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)

Spaces within citations

Spaces in various places within citations are not rendered, but make the source code easier to read. Where there are multiple citations, a space before the closing </ref> tag makes it easier for the human eye to see the end of one citation and the start of the next one. However, WP:GENFIXES currently removes it. Worse, some bots also remove spaces adjacent to pipes within citation templates such as {{cite web}}. Is this the place to discuss a change in GENFIXES (or other rules) with the goal of retaining such spaces? (An editor at WT:Manual of Style suggested that I bring this here.) – Fayenatic (talk) 22:12, 17 January 2012 (UTC)

Can you give me a diff of this happening in an actual article? The things I've found in the list (e.g., WP:GENFIXES#FixReferenceTags_.28FixReferenceTags.29) don't seem to do what you describe. For example, you complain about this:

<ref>Here is my text. </ref>

being changed to

<ref>Here is my text.</ref>

but what I see listed is ^{[1] [2]} being changed to ^[1][2]. WhatamIdoing (talk) 22:44, 17 January 2012 (UTC)

If you want to see a space between and , then see Help:Reference display customization. ---— Gadget850 (Ed) ^talk 23:03, 17 January 2012 (UTC)

No thanks, Gadget850, that's not what I want. Yes, WhatamIdoing: here is an example of a bot removing spaces before the closing /ref tag. The bot's owner says the space is removed by general fixes. – Fayenatic (talk) 09:47, 19 January 2012 (UTC)

It is marginally useful to keep those spaces. Much more useful in my view is to keep, or indeed insert (as I often do if I am editing in that paragraph anyway), spaces adjacent to pipes within reference templates. They not only make the contents of the edit window easier to read, they also allow line breaks to appear in sensible places in the edit window. -- Alarics (talk) 10:26, 19 January 2012 (UTC)

I did a bit of testing and don't see that adding a space before </ref> helps in any way. A space in citation templates before parameter pipes does help to wrap text in the edit window, especially for URLs. ---— Gadget850 (Ed) ^talk 11:55, 19 January 2012 (UTC)

Extra whitesspace (spaces, but also linebreaks) can very definitely make raw text more readable. If you don't find it so, fine, no one is required to put them in. But taking them out because someone wants to a save a very little space is more prissy than useful. I put them in for a reason, and it is rude (at the least) for bot-drivers to wipe them out without even a by-your-leave. ~ J. Johnson (JJ) (talk) 20:18, 19 January 2012 (UTC)

Thanks for the diff. I don't normally leave trailing spaces, but I wouldn't bother removing them. I agree with Fayenatic's statement at the bot page that twiddling the spaces makes the diffs needlessly complex, and it provides no apparent benefit (since, as Gadget points out, the space isn't rendered). Adding it might not help anything except the editor, but removing it definitely doesn't help anyone in any way. I think this "fix" should be removed from the list of automatic changes. WhatamIdoing (talk) 01:37, 20 January 2012 (UTC)

I was thinking that collapsing multiples spaces into a single space might be okay, then realized that I have done citations with long lists of authors, and found it quite helpful (as in catching a few errors) to add multiple spaces to align names vertically. So I would say don't mess with the spaces in any way, neither adding nor removing. ~ J. Johnson (JJ) (talk) 21:34, 20 January 2012 (UTC)

Thanks for the support here. In response to Gadget850, the benefit of a space before the closing </ref> arises only if there are adjacent citations. It is rightly not allowed to put a space between them, i.e. </ref> <ref> should be </ref><ref>, so the best way to make the end of the first citation more visible is to leave white space before the closing tag.

Would any of you experienced hands care to raise a bug report or two for AWB as this user did for some similar points? – Fayenatic (talk) 20:41, 23 January 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)

Harvard references

The {{harv}} template is up for deletion: see Wikipedia:Templates for discussion/Log/2012 January 24#Template:Harvard citation. Please comment there. --Redrose64 (talk) 00:01, 25 January 2012 (UTC)

Tool needed

I seem to recall seeing a tool that would find duplicate citations and combine them into a shared citation with a name. For instance, it would convert:

blah blah<ref>Smith, p 32</ref> more blah.<ref>Smith, p 32</ref>

into

blah blah<ref name=Smith32>Smith, p 32</ref> more blah.<ref name=Smith32/>

Was I dreaming? --Noleander (talk) 04:53, 25 January 2012 (UTC)

I found it. it is

http://toolserver.org/~verisimilus/Bot/DOI_bot/doibot.php?edit=toolbar&slow=1&page=ARTICLENAMEHERE

--Noleander (talk) 18:13, 25 January 2012 (UTC)

AFAIR, {{sfn}} provides a better solution. — Dmitrij D. Czarkoff (talk) 19:03, 26 January 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)

Wikipedia link in lieu of citation

Somewhere I hope it says that linking to a Wikipedia article, in lieu of a citation, is insufficient, since each article is edited by different editors with different standards. And often, the Wikipedia article isn't cited either, or the cite is not WP:RS. Student7 (talk) 21:57, 26 January 2012 (UTC)

WP:SECONDARY: "Wikipedia is a tertiary source.... Some tertiary sources are more reliable than others, and within any given tertiary source, some articles may be more reliable than others. Wikipedia articles may not be used as tertiary sources in other Wikipedia articles, but are sometimes used as primary sources in articles about Wikipedia itself." --Orange Mike | Talk 22:10, 26 January 2012 (UTC)

Also WP:CIRCULAR. --Redrose64 (talk) 22:14, 26 January 2012 (UTC)