versa. This philosphical debate will likely continue for eternity. My vote is delete the crap and don't write even a stub if you can't provide a source.Alan.ca 05:39, 11 December 2006 (UTC)

14 days is less deletionist than five days

Let's say I put an article with sources up for AFD on the basis that the sources are not reliable, and I describe a diligent inquiry into whether sources exist, the article could be gone within five days if no one comes up with a reliable source.

If current policy (including Wikipedia:Deletion guidelines for administrators) is followed, the closing admin should ignore any "keep" except an actual source or a reasonable argument that the existing sources are sufficient. The article could be gone in five days.

Under this proposed process, the article has at least 14 days from tagging. I think that is a good choice. AFD is five days, which might be too short. We would end up either deleting salvagable articles or subverting the process in some way.

Now, some have suggested allowing an article based entirely on self-published and/or fraudulent sources to be tagged; others would not. This can be tweaked, but the proposal is actually less deletionist than using AFD and enforcing existing policy. 208.20.251.27 17:24, 15 November 2006 (UTC)

Not really. AfD guarantees many, many eyes seeing the article and the issues surrounding it, and requires a consensus to act. This requires only two sets of eyes - one to tag, one to delete, and that's if the 14 day situaiton applies, and assumes that the person tagging isn't deleting 14 days later. More eyes = better sources = better articles. Consensus = important. --badlydrawnjeff talk 17:35, 15 November 2006 (UTC)

We should have a log of pages so tagged, like PROD, so that editors can check for articles that they can add reference to. Night Gyr (talk/Oy) 17:47, 15 November 2006 (UTC)

It wouldn't work the same way. Prod is very easy, this is not. --badlydrawnjeff talk 17:48, 15 November 2006 (UTC)

How is it more difficult than prod? - Taxman ^Talk 17:51, 15 November 2006 (UTC)

Prod involves nothing more than a single judgement call. This requires much, much more, including not only finding a source for somehting that may not be easy to find sources on to save aggressive, moronic deletion, but finding a source that an admin who wants to delete it will accept as reasonable. --badlydrawnjeff talk 19:10, 15 November 2006 (UTC)

I'm not sure if you get the point of this. It doesn't require sources for every statement, just a minimum of a single source for the entire article. Any source is sufficient to prevent speedy deletion under this, even an unreliable webpage. We shouldn't be creating articles without a source in the first place anyway. Disputed sources would need AfD. Night Gyr (talk/Oy) 19:03, 15 November 2006 (UTC)

No, I get the point, and I find the point to be transparent and disturbing. I've seen too many CSds get expanded and implemented and abused, and this will be no different, and this is especially unnecessary. You say it'll only need one source, yet A7s are consistently deleted with assertions of notability. Guaranteed, within 6 months of implementation, someone will propose we limit it to reliable sources, and then the spree begins again. This should never, ever be allowed to happen. --badlydrawnjeff talk 19:10, 15 November 2006 (UTC)

Wow, red herrings all around. This involves nothing more than a single judgement call: is there a source or isn't there. That's not as drastically hard as you're making it out to be, and is much more objective than deciding if an article asserts notability or not. Allowing unsourced additions does have the effect of making the creation of a reliable reference work much more difficult. But now that you've stepped into emotionalized language and name calling, it's clear you're not trying to evaluate the facts and aren't willing to re-assess your position. - Taxman ^Talk 19:31, 15 November 2006 (UTC)

I'm forced to believe that aren't all that involved in deletion discussion, CSD implementation, or deletion review. If you were, you probably wouldn't dismiss my legitimate observations as "red herrings" and accuse me of "name calling." There's nothing to re-assess, not that it appears you're doing much soul-searching on the issue - this is a horrible, horrible idea. --badlydrawnjeff talk 19:59, 15 November 2006 (UTC)

When the name calling is three comments up from mine, it's not an accusation, it's a fact. And "There's nothing to re-assess" is probably the most problematic of anything you've said on this page. There's always reason to re-assess and see if the assumptions we are making are valuable or not, and whether the other side has points that should be conceded. Not being willing to do that is what makes improving this project more difficult than it needs to be. - Taxman ^Talk 21:03, 15 November 2006 (UTC)

If you're going to accuse me of name-calling, you'd better have some evidence, otherwise it's not very civil of you. Back on track, this is a mistake, plain and simple, and I'm disappointed to see it getting any traction, never mind that no one seems to be listening to the anti- side of this. --badlydrawnjeff talk 21:10, 15 November 2006 (UTC)

Nah, the median AfD gets what 4-5 votes? And saying AfD should be used assumes people will vote the right way. As established they don't, they just push it off the to hopefull someday the article will get a source. A system for tracking articles that will get deleted if not sourced after 14 days will put the needed sense of urgency behind the problem and allow the needed eyes. - Taxman ^Talk 17:51, 15 November 2006 (UTC)

I read every AfD I see, and comment on maybe 4 or 5 of them per day. The amount of votes has nothing to do with the amount of eyes that see it. --badlydrawnjeff talk 19:01, 15 November 2006 (UTC)

Perhaps, but that applies just as much to this process. - Taxman ^Talk 19:31, 15 November 2006 (UTC)

No, it doesn't, as I've already pointed out. --badlydrawnjeff talk 19:59, 15 November 2006 (UTC)

comment from Night Gyr (talk/Oy)

I support something similar to this, although I prefer a longer timeframe, mainly because a lot of articles are low-traffic and don't have many editors to work on them. I see the main purpose behind doing something like this not to get rid of articles, but to light a fire under editors who create articles without sources to add them. I proposed on template talk:unreferenced that it be made to categorize by month, with articles that have had it for three months eligible for deletion. This gives plenty of time to add sources, but also has strict criteria for eligibility--{{unreferenced}} only goes on article with zero sources whatsoever. If it has unreliable sources, that kind of judgement should be made by AFD, but a total lack of sources should be obvious. Also, I don't like calling this speedy because we obviously plan to give them some time before deleting. As for the concerns that editors are deleting without looking for sources, obligation to source is on the editor adding material, not the one removing it.

To summarize:

• This is intended first to light a fire under editors, not delete articles
• Only articles with zero sources whatsoever should face deletion without AfD

Perhaps we could extend this to pre-existing articles, but with a longer timeframe, like three or six months? I'd also suggest a notification like we currently have in place for {{nsd}}.

Night Gyr (talk/Oy) 17:54, 15 November 2006 (UTC)

No one should delete or nominate anything for deletion that isn't already speedy (WITHOUT this expansion) without first doing a ten second check on the search engine of their choice, even if they tagged it "need sources" ten years ago. Doing so is lazy and disrespectful of previous editors, and violates WP:AGF. Unfocused 18:04, 15 November 2006 (UTC)

Another comment: Our longstanding base requirement for notability has been coverage in independent, reliable sources. What better way to prove it than cite those sources in the article? Night Gyr (talk/Oy) 18:11, 15 November 2006 (UTC)

Something I haven't seen addressed: How can we know added information is reliable unless it's sourced? Do we want to allow people to add unreliable information? Night Gyr (talk/Oy) 20:15, 15 November 2006 (UTC)

• How can we know added information is reliable if it is sourced? We can't, unless we consult the source; and then we will often be surprised to find that the source says something quite different.
• How do I know simply connected space is right? Because I know what a simply connected space is, and that's it.
• How can you know? By relying on the consensus and good faith of the editors + Splash and myself. This is the lowest grade of confirmation; but it's not nothing. Septentrionalis 02:56, 16 November 2006 (UTC)

Comments by Bkonrad

I'm afraid I don't see why this is a proposal for CSD. I thought the idea behind CSD was a sort of shoot-on-sight to get rid of complete crap. This proposal is not about speedy deletion (at least not in the same way that all the other CSD are). I don't see why this couldn't be treated more as a variant of PROD. But beyond that, I'm very uncomfortable that combining unreferenced with deletion sends mixed messages. The unreferenced template is (was) supposed to be a clean-up variety of tag. When someone puts the unreferenced template on an article, are they saying "I think this article should be deleted" or are they saying "I think this article seems reasonable, but it doesn't cite it's sources and there is no way to verify the content"? older ≠ wiser 18:32, 15 November 2006 (UTC)

Yeah, I don't like including it with CSD either. I'd rather think of it as a parallel deletion system like PROD. The trouble with being nice is that "this article seems reasonable, but it doesn't cite it's sources and there is no way to verify the content" is right smack against WP:V. Night Gyr (talk/Oy) 18:38, 15 November 2006 (UTC)

Note that 'speedy' deletion for unsourced and/or unlicensed images has a seven day waiting period, so there is a prededent for calling this 'speedy'. -- Donald Albury 19:26, 15 November 2006 (UTC)

Except the current hard-line version of the Verifiability policy is a fairly recent interpretation. There was always a bit of ambiguity about what "verifiable" meant. I've always consiered the policy to be that it must be possible to verify information added to WP and that appropriate sources should be provided if challenged or requested. Personally, I find the voluminous footnoting of mundane details detract from Wikipedia more than they add. I'd much prefer it if articles provided a list of general references/sources for a topic and then only footnote details that are actually remarkable or contested in some way. older ≠ wiser 18:53, 15 November 2006 (UTC)

The hardline has come pretty much straight from jimbo. Also, this criteria doesn't say you need footnotes or individual citations, even just one general reference is sufficient. Night Gyr (talk/Oy) 18:54, 15 November 2006 (UTC)

I think taking Jimbo's words out of context to be universal truths is not a very sound basis for making policy. I completely agree that any unsourced edit that seems in any way dubious should be challenged and removed if no source is provided. Whether it is removed immediately or only after solicitations on talk pages depends on how dubious or defamatory the edits are. I also understand that this proposal is not about footnotes in general, I was only venting a bit about a cultural shift that in some cases may be taking a good thing to too much of an extreme. older ≠ wiser 19:09, 15 November 2006 (UTC)

But I might hasten to add, that while I'm not sure about using the unreferenced tag for this purpose, I would support using some form of prod-like mechanism for deleting articles that are indeed completely unreferenced (and for which no sources can be found with some minimal amount of due diligence by the deleting admin). older ≠ wiser 19:23, 15 November 2006 (UTC)

• Unless NightGyr is referring to some other comments from Jimbo, in which case they should be added, his hardline comments refer to living people. This page is much broader in scope. Septentrionalis 19:15, 15 November 2006 (UTC)

See here, here, and here. Instead of leaving unsourced (and therefore unreliable) material in wikipedia with a "source needed" tag, he says we might as well remove it. Night Gyr (talk/Oy) 19:24, 15 November 2006 (UTC)

The second one of these is not on-topic; does NightGyr intend another post from the same thread? The other two support removal, not deletion. Over-enthusiatic removal can be reversed by anyone, with the page history; deletion is much harder to reverse. Septentrionalis 19:42, 15 November 2006 (UTC)

When you have no sources, removing the unsourced information is the same as deleting the article. Night Gyr (talk/Oy) 19:47, 15 November 2006 (UTC)

No, it's the same as blanking the article. This is exactly the sort of confusion that makes this proposal dangerous. Septentrionalis 19:56, 15 November 2006 (UTC)

(back over) And blank articles are deletable, and under this proposal any editor who provides a source can have the article undeleted, so the material can come back. Night Gyr (talk/Oy) 20:04, 15 November 2006 (UTC)

Thank you. I strongly oppose any proposal which can be interpreted as Night Gyr interprets this one. Septentrionalis 20:15, 15 November 2006 (UTC)

I only think this should be policy for new articles, having gone back and looked over older ones, but I still see no reason we should allow people to write new articles without a single reference. Night Gyr (talk/Oy) 20:17, 15 November 2006 (UTC)

I'm afraid that I also must disagree with Night Gyr's interpretation above. Removing content is not at all the same as deleting a page. Reverting an edit preserves the page-history in a way that any future reader/editor can review and, if necessary, correct. It requires no special admin-powers to improve the article. Deleting a page, on the other hand, secures the page-history from general view and does require special permissions to reverse. It dramatically reduces the number of users who would have the knowledge and ability to improve the project in the case of an unsourced but sourcable page. Rossami (talk) 22:04, 15 November 2006 (UTC)

Which is why we have a waiting period during which the article is placed in a highly-visible log of unsourced pages for any interested editors. Or should we keep unsourced content around indefinitely until someone gets around to finding out whether it's actually true or not? Night Gyr (talk/Oy) 22:11, 15 November 2006 (UTC)

So if you've already tagged it as unsourced and implicitly disclaimed the accuracy of the content, why 14 days? Why not 3 months? Or 3 years? As long as it's properly tagged and in a cleanup log where interested editors can find and work on it, what's wrong with leaving it until someone has the time and interest to prove or disprove the point with facts? Shoot, even if it's not tagged, readers can see just by looking at the page whether it's sourced or not and can evaluate the page's reliability for themselves. Why is it necessary to create a hard and very short timeline for this work to get done? Rossami (talk) 22:40, 15 November 2006 (UTC)

Note that "speedy" deletion doesn't really mean a deletion is fast (if that makes sense) but that any administrator may delete an article that qualifies without debate. This is why this would be considered speedy, just like the various image criteria. It's really semantics, though, since PROD works the same way. I'm not sure if there's any use spending time on deciding whether the concept should be called speedy or not. Dmcdevit·t 23:58, 15 November 2006 (UTC)

Thoughts of a compromise wording

Okay, so this came about in my head just now. I still think this is absolutely horrific, and I don't consider the compromise much better, but accepting middle ground is a good thing, right? I'm going to use 1 January 2007 as a starting point:

Any article created after 1 January 2007 that does not cite any sources or references is tagged with a {{sourcedeletion}} template. If, after 14 days, the article can be AfD'd with this tag in mind and deleted if the consensus of the community is that no references are attainable.

Obviously, the language needs to be cleaned up, but this is a good idea:

1. It maximizes the amount of eyes that see the article before it's deleted.
2. It takes the power away from administrators who will undoubtedly, either purposefully or accidentally, abuse this tag, either by not waiting the 14 days, deleting articles with sources, or not approving of articles with sources. It happens now with CSD A7 almost daily, so don't come out with "oh, that will never happen."
3. It reduces, if not eliminates, the chances of instruction creep. I find instruction creep arguments to be completely without merit, but this is rather airtight, and always ends in the same way, with a full hearing at a deletion process.

I'd much rather slap a historical tag on this monstrosity and be done with it, but any thoughts on this wording? --badlydrawnjeff talk 19:28, 15 November 2006 (UTC)

• ...Slap a historical tag? It's barely been in discussion for three days. The trouble with your proposal is that it says to go to AfD...which is exactly what we already do. Night Gyr (talk/Oy) 19:31, 15 November 2006 (UTC)
  • Right now, nothing (thankfully) forces anything into AfD. We simply do not delete due to lack of sources. This is a change to that. And yes, historical as opposed to outright delete. Kill it with fire as far as I'm concerned. --badlydrawnjeff talk 19:55, 15 November 2006 (UTC)
• Unless that's changed to attained then this amounts to nothing more than a massive increase in overhead. If it's so important to include the given material people can accomplish that by helping to find sources for it. - Taxman ^Talk 19:41, 15 November 2006 (UTC)
  • This is already a massive increase in overhead, I don't see what this has to do with anyhting.
• If it goes to AfD, the only way it should be kept is if sources are provided (not if sources could be provided). But that's not really AfD. I do think it would be a good idea to have a maintenance page for to-be-deleted articles that it should be easy to find sources for. --Interiot 19:48, 15 November 2006 (UTC)
  • I wouldn't be against the maintenence page, but the 'send to AfD' is meant to be a compromise to those who can't get rid of good content fast enough. --badlydrawnjeff talk 19:55, 15 November 2006 (UTC)
    • And thus the extra overhead. If the change is made that the article will be deleted if there are no sources added then the extra overhead will be worth it because it should help aid in finding sources. If we create a place with a sense of urgency to find more sources for articles I can't think of anything more valuable. The last half of your statement isn't helpful. No one is trying to get rid of good content and you know it. It's just that unsourced content isn't good content because the reader has no way of knowing. The other side people are missing is that once this policy becomes well known, the number of articles that fall under it will reduce. Just think of how many headlines will be generated when Wikipedia shuts it's metaphorical doors to articles with no sources. Our percieved reliability will go up and so will the actual reliability because more of the type of articles we really need will be submitted. - Taxman ^Talk 21:19, 15 November 2006 (UTC)
• Though maybe it would be a good idea to make it mandatory to run grandfathered articles through AfD before deleting them, just because there's quite a few unsourced articles with a long edit history. --Interiot 19:52, 15 November 2006 (UTC)

We could say that articles that have an edit history of more than 50 are exempt. Addhoc 23:47, 15 November 2006 (UTC)

Simply connected space has just less than 50 edits, but it was also pretty good when created. This also raises the problem that dividing an article is, by the system, creation of a new article; and those often appear unsourced (even when the old article had one, like the 1911 Britannica.) Septentrionalis 00:12, 16 November 2006 (UTC)

• One solution would be to permit appeals of the tag to AfD: "I know simply connected space is good; I could just add a random topology textbook to it, but it's not my field. Should we really delete this?" Septentrionalis 00:12, 16 November 2006 (UTC)

• I like the compromise of sending it to an AFD type discussion at the end of the period, and running that pseudo-AFD discussion about the sourcing of the article. One of my two major concerns was that this tries to move the culture too fast, and thus would inherently lead to a large number of controversial deletions. We don't need more activity at WP:DRV in the amount I fear that this would bring as a criteria for speedy deletion. Using a pseudo-AFD has several advantages: 1) it changes the culture more slowly, 2) it puts more eyes on the article/subject prior to disposal, and 3) it gives more time and discussion, creating less of a WP:BITE problem.
• Iff that is working well after a while, we could move to a pseudo-prod, and iff that works well eventually to a pseudo-CSD. But starting as a pseudo-CSD is starting at the wrong end of the speed chain. GRBerry 01:36, 16 November 2006 (UTC)
  • I think just dated tags with a message to notify people who add unsourced information and point people to WP:V and WP:CITE would be a great start. If editors know that they shouldn't be writing articles without sources, even stubs, then they will add them in the future and the problem will shrink even before we implement any deletion. Night Gyr (talk/Oy) 03:45, 16 November 2006 (UTC)
    • Great. Now if everybody would do it (which I can guarantee won't always happen). And except that they won't know; I daresay that the majority of Wikipedia editors can't state a single Wikipedia policy off the top of their head. ColourBurst 15:59, 16 November 2006 (UTC)

Comment by Messedrocker

Apparently everyone's doing it this way, so I thought I would too. (Jim! How dare you do things like everyone else! You tool!) Erm, in any case, this can be a good idea but should be approached carefully. For starters I like the idea of the grace period that'll allow people to find sources. I also like that it would not be applied retroactively -- that would cause most of the encyclopedia to disappear. One thing we could do is categorize all articles nominated under this criterion under "Articles immediately needed sources" which people would watch and try to fill in sources for articles. In regards to there being a WP:RS-related dispute, at first we should consider the definition of a source liberally, then get stricter as we progress. ★MESSEDROCKER★ 20:21, 15 November 2006 (UTC)

A list, like the AfD log, might be better; categories aren't watchable. Septentrionalis 20:25, 15 November 2006 (UTC)

Perhaps we could have a category and then a bot who adds the latest editions to the category to a list maintained by it, because expecting humans to maintain what will be an enormous list is too much to ask for. ★MESSEDROCKER★ 00:24, 17 November 2006 (UTC)

Strict interpretation

Criteria for speedy deletion are written strictly, so that they may be uncontestable: it should be the case that almost all articles that can be deleted using the rule, should be deleted, according to general consensus. If a rule paves the path for deletions that will cause controversy, it probably needs to be restricted.

No criterion which permits IAR satisfies this, and unless it is explicitly so restricted, I oppose it. Septentrionalis 21:16, 15 November 2006 (UTC)

I assert the following text is the meaning of may.

• "may be deleted" It is neither required nor desirable to nominate all articles which qualify under this policy; it is a matter of judgment whether an article is evidently sourceable.

Should we write a tag for the purpose?

Please discuss. Septentrionalis 21:25, 15 November 2006 (UTC)

This tag isn't about "sourceability", it's sourced or it isn't. Unsourced content may be removed by any editor. Uncle G has a good essay on how everything we should include is sourceable. Night Gyr (talk/Oy) 21:47, 15 November 2006 (UTC)

See User:Uncle_G/On_sources_and_content#There_are_no_exceptions_to_everything Night Gyr (talk/Oy) 21:49, 15 November 2006 (UTC)

Could someone please give me an example of an article that doesn't need a single source? Night Gyr (talk/Oy) 21:50, 15 November 2006 (UTC)

The more important question is not such much "need" as "can do without" (Uncle G's argument has several problems). See Simply connected space, for example. Now it needs a source, really, I agree, but being already in possession of the knowledge that the article is right, should we/I delete it if I come across it? Do I improve Wikipedia by doing so? Do I do the readers a service? Or do I merely follow fashionable rhetoric built up from a slightly hot-headed series of posts on a mailing list? And do note that I am quite firmly in favour of well-sourced articles, too. Splash - tk 22:24, 15 November 2006 (UTC)

We don't delete it (right away), we slap a tag on it that says it needs sources or we go find a source for it ourselves. Now, the editor who originally added that article may not still be around, so we can't trust that we'll have the original author to provide sources. On newly-created articles, however, the original author will be right there so we can say "ok, nice article, but you need to tell us your sources or it can't stay." Night Gyr (talk/Oy) 22:30, 15 November 2006 (UTC)

So the answer to my question is "yes, we should delete it". Oh, and just to close the wriggle room, it doesn't count to now go and add a source to the article (good as that would be!), because this proposal deals with what should happen to the article if it were still unsourced at the speedy guillotine. Splash - tk 22:33, 15 November 2006 (UTC)

That's a very good example. I fully support this proposal, however there are some exceptions probably worth discussing. In this case, the article exists in five other languages, which possibly could be enough for AfD instead of speedy. Regarding wiggle room, don't worry there are over 2500 similar articles. Addhoc 22:35, 15 November 2006 (UTC)

Objections

Arguing for the inclusion of unsourced information in wikipedia -- That isn't a strawman, that's exactly what opposition to this policy advocates. Why? These are what I see:

• Too many pre existing articles -- this ignores preexisting articles, only applies to newly created content.
• Will lead to deletion of good content -- unsourced content isn't good content, and we're going to set up a system to give many chances for sources to be added before an article gets deleted (two week wait, listed in a log, etc.)
• Will be abused to delete articles subjectively -- CSD has been abused before -- G11 was used to delete a bunch of valid articles on product when they first came out, but this one is pretty clear cut. no source = delete, source = no delete. That's less of a judgement call than advertising or notability claims.

Are there any I'm missing? Night Gyr (talk/Oy) 22:02, 15 November 2006 (UTC)

Opposition to this policy does not mean that the opponents are all "arguing for the inclusion of unsourced information in wikipedia". Many of the opponents are merely arguing that speedy-deletion (even with all the proposed qualifiers and controls) is not the optimal process for doing so - that other mechanisms could be better at either finding sources or removing unsourcable information or that those other mechanisms would have a lower probability of unintended consequences.

I also suspect from the comments above that a fair number of experienced users would be very uncomfortable with the sentiment expressed in your premise that "unsourced content isn't good content". In context, your premise comes across as "unsourced content is axiomatically bad content". The fact that something is not yet sourced does not mean that it is unsourcable or that the project will be irreparably damaged by the delay. Rossami (talk) 22:28, 15 November 2006 (UTC)

The main issues of unreliability come about with content that can't be supported by sources. Entirely unsourced articles are obviously unreliable. I've learned a lot of things from them, but I wouldn't trust them for anything, and I can't go anywhere more reliable to check without a source indicated.

This policy wouldn't remove content that's already here and just didn't get sourced the first time, it just calls for all new content to be sourced, which makes sense as I said in the last section. Night Gyr (talk/Oy) 22:32, 15 November 2006 (UTC)

Short version: unsourced = unreliable. unreliable = Bad Thing. Night Gyr (talk/Oy) 22:42, 15 November 2006 (UTC)

Unless an alternative is proposed that is equally successful in removing unsourced new articles or ensuring they are sourced, then yes, opposition to this does have the effect of "arguing for the inclusion of unsourced information in wikipedia". The project isn't terribly damaged by the lone unsourced statement, but on the whole it is of vastly lower quality than it could be due to the extremely high percentage of unsourced articles. Again, as in the other section, consider the power of everyone knowing that only sourced articles are allowed in. - Taxman ^Talk 23:01, 15 November 2006 (UTC)

Specifically, though, opposition to the proposal to delete any articles containing unsourced information means that someone thinks the article should not be deleted. It does not mean that the unsourced should not be dealt with in some way, perhaps by editing it out. It's just the nearly-nuclear option of deleting the entire article that's getting opposed, and that for a variety of reasons not all relying on a desire to see unsourced information in articles. Splash - tk 23:11, 15 November 2006 (UTC)

The crux comes down to articles that have no sources. Is it acceptable to blank the article until sources can be found? - Taxman ^Talk 00:31, 16 November 2006 (UTC)

It's acceptable to stub them. We do it with WP:OFFICE stuff all the time. --badlydrawnjeff talk 00:32, 16 November 2006 (UTC)

• I don't think you're accurately representing the opposition. We're not advocating unsourced material, we're opposed to wholesale deletion of it and further expansion of administrator power to delete information without consensus. --badlydrawnjeff talk 23:58, 15 November 2006 (UTC)

• I was careful to say that you're not advocating unsourced material, but I did correctly state that arguing against this proposal does have the same effect, more unsourced information in articles, unless an equally effective alternative is produced. - Taxman ^Talk 00:31, 16 November 2006 (UTC)

• Not at all. We simply oppose deletion of entire articles as part of the battle against unsourced information. --badlydrawnjeff talk 00:32, 16 November 2006 (UTC)

• That the above mentioned is the effect is really easy to establish. So much so I'm surprised you would try to deny it. - Taxman ^Talk 04:59, 16 November 2006 (UTC)

• So anyone who disagrees is a heretic? How many politicians have used the old "Whoever isn't with us is against us" chestnut? I disagree with this proposal because of its destructive nature. Improve articles, don't destroy them. Articles that are unverifiable can be legitimately deleted. Deleting unverified articles is, in my opinion, short-sighted and very high-handed for a project that relies on volunteers. Mallanox 00:06, 16 November 2006 (UTC)

Comment by PBS

• AFAICT, There is currently no definition in a WP:Policy of what is a reliable source. (there are descriptions of what are not reliable sources, but not what is a reliable source). --Philip Baird Shearer 22:17, 15 November 2006 (UTC)
  • This doesn't care about reliability, just whether a source is provided at all. Night Gyr (talk/Oy) 22:21, 15 November 2006 (UTC)
    • Well no, what it really cares about is reliability, but as a safety latch, it only acts on articles with no sources at all. - Taxman ^Talk 23:02, 15 November 2006 (UTC)
    • For a source to be provided it must be reliable. Othewise this suggestion is not worth having. --Philip Baird Shearer 08:06, 16 November 2006 (UTC)

• What about short stub articles. I see no reason why they have to have source. After all they can be a round about request for a more detailed article.--Philip Baird Shearer 22:17, 15 November 2006 (UTC)
  • Stubs contain some information. Information has to come from somewhere. Night Gyr (talk/Oy) 22:21, 15 November 2006 (UTC)
    • Tagging won't be automatic. A stub that has no sources will probably not be tagged for a couple of weeks, and then it will have another 14 days to grow into a one-source superstub. If the stub creator doesn't have a source and doesn't know where to look, why should we trust his stub? Robert A.West (Talk) 23:00, 15 November 2006 (UTC)
      • We have articles tagged for CSD within minutes of their creation. New pages patrol will be all over this. --badlydrawnjeff talk 00:33, 16 November 2006 (UTC)

• On the contrary, WP:V#Sources of dubious reliability (one of our most important policies) covers it. Wikipedia:Reliable sources (a guideline) expounds on it more, of course. If this becomes a CSD, I think we'd want to create a simplified version of WP:RS for new users to quickly look over.
• I don't think any reference should be allowed... people shouldn't be able to trivially game the system by tossing in a random link that has little to nothing to do with the article. Common sense can be used, especially if it's extremely obvious someone is trying to game the system. --Interiot 01:56, 16 November 2006 (UTC)
  • What definition of "reference" do you want to apply? It should satisfy the CSD conditions objective and uncontestable. Septentrionalis 02:15, 16 November 2006 (UTC)
    • One that is covered by a definition in the Wikipedia Policy pages and AFAICT this is a very active area of debate at the moment. I think that defintion has to be decided before one concludes a debate on this proposal. "First catch your hare" --Philip Baird Shearer 08:14, 16 November 2006 (UTC)

The verification that the information in a stub is valid, can come from links to other Wikipidia articles, and internal links do not count as sources. For example think of an timeline which starts as little more than a disambiguation page, providing the links to the items include the date put into the timeline there is no need to provide an alternative source. BTW would this suggestion as it stands mark disambiguation pages for deletion as they do not have sources? --Philip Baird Shearer 08:06, 16 November 2006 (UTC)

Law of unintended consequences, and other realities

Ais523 did a search though Random articles, and came up with two "unsourced articles", which exemplify the problems with this proposal.

• John Powers (poet) is a vanity article, and I support Ais's prod. But it's not unreferenced: the text contains two external links, partly substantiating it. It's not subject to this CSD, but it's c;ear that Ais would have tagged it. What happened here could happen to a better article; and will, if this passes.
  • (inserting comment here) What actually happened was that the author of the article showed up on my talk page with a source; if {{prod}} can do this, imagine what {{nosourcedel}} can do. --ais523 08:59, 16 November 2006 (UTC)
• Mirko Derenčin is a more serious case. It contains two facts: he was one of the Bans of Croatia in 1493, and he died at the Battle of Krbava field. The battle has an article, which mentions his death, and has a source; a general history of Croatia.

Now consider the following all-too-plausible scenario (A):

• New article patrol tags Mirko Derenčin with {{source or die}}.
• Loyal Croat, watching the page, does the research I did, and then copies the history from one article to the other, without consulting it first. He is very likely, but not certainly, right to conclude that Mirko's banship and death are in fact mentioned; and if LC is in Croatia, he may not have access to the book in 14 days.
• LC removes the tag.

Has Wikipedia benefitted from this exchange? Not really; the article is still unchecked, and it looks better than it is; and it may be bearing a misleading source.

How often will such things happen? Any proposal like this is a balance between several possibilities; two more for Mirko are:

B: Mirko Derenčin was in fact written from a sound source, which was mislaid when the article on Bans of Croatia was divided into separate articles, No one is watching it, and it is deleted.

C:Mirko is being watched by someone who can source it in a hurry. The source he can find right now is in Croatian.

A is a definite, small, loss; B is a real loss; C is a gain, but we would have had a greater gain by tagging it {{unreferenced}} and waiting. We should consider such things before enacting policies and guidelines, rather than assuming the best result (everyone is scared into sourcing every article they start) will happen automatically. It won't. Septentrionalis 02:15, 16 November 2006 (UTC)

I whole-heartedly agree. Thanks for taking the time to reason that through. Mallanox 02:20, 16 November 2006 (UTC)

No major loss, it can be recovered trivially. I find it really funny that people are willing to say be patient, articles will eventually get sources (even though it's easy to establish that happens extremely slowly on the whole, much much more slowly than articles get created in fact), but stand agast that we should wait a bit longer for an article to be created (or recreated) with sources. In any process there will be some errors. We already knew that, so pointing out an error does nothing. Should we can the whole deletion process because there are some errors? Sure, then we'll have an untold flood of junk. Same goes here, instead we just need to flip the switch to avoid the junk instead of failing to do so out of fear. Eventualism swings both ways. - Taxman ^Talk 05:13, 16 November 2006 (UTC)

There is little I disagree with more than the perennial assumption that undeletion is trivial. It isn't; and if no-one is watching Mirko Derenčin to know it's gone, who will have reason to rewrite the article? Septentrionalis 05:20, 16 November 2006 (UTC)

Amen. Undeletion is trivial for admins and long-time Wikipedians. Casual and new (and probably not-so-new, but I can't prove it) editors don't understand what undeletion is, how it's possible, where to go to get it, or anything else. Sometimes they think there's something wrong with our servers; more often they think someone hates them or their article's subject or us on a powertrip. Think back to what it was like when you started here, and tell me it wasn't a confusing, noisy place. Now multiply that by what we've got now in terms of userbase, user tools, and conflicting essays/guidelines/policies. --nae'blis 15:31, 16 November 2006 (UTC)

Recovered trivially? It's never easy to get an article undeleted, and it's even harder with some admins than others. --badlydrawnjeff talk 15:42, 16 November 2006 (UTC)

A may become a gain next time the article is checked; anyone who checks the source will find either that it's reliable, which is fine, or that it isn't or doesn't substantiate the article, in which case at least we know there's a problem; probably enough of a problem to put the article on AfD (which wouldn't happen at the moment). B is a loss at the moment; this rule would discourage that situation from happening. As for C, tagging it {{unref}} and waiting wouldn't be guaranteed to have results at all. --ais523 09:07, 16 November 2006 (UTC)

The whole point of this proposal is that articles tagged as unsourced aren't checked. To now discuss the checking of case A (which has been provided with an obvious and plausible source) as though it were likely to happen is to beg the question. Septentrionalis 19:34, 16 November 2006 (UTC)

Best I can tell the whole point to this proposal is too put all the article's in catagory C there. If a source doesn't turn up within two weeks I highly doubt it ever will. Tagging something indeffinetlly is just passing the buck, a source or deletiong senario is going to encourege sourcing much more --T-rex 16:39, 28 November 2006 (UTC)

Fake sources

I just thought of something else: false sources.

User A makes a new article about, say, an obscure piece of art. User B on newpage patrol sees that there's no source, and tags it. User A freaks out, and decides he's going to put "The Obscure Book of Art, by Alfred P. Artist, 1943" under "references." Tag removed, everyone runs along on their merry way.

I'd like to think this wouldn't happen, but how would this possibly be dealt with? Especially when "any source" can be used. Hell, why not a webpage that never really existed? The possibilities here are unfortunately endless, and we know what lengths people will go to in order to justify their article's existence. --badlydrawnjeff talk 03:55, 16 November 2006 (UTC)

Because this could happen anyway, when the thing goes up for deletion under any other criteria. Night Gyr (talk/Oy) 03:59, 16 November 2006 (UTC)

It could, yes, but the current situation doesn't encourage false sourcing. Currently, if you AfD due to lack of sources, many many many people see it and vouch for it. In the proposed one, two people see it and it likely never gets questioned again. --badlydrawnjeff talk 04:01, 16 November 2006 (UTC)

But in the current situation it'd never get questioned ever. If someone wouldn't bother to bring it to AfD now, why would they before? Night Gyr (talk/Oy) 04:36, 16 November 2006 (UTC)

And under the current situation, there is very little incentive ever to make a fake source. Unless the article is heading for FA, in which case it may well get caught, there is no requirement for one. This proposal combines maximum incentive and minimum checking, since the admin who evaluates the {{source or die}} (assuming one does) will say, "Yep, that's a reference; I never heard of Alfred P. Artist, but what do I know about decorated pin-cushions?". Septentrionalis 05:01, 16 November 2006 (UTC)

If we put in place a process which encourages the addition of false references, the the credibility of Wikipedia will be underminded the first time a journalist writes an article on Wikipedia which highlights a few articles with this problem, particularly if they can point to a process within Wikipedia which makes it systemic. At the moment as Septentrionalis points out there is no incentive for people to do this, so usually the compliant is that an article on Wikipedia is unreliable because it is not sourced. If it becomes articles on Wikipedia are unreliable because the references my be fake then we have a much more serious problem as it can not be fixed by just adding true sources. --Philip Baird Shearer 08:29, 16 November 2006 (UTC)

Checking the sources is a good way to check the article. If the wiki process works properly, the first person to check the source will think "Hey, that's not a genuine source", and remove the source/warn the person who added it/ask what to do at the Help Desk, etc. An erroneous source at least makes it easy to AfD an article (this article was written based on incorrect sources); no source at all makes it hard to do anything. --ais523 09:03, 16 November 2006 (UTC)

Contrary to what is asserted above, sources are currently required, articles without them can be and are deleted, and edits without them are routinely removed. Thus, we already have plenty of incentive to fake sources: we call them edit wars. Granted, this usually takes the form of a citation to a source for an assertion it does not contain, but that is just as damaging as a completely fabricated source, and it is a lot harder to prove before Arbcom. Robert A.West (Talk) 11:38, 16 November 2006 (UTC)

In my FA candidate right now, Kroger Babb, I use a lot of real sources that are not easy to find. Because of that difficulty, I've made it clear as to where I've retrieved them from, although I'm under no obligation to do so. Now, if I have random newspaper clippings and transcripts of speeches that don't appear online, in books, or anywhere else but the library that they came from, you have no viable way of checking on these if I don't tell you where they're from, nor am I compelled or forced to tell you where I got them. Now, some dishonest person creates an article on our artist, and says "Hey, I have this clipping from the Worcester Weekly that tells about this piece of artwork, so I added the source, don't delete." The publication doesn't exist, but what admin is going to know that? So the user gets to keep his article, and it remains not only unsourced, but falsely sourced, and no one's the wiser. Now, in terms of "what's worse": the Wikipedia entry that contains no sources, or the Wikipedia entry that contains false sources? We don't have a widespread issue with false sources currently, thank goodness. This will indeed swing the door wide open for such abuses, and there will be no realistic way to stop it. --badlydrawnjeff talk 13:13, 16 November 2006 (UTC)

First, almost anything that has happened in the past 20 years is either going to have online sources, in which case faking is pointless, or not, in which case it's going to be glaringly obvious that the source is faked. Yes, there will be a few exceptions, but the purpose of the proposed policy is to reduce the volume of junk sharply, not to be 100% accurate.

Second, the major problem with pointless wikipedia articles isn't with articles about old, historic matters -- it's with aticles about current events, people, bands, etc. posted by people who (mostly) have no clue about wikipedia policies. For example, vanity articles are almost always about current events. This policy deals with that major problem.

Third, an editor who makes up sources is at high risk of losing a lot of credibility. Which means that frequent contributors - like you - won't, while the hit-and-run artist who does will be fairly obvious because of very few edits.

No policy is perfect; anything that solves a major problem can still be expected to create minor ones. If this policy results in 100 articles being taken out of AfD, where they require work by a lot of editors and administrators to close out, and 5 more cases where editors put up fake sources in articles (presumably articles about relative obscure subjects, non-vanity) that need to be challenged, or even go unchallenged because no one reads such articles, then I think that's a fair tradeoff. John Broughton | Talk 13:32, 16 November 2006 (UTC)

We're not talking about the last 20 years, here. We have articles from the last 2000 years, for god's sake. Regardless of if it's 2, 20, or 200 years ago, the ability to fake a source is incredibly easy, and this makes it very enticing to do so to save one's article without any easy repurcussions. You say "the purpose...is to reduce the volume of junk." I think that keeping 100 unsourced articles in inherently better than deleting 99 unsourced articles and have one kept around that is sourced falsely. --badlydrawnjeff talk 14:07, 16 November 2006 (UTC)

I would hope that an editor who puts in fake citations is risking a lot more than credibility. Such an editor is a clear and present danger to Wikipedia. They should be risking a permaban.

W. Edwards Demming pointed out that, for any metric of quality, some people will find it easier to crock the metric than to do what the metric was designed to encourage. The fake citation defect is a risk with any attempt to reduce the number of unverifiable articles, regardless of methodology. That doesn't mean we shouldn't make the attempt. Robert A.West (Talk) 15:29, 16 November 2006 (UTC)

Those against this criteria are not against reducing the amount of unsourced material on Wikipedia. The difference is fostering a situation that encourages real sourcing instead of creating a situation where adding a false source that is difficult to prove a hoax is worth doing. There are plenty of ways we can encourage sourcing without the added affect of encouraging hoaxes. --badlydrawnjeff talk 15:38, 16 November 2006 (UTC)

Except 1) the ratio won't be 100 to 5. It would probably be more like 100 to 50; when cornered, people act strangely. 2) "Everything created within the last 20 years would be online" is systemic biasish; you're discounting countries where they don't even have widespread internet access for 20 years (heck, some parts of the world don't even have widespread internet access now.) ColourBurst 17:12, 16 November 2006 (UTC)

This will not remove that many articles from AfD, and most of the ones it will remove are not much work. John Powers (poet) has acquired a source, such as it is; the next step is to put it up for AfD as non-notable, which should be a snowball, but is going to be as much work as it ever was. (I'll let someone else do that). As for the tradeoff, all I can say is that I firmly disagree; easing AfD is not worth introducing positive errors, and deleting stubs like Mirko, which are hard to replace. Septentrionalis 19:50, 16 November 2006 (UTC)

Proposed Exemptions

Could I suggest the following cases should be exempt from this proposal:

• Where the article has survived an AfD.
• Where there have been more than 50 edits to the article.
• Where the article exists in other language Wikipedias.

Addhoc 13:11, 16 November 2006 (UTC)

The first case is what this proposal was originally intended to prevent! (Dmcdevit seems to want this to prevent 'keep unless sourced' results to AfDs; correct me if I misunderstand.) The second case seems reasonable, but is possibly instruction creep. The third case is undesirable; just because an interwiki exists doesn't mean that the article's any more verifiable; either it's sourced in the other language, in which case you can just borrow the same sources, or it isn't, in which case the other language article isn't any more reliable than the English version. --ais523 13:31, 16 November 2006 (UTC)

And I am not sure about the 50 edits rule either. It would further encourage edit-count padding by saving once per sentence, and twice at the end of a paragraph. If the topic has no sources, having 50 edits doesn't magically create them. Robert A.West (Talk) 15:38, 16 November 2006 (UTC)

At least three substantive editors would serve the same purpose, although it has the weakness that "substantive" might be a judgment call. The purpose would be to exclude articles on which there is editorial consensus; these should be sourced, not deleted. Deleting simply connected space would be cutting off our nose to spite our face. Septentrionalis 16:49, 17 November 2006 (UTC)

So it's to be a 14 day death march, then?

Keep in mind that every power we've ever granted editors to delete or otherwise pressure other editors to conform to their will has been used enthusiastically upon introduction. Jadwiga_of_Poland is a front page article, yet has not a single source. Are we gonna speedy delete that if some other editor doesn't comply with our demands within fourteen days? This entire proposal is a complete violation of assume good faith and pretty much ruins the project for casual editors.

The way forward is by persuasion, not by putting people's articles on a death march. A majority of editors here have a real life, and don't check Wikipedia obsessively. They come, they contribute, the leave. If there's an unsourced article whose topic is not clearly and obviously contrary to our mission, the deletion criteria should be "Well, I personally tried to find sources, and here's where I looked. I couldn't find any, perhaps we should delete this. As in AfD." Unfocused 13:29, 16 November 2006 (UTC)

This sort of thing really shouldn't happen, especially in an 'On this day' article; it should definitely be sourced by now. (I've added an {{unref}} tag; there are sources in the German and Polish interwikis if anyone with the language skills wants to follow those up.) If this proposal was in place the article would have been sourced when it was created, preventing this sort of problem cropping up. The criterion is not designed to delete such articles (that's why there's a 'grandfathering' clause), but to get people to provide the sources when it's written in the first place. Otherwise, there's no way of knowing whether the information is reliable. If that article was a hoax, would a user without a detailed knowledge of Polish history know? --ais523 13:38, 16 November 2006 (UTC)

It's obvious that the example provided is a good contribution. If the original authors still don't add sources, it's still a good contribution. Good contributions should always be welcomed, not rejected simply because they aren't spectacular contributions. No one should ever even consider deleting good contributions until they personally have made a good faith attempt to correct the flaws. Unfocused 13:57, 16 November 2006 (UTC)

Another very good example, the article has well over 50 edits and exists in several other language Wikipedias, I dont believe articles such as this should be speedy deleted. Addhoc 15:15, 16 November 2006 (UTC)

I've added a good source to the Jadwiga of Poland article. It took me about fifteen minutes to find and double-check it, but it wasn't really that difficult. I realize the article still needs additional references and inline citations, but I don't think that providing a single, solitary source is really that high of a bar for new articles. This policy might just encourage more editors to become better researchers. -- Satori Son 15:49, 16 November 2006 (UTC)

Or use google books more often ;). [1] Addhoc 16:18, 16 November 2006 (UTC)

I missed those - nice job! -- Satori Son 16:26, 16 November 2006 (UTC)

Glad to see the article get attention. However, as far as source quality goes, this only verifies a tiny fraction of the articles' content. Care to address the other 900,000 or so articles that need attention, and can you promise you'll do them all faster than the authoritarians who demand "do this now, I command it!" can nominate them for speedy deletion under this proposed policy?

I've found that greater than 10% of WP:PROD is stuff that I'm shocked to find nominated for deletion in the first place. By the time five days have past, a lot has been claimed from the dumpster, so if you want to see what I mean, look at the current day's PROD as it comes in. Even dumpster diving PROD doesn't address the fact that much of this stuff being in PROD is a complete waste of time. If the nominators would the proper review before nominating, such as checking the article history and talk pages, and doing a ten second search, it wouldn't be in PROD.

Why did I discuss the shortcomings of PROD here? Don't count on people being any better at checking first under THIS proposal. We'll lose a lot of good content, and we'll lose a lot of good editors over this if adopted. And we'll waste a lot of the time the editors who don't leave.

Do you think UncleG would've thought to write an article on Toe cleavage had it not been posted as an unsourced contribution? Unfocused 16:55, 16 November 2006 (UTC)

So let me see if I understand. You criticize my admittedly very humble contribution of providing a single source for Jadwiga of Poland as woefully inadequate, yet you strongly oppose a policy that simply requires that as a bare minimum?? As I stated, the requirement that a new article provide at least one single source, regardless of quality, is a low threshold indeed, yet one we should have embraced long ago. -- Satori Son 18:28, 16 November 2006 (UTC)

It was not my intention to criticize your good contribution, for that impression I am sorry. What I wanted to say is that even adding a good source doesn't necessarily solve the problem that this proposes to address: unsourced edits. There are hundreds of other edits there in the Jadwiga article not covered by your source. No proposal will be effective at addressing the real problem without destroying good content unless it is based on a "see what you can contribute" attitude rather than "see what you can cut out" attitude. Unfocused 19:50, 16 November 2006 (UTC)

THIS IS ONLY FOR NEW ARTICLES. We're not going to speedy delete every unreferenced thing in wikipedia. Night Gyr (talk/Oy) 17:35, 16 November 2006 (UTC)

So it's been changed. Most of my points still remain; this proposal changes the culture from open and accepting to closed and demanding. That's an enormous shift in attitude! There's already a backlog of articles that need sources. People who think unsourced articles are such an urgent problem should start by working on the backlog instead of looking for new ways to delete good contributions. The contributions that aren't good will find their way to a deletion forum anyway; we don't need to be so oppressive and bureaucratic about it. Unfocused 18:12, 16 November 2006 (UTC)

We've always had policies and guidelines that restrict what can be included. This is just saying that people shouldn't be allowed to add more unsourced material, just as we wouldn't allow them to add material that violated any other policy or guideline. Night Gyr (talk/Oy) 21:15, 16 November 2006 (UTC)

(Deleted your personal attack.) Sorry, I think assume good faith is far more important to making this project such a success than giving admins more power to delete willy-nilly. We still 'want people to add information. Unfocused 16:52, 19 November 2006 (UTC)

"Keep an open mind, but not so open that your brains fall out," which you removed, was not a personal attack, it was a comment on your advocated "open and accepting" culture. We allow in anything that meets our policies and guidelines, but we aren't an "indiscriminate collection of information," we have standards, and we don't let things in that violate policy already. This isn't really a change in culture. Night Gyr (talk/Oy) 18:33, 19 November 2006 (UTC)

Oh, please. What you said equates to "Don't disagree with me too much, or you might be retarded". That isn't a personal attack? Don't be so full of your own opinion that you can't see when you're being insulting. Unfocused 16:02, 20 November 2006 (UTC)

It's nothing personal, it's a widely circulated phrase among skeptics. "One should always keep an open mind, but not so open that one's brains fall out." If we're too open and accepting, we'll accept a lot of stupid crap, like all kinds of pseudoscience. Night Gyr (talk/Oy) 17:43, 20 November 2006 (UTC)

But we should have articles on pseudoscience; which include the consensus opinion that they are pseudoscience, and the reasons why. For my best try at what can be done with crank theories (although strictly this is pseudohistory), see Novus Ordo Seclorum#Conspiracy theory. Septentrionalis 18:31, 20 November 2006 (UTC)

Of course, but they, like all our other articles, need to meet certain standards, just like our thinking about them should follow basic standards of reason and evidence. Hence, not being too open. Night Gyr (talk/Oy) 18:40, 20 November 2006 (UTC)

It was still mildly insulting, and I support its removal (of course, now that we're discussing it in terms of whether it was insulting or not, the point is sort of moot). Please try to maintain a more civil level of discourse, especially when we're talking about making our encyclopedia more substantiated. -- nae'blis 18:50, 20 November 2006 (UTC)

I don't care how "widely circulated" a phrase is, if it's used in context as an insult, it's a personal attack. In this context, I think its obviously used as an insult. Also, I agree with Septentrionalis that it's important to have articles on pseudoscience. Unfocused 18:49, 20 November 2006 (UTC)

Another possible compromise

...which will remove all the talk of deletionists using this to get rid of articles. All we need is a clause that prevents this being used on any but very new articles (i.e. ones where the author might be reasonably expected to see the tag). The '50 edits' above is one idea, but I'd prefer a time delay; a couple of days would probably be fine, but this would work for my purposes (although probably not Dmcdevit's) if the period in which the tag could be added was as short as 15 minutes. (In that case, the 14-day requirement could probably be reduced.) Any thoughts? (I'm not sure I support this myself, but I'll at least submit it for consideration.) --ais523 13:48, 16 November 2006 (UTC)

Not sure if I understand you correctly, you're saying that articles older than 2 weeks should be exempt? If so, then such a modification would certainly result in a much less 'shock' to the community and could be used to phase this in gradually. That said, not sure I would vote for it either, as it possibly is too gradual, I think either 50 edits or 3 months could be more appropriate. Addhoc 15:39, 16 November 2006 (UTC)

That's the idea; articles older than a certain age (rather than just 'created before this rule came into force') would be exempt. The problem would be finding the correct age to strike a balance between making the limit useless and making the CSD useless. --ais523 17:12, 16 November 2006 (UTC)

dated unreferenced tags

Okay, I've done plenty of bitching about the idea above, but in no way do I disagree with the postulation that we need more sources on our articles. I just don't think a sword of Damocles is the way to do it.

So can we start with a system of dated {{unreferenced}} tags, similar to {{cleanup-date}}? It's embarassing to me that we have articles still in Category:Cleanup from August 2005, and if I knew more about some of those subjects (or had more time), I would help. Being able to marshall our forces to tackle the oldest tagged unreferenced articles seems like a more productive use of our energy than ramming through a speedy deletion criterion or overloading AFD with more articles we don't actually want deleted. I can probably make the category and template system myself, but I want to know I'm not wasting my time first. I think there's a great potential for positive change to come out of this proposal. -- nae'blis 15:43, 16 November 2006 (UTC)

• Regarding your example, I probably agree with Guy, it should have been deleted. Overall, I would suggest the consensus is that some form of proposal should be implemented. My only concern is that it should be very gradually phased in to avoid demoralizing editors who have worked hard writing articles. Addhoc 15:58, 16 November 2006 (UTC)
  • I will add that it seems clear that of 10 articles like Turun Balleti, nine will be sourced. Most people writing about their favorite club have access either to its website or an article in the local paper; and will put it in, as advertising. Septentrionalis 19:55, 16 November 2006 (UTC)
• My opinion is that dated {{unreferenced}} tags would be more successful than an expansion of speedy deletion criteria. {{unreferenced}} is a cleanup tag, not a deletion tag. It encourages one to attempt to clean the article up in the appropriate manner (i.e. add references), and only consider deletion if one then finds, as a result of the attempt, that such cleanup is impossible.

  That, in turn, leads to good AFD nominations ("The article cites no sources. I went and looked for sources (here) but couldn't find any.") instead of bad ones. It has been stated that part of the motivation for this criterion is improving AFD discussions. Making nominations like that is how one improves AFD discussions. No amount of "I believe that it could be sourced, keep." from editors who don't do any research is going to withstand a nomination where an editor shows that unlike those editors xe actually did the research and found that it couldn't in fact be sourced, and rationales from other editors who independently also did the research and came to the same conclusion. Equally, no amount of "It's unverifiable.", from editors who didn't do any research but simply looked to see if the article had any sources cited, withstands a rationale where an editor actually does the research and cites the sources that xe finds. Uncle G 18:57, 16 November 2006 (UTC)

I think this is a superb idea. Build up, don't knock down. The last think Wikipedia needs is a reputation for being a bureaucratic nightmare, I fear that's what it's getting though. Mallanox 19:39, 16 November 2006 (UTC)

Two alternatives

The following are not mutually exclusive. The first was my original proposal that led me to this (already existing) proposal. The second is a way to use AFD without too much instruction creep: other than allowing 14 days, I think it expresses current policy. Robert A.West (Talk) 17:20, 16 November 2006 (UTC)

Alternative #1: Limit this proposal to BLP

The above discussion convinces me that I had the right idea in the first place, so I will propose it again. The speedy delete for unverifiability should be limited to biographies of living persons, and the criteria expanded to include articles sourced only from self-published and/or notoriously unreliable sources (ala The National Enquirer).. The grace period should be shorter as well.

Why BLP? Because unsourced biographies are probably the biggest problem area in this regard. If we get rid of them, we probably get rid of half the actual junk. Also, because a false statement can harm a living person, even if it is not obviously defamatory.

Why only BLP? People who want this to apply to everything will get a chance to demonstrate that this can be done responsibly. People who think this will be a horror should see that there is a stronger case for BLPs, and if they are right, we will have done less damage. People in the middle can wait and see.

Also, I recommend a mandatory message, perhaps on a template, to be placed on the principle author (or last substantial contributor's) talk page.

Your unsourced biography of Jane Doe Smith was deleted because it was unsourced. To avoid possible legal problems, and out of a sense of decency, Wikipedia applies its policies on verifiability and citing sources very strictly to the biographies of living persons. If you believe that this action was taken in error, you can ask for deletion review. If you simply neglected to name sources, but have them, you can ask for undeletion so you can add them. If you are unfamiliar with how to do these things, or have questions, please contact the administrator leaving this message.

Even a rank newbie should be able to clue in to the fact that a biography of a living person is different from a treatise on the Phaistos disk. They should feel less bitten. Robert A.West (Talk) 17:20, 16 November 2006 (UTC)

Agree with the requirement for a warning (I put a link to the currently fictitious {{nsdel-warn}} on the {{nosourcedel}} template I'm proposing to use). The message should say how to contact the user who left the message (I don't see why they'd necessarily be an admin), to avoid confusing new users who don't know how. One last BLP-related problem that I've come across occasionally is when there isn't enough info in the article to know if a person is living or not, and there are no sources given to check... --ais523 17:15, 16 November 2006 (UTC)

We should make the warning mandatory, especially if we are requiring action (as opposed to simply removing a {{prod}} tag. This would be a different system, much lighter than AFD, but harder to blow off than prodding. Septentrionalis 20:00, 16 November 2006 (UTC)

I think this would be an excellent start, because it'll get a structure in place so we can see how well it works, and we're going to be applying it to articles that aren't really harmless to leave around unsourced, unlike something like a math concept. Perhaps we should do it for all biography articles, instead of just living people ones? Night Gyr (talk/Oy) 17:30, 16 November 2006 (UTC)

I agree that having a trial run with biographies would be a good idea... Addhoc 17:39, 16 November 2006 (UTC)

If we need to restrict this to biographies to start with, so be it. It will give us a better idea of whether this can be sensibly extended to maths articles. --ais523 17:44, 16 November 2006 (UTC)

Fully support starting with this and seeing what happens. It will be a good start. John Broughton | Talk 18:29, 16 November 2006 (UTC)

Oppose major change in wiki-culture because editing and PRODding the articles in question already accomplishes the same goal. Unfocused 18:58, 16 November 2006 (UTC)

Alternative #2: New procedure within AFD

There are four problems with using AFD to delete articles that are unverifiable.

1. There are no agreed upon criteria to demonstrate unverifiability, as opposed to merely lacking sources.
2. There is no agreement on the amount of effort the nominator should invest before nominating. I have seen AFD's go through where the nominator admitted to only a quick Google search. Others have been accused of bad faith and insufficient time despite extensive searches.
3. If an article has been left by its creator for some time, five days may not be enough time to find someone who knows how to find a source.
4. My pet peeve of "Keep and find the sources somewhere" response.

So, we can fix these, without creating a whole new procedure.

Articles nominated on AFD as unverifiable and/or original research may be specially tagged (using a template TBD), in which case the following special procedures apply.

1. To reduce confusion, there should be no other reason for the nomination. If there are other problems with the article mandating deletion, it is not bad faith to renominate for those reasons if the article is determined to be verifiable. An article may be nominated even if it has sources, if the elimination of unsourced statements would not leave a meaningful article.
2. The nominator must have reason to believe that no sources exist. Generally, this means one of the following.
   1. The nominator has made a significant effort to find sources, or is aware of such efforts by others. Efforts need not be heroic, but should be more than a single Google search.
   2. The article has been tagged as lacking sources or as original research for at least 90 days.
   3. The nominator can cite sources (including a well-sourced Wikipedia article) that cast serious doubt on the legitimacy of the subject matter.
3. The burden is on those who would keep the article to add sources, or demonstrate that the current sources are adequate. Arguments based on considerations other than verifiability and the avoiding of original research are out of place and should be ignored by the closing admin.
4. The case will be considered for 14 days, rather than the standard five. Extreme care should be exercised when closing such a case prematurely.
5. Misuse or subversion of this process will be treated as a serious disruption of Wikipedia.

This buys into the existing AFD structure, gives enough time for consideration, establishes some criteria and doesn't raise the spectre of "speedy" deletion. Submitted for your approval or brickbats. Robert A.West (Talk) 17:18, 16 November 2006 (UTC)

This process looks pretty much independent of AfD to me, although it's similar. DfV ('Deletions for Verifiability') was proposed as a name for a process like this on WT:CSD, so that seems reasonable. I have no idea how many people will bother to participate in something like this, so it might be a failure, but I don't think it could do any harm. As for your last point about misuse or subversion, you can probably leave that out as it's true for just about any process... --ais523 17:22, 16 November 2006 (UTC)

DfV was a winking reference to the old VfD. My reason for deciding that it should not be a separate process from AFD is twofold. (1) Dividing attention between AfD and DfV would make mistakes more likely in each. That split can always be done later if volume demands it. (2) This almost describes the current AFD policy when unverifiability/OR is the reason for nomination. The only major new feature is a 14-day period. It could be dispensed with, but we would get reasonable pleas for additional time. One can still nominate for unverifiability and other reasons, but the 14-day period doesn't apply and other arguments are obviously valid. I have found it useful to avoid other considerations when I think an article is unverifiable -- I need no other argument, and it gives me a basis to challenge a keep for no reason in DRV. The requirement that the nominator take some responsibility for checking if sources can easily be found is just, in my mind, decent behavior. Robert A.West (Talk) 17:46, 16 November 2006 (UTC)

There are some key differences: DfV is forced to stay on topic, nominators would be forced to concentrate on the nom due to policy, rather than custom, and it would default to delete, not keep; this last is a very significant difference. --ais523 17:49, 16 November 2006 (UTC)

I quote from Wikipedia:Deletion guidelines for administrators, "Note also that the three key policies, which warrant that articles and information be verifiable, avoid being original research, and be written from a neutral point of view are held to be non-negotiable and cannot be superseded by any other guidelines or by editors' consensus. A closing admin must determine whether any article violates such policies, and where it is impossible that an article on any topic can exist without breaching these three policies, such policies must again be respected above other opinions." I suppose that this proposal is designed to convince people not to ignore the rule, but IAR is for emergencies. Robert A.West (Talk) 21:58, 16 November 2006 (UTC)

• It depends what one means by "default". If the nominator fails to make a prima facie case for lack of sources, the article should be kept. Septentrionalis 23:09, 16 November 2006 (UTC)

• How about we just log things onto a page, as with WP:IFD or WP:TFD, and if any discussion is necessary, it can take place under the entry? There aren't going to be as many unsourced articles as unsourced images, so we don't need to worry about an overwhelming number of entries. Night Gyr (talk/Oy) 18:22, 16 November 2006 (UTC)

Changing the default from keep to delete changes Wiki culture from "assume good faith" to "prove you're good enough first" We'd better have a damned good reason to do so. For biographies of living persons, the threat of legal action does require immediate action when potentially negative content is added, but even in the case of biographies, editing the article already solves the problem! Unfocused 18:27, 16 November 2006 (UTC)

• AGF has nothing to do with it. It is the simple fact that you can't prove a negative. I refer you to WP:V#Burden of evidence. This is long-standing policy. As to editing, note the criterion, "If the elimination of unsourced statements would not leave a meaningful article." If editing will work, none of the forms of the proposal will apply. Nomination is an assertion that, after editing out policy violations, the article would be a speedy deletion candidate. Robert A.West (Talk) 21:33, 16 November 2006 (UTC)

Quaere: What about Simply connected space, a nomination of which deserves the reply: "See any elementary point-set topology textbook"? Septentrionalis 20:05, 16 November 2006 (UTC)

• • If the article was just created, I'd say "cite the one you used." Night Gyr (talk/Oy) 21:13, 16 November 2006 (UTC)
    • In the long run, even if this proceedure is not expanded, it will be applied to articles which are long established. If this were strictly to apply to new articles, that would decrease the costs. Furthermore, it is quite likely that this article wasn't created from a textbook; it's probably a segment of a longer article. Septentrionalis 23:06, 16 November 2006 (UTC)
  • I would recommend a "Suggested Reading" section at the end, pointing the reader to several decent elementary textbooks. That is the method used by other encyclopedias. No one is suggesting forcing such articles to use footnotes, which is, I believe, your real fear. Robert A.West (Talk) 21:33, 16 November 2006 (UTC)
    • Footnotes would be dumb; but I think there is a reason why a random textbook hasn't been flung down at the end of this article: which one? Septentrionalis 23:06, 16 November 2006 (UTC)
      • So, like A.A. Milne's shipwrecked sailor, our mathematicians find themselves unable to decide which source to cite first, and so threaten to cite none at all until the article is deleted? I'm whelmed. Robert A.West (Talk) 04:47, 17 November 2006 (UTC)

Mathematical articles

Simply connected space has ingredients not uncommon for mathematical articles:

• 90% of it asserts propositions obvious to any person learned in the art.
  • Of this 60% or 70% consists of rigorous mathematics
  • 20% to 30% consists of informal ways of conveying the general idea to readers who don't want to go through the 60%
• 10% is important short general theorems, which are not obvious.

There are dozens of elementary topology texts which contain much or all of the 60%; there are dozens of popular books which contain much or all of the 30%. I could date myself by putting down the textbook I learned this from; but why should I put down that one, when any other, including books I haven't seen, would do as well?

Few books, if any, will include both the 60% and the 30%; if the author is going to "do it right", he's not going to spend much time on vague metaphors. We have to do both, in one article. Septentrionalis 23:06, 16 November 2006 (UTC)

There are two reasons that pointing out the existence of an article that appears not to need a reference misses the point, and is not relevant to this discussion.

• I have yet to see a good example of that article. What you're asking for here is a redefinition of our WP:V/WP:CITE policies. For instance, the content of that article is certainly not obvious to me, and not everyone reading an article is learned in the art. "'Verifiable' in this context means that any reader should be able to check that material added to Wikipedia has already been published by a reliable source." To answer your question: yes, any of those sources would work, please put one. That all mathematical texts say it does not logically mean that we need not give any as a source; you still have to read the text to learn it.
• This criterion, like all criteria, does not mandate the deletion of all articles it may apply to. Rather, it permits the deletion of such articles by administrators, if they deem it appropriate. If any administrator were to come across such an acceptable unsourced article, they need only remove the tag and not delete. Administrator's have the button because they are trusted with the judgment to do so. Dmcdevit·t 00:37, 17 November 2006 (UTC)
  • I trust your judgment, and that of most admins. Septentrionalis 04:40, 17 November 2006 (UTC)
    • Readers who will find the article informative will benefit from the article will benefit from a bibliography. If the article is being written for other skilled mathematicians, then the article has the wrong audience in mind. As for not finding the 60% and the 30% in one source, articles are supposed to have multiple sources, so I don't see what the issue is. Robert A.West (Talk) 04:58, 17 November 2006 (UTC)

175kb with no consensus in sight

The commentary on this proposal currently stands at over 175kb on this page alone - more when you include the discussion at WT:CSD. The discussion seems to be stalemated to the point that we are repeating arguments and rebuttals. It's an awful lot for anyone new to the conversation to wade through. I have attempted to distill the pros and cons down to a single summary. Please make changes or corrections as necessary to represent the arguments as you see them.

Please be concise and please don't change or delete an argument just because you don't agree with it. Rossami (talk) 01:28, 17 November 2006 (UTC)

Arguments in favor Unsourced articles are an urgent issue needing immediate resolution Unsourced articles jeopardize the credibility of the project Requests for sources often go unanswered Current processes default to keeping unsourced articles with no mechanism for mandating sources The threat of deletion will result in the sourcing of many articles This puts the burden of sourcing on the original contributor -it encourages editors to do the work of sourcing themselves, rather than expecting others to do it later. AFD is too slow or is ineffective at dealing with this problem We need to change the culture of Wikipedia in this area Topics which can be sourced can be easily undeleted It is a logical extension of the WP:V policy that states "The obligation to provide a reliable source lies with the editors wishing to include the material, not on those seeking to challenge and/or remove it."

Arguments opposed Unsourced content is a real but not urgent problem Deletion is too severe a remedy for the problem - lack of sources should be dealt with via cleanup "Speedy"-deletion does not put enough eyes on the problem Unsourceable or persistently unsourced articles can be deleted through the Prod or AFD processes - this new proposal is unnecessary Some users enjoy finding and adding academic sources, others find such work boring and will leave the project if forced to do such work Some topics are unnecessary to cite to a single source because they are widely known by any practitioner in the field If mistakes are made, undeletion is not easy and will confuse or drive away some proportion of new users Some articles will be deleted despite being sourced or sourceable, even with the best intentions Represents a negative shift in openness of Wikipedia's culture May produce inaccurate, careless, or false citations Will be used more aggressively than intended Encourages editors to demand work from other editors rather than encouraging them to do it themselves While the spirit of the plan is good, it is wrong to start with a CSD variant. It would be better to start with an AFD variant, as that will more effectively and less contentiously move the community culture toward sourcing. Incomplete articles are where new users start. Often new users won't feel comfortable editing a fully developed and referenced article.

Seems like an adequate summary. I would point out that the amount of people here raising objections still seems to be small, so "no consensus in sight" might be a bit overraction. Some of the objections are good, and can be solved by writing the concerns into the proposed wording. For example, we can add a sentence ensuring that this criterion is inapplicable to an article with any reference or external link at all. If you are weighing a source's reliability, you aren't using this criterion. Some other objections are either fundamental or missing the point, but there's no reason to think they couldn't (or might not already be) outweighed by the support of others. Dmcdevit·t 01:49, 17 November 2006 (UTC)

How did you quantify "I would point out that the amount of people here raising objections still seems to be small"? And "but there's no reason to think they couldn't (or might not already be) outweighed by the support of others" does not seem to fit in with the usual Wikipedia definition of consensus. --Philip Baird Shearer 11:27, 17 November 2006 (UTC)

Regarding the suggestion we should give this proposal a trial run in regard to BLP articles, I would suggest there is a consensus this seems to be a worthwhile idea. Addhoc 11:37, 17 November 2006 (UTC)

I haven't really weighed in regarding a trial run, but I am now - I'm opposed to it. There's no pressing need for it, and I think there's a lot of issues that need to be addressed better before we dive in. --badlydrawnjeff talk 11:43, 17 November 2006 (UTC)

I'd be LESS opposed to a trial run in BLP than elsewhere, but I don't think it's necessarily a good idea there either. However with CS #A7 being used so strongly against articles without claims of notability, it would be a relatively smaller expansion of scope. Still not sure that you can demonstrate consensus; where has this been advertised? Is it on the Centralized Discussion page yet? I just yesterday mentioned it to the Fact and Reference Check WikiProject. -- nae'blis 15:00, 17 November 2006 (UTC)

Increase the number of people raising objections by one. While I have nothing to add that hasn't been stated more eloquently already, I am firmly and eternally opposed to this proposal. -Toptomcat 17:47, 11 December 2006 (UTC)

Advertised, or at least linked, at the following venues (not exhaustive):

• Wikipedia:Requests for comment/Policies
• Wikipedia:Village pump (policy)
• Wikipedia:Village pump (proposals)
• Wikipedia:Village pump (all)
• Wikipedia talk:Criteria for speedy deletion
• Template talk:Unreferenced
• Wikipedia:Requests for comment/All

Addhoc 16:02, 17 November 2006 (UTC)

What clauses?

I'm trying to systematically enumerate various proposals for what the criterion should say here. Each one should probably be checked to see what effect it would have, whether it would be useful, whether it would be instruction creep, and which of the opposes/supports in the table above it would invalidate. I invite discussion on each individual criterion and additions. --ais523 13:28, 17 November 2006 (UTC)

The basic ideas

Any article without references may be deleted.

The basic idea of the criterion. Each clause below limits the ability of this to apply in certain ways.

Restrictions in the current proposal

The article must have had a specific tag for 14 days.

The article must have been created after this criterion was adopted.

Other proposed restrictions

• 1. The article must have less than 50 edits.
  • or 1A: The article must only have one or two substantive editors.
• 2. The article must be newer than two weeks old.
• 3. The article must not have survived an AfD.
• 4. The article must have been listed on an AfD-like debate for 14 days.
  • or 4A. Nominations under this proceedure may be replaced by AfD nominations, if an editor has a reason why an unsourced article should be kept.
• 5. The article must not have interwiki links.
• 6. The article must be a BLP.
• 7. The article must be a biography which gives no reason to suppose the subject is dead, such as being alive before 1900. amended 21:04, 17 November 2006 (UTC)
• 8. Articles which 'don't need sources' in some sense are exempt.
  • 8A In particular, disambiguation pages are not covered.
• 9. The user who created the article must be warned on their Talk page.
• 10. The article can't be deleted, only stubbified.
• 11. An article that could be deleted under the other restrictions must be AfD'd first.
• 12. An article deleted under this proposal is undeleteable as soon as a source is given.
• 13. The tag can only be placed/AfD-like debate started if (a) there is doubt about the source, (b) the nominator has made a proper search to find sources, or (c) the article has been unreferenced for 90 days.

Proposed generalizations

*14. The article needs at least 3 sources. This was my (extreme, admittedly) proposal, and I withdraw it. John Broughton | Talk 21:50, 17 November 2006 (UTC)

• 15. The sources must be 'reliable' in some sense.

Comments on "What clauses?"

Clause comments by John Broughton

For the generalizations (14, 15), the first is mine, and in the spirit of starting with the least controversial implementation of this proposal, I think it should be dropped. For 15, I think that should also be dropped. An editor always has the option of deleting a source that fails WP:RS and then tagging an article as lacking sources.

For the restrictions:

• 2 - I can't find that in the text above, and in any case I don't think it makes much sense, so I suggest it be dropped.
• 4 - would it suffice simply to have a page where all tagged pages (by date) are listed? The AfD-like "process" would simply be that any editor can add a source and remove the tag, in that case. (The devil is always in the details; such an AfD "process" would have minimal overhead and not compromise the proposal, while an "AfD-like process" that involved, say, voting, would radically change the proposal.)
• 6 and 7 seem essentially the same - clarification?
• 8 clearly needs a lot more explanation; in general, seems unworkable as is.
• 10 isn't consistent with this proposal and I suggest it be dropped - it's not a restriction, and it's not clear who would "stubbify" articles.
• 11 - needs further elaboration (at minimum, an example); seems to unduly complicate this in any case
• 13 - (a) isn't consistent with the proposal - if there is a source, an article shouldn't get this tag. (see also 15) Suggest dropping. I do like the combination of (b) and (c) - if the article is less than 90 days, the nominator has to do a proper search or give the article more time to ripen. (A proper search might find a few sources that indicate that the subject is not notable, leading to a regular AfD, which would still be a useful expenditure of an editor's time).

Thank you for taking the time to compile the list - very helpful, I think. John Broughton | Talk 14:04, 17 November 2006 (UTC)

Yes, thanks also. I would suggest this overall proposal is fairly radical and consequently many of the suggestions relate to initially restricting how this would function, to gauge community reaction. Addhoc 14:43, 17 November 2006 (UTC)

:The second point is for legacy articles (which there are many) - it's to prevent this process from getting overloaded with by someone nominating everything in sight, thus getting some articles deleted just by sheer weight of numbers. ColourBurst 15:19, 17 November 2006 (UTC) The main point addresses this, so I'm striking out my comment. ColourBurst 15:53, 17 November 2006 (UTC)

Clause comments by nae'blis

As much as I dislike this proposal, I'll weigh in on the above proposed restrictions/clauses:

• 1. The article must have less than 50 edits. - No. Not sure why this is relevant.
• 2. The article must be newer than two weeks old. - Undecided; "Created after January 1, 2007" seemed a better criterion, but this may actually be overly restrictive if the goal is to deal with the thousands of unsourced articles already in existence. However better to deal with newer articles first and prove it works constructively than open the floodgates immediately.
• 3. The article must not have survived an AfD. - No, due to the "Keep but source" outcome people above have objected to. Unless we change the culture of AFD to neutralize that possible outcome, this may backfire. On the other hand, a "clear keep" might be a good escape clause.
• 4. The article must have been listed on an AfD-like debate for 14 days. - Neutral.
• 5. The article must not have interwiki links. - Irrelevant.
• 6. The article must be a BLP. - Maybe. (see above regarding trial runs)
• 7. The article must be a biography. - See #6.
• 8. Articles which 'don't need sources' in some sense are exempt. - No, so long as we're counting "Further reading" and "References" as sources, not just explicit footnotes. All articles can and should have some sort of basis outside of the encyclopedia. Edit: Per conversation below, a common-sense exception for disambiguation pages wouldn't hurt, but seems superfluous, as they're not really articles in the same sense that redirects aren't articles.
• 9. The user who created the article must be warned on their Talk page. - Somewhat against this; better phrasing might be "major contributors" but Watchlists/sufficient lead time for people to notice the CSD tag should make this moot. On the other hand, we are supposed to warn image uploaders, but those are much more difficult to "edit" collaboratively.
• 10. The article can't be deleted, only stubbified. - No. If you can save something for a stub, it should be sourceable too, which throws the article/stub beyond the scope of this CSD, technically.
• 11. An article that could be deleted under the other restrictions must be AfD'd first. - Huh?
• 12. An article deleted under this proposal is undeleteable as soon as a source is given. - Yes.
• 13. The tag can only be placed/AfD-like debate started if (a) there is doubt about the source, (b) the nominator has made a proper search to find sources, or (c) the article has been unreferenced for 90 days. - No. For it to be a CSD there should be relatively little wiggle room. "Doubt about the source" is the exact sort of scope creep some opposers have referred to. Even A7 doesn't rely a sourced claim of notability, just a claim.
  1. 14 and #15 have has already been withdrawn so no comment there. -- nae'blis 15:12, 17 November 2006 (UTC)
• 15. The sources must be 'reliable' in some sense. - No (since it was re-added). We don't speedy hoaxes unless they are so badly written to rise to the level of pure vandalism. We don't speedy A7-candidates if we don't judge the claim to be "notable enough"; this is too much discretion that can and will be misused, I believe. Look at how T1 and G11 are used, and they're supposed to be pretty clearcut! -- nae'blis 21:41, 21 November 2006 (UTC)

You're never going to be able to deal with No. 2 unless you remove the restriction that articles created before the policy are exempt, and that's never going to happen (because that's going to remove quite a bit of content). I do think that newbies are going to complain when you remove their article without a source, and they're inevitably going to point to an article that's already existing and say "look, there's an article without a source and it's not been deleted!" ColourBurst 16:00, 17 November 2006 (UTC)

Yep, hence my comments above that this would be better as a non-deletion process. -- nae'blis 16:21, 17 November 2006 (UTC)

#8: "Articles which 'don't need sources' in some sense are exempt." I thought this was about e.g. disambiguation pages. Are you really saying these need sources too? Eugène van der Pijll 16:33, 17 November 2006 (UTC)

Ahh, I was thinking of the simply connected space example from above. I don't consider disambiguation pages to be "articles" at all, really, but you're right that something like A3's exception wouldn't hurt. -- nae'blis 16:55, 17 November 2006 (UTC)

Yes, but A3's exception is not quite enough. Mentioned above are e.g. timelines, and I can add lists to that. Some lists definately need sources (often lists of (living) people), others are merely navigational. For example: none of the LoPbN lists hav sources, and they don't need it either. Eugène van der Pijll 17:03, 17 November 2006 (UTC)

Clause comments by ais523

After taking the time to compile the list, I'll comment on it now.

• 1: Too easy to game, not very relevant.
• 2: Would change the whole nature of the proposal; this would be useful in getting authors to provide sources, but not otherwise.
• 3: I don't think this meshes much with the spirit behind the proposal.
• 4: Overkill for me. There's a PROD-like process behind the tag which creates a central listing.
• 5: Either it's sourced at the other wiki, in which case copy the sources, or it isn't, in which case this isn't useful.
• 6 and 7: BLPs are more urgent. The '7' variation (which I proposed above) is because in some cases there's little context in a biography and no source, so you can't tell whether the person is living or not.
• 8: I'm not too clear on what this means either. The dabpage point above is a good one; dabs and redirects probably don't count as articles (tellingly, the CSD for a 1-link dab page is G6, not A-anything, and redirects have their own CSDs).
• 9: I fully support this one, to help avoid WP:BITE problems, and because it's the original author who was using sources in the first place. I'm happy with wording tweaks, however.
• 10: Defeats the point of a CSD.
• 11: Why not just AfD the page anyway?
• 12: Similar 'safety valve' to PROD; not allowing this would make no sense to me (in fact, I placed 12 in the proposal for a while).
• 13: Too uncontroversial for a CSD.
• 14/15: Withdrawn, but would have been bad ideas anyway; there's no point in trying to expand this too early.

--ais523 17:05, 17 November 2006 (UTC)

On #13, do you mean too controversial? -- nae'blis 19:02, 17 November 2006 (UTC)

Thanks, and now corrected. --ais523 09:13, 20 November 2006 (UTC)

Other comments by User:Pmanderson

• 2. The article must be newer than two weeks old. Useful; on January 15 it will be more restrictive than created after January 1, 2007. This increases the chance of getting the creator, who knows where he got it, to source it. I would deprecate using this as a way to delete unwatched articles; that's one of the costs of {{prod}} - an acceptable one there.
• 3. The article must not have survived an AfD. Probably desirable; CSD's should be for articles which have a snowball's chance of surviving AfD. Since it's prospective, no article which has now had a "Surely, there must be sources somewhere" AfD falls under this process anyway.
• 4. The article must have been listed on an AfD-like debate for 14 days. No, this should be lighter than AFD, or what's the point?
• 4A Nominations under this policy may be replaced by AfD nominations, so that reasons to keep can be explained.
• 6. The article must fall under BLP. This is the obvious first step, and where this is most needed. Let's start here.
• 7. The article must be a biography. Amended 7 per Ais523's explanation; I suspect this would still fall under BLP.
• 8A. Disambiguation pages are not subject to this policy. If we don't say this, some good soul will do it.
• 9. The user who created the article must be warned on their Talk page. The proposed tag provides a mechanism for this; let's use it. However, all we can do, in practice, is recommend; I don't think this is enforceable.
• 12. An article deleted under this proposal is undeleteable as soon as a source is given.Are we agreed on this? It's in the proposal now.
• 13. The tag can only be placed/AfD-like debate started if (a) there is doubt about the source, (b) the nominator has made a proper search to find sources, or (c) the article has been unreferenced for 90 days. Unenforceable; but the wording should suggest looking for sources before tagging; as more helpful to WP.

I have two substitute suggestions, which fulfill some of the purposes of some of the above.

• 14/15 If withdrawn, they should be struck. I am opposed to both: Some articles are completely sourced from two references; whether such articles should be merged or deleted is a judgment call, and should be done by discussion. Whether sources are reliable is also a matter of judgment, and should also be done by discussion. Septentrionalis 20:49, 17 November 2006 (UTC)

Comments by Colourburst

(only 2 I can think of, most people have gotten the rest)

• 5: interwiki links can be a good place for sources, but they can't be the sole support for an article. I don't think we have enough cross-wiki communication at the moment, though, it might be a good point. ColourBurst 05:23, 18 November 2006 (UTC)
• 15: "reliable" has no place in an undebatable deletion process. If it were to be debated in the 14-day process instead, it would be just another AfD.

That's it, pretty much. ColourBurst 05:24, 18 November 2006 (UTC)

General comment on clauses by Sjakkalle

If a rule has too many clauses, people will not them, see m:instruction creep. People will ignore the clauses and go ahead and do what they like with the criterion. Sjakkalle (Check!) 10:34, 18 November 2006 (UTC)

Compromise wording

One of the strongest concerns expressed so far seems to be that articles that validly exist without sources are still in danger according to the current wording. My response has been that such articles aren't really in danger because administrators may delete articles that qualify for CSD, but are not required to. We would expect anadministrator to recognize such articles and remove the tag without deleting. There's no reason not to make this explicit, however, and I hope my last edit goes some ways towards allaying such fears: [2]. Dmcdevit·t 06:05, 18 November 2006 (UTC)

With about 500 admins on Wikipedia, chances are that someone will interpret this more liberally than ever intended and go ahead and delete things where the proper remedy was to look for a source. Some weeks back we had one admin who interpreted the "spam" criterion G11 as a license to delete all sorts of cookie products without any discussion, even if the article was worded neutrally. Sjakkalle (Check!) 10:13, 18 November 2006 (UTC)

If 90 percent of the articles that get the tag actually deserve the tag, and for the 10% that don't, 90% of the time an editor recognizes that and removes the tag, then 99% of the time, an article that should be deleted for lack of sources WILL be deleted, with a 1% error rate. (You can play with the %s, of course - 95% correct tag and 80% correct removal, for example, is also 99% overall correct rate.)

And that if that 1% error rate involves articles that (a) are retrievable by some process and (b) were short, so not a lot of content was lost in any case, then I think there is a strong case that the benefits of this policy (freeing up editors from involvement in AfDs, so that they can spend their time more productively elsewhere) is much greater than the negatives here. John Broughton | Talk 16:43, 18 November 2006 (UTC)

Please slow down and rethink your statistics; you've confused 90% of the articles that get tagged should be deleted with 90% of the articles that should be deleted do get tagged, a very different claim. In addition, this may not actually do much to reduce cruft: John Powers (poet) was effectively tagged {{source or die}} and is now cruft with a source, immune to this CSD. Septentrionalis 16:58, 18 November 2006 (UTC)

Trial run

Much of the concerns, and support, are necessarily speculative. I want to propose that we begin a month-long test run sometime in the future, to observe whether the problems we fear and the benefits we look forward to occur. We are arguing whether or not things will happen certain ways right now, when we could simply find out. At the end of the trial period, the application of the CSD should stop, without exception, to allow assessment of the criterion to resume.

Problems we might look for that only a trial period is likely to reveal:

• How easy is undeletion for recreated articles?
• Will it be used more aggressively than intended?
• How often will mistakes (articles for which sources are not necessary) be made?
• Does biting appear to be at an unacceptable level?

Thoughts? Dmcdevit·t 06:25, 18 November 2006 (UTC)

• Support, the only path to finding and fixing the problems. --May the Force be with you! Shreshth91 06:37, 18 November 2006 (UTC)
• I'll get a template etc going. Can't get it to work, but support none-the-less. Daniel.Bryant ^{[ T · C ]} 06:56, 18 November 2006 (UTC)
• Support trial run. --Connel MacKenzie - wikt 08:23, 18 November 2006 (UTC)
• Oppose. Will do too much damage to the encyclopedia's integrity, even if it's just a trial run. I've just recently come across this discussion, and I agree with the people opposing this. People will be laughing at us if we start deleting new articles on real towns and cities just because there happens to be no source or external link provided when a quick Google check will verify at least the existence for you. WP:V is the core policy, this enforcement is an overkill. Sjakkalle (Check!) 10:11, 18 November 2006 (UTC)
  • We don't jst aim for articles on things that exist, but accurate and neutral articls on them. Also, proof that they exist is nice. If a simple Google check will verify it, then 14 days is plenty for the source on Google to be added. What's the problem? Dmcdevit·t 11:02, 18 November 2006 (UTC)
    • If a simple google search will provide a source, it's not exactly a burden for the creator of the article to find one. And as soon as a trial run starts, this is exactly what will happen. jguk 11:24, 18 November 2006 (UTC)
• Oppose. I think this proposal is not in accordance with some of the most basic fundaments of Wikipedia. Although citing sources is undesirable, it has not been proved that uncited articles (other than those that can be deleted using current CSD rules) are likely to be incorrect. Just marking them unreferenced is of more service to our readers. Eugène van der Pijll 10:49, 18 November 2006 (UTC)
• Support. A trial run, which is all this is a request for, is likely to prove that the concerns raised are unfounded. As we can undelete things easily if necessary, there really is little downside risk in it. jguk 10:57, 18 November 2006 (UTC)
• Oppose even the trial run. We already know from our parallel experience with other deletion criteria that some of the concerns stated above are serious and real. The case has not yet been made that this problem requires such draconian measures. Rossami (talk) 16:29, 18 November 2006 (UTC)
  • What do you think will happen if we tag and then, if necessary, delete articles that don't provide a reference? Creators will soon get into the habit of always providing a reference, won't they? Within 4 weeks, very, very few articles will ever be deleted under this rule. jguk 16:51, 18 November 2006 (UTC)
    • I think that you will rapidly select for those editors who enjoy writing academic articles and providing painstaking detail. Many other good contributors will be put off by this overly-harsh response to their good-faith contributions and will not return to the project. While we need detail-oriented editors, the project is better off when we can take full advantage of the diversity of our good contributors. Rossami (talk) 21:52, 18 November 2006 (UTC)
      • Maybe so, but there is no way to know that without a test run. Considering that the lag is 14 days, a sort trial can hardly do much damage under the worst of circumstances. If it turns out badly, the trial will show that and it will be scrapped easily. We might argue what could happen endlessly. Dmcdevit·t 22:47, 18 November 2006 (UTC)
        • I'm not sure that the trial will be able to prove or disprove the points in question. How do you propose to measure the impact on culture? I and others have raised the concern that it could be severe. Supporters think it will be negligible. Regardless of who's right, any impact will be delayed. How do you propose to measure it? How will we know how many good editors were run off during the pilot? Rossami (talk) 23:07, 18 November 2006 (UTC)
• Completely Support the trial run. The proposal seems quite flawless in theory and having a practical trial will iron out any minor unforseen glitches that come up. --Srikeit 19:16, 18 November 2006 (UTC)
• Support. From reading the discussion I feel that a lot of the concern is unfounded and will prove as such in a trial run. An article without any sources is qualitatively different from an article which has 1 sourced statement and many unsourced ones. For a town, a link to Google maps showing its existence would probably count as a reference (though it gives 0 information beyond location).--Nilfanion (talk) 22:58, 18 November 2006 (UTC)
• Support with qualifications: Let's make this as narrow as possible: BLP articles only, and lack of any wikilink or external link whatsoever, in whatever form. But in general, we can debate until the sun burns out on what a proposal will do, but if it looks like the benefits are likely to significantly outweigh the costs (as I believe is the case here), it's reasonable to actually go and test it and found out whose hopes and fears and concerns and expectations are right and whose are wrong, and maybe even things that hadn't been anticipated. And we can learn from that, and continue to improve, and if this doesn't work, then we can look for other ways to get to a higher quality encyclopedia. John Broughton | Talk 00:38, 19 November 2006 (UTC)
• Oppose as I did above. Way too many questions at this point to even think about a trial run. --badlydrawnjeff talk 04:19, 19 November 2006 (UTC)
• Support for biographies, the required level of referencing for articles about cases, books, and bands is iffy, so I think we'd be best off starting with articles that definitely need sources. Night Gyr (talk/Oy) 04:21, 19 November 2006 (UTC)
• Strongly oppose. This is bureaucracy at its very worst, and hasn't come anywhere near consensus. Unfocused 16:47, 19 November 2006 (UTC)
• Support would prefer a limited trial run for biographies to gauge the wider community reaction, but would still support a full trial on the understanding if there were serious concerns we could pull the plug. Addhoc 16:55, 19 November 2006 (UTC)
• Oppose. Far too much deletion going on already, without encouraging it. Create, don't destroy. We're here to build an encyclopaedia, not to provide as many reasons as possible why people's hard work should be deleted. -- Necrothesp 17:04, 19 November 2006 (UTC)
  • The point isn't to actually get rid of anything, if possible, it's to remind creators to include sources, which leads to building a better encyclopedia, not a worse one. Night Gyr (talk/Oy) 18:16, 19 November 2006 (UTC)
    • "We're not nominating for deletion in order to get rid of anything". Does this sound like WP:POINT to anybody but me? ;->Septentrionalis 23:38, 19 November 2006 (UTC)
• Oppose. It just encourages the creation of spurious citing, and the deletion of articles that while important, no one is watching. Titoxd^(?!?) 21:56, 19 November 2006 (UTC)
• Oppose --Francis Schonken 22:27, 19 November 2006 (UTC)
• Oppose for any articles other than BLPs. There's too much potential for damaging the quality of the project if we get this wrong. JulesH 23:09, 19 November 2006 (UTC)
• Oppose except on conditions The fundamental assumption here is that article creators will learn to provide sources if we are draconian enough. A factoid has been going through WP that most of our content, especially new articles, is provided by newbies and anons. They aren't in a position to learn anything; they'll just wonder why their article disappeared. If this is done at all, it must be for BLPs and it must apply to articles less than 14 days old, so we have some hope of catching the newbie before he gets bored with WP. Septentrionalis 23:15, 19 November 2006 (UTC)
• Oppose. ^BlankVerse 12:11, 21 November 2006 (UTC)
• Oppose - trial run of which version?? Too many questions still to even try this out, and I have concerns about what methodology would be used. -- nae'blis 21:43, 21 November 2006 (UTC)
• Oppose any run, trial or otherwise. I believe this proposal is contrary to the purpose of the principles of verifiability and no original research. These principles are intended to prevent the inclusion of crackpot science, original historical analysis and other such junk, not to promote the deletion of material that is accurate and uncontroversial.dryguy 23:13, 21 November 2006 (UTC)
• Support CSD for BLP only. There it is an urgent and strong matter. On consideration, I think an alternative XfD process is better for general verifiability/OR problems. Robert A.West (Talk) 00:13, 22 November 2006 (UTC)

• Support We are being overwhelmed by unsourced articles, and the threat of deletion is probably the only thing that will cause most editors to bother adding a source. -- Donald Albury 21:29, 22 November 2006 (UTC)
  • Then threaten; see the proposal at MediaWiki_talk:Copyrightwarning#Unsourced_content. Most of the people writing unsourced articles won't know about this. Septentrionalis 18:00, 23 November 2006 (UTC)
    • {{unsourced}} and {{cn}} tags in articles now are more likely to be ignored, removed and/or provoke complaints from editors about why am I cluttering up the page than to actually result in sources being added. I do remove material that has been tagged, after waiting for a month or more, but what do I do about an article that has NO sources, even after the {{unsourced}} tag has been sitting at the top of the page for months. I have found sources for a few articles like that, but I have trouble finding time to work on the articles I want to do, and keep up with my watch list. Just what threat can I use to get editors to do something about missing sources? Maybe I'll just start cutting such articles back to stibs until sources are produced. -- Donald Albury 21:25, 23 November 2006 (UTC)
• Oppose. Semi-related external links, that will be added as "references" only to prevent deletion of articles, are not the answer to the high number of unsourced articles. Trial run is not necessary. Prolog 12:27, 23 November 2006 (UTC)
• Reluctant oppose I support the CSD, but I won't support a trial run without wider acceptance in the community. --ais523 13:26, 23 November 2006 (UTC)
• Support trial run provided that only articles created after the run begins may be tagged. Seraphimblade 13:37, 23 November 2006 (UTC)
• Strong support. Rebecca 04:12, 24 November 2006 (UTC)
• Oppose per reasons explained here. John254 06:05, 24 November 2006 (UTC)
• Oppose as long as the notion of "reputable, reliable, third-party sources" is not spelled out in a clear,consistent and uncontroversial way, which btw I regard as impossible. Stammer 09:14, 24 November 2006 (UTC)
  • This problem is answered in the proposal; any source is enough to prevent the CSD from working (AfD or prod has to be used if the sources are deemed unreliable). --ais523 10:16, 24 November 2006 (UTC)
    • Yes, but that does not address the underlying issue. Let me give you a concrete example. If you start checking links related to, say, Somalia and its civil war, you'll find that many of them are poorly sourced or not sourced at all. Still, they may provide valuable information on what's the situation on the field. Even an unsourced, rabidly WP:POV article about some Somali clan may still contain useful information that can be filtered by a trained eye. I hate the idea that such information may be be just thrown out of the window, without discussion. I have repeatedly seen appropriate sources being added within the AfD process and ugly ducks become swans (well, almost ...) . IMO Wikipedia is a collective endeavour, where every participant is invited to contribute. Raising barriers to entry may just destroy value. Stammer 11:57, 24 November 2006 (UTC)

Comment by Sjakkalle

As per my two comments above, I feel that the proposal is too far reaching, and too draconian. Several concerns have been attempted addressed in a "may delete" clause, but that is wide open too interpretation. If 563 admins look at the page and decide that they will not delete even if its unsourced, and one admin thinks it should be deleted, the article will wind up deleted. After all, I frequently see articles tagged for deletion as A7 even when there is some assertion of notability (some people think running as a major party candidate for congress in the US is not even an assertion), I have no confidence that we won't have admins deleting everything unsourced (even if it's easily sourced) if we implement a criterion like this.

Then there are some cases where adding sources just looks silly. THe example which comes to mind are disambiguation and other navigational pages. I strongly doubt that JFK (disambiguation) would be a much better page if each entry there were sourced (if you want a source, click each entry and you will find the sources). So someone has suggested a clause to prevent such pages being targetted, called a clearly unneeded clause. But what ishttp://en.wikipedia.org/w/index.php?title=Wikipedia_talk:Speedy_deletion_criterion_for_unsourced_articles&action=edit&section=74 clearly supposed to mean? What if we have an admin who thinks JFK (disambiguation) clearly needs sources? Is it that obvious that people refer to John F. Kennedy as "JFK"? Maybe someone has just made up all these abbreviations! After all, there was an admin who interpreted the spam criterion G11 to mean that all the neutrally worded articles on cookie brands could be deleted.

Now, I wonder why we cannot simply use the existing processes WP:PROD and WP:AFD. One advantage (well, sometimes it's a disadvantage) of WP:PROD is that once something is deleted, it can be immediately undeleted without discussion. So if something is deleted by WP:PROD as "possibly unverifiable, no sources provided", it is not too late to restore the article with sources. Alright, let's say WP:PROD is not possible, perhaps a bad faith vandal removed the extremely well reasoned prod-tag. Still, we have no problems getting the article deleted if it's truly unverifiable. A concern that the article is unverifiable carries a huge weight on AFD discussions.

My basic question is this. What is this criterion supposed to achieve that the current WP:PROD and WP:AFD cannot do already? (AFD with a 5 or 6 day time scale is after all faster than this two-week "speedy" criterion.) Sjakkalle (Check!) 10:31, 18 November 2006 (UTC)

"Then there are some cases where adding sources just looks silly." Can you give a more reasonable example? No one at all considers that this criterion would be useful for disambiguation pages, and if you think an admin would delete a disambiguation page based on this, they don't deserve to be an administrator. Furthermore, disambiguation pages are explicitly mentioned in the wording. The argument sounds like grasping at straws.

This criterion is supposed to achieve precisely what AFD and PROD don't. That would be sourcing articles. Do you honestly think that either of those deletion processes causes articles to be sourced as a result, and that they don't just keep any unsourced articles indefinitely, with no way to cause sourcing? They have no mechanism for requiring them, since they don't delete for lack of sources, and there's no oversight after a "keep and cleanup" or de-prodding. In fact, not too long ago, I prodded dozens of band articles that had no sources other than official websites/MySpace/YouTube links. The results: "will dig up some more sources shortly" on Oct 20, none ever came, "will hunt down sources shortly." none ever came , no comment, no comment, no sources provided, no sources provided, no sources provided, no sources provided, no sources provided, no sources provided, no sources provided, no comment, no sources provided, no sources provided, no source provided, I could go on. Most of these are still bad sourceless articles: our current mechanisms are insufficient for requiring sourcing. Dmcdevit·t 10:59, 18 November 2006 (UTC)

OK, does Brothers Majere need a source? Sjakkalle (Check!) 11:01, 18 November 2006 (UTC)

I think it should be easy to list out which pages do not need references. Unless I'm missing something, it is disambiguation pages and redirects. I've amended the text accordingly. Do check to see if I've missed anything though, jguk 11:11, 18 November 2006 (UTC)

The example with Brothers Majere is a book. Is the book itself a reference for its own existence? (Primary source, but there is no trouble verifying the book's existence online if you want to). Sjakkalle (Check!) 11:18, 18 November 2006 (UTC)

Yes. Some reviews and commentary about the work would be nice (there should be some). WP:FICT says that just a plot summary is highly discouraged. (I mean, what's to stop somebody from claiming that their self-published work "doesn't require a source" either because this book doesn't have it?) ColourBurst 15:05, 18 November 2006 (UTC)

If the article only covers things in the book, the book is a reference. In this case we have the author, date of publication, ISBN and publisher, so it's easy to find the book if you want to check the details in the article. So yes, I would say that that article is properly referenced. To give another example, I have recently removed a request for sources from North Carolina v. Mann, on the grounds that as the article gives a full case reference, the case itself is a source for the content of the article, jguk 15:08, 18 November 2006 (UTC)

When it comes to those un-prodded music articles, then they would not be covered by the proposed speedy criterion here becuase they have a link, albeit only to their own site. Still a sort of reference though whowing that it was not made up for Wikipedia without anything. (Incidentally, I am an utter idiot when it comes to music, and am not entirely sure how to determine if WP:MUSIC's inclusion criterion 4 is satisfied, but it seems to me that Bigwig (band) has released some CDs on record labels which are widely available.) Sjakkalle (Check!) 11:14, 18 November 2006 (UTC)

Yes, a few of them wouldn't be covered because this criterion is constructed too conservatively to cover them. How is this an argument that the criterion is too "far-reaching" and "draconian"? Things like "it seems to me that Bigwig (band) has released some CDs on record labels which are widely available" are precisely the problem. It doesn't matter what it seems to you, whet we need is an independent, reliable source saying that. The unsuccessful prods show that our current mechanisms can't produce that. Dmcdevit·t 05:09, 19 November 2006 (UTC)

You need an independent source to tell you that something is for sale? I don't understand, isn't a listing on Amazon, for example, proof enough that something is widely available, given that Amazon ships to almost anywhere? In this specific case, just take a look for yourself if something's there, and if not, AfD it. How we record your verification is a different question for a different discussion, since things can go out of print and unavailable without much warning. Unfocused 18:27, 19 November 2006 (UTC)

By the way Dmc, I fully agree with you that anyone who wanted to delete JFK (disambiguation) because it is unsourced should not be an admin, and the example was extreme and therefore perhaps not the best one. However there are a number of list articles which are also unsourced because they are primarily intended to be navigational aids. Yet there are people who might condemn them as "listcruft", but they cannot go ahead and delete them because they are not speedy deletion candidates. Is it possible that someone would use this speedy criterion to delete a list with the official reason "unsourced" while the real reason is "listcruft"? Sjakkalle (Check!) 11:18, 18 November 2006 (UTC)

Even a relatively short list can contain far more information than a short article. It's not unreasonable to require it to be sourced, jguk 11:19, 18 November 2006 (UTC)

An example of an unsourced list is List of religions. If someone wants to make this a category instead, then fine. However, I don't think speedy delete is appropriate. That said, admins are, of course, required to use their common sense. Possibly we could rephrase and say "navigational pages" instead of "disambiguation and redirects". Addhoc 12:43, 18 November 2006 (UTC)

Dmcdevit, this process won't nab those articles, because they would have at least a myspace/youtube/official site external link. Unless you specifically exclude those from being "reliable sources" most of those are going to stay still. ColourBurst 15:03, 18 November 2006 (UTC)

Why not just use AfD and PROD? Well, AfDs are overkill - way too much time by editors - for cruft, and besides, the point of the tag is to try to get someone to improve the article as much as it is to delete it, while an AfD often devolves into arguments. As for PRODs, if there was an automated system that automatically stuck a PROD on every unsourced article, would you support that? And, of course, anyone can remove a PROD without actually improving an article. I see this policy as filling the gap between speedy delete and PROD/AfD, removing minor articles that no one cares enough about (within 14 days) to actually add a source. John Broughton | Talk 16:51, 18 November 2006 (UTC)

Some of those examples I gave would not be covered with this criterion, becuae it is written more conservatively. My point with those examples is to demonstrate that our current processes have no mechanism for requiring sourcing. Saying AfD and PROD are sufficient misses the point. Dmcdevit·t 20:04, 18 November 2006 (UTC)

Problem: you say they are "all unsourced". A newbie that knows nothing about policy will say that all but 5 are sourced (they're all primary sources, or youtube, or myspace). This process will only delete those 5 but not the rest (remember we're not judging the quality of the sources there, but if they exist). So that's around a 25% hit rate, pretty abysmal. You'll still need to prod/AfD the rest unless you place a "reliable sources" clause in there, but that's dangerous. ColourBurst 21:11, 18 November 2006 (UTC)

Since when was 25% worse than 0%, the alternative you suggest? "You'll still need to prod/AfD"—no, this is the problem. Those were PRODs, and they survived, and were never sourced anyway. AfD and PROD can't fix them. Dmcdevit·t 21:30, 18 November 2006 (UTC)

Requiring references really shouldn't be a big deal

I really see only improvements arising from what really is a modest proposal.

It's a quite reasonable request to ask the author of an article - how do you know that? can you give me a reference? And it's not a particular onerous question to answer provided you really are writing about something you know about. After all, if you can't tell me why you believe it to be true, why should you present it to me as true.

Very quickly after implementation, those creators of articles who are not yet providing references will get into the habit of adding them, which will improve the reliability of Wikipedia. When occasionally they make a mistake, the article will be tagged, and they can go back and add the references. And as long as they are informed of the situation, that will happen. jguk 11:18, 18 November 2006 (UTC)

• Requiring references is completely uncontroverial. It's when the extreme position of speedy deletion (or delayed speedy deletion) is brought into the fold that it becomes a very, very big deal. --badlydrawnjeff talk 04:18, 19 November 2006 (UTC)
  • What's a requirement without enforcement? Night Gyr (talk/Oy) 04:36, 19 November 2006 (UTC)
    • Badlydrawnjeff, WP:V already allows the removal of content where a reference has been requested but not provided. Current policy therefore already permits page blanking in this situation. A blank page would then be a candidate for speedy deletion under other rules. jguk 17:00, 19 November 2006 (UTC)

Challenged

Granding that truth is not a defense against nonverification, I do note that WP:V say "Any article that is challenged or likely to be challenged needs a reliable source". The great majority of statements are not likely to be challenged by any reasonable person. For instance, take the statement "A. C. Smith was commander of XV Corps of the British Army of the Rhine from 1957 to 1959." Who would challenge that? To challenge it, you would have to have some reasonable cause to suspect that it isn't true. But who would make something like that up? And the great majority of unverified statements are of this type.

Of course, it's possible that the statement was made up by a sneaky vandal. But very, very few such statements would be so - vandals don't work like that. (For that matter that vandal could just a source that doesn't support the statement and, provided it isn't online, no one would be likely ever be the wiser.)

It's also possible that the person made a mistake - it's a different A. C. Smith, maybe, that commanded XV corps. Or that his source (whatever it was) is itself in error. In this case, referencing isn't going to help much... the reference would point to a valid source that stated that one A. C. Smith was indeed commander of XV Corps, and absent further digging the reader has no way to find out that it's a different Smith or that the source is wrong.

I do enforce very strict verification standards on the articles I mainly work on, because they are contentious and open to POV and OR statements, but for stuff like the example here, which is most of our stuff - I don't see why the spirit of "...that is challenged or likely to be challenged..." shouldn't be the standard for demanding verification. Herostratus 13:53, 18 November 2006 (UTC)

This proposal is not really about that. It only asks for one reference for each new article. Though a reference for your claim really would be beneficial. Maybe it was CA Smith, or the XIV Corps or from 1947 to 1949. I'd therefore recommend giving a source for everything so anyone so minded can check it. But as noted before, this proposal does not go anywhere near as far as that - just asking for one reference for each article (which will at least provide evidence that the subject matter exists), jguk 14:59, 18 November 2006 (UTC)

• • Well, some vandals do operate that way. [3] This is all the more likely when POV-pushing is possible. Of course, this proposal does not directly address that problem, and indirectly only to the extent that it may encourage more fact checking. Robert A.West (Talk) 20:35, 26 November 2006 (UTC)

Fish where the fish are

This strikes me as an ill-conceived proposal. Plenty of basically decent articles, especially short articles, appear in Wikipedia without any reference apparatus. And plenty of articles that provide quite a few references nonetheless include masses of crap. The issue ought to be to identify statements that need to be removed because they are unreferenced and someone actually has doubts about them, not to remove articles that lack proper references but appear to be correct. - Jmabel | Talk 01:23, 19 November 2006 (UTC)

WP:V already allows for the challenging and removal of dubious, uncited statements by any editor at any time. This is intending to encourage editors who create new articles to remember to include references. Night Gyr (talk/Oy) 04:15, 19 November 2006 (UTC)

To my mind "add a source or it gets deleted" isn't encouragement, it's wikipolicing. That's like saying "we encourage you to pay taxes"! It's another rule in an already very litigious environment. Any half-clever vandal will provide a source that means his load of rubbish stays while a well-meaning novice creates a valid article and it gets deleted. Why do so many wikipedians concentrate on deletion rather than improvement? Mallanox 16:26, 19 November 2006 (UTC)

Well said. There seem to be far too many people coming to Wikipedia now who never write anything of their own, but only go round looking for other people's work to delete, usually sneering at it in the process (just look at the attitude taken on many AfDs and PROD notices if you don't believe me). This is a tendency that needs to be nipped in the bud now before these people damage the project irrevocably. Despite their arrogant claims that only they are the true guardians of Wikipedia and their sneering at we lowly beings who actually try to create articles, they are in fact the true enemies of this encyclopaedia - far more dangerous than any vandal and certainly much more dangerous than the unfortunate editor (sorry, heretic) who doesn't reference every single piece of information in his article. -- Necrothesp 23:54, 19 November 2006 (UTC)

Ok, I support this proposal and since posting on this page I've started two articles for the WP:MISSING project. Could I enquire how many articles have you have started recently? Addhoc 00:21, 20 November 2006 (UTC)

I think if you look on my userpage you'll find the answer to how many articles I've started! -- Necrothesp 16:07, 20 November 2006 (UTC)

Night Gyr makes a good point here. Current policy would allow users to go to every unreferenced article and quite reasonably challenge all of its content on the grounds that as there are no references, they can't be sure that the subject matter exists. If the creator then did not supply a reference within a reasonable time, the content would be removed, at which point the article would be deleted as it contained no content. All this proposal is doing is regularising this. jguk 16:58, 19 November 2006 (UTC)

Instead of tagging the article, why does the challenger not look for sources and AfD if none can be found? Mallanox 17:05, 19 November 2006 (UTC)

How long does it take to find a source that backs up the text, and then assess the source to see whether it is reliable (which can be difficult even if you are familiar with the subject in hand)? Then, possibly, add the time it takes to do an AfD, which is likely to fail as soon as the creator gives one reference. Then multiply that by however many new articles are created that currently do not have references. Compare that with the time it would take each article creator just to make a note of why he believes what he has written is true when he is creating the article. jguk 17:17, 19 November 2006 (UTC)

Personally, I agree that looking for references should be the first stage. The second stage is to give everyone else the opportunity to find references. If after a reasonable period, none can be found, the article should be speedily deleted. This seems common sense to me. Addhoc 17:26, 19 November 2006 (UTC)

New users aren't going to know to do this. It may be common sense to you (common sense is not always objective, and invoking its name doesn't guarantee instantly everyone will agree) but it's not going to be comon sense to a new user. Do we want to make it too much of a chore for new people to be able to contribute in a constructive way? That's the way things are heading. We need to look at simplifying existing rules before we start adding more, convoluted, ones. Mallanox 21:25, 19 November 2006 (UTC)

So autoload new article pages with a ==References== header (and probably a {{uncategorized}} tag, but that's another argument). make it easier for people to add sources, rather than just making it easier to delete overlooked articles. We already have all the policies and procedures in place to deal with truly bad actors. -- nae'blis 19:19, 20 November 2006 (UTC)

Jmabel has it exactly right. I previously estimated the percentage of unsourced statements and the percentage of correct statements on Wikipedia. By coincidence, I got 99.6% for each (not the same 99.6%, of course). I would guess that a large percent of material that would be deleted under this proposed policy would be factually accurate, uncontroversial material. The real cancer is the inaccurate material, especially illegitimately-sourced inaccurate material, which this proposal doesn't address. dryguy 23:10, 20 November 2006 (UTC)

You really believe that 99.6% of all statements on Wikipedia are inaccurate? If not, what did you mean to say?

In any event, the discussion cited is not really relevant. Here, we are not discussing {{fact}} tags, but completely unreferenced articles, most of which are not tagged in any way. (I also dispute the methodology used to come up the the figure, but that is a different matter.) The complaint that a change to deal with problem X does not also solve problem Y is a form of false dilemma. All variants of this proposal (and DfV or UfD is actually less severe than current policy as written) are designed to deal with a problem that is relatively easy to identify and solve. If, as you suspect, most unsourced articles are sourcable, the result will just be to put urgency on sourcing them, which will improve Wikipedia.

Finding inaccuracies and then getting the corrections sourced is a much harder and slower process, and requires knowledgable editors. Policy pretty much can't do anything about it, unless we start taking people to Arbcom for putting in fake sources and grossly misusing sources. Robert A.West (Talk) 13:27, 21 November 2006 (UTC)

Whoops. 99.6% accurate. I don't claim that this proposal fails to address a problem that it isn't intended to address. It is meant to improve sourcing of articles, yes? I just don't see the benefit of achieving that by tossing a large percentage of otherwise good material. It would do more harm than good. And you can quibble with my estimation, but whether it is 99.6%, 90% or even 80%, you are talking about a proposal that would throw the power to delete a large percentage of Wikipedia into the hands of anyone who has an axe to grind about nearly any article. dryguy 13:31, 21 November 2006 (UTC)

Effectiveness

While I appreciate the motivation for this proposal, the more I think about it, I'm most concerned about its effectiveness. If any external reference will suffice to thwart off this CSD, then what is it really buying? People will just start adding their own personal home pages or the first entry in the results of a Google search. How is that helping? If this proposal were to require secondary sources to establish some base notability, that would have some teeth behind it. I'd be all for it. The way it's proposed now, I wouldn't be surprised if it resulted in neither the deletion of any article nor the overall improvement of Wikipedia. —Wknight94 (talk) 19:08, 19 November 2006 (UTC)

You can't create articles anonymously. An editor creating lots of articles with clearly inappropriate references will soon be sussed out, and appropriate measures taken, jguk 21:43, 19 November 2006 (UTC)

Inappropriate how? What is this going to catch that other CSD won't? Frankly, if an article looks like borderline notability and is unreferenced, I'll just bring it to AFD - it's faster. —Wknight94 (talk) 21:47, 19 November 2006 (UTC)

• And how will he be sussed out? Quite possibly only the nominators and the deleting admins will have seen the nominations. On AfD there's more eyes to notice that User:wasteoftime keeps showing up to defend cruft. Septentrionalis 23:07, 19 November 2006 (UTC)

Let's be direct

Instead of this, why don't we add Do show where you got your facts or even Articles without sources may be deleted to the edit screen. (As NightGyr keeps pointing out, the second is possible under present policy.) If we're trying to scare people straight, let's make sure that it's a message new editors will actually see.

I don't say anything for either wording above, except that it's short. Septentrionalis 23:23, 19 November 2006 (UTC)

I'd suggest taking this to MediaWiki talk:Copyrightwarning, where this change will have to be discussed anyway before being implemented. (I suggested something similar on WT:CSD.) --ais523 11:26, 20 November 2006 (UTC)

• • Either this has just been done, or it has been there all along. "Articles that do not cite reliable published sources will be deleted." [4]. Robert A.West (Talk) 22:00, 26 November 2006 (UTC)

A horrible example

As an example of why this proposal should not be tightened to require any more than the existence of a source, I present Wikipedia:Articles for deletion/Abba Samuel of Dabra Wagag, where an a editor nominated an article on the grounds This is almost utter nonsense. Also, the only source is in Latin, so no one can really verify it. Finally, this could be a possible violation of WP:HOAX.

The article is about an Ethiopian saint, and summarizes his Life. (The "nonsense" is mostly bad English by the author of the article; also a couple miracles.) It is sourced from the translation of that hagiography (into French, not Latin). It was contributed by a very new contributor, who hasn't been back in three weeks. (It also needs a lot of work.)

Now suppose that the nom had used this proposal, extended to judge sources. It is clear that he considers the source unreliable; so he would tag it. I doubt anyone (except perhaps the creator) was watching the article, so the tag would probably stick, and we would lose this article, which might not be replaced for years. (Who knows Ge'ez saints?) Fortunately, he brought it to AfD instead. Septentrionalis 23:07, 19 November 2006 (UTC)

What does that "example" have to do with the current proposal. The article had a source; there is already no substantial support for including articles with dubious sources in this proposal. Dmcdevit·t 23:27, 19 November 2006 (UTC)

I still see substantial support for this among the support !votes; but if the horse is dead, so much the better. Septentrionalis 23:32, 19 November 2006 (UTC)

I believe that the support for evaluating sources comes almost entirely from the discussion about an XfD process, where there are more eyes and a chance to gain consensus on the validity of a source. Of course, one could remove fake, irrelevant and unreliable sources and then nominate under CSD, but unless the case for removal was clear and strong, the admin should revert to the "sourced" version. Robert A.West (Talk) 13:37, 21 November 2006 (UTC)

Comment by JulesH

I strongly oppose this proposal in its current form. My reasons are:

• A very large majority of articles are unsourced, and generally remain unsourced for long periods of time.
• An unsourced article is useful -- it is a basis off which an editor can work to produce a usefully sourced one.
• An article deleted is nearly useless -- it will only ever be useful if the editor who eventually fixes the issues with it actually knows about it, and can persuade an admin to undelete it. This won't happen very often, largely because it's highly unlikely that the user who would fix the problem would know about the existence of the deleted article.
• Most completely unsourced articles are written by newbies. This makes the entire proposal sound somewhat like WP:BITE to me. Certainly, I think a requirement to provide sources with all new content would deter new editors.
• Finding suitable sources for an article can sometimes be a long process that requires the interaction of multiple editors over a long period of time. See, for instance, the edit history of Disemvoweling, an article that took three editors working together for several weeks to find a source that overcame the objections of one of the other editors.

But:

• Unsourced information is a big problem. It decreases public opinion of the project, which is an important issue.
• Unsourced information on BLPs is a serious problem; it opens up the foundation to potential legal issues.

Therefore, I think something should be done. My suggestion:

• Apply something like this proposal to BLPs. I think, in this case, waiting 14 days even is not appropriate; the material should be removed as soon as possible. Perhaps the 5 days currently allowed for prod would be appropriate.
• For other articles, there should be a large, friendly banner, that explains clearly that the article has no sources and could therefore be incorrect, and advises readers that they shouldn't trust its content, and
• There should also be another banner which states that information has been removed from the article because it lacks sources. Then, WP:V (or WP:ATT) could be rephrased to suggest to readers encountering information that has remained unsourced for a long period of time that it should be removed, and such a banner placed on the article's talk page. It doesn't even matter if the article ends up empty (although in that case the banner should probably be on the article page, not the talk page, so that readers don't get totally confused...)
• I also like the proposal above for a dated "unsourced article" system similar to the current cleanup system. JulesH 23:34, 19 November 2006 (UTC)
  • And a modified edit screen. See #let's be direct above. Septentrionalis 23:41, 19 November 2006 (UTC)