Wikipedia talk:Citing sources/Archive 44

Archive 40

←

Archive 42

Archive 43

Archive 44

Archive 45

Archive 46

→

Archive 50

Citing photographs held by GLAM institutions

I'm not sure if this is the right place to ask this, but why was this proposal to create a template for citing photographs held by GLAM institutions never taken further? It strikes me as eminently sensible to allow for citing an historical photograph as an object in an of itself, as opposed to merely citing the webpage that carries its collection record. — Hugh (talk) 00:56, 19 October 2017 (UTC)

The same reason you can't cite the original manuscript of the Book of Kells. WP cites published works, not original manuscripts. There is no reason this would magically not apply to graphical manuscript materials.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  05:21, 19 October 2017 (UTC)

Citation proposal at Village pump

Please see WP:Village pump (proposals)#Legal Citations in articles dealing with U.S. law Jc3s5h (talk) 10:36, 26 October 2017 (UTC)

Partial revert

@Nikkimaria: can you say why you reverted this? Edition number is definitely optional. The Google Books page examples just showed another two ways of writing them. And the page number edit tried to explain when page numbers are needed. SarahSV ^(talk) 20:07, 11 November 2017 (UTC)

Given that the usual reason for a new edition number is a nontrivial change or revision from one edition to the next, and given that different editions often have different page numbering, why would the edition number be generally optional? I don't understand why the edition number would be omitted. The other changes all seemed fine to me, and I don't see why they would be undone.— Carl (CBM · talk) 21:22, 11 November 2017 (UTC)

Given that the usual reason for a new edition number is a nontrivial change or revision from one edition to the next You must not be familiar with the textbook industry. --Izno (talk) 21:53, 11 November 2017 (UTC)

1. Volume is not always denoted by a number, or not solely
2. Edition per CBM
3. The explanation of when page numbers are needed is IMO not accurate - for example, sometimes long citations inline don't need page numbers (eg. if the whole book is being cited)
4. My understanding is that that formatting for GBooks pages is best avoided, due to metadata issues. Nikkimaria (talk) 23:10, 11 November 2017 (UTC)

Re: edition number, if you have the year of publication, that's usually enough. If there are two editions in the same year by the same publisher, then the edition details can be added. I don't understand your final point about formatting for GBooks pages. The guideline already includes this. All I did was add two other ways of writing them. SarahSV ^(talk) 23:41, 11 November 2017 (UTC)

The guideline already includes GBooks links - but it doesn't, absent your edits, include linking them within the page parameter. Nikkimaria (talk) 02:22, 12 November 2017 (UTC)

People do link them, and my edit showed how to do it for people who want to. SarahSV ^(talk) 04:09, 12 November 2017 (UTC)

Re edition number - people typically know the edition number of a text without knowing the year of reference, rather than vice versa. So if we wanted to omit something, the year would make just as much sense to omit. But the goal of a citation is not just to include minimal sufficient information - we also want the citation to be nicely informative to readers. — Carl (CBM · talk) 04:01, 12 November 2017 (UTC)

CBM, the edit didn't say people shouldn't add edition number, just that it's optional. I almost never see citations with edition number. "Smith 2017, p. 1" is standard, rather than "Smith, 3rd edition, p. 1". Ditto with long citations; there's only sometimes an edition number, so people are treating it as optional. SarahSV ^(talk) 04:09, 12 November 2017 (UTC)

As indicated at the top of that section, we don't intend for people to include every component of a full citation in a short cite; that doesn't mean everything not typically included in a short cite is optional for a long cite. This page should reflect best practices - for example, it discourages bare links as citations, even though people sometimes do that too. Nikkimaria (talk) 04:16, 12 November 2017 (UTC)

Adding place of publication is best practice too, but it's also optional. SarahSV ^(talk) 04:25, 12 November 2017 (UTC)

I don't particularly agree - IMO the edition is identifying for the reasons outlined by CBM whereas the publication location is not. Nikkimaria (talk) 04:37, 12 November 2017 (UTC)

I don't expect to see the edition in short citations, only in the full citation. But I don't see the argument to explicitly say that the edition number is optional in the full citation - I would see it as just as crucial as the title of the work. — Carl (CBM · talk) 05:08, 12 November 2017 (UTC)

Sometimes the information that ordinarily would be conveyed by an edition number is included in the title. For example, Robert's Rules of Order, Newly Revised or Astronomical Almanac for the Year 2017. The front matter for these works doesn't contain any clear statement of an edition number. (In the latter case it's debatable whether to treat it as a book or a serial). Jc3s5h (talk) 12:57, 12 November 2017 (UTC)

Robert's Rules is an interesting case - some bylaws explicitly say that the "latest edition" holds, while others refer to a specific edition. It is a big deal when a new edition is published because any minute changes have to be considered and taken into account (the current edition is 11). The edition is shown clearly on the cover image we currently show in Robert's Rules of Order and I have never had trouble identifying an edition when I had one in my hands. (The edition is also on the copyright page and on a dedicated page showing the history of the editions.) So I would personally expect that our articles should name the edition by number if they cite it. — Carl (CBM · talk) 13:30, 12 November 2017 (UTC)

I have the 8th edition. As CBM says, it can be important to know which edition you're using. Sometime decades ago I figured out mine was the 8th, and I wrote it on the cover. That information is nowhere to be found on the title or copyright page. I certainly wouldn't expect an editor to add an edition number if considerable effort was needed to figure out what it is. Jc3s5h (talk) 13:46, 12 November 2017 (UTC)

Agreed edition number should not be flagged as optional; it matters, and not all book-related site and such agree on publication dates (some use the date as stated in the fine print in the book, others use the literal date of release; near a year boundary, these can result in different years).  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  19:04, 12 November 2017 (UTC)

Linking to preprints of published articles

It is quite commonly the case (especially in physics and mathematics) that the final peer-reviewed published version of a paper will be pay-walled, but the preprint version of the same paper will be openly available to read on arXiv. In such cases, should we link to the final pay-walled version or the preprint version? Kaldari (talk) 00:30, 25 October 2017 (UTC)

Both? The DOI will generally link to the published version (which may be paywalled or not), so a separate link to the preprint can also be included easily if desired. — Carl (CBM · talk) 01:17, 25 October 2017 (UTC)

It looks like some of the citation templates offer a separate arxiv parameter, so I guess you're right, we can just link to both :) Kaldari (talk) 01:44, 25 October 2017 (UTC)

Yes, we can do both. Just be aware that pre-press versions do not always match the final one, since the pre-press arXiv stuff usually has not had peer review yet.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  10:14, 25 October 2017 (UTC)

Any pre-press version (and they're not all found at the several arXiv sites) should be identified as such. ~ J. Johnson (JJ) (talk) 22:22, 12 November 2017 (UTC)

Yep. It's a low-quality primary source. Not impermissible for certain things, but cannot be used (by itself) for anything we'd expect a secondary or even high-quality primary source for. The main benefit to ever including one at all is providing full-text access to [a version of] to research which has been treated in secondary sourcing (e.g. in a literature review) but the final version of which is paywalled. Pre-press stuff is frequently misused in citations as a sole citation, which fails WP:SPS in most cases (i.e., any cases other than those for which a self-published source is permissible, like WP:ABOUTSELF, and sourcing and attributing a quotation).  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  23:10, 12 November 2017 (UTC)

Quick question

Where do I report complaints about users (including admins) who revert my introduction of verifiable citations for material? 146.229.240.200 (talk) 09:56, 13 November 2017 (UTC)

You can report it at WP:ANI... but before you do, have you asked WHY the citations are being reverted? Have you engaged the editors in discussion on the article talk page? Blueboar (talk) 11:33, 13 November 2017 (UTC)

I'll ask them if it happens again. Looks more like a combination of WikiBullying and illusory superiority to me. 146.229.240.200 (talk) 12:08, 13 November 2017 (UTC)

It could be... or it could be that the cited sources were not considered reliable (or that there was some other problem). Without looking at the specifics, we can not determine why your edits were reverted. There are lots of legitimate reasons why someone might revert an edit. Don’t take it personally. It happens all the time. Blueboar (talk) 12:41, 13 November 2017 (UTC)

Commercial Radio Australia - Refs working for me (on multiple systems) but not for another user

I do not see any issues on Commercial Radio Australia either on computer desktop view or phone mobile view or phone desktop view. Nor do I see any issues with a direct link to the current or previous edits. Atsme, on the other hand (ref User_talk:Hydronium_Hydroxide#Commercial_Radio_Australia) sees:

• ^ Inglis 2006, p. 8. Harv error: link from #CITEREFGriffen-Foley2006 doesn't point to any citation.
• ^ a b Griffen-Foley 2009, p. 13. Harv error: link from #CITEREFGriffen-Foley2009 doesn't point to any citation.
• ^ Griffen-Foley 2009, pp. 25-26. Harv error: link from #CITEREFGriffen-Foley2009 doesn't point to any citation.
• ^ Griffen-Foley 2009, pp. 26,37. Harv error: link from #CITEREFGriffen-Foley2009 doesn't point to any citation.
• ^ Moran & Keating 2007, p. 103. Harv error: link from #CITEREFMoran_.26_Keating2007 doesn't point to any citation.
• ^ Alysen 2012, p. 238. Harv error: link from #CITEREFAlysen2012 doesn't point to any citation.
• ^ Harrison 2013, p. 61. Harv error: link from #CITEREFHarrison2013 doesn't point to any citation.

Anyone have any idea on where the problem is and how to fix it? ^~Hydronium^~Hydroxide^~(Talk)~ 04:23, 17 November 2017 (UTC)

I see:

• ^ Ingles 2006, p. 8
• ^ ^{a b} Griffen-Foley 2009, p. 13

etc. ^~Hydronium^~Hydroxide^~(Talk)~ 04:31, 17 November 2017 (UTC)

The problem is with the bibliography entries. The "cite-book", "cite-web", "cite-journal" (etc.) templates require a "ref=harv" parameter for the inline citations to link. My browser (Safari 11.0.1) displays the footnotes normally, but does not link when clicked. Atsme seems to be using a browser that at least gives a clue as to why these links are not working. I have already fixed a couple of them, and will now return and fix any remaining problems.—Jerome Kohl (talk) 04:45, 17 November 2017 (UTC)

Check out Template:Harvard_citation_documentation#Shortened_footnote. ^{Atsme📞📧} 04:48, 17 November 2017 (UTC)

Also if you add User:Ucucha/HarvErrors.js to your appearance preferences these problems become much more visible. —David Eppstein (talk) 04:51, 17 November 2017 (UTC)

Thanks all -- will do. ^~Hydronium^~Hydroxide^~(Talk)~ 04:54, 17 November 2017 (UTC)

WP Shortcuts -- question

These links in the box below (copied from the article) lead to short pages stating "This is a redirect" etc, rather than to in-depth explanatory pages. Is this intended?

Thanks, GeeBee60 (talk) 17:40, 28 November 2017 (UTC)

GeeBee60, that may be because they are in a template to be used in the article itself. WP:CS, WP:CITE, and WP:REF written elsewhere go where they are intended. StarryGrandma (talk) 02:18, 29 November 2017 (UTC)

Title case?

When adding a footnote that references an article from a news source or other web site, is it preferable to put the title of the article in title case? Or is it better to use the same case that the cited website does, which is often sentence case? I somehow got the impression that the Wikipedia Manual of Style recommends title case, but I can't find that anywhere.

References

— Mudwater ^(Talk) 11:05, 15 November 2017 (UTC)

If our MOS takes a position on that, I've yet to see it. I've always copied from the source with only two changes: De-shouting all-caps unless there is good reason for them (such as acronyms), and changing curly quotes to straight. If the entire title is all-caps, I convert to title case or sentence case, depending on the title content and my mood. I've never seen any other changes, and I obsessively notice little stuff like that. ―Mandruss ☎ 11:18, 15 November 2017 (UTC)

That's exactly what I do, and I haven't noticed any other treatment either. -- Michael Bednarek (talk) 12:43, 15 November 2017 (UTC)

In my view, this is the proper page to consider this issue, WP:MOS would be the wrong page to address it. This page allows any consistent citation style in an article. I try to identify which style seems to be in use in an article, such as Help:Citation Style 1, APA Style, Chicago Manual of Style, or something some editor cooked up for the occasion. I match the title capitalization of the other citations in the article.

Help:Citation Style 1 seems to be the most popular style on Wikipedia; that help page says "Use title case unless the cited source covers a scientific, legal or other technical topic and sentence case is the predominant style in journals on that topic." But of course that only applies when those templates are being used. Jc3s5h (talk) 14:28, 15 November 2017 (UTC)

Yeah, this is where it gets messy. The page you linked begins with: "It is not one of Wikipedia's policies or guidelines, as it has not been thoroughly vetted by the community." Even if it were a guideline, Wikipedia:Policies and guidelines says: "Whether a policy or guideline is an accurate description of best practice is determined by the community through consensus." Absent a community-level discussion, the consensus is what editors do. Neither I nor User:Michael Bednarek have ever converted to title case or seen anybody do that, in 16 years and 86K edits. Unless a herd of editors show up and say they do what that page says, I think it needs to be brought in line with consensus. ―Mandruss ☎ 16:11, 15 November 2017 (UTC)

The consensus being where? :D If a template wants to have documentation which is not obviously out of sync with an established policy or guideline, there is nothing prohibiting such a practice. To me, regardless of anything else, the important part is ensuring titles aren't in WP:ALLCAPS. --Izno (talk) 16:59, 15 November 2017 (UTC)

The consensus being where? :D - "Absent a community-level discussion, the consensus is what editors do." If a template wants to have documentation - The point is that the documentation lacks community consensus - either by discussion or, as far as we can tell at this point, by common practice. The doc itself states that it has not been vetted by the community. Its proper weight approximates zero. ―Mandruss ☎ 17:45, 15 November 2017 (UTC)

I do the same as Mandruss and Michael Bednarek, other than I start any subtitle with a capital letter ("Lorem ipsum: Foo bar baz quux"). We don't need a rule imposing sentence or title case on article and chapter titles. They're already used (per MOS:TITLES) on titles on major works such as books, films, journals, etc., and that seems sufficient. It would be a tremendous pain the backside to go apply title case to millions of mostly-lowercase article titles in citations, for no real benefit. It would also be a disruptive pain the backside to go around forcibly de-capitalizing them when inserted in sentence case (which plenty of periodicals and such actually do use) by editors copy-pasting or who simply prefer it that way; triggering a zillion watchlists for pointless trivial changes like that will not actually improve the encyclopedia.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  17:50, 15 November 2017 (UTC)

@SMcCandlish: Your statement "We don't need a rule imposing sentence or title case on article and chapter titles" must be interpreted to apply to titles of articles and chapters outside of Wikipedia, since Wikipedia does not have chapters. Because of WP:CITEVAR, MOS:TITLES can only apply to titles of external works in the running text of a Wikipedia article (or other areas outside the citations). The citation section is subject to WP:CITEVAR. Jc3s5h (talk) 18:23, 15 November 2017 (UTC); fixed link 15 Nov 2017 18:44 UT

Huh? This entire discussion is about and only about titles of works we're citing. It has nothing to do with WP-internal naming conventions for Wikipedia articles, which are governed by WP:AT, WP:MOS, and topical guidelines in Category:Wikipedia naming conventions.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  18:34, 15 November 2017 (UTC)

Perhaps I wasn't clear. MOS:TITLES has nothing to do with citations and we should completely ignore it when discussing citations. Jc3s5h (talk) 18:41, 15 November 2017 (UTC); fixed link 15 Nov 2017 18:46 UT

Your wish that it didn't apply is your own idea and is not reflected at all in actual community behavior. We reglularly normalize book and journal titles (just not chapter and article titles) to title case, whether you desire that this be true or not, and whether you think that ultimately has anything to do with MOS:TITLES or not. I refuse to get into another fruitless "WP:CITE regular versus WP:MOS regulars" circular pissing match with you, especially since its a false dichotomy – we're mostly the same people at this point.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  19:36, 15 November 2017 (UTC)

I think it's preferable to recase all sources consistently rather than copying inconsistent casings from different sources. But I don't have a strong opinion on whether it should be title case or sentence case; I think that's a style variation that we shouldn't prescribe. —David Eppstein (talk) 18:29, 15 November 2017 (UTC)

That would require an RfC, which would fail for the reasons SMcC cites. We're certainly not going to decide that in this thread. ―Mandruss ☎ 18:35, 15 November 2017 (UTC)

It would require an RFC to suggest that titles be recased consistently within an article? Really? Why do you think this is not already what WP:CITEVAR "citations within any given article should follow a consistent style" means? —David Eppstein (talk) 18:43, 15 November 2017 (UTC)

To my knowledge, the community has decided that CITEVAR applies to the choice between citation styles, and nothing else except for embedded vs. list-defined references. If you want to devote your editing time to converting citation titles, you won't get any objection from me. But I would object to a guideline to that effect, since I follow guidelines whether I agree with them or not. ―Mandruss ☎ 18:46, 15 November 2017 (UTC)

Most published citation styles do specify how the titles of works should be capitalized, although Help:Citation Style 1 is rather wishy-washy on the subject. As an example, APA Style requires sentence-case capitalization of titles. So choosing a citation style usually is choosing a title capitalization style. Jc3s5h (talk) 18:56, 15 November 2017 (UTC)

Fair enough. I do 100% of my work in CS1, so it's all I care about. CS1 is completely home-grown, is it not? If that's the case, we have no external standard to comply with. As I've said multiple times, we have no community-vetted internal standard in this area either. ―Mandruss ☎ 19:01, 15 November 2017 (UTC)

Right. CS1 doesn't impose a titling style. Some external citations style do so. Thus, if one is assiduously following one of the latter, it will amount to also selecting a title capitalization style for the duration of working within that particular citation style. And really no one cares. Wikipedians should not spend any further time arguing about this stuff. To the extent that WP:CITE contains citation style advice (rather than deferring to editorial whim), its intent is to forestall dispute not perpetuate it. If someone's editwarring to force-capitalize, or impose lower-casing on, all article/chapter titles then they need to stop.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  19:36, 15 November 2017 (UTC)

Titles of works are to be titlecased (journals, booktitles, etc...) per WP:MOSTITLES, which applies everywhere (including citations), while parts of those works (e.g. chapters, articles, etc...) can be either titlecased, or sentencecased. Doesn't matter which as long as the article is consistent with its choice. I prefer sentence casing myself, but if the dominant style used is title case, then I'll use title case. Headbomb {t · c · p · b} 19:04, 15 November 2017 (UTC)

I dispute the assertion by Headbomb that Wikipedia:Manual of Style/Titles applies everywhere. Scanning the thread titles in the talk page archive for that page, I see no discussion of citations. If it purported to apply everywhere, it would be in obvious conflict with WP:Citing sources. Jc3s5h (talk) 19:17, 15 November 2017 (UTC)

You've been disputing this for years, and the rest of the site doesn't pay any attention, and instead continues a) capitalizing titles of major works in title case, and b) accepting sub-works' titles in the case the were published in, within reason, because it's just too tedious to normalize them unless they're SCREAMING ALL-CAPS.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  19:36, 15 November 2017 (UTC)

"accepting sub-works' titles in the case the were published in" This is mostly bot-assisted (or copy-pasted) laziness, to clarify. There are zero style guides out there advocating inconsistency in casing. For foreign languages, sometimes people use sentence/title case as is standard in the original language, but people will often use the English conventions just as often if not moreso. Headbomb {t · c · p · b} 21:49, 15 November 2017 (UTC)

We all have limited time, and manually case-perfecting a copy-pasted citation title like "Mutation in and lack of expression of tyrosinase-related protein-1 (TRP-1) in melanocytes from an individual with brown oculocutaneous albinism: A new subtype of albinism classified as 'OCA3'" is time-consuming for very little benefit, is apt to causing whining and even warring from people convinced (wrongly) that WP "must" use the casing of the original publication, and (if one is not careful) is also apt to trigger objections from someone insisting on some really particular off-WP citation style that one didn't notice was in effect, if it is one that expects article titles in sentence case. I'm not sure what you mean by "There are zero style guides out there advocating inconsistency in casing". Sure, there are none that advocate inconsistency from cite to cite, but some give article and chapter titles in the format of my copy-pasted example, while giving book and journal titles in the format American Journal of Human Genetics and War and Peace; it appears to be the dominant style in academic journals right now. Regardless, what I'm suggesting is that we shouldn't care. As long as the main work's title isn't confusingly given a form like American journal of human genetics, which looks like a description not a title, there's no particular reason to give a damn if someone copy-pasted in "Mutation in and lack of expression of tyrosinase-related protein-1 (TRP-1) in [blah blah blah]" exactly as published. We probably can't expect a bot to figure out capitalization "correction" without screwing it up, and humans have better things to do with their time here. If someone really does want to title-case that, people shouldn't fight them on it, but wouldn't we rather that they did more useful thingse, like looking for missing citations and broken ones (dead URLs, etc.)?  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  22:27, 15 November 2017 (UTC)

If we're advocating letting lazy editors not manually re-case things if we don't want to (a position I'm ok with as long as we don't disallow re-casing by non-lazy editors) I should point out that MathSciNet uses sentence-case even for book titles (but not for journal names). So if a citation is copy-pasted from there, it may not title-case the main work. —David Eppstein (talk) 23:56, 15 November 2017 (UTC)

Yeah, I clean up after that all the time. I think what I'm getting at is there's really no way to force people to title-case properly, and it's more important that people add cites than format them, so the work is largely left to gnomes, if they want to bother. I do it, but only when otherwise improving the same citation. People aren't going to want to see watchlist hits just for correction of, e.g., Guns, germs and steel to Guns, Germs and Steel. The "as long as we don't disallow re-casing by non-lazy editors" matter is of course the key thing, but it doesn't seem to be a problem. I'm not regularly running into any such resistance. I also don't frequently change article/chapter style, because the real world seems to be moving more and more to using sentence case for those, and title case for "cover titles", so it just seems pointless to resist.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  00:40, 16 November 2017 (UTC)

For a long time I've considered tweaking the cs1|2 modules so that they will recase SHOUTING in |title= and |chapter= when the SHOUTING the whole title and IS MORE THAN ONE WORD. The module would drop pages with recased titles into a maintenance category so that gnomes and bots could apply appropriate fixes.

—Trappist the monk (talk) 01:25, 16 November 2017 (UTC)

That would be a bad idea, I think. It is too common to refer to the titles of computer science conferences by the acronyms of their sponsoring society and conference name ("ACM STOC 2016", "IEEE FOCS 2017") and that would be indistinguishable from shouting according to this test. In some cases (SIGGRAPH) the all-caps version is the official name of the conference and it would be incorrect to downcase it. —David Eppstein (talk) 02:16, 16 November 2017 (UTC)

Yeah, but a) just categorizing them for potential cleanup is harmless, as long as false positives can be flagged somehow; and b) it's actually more helpful for the readers, anyway, to expand some of these acronyms into something that's informative to someone other that people deeply embedded in the field in question; WP isn't written with experts in mind. Seems like a win-win and a no-harm-no-foul situation to me.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  03:32, 16 November 2017 (UTC)

Adding them to a maintenance category is harmless enough but Trappist was suggesting automatically recasing them. —David Eppstein (talk) 20:01, 16 November 2017 (UTC)

I would love to see something like Trappist suggests if one could simply indicate a preferred consistent usage throughout the citing article as is done for {{Use dmy dates}}. It will be far simpler to store |title= in sentence case and calculate the title case as opposed to doing the reverse. It can be difficult for a program to reliably determine whether a capitalized word in title case is a proper noun or not, but by storing the sentence case the difficulty is mooted. LeadSongDog come howl! 21:33, 29 November 2017 (UTC)

Is this a meme?

I know I sound dumb, but is this whole "Citation needed" thing a meme on wikipedia? — Preceding unsigned comment added by 65.175.240.110 (talk) 19:49, 24 November 2017 (UTC)

Originally a meme that rapidly became a bedrock core policy (see WP:V). Boghog (talk) 19:58, 24 November 2017 (UTC)

Well, I suppose xkcd 285 may have helped it along, but really attribution is pretty basic to academic honesty. It's just more important in a tertiary work by anonymous editors where you can't really punish bad behaviour. LeadSongDog come howl! 21:42, 29 November 2017 (UTC)

Inline citations on Talk pages

When inline citations appear on talk pages, the references often appear as part of the last comment section. If the usual ==References== <references/> is added at the end, then the references themselves appear in a separate section, and perhaps other comments will follow. This article should provide help for editors for that situation.--Dthomsen8 (talk) 01:54, 5 December 2017 (UTC)

This guideline is about citing sources in article space. Information about how footnotes appear in non-article-space would be doubly off-topic, because footnotes and source citations are two different things and because of the namespace. So I'm not convinced this would be a worthwhile addition. But as an aside, I think it's preferable to use {{reflist-talk}} than to use <references/>, just as it is preferable in articles to use {{reflist}}. And adding a new top-level section for references in a talk page is usually a bad idea; the references should appear in the same section as the comment that includes them. —David Eppstein (talk) 02:29, 5 December 2017 (UTC)

Strongly agree with David. Furthermore it should normally be the last item in the sections. Other contributions should be placed above it. Martin of Sheffield (talk) 13:23, 5 December 2017 (UTC)

Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum. Martin of Sheffield (talk) 13:23, 5 December 2017 (UTC)

Yes exactly. Paul August ☎ 15:35, 5 December 2017 (UTC)

References

In long discussions such as at AfD or elsewhere it is better to have a reflist for each comment using references. See Wikipedia:Articles for deletion/Constructal law. StarryGrandma (talk) 15:47, 5 December 2017 (UTC)

Are non-visual spacing modifications covered by CITEVAR

It is not clear by CITEVAR but would the bulk addition and/or removal of the wikitext spaces between the elements between a citation template parameters? Eg, an edit like this one [1] where the bulk of the article used spaces generally between the "|" and "=" elements of the cite templates with a few exceptions, but this edit principally removed them all. (If it were bringing a handful of refs in line to the same spacing style, I would agree, but its clear the spacing version was predominate). --MASEM (t) 14:29, 4 December 2017 (UTC)

• Does it really matter whether there are spaces or not? Blueboar (talk) 16:27, 4 December 2017 (UTC)
  • I know that the produced markup, it makes no difference, but for myself, I prefer to have the spaces to easily parse references in text. Arguably, the same distinction for list-defined vs inline references - the end use knows no better what happens, but it can affect how editing happens. (And this is also issues that this is stuff that we discourage the use of bots for without good purpose). --MASEM (t) 16:32, 4 December 2017 (UTC)

  So why are the spaces being removed? Blueboar (talk) 16:58, 4 December 2017 (UTC)

  I don't know, and I don't know if that's "improper" per CITEVAR. It seems like one of those things not to change en masse per CITEVAR/DATERET-type logic, but this is also not clear in P&Gs. --MASEM (t) 17:01, 4 December 2017 (UTC)

• There was this rfc that determined that WP:CITEVAR does not control vertical and horizontal template format which difference is, in my opinion, merely a matter of spacing between template parameters. But, that rfc doesn't say that it's ok to switch from one 'spacing' to the other without a consensus to do so. Without that consensus, it would seem to me that Editor Lordtobi may be in the wrong (pinged here because you should when discussing another editor's work).—Trappist the monk (talk) 17:08, 4 December 2017 (UTC)

  As far as I can see, the "tidy" citation code format fits best to those seeking to edit the code, and I do not really see a disadvantage in applying it. As Masem states above, the primary target is to bring all ref styles inline, though I usually stick with the afformentioned tidy format (for prior-given reasons), and I hadn't been issued about this before (until now, even if not directly), wherefore I was not aware it was being considered harmful or disruptive in any way. I'll go ahead and reset the citation format on that article to its original. Lordtobi (✉) 17:49, 4 December 2017 (UTC)

  I want to be clear that I wasn't sure if it was disruptive or not, hence just seeing if this fell under CITEVAR or not. I do think that we should have some efforts to harmonize whitespace markup (eg if someone has used vertical citations, and another editors adds a horizontal one, that last one should be converted), but it doesn't sound like this has the same weight as, say, switching from list to inline , or using one date format over another. --MASEM (t) 20:28, 4 December 2017 (UTC)

I would suggest that the first thing you should do is contact the editor who removed all the spaces, ask some questions, and start a dialogue. It could be that the other editor will not object to the spaces, once you explain WHY you used them. If so... there is no need to ask whether it is covered by the “rules”. Blueboar (talk) 22:13, 4 December 2017 (UTC)

As this isn't the first instance I've seen about someone objecting to someone else changing spaces in citations, I suspect that this actually is a question that needs to be asked, particularly if it's something people want to incorporate into scripts. Nikkimaria (talk) 22:16, 4 December 2017 (UTC)

Well, Lordtobi's edit summary on the diff is rather clear: space normalization within ciations. I can completely understand some type of normalization of citation format wikitext, that's a very valid goal particularly if it can be done with scripts and be pure busywork. But as Nikkimaria has pointed out, whether changing spaces is considered part of CITEVAR is very unclear, and it would not seem productive to just have this discussion with one editor. I think we should establish whether it is or isn't meant to be covered wiki-wide (or at least on this page where it is covered to some degree). --MASEM (t) 22:24, 4 December 2017 (UTC)

As someone who occasionally edits on an iOS device using desktop mode (which is to say I prefer to edit the wikitext) I can certainly testify that it can be difficult to place the edit cursor on an exact spot rather than selecting an entire word: leaving whitespace makes work much more productive and less error-prone. I can only imagine how tough that would be made for someone with a condition such as palsy. It can also be tough in many fonts to visually distinguish the pipe character from an adjoining l, I, or 1, leading to puzzling errors for low-visibility users. WP:ACCESSIBLE would argue for having the whitespace. LeadSongDog come howl! 16:29, 6 December 2017 (UTC)

FWIW, the most common format is {{cite foo |last=Surname |first=Forename |title=Whatever}} While there are many instances of an additional space after each pipe, and of spaces after or on both sides of the equals signs, and of all such spaces removed for maximum compression, none of those are as frequent as just the space before the pipe, which groups the parameter name and its value as a unit, and separates each such unit from adjacent ones. It's also common to insert a space after = and before a URL, to provide a line-wrapping point (many URLs are long).

Whether this is accessible enough is an open question; it probably has a lot to do with individual browsers and how they define a "word" when selecting one. E.g., I've noticed that Chrome has inconsistent behavior between how it treats text in a text entry box like the one I'm editing now and how it treats text in the URL and search entry bar at the top of the window, and neither of these are entirely consistent with what Safari or Firefox do, or what is done by various "stock" Web browsers that ship with Androids. Does a : divide "words"? What about a / or a | or a =? And so on.

Anyway, in WP:LDR blocks, each citation template parameter is typically on its own line with its value, though sometimes author info is grouped on one line. That vertical formatting is useful for refs grouped at the end of an article, but terrible in mid-article, since it interferes with the ability to easily get a sense of paragraphization.
 — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  05:30, 9 December 2017 (UTC)

Khai khrop

Hello - I would appreciate opinions on a situation I've encountered in the past couple of days, involving the article Khai khrop. It came up during New Page reviewing, and I tagged it for needing improvements to citation style as it had no inline citations (still doesn't at this time). After a bit of back-and-forth with an editor who is building the article, s/he removed the citation style tag with the edit comment "rm pointless maintenance tag: article uses a perfectly normal text-only inline variant of the harv style: if anyone dislikes it feel free to change it any way you want". The article is something I know nothing about, so I'm not about to get more involved in it.

References have been supplied in this article, but as general footnotes, without inline citations. I would appreciate more eyes on this little situation - if the editor's opinion is valid according to Wikipedia policy, then great - file closed. However, if his/her opinion on not needing inline citations is offside, can somebody with more gravitas please step in to 'advise' the editor to not remove a reasonable 'citation style' tag? Thanks in advance! PKT(alk) 22:20, 8 December 2017 (UTC)

PKT is mistaken. Parenthetical referencing is an acceptable method of providing inline citations, and this style of referencing is being used in the "Khai khrop" article. Jc3s5h (talk) 23:27, 8 December 2017 (UTC)

(edit conflict) Citations ought to use the relevant templates to ensure that both formatting and links are in a standard style. Regardless of this though, the article needs inline references such as {{sfn}} to link the assertions to the cited sources. It would also help, but is not required, if the sources were in English - it's pretty hard to check them if they use a differnt script and language. I've reformatted one of the citations (because it was the only English one) and use {{sfn}} to point to it. I trust that will be an example for all to follow. Martin of Sheffield (talk) 23:44, 8 December 2017 (UTC)

But see WP:CITECONSENSUS: "... The use of citation templates is neither encouraged nor discouraged ...". Wtmitchell (talk) (earlier Boracay Bill) 01:02, 9 December 2017 (UTC)

• I guess I'm the ultimate culprit here: I'm the one who removed the maintenance template that PKT had diligently placed on the article: I simply don't think the prime real estate at the top of an article should be wasted on big warning signs about exceedingly minor issues. Yes, the citation system is a bit wonky, but it's still clear and within reasonable distance of the standards. Martin of Sheffield, thank you for showing the world that snf exists! You're absolutely welcome to convert the remaining four citations to use it, but what you have done so far is to simply introduce inconsistency. Now, I just started looking for the guidelines that talk about consistency in citation style, but hey, we're already on their talk page! – Uanfala (talk) 01:33, 9 December 2017 (UTC)

Feel free to revert, I'm trying to help here. The OP asked for opinions on the situation due to a minor disagreement. There are several advantages in using the citations as well as consistency; for instance is "Mthai food" retrieved from http://food.mthai.com or is it from http://food.mthai.com/food-inbox/90416.html? The citation templates can also be used to add internal links. As regards {{sfn}}, again feel free to use any other method including moving the citations inline, but you do need to provide the link from assertion to citation. Don't let your sarcasm blind you to the need to do this somehow. "(ครูฑูรย์, 2012)" is one way to do it, but do remember that this is the English language wiki and very few readers will be familiar with Thai script. See WP:NOENG. You also need to aim for at least one reference per paragraph. Regards, Martin of Sheffield (talk) 12:07, 9 December 2017 (UTC)

Consistency in citation style

Does consistency in citation style imply that if one citation uses author last name and initials, one should not use last name and full first names for other citations in the same article?

Not to me. Suppose you decide on last name and full names as the style: you can't always use this style because the full names aren't always given in the source. If you decide the other way, and settle on initials as the style, then there's a problem with Chinese names (and some other Asian names). Consider a real example. One contribution to the Flora of China has its authors given as "Zhanhe Ji & Alan W. Meerow". We can turn "Alan W. Meerow" into "Meerow, A.W." (with or without a space between the initials) but different sources use "Ji, Z.", "Ji, Z-H.", "Ji, Z.-H." or "Ji, Z.H." when converting two "word" Chinese first names into initials. The IPNI – the official source for the names of botanical authorities – generally splits such names (see here for example), so "Ji, Z.H." is the more "official" for botanists.

The best solution in my view is to reproduce the usage in each source, and ignore minor inconsistencies. Peter coxhead (talk) 09:48, 14 December 2017 (UTC)

Both of those suggestions appear to be both practical and logical, and pretty much what I do anyway. However, there are occasionally GA and FA reviewers who claim that consistency in citation style does imply that all citations in a given article must either give full names or initials only, and that this is written somewhere. I am still waiting for one of them to link me to the policy, rule, criteria or guidance where this is specified, and asked here in case there actually is something I have missed. It seems that if there is then you two have also missed it... Cheers

It is certainly a consistent citation style to always use the author names as they were written in their publications, regardless of the fact that some publications might use initials and others might not. That's how MathSciNet generally lists its citations, for instance. —David Eppstein (talk) 11:43, 14 December 2017 (UTC)

Agreed. And any other approach, as I've argued above, is simply not possible in all cases. Peter coxhead (talk) 11:46, 14 December 2017 (UTC)

It doesn't have to be possible in all cases to be possible within a single article. —David Eppstein (talk) 11:57, 14 December 2017 (UTC)

RFC: Accurate dates in citation metadata

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

---

For citations that emit metadata for publication dates, is it necessary that the value of the date agree with the calendar that the metadata purports to use? For example, if a publication is 1 July 1750 Julian calendar, and the metadata emits it as 1 July 1750, Gregorian calendar, is this acceptable? Jc3s5h (talk) 14:54, 26 September 2017 (UTC)

Discussion of Accurate dates in citation metadata

This issue was raised at Help talk:Citation Style 1#Truthful publication dates but I believe it should be discussed by a wider audience. I believe metadata should be either accurate or absent. The simplest fix for the problem is to avoid emitting publication dates with a precision of a day or month if the year of publication is earlier than 1924, since Greece was the last country to change its civil calendar from Julian to Gregorian. Jc3s5h (talk) 14:54, 26 September 2017 (UTC)

• Of course the software should not be publishing wrong dates because of a technical error. This is the kind of mistakes that could be copied by others and perpetuated forever, becoming incredibly difficult to fix. Why do we need an RfC for this? If an article contains a date in the Julian calendar, a metadata system that can't handle Julian dates shouldn't be reading it. Diego (talk) 15:34, 26 September 2017 (UTC)
• Whatever date format the underlying database uses, the software should be capable of converting a date in any calendar and style to and from the database format. The citation templates don't have a way to specify what calendar is used in the date parameter; for accuracy, an extra paramater is needed, and quite a lot of conversion code, and validation. — Stanning (talk) 15:52, 26 September 2017 (UTC)
• A date on a publication is just a black box. If the source itself says "July 10, 1812" then our citation templates should also include the date as "July 10, 1812". It is incorrect to view the date string as being in particular calendar. It is just a string, like the author's name, which we should repeat without changing, for the sake of accuracy. It would be inaccurate for us to change the date to something other than what is written on the item itself. It is up to the person who reads the citation to interpret the date in whatever way they prefer, and this is facilitated only by us presenting the date exactly as the publication itself does. We do not need to hamstring ourselves based on COINS; we can simply say that we emit COINS-like citation information, which a consumer needs to interpret correctly. We certainly should not avoid stating the date at all solely because some documentation for COINS refers to a particular calendar. — Carl (CBM · talk) 21:51, 26 September 2017 (UTC)

• I submit that since most of the COinS data is read automatically rather than by human eyes, creating data in the COinS format is an irrefutable statement that the data follows the COinS standards, and no footnote on some obscure page in "Wikipedia:" space can turn a false date into a true date.

I do agree that we shouldn't change the date; if the source says July 10, 1812, then we should not change it to July 22, 1812. (Although it isn't just a string; changing it to 10 July 1812 would be acceptable). Let the COinS community fix their broken standard to support an unspecified calendar. An unspecified calendar is best for our purposes, since we may not know which calendar a particular date is stated in. Jc3s5h (talk) 22:16, 26 September 2017 (UTC)

If we just emit the literal date entered by the user, I don't see that as a false date - I see it as our best effort to specify the date. For most of the uses of Coins, which I think are to automatically populate our citation info into other bibliography managers that will go on to treat the date as a black box, I don't see any likely problems from just passing the date unchanged. In the relatively rare cases where a person cares what calendar a citation to an old work is in, they would have to work it out manually, of course. After all, we are not talking about using these dates to synchronize clocks or anything like that. — Carl (CBM · talk) 23:11, 26 September 2017 (UTC)

A problem I see is the people who write standards and software do not necessarily work with a wide range of dates and documents. My personal suspicion is they like to work with documents that were "born on the web", airline tickets, hotel reservations, and the like. My concern is that someone like that will blindly look at the specifications, have no personal knowledge of how they are used outside the 21st century, and blindly use the data as if it were correct. If you don't give them wrong data, they might just notice and look into the issue. Jc3s5h (talk) 23:30, 26 September 2017 (UTC)

I agree. But would they be more likely to notice a void of data, or data that blows up on them? Or (as I suggest below) embargoing all use of COinS? ~ J. Johnson (JJ) (talk) 22:21, 27 September 2017 (UTC)

• No. "Agreement" with a supposedly "purported" calendar is not necessary, and this idea of "accurate dates" is incorrect.

I don't like this notion that it is "up to the person who reads the citation to interpret the date in whatever way they prefer". Where publications (sources) have dates more specific than just year – typically newspapers, magazines, government bulletins, military orders, etc. – that is a definite datum not open for interpretation or alteration, and which we should ALWAYS supply as specifically and completely as possible, as otherwise we greatly impair verifiability. (Note that omitting the day or month does not make the remaining part less accurate, only less precise.) That a date may be incorrect if the wrong calendar is assumed is not an issue of date "accuracy", but in the assumption of a calendar, and we should not be coercing dates into any specific calendar

But perhaps what Carl means is that without any indication of which calendar a date is based on a reader may have to use some judgment. That should not be any problem, as most commonly there are only two alternatives. It is certainly a lesser problem than not having a complete date.

Whether the "metadata system" can "handle" Julian dates is not an issue, as all of Julian/Gregorian/cs1/COINS "handle" date strings as years, months, and days. I don't know that any metadata "purports" to use any calendar, and I don't believe it really matters, just as a book that says it was published in September but didn't come off the presses until November would still have September as its publication date. The only exception I can think of would be a source that claimed dual Old Style and New Style dates, but that is such a special case we need not worry about it. ~ J. Johnson (JJ) (talk) 23:33, 26 September 2017 (UTC)

Anybody who follows a complex standard well enough that many external projects, like Zotero can read it, is sending a strong message to the world that they are following the standard. COinS is a way to encode OpenURL, which in turn is specified by ANSI/NISO Z39.88-2004 (R2010) The OpenURL Framework for Context-Sensitive Services. That standard says a date follows W3C Date and Time Formats Submitted to W3C 15 September 1997. That in turn says it is a profile (that is, subset) of ISO 8601, which in turn requires the Gregorian calendar. So COinS cannot handle Julian dates; they are forbidden. Jc3s5h (talk) 00:49, 27 September 2017 (UTC)

I don't see that ISO 8601 forbids all use of Julian dates, only that it sets the Gregorian calendar as the common reference for comparison. The problem seems to be that CoinS wants everything referenced to Gregorian. More on this below. ~ J. Johnson (JJ) (talk) 22:10, 27 September 2017 (UTC)

I don't have a copy of ISO 8601 that I'm free to copy excerpts from. But I can point you to a page linking to a discussion draft of the next ISO 8601. It claims that Part 1 is essentially the same as 8601-2004 (with some corrections). Part 1 of the discussion draft "is essentially the same as 8601-2004 (with some corrections)". Part 1 states on page 7

This International Standard is applicable whenever representation of dates in the Gregorian calendar, times in the 24-hour timekeeping system, time intervals and recurring time intervals or of the formats of these representations are included in information interchange.

This implies the standard is inapplicable whenever other representations are to be included in information interchange, such as dates in the Julian calendar, or times stated with a 12 hour clock (e.g. 3 PM). Jc3s5h (talk) 10:40, 29 September 2017 (UTC)

Not exactly, unless there is some language that "inapplicable" applies to all cases not explicitly included. Lacking that, applicability for non-Gregorian calendars might be "undefined". Though I suspect it is the intent of 8601 to have all representations of time absolutely referenced to a common system of type "Gregorian". However, the problem is not with 8601, but with COinS: was it intended to coerce all dates into the Gregorian calendar? Or is that an unforseen and unintended consequence of adopting (albeit implicitly) 8601? The intent may have been to rely on 8601's tremendously useful standardized representation of a date for the purpose of information interchange, which works just as fine for Julian dates. But without requiring conversion to an absolute reference, which is not needed for citation work.

Of possible interest here: have any other "emitters" of COinS data dealt with Julian dates? Or for that matter, with Iranian dates? (I seem to recall seeing some modern works dated per the Iranian calendar.) Even better: how do libraries handle Julian (etc.) dates? (See below.) ~ J. Johnson (JJ) (talk) 22:03, 29 September 2017 (UTC)

• I think the cleanest solution is to add parameters for other calendars, not used by default, and always do a one-way conversion to Gregorian if one of them is used. Emit an error if two are used or one is used and |date= is manually filled and does not match the auto-calculation. Use that calculation to emit COinS Gregorian dates, which is the only format that spec can handle. Displayed to our users should be something like DD Month YYYY (Julian: DD Month YYYY) or Month DD, YYYY (Hijri: Month DD, YYYY), always putting modern, Western, Gregorian date first to agree with all the other date presentation on this site. No alternative dates should be shown if one taken from the source wasn't supplied to the template, since our goal is not to show converted dates in all the various calendars. The goals are and only are to a) have a consistent Gregorian-calendar date for our users (even if not a consistent date format, MDY vs. DMY) and for COinS users, and b) preserve the in-source date in another calendar if and only if that source used one (and didn't also provide a Julian date). In other words, avoid adding date-conversion trivia.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  14:45, 27 September 2017 (UTC)

• After reading the discussion so far, my idea of the cleanest solution is to add a language calendar parameter. For the time being, if the calendar parameter is absent or set to anything other than "Gregorian", the precision of the date is better than year, and the year is earlier than 1924, the metadata for the publication date is suppressed. This avoids emitting a converted date, which will confuse readers when the citation says 9 December 1745 but when they get an image of the Stamford Mercury from an archive it says 28 November 1745.

Later, after the COinS community fixes their standard to properly accommodate non-Gregorian calendars, the citation templates can be edited to follow whatever the improved COinS standard says to do, and all the citations will be instantly improved. Jc3s5h (talk) 16:54, 27 September 2017 (UTC). Corrected 06:11, 29 September 2017 (UT) Jc3s5h.

Like you say, converted dates are confusing. But they can be useful. Which leads to a somewhat subtle difference in the use of dates for comparison, and for identification.

In some cases – e.g., examining press coverage of the opening days of World War I in various countries using different calendars – it helps to have all dates referenced to a common calendar (typically the Gregorian). (Though I think a chart would be better.) But in bibliographical work a publication date is more of an identifier. If a source asserts a certain publication date then the asserted date is the controlling identifier, even if it is known that the "real" date was months later. It is analogous to pseudonyms: while we know that the "author" of The Adventures of Tom Sawyer is really Samuel Clemens, what the book actually says is "Mark Twain".

That COinS wants (?) to coerce all dates into Gregorian calendar therefore undermines the use of publication date as a bibliographical datum. Which suggests another option for us: embargo all COinS metadata until that standard is fixed.  :-)

For us, setting a calendar parameter has some merit. But there will always be cases where the calendar is not known (and arguably not even relevant), and making the default an assumed Gregorian leads to error. It might be better that the default (at least for dates prior to 1924) be "undefined". And let COinS choke on that. ~ J. Johnson (JJ) (talk) 22:16, 27 September 2017 (UTC)

Agreed. We agonize way too much over a crappy metadata spec that is a minor side matter when it comes to WP's purpose. This dates-as-identifiers rationale is also a good reason to replace unnecessarily precise publications dates (copy-pasted from Amazon) with just years for books. We do not need a |date=13 September 2015 for a book.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  22:38, 18 October 2017 (UTC)

You know what? Unless Jc3s5h is proposing to do this work himself, then I think that whatever User:Trappist the monk wants to do with the CS1/2 modules is fine with me. Nobody really wants inaccurate dates, but – we're WP:VOLUNTEERs, and this feels a bit like "I volunteer someone else to do all the hard bits". WhatamIdoing (talk) 01:54, 29 September 2017 (UTC)

Hold on, Trappist the monk has no power there, does not own the cs1|2 modules. Trappist the monk may or may not choose to do the work to implement whatever decision arises from this discussion but that is all.

—Trappist the monk (talk) 11:15, 29 September 2017 (UTC)

• Remember that we are talking about citations here. The only reason for including metadata (such as dates of publication) in a citation is to assist those who wish to read the source to find it (for example when searching in a library card catalog, or looking for a scanned copy in an on line database).

I am concerned that "correcting" a date would interfere with this. The question we should be asking isn't "Is this date accrurate?" but "What date is likely to be listed in a library's card catalog?". I suspect that most libraries would simply give the date printed in the source itself... and if I am correct about this, then we also need to be faithful to what is printed in the source itself. When it comes to citations being "faithful" to the source trumps being "historically accurate".

Now, in a different context ... for example when discussing the book in the text of an article about the book (or in the infobox of such an article)... then it might be appropriate to "correct" the date (or to give both Julian and Gregorian dates). But NOT in a citation. Blueboar (talk) 11:48, 29 September 2017 (UTC)

I think the WP:V policy would expect us to first, tell no lies. If what we want to express can't be expressed in a certain context, we shouldn't express it.

As for different contexts, with all the automated software roaming around cyberspace stitching together information, we can't expect that information given in the context of a citation will remain in that context. Consider automation that is assigned the task of finding out when a certain fact was first published. The automated process determines it was first published on January 1 in a Russian newspaper, but that's a Julian calendar date and the first publication was actually January 10, Gregorian calendar, in an English newspaper.

Of course, the automation could make the mistake anyway by screen scraping or parsing the |date= parameter in our wikitext, but at least we shouldn't lie by claiming it was a Gregorian date. Jc3s5h (talk) 14:24, 29 September 2017 (UTC)

WP:V does not really apply to the metadata of citations (if it did, we would have to cite our citations... and then cite those citations... ad infinitum). To put this another way - citations are the means by which we carry out WP:V. They are the means to verification of content.... not the content itself. (does that makes sense?). In terms of citation, it does not matter what the actual publication date of a source is... what matters is the listed publication date. Blueboar (talk) 15:59, 29 September 2017 (UTC)

The material in a citation can ordinarily be verified by consulting the work cited. If one were to annotate a citation with material that wasn't obvious, or contained in the cited work, it would indeed need a citation. And of course if the annotation (or in this case, metadata) is false, one won't be able to find a reliable source to support it. Verification is really a multi-step process: noticing the information in the citation, finding the source, reading the source, and comparing the information in the source to the article. A citation facilitates verification but is not, by itself, verification. Jc3s5h (talk) 19:22, 29 September 2017 (UTC)

As Blueboar says: yes, how do libraries catalog Julian dates? (Or other calendars.) Should we be faithful ("tell no lies") about what the source actually says is its publication date? Or the "real" time the asserted date corresponds to? Are we "lying" if we do not assert a calendar where COinS has not explicitly required that?

As to not expressing any date: isn't that kind of like cutting off one's face because one's nose is crooked? As a user of this data, it is going to be a lot easier to find my Russian newspaper article knowing it was published in January, 1914, (even if the day is "wrong") than just "1914". ~ J. Johnson (JJ) (talk) 22:07, 29 September 2017 (UTC)

J. Johnson (JJ) asked 'Are we "lying" if we do not assert a calendar where COinS has not explicitly required that?' If you use COinS, just by using it you're asserting all dates are Gregorian. Jc3s5h (talk) 22:13, 29 September 2017 (UTC)

Not at all! Assertions are positive, definite statements; some would say "emphatic" or even "forceful". Some thing I don't say can hardly be "asserted", no? Or is there some kind of hidden End-user license agreement here that legally inverts this distinction? ~ J. Johnson (JJ) (talk) 22:56, 29 September 2017 (UTC)

A small point about the "Russian newspaper" example. We are now used to news being reported immediately and therefore taking the reporting date as the event date. This just wasn't true in the past. An event might happen on Monday, the editor hears on Wednesday and sends a reporter and artist who travels by train on Thursday, returning Friday. The following Monday he writes up the article and the artist finishes the engraving by Tuesday missing the press deadline, so it is reported the following Thursday, 2.5 weeks late. If the event is further away or infrastructure breaks down, or it is before trains, then it might be 2.5 months. Now, is an argument about 1 Jan versus 10 Jan that relevant? Yesterday I was checking up on some death notices in the local weekly paper from c.1896 and they were often a week or two post mortem. Martin of Sheffield (talk) 09:31, 30 September 2017 (UTC)

Very true... and someone looking to read the newspaper account of the event will need to know the date of publication (i.e. the date printed on the newspaper where the account appears), not the date of the actual event itself. Blueboar (talk) 11:47, 30 September 2017 (UTC)

Just to illustrate the difference between dating in a citation, and dating in text... prior to about 1750 (I forget the exact year), England still started the new year in March... which means that a newspaper with a publication date of "February 15, 1735" and giving an account of an event that happened the day before, was actually discussing an event in (what we would call) 1736.

Now, which date should we use? It can get complicated... certainly when discussing the event in the main body of an article's text we should use (the updated) 1736, as that is the date of the event by modern standards. Writing the date in modern form helps the reader put the event in context with other events (otherwise the reader may get confused by the fact that an event occurring in February 1735 actually happened two months after an event that happened on December 1735... not several months before as you might assime)

However, If you go to a library and try to read this newspaper account, you will not find it if we used the modern form (1736) in our citation. The library database does not take the shift in dating into account. It uses the date printed on the newspaper (1735). So... For the purposes of citation we also need to use 1735... so that someone searching for the newspaper in a library can find it. Updating the date in a citation would be a disservice to our readers.

So... yes... an event that we say occurred in 1736 should indeed be cited to a newspaper with a date of 1735. Time travel! Blueboar (talk) 12:50, 30 September 2017 (UTC)

---

If we are all clear that the date of an event is not to be confused with the date of publication of a source: could we get back on-topic?

I should like to learn: how do libraries handle Julian dates? How do other generators of COinS data handle Julian or other non-Gregorian dates? ~ J. Johnson (JJ) (talk) 23:23, 3 October 2017 (UTC)

• The normal way of indicating a Julian year is either to say O.S. (for old style ) or use a dual date, e.g. 1698/9. See Old Style and New Style dates . As mentioned above, it is not just the shift of 10 or 11 days from Julian to Gregorian, but the usually simultaneous change in the Month when the year is said to begin (for example, for most purposes in England March until 1753 when it changed to January.) To the best of my knowledge, the articles I link to explain all this in accurate detail.

At least when I was a librarian, for most purposes libraries just try to get the year approximately right. Since conventional book publishing takes a considerable amount of time, there is always a period of several months at the turn of a year when the date on the book might be 1990 and the book actually published in 1991, or vice versa. Normally the cataloging record will just say whatever in on the book unless the discrepancy is more than one year. For journals, sometimes publication can be several years behind-- the papers for the volume intended for 1990 might not actually get published until several years later. (For purposes of academic priority, the key date is usually the date accepted for publication; for copyright and patents, when it actually appears.) For exact bibliographic purposes, or rare book cataloging, one tries to determine the exact date of publication and reports it will all necessary qualifier and explanations. There a large bibliographic literature on how to determine it, and on the various things that discrepancies can indicate.

Nice! Old Style and New Style dates is why I love Wikipedia: stuff I'd really like to know, but I can't do everything myself. And your comments are quite apropos here. Thank you. ~ J. Johnson (JJ) (talk) 21:35, 18 October 2017 (UTC)

For WP purposes, it doesn't much matter. When inserting a reference, I use whatever date in on the OCLC catalog record.

DGG ( talk ) 16:20, 17 October 2017 (UTC)

The "1698/9" style is going to be confusing for readers, as it's more commonly used (here, anyway) to indicate a span crossing a year boundary in the same calendar without exceeding an actual yaer, e.g. the 2009/2010 fiscal year, the 1998/1999 snooker season, etc. We could use "OS", I suppose, if it was linked to the Julian calendar article, e.g. OS. Or are we only concerned here with the COinS output?  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  21:48, 17 October 2017 (UTC)

The RFC is about metadata, so we are only interested in COinS. Since COinS only supports ISO 8601, the calendar is always Gregorian and the year always begins January 1. For the publication date presented in the citation read by a human reader, rather than a machine, WP:CITE allows any citation style, so one would try to discern which citation style is being followed (unless it's an ad hoc style just for that article) and read the associated manual or documentation, such as Chicago Manual of Style or Help:Citation Style 1. Frequently those resources will be silent on the issue, so the procedure adopted should be explained in a footnote. Jc3s5h (talk) 22:42, 17 October 2017 (UTC)

From DGG's comments I am more confident in my theory that, for bibliographic purposes, the importance of a "publication date" is not in capturing the exact moment in "time" (however it is pinned down) a source is "published", as for giving librarians, scholars, etc., an unambiguous identifier for (primarily) distinguishing different versions or editions of a work, and (secondarily) placing a given publication in the relative context of publishing history. As such, publication dates of "1776" and "1789" suffice to distinguish the first and fifth edtions of Smith's Wealth of Nations, and it matters not a wit whether those dates are New Style or Old Styler. Or even if the fifth edition was delayed and did not come until 1790. For bibliographic purposes "date" is simply a system of ordered identifiers. We don't need the precise number of seconds a date comes after (or possiby before) some event in 1970; that is entirely irrelevant.

On the otherhand, ISO 8601 is concerned with having an unambiguous data representation of actual time intervals. Which it pins to the Gregorian calendar. Jc3s5h has framed this discussion as a matter of accurate dates, and whether the value of a publication date should agree with "the calendar that the metadata purports to use" (i.e., the Gregorian calendar). But is this to be "accurate" in respect of a clock? Or of the "date" specified on the title page, by which a book is identified and catalogued?W

At 10:40 29 Sep. Jc3s5h quoted 8601 that it "is applicable whenever dates in the Gregorian calendar" are used, which he summarized as implying that 8601 is "inapplicable" (his emphasis) for other representations or calendars. If COinS requires (as he implies) Gregorian dates, then feeding it Julian dates is wrong. But so would be alteration of the data, or providing incomplete or misleading data.

I think there is a simpler solution: don't generate COinS data when an Old Style or non-Gregorian date is involved. That avoids any conflict between a COinS date and what is on the title page. ~ J. Johnson (JJ) (talk) 21:41, 18 October 2017 (UTC)

The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Wikilinking authors in references

I noticed some wikignomes remove authorlinks from footnotes when a particular author is mentioned somewhere in the text. In my opinion, a wikilink to an author is useful in the "References" section, since the section is usually remote from article text where the name may have possibly be mentioned, and usually when I see a book, I am curious what else the author wrote, regardless article content. What is your opinion? Staszek Lem (talk) 00:19, 21 December 2017 (UTC)

Agreed. A related question is whether we should link only the first entry in the bibliography: there is a case for linking them all. When I end up looking at a bibliography item it's most often because I've been taken there by a link from the short citation: with this entry at the top of the screen I can't see without having to scroll up if there are any previous entries by the same author. – Uanfala (talk) 00:27, 21 December 2017 (UTC)

An example would be nice to see if there are any extenuating issues, but in principle I agree as well. PKT(alk) 01:14, 21 December 2017 (UTC)

My own tendency is to link all occurrences of the authors, except that I avoid linking an author of a publication within an article about that author. It makes it easier to copy and paste references that way, and also easier to tell which authors are missing articles (and might be a candidate for a new article). —David Eppstein (talk) 01:20, 21 December 2017 (UTC)

I agree that authors ought to be linked in footnotes, even if they are linked in the article text already. I think that in bibliographies, which should be in alphabetical order, per WP:REPEATLINK they should be linked on 1st occurrence only (possibly using |author-mask=). If the authorlinks occur in inline references, they should usually be linked every time. Linking when author=subject is obviously a bad idea. -- Michael Bednarek (talk) 04:55, 21 December 2017 (UTC)

Linking when author=subject is obviously a bad idea. Not necessarily so. In the cs1|2 templates, an |author-link= to the author page on the author page renders in plain text; it does not link to the current page nor does it bold the author name. There is some benefit to this because that template can be copied elsewhere and the link to the author's article works without need for tweaking the template. Additionally, |author-link= in combination with |author-mask= does not link the underscores that mask the author's name. And one last thing, setting |display-author=0 hides a linked author name; this can be useful when listing the author's writings (where the author is the only author).

—Trappist the monk (talk) 10:21, 21 December 2017 (UTC)

(squeeze) a) I didn't know that about CS1|2 templates; b) there are plenty of citations that don't use them, and that's what I, and I suspect David Eppstein, was referring to. -- Michael Bednarek (talk) 10:50, 21 December 2017 (UTC)

I sometimes link authors in references...if they have a Wikipedia article, of course. Flyer22 Reborn (talk) 05:06, 21 December 2017 (UTC)

And I commonly link the publishers. Flyer22 Reborn (talk) 05:07, 21 December 2017 (UTC)

To "link all occurrences of the authors" because "[i]t makes it easier to ... tell which authors are missing articles (and might be a candidate for a new article)" is a terrible "sea of red" idea and against our linking guidelines. We should never red-link anything not likely to be notable, and approximately 99.99% of writers are not notable, the vast majority of them being random news journalists, and most of those who are not being random academics. Please use WP:Common sense and link blue-links, and red-link cases where you are pretty sure we should have an article on the person.  — SMcCandlish ☏ ¢ >^ʌⱷ҅_ᴥⱷ^ʌ<  02:25, 23 December 2017 (UTC)

I meant only linking the authors with article; I tend to leave authors without articles unlinked, even when I'm convinced they're notable. —David Eppstein (talk) 02:39, 23 December 2017 (UTC)

General References

are articles allowed to consist only of general references? Are inline references essential?Egaoblai (talk) 19:07, 27 December 2017 (UTC)

As outlined at WP:GENREF, general references are allowed eg. when all of the content has a single source; generally better-developed articles use inline references. Nikkimaria (talk) 19:12, 27 December 2017 (UTC)

I am quite dubious of any article where "all of the content has a single source". (And especially if all of the content is supported from just, say, a single page.) The role I see for "general" references is where some source has something that applies "generally" to the whole article. (Though I have yet to see a clear instance of that.) But all the specific content in an article should be specifically cited. Which implies having inline "references". ~ J. Johnson (JJ) (talk) 22:44, 28 December 2017 (UTC)

Agree with the above answers. Or to put it another way: the answer to both of the questions asked is: Yes.

Articles are indeed “allowed” to be based purely on general references... however, inline references are indeed “essential” for any half way decent article to grow beyond a basic stub. Blueboar (talk) 00:52, 29 December 2017 (UTC)

A stub of a single sentence – and a short sentence at that, of a single fact or assertion – would be the limiting case of where a single "reference" specific to that content is effectively a "general" reference. But: where different points in the content are attributed to different sources, or even different locations within the same source, there needs to be in-line citations. We should not tolerate letting editors say (essentially) "I took material from these sources, and if you study them you should be able to determine from which one, and where." Wherefore I find any article (aside from a trivial stub) "based purely on general references" – that is, without any in-line citations – to be deficient. ~ J. Johnson (JJ) (talk) 23:26, 29 December 2017 (UTC)

Meh... it is possible to write more than a single sentence based on one single “general” reference... for example, on many topics you could write a multi paragraph “start” level article, cited entirely to the Encyclopedia Britanica entry on the same topic. Sure, we encourage editors to do better than that... and as the article grows we SHOULD do better... but something like that is acceptable to do in the early stages of a writing a new article. Blueboar (talk) 23:50, 29 December 2017 (UTC)

It is not forbidden, but it is creating a problem for later editors who will not know what all is attributed to that/those general references. The practice should be strongly discouraged as it is trivially easy to add inline refs to a general ref that is already defined. I think this is a remnant of very old (for Wikipedia) policy dating from the days before inline references were required. It is only marginally better than no references at all, and often practically equivalent. There are a lot of articles with general references listed in addition to inline, but they are for practical purposes useless for verification. It may be time to get rid of the option and only allow inline. · · · Peter (Southwood) ^(talk): 16:34, 31 December 2017 (UTC)

Sure, sometimes it takes more than one sentence to explain some item from a source, but that is somewhat of a special case. My point is that in the limiting case of where a stub contains only a single fact or assertion – perhaps I should qualify that with "no matter how many sentences it takes to explain that single fact or assertion" — then a single citation ("reference") suffices for verifiability, and it seems to hardly matter if it is in-line or "general".

(Except that it does matter. Addition of content outside of the scope of a "general" reference makes it no longer general. And if it is not made specific – i.e., in-line – it takes on a ghostly character.)

The problem with "general references" is where there are multiple elements that need to be cited, to either different sources, or different locations within a source, but an editor can't be troubled to add those details. (Lack of specificity is a slightly different but related problem.) As Peter says: we can't tell "what all is attributed to that/those general references." I can see sources being useful generally for understanding a topic, and perhaps even heavily relied on by an editor in evaluating and balancing the content, and therefore ought be referenced. But for purposes of verification specific content needs specific citation. Which is to say: in-line citation.

I agree with Peter that the typical usage of "general references" should be strongly discouraged. And, for the purpose of verifiability, should no longer be condoned. ~ J. Johnson (JJ) (talk) 21:21, 31 December 2017 (UTC)

"except to the extent that this affects page numbering"

In the section WP:SAYWHEREYOUREADIT, it says

"So long as you are confident that you read a true and accurate copy, it does not matter whether you read the book using an online service like Google Books; using preview options at a bookseller's website like Amazon; on an e-reader (except to the extent that this affects page numbering); through your library; via online paid databases of scanned publications, such as JSTOR; using reading machines; or any other method."

Now I read the above as telling me that I should cite the book even if I have never seen the physical book but instead read it on an e-reader -- but only if I am confident that my e-reader copy is showing me the same material that I would see if I had the physical book. I also read the above as telling me that I should use the page number from the physical book, not whatever page numbers my e-reader assigns to the material.

Over at Wikipedia:Village pump (miscellaneous)/Archive 57#Kindle location instead of page numbers there are several editors who read the above as telling them that if they read it on a kindle they should cite kindle and use kindle page numbers. When I asked

"Would it be acceptable for me to start citing sources using the page numbers from a Telcon Zorba?[2] Because that's about how useful those Kindle page numbers will be 20 years after they stop making Kindles."[3]

I got the reply

"If YOU read it using a Zorba, yes. That’s what you should say on the citation."[4]

One of the two interpretations of this policy is wrong. Either I should cite the physical book with the page numbers from the physical book even though I read an electronic version of it on a Zorba, or I should cite the file on the Zorba and use the page numbers the Zorba assigns to the document. Could we have some clarification as to which is correct? Also, would some minor rewording of the guideline help prevent future confusion? --Guy Macon (talk) 16:04, 16 January 2018 (UTC)

False cite tag ?

Need a template tag for false citations - citations which are said to support the writing but which do not. -Inowen (talk) 00:59, 24 January 2018 (UTC)

I agree. It looks like many or most of Supasun's meteorology-related contributions seem to link to a general page and thereby lack a proper source [5] Geographyinitiative (talk) 01:11, 24 January 2018 (UTC)

Would one of the templates in Category:Inline citation and verifiability dispute templates do the job? – Uanfala (talk) 01:23, 24 January 2018 (UTC)

In the case the OP is requesting, Template:Failed verification and Template:Irrelevant citation would seem to be relevant. --Izno (talk) 02:06, 24 January 2018 (UTC)

There is a list of inline templates at Citation underkill#Templates. QuackGuru (talk) 14:20, 24 January 2018 (UTC)

Rfc on reference date format at 'Tesla Model S'

I have started an RFC at Talk:Tesla Model S#RfC about date format in references about whether an article using MDY date format in the text is allowed to have yyyy-mm-dd date format in references or not. There was also discussion in the talk topic just above it at Talk:Tesla Model S#Date format. Please answer there, not here.  Stepho  talk  04:59, 28 January 2018 (UTC)

Is social media (Facebook, Twitter, etc) a ok source if the profile has been verified?

For example on a famous person's verified Facebook profile if he says his birthplace was so and so is it reasonable to write on Wikipedia what their birthplace is and provide their Facebook post for a reference? Facebook has the "verified page" checkmark which tells that facebook confirmed a page actually belongs to a famous person. A145029 (talk) 05:34, 21 February 2018 (UTC)

A secondary source would be best....many people lie about their age eg. Joan Crawford#Notes.--Moxy (talk) 05:39, 21 February 2018 (UTC)

According to WP:BLP which clearly mentions not to use tweets. An excerpt from that page:

Never use self-published sources—including but not limited to books, zines, websites, blogs, and tweets—as sources of material about a living person, unless written or published by the subject of the article. "Self-published blogs" in this context refers to personal and group blogs. Some news organizations host online columns that they call blogs, and these may be acceptable as sources so long as the writers are professionals and the blog is subject to the newspaper's full editorial control. Posts left by readers are never acceptable as sources.[2] See #Images for our policy on self-published images.

Fylindfotberserk (talk) 10:08, 21 February 2018 (UTC)

Discussion at WT:JAPAN#Date formats

 You are invited to join the discussion at WT:JAPAN#Date formats. -- Marchjuly (talk) 05:28, 1 March 2018 (UTC)

Collapsible references

Is it against Wikipedia rules to put the Reflist for a page inside of a collapsing frame?
Something like:

References

• Reference 1
• Reference 2

• Reference 3
• ...

• Reference 349
• Reference 350

<div class="NavFrame collapsed">
<div class="NavHead"> References</div>
<div class="NavContent" style="text-align: left;">
{{Reflist|3}}
</div>
</div>
I'm thinking specifically for a sports transaction page I have been helping with that has nearly 350 references, and the list just takes up so much space. –uncleben85 (talk) 04:24, 27 February 2018 (UTC)

I don't think that's advisable because it's similar to a scrollbox: Wikipedia:Citing sources#Footnotes. DrKay (talk) 10:14, 27 February 2018 (UTC)

Per MOS:DONTHIDE - "Collapsible templates should not conceal article content by default upon page loading. This includes reference lists, tables and lists of article content, image galleries, and image captions." Apart from anything else it's an accessibility issue - some users/devices with no javascript support would not be able to "un-collapse" the content, and therefore could never access it. -- Begoon 10:42, 27 February 2018 (UTC)

Web pages have no natural maximum height, so there is no reason to try to conserve vertical space by hiding the references. They are a key part of the article, not a kind of fine print to skip over. — Carl (CBM · talk) 13:22, 27 February 2018 (UTC)

I agree; references should not be hidden. Peter coxhead (talk) 21:26, 27 February 2018 (UTC)

Likewise. I looked at the article in question: it simply has a lot of data, which needs a lot citing, and that's how it is. I did try giving it a slightly more compact format, but that doesn't really change anything. (Feel free to revert.) ~ J. Johnson (JJ) (talk) 20:07, 1 March 2018 (UTC)

Citing via ProQuest

I'm seeing a lot of citations recently that reference sources via ProQuest, with a URL of the form https://search-proquest-com.ezproxy.sl.nsw.gov.au/docview/107708906?accountid=13902. This takes you to a login page, which asks for an account number. I've discovered that you can enter the account in the URL (always 13902), and that will get you through to the document. However, it seems a little questionable. Is this account something it's legit for any Wikipedia reader to use, or are we piggybacking on some individual's account? Should these citations be marked as 'subscription/registration required'? Colonies Chris (talk) 15:58, 4 March 2018 (UTC)

Title parameter in magazine citing

Just to clarify, the |title= parameter in {{Cite magazine}} is the title of the magazine issue? I've been looking through some magazine citations and noticed that some citations are using the title parameter for the section title where the cited material came from, which is problematic for consolidating multiple citations of the same magazine issue.

More generally, for some magazines, I'm unsure what would actually be the title; the front of the magazine does not necessarily have a clearly distinguished title. E to the Pi times i (talk | contribs) 17:27, 11 March 2018 (UTC)

|title= gets magazine article name in the same way that |title= in {{cite journal}} gets article name. For cs1, the distinction between {{cite magazine}} and {{cite journal}} is how |volume= and |issue= are handled; otherwise, they are more-or-less the same. |magazine= holds the name of the magazine.

—Trappist the monk (talk) 17:39, 11 March 2018 (UTC)

So there's probably no point (and no standard way) in creating a reference at the bottom for the whole magazine, if different articles from the same magazine are used? I guess that makes some sense, since magazines are more article-based, and it's not necessarily cohesive material that all connects together. E to the Pi times i (talk | contribs) 17:55, 11 March 2018 (UTC)

If I understand what you are asking, and I'm not sure that I do, because a magazine is typically composed of multiple articles, each with its own author(s) and pagination, then I think not. But, if you have a real-life example, that is usually helpful in answering these kinds of questions.

—Trappist the monk (talk) 18:15, 11 March 2018 (UTC)

Which page number to use when citing PDFs?

I'm citing http://www.nyc.gov/html/dot/downloads/pdf/nyc-bridge-traffic-report-2016.pdf for Brooklyn Bridge. The information I'm using is located on page 21 according to the PDF viewer and page 7 according to the PDF (see image).

So which one do I use?

--Annoyedhumanoid (talk) 16:06, 11 March 2018 (UTC)

Whichever is in the source (your viewer is not the same as my viewer)--in this case, page 7. --Izno (talk) 16:18, 11 March 2018 (UTC)

When the page number used in navigating in Adobe Reader or equivalent software disagrees with the number that appears on the page, I always use the number that appears on the page. Frequently, the PDF is a scan of a book, and some readers may buy, or borrow from a library, the paper book rather than viewing the PDF. Likewise, some readers may print the PDF on paper rather than reading it on a screen. The readers who are working with a paper version cannot use the electronic page number, but readers reading on a screen can use either kind of number. Jc3s5h (talk) 16:20, 11 March 2018 (UTC)

Use the number that appears on the page, not whatever the PDF reader says. This makes the citation useful for someone who has a printed version of the source. This comes up often with journal articles - the article might be on pages 200 to 220 of the journal, so saying that you say something on page 3 doesn't make much sense. Remember that you are citing the source, not the PDF of the source - you just happen to be looking at the source in PDF instead of paper. — Carl (CBM · talk) 16:44, 11 March 2018 (UTC)

Yes.

There are cases where a pdf incorporates multiple sources with their own pagination, and it is a convenience to the reader provide the pdf "page" (frame?) number, but that is secondary to the pagination of the source. ~ J. Johnson (JJ) (talk) 22:02, 11 March 2018 (UTC)

To qualify as a "reliable source" does the document have to be published as a book?

What about an interpretation board in the official museum of the history of the Salvation Army, are you expecting me to upload photos of what's published as visitor information in the visitor centre in the International Heritage Centre or merely state "this information can be verified by sending an email to (Salvation Army employee email address)" Adrian816 (talk) 18:55, 11 March 2018 (UTC)

@Adrian816: See WP:Published on what Wikipedia considers a published source: "created for distribution and actually distributed to the general public". The interpretation board is not distributed to the public. Neither are emails that we'd have to request. Also, Wikipedia has millions of readers - just how many emails would the Salvation Army be willing to answer? Furthermore, in your particular case, the Salvation Army wouldn't be an independent source, whether that's its official museum or any emails they may send. On an unrelated note, this page is meant to discuss the improvement of WP:Citing sources. For questions about sources, please use the WP:Help desk or the WP:Teahouse. Huon (talk) 19:23, 11 March 2018 (UTC)

For question about source WP:RSN is more specific if the above two don't help. --Emir of Wikipedia (talk) 20:05, 11 March 2018 (UTC)

"Reliable source" and "publication" are two independent aspects, both of which get quite murky at the edges. A source is considered reliable (in some respect) when it is considered "consistently good in quality or performance; able to be trusted", or there are reasons to expect that. This is usually the result of a process, such as the enforcement of standards (like WP:V), and the intent and commitment (or not) of those setting up or driving the process. Reliability is a matter of the reputation of (mainly) the publisher, which is independent of the form of publication. ~ J. Johnson (JJ) (talk) 21:55, 11 March 2018 (UTC)

If an "information board" means a plaque or sign of some sort that's a semi-permanent part of the Centre, and the place it's located is accessible to the general public (admission fees are allowed), it is "published" and you could use {{cite sign}} to cite it. A photo isn't necessary, but might help if the sign is later removed (e.g. in a remodel). Anomie⚔ 11:12, 12 March 2018 (UTC)

Way to cite pages behind JavaScript forms?

This is a pretty basic question, but I can't find an answer anywhere. Let's say you want to cite a reference which is behind a JavaScript action, for example if you wanted to cite a specific corporate filing found using the search form at this URL. Once you make a choice on that starting page, then enter a search term on the next page, you get search results, and then you can click one to display a filing. The problem is that all of these pages are under the same URL. Is there a suitable way to cite a live URL like this if it requires the reader to supply his own input to get the actual intended citation? Or is there a way to cite an archived page of the actual filing page? It seems like this must have been discussed multiple times on Wikipedia, but I don't know where. A lot of potential citations are behind "walls" like this. --Iritscen (talk) 17:31, 9 March 2018 (UTC)

Assuming that such a thing passes sourcing policy, I don't know of a way to handle it besides the following:

References

―Mandruss ☎ 17:53, 9 March 2018 (UTC)

Your "pages behind JavaScript forms" confused me, until I realized you are accessing a database of reports. Some databases are set up to use a specfic url for each item, others use some kind of query language. Which could be done in JavaScript. Or any other language.

But are you perhaps confounding "citation" with a mere Uniform Resource Locator? That is, don't these documents you are accessing all have titles? Dates? Some kind of index or filing code? If they existed only as paper documents, how would you cite them? With full bibliographic details, of course, such as title, date, etc. The url only specifies an on-line location, without any other description or characterization of the document, and is only one part of a citation. So one answer to your question is: cite those documents like any other documents. As the Illinois Secretary of State presumably has many certificates of good standing, and you want only the one for the John Deere Foundation, that should be specified in the citation. Possibly that could be the title, with "CORP/LLC - Certificate of Good Standing" the series. Or "John Deere Foundation" could be an identifier. If sufficient information is given, I don't think you need to give the reader explicit instuctions. (E.g.: "click 'continue. Enter ....") You might look for other cases where these items have been cited, and see how they do it. ~ J. Johnson (JJ) (talk) 22:47, 10 March 2018 (UTC)

Thanks for both of your input. I searched WP's external links for previous uses of the Illinois SOS URL and found that there was some precedent for applying it as a citation for incorporation dates. The state's listing for each corporation filing provides a file number which can be supplied to the Citation template as a parameter, telling the reader which listing was the basis for the citation. --Iritscen (talk) 14:43, 12 March 2018 (UTC)

A few questions

I have a few things I'm not sure about while I've been formatting some citations for a video game article. My citation formatting has taken place here, but I have three specific concerns below (although you are certainly welcome to go and correct things there, whether or not I've mentioned them here.)

1. When should {{Cite news}} be used? Since I'm doing an article about video games, I don't know if some of the sources technically count as news. My choice of when to use Cite web over cite news has been largely based on whether the information in question has a date, since most things with dates were published in a news-like way, while webpages without dates haven't. Is this news or web:

Keiser, Joe (August 2, 2006). "The Century's Top 50 Handheld Games". Next Generation. Archived from the original on October 10, 2007. {{cite news}}: Unknown parameter |deadurl= ignored (|url-status= suggested) (help)

2. Should pseudonymous authors be cited under their pseudonyms:

"GameSpy's 2003 list of the 25 most overrated games of all time". GameSpy. Archived from the original on 24 February 2006. Retrieved 1 May 2015. {{cite web}}: Unknown parameter |dead-url= ignored (|url-status= suggested) (help)

3. And what should I do when I want to use a source in a short citation ({{sfn}}), but it doesn't have an author:

"1UP's 2005 list of the 10 most overrated games". 1UP.com. Ziff Davis. 4 April 2005. Archived from the original on 12 May 2012. Retrieved 14 October 2007. {{cite news}}: Unknown parameter |deadurl= ignored (|url-status= suggested) (help)

E to the Pi times i (talk | contribs) 14:42, 13 March 2018 (UTC)

1. {{cite news}} and {{cite web}} are equivalent for all the parameters of interest for web-only sources. As this is the case, "dated things go with cite news" is a pretty decent rule of thumb in this domain.
2. Give the author that is directly in the article cited (WP:SAYWHEREYOUGOTIT principle). (You might consider including a comment at the end of the template with the author's real name, since I do not know if the template metadata will be screwy if you include it elsewhere, and it certainly will be screwy if you include it as actual data in the template.)
3. Template:Sfn#No author name in citation template.

--Izno (talk) 14:59, 13 March 2018 (UTC)

Further to #1: Next Generation (magazine) so {{cite magazine}}

—Trappist the monk (talk) 15:03, 13 March 2018 (UTC)

While actually, that is linked to that in the interrim, but I think that link is wrong, because the actual magazine was discontinued before that article was written. So I will probably just remove the link, and maybe do a bit of research on if the website is a reliable source.

@Izno and Trappist the monk: Much thanks for the guidance. I have incorporated your feedback where relevant. E to the Pi times i (talk | contribs) 19:01, 13 March 2018 (UTC)

How Best to Cite This Raw Data as a Reference

It would be easiest to provide the example: Someone has asserted on http://en.wiki.x.io/wiki/Donna_(given_name), that Donna occasionally, albeit rarely as a surname, and there was an indicator there saying a reference was needed. I went out to the Census Bureau reports on the Census Bureau site, and they serve a spreadsheet with every surname that they saw more than 100 times in the 2010 Census. That CSV or XLS spreadsheet supports the above statement, but is a file served up to be downloaded. How can we, if possible, use this data as a reference and cite it? Spawn777 (talk) 22:01, 26 March 2018 (UTC)

Can you supply the URL of the web page that serves the spreadsheet. More than likely a {{cite web}} of that page will do. Note that every census bureau that I know of is limited to a single country, so you will need to qualify the reference with the |location= parameter.  Stepho  talk  22:20, 26 March 2018 (UTC)

{{cite report}} is perfect for this kind of thing, since what you're describing is a report from the census bureau. Even though it's not in written form, this is the most appropriate citation template (since {{cite web}} is intended for more miscellaneous things.) E to the Pi times i (talk | contribs) 23:50, 26 March 2018 (UTC)

On the publisher field in cite web (historical versus current)

Hey. I've just pondered something I've never actually thought about before. I'm adding web sources from sites/periodicals that are now owned by a different publisher than when the content was written—for example, Engadget was owned by Weblogs, now it's owned by Oath. Does it make the most sense to use the current publisher or the publisher of the content when it was written? Der Wohltemperierte Fuchs^(talk) 18:00, 28 March 2018 (UTC)

Personally I would always cite the original publisher. This is how it would be handled for print copies, so I would favor consistency. To look at it another way, what if the website ceased to exist? You would have to cite the original publisher in such cases. Betty Logan (talk) 18:26, 28 March 2018 (UTC)

Cite the material as it appeared on the access date (if provided) or in the archive URL (if provided; preferred IMO), per WP:SAYWHEREYOUGOTIT. --Izno (talk) 04:35, 29 March 2018 (UTC)

For periodicals, I tend to leave off the publisher and just make sure that the work/magazine/journal field is filled in. The publisher field is more important for books. For websites like Engadget I just fill in the work or website field.  Stepho  talk  10:43, 29 March 2018 (UTC)

Thanks for the responses. Der Wohltemperierte Fuchs^(talk) 15:34, 29 March 2018 (UTC)

To what extent does MOS:CAPS apply to reference titles?

I have asked this question at Wikipedia talk:Manual of Style/Titles, so I hope this doesn't come across as canvassing, but I didn't really know which project was best placed to answer the question. I am interested as to what extent MOS:TITLECAPS applies to citations and external links. To take an example, MOS:TITLECAPS instucts that the preposition "with" is not capitalised in titles of works, and on that basis "with" is not capitalised at Gone with the Wind (novel). However, many of the sources used in the article do capitalise the "with" such as The Making of Gone With The Wind, Part I. How does MOS apply to citations and external links? MOS:TITLECAPS states that "WP:Citing sources § Citation style permits the use of pre-defined, off-Wikipedia citation styles within Wikipedia, and some of these expect sentence case for certain titles" which implies we adopt the case appropriate to the citation style." I am trying to figure out how this applies to Wikipedia articles. I presume if a certain citation style uses a specific case and that is established in the article we continue using that style? If there isn't a specific citation style/case then would we apply title case apply to our citations (i.e. the title is thus cited as "The Making of Gone with the Wind, Part I"), or should it retain source purity and mimic the style of the source (i.e. "The Making of Gone With The Wind, Part I")? Betty Logan (talk) 15:13, 29 March 2018 (UTC)

It's not unusual to alter the capitalization of a title when citing the work. For example, some titles appear in all-caps in the original source, but are cited in mixed-case.

If the article uses the CS1 family of citation templates the documentation states

Use title case unless the cited source covers a scientific, legal or other technical topic and sentence case is the predominant style in journals on that topic. Use either title case or sentence case consistently throughout the article.

Of course, other citation styles are found in some Wikipedia articles. Jc3s5h (talk) 15:30, 29 March 2018 (UTC)

Wow, thanks, I have never even seen that help page before! I will bookmark it for future reference. Betty Logan (talk) 15:43, 29 March 2018 (UTC)

Of course, it is possible for articles to use CS1 templates without actually using the CS1 style. Many articles used the templates of the same name before the term "CS1" existed, for example. CS1 is one of many styles that can be achieved with the CS1 templates. — Carl (CBM · talk) 16:23, 29 March 2018 (UTC)

There are two obvious styles that can be achieved with the CS1 templates, CS1 and CS2. CS2 is achieved with the mode parameter. I would feel free to make an article that uses CS1 (without the mode parameter) conform to the current CS1 documentation, just as I would feel free to make an article that uses The Chicago Manual of Style conform to the 17th edition, even if the last edit to the article was made when the 14th edition was current. It is unrealistic to expect editors to figure out what outdated version of style documentation was used to create an article, or to shop at used book bookstores to obtain such documentation. Jc3s5h (talk) 16:31, 29 March 2018 (UTC)

Editors are free to choose any citation style for an article; it often happens that the styles within references may differ from the styles in the running text. MOS:CAPS is much more focused on running text. There is no reason that editors would be forced to change the citation style of an article because of it. — Carl (CBM · talk) 16:21, 29 March 2018 (UTC)

However, consistency is still key. A haphazard mix of sentence case/title case should be standardized to a specific style. Headbomb {t · c · p · b} 16:32, 29 March 2018 (UTC)

How would you cite an e-mail?

I just recieved an e-mail from the Food and Drug Administration’s Food and Cosmetic Information Center in response to my inquiry about Miraculin and how it was banned from importation as an additive. The e-mail contains some useful information that I believe would deem necessary to put on the Miraculin article. I want to add this info, but I have no idea how I should cite it. Should I cite it at all? Can you cite e-mails? Thanks. OblivionOfficial (talk) 19:04, 2 April 2018 (UTC)

The sources cited in Wikipedia must be published, that is, made available to the public in a reliable way. Emails are not published. You can't cite the email in Wikipedia. Jc3s5h (talk) 20:24, 2 April 2018 (UTC)

Should I cite it at all? - I would not add the information unsourced. If it exists nowhere on the vast World Wide Web, that is to me an indication that (1) it is false, or (2) the world deems it insignificant. Either way, it shouldn't be in a Wikipedia article. I think in this case I would reply asking where this information is found in published sources. ―Mandruss ☎ 22:50, 2 April 2018 (UTC)

Ask them if they can publish the email publicly if possible. Emir of Wikipedia (talk) 22:52, 2 April 2018 (UTC)

Agree with others that a private email is inappropriate to cite. There are email-based group discussion forums that have published their messages on a web page; then it effectively becomes a {{cite web}} use case. Even then it would be exceptional to cite it, because such content is usually not editorially managed, so it isn't reliable. --RL0919 (talk) 18:39, 3 April 2018 (UTC)

The issue is not whether an e-mail is private or not, but publication. Simply sending an e-mail (or printing it on paper) does not constitute publication. Unless an e-mail is published in some other work it is not citable.

But in this instance there is another aspect: Original research. You made an enquiry, and the FDA responded directly to you. That information comes from your personal communication, not what is already publicly available. That might be fine for a journalist, but keep in mind: Wikipedia is not a newspaper. As encyclopediasts we don't ferret out new information, we summarize from the existing stock of human knowledge. ♦ J. Johnson (JJ) (talk) 19:16, 3 April 2018 (UTC)

Paywall and subscription/registration url ettiquette

I have searched around some and do not find any information on my question. As an active participant in fighting link rot and citation errors within WikiProject: Oregon I have run into this with newspaper archives most commonly, but other referenced sources have been affected similarly. I would like to know what the community's position would be on the matter.

It has become my practice when a source is behind a paywall to set the url= to a preview of the source or a pre-defined specific search result giving the reader the choice to click for paywall yet verifying the source is real and obtainable.

• Example Multi-result pre-defined search by url with preview.
• Example - Multi-result pre-defined search by url list only.
• Example - Single result pre-defined search by url.

However, some archive sites does not have a simple and direct way link to a result or preview page, such as the Oregonian newspaper for articles from 1923 to 1987.

• Example Paywall

The only linkable pages that can be offered the reader are the empty search page for entering the search data themselves or a link direct to a paywall login screen with no other information.

• Example - Blank Search Page (if you search anything you can not link the results in any way I know how to do)

What I want to discuss is:

When a reasonable url of a source location can not be produced between search and paywall, what would the community prefer as a solution? (Thinking like there was such a thing as MOS:Cite Paywall... ect)

1) No url at all. Link publisher= if appropriate and treat as a standard citation.

2) A url to the blank search page of the publisher for locating the source. (I would ask to then consider adding a template something like {self search} to label as such. See {closed access}.)

3) A url direct to paywall screen of publisher, label {closed access} and or set subscription=yes . (I use both)

4) (Intentionally left for community use)

5) (see # 4)

I ask this to provide a guideline as to what the community would like done about it. There are many of these citations and each way of doing this has its pros and cons. I feel it bears some discussion for a consensus. If this has already been done elsewhere and I have just missed it in my searching... please advise. Darryl.P.Pike (talk) 20:28, 9 April 2018 (UTC)

In the case of a paywall/subscription there is parameters as shown at Template:Cite_web#Subscription_or_registration_required or the closed access label. With regards to searching we have a few options, there is either Template:Link note or |others= parameter which is documented to be for To record other contributors to the work, including illustrators. For the parameter value, write Illustrated by John Smith. but I have seen it be used to mention that you should search. Emir of Wikipedia (talk) 21:04, 9 April 2018 (UTC)

Links to search engine results violate WP:ELNO #9. I think it is much preferable, and much more robust against future changes of the search results, to link to the actual article you are citing, rather than to the search results where you found the article. If it requires a subscription then tag it as requiring a subscription, but don't try to censor the link. So to me only #3 (direct url) is acceptable among the choices you list. —David Eppstein (talk) 21:05, 9 April 2018 (UTC)

@David Eppstein: WP:ELPOINTS: WP:EL does not apply to citations. --Izno (talk) 23:40, 9 April 2018 (UTC)

Well, but WP:RS does, and certainly it does not permit search engine results to be used as reliable sources. I would argue that WP:EL does have some applicability to the choice of courtesy urls used for sources that are reliably published in some other form. —David Eppstein (talk) 23:42, 9 April 2018 (UTC)

Responding to both issues raised. I am not asking how label the link in the citation as a subscription. I am asking, if you had to pick one of the choices in the list, WHICH type of url is appropriate to use, a blank search for yourself to complete and find the document, a direct link to the article that hits you in the face with nothing more than a paywall to be informed, or no url at all? (adding an option is perfectly acceptable.) I would argue displaying a search result from an archive's search of itself to display an exact location of a referenced source material is not "Search Engine Results" in any fashion of the interpretation, but these are the things I wanted discussed. — Preceding unsigned comment added by Darryl.P.Pike (talk • contribs) 00:24, 10 April 2018 (UTC) Bah! Forgot to sign it.. Darryl.P.Pike (talk) 00:33, 10 April 2018 (UTC)

If I ever see you putting search engine result links into citations I will revert you. Is that clear enough? Put the link to the actual citation there if you read the source online (even if it's paywalled); leave out a link if the source is offline (a book you read on paper, say). If readers want to use search engines to find alternative copies they can do so themselves. But the point of a citation is to say where you got the information. You should not be getting the information in a citation from the summary that a search engine tells you about it, so that is not the link you should use to say where the information comes from. —David Eppstein (talk) 00:46, 10 April 2018 (UTC)

You pick # 1. Noted and thank you for the input to it. It is a wiki and you can make any changes to you like as I understand it. "Be bold", it says. "Its a community", it says. I posted here for guidance and discussion. Seems I am still looking for that, ...in case anyone else might be listening in. Darryl.P.Pike (talk) 01:16, 10 April 2018 (UTC)

No, I pick #3 for sources that you read online. You can use either #1 or #3 for sources that you read on paper but that can be found online. And you should go read WP:SAYWHEREYOUREADIT. —David Eppstein (talk) 01:28, 10 April 2018 (UTC)

*Hannah-Jones, Nikole (October 28, 2009). "County Library Moves to Curtail Thefts". Oregonian. Retrieved April 7, 2018. {{cite web}}: Unknown parameter |subscription= ignored (|url-access= suggested) (help)CS1 maint: multiple names: authors list (link) CS1 maint: numeric names: authors list (link)
Click that link then read the rest of this. Is that acceptable as a link destination within a citation to referenced a source, by your standards? If it is, it is a search result of the exact full date of the publication with the exact full worded title of the cited source as the search criteria within the url= of the citation to purposly generate a single precise result from THIS page. The search and the presented results are both by the publisher (or the archive) themselves upon their own database only to show what you have asked for. It is not provided by any "enigne" site. It is the exact steps a reader would take if going to look for the exact article independently from the same site. I read WP:RS and does not contain the word "search" or "engine" except in the links at the bottom on how to locate a reliable source. There is no mention of usage in which I inquire and had there been I would never have posted this to begin with. It was the first document I went to for an answer. The only thing I would mention is that what I am trying to do with my question is to be able to bring the reader as close to WP:SAYWHEREYOUREADIT as possible to a source behind a paywall and still present some information that the source is really there and in some way obtainable.

I feel in the cases of which something like the above is not possible, linking the default internal search of the archive itself (NOT a google search, ever), and marked as such giving the reader foreknowledge of the link type they were clicking and how much participation they will have to have, is a much greater service to the reader than leaving it blank to figure it out for themselves. Sending them to an impassable empty login screen with the source supposedly on the other side seems silly, abuse-able, and of no real use than saying it is accurate. I was not happy with it the first time I got one, which has lead over time to this thinking. I desire to tell the reader more about where the source can be found, what kind of storage the source is contained in, what amount of participation they will have when they get there to see it for themselves, all while trying to place them within as few clicks as possible of it and yet still show something of value in the effort. The internet has changed citations presenting the need to show what was available online and where. Now the internet and its ways of doing things has changed and I feel citations (on this site) will need to be adapted to expand and present that information to the reader. Darryl.P.Pike (talk) 03:04, 10 April 2018 (UTC)

Let's put it another way. How likely do you think it is that repeating that search a month from now, a year from now, or five years from now will actually produce the same result? —David Eppstein (talk) 04:30, 10 April 2018 (UTC)

With all due respect, I do not understand the perception you are trying to give me with that statement, nor how it applies to my original inquiry and/or my particular position on such. Again, I am attempting to provide a reader the most direct, fruitful, straight-forward, result of the question "How can I see this for myself as simply as possible and know something of what I am in for when I click this?". To even be in the article the source itself has already passed WP:RS (2nd party, archived elsewhere, publicly accepted, ect, et al,... ) or I would not be searching an archive wanting me to pay for the ability to see it to begin with. I would think something calling itself an archive that wasn't would have the internet coming down it like a ton of bricks and not something for me to worry about. The cases I am referring to 80% of it is newspapers published between 1922-ish and 1987-ish (public domain, copyright, its really SUCH a mess) and Federal Census data made searchable by genealogy sites such as Ancestry.com makes up most of the rest. Six decades of the nations newspapers and magazines can be viewed only behind paywalls of an increasing type and approach. My above example citation contains everything I can tell you about that link you clicked pertaining to that fact. The first response to this discussion demonstrates what the community does currently to solve the "You have to search for this yourself when you get there" problem using others= or another parameter incorrectly causing citation clutter, which is an MOS issue on is own indicating a need for attention of some kind. Darryl.P.Pike (talk) 08:25, 10 April 2018 (UTC)