Wikipedia talk:Citing sources

Dead links and BLPs

The Preventing and repairing dead links subsection says "Do not delete a citation merely because the URL is not working today." What about dead links in Biographies of Living Persons?

For example, if you find a dead link on a particularly contentious assertion in a BLP, should you remove the superscripted citation itself immediately (after checking the archives; see WP:DEADREF)? Or should you wait until the 24 months have expired before eliminating the reference? Remember, this is in a BLP. (Should archived citations be valid for BLP’s at all?)

Much more important, if the dead link is the only citation for a particular assertion, should the statement that the citation references be removed as well? Even if the statement is only mildly debatable? (I’m assuming that you’ve tried to find an equivalent reference and failed.) For example, “Hitler suffered from...tinnitus.^[123]” If citation 123 is lost and we can’t find a substitute, should we entirely remove the part about tinnitus (ringing in the ear) from the Hitler article?

Remember, it's possible that the reason why the link was removed is that it was simply wrong and/or may be under threat of legal action. On the other hand, it may be due to normal link rot and the assertion is perfectly OK. Since the link is dead, we can’t say which is correct.

But, whatever we decide, I think that we should have something about dead links in BLP and something about BLP under WP:DEADREF. --RoyGoldsmith (talk) 13:42, 14 January 2012 (UTC)

I would specifically suggest that where there is any source suggesting that the material has been removed as a result of court action or editorial correction or retraction, that we not use such material (and yes - this sort of thing has occurred in the past where an editor has argued that if it is archived anywhere that it is still a valid cite). Collect (talk) 14:01, 14 January 2012 (UTC)

If the material is supported by a single cite and the material violates BLP if untrue, I would be inclined to remove the material. If no one can come up with a reliable source to replace the dead link, then we have no way of verifying it. If someone can state that they have a paper source to replace the online source, that would be fine, but, otherwise, I'm not willing to accept the proposition that it was legitimate when added. My view is not necessarily the majority view on this issue.--Bbb23 (talk) 19:18, 14 January 2012 (UTC)

It's also a matter of WP:WEIGHT. If only one source exists – but it seems reliable – then it would presumably be authoritative. But if material is supported by only one source, and contradicted by a dozen other sources, then it is, at best, a minority view. ~ J. Johnson (JJ) (talk) 20:46, 14 January 2012 (UTC)

Just to be clear, my comments were directed at one situation - where you have only one source related to the material and that source is dead.--Bbb23 (talk) 20:58, 14 January 2012 (UTC)

Absent any indication of problems, I see no reason a link that has been archived should be treated any differently than one which is offline or behind a paywall. Like most things with BLPs, once there is an indication of a problem, things change, and are probably best dealt with on a case-by-case basis. I've seen plenty of RS print corrections inserted into articles after publication and plenty of RS articles correcting inaccuracies in prior articles, but I know of none which have taken down stories merely to correct errors. As such, I think it unhelpful to presume that a no-longer-accessible article should be assumed to be no longer correct. Jclemens (talk) 17:56, 15 January 2012 (UTC)

I was more worried about the 24-month rule from WP:DEADREF: "Most archives currently operate with a delay of ~18 months before a link is made public. As a result, editors should wait ~24 months after the link is first tagged as dead before declaring that no web archive exists." This says that a link must be dead for 2 years before the material it supports should be regarded as unverifiable. Does this apply to BLP's, the policy for which states "Contentious material about living persons that is unsourced or poorly sourced...should be removed immediately and without waiting for discussion."?

There seems to be a conflict between the policy for BLP's and this Citing sources guideline (see WP:POLICY). Since policies normally take precedence over guidelines, I would suggest something like following sentence be added to the Preventing and repairing dead links subsection on the project page:

Contentious material about living persons for which a dead link is the only source may be removed immediately without waiting to see if the reference becomes archived.

--RoyGoldsmith (talk) 13:51, 17 January 2012 (UTC)

The first question in my mind is whether the source is web-only. I'm sure we can all agree that it would be pretty silly to remove a citation to a story that ran on the front page of The London Times newspaper merely because they re-organized their website.

Even for web-only sources, your sentence won't work. Too many editors will interpret "a dead link is the only source" as meaning "a dead link is the only source that has already been typed into the article", rather than " a dead link is the only source that has ever been published in the real world", which is our actual standard. WhatamIdoing (talk) 22:38, 17 January 2012 (UTC)

But in BLPs (for contentious material), the standard is "a dead link is the only source that has already been typed into the article". The problem is that, if we don't insist on reliable, accessible sources in BLPs, then the possibility of harm to living subjects might not be considered.

For example, the article on Bashar al-Assad states (on 1/23/12 in the first paragraph of Presidency): "On 27 May 2007, Bashar was approved as president for another seven-year term, with the official result of 97.6% of the votes in a referendum without another candidate.^{[citation needed]}" The CN tag has existed since October 2011.

Let's say that the source of that statement was a front-page article in The London Times. However, due to a mix-up, it isn't accessible over the web and the editor who introduced the statement can't be contacted. Therefore, unless you have the issue of the newspaper itself, we have no knowledge of where it came from. It could be the London Times, it could be a blog associated with the Syrian opposition, it could be made up out of whole cloth. Let's further say that, although there are many sources that report Bashar's winning on 27 May 2007, we can't find any reference about the 97.6 percent of the vote. This is contentious because it implies a fixed election.

The rules of BLP say that we must enter the new citation and remove the part about the 97.6% (until we can find a reliable source): "The burden of evidence for any edit on Wikipedia [to a BLP] rests with the person who adds or restores material." Under this rule, we really are not obligated to do any research: anyone can remove the whole sentence in its entirety.

It seems to me that the same applies to a dead link (instead of a CN tag) for this statement. If you can cite the issue, date and page number of the particular Times article, you don't need a URL; anyone who doubts you can go to a library. But if you can't, the 97.6% should disappear.

What do you say? --RoyGoldsmith (talk) 02:10, 24 January 2012 (UTC)

Distinguishing between a dead URL that begins with www.thetimes.co.uk/ and a dead URL to someone's blog is not usually a problem for Wikipedia editors.

A statement supported by a citation whose URL is broken or dead is not an unsupported statement, and is not treated like an unsupported statement.

BLP does not require editors to use sources that are available online. If you can't be bothered to find a copy of that issue of The Times on paper (and I acknowledge that finding offline sources can be an enormous bother), then that's your own problem, not Wikipedia's. See WP:SOURCEACCESS for the official policy, and note that there is no exception for BLP-related issues. WhatamIdoing (talk) 05:26, 7 February 2012 (UTC)

British vs American dates

Wikipedia editors from Europe put the day before the month, but editors from America put the month before the date in the citations. Is there a WP rule that makes it less confusing for us ? Pass a Method talk 10:16, 23 January 2012 (UTC)

This has been debated many times. The consensus compromise is that, in articles on a subject of worldwide scope, either method will do, as long as it is consistent within any one article. The British method is preferred in articles about British or Commonwealth subjects. -- Alarics (talk) 10:21, 23 January 2012 (UTC)

In addition, if a particular article follows a printed style manual for citations, such as APA style, the date format specified in the manual for citations should be followed. Jc3s5h (talk) 15:12, 23 January 2012 (UTC)

What the fuck? Thats pretty pointless to me. I'm still clueless whether 3.11.2011 indicates March or November. Thats a major discrepency. Pass a Method talk 16:21, 23 January 2012 (UTC)

Are there no other non-ambiguous dates anywhere in the article? (E.g. 23/5/2005 is clearly the 23rd of May) What variety of English is the article written in? If it is really not possible to deduce the format just ask on the Talk age. Roger (talk) 16:41, 23 January 2012 (UTC)

The only acceptable all-numeric date format in Wikipedia is YYYY-MM-DD, for example, 2012-01-23. That is only acceptable where space is limited. It is allowed in citations if used consistently; it's been decided that is an area where space is limited (although I disagree). In the body of the article the month should be spelled out. Pass a Method's example of 3.11.2011 is always unacceptable and should be fixed, if one can figure out what it means. Jc3s5h (talk) 16:48, 23 January 2012 (UTC)

One extremely helpful aspect of using the YYYY-MM-DD format in citations is that this makes it much easier to check an article's prose for date formatting inconsistencies (because searches for names of months don't result in dozens of hits within citations). —David Levy 16:58, 23 January 2012 (UTC)

This supposes one would want to check only the prose for correct format and not the citations. Jc3s5h (talk) 19:05, 23 January 2012 (UTC)

No, I'm referring to consistency (i.e. the use of a single date format of the two sanctioned within prose). In other words, apart from direct quotations, the prose should contain either "January 23" or "23 January" (not both).

Of course, the citations should have consistent date formatting too. If they don't, those containing month names will show up as hits. Assuming that no unsanctioned formats are present, whichever ones aren't found via searches for month names inherently use the "2012-01-23" format. It's much easier if most fall into the latter category.

If the one or more unsanctioned date formats (e.g. "01-23-2012", "23-01-2012", "January 23rd", "23rd of January") are present, that's a separate issue requiring attention. (They either will or won't appear in searches for month names, irrespective of what style is preferred for prose/citations.) —David Levy 20:03, 23 January 2012 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘Just to be clear, in my initial response to Pass a Method I was assuming the question was about dates in which the month is written out as a word, i.e. the difference between 23 January 2012 (British) and January 23, 2012 (American). From his/her supplementary remark above, I now infer that he/she was actually talking about all-numeric dates. These are a different can of worms altogether, and the settled consensus on WP is that, as Jc3s5h notes, only the YYYY-MM-DD format may be used, e.g. 2012-01-23 (because any other format is ambiguous) and only then if necessary for space reasons. In response to David Levy, there is a script that anyone can use for making the date format consistent in any one article. -- Alarics (talk) 19:49, 23 January 2012 (UTC)

I've never been fond of scripts (though I'd be willing to give it a try). —David Levy 20:03, 23 January 2012 (UTC)

I have been using User:Ohconfucius/script/MOSNUM dates for ages. It scrubs the entire article and converts the dates to on format. See also User:Gadget850/FAQ/YYYY-MM-DD dates. ---— Gadget850 (Ed) ^talk 21:41, 23 January 2012 (UTC)

Most of my engineering career was spent in large companies with subsidiaries all over the world. Drawings came in all different languages and formats; some dating back to the late 1800's. All that to say, our general rule of thumb was dates that were NOT obvious (eg, Nov 03, 1912 or 11 Mar 1912) should be interpreted generally like this (in the absence of conflicting information):

- dates with dashes or slashes were assumed to be mm-dd-yyyy (eg, 11-03-12 would be 03 Nov 1912).

- dates with dots were assumed to be dd.mm.yyyy (eg, 11.03.12 would be 11 Mar 1912).

The corporate standard was dd mmm yyyy so newer drawings were unambiguous. JimScott (talk) 06:04, 6 February 2012 (UTC)

Here in the UK it's common practice to write all-numeric dates with slashes, although dashes are sometimes used, dots are rare. So, given that (when lazy) I'd write a date as 06/02/12, or 06-02-12, what form is that? --Redrose64 (talk) 14:40, 6 February 2012 (UTC)

FWIW I don't agree that the use of dots is rare in UK. I have been doing it for over 50 years, myself. It seems to be quite common in publishing at least. And I have never before heard of the idea that whether it is dashes/slashes or dots might help to work out what is what in an otherwise ambiguous date. But really all that is beside the point, which is that all forms of all-numeric dates are potentially ambiguous in an international setting and they should therefore not be used in Wikipedia. -- Alarics (talk) 21:35, 6 February 2012 (UTC)

Wikilinking in cites. What is too much?

Should everything linkable in a citation be linked to an article? The work? The publisher? The location city? The names of the authors and editors?

If the above answer is yes for even one parameter, should such wikilinks in citations be repeated in each similar citation in an article, or just in the first one? For instance, do we wikilink to The New York Times once, or each time it shows up as a citation? Binksternet (talk) 22:20, 25 January 2012 (UTC)

Certainly not the location city, and in my own view not the names of mainstream newspapers if the location is given (or is included in the title, as in The New York Times). If there have to be links, it should only be the first time it appears in the article. Authorlink is OK. I am not sure whether there is any point in wikilinking the names of publishers of books. I think many editors wikilink too much where no purpose is served. Getting the citation right and including all the details that are required seems to me much more important. I have seen people wikilinking the name of an extremely well-known newspaper but not bothering to include the publication date, which is absurd. -- Alarics (talk) 22:41, 25 January 2012 (UTC)

Use your judgement. Alarics is spot on here. I would link obscure sources (e.g. Field Artillery) or sources that might be confused with others on the first use. ---— Gadget850 (Ed) ^talk 23:06, 25 January 2012 (UTC)

I would actually link The New York Times to show that this newspaper is really known, not some obscure newspaper that mimics the name of a known one. Also I wouldn't assume the fact that any newspaper is well-known, given that editors are coming from various places. I've actually never seen the The New York Times in the real life, and I'm not quite sure that I would spot difference if the actual source was eg. The New York Time newspaper. — Dmitrij D. Czarkoff (talk) 06:09, 26 January 2012 (UTC)

I think we just had this discussion (or part of it) on WT:LINKING. But, consider that links should be made for relevant information, I would never link a publication's city or conference city as part of a citation, even if the city's not well known, because the link isn't helping the reader to understand the citation, it's just a convention used to assure we've got the right publisher or city for a reference. Linking other fields like author, publisher, or work are germaine to the reference information (it helps you know what the work or person is as an authority), and thus while up the editors of the page, these are recommended. If they are used, they need to be used on all references when possible to be consistent within the citation list, as to make it easy to the reader to find out more on the author/publisher, and to avoid having to relink when reference use is moved around on a page. --MASEM (t) 23:54, 25 January 2012 (UTC)

I was the other side of a short discussion on the matter. And my rationale was exactly this one. I would also note, that I believe one should wikilink author, work, publisher and book (provided that targets exist) for each reference, as the references are normally accessed from <ref></ref> by jumping from the article and back. The exception I would also like to note is that authors, works and publishers that are wikilinked in article's body shouldn't be linked in references. All of this is my opinion, and I'm not aware of any policy or guideline detailing this issue. — Dmitrij D. Czarkoff (talk) 06:09, 26 January 2012 (UTC)

The citation block is generally considered it's "own" part of the article; while we avoid relinking in prose of the article body, the citation block should be linked appropriately regardless if the names are mentioned previously. Again, for the reason that readers may jump from a reference link in the article body to the cite block, and if they haven't seen the linked term yet in the prose by that point, they may not know what it is within the citation. It also just makes it a lot easier to manage citations and links within them and to ignore any problems with overlinking in the prose itself. --MASEM (t) 15:02, 26 January 2012 (UTC)

There is a particular danger about encouraging the wikilinking of names of newspapers: many newspapers have the same name as other newspapers, and only the "location" (city of publication) uniquely identifies it. For instance, there is a Daily Telegraph in Sydney, which is a major paper in Australian terms, as well as the one in London. There is a The Guardian in Dar es Salaam as well as the one in London. There are dozens of newspapers around the world called The Times. And so on. Editors who are insufficiently aware of this might just bung double square brackets round the name and, seeing that it produces a blue link, assume that they have linked to the article about the correct paper when in fact they have got the wrong one. This is not just a theoretical problem: I have seen it done. Rather than encouraging linking, I think it is better to just insist on the inclusion of the "location" (city of publication) except, of course, where that is already part of the name of the publication. -- Alarics (talk) 12:37, 29 January 2012 (UTC)

I would say that it is a good reason to encourage wikilinking: when done right, it provides the needed information in a more compact and "wikified" way; when done wrong, it may be fixed.In fact, I would encourage editors to create red links for newspapers, journals and alike (eg., The Times (Kiev)), which would help both establishing the reliability of source and the need for article (via "What links here"). — Dmitrij D. Czarkoff (talk) 14:34, 12 February 2012 (UTC)

{{Cite News}} Question

I have read the project page but am still a bit confused as to the proper way to cite different content contained within a single, lengthy news article.

1. When citing different content from a single article, should an individual citation be built incorporating a "relevant quote" as appears to be suggested by the template guidance?
2. Where online page numbers are available, should those page numbers also be contained within the citation or is that meant to refer to "print" media only?

Thanks JakeInJoisey (talk) 13:31, 6 February 2012 (UTC)

If there are page numbers, I would go with those. It avoids the risk of the relevant quote appearing at two spots in the article, and the possibility that the reader might print the online source before reading it (in which case searching for quotes would be hard). Or, you could do both. Jc3s5h (talk) 15:40, 6 February 2012 (UTC)

If I might pursue this a bit further, I think you are saying that, where there are multiple page numbers presented in an online newspaper article (which may not necessarily reflect the page numbering of a "print" edition), new citations reflecting a different online page number should be constructed for each. Is that correct? JakeInJoisey (talk) 15:54, 6 February 2012 (UTC)

Ordinary websites that are simply HTML documents do not have page numbers. In my experience, an online news article will not usually have a page number unless it is a PDF document representing an exact copy of a (real or notional) printed newspaper or magazine. So the question does not arise very often, but when it does, including the page number is helpful. BUT one or two US newspapers' website articles sometimes say things like "this article appeared on page x of the Blah Times on such and such a date" and in that rather special case I think one might also cite the page number. -- Alarics (talk) 21:48, 6 February 2012 (UTC)

OK. Now let me see if I can correctly summarize your position...

1. Where "online page numbers" are provided (eg. see [1]), a citation should cite a specific page.
2. Where specific information on "print" pages are provided, a citation should cite the printed page information.

Have I got that right? Assuming I do, in the case of #1, should the "relevant text" from the source page be incorporated in the "quote" parameter? JakeInJoisey (talk) 03:21, 7 February 2012 (UTC)

Including a quotation is entirely optional; one is typically included only when the source is not easily accessible or the fact being cited is controversial. --Cybercobra (talk) 05:08, 7 February 2012 (UTC)

We have some confusion here. That New York Times example that you give is not at all what I meant by online page numbers. That is simply an article that has been spread over several screens. Rather than citing those numbers I would give the URL for the version of the article that is all on one web page (click on "single-page" in the right-hand column), since the NYT happens to provide that option. Where that option doesn't exist, I don't think we should get into the business of mentioning the different "page" numbers, which will bear no resemblance to the real page number of the article in the actual physical paper. No, what I meant by online page numbers is for instance this [2], where the website provides a PDF of the original newspaper page. A more recent example is this kind of thing [3], where the website provides a PDF of the physical magazine. As I said before, this will not apply in most cases. (I am leaving out of account here services such as [4] because the items are kept there for only 3 months so would not be worth citing on WP.) -- Alarics (talk) 09:04, 7 February 2012 (UTC)

I am fully prepared to embrace any determination as to the appropriate use of the "quote" parameter, but the expression of differing "opinions" here as to it's appropriate use demonstrates, I think, the "shades of grey" problem I'm attempting to address. I believe the project would benefit from some greater clarification as to this parameter, but If it is the consensus of the community to allow significant editor latitude in the use of "quote", then I'll simply retire and be on about my citation building business interpreting "quote" as I deem appropriate. JakeInJoisey (talk) 15:53, 11 February 2012 (UTC)

Citation templates in editing non-functioning

I think the title says it all. Who's messing with the coding? - Jack Sebastian (talk) 03:00, 7 February 2012 (UTC)

Don't see any issues. Details? ---— Gadget850 (Ed) ^talk 10:17, 7 February 2012 (UTC)

Weird; now it's working. I was looking to add a citation to Kershaw Knives, and I couldn't get the drop-down box for the different citation templates. Was something done in the overnight? - Jack Sebastian (talk) 16:24, 7 February 2012 (UTC)

There have been no recent changes to the RefToolbar scripts. ---— Gadget850 (Ed) ^talk 16:29, 7 February 2012 (UTC)

What should I be looking for, troubleshooting-wise, if it happens again? - Jack Sebastian (talk) 16:57, 7 February 2012 (UTC)

Consistent style

I propose to remove "(Note that templates should not be added without consensus to an article that already uses a consistent referencing style.)" Adding templates does not necessarily alter the style used and altering style is covered in the section Citation style and applies just as much to non template citations that change the style as to those using templates.

For example if an article uses parenthetical referencing the introducing the template {{harv}} does not change the style. But introducing <ref>A footnoted reference</ref> does change the style. -- PBS (talk) 08:19, 12 February 2012 (UTC)

No; this is a style change. Fifelfoo (talk) 08:29, 12 February 2012 (UTC)

I am sorry what are you saying is a style change? -- PBS (talk) 09:11, 12 February 2012 (UTC)

I'd agree with Filfelfoo. There are also multiple parenthetical referencing styles, and several different templates out there, some of which change over time. Hchc2009 (talk) 09:15, 12 February 2012 (UTC)

I think he is discussing a dichotomy of style: presentation and markup, which has been touched on before. By presentation, I mean the visual appearance of the rendered page; by markup, the markup used to render the page. These two examples use different markup, but have the same presentation:

Markup	Renders as
<ref>[[#CITEREFElk1972|Elk 1972]], p.5.</ref> <references />

{{sfn|Elk|1972|p=5}} {{reflist}}

---— Gadget850 (Ed) ^talk 12:30, 12 February 2012 (UTC)

The goal of WP:CITEVAR is to discourage several things:

• Changing the appearance of references in the rendered page (e.g. changing from footnotes to parenthetical citations, or changing from MLA to APA style)
• Changing the way that the references are entered in the source code (e.g. changing from no-templates to templates, or vice-versa, or changing from one series of citation templates to another). This applies even if the rendered page looks identical.

— Carl (CBM · talk) 12:44, 12 February 2012 (UTC)

I can understand the first, but what is the reasoning behind the second? ---— Gadget850 (Ed) ^talk 12:54, 12 February 2012 (UTC)

Some editors HATE templates, and they're right. See this failing example. One of the many places this dead horse has been beaten is Wikipedia:Centralized_discussion/Citation discussion. I wish I could find the place where one of the long-time Wikipedia developers lists the introduction of citation templates among his biggest mistakes. Jc3s5h (talk) 13:09, 12 February 2012 (UTC)

Ah... hate. So, performance doesn't enter into it. Understood.---— Gadget850 (Ed) ^talk 13:17, 12 February 2012 (UTC)

Poor performance and brokenness are valid reasons to hate computer hardware or software; performance does enter into it. Jc3s5h (talk) 13:43, 12 February 2012 (UTC)

OK, for both those reasons I hate anything produced by Microsoft since 1995. It won't stop people from switching to the latest version of Windows as soon as it appears though. --Redrose64 (talk) 14:32, 12 February 2012 (UTC)