Wikipedia talk:Citing sources/Archive 47

Archive 40

←

Archive 45

Archive 46

Archive 47

Archive 48

Archive 49

Archive 50

Bare urls and the use of reFill

At WP:ANI#Use of semi-automated tools there is a disagreement between me and User:SheriffIsInTown about their use of WP:REFILL, the semi-automated tool that formats bare urls into full citations. Apparently, they don't always find it convenient to check or fix the output of this tool, so if the url goes to a different page from the one that was initially used as a reference, they end up fetching bibliographic information from the wrong page. This typically happens when the url has been redirected to the website's home page (like here), or when it returns a "not found" error message. My opinion is that having bare urls, though not optimal, is better than formatting citations to the wrong pages, and that editors shouldn't be expanding references if they aren't willing to check they aren't doing it wrong. Their opinion is that expanding bare urls is always an improvement, and that if it turns out wrong someone else should come around and fix it. – Uanfala (talk) 01:46, 2 November 2018 (UTC)

That is a wrong interpretation of my stance. My stance is that most of the time I notice the problem in preview and make a note to go back to correct. For example when the title clearly says "Page not found" or it says something like "search" or "home" in it, it is easy to notice. If tool changes the url that can be noticeable but if url stays same and it is just redirecting at the backend and title is benign as well from which you cannot guess if it is redirecting at the backend then it obviously goes unnoticed. The benefit of filling the urls despite the title saying "Page not found" is that it is more noticeable for others to see and gives an opportunity for others to correct it if I do not have time right away to do that. Leaving them bare will not change the fact of actual url going to "Page not found" page but nobody will ever notice it or fix it so it would stay there for ages like that. We all edit voluntarily, if I somehow cannot go back to correct it due to real life issues or just simply not noticing the issue, anyone who notices the issue is also editing voluntarily, they can either correct it themselves or notify me on my talk. What is the big deal about doing it? Harassing someone again and again for their good faith contributions does not make sense. The only thing will happen out of this is that I will stop editing. Then, still someone will have to fill those urls or leave them bare, rotting and pointing to wrong pages. BTW, there are 247,478 articles with bare links in them! Sheriff | ☎ 911 | 02:45, 2 November 2018 (UTC)

I use Refill sometimes, and whenever I use reFill, I always preview the edits before actually making them, Often I abandon the edits by this tool if its not adding any improvement. The above problematic edits are a clear indication of indiscriminate and mindless use of the tool by this user SheriffIsInTown. Had it been a Bot it would have been blocked already by now. Another example of WP:CIR by the above user who simply failing to understand the negative consequences of his edits. --DBigXrayᗙ 10:12, 2 November 2018 (UTC)

As this matter is/was discussed already elsewhere, I'm just suggesting the use of Wikipedia:ProveIt (gadget that simplifies adding/editing refs). I use both. --Pelmeen10 (talk) 22:36, 2 November 2018 (UTC)

Citing a website which forbids deep links?

A question has come up at Draft talk:Michigan State Law Review of how to cite a service (specifically, Washington & Lee legal database) which does not allow deep linking to a specific search result. The citation in Draft:Michigan State Law Review just links to the search page, which I claim is not acceptable. Any suggestions on how to handle this? -- RoySmith (talk) 19:57, 8 November 2018 (UTC)

Looking at the discussion at Draft talk:Michigan State Law Review and Washington & Lee legal database), it seems by "allow" you mean that providing a URL to a particular database entry is not possible, rather than meaning the database provider asserts it would be a copyright violation to do so.

I believe it is perfectly appropriate to cite a database entry. The citation would include a brief description of how to search for the entry. This is no different than citing American Heritage Dictionary 2nd ed. s.v. aardvark. Jc3s5h (talk) 21:06, 8 November 2018 (UTC)

Yes, my apologies for being unclear. By "allow", I meant there's no mechanism to construct a URL which encodes the search terms. -- RoySmith (talk) 02:20, 9 November 2018 (UTC)

My 2 cents, if you have a reference that directly points towards the needed information, I would prefer that over such a database. However if the database you describe is needed and publicly searchable and preferably if a pointer is given how to locate the relevant information (as suggested by Jc3s5h), I see little problems with such a source. I can even see the analogy with sources available in hardcopy only (usually older sources that have (not yet) been transferred to an e-version) which require the reader to locate a library who holds the source to gain access. Arnoutf (talk) 07:03, 9 November 2018 (UTC)

I am always deeply skeptical of needing to cite this kind of source as it usually points to a lack of significant coverage in a reliable, independent source. --Izno (talk) 13:16, 9 November 2018 (UTC)

Significant coverage in reliable independent sources is only a concern in establishing the article covers a notable topic. There is no particular reason to avoid databases to support individual facts within an article, such as the geographical coordinates of a place described in an article. Jc3s5h (talk) 13:36, 9 November 2018 (UTC)

Not really--WP:WEIGHT is important for setting the contents of the article, which relies heavily on the same concepts. As I said, I am skeptical, and might care to remove such references and their content regarding topics I am interested in, but wouldn't do so elsewhere. --Izno (talk) 14:28, 9 November 2018 (UTC)

Visual Editor and Numerical reference names

This guideline, discussing naming of re-used references at Wikipedia:Citing_sources#Repeated_citations, says: "The text of the name can be almost anything‍—‌apart from being completely numeric."

Help:Footnotes#Footnotes:_using_a_source_more_than_once says "Names must not be purely numeric; they should have semantic value so that they can be more easily distinguished from each other by human editors."

But Visual Editor produces references with "names" with are numbers, starting with "0". That contravenes our long-standing guidelines on reference names. (We also have at least one editor who prefers to use Roman numbers i, ii, etc for their reference names, but that's a different story).

It's a mess. I don't know whether it has been raised anywhere already, but I've reported it as a bug in VE at https://phabricator.wikimedia.org/T208781. PamD 20:27, 5 November 2018 (UTC)

ve doesn't use purely numerical ref names but they are just about a close as can be gotten to purely numerical: <ref name=:0>; there is a colon there.

—Trappist the monk (talk) 20:51, 5 November 2018 (UTC)

Citoid gets around the limit on numeric names with the text :n; in this regard, the Citing sources section is not at issue, as is the first clause of Help:Footnotes.

Regarding "semantic values", we've tried to get the developers to add the ability to name the reference. That's phab:T52568, which is currently listed as being in the "Freezer" on the workboard of interest. --Izno (talk) 20:53, 5 November 2018 (UTC)

And I found a ticket for "make better auto names", which I've also listed in the ticket (and which is also "frozen"). Please feel free to merge as a duplicate to one or the other tickets. --Izno (talk) 20:56, 5 November 2018 (UTC)

Probably too much to ask that the "ref name" correspond with the CITEREF generated by the cs1 templates. ♦ J. Johnson (JJ) (talk) 23:59, 5 November 2018 (UTC)

SoWhy has again put in a request for Visual Editor to do better with named references at m:Community Wishlist Survey 2019/Citations#VisualEditor: Allow references to be named. Voting begins November 16. StarryGrandma (talk) 01:25, 14 November 2018 (UTC)

RfC on amending this guideline

There is a RfC on amending this guideline at Wikipedia:Village pump (policy)#RfC: Citation styles. StarryGrandma (talk) 16:40, 18 November 2018 (UTC)

Example with DOI link

IMHO the example in the section Links and ID numbers is a bad one. Instead of using http://dx.doi.org/10.1371/journal.pmed.0030496 as an ordinary URL, the DOI should rather be formatted as follows: doi:10.1371/journal.pmed.0030496. --Leyo 00:04, 1 November 2018 (UTC)

I replaced the example. --Leyo 13:40, 21 November 2018 (UTC)

Discussion: Citation bot removal of publisher and location in cite journal

I have begun an RFC at Help talk:CS1 regarding Citation bot's activity for cite journal publisher and location. Please provide input. --Izno (talk) 16:05, 27 November 2018 (UTC)

" Preventing and repairing dead links"

This section seems a little outdated. Good info to have for DIY editors but if a new editor asked me how to resuscitate dead links, I'd sooner point them to toollabs:iabot and say to check the checkbox when analyzing a single article. The manual advice, as written, would be secondary and methinks it would sit better in a subarticle than as a section in this one. (not watching, please {{ping}}) czar 17:13, 20 December 2018 (UTC)

@Czar: What is the cue in a page history that the bot has been run? --User:Ceyockey (talk to me) 20:16, 20 December 2018 (UTC)

Ceyockey, it leaves a summary like this. I tried iabot after seeing this. It ignored the dead link that I wanted to test it on, although it added archive links to 54 citations that, as far as I have checked, were still live. I'm not sure what its usefulness is. - Donald Albury 21:46, 20 December 2018 (UTC)

• Thanks, @Donald Albury:. I was hoping it would add a tag, and so it did: "IABotManagementConsole [1.1]" (actual tag name is "OAuth CID: 678"). Says it's made ~36,000 edits to date. --User:Ceyockey (talk to me) 15:41, 24 December 2018 (UTC)

Eh, but that dead link is an edge case—I think iabot ignores sciencedirect.com links because they only index the publication. You'd want to pull the full citation for that example.

iabot runs on all <ref>-enclosed links. So say an article hasn't been touched since 2010 and all of its links are potentially dead. The bot crawls the refs, determines which are dead and adds links to archives whenever available. (It can also preemptively add archive links, via the aforementioned checkbox.) Works in the vast majority of cases and given how this page is for beginners, I'd recommend that tool over manually figuring out archive services, at least for getting the job done the fastest. czar 07:57, 22 December 2018 (UTC)

I've been manually repairing dead links for years. I was hoping for something that would make that easier, especially when there are multiple dead links in an article. I'll give it another try. And then there are the cases where the Wayback Machine never archived a page, and I have to search for a suitable replacement for the source. It would be nice if there was a bot for that.:) - Donald Albury 15:19, 22 December 2018 (UTC)

Author prominent citations

I think the citation style guide needs to do more to distinguish between parenthetical referencing, footnotes, and author prominent citations, where the author is specifically mentioned in the text. Footnotes are the more commonly-used style in Wikipedia than parenthetical referencing for good reason. It's more than a preference, and the difference should be explained here. [Here] is an article on academic writing style. It refers to "information prominent citations" and "author prominent citations". In particular, we should discourage the use of inappropriate author prominent citations. They are for when "the focus is on the author as the source of some original idea or information". Such references are often deleted by Wikipedia editors because they appear to give undue weight to one author's point of view. And then an edit war starts. Acedemic papers use parenthetical citation much more often than Wikipedia, because academics write research papers with their peers as the audience, and the style is such that they are contributing to an ongoing dialog in their field. The author is more prominent when the dialog itself and the new contribution the paper is making in the dialog is important. In Wikipedia, the contribution to an academic discourse isn't usually the point. It's a reference on the topic based on currently accepted knowledge. I think it would be helpful to point out that parenthetical citations and author prominent citations are a more academic style for this reason and that footnotes are favored. Only in cases where the topic is discussed in the context of an evolving discourse is the parenthetical or author prominent citation more suited. To simply say it's a matter of preference misses an important distinction between Wikipedia and academic writing, and ignores that fact that footnote style is predominent in Wikipedia articles. Coastside (talk) 16:30, 23 January 2019 (UTC)

I think you are splitting hairs. I have come across plenty of reference and academic books that do references inline, as footnotes at the bottom of the page, as notes at the end of a chapter or as notes at the end of the book. Academic papers tend to use either inline or footnotes at the end of the paper. All of these are purely a stylistic choice of the publisher (ie a house style), although notes at the end of the book/paper tends to be a bit more common. Likewise, each article in WP chooses the style based on the original editor's stylistic choice.  Stepho  talk  03:50, 2 February 2019 (UTC)

User:Meteor sandwich yum/Tidy citations.js

Hi, while using the above script in this edit, in Michael Jackson article, I am continuously being reverted by user Synthwave.94 who thinks that its changing the citation style. However, all this script does is remove extra whitespaces from references and standardizes it. Doesn't it fall within WP:CITEVAR since it does not change any reference formatting but reduces whitespaces thereby reducing size? I think the issue is from the amount of whitespace present in some article which users think that maybe content is being removed. Can someone add their comments? —IB ^{[ Poke ]} 13:12, 21 November 2018 (UTC)

Removing white space from citations makes the line wrapping of the Wikimedia source editors, and desktop-based editors that editors may like for large articles, perform poorly. I think the basic goal of the script is misguided and the whole script should just be scrapped.

In addition I reviewed the article history since April 2017 and observed you only edited the article yesterday and today. I suggest you not try to impose your source formatting views upon articles that you have no long-term interest in. Jc3s5h (talk) 17:14, 21 November 2018 (UTC)

IB: I see an issue there, and it is not about "the amount of whitespace present" – which is an extremely trivial matter – but about formatting. Particularly, what has been called "pretty printing". This was a big issue in the programming community back in the 1960s, where one camp (largely managers) claimed that programs with source code that had been "prettily" formatted with a lot of extra whitespace were more reliable and easier to maintain. On the other side, many programmers scoffed at the idea, could even disprove, that adding or removing whitespace made any difference whatsoever in the code the compiler generated. (Readers of an analytical bent will notice that the two positions don't actually intersect.) The issue was settled when it was recognized that the essential factor was clarity: code suitably formatted with extra whitespace is easier to read, easier to comprehend, and makes errors more visible.

Same for WP citation templates: "|last1=Abrahanson|first1=Caroline|last2=Beaudinger|first2=Lionel|last3=Smithson|first3=Lawrence M." is not as easy to read as "|last1= Abrahanson |first1= Caroline |last2= Beaudinger |first2= Lionel |last3= Smithson |first3= Lawrence M."

Which is why I always include the whitespace, and would revert any removal.

The amount of whitespace in an article is a very trivial matter. But its use can be very significant. I agree with Jc3s5h that the basic goal of that script is misguided. If you don't understand why seemingly redundant whitespace is present it is best to not remove it. ♦ J. Johnson (JJ) (talk) 21:16, 21 November 2018 (UTC)

Concur with J. Johnson. Before removing whitespace one should consider why it is there. Automated removal does not consider whether the whitespace in a citation template makes it more readable. · · · Peter (Southwood) ^(talk): 20:11, 27 November 2018 (UTC)

Additional support for J. Johnson. Some editors are of the mistaken belief that removing whitespace somehow saves space. They are mistaken to worry about performance WP:DWAP since Wiki pages are compressed at various levels. Anyone concerned about the size of the article should look at Wikipedia:Article size instead of stripping whitespace. -- 109.78.238.236 (talk) 15:12, 12 January 2019 (UTC)

The 2 extremes are runningeverythingtogetherwithnospacesatall and p u t t i n g s p a c e s e v e r y w h e r e. Both are damn hard hard to read and should be avoided. Personally I like |url=https:xxx.com/xxx |title=Title goes here |first=Moe |last=Lester |year=2000 because each item is separated from other items but not lost in a sea of spaces around the = characters. As mentioned above, the WP servers don't care (in fact each revert uses more than the potential saving). But editors who have to read/modify what you wrote will care.  Stepho  talk  04:16, 2 February 2019 (UTC)

Wikilinks in references?

What's the correct style guideline for wikilinks in references? In Talk:Anna Maria Erdődy#Wikipedia is not a WP:RS, I complained about references that were just wikilinks, but it turns out I was simply confused. A reference such as is actually a reference to a book by Maynard Solomon, but looks (at least at first glance) like it's just a wikilink to our own article on that author. What's the right way to handle this? -- RoySmith (talk) 02:07, 2 February 2019 (UTC)

A better (but not mandatory) way is this.  Stepho  talk  03:45, 2 February 2019 (UTC)

At first glance, the article "Anna Maria Erdődy" does not use citation templates, so in accordance with WP:CITEVAR it wouldn't be appropriate to use citation templates for any new sources that might be added. But on closer examination, the existing sources do not seem to follow any consistent style, so an overhaul of the citations would be in order. If one were to look through the article history, one might find there was a consistent style at some point, but after that editors added sources using any old style. If so, the consistent style would be the one to use. Jc3s5h (talk) 12:19, 2 February 2019 (UTC)

True, we cannot force the use of the {{cite}} templates. But we can propose it on the talk page. If there is consensus to use the templates (or at least no objections) then we can change them all. If there is no consensus then your original method above works.  Stepho  talk  14:09, 2 February 2019 (UTC)

References

Multiple values for date attribute

In the article 1936 Baseball Hall of Fame balloting, how can the multiple date issue be resolved?Adithyak1997 (talk) 17:15, 12 February 2019 (UTC)

Is it one publication or two? If two, it should be two citations. If 1, the document in question will have been published on one of the days, not both. Pick one. --Izno (talk) 18:24, 12 February 2019 (UTC)

@Izno: According to the given info in that article, it's the same source in which the same data was posted on two different days.Adithyak1997 (talk) 18:47, 12 February 2019 (UTC)

So, cite the one that was referenced. --Izno (talk) 21:17, 12 February 2019 (UTC)

Is there a better way to keep the editing text clean?

I'm currently working on a new article where I cite the same handful of sources several times. What want to do is group all of the full-template references together (like in a references section) and just use named references in the body. The benefits of this are that it's much easier to see what I'm writing without all of the markup, and also it's much easier to later unify all of the reference parameters for the article since they're nicely grouped together.

I have done this and found that it works (with one caveat). Here's an example of what my method looks like in the editing window:

Here's me introducing the subject.<ref name="name1" /> Then I go on to talk about more things.<ref name="name2" />. And I continue to add clarification.<ref name="name1" />. And finally I'm done.<ref name="name2" /><ref name="name3" />

References Section:
<ref name="name1">lots of parameters here, with piping and everything</ref>
<ref name="name2">lots of parameters here, with piping and everything</ref>
<ref name="name3">lots of parameters here, with piping and everything</ref>
{{reflist}}

Bascially I don't want to see all the messy template markup in the body of the editing window text. The problem this creates is an extra set of footnote markers when I spell out the references. Is there any other better way to do this? I read through the article and couldn't find any discussion of this.Teresacurl (talk) 01:53, 19 February 2019 (UTC)

​You want to use {{reflist}}'s |refs= like this. — JJMC89 (T·C) 02:29, 19 February 2019 (UTC)

(Aka WP:LDR. --Izno (talk) 02:56, 19 February 2019 (UTC))

@Inzo: Oh I see that now in the "avoiding clutter" section - somehow I missed it on the first read-through (I saw the bit about paranthetical references and knew I didn't want that so skimmed the rest I guess). Thank you! Teresacurl (talk) 03:40, 19 February 2019 (UTC)

@Izno: (misspelled, sorry!) Teresacurl (talk) 03:42, 19 February 2019 (UTC)

The problem this creates is an extra set of footnote markers when I spell out the references. Well, you have to define the references somewhere. Pick your poison. --Izno (talk) 02:56, 19 February 2019 (UTC)

​I believe Teresacurl was referring to the [1] [2] [3] [4] [5] seen here. — JJMC89 (T·C) 03:02, 19 February 2019 (UTC)

Hah, of course. --Izno (talk) 03:10, 19 February 2019 (UTC)

@JJMC89: Yes, that's exactly it, thank you so much! Teresacurl (talk) 03:18, 19 February 2019 (UTC)

@Teresacurl: "Full-template references" is not a good term. The bibliographic construct you are referring to is a full citation (containing the full bibliographic details of a source), of which there should be exactly one per source. These can be created with or without templates; it is preferred that templates be used. The principal full-citation templates are {{citation}}, and the {{cite xxx}} family.

"References" is a very unuseful term because it has been used so loosely, and so ambiguously, that it induces more confusion (and argumentation) than it resolves. A common usage is to refer to the notes (a.k.a. footnotes, endnotes) created with the <ref>...</ref> tags. But many editors fail to distinguish the notes ("footnotes") from the content of the notes, such citations of all sorts.

The problem of how to "cite the same handful of sources several times" comes up frequently, often described as how to "reuse" the "references" (more precisely, the full citations). I recommend not using named-refs (the "<ref name=" construction). It is a cute trick for making a note (typically containing a full citation) appear in more than one place, but when used as the required in-line citation it does not identify where in the source the specific material is to be found, nor accommodate multiple sources or any comments. Instead, use short-cites in the form of author's last name and the year of publication (e.g., "Smith, 2018"). This can be done in several ways; I recommend the use of {{harv}} templates. If the full citations are done with {citation} templates, or {cite} templates with an added |ref=harv parameter, the {harv} templates will automagically link to the full citations. ♦ J. Johnson (JJ) (talk) 21:33, 19 February 2019 (UTC)

Fully concur with J. Johnson. However rather than <ref>{{harv|Smith|1945|p=5}}</ref> I prefer the rather more succinct {{sfn|Smith|1945|p=5}} Martin of Sheffield (talk) 22:42, 19 February 2019 (UTC)

The problem with {{sfn}} is that it wraps the short-cite in <ref> tags. It does give you the in-source specifier (page number), but that's all. I think there is another problem, but at the moment I can't think of what it is. ♦ J. Johnson (JJ) (talk) 00:34, 20 February 2019 (UTC)

I don't see the automatic tags as a problem, more of a blessing. If you need to make an textual reference then use {{harvtxt}}. The documentation is the same for both ({{Harvard citation documentation}}) and it shows the same parameters (p, pp, loc, ps). Section 9.4 explains what is happening. When I have a free hand (principally article creation) I put all my citations into a bibliography and populate references and footnotes exclusively by {{sfn}} and {{efn}}. See Subhas Chandra Bose as an example of this style. Martin of Sheffield (talk) 08:31, 20 February 2019 (UTC)

Which (incidentally) is a very nice looking article. And while I maintain that "references" is an extremely unuseful term whose use ought to be abjured, yet if it must be used I rather like the way you have split your "References" section into "Notes", "Citations", and "Sources". The latter, of course is your "bibliography" section, which I strongly recommend. But I find "Citations" not accurate, as the contents of "Sources" are also citations. What the "Citations" section contains are what I explicitly call short-cites (distinguished from full citations, and even shortened citations), but as that is more of a technical term I think it is not appropriate for a section title. I don't know what else it could be called; but it's not a problem for me as I combine "explanatory notes" and "citation notes" into a single "Notes" section. I think it is best not to split them as it is often useful to explain something about the citation. ♦ J. Johnson (JJ) (talk) 21:47, 20 February 2019 (UTC)

That's some good points which I'll take back to the article. I haven't written any of the article itself, I wasa asked for advice and help with an horrendous mismash of citations, quotes, repeated citations, non-alphabetic order and practically every other problem under the sun. Over the last couple of years I've kept an eye on it in a purely gnomish manner, I'm no expert on Indian history. You are quite correct to point out that the "citation" section is actually a "references" section, and that the sources/bibliography holds the actual citations. As soon as I've finished typing this I'm heading back there for a bit of revision! Regards, Martin of Sheffield (talk) 23:15, 20 February 2019 (UTC)

Not "references", please! (Don't want to set a bad example.) And "citations" should be clarified as "full citations". But other than that, sure. Good luck with the article. ♦ J. Johnson (JJ) (talk) 00:55, 21 February 2019 (UTC)

Bundling citations with semicolons?

Wikipedia:Citing sources#Bundling citations says three styles are available, the last being "paragraph" form. However I have seen in multiple places semicolons being used to string them together in one <ref> pair (for example see the ref in William of Ockham#Literary Ockhamism/nominalism). Should this be added as an available style, since it isn't technically a "paragraph" and doesn't match the example shown? —DIYeditor (talk) 15:59, 22 February 2019 (UTC)

The "paragraph" example matches that style, with the distinction that it uses commas instead of semicolons. The name is a bit confusing, but I guess what it's going for is that the references are written in one (or more) prose-like paragraphs. – Finnusertop (talk ⋅ contribs) 19:05, 22 February 2019 (UTC)

Stuff like that makes my eyes bleed; too much wall-of-text-like. Semicolons, commas, periods are all commonly used within citations so some other mechanism that distinctly separates the individual citations in the bundle from each other is worth considering. One might prefix the first with • (the dot followed by a space character) and the others with <br />• . The other thing that make my eyes bleed in that bundle is the lack of consistency in layout.

—Trappist the monk (talk) 19:48, 22 February 2019 (UTC)

Semi-colons are the standard way to separate cites in a bundle: Smith (2018), p. 1; Jones (2019), p. 2. If the citations are very long and a semi-colon would be confusing for that or some other reason, other styles can be used, but using semi-colons isn't wrong. SarahSV ^(talk) 20:00, 22 February 2019 (UTC)

I am pretty sure that I did not say that use of the semicolon as a separator character is wrong. As is illustrated by the minimal punctuation that your example uses, semicolons work quite well when bundling simple, short-form citations. My objection arises from the use of semicolons (or any other common punctuation character) when bundling long-form citations in the manner exemplified by Editor DIYeditor's example, because such long-form citations generally contain all sorts of punctuation characters (cs1|2 templates can and do use commas, periods, semicolons, colons). Using a separator character that is commonly used within individual long-form citations to separate one citation from another is just confusing for me as a reader and perhaps for other readers as well. We should not be making life harder for readers.

—Trappist the monk (talk) 20:22, 22 February 2019 (UTC)

If this is standard, would there be any objection to adding a 4th style of "semicolon list" to the examples? I am not too happy with the current "paragraph" example because the periods delimit individual parameters as well as each reference - would like to see some consistency there. I tend to agree with Trappist that this can be confusing (either with periods or semicolons) because they may appear otherwise in the individual references, and to my mind this should not happen with delimiters. —DIYeditor (talk) 20:38, 22 February 2019 (UTC)

I would have no objection if you were to add semi-colons. SarahSV ^(talk) 21:24, 22 February 2019 (UTC)

What are the arguments against use of templates, aside from ease of use?

There's an ANI thread right now concerning, in some part, CITEVAR, but an application of CITEVAR I've not encountered. I'm opening this here and trying to be vague because I have no stake in that thread and don't really want to get into that dispute.

I've seen lots of arguments in favor of using templates to format citations. The only arguments against use of templates that I've seen is that they're difficult to use/confusing. My impression is that the general consensus is that structured, formatted citations is just better, but that we don't force people to use templates because they're hard to use. If someone creates refs that don't use templates and someone else is willing to do the formatting to add templates, I cannot think of a reason why one would argue against that. What am I missing? — Rhododendrites ^talk \\ 04:09, 23 February 2019 (UTC)

Just playing devil's advocate. Say user X has been working on article Y, not using citation templates and is building a really fine, unbiased, complex article with well attributed citations to sources. (User X, by the way, is a 79-year old retired history professor who has written five books but can barely use email). Now, we all know that there's this aspirational malarkey that fine Wikipedia articles are written by lots of different people slowly over time accreting contributions until magically one day a fine, polished end product exists, but that in reality, but for topics of very wide interest, every fine article we see, e.g., all featured and good articles and so forth, are the product of one or a few editors taking up the reigns to tackle the hard work, organizational focus and deep dive into sources to turn out an excellent article. Meanwhile X, in fact, is planning to work on Y, an obscure topic, until it's gleaming. No one else has ever come along to do that work, and it's unlikely, in her absence that anyone will. The Plot Twist: one day X comes back to Y and finds that the sourcing scheme she had with great difficulty learned to place with ref tags had been turned into this unrecognizable, alien computer code that might as well be written in Sanskrit for all she cared. I think you get the drift of where I'm going with this. Or, let's just say that X immediately commits harakiri. The end.--Fuhghettaboutit (talk) 08:25, 23 February 2019 (UTC)

That sounds reasonable. While I don't think shielding someone from having to see template wikicode is a good reason to do anything, there's something to be said for the scenario for when X goes back to change one of those refs and finds she has to navigate the template. Frankly, X seems like the ideal candidate for using the Visual Editor, but that's a whole other thing. :) — Rhododendrites ^talk \\ 17:34, 23 February 2019 (UTC)

Explicit structure requires effort, while writing free text with implicit structure is effortless for those that have learned the style well. It is a significant difference in effort that the community has decided is not justified by the benefits. There are also a lot of people that care about such minutia as whether to separate parts of a citation with a period or comma (hence why some the citation templates support a parameter to switch between the two). While templates make it easier to switch between major styles, they are also rigid and cannot support every such quirk. But the bottom line is that CITEVAR isn't about which is "better": it reflects the community's sense that we can live with either just fine, and any possible relative benefits are outweighed by the downsides of forcing everyone into the same straightjacket. And to those who have a preference, having someone else come in and impose their preference feels incredibly disrespectful and arrogant.

That being said, give it ten more years and we'll be at 99% citation templates and 99% of a single universal citation style. Even the most outspoken proponents of manual citations are slowly coming around, and in volume of content created, manual citations are dropping. Which is good because, in my opinion, the actual critical problem we have is quality of citations, not the style or what syntax we use to edit them. Picking a number out of… the air… I'd estimate we're at the order of 50-80% of new citations being so flawed that they either now fail WP:V or soon will through linkrot and similar factors. --Xover (talk) 10:34, 23 February 2019 (UTC)

Actually "ease of use" is decisive argument. Yes we should have nicely formatted and structured, but they need to easy to use and standardised. The latter in particular means they should be part of the Wikimedia syntax. Templates (contrary to the Wikimedia syntax) usually don't work across other language WIkipedias, which makes it hassle for people working that across several languages when they write and maintain articles or content. It also forces new editors potentially to learn additional syntax, which may lead to them not correcting/improving references and associated (as in not touching what you don't fully understand).

As far as (automatic) maintenance goes yes templates offer signifcant advantages to "dumb" software with regard to link rot for instance. However the better solution in the long run might be smarter software. The same way as better structured or formatted references should be handled by an improved Wikisyntax rather than a template zoo on top of it.--Kmhkmh (talk) 10:23, 23 February 2019 (UTC)

Like most things involving computers, CITEVAR came out of the history of citation templates. In the past, they didn't work as well as they do today; one of the glaring problems was that you couldn't have too many citations on a page or the page woudn't display. For that reason, in addition to the reasons mentioned above, and other reasons that can be found in the talk page archives, CITEVAR was established. Several attempts have been made over the years to overthrow it with well-advertised RFCs, and those attempts failed to gain consensus. See the archives.

Even if CITEVAR were modified to prefer templates over manual citations, there would still be an issue over which templates, and which method of using them (parenthetical referencing, short citations, list-defined references, etc).

One clear-cut situation for using a manual citation, even in an article that otherwise uses templates, is a source that can't be represented with a citation template. The suite of citation templates only support various fairly-popular categories of sources; now and then there will be a reliable source that can't be described with a citation template. Certainly the lack of a suitable template is not an excuse to suppress a relevant reliable source. Jc3s5h (talk) 12:22, 23 February 2019 (UTC). Added missing word 16:29 UT.

The cross-language issue is an interesting point I hadn't considered, but seems easy to implement if not for this or that community's "stop trying to force us to use templates" (for better or worse). I feel like the advances in the Visual Editor go a long way towards standardization of templates, though, at least with enwiki. Using VE it's more work not to use a template for a citation, and new users love VE. — Rhododendrites ^talk \\ 17:40, 23 February 2019 (UTC)

With a new generation of editors that have grown up on click bait news and not academic publications... link rot is a bigger problem then ever.....anything we can do to help this should be implemented....and is why we have tools to combat this. --Moxy (talk) 15:06, 23 February 2019 (UTC)

I really don't see where that requires templates.--Kmhkmh (talk) 17:03, 23 February 2019 (UTC)

Just makes it a lot easier for bots to add archive urls. Boghog (talk) 17:33, 23 February 2019 (UTC)

I don't really see that either, since detecting of an url should be as easy as detecting a template.--Kmhkmh (talk) 18:47, 23 February 2019 (UTC)

Detection is easy. It is what you do after that. User:InternetArchiveBot has been replacing dead URLs with {{webarchive}}. It does a better job if the citation is already templated with a |access-date= parameter. That way, the archive link date can be made as close as possible to the access date. Also it is a bit strange to be inserting a {{webarchive}} template in the middle of a un-templated cite. If the originally citation is templated, then the insertion of a |archive-url= is cleaner and more accurate. Boghog (talk) 21:02, 23 February 2019 (UTC)

Yes, this is one of the big reasons to use templates. Anything structured makes maintenance easy. We defer to individual editors for their own citation preferences, but it's hard to depend on each and every one of them to update/fix issues down the road (and they shouldn't be expected to). A clever AWB user can fix a lot, but having things structured makes it so much easier. See also: meta:WikiCite. — Rhododendrites ^talk \\ 17:43, 23 February 2019 (UTC)

Rhododendrites, the citation templates are fiddly and take longer to write than a manual cite. Writing Jones, Paula (2019). Name of Book. London: Publisher, p. 1. couldn't be easier. But one typo in a template can be enough to make it not work. I recently spent ages staring at a red error message telling me that a short cite had no corresponding long cite, when I knew that it did, but for the life of me I couldn't see what I'd done wrong. It turned out I had written "Friendländer" in the short cite and "Friedländer" in the long one. As a result of that kind of thing, everything takes longer, and when you're writing a whole article, all those bits of time fighting with the templates adds up. It would be extremely unfair to expect someone who had written an article using manual cites to continue to maintain it after a passing editor had added templates. The original author would most likely abandon it, so you'd have structured citations and no further article development. SarahSV ^(talk) 18:19, 23 February 2019 (UTC)

I completely agree. Even more templates in such case have advantages for automated format maintenance, but they can have disadvantages for people doing the actual content maintenance and expansion.--Kmhkmh (talk) 18:53, 23 February 2019 (UTC)

Sarah: I mis-type a lot. Which was a major problem when I started programming, and is a continuing problem when I do citations. But knowing that, I also know to look for exactly that kind of problem when ever there's a problem with a short-cite. We could have a tool to check this kind of problem, but I haven't figured out yet how to do that. ♦ J. Johnson (JJ) (talk) 21:01, 23 February 2019 (UTC)

@J. Johnson: a tool would be wonderful. I often dread hitting preview after adding a lot of citation templates in case it's a sea of red. Maybe I've forgotten to add "ref=harv", or I've typed "hrv" or some other mispelling. Time I wanted to spend writing the article is devoted to "hunt the typo". SarahSV ^(talk) 21:21, 23 February 2019 (UTC)

The templates are fiddly, yes. But if someone else is willing to do that fiddling for you, then all you have to do is that basic formatting. That's where I got hung up. But your scenario seems similar to Fuhghettaboutit's above, and it makes a good point: if the original editor has cause to go back and mess with old citations, they could have difficulty doing so if someone has converted them to templates. It's not often that I get a citation right and then ever feel the need to modify it ever again, but my citations are generally pretty straightforward, and I don't often work with outlier and/or very old sources, so it's easier to decide that it's "right" and then forget about it. This scenario satisfies my curiosity, more or less.

My one remaining question/discussion point is about Visual Editor and the future, though. I work with an awful lot of new users IRL, and it's been a long time since I've had to help some fiddle with a citation template because VE just makes it so easy. Sometimes we'll go back and I'll help them to hunt down a particular parameter, but the functionality will only get better. At some point, the difficulty of using the template wikicode is no longer a valid excuse because there's a tool that does it for you. It's not quite good enough to make that argument right now, but I don't think we're far off. — Rhododendrites ^talk \\ 18:50, 23 February 2019 (UTC)

This discussion is related to the one above, related to references formatted as external links, only.

That said, I do not see the need to require templates for citations as long as 1) as they end up presented they are all equivalent in terms of what information, the order of that information, and the formatting style is 100% consistent, and 2) the style minimially includes what are considered "essential" elements for a reference to meet WP:V. If the editor that has the priority on citation format keeps to these, then there is no reason to force citations. We should strongly encourage newer editors towards cite templates because they reduce a lot of hassle, but they should never be required as long as consistency and bare WP:V infomrmation is thre. --Masem (t) 18:30, 23 February 2019 (UTC)

It is often said (in various forms) that templates are hard, and "difficult to use/confusing". This is not exactly true. They are structured, and as Xover said: "Explicit structure requires effort, while writing free text with implicit structure is effortless for those that have learned the style well." More precisely, templates require some discipline (not som apoximations that hoomans see as clos enuff), which is hard for people not used to it. And there is a bit of a learning curve. But once on top of the horse, I think it is easier to let the template handle the piddly details of formatting, checking, linking, and even categorizing. The "ease of use" argument is really just a passing matter of ease of learning. I would say there are no arguments against use of templates, only an argument to not coerce editors into using them.

@Jc3s5h: I don't recall having ever seen a source that could not be represented with a "citation" (presumably you mean "citation template"). Could you provide an example or two? ♦ J. Johnson (JJ) (talk) 21:06, 23 February 2019 (UTC)

Going back to the opening question at the top, there are of course 57 varieties of cite template, and many of them have all sorts of specific problems (the dreaded sfn for example). Some top general disadvantages include:

1. They frighten the hell out of new editors, especially the sort who would actually add content. Not just doing new ones, but working out wtf an editing mode screen is showing. Anyone who has done in-person editor training knows this. Of course some new editors take to them with great enthusiasm, but these are rarely content-adders. This by itself would be a clincher for never ever requiring them. The decline in editor numbers tracks the rise in citation templates pretty well.
2. All or most types handle bundled references very badly or not at all. Hence the appalling taxi-ranks of refs that disfigure so many articles, and make the editing view impenetrable.
3. They encourage web-ref only articles, to the great detriment of quality. Johnbod (talk) 21:21, 23 February 2019 (UTC)

I'm not sure why you refer to "the dreaded sfn". It is simple to use and less error prone than <ref>{{harv|author|year}}</ref>. As regards the multiplicity of {{cite XYZ ..., I agree that they are confusing. There is a simple solution though, just use {{citation}}. Fill in all details you know, ignore those you don't and it sorts everything out for you. Keeping the citations away from the footnotes/references means that they can be kept in alphabetical order, easy to find, easy to edit and unlikely to lead to duplication. Burying a citation in a random position amongst hundred or two others is not nice to editors or readers. And as for the chaos of working on named refs where you don't know which of the half-dozen holds the actual citation - Yuk!

A few years back someone pointed me to a simple way to sort out harv errors. Simply place the line

 importScript('User:Ucucha/HarvErrors.js');

in your common.js file and the system does the hard work for you. Martin of Sheffield (talk) 22:33, 23 February 2019 (UTC)

Oooh, looks like a tool--that I keep forgetting to try out. Thanks for the reminder, Martin.

A couple of points. I very much doubt that the decline in editor numbers is due to editors being intimidated by citation templates. "Citation" as a whole is a problem, but here we are discussing only the use of templates. Also, "bundled references" (which ought to be deprecated) are not a problem involving any kind of citation templates, it is entirely a problem of how editors try to use <ref>...</ref> tags. This pseudo-problem comes about solely because of confusion in the use of the term "references", and in the broader concept of "citation". ♦ J. Johnson (JJ) (talk) 22:58, 23 February 2019 (UTC)

If you think bundled references ought to be deprecated, I've pretty much lost interest in anything else you may have to say on references. On the contrary, correctly used they should be strongly encouraeged (but generally aren't). Johnbod (talk) 23:09, 23 February 2019 (UTC)

My point is that any problem with bundling "references" really has nothing to do with the topic here, which is about the use of templates. ♦ J. Johnson (JJ) (talk) 00:06, 24 February 2019 (UTC)

While that is helpful for people wanting to use that template, it is in the big picture however yet another argumrnt against templates as we need solutions that work for reasonable well for the default settings. I. e. anything requiring a special tool or script installed to work reasonably well is imho a no-go for a general approach.--Kmhkmh (talk) 23:02, 23 February 2019 (UTC)

Wrong. Use of citation templates generally (I am not aware of any exceptions) do NOT "requir[e] a special tool or script installed to work reasonably well"; they work quite "reasonably well" without any such aids. What Sarah commented on was how to find a problem where she had mis-typed; Martin pointed out a script that can aid the user in such a case. But such tools are not required. ♦ J. Johnson (JJ) (talk) 23:30, 25 February 2019 (UTC)

@Johnbod: (edit conflict) Have you tried the Visual Editor recently? I train or otherwise help some hundreds of newbies/year through various events/programs. For a long while the VE was painful -- something people were tempted to use but which ultimately wasn't up for the task and wound up creating as many problems as it solved. Now, there's scarcely any reason to talk about wikicode save for what's needed to use talk pages, and I would never go back. Newbies love the Visual Editor -- and why wouldn't they? It's much more similar to anything at all they've user before. Because of that, while I would agree that a new user who jumps into the wikicode editor is easily frustrated by, if not scared off by, the templates, that's just not something that they need to do anymore except for unusual circumstances. It's the structure of the templates that makes that possible. If a newbie goes into an article and tries to edit an untemplated citation with Visual Editor, it makes things a lot harder for them because the structure isn't there for VE to use. I was slow to get on the VE train, but while I almost never use it in my own editing, I'm squarely on the VE train when it comes to new users. TL;DR - Because the Visual Editor is so useful now, templates help, not hinder newbies.

I don't agree that they encourage web refs. The web itself encourages web refs, but using the Visual Editor it prompts the user for either a url or doi, isbn, pmid, or other identifier. Doesn't work for old books, and can't always pull all the information you need, but it starts with some fields populated and lets you search for other fields.

Where it's hard to argue is when it comes to trickier citation matters. Like bundled references, as you say, or complicated formats that combine things like short refs and notes.

That brings me to types of editors. The people I work with aren't just edit-a-thon sorts who add a few sentences and go on their way. Also people who jump in to make substantial edits, with some GAs coming out of the various programs/processes. I don't think it's fair to make a generalization that people who add content will prefer one format and people who do other stuff will like another. It may be the case that there is a correlation between people who do a lot of work on FAs presently and people who prefer to use untemplated formatting, and it may even be the case that FA-level work is going to demand more flexibility in citations that templates don't always work with. I don't know. I'm not arguing in favor of forcing the use of templates in this thread FWIW -- just trying to understand the arguments, because I realized I didn't know why it was controversial. — Rhododendrites ^talk \\ 23:01, 23 February 2019 (UTC)

Short answer, no. People still tell me it's more trouble than it's worth. Since I never template refs anyway (except for when I used to do medical articles, & the MEDRS tool I used has vanished, and now I wouldn't know how to do those) and I use quite a few "old books", it doesn't seem to offer much. Johnbod (talk) 23:09, 23 February 2019 (UTC)

Just to be clear, "old books" wasn't intended as hand-wavey/dismissive. Just a note that, as I had just mentioned that the VE citation tool asks for isbn, etc., the implication is that it's limited to the span of time in which books, etc. have been assigned such things. — Rhododendrites ^talk \\ 23:14, 23 February 2019 (UTC)

Sure, but from what you're saying, it's another nail in the coffin for the claimed "ease of use"! Johnbod (talk) 23:20, 23 February 2019 (UTC)

• Incorporating the reflist template with the section preview function would encourage fuller, manageable referencing. There is no argument beyond ease of use/time spent. ~ R.T.G 00:02, 26 February 2019 (UTC)

Duplicate sources with quotes

In the case that an article has named references which are duplicated elsewhere with an added quote (See Detransition for specifics), should the references be merged and how? The references are identical save for differences in the "Quote" section. Using guidelines from WP:DUPECITE and H:CITEMERGE, I personally would combine the citations and add a short explanatory comment on each quote that points to which in-text citation it refers to. So this:

1. ^Herzog, Katie (28 June 2017). "The Detransitioners: They Were Transgender, Until They Weren't". The Stranger.
2. ^Herzog, Katie (28 June 2017). "The Detransitioners: They Were Transgender, Until They Weren't". The Stranger. "A 50-year study out of Sweden found that only 2.2 percent of people who medically transitioned later experienced 'transition regret'. … There have, however, been almost a dozen studies looking at the rate of desistance among trans-identified kids [which] James Cantor summarized [as] '[V]ery few trans kids…transition by the time they are adults'. The exact rate of desistance varied by study, but overall they concluded that about 80%…identified as their sex at birth. … [T]he most recent study…found that two-thirds ultimately identified as the gender they were assigned at birth."
3. ^Herzog, Katie (28 June 2017). "The Detransitioners: They Were Transgender, Until They Weren't". The Stranger. "Cass recently started posting videos to YouTube, where there are a growing number of detransitioning confessionals."

Would become this:
1. a b c^Herzog, Katie (28 June 2017). "The Detransitioners: They Were Transgender, Until They Weren't". The Stranger. On frequency of detransition: "A 50-year study out of Sweden found that only 2.2 percent of people who medically transitioned later experienced 'transition regret'. … There have, however, been almost a dozen studies looking at the rate of desistance among trans-identified kids [which] James Cantor summarized [as] '[V]ery few trans kids…transition by the time they are adults'. The exact rate of desistance varied by study, but overall they concluded that about 80%…identified as their sex at birth. … [T]he most recent study…found that two-thirds ultimately identified as the gender they were assigned at birth." On the growing number of people who detransition: "Cass recently started posting videos to YouTube, where there are a growing number of detransitioning confessionals."
Well, that didn't show up right. Let me see if I can work this. Mooeena ● 💌 ● ✒️ ● ❓ 23:47, 5 March 2019 (UTC) Okay, I fudged it and pasted the text in because I guess Template:cite doesn't work on talk pages? Anyway, I hope the point comes across. Mooeena ● 💌 ● ✒️ ● ❓ 00:12, 6 March 2019 (UTC)

You have run into the basic flaw of "named-refs": even though a named-ref magically appears in many places, it is still just a single "ref", constrained to one-instance-serves-all, whether it "fits" or not; customization (such as you have in mind) is contra-indicated. What you want can be done, but it requires a better conceptual framework and more careful use of terms.

"Citation" (in regard of theory, practice, and discussion) is much easier with the following conceptual framework. First of all: "reference" is an ambiguous term, best avoided. What the <ref>...</ref> tags create are notes. Which often contain citations, but can contain other content (but not other "<ref> notes).

Second: all sources require exactly one "full citation", with all the bibliographic details (just as you learned in college, and even high-school). These are typically done using either the {{citation}} template, or one of the {{cite xxx}} family of templates. These are often seen in the notes (leading to the "ref[erence]/citation" confusion), but can be anywhere. I recommend a dedicated section – typically named "Sources" – with all the sources in a sorted list, but this is not required.

Given the forgoing, the challenge is how to fulfill the requirement for linking specific content to source's full citation. There are many ways of doing this; the most common is a short-cite, typically the author-date form (e.g.: "Smith 2001"). On Wikipedia these are most commonly (and quite handily) created using the {{harv}} family of templates. These can be in the text (as above), or in a note (like here Harv templates can also take a page number for the exact in-source location: Smith 2001, pp. 47, 413–1423.

Note that if you use {{cite xxx}} templates you have to add the |mode=cs2 parameter. I have some more tips for you I will put on Talk:Detransition. ♦ J. Johnson (JJ) (talk) 23:03, 6 March 2019 (UTC)

Notes

Sources

• Herzog, Katie (29 January 2019), "The Detransitioners: They Were Transgender, Until They Weren't", The Stranger, Seattle.
• Smith, Tomasina (2001), A Big Book.

Fixed the Herzog {{harvnb}} template.

Note that if you use {{cite xxx}} templates you have to add the |mode=cs2 parameter. This is not quite accurate. You can use |mode=cs2 as one way to enable the links between {{sfn}} / {{harv}}-family templates and cs1 templates, another is to use |ref=harv which preserves the cs1 styling of the {{cite xxx}} templates.

—Trappist the monk (talk) 23:25, 6 March 2019 (UTC)

That's another option. But for an introduction it is better to show a way getting something to work than than showing the details of the many ways it can work. ♦ J. Johnson (JJ) (talk) 19:54, 7 March 2019 (UTC)

Are "external link" -formatted citations valid citations?

This is coming up at a situation related to ANI/AN3, but there is a fair question.

Is a reference that is formated like [https://www.epa.gov/laws-regulations/summary-clean-water-act US Environmental Protection Agency] considered a valid citation format that would be subject to the concerns of WP:CITEVAR? I don't see a problem when this formats being added by new users or preliminary development of the article, but as development progresses, towards GA/FA quality, is this a format that should be considered standard and subject to WP:CITEVAR. Arguably, missing features like accessdate among others would seem to go against the valid citation formats WP:CITEVAR suggests. --Masem (t) 21:43, 22 February 2019 (UTC)

If all or most of the citations in an article consist of bare URLs, or otherwise fail to provide needed bibliographic data – such as the name of the source, the title of the article or web page consulted, the author (if known), the publication date (if known), and the page numbers (where relevant) – then that would not count as a "consistent citation style" and can be changed freely to insert such data. Seems to say that does not count as a consistent style? Or am I misunderstanding? —DIYeditor (talk) 21:54, 22 February 2019 (UTC)

Yes i thought that was clear from the discussion above. Bare urls do not represent a valid citation style.--Kmhkmh (talk) 23:15, 9 March 2019 (UTC)

I read it that way too, but I should clarify that what if it is the case that these external link citations are not "all or most", but still more than one or two? That could be read another way. --Masem (t) 21:57, 22 February 2019 (UTC)

I can think of two cases. In the first case, most of the citations follow a single style, and that style provides all or most of the bibliographic information called for in this guideline. But some citations are of the format Masem mentioned, providing little more than a URL. So editors should either improve the URL-only citations to full citations, and follow the style already established. The alternative would be to discuss on the talk page changing all the citations to some style that is better-suited to the article.

The second case I can think of is there is no discernible style; even the early versions of the article show a random scattering of different citation formats. In that case an editor could just go ahead and pick a style and convert all the citations. Jc3s5h (talk) 23:16, 22 February 2019 (UTC)

The essence of CITEVAR is "Don't edit-war over citation format!" by way of "Respect the original citation style if there is one.". Anything else is effectively just wikilawyering. In particular, the list of exceptions are meant to be common-sense: it's not a strict definition (bright-line rule) of what constitutes "consistent". It also applies to all types of citations, not just web citations, so difference between book citations and web citations can not be used to argue that the web citations are not consistent (the citation templates also have different formatting for books and websites): so long as all the web citations are internally consistent, the style is consistent for the purposes of CITEVAR. It is entirely valid to argue that all that's actually needed for a web citation is just the URL, and the title only because the URL isn't human readable (not because it needed in bibliographic terms or for WP:V).

So far for CITEVAR. Personally I would argue, emphatically, that all citations should contain as much bibliographic data as humanly possible. I do a lot of work cleaning up citations and have at one time or another found every piece of bibliographic data essential for tracking down a source. Including for web cites: URLs go dead and the ability to google for title or author, browse the publication's archives, etc. etc. is then essential to fix it. I also work in an area that is plagued by FRINGE pushing, and find stuff like publication name and publisher essential for assessing reliability of sources. But the point is that I would argue all that on the relevant article's talk page. If there is an apparent existing style (even if in my opinion flawed or sub-optimal), or if a BOLD attempt to improve it is challenged (that is, by the first revert; not keep pushing until, say, the fourth), the way forward is the talk page. And if there is a single recalitrant OWNer, you can ask for a third opinion, invite relevant WikiProjects, or even start an RfC (otherwise known as WP:DR).

If the citations are, in your opinion, so bad that CITEVAR shouldn't apply, then getting consensus for change through DR should be pretty easy. But for the case at issue here I wouldn't like to place any bets on the outcome: the citations are sufficiently consistent that I believe it likely the wider community would consider it an issue goverened by CITEVAR. Absent other problems I probably would not have bothered. --Xover (talk) 08:42, 23 February 2019 (UTC)

Addendum: Since I made the mistake of referencing the wider context of the question here (which I should not have done: that mess should stay confined to where it's currently playing out) I feel I have to be clear that my comment above about pursuing a change to the point of 4RR was in no way intended to acquit an editor reverting up to 4RR! 3RR is a bright-line rule and use of DR to resolve such issues applies equally to all parties. Please ignore that part of my response (stricken above). --Xover (talk) 09:09, 23 February 2019 (UTC)

As penance, please flagellate yourself with a wet noodle. Twice. :-) ♦ J. Johnson (JJ) (talk) 22:02, 23 February 2019 (UTC)

• I don't care enough about citation style to have an opinion here (I almost always let the software do all the formatting), but there is one situation where I would deliberately use bare urls – when citing some random web source that I know should ultimately be replaced with something better, even though the source isn't necessarily bad so as to warrant tagging with something like {{better source}}. The bare url then provides a nice contrast with the rest of the citations in the article (provided they're properly formatted) and so quietly flags it up for attention in future rewrites. I have to say though that's not as useful as it used to be, what with all these editors using automated tools like reFill() to expand such bare urls all over the place. – Uanfala (talk) 18:02, 23 February 2019 (UTC)
  • I should stress that bare URL or ELs are absolutely fine to add information (and particularly fine if to get at least new editors involved). It's when editors are trying to improve the quality of the article is when whether these types of reference styles are appropriate or not. --Masem (t) 18:25, 23 February 2019 (UTC)
• People were not told at the start here that these are cases where the important identifying information that a full template would add was already in the text being referenced. Thus "improvement" was essentially limited to the access date - not very helpful if the website has migrated its urls. Johnbod (talk) 21:27, 23 February 2019 (UTC)

Absolutely not. URLs, whether "bare" (e.g., "http:sample.com") or "labeled" (e.g., "[http:sample.com Sample]"), are not citations, in any sense, and certainly not any kind of "style" of citation. Which does not preclude their use in pointing the reader somewhere (typically as a "see also"). But a citation (more precisely, a full citation), as required in support of content, requires certain elements, including authorship (who is responsible for that material, either an author, or a publisher), date (possible timeliness, but also disambiguates other work by the same author(s)), and a title (to identify the work). (Other information for identifiying, locating, and disambiguating are also strongly recommended.)

This is not matter of "allowed per CITEVAR", as URLs simply fail as citations.

And absolutely not fine, not even provisionally. First and foremost, we should know who is responsible for the material, and if an editor can't identify that then the material really should not cited. As I said above, there is no problem pointing the reader some where. But we should not incorporate, or rely on, material whose authorship cannot determined. If an editor has not studied the source enough to determine the basic bibliographic details, then he (or she) really has no basis for determining that adding the source will "improve the quality of the article." ♦ J. Johnson (JJ) (talk) 21:59, 23 February 2019 (UTC)

Ok, you seem to come from a hard science "cite lots of abstracts of papers and cite them once" sort of context, and perhaps have difficulty understanding the whole world of knowledge isn't like that. The citations that gave rise to all this were mostly to museum pages on paintings in their collection, merely referencing the existence, ownership and date of the work, and of course giving an image if the article didn't already have one. These are entirely RS, but never normally carry personal authorship info. The institution name was in the citation, and the author & title in the text being referenced. Johnbod (talk) 23:28, 23 February 2019 (UTC)

I do come from a hard science, but I don't cite the abstract of a paper in lieu of reading the whole paper (or at least the relevant parts), and I often cite a source more than once. I have no idea what you mean by "whole world knowledge", but I am familiar with citation practice even in non-sciences.

Please note that I did not (and do not) say that personal authorship info is required. I referred to "who is responsible for that material, either an author, or a publisher". In the case of a museum catalog published by the museum, the museum is considered responsible for the work, and it is not necessary to name a particular person. Same situation applies to many newspaper articles. If you bundle this additional information with the url then, presumably, you do have a citation. On whether this requires a template, see the following discussion. ♦ J. Johnson (JJ) (talk) 23:54, 23 February 2019 (UTC)

Actually museum catalogues almost always do have authors, who should certainly be named (see my user page), but not object webpages like these. Since, to quote myself, "The institution name was in the citation, and the author & title in the text being referenced", the citations seem to be ok, per you. Johnbod (talk) 18:19, 24 February 2019 (UTC)

There's always an author, of some sort; it's just a matter of identifying who (or what). Please note that I am not saying that your "citations are okay" (( haven't even seen them). What I am saying is that a bare URL (lacking certain essential information) is not a citation. With additional information you presumably have a citation, but that does not mean it is a good citation. ♦ J. Johnson (JJ) (talk) 20:06, 24 February 2019 (UTC)

Is this dispute now about valid source or valid citation (format)? And if it is about both, it should be split in two separate threads.--Kmhkmh (talk) 20:58, 24 February 2019 (UTC)

I think there are three issues

1. Is a source valid?
2. If an article developed early on with citations mostly like the one in the original post, [https://www.epa.gov/laws-regulations/summary-clean-water-act US Environmental Protection Agency], and most of the citations are still like that, does the article have a consistent citation style, so that CITEVAR applies?
3. If a citation looks like [https://www.epa.gov/laws-regulations/summary-clean-water-act US Environmental Protection Agency] is it a valid citation?

I think this thread is not about question 1.

I think the thread does involve question 2 and the answer to question 2 is no, CITEVAR does not apply to such an article.

I think the thread does involve question 3, and it is a valid citation as long as the link still works, or can be used to identify a successor website or archive, but once the linkrot is so severe that the content of the source can't be found, the claim in the article becomes unsourced and subject to removal.

One clarification: a citation consists of all the information in the article that identifies the source, not just what is between the <ref></ref> tags. So if the author and publisher are mentioned in the running text and other necessary information in the footnote, it is a valid and correct citation (provided that's the style being used throughout the article). Several printed style manuals call for not repeating anything in the footnote that is mentioned in the running text. Jc3s5h (talk) 23:34, 24 February 2019 (UTC)

Well the point of citation is to (clearly) identity a source, an url or ISBN + page information already deliver that. In that sense i'd consider valid but incomplete ciations. Complete citations tend to have redundant information as far as identification of the source is concerned.--Kmhkmh (talk) 06:28, 25 February 2019 (UTC)

Whether the essential information about a source should be in a compact unit is arguable, but that is not the issue presented here. The question presented here is whether "external links" (URLs in the form of [http:sample.com Sample]") are "valid citations". Not "if they come with additional information", just the labeled URL by itself. The argument that "it's okay, because the rest of essential information in in the text" does not make the URL a "valid citation". Even where the additional information is included, the URL is, at best, part of the implied citation, and not a "valid citation" itself. This is a crucial distinction, because many inexperienced editors (and apparently a few experienced editors, such as Kmhkmh) seem to think that merely supplying a link (whether a URL, DOI, ISBN, JSTOR, PMID, ZBL, etc.) that points to the bibliographic information is as good as the information itself. I beg to differ: that information needs to be in the article itself, as a full citation.

The problem I frequently see is where a URL ("bare" or "labeled") is used as a citation, but without any other information. Such usage must be clearly distinguished from cases where a URL (or other link) is part of citation. ♦ J. Johnson (JJ) (talk) 23:09, 25 February 2019 (UTC)

Nobody argued there shouldn't be a full citation or the citation information should be as compact as possible. I said such an incomplete citation (like an url) is nevertheless enough to identify and locate the source (which is the primary goal of a citation albeit not the only one).

The question here is really in which context and with which intended consequence this "invalid citation" arises. If the point is to recommend editors to use full citations including, author, date, publisher rather than just an url, I don't think anybody will disagree with that.

If "invalid citation" however arises as an argument to suggest the some content is not sufficiently sourced and needs to be removed including its "invalid citation", then I'd call that formalistic nonsense.--Kmhkmh (talk) 04:16, 26 February 2019 (UTC)

I want to stress the issue that I raised has nothing to do with whether a URL-only or URL-with-label is reliable or not. This should be strictly seen from a MOS level - how to format citations as the article is being improved towards GA or FA. If a new editor only knows how to add a citation by a URL-only but still adding useful/appropriate information, great! But we do need to deal with that citation as the article improves. --Masem (t) 04:22, 26 February 2019 (UTC)

Ok but now i'm a bit at a loss. here. Do we have people arguing extending an url into a full citation is not desired or should not be done due to keeping the existing citation style? In that case sure I'd agree that it is "invalid citation" and that it should be completed/improved and that this not to be understood as an unwanted style change. However this does not necessarily imply that that this completion needs to be done with a specific template or a template at all.--Kmhkmh (talk) 04:56, 26 February 2019 (UTC)

There are two kinds of "invalid" here. To say that bare URLs are "invalid citations" is to allow that they can be considered as citations. In that form the question is whether a URL, considered as a "citation", is valid. I say that the fundamental and antecedent issue is whether it is valid to consider URLs as "citations". I quote the latter term because (unfortunately) it seems to be getting used as a casual term for "pointing to some information some where", whereas we should be using it more precisely. In that regard, URLs, whether labeled or not, are absolutely not citations, and (mindful of the Confucian "rectification of names") it is invalid to call them citations. Which should resolve the issue presented.

What has been raised here is an underlying concern (per Kmhkmh) about the "context and ... consequence" of lone URLs. (Perhaps "citation fragments", but not "invalid citations".) That is, how should they be handled?

The ideal handling of any citation fragment – whether a URL, or an author's name, a DOI, ISBN, etc. — is to build it into a complete citation. (Understood to be either a full citation, or a short-cite that links to a full citation.) While I would not shoot on sight any material supported only with a citation fragment (assuming that there are no problems with the source, nor WP:BLP issues), yet it is a reasonable interpretation of WP:V that full citations are required. (And incomplete citation is not a "style" that is protected by CITEVAR.) I do not agree that "bare URL or ELs are absolutely fine to add information" because they are not necessarily an improvement. That an editor doesn't extract and supply the minimal essential elements of a full citation suggests a lack of comprehension of the source material, and its addition is quite likely a disimprovement. If an editor finds a possibly useful source but cannot assess its actual value, the proper action is not to add it to an article in the hope it might be an improvement, but to post it to the Talk page for somone else to follow up. ♦ J. Johnson (JJ) (talk) 03:40, 27 February 2019 (UTC)

Maybe the better way to think of this and reflect that in CITEVAR is that 1) the url+single label citation is not a complete citation as it lacks key WP:V components and with that 2) CITEVAR should prevent one from reformatting the set of complete citations from one format to another, but it is acceptable to bring incomplete citations to the article's standard complete format. --Masem (t) 04:00, 27 February 2019 (UTC)

Well I gotta disagree again and I really don't care about semantics or naming conventions. I. e. doesn't really matter whether it is in an "invalid citation", a "incomplete citation", a "citation fragment" or whatever. The most important aspects for writing articles or content on WP are whether you have reliable source and whether you extract content from it appropriately. Whether you extract the citation information correctly or fully is secondary and comparatively minor issue, provided you still gave enough information to identify the source. The primary article improvement comes from extracting (and verifying) content not citations.

So if someone adds a bare url pointing to a reliable source to article content that was unsourced so far then that is an improvement.

Now while adding a bare url as above is an improvement, it is clearly not the best way to do it and warrants further improvement, i.e. turning the url into a full citation. So if you come across (regular) editor who are routinely adding bare urls only, ask them politely to use full citations and direct them to recommended formats and templates.

We shouldn't create the impression that we do not want people to source, expand and correct our articles if they don't know how to use "valid citations" (yet). That would be the wrong approach to go about it.--Kmhkmh (talk) 11:01, 27 February 2019 (UTC)

Masem: are you suggesting that "URLs only" is a citation "style" that, once established in an article, subsequent citation must conform to? Can the first and/or major contributor to an article set a "standard" of incompleteness that, per CITEVAR, other editors are barred from improving?

It seems to me that you may be troubled by some particular instances (likely involving CITEVAR), but I can't address what you have not presented. My general argument is that lone URLs are not citations, are not a citation "style" covered by CITEVAR, and carry no weight in determining what an article's citation style is. If an article's complete citations establish a style, then any "completion" of the citation fragments should be in the established style. (Just as Jc3s5h said at the top.) If not, then (again as Jc3s5h said) who ever improves them gets to choose the "style".

KmhKmh: you should care about semantics because the lack of clear semantics is a major cause of extended, tedious, repetitive discussions that hardly go anywhere.

I disagree with your contention that extracting content from reliable sources is the most important aspect for writing articles. It is essential, for sure, but I say that even more important than that is our foremost principle of WP:Verification. It does not matter how true, or accurate, important, or interesting something is, if it cannot be verified. And that requires (as a first step) citation, which in turn requires certain essential information (such as authorship). The essence of Wikipedia is not content – the Web is full of content, and the worse for it – but verifiable content. We don't require people to use templates, and we don't expect absolutely complete bibliographic data (that's what bots do, right?). But people that can't figure out the basics of author, date, and title quite likely have no grasp of the quality of the material. Some competency is required. If someone really has found a super excellent source, but lacks the competency, or perhaps only the time or interest, to minimally identify that source, then the best course is (as I have already said) to mention it on the Talk page. ♦ J. Johnson (JJ) (talk) 19:35, 28 February 2019 (UTC)

No, I would strongly argue that URL-only is not a style the article should conform to under CITEVAR, because it is not a complete citation, and this is a concept I am suggestion we add to CITEVAR to still allow for any reasonable citation style as long as it is complete to avoid similar problems. We should define what a complete citation is for various works, so for example, a web-based only resource, it should minimally contain the URL, page title, the website it is on, and the accessdate it was used to be considered a complete citation. (there may be more). Any format that contains minimally date should be considered a "complete" citation, and if the article's original style or most dominate style is that, then that's the style it should use. So clearly a URL-only citation is not "complete", so any attempt to standardize on the URL-only style would not be appropriate per CITEVAR, and more specifically users should be free to convert those to the actual complete style used by the page per CITEVAR as well. Same with labeled URLs. --Masem (t) 23:07, 28 February 2019 (UTC)

Masem: So I would agree that: (1) a valid citation requires certain minimal information, including identification of authorship (personal or organizational), effective date, and a title; (2) lone URLs, with or without a label, are not valid citations, and (3) cannot be considered a valid citation "style" per CITEVAR; and (4) augmentation of lone URLs in an article should be in accordance to any established style per CITEVAR. There a couple of minor points we could chew on, but I think these are the key points. Are we in agreement on these points? ♦ J. Johnson (JJ) (talk) 21:58, 2 March 2019 (UTC)

Yes, WP:Verification is important, it just doesn't require what you claim. Verification just requires the identification and location of a reliable source for which a bare url (or citation fragment) is already sufficient in most cases. What the full citation gives you here however is a first assessment of a source without actually looking it up or reading it. You probably can argue that on some level that makes a superficial "verification" easier, but from another aspect it makes it even worse because it is just that superficial. In fact it is one of greatest weaknesses of WP that all to often article get copy edited and get the right citation format, while barely anybody bothers to carefully read the content and compared it towards the given sources.--Kmhkmh (talk) 11:40, 1 March 2019 (UTC)

Kmhkmh: I agree that all too often citations get copied without examination of the source (even in the professional literature). But I disagree that requiring full citations "makes it even worse". I think you actually have the argument upside down: there is no way that requiring full citations makes the copying of material you describe any easier, or verification any harder. Perhaps what you mean is that a properly done citation looks more plausible, and therefore less likely to invite verification? Which is, to a large extent, true: I definitely mistrust articles and editors that don't bother with proper citations. But a superficially plausible citation does not evade verification, and may even trigger it. E.g., in a topic with which I am familiar just seeing which key sources are used gives me some idea of how well topic is covered, and any irregularities, or significant deficiencies, will prompt a closer look. Having to run down a bunch of URLs to identify the sources deters such assessment, and where editors are too lazy to provide the bibliographic details we should just slap their URLs with "full citation needed" and move on.

To your particular point that a bare URL is sufficient: no. All a URL does is point somewhere, and even for permanent sources the URLs often change. There is absolutely no advantage in "URL-only citations" (except for editors that can't or won't do a full citation), and all citation practice says to include the fullest bibliographic details in the citation. (See any manual of citation style.) ♦ J. Johnson (JJ) (talk) 22:06, 2 March 2019 (UTC)

I'm not arguing full citation make it worse and I prefer full citations. However I do believe (contrary to you?), that adding or correcting content with an identifiable reliable source is usually an improvement (no matter whether that reliable source has full citation or is just a fragment sufficient for identification/locating). Our priorities should be (in order):

• 1.) correct (verifiable) content 2.) (identifiable) reliable source 3.)valid citation

and not

• 1.) valid citation 2.) (identifiable) reliable source 3.) correct (verifiable) content

To some degree I see citation fragment like spelling or grammar mistakes in a content extension or correction. They are to be fixed but it is more important to have our content correct than our spelling.--Kmhkmh (talk) 11:33, 3 March 2019 (UTC)

You did, indeed, argue that full citation leads to "superficial" verification, which "makes it even worse". If that argument is no longer operative perhaps you might strike it?

You trivialize citations when you regard their absence as a mere spelling mistake. Correcting a misspelling in an author's name or a title is usually trivial, given the nearly correct form in the first place. But given no author, no title, no identifiers whatsoever except the very slender reed of a possibly broken URL, no, it is not so trivial to determine the "correct spelling" of something you don't have. You might as well say "I can't be troubled to write down the title of the book I used or who wrote it, but here is its call number at the local library, and it's sure a pissaroo that barely anybody bothers to verify this stuff."

What you seem to not understand is that URLs do NOT identify sources; they only point somewhere. Also: it is not a matter of priorities. Per WP:V: "All material ... must be verifiable", and verifiability requires citation. Citation is not a "priority", it is a requirement, essential for determining (among other points) whether a sourde is reliable. ♦ J. Johnson (JJ) (talk) 23:53, 4 March 2019 (UTC)

Yes it is matter of priorities and they are as I stated above! The verifiability isn't affected by that at all, strictly speaking verifiability might not even require a given source nevermind a full citation depending on the context. It requires that a source to verify the content exists. I see absolute no value is discussing whether an url does "identify" a source or not as long as it does in practical terms, that is the url is still functional or available on archive.org. The proper proper action in such cases to simply expand the url into full citations and by all means ask the editors to use full citations rather than bare urls in the future. Simply deleting content on the base of just being sourced by a bare url is not a proper action.--Kmhkmh (talk) 23:10, 9 March 2019 (UTC)

Where was anything said about deleting content? That's a "citation needed" matter. ♦ J. Johnson (JJ) (talk) 00:14, 10 March 2019 (UTC)

Break, and leading into "content is alway good"

What I want to stress is that we should not discourage editors from adding information because they don't know how to format a ref and we chastize them for a bare-url or labeled-url style ref. As long as we can follow that URL to validate, the ref can be improved and the content addition should be welcome. That's a different issue than CITEVAR, where now we need to make sure that for quality purposes those refs are complete, and formatted in a consistent style. But CITEVAR should only be evoked on the purposes of improving the quality of the article, not as a stick to poke novice editors trying to help. --Masem (t) 01:07, 5 March 2019 (UTC)

This notion that adding content is always good would, by itself, be off-topic here, except for the argument (sometimes only implied) that anything that impedes addition of content – such as a requirement for full citations — is therefore bad.

That is a poorly considered idea whose elevation into a quasi-principle is unwarranted, and harms the encyclopedia. Sure, we are all pleased to see WP grow, taking that as some kind of measure of success. But growth merely for the sake of growth itself is never useful, and being simply the most popular container of interesting stuff on the Internet is rather pointless. Wikipedia's growth is a readily graspable factoid, but less readily comprehended is that quantity of articles is rather worthless without some leavening of quality. If mere volume of alleged "information" is the goal, then the World Wide Web suffices, and Wikipedia is not necessary.

The argument that new editors should exempted from providing a citation because citation templates are too hard to use is entirely bogus, because there is no requirement to use templates. If someone can copy in text, they should be able to copy in other details such as the author's name. If they can't get it template perfect, no problem, there are plenty of editors that will do that for them. ♦ J. Johnson (JJ) (talk) 22:41, 5 March 2019 (UTC)

But at the end of the day, that is all "busy work". There are wikignomes including bots that can expand a bare url into something close to a more proper citation.

This concept is coming up again at a current WP:AE, and I think we need to be very clear between the concept of a "valid" citation and a "complete" citation, both which are getting past the CITEVAR aspects. And for purposes here ,I'm focusing only on web-based sources.

• A valid citation for a web based source should point to the page that has that information directly on it that supports the claims in the Wikitext. Pointing to epa.gov top level as a source for information on the Safe Drinking Water act is not "valid", while a link to an online summary of the act at epa.gov is fine, even if left as a bare URL. This is sufficient for any other editor to come through and expand out the citation; should the source change between its addition and the citation expansion, that might require applying the accessdate when the URL was added, finding an archive, or removing the material altogether if another source can't be found.
• A complete citation is one either templated or not that includes the title, authorship, publisher, date, and, importantly, accessdate for the web-based work. This is necessary for article quality so that if a user was reading the article off-line or in print they can still understand the citation and look it up. CITEVAR (eg retaining formats) should only apply to citations in completed form, and bring otherwise valid citations not in a complete citation form to spec. --Masem (t) 16:45, 8 March 2019 (UTC)

Let's not say "complete citation", as the "citation" part is ambiguous. We are really talking about full citations. But even "complete" is a little shfity, as that could be applied to a citation having all possible bibliographic data. I think a better formulation would be that the minimum requirement is to provide title, date, and authorship. Other information is desireable, and in some cases may also be required.

"Publisher" isn't always useful (let alone necessary), except in cases such as newspapers and organizations where the publisher takes on the responsibility for the work, and the actual author(s) might not be identified. That is why I prefer to say "authorship", which is a bit more expansive as a concept, which can include the publisher.

As before, I can't address your concern with CITEVAR because I am not certain I understand what these "CITEVAR aspects" are. Certainly no one should argue that CITEVAR protects "incompleteness" as a "style", and prohibits expansion. But that goes beyond the discussion of URLS as "citations".

I think you don't understand my objection that omission of certain key data is not a minor, bot-fixable detail. It indicates a likely editor failure to assess the material being added that belies any assumption of goodness. Also: "valid citation" is itself invalid, as URLs (bare or labeled) are not citations (not by any definition I have seen), a point I thought I had adequately addressed at the top of this discussion. But I begin to suspect that all my hot air can't melt your frozen fixation that URLs are acceptable as citation. ♦ J. Johnson (JJ) (talk) 21:58, 9 March 2019 (UTC)

Maybe not complete, but "sufficient" information. The other points are good to think about.

Core here is bascally, we don't want CITEVAR or other parts of this same page to be used to discourage editors that want to contribute and can provide URL links to the right pages, but have zero interest in learning the reference system, to require them to package those URLs into sufficient citations or ones that match the page --- while at the same time recognize that in time, bare URL or labeled URL "citations" do need to be fixed. The fixing is a gnomish activity , so there's no reason to force new editors to deal with that. But we do want to make it clear to experienced editors that bare-url/labeled-URL "citations" are definitely not sufficient for quality checks that use CITEVAR (ala GA/FA). --Masem (t) 23:02, 9 March 2019 (UTC)

I agree--Kmhkmh (talk) 23:20, 9 March 2019 (UTC)

How is CITEVAR being "used to discourage editors"? Please provide an example of CITEVAR being used this way. Your "require them to package those URLs into sufficient citations" is a red herring, because (AS I HAVE SAID BEFORE) the requirement for citation does NOT require "packaging". (Which is, after all, a "gnomish activity", right?) CITEVAR requires consistency in citations, but that is different from the WP:V requirements. It seems to me you want two kinds of "citation": one sufficient for newbies, and a different kind for the rest of us. What both of you seem to have missed is that if URLs are deemed "citations", then CITEVAR can be used to beat new editors about the ears. Which I believe is your concern. On the other hand, if URLs are not citations, then CITEVAR does not apply. If you want to beat up on a newbie you have to use {cn} or {fcn}; CITEVAR is no longer available for that purpose. ♦ J. Johnson (JJ) (talk) 00:21, 10 March 2019 (UTC)

So this is still going on! Probably time to restate the reminder that the original query related mostly to citations where the "missing" information was in the text being referenced - a point all commenters seem to have strenuously avoided considering, but which is actually rather common. Johnbod (talk) 01:51, 10 March 2019 (UTC)

As you have said yourself (21:27, 23 Feb), that is more of what "People were not told at the start". As a point of fact, Masem's original query did not mention any "missing" information other than accessdates, nor to anything being "in the text". I think what some of us are strenously trying to avoid is going off on tangents. We could avoid a lot of traipsing over the discursive landscape if queries (or even points for debate) were carefully qualified at the outset to trim potential tangents. ♦ J. Johnson (JJ) (talk) 22:30, 10 March 2019 (UTC)

cs1|2 auto date formatting

At Help talk:Citation Style 1#auto date formatting it is proposed that {{use mdy dates}} and {{use dmy dates}} be read by Module:Citation/CS1/Configuration (the engine that renders {{cite book}}, {{citation}}, etc) to determine the dmy/mdy form used by the article and automatically conform the dates in cs1|2 citations to that form. Additionally, it is proposed to add a non-functioning parameter, |cs1-dates=<keywords>, to instances of the {{use xxx dates}} templates that will direct the cs1|2 module to render dates in all of the formats permitted by MOS:DATEUNIFY. I recognize that that conversation may be a bit too technical for some readers, and I'm sorry for that. Questions and comments sought at the linked discussion please.

—Trappist the monk (talk) 14:24, 23 March 2019 (UTC)

Discussion at WT:NCPEOPLE

Editors watching this page may be interested in this NCPEOPLE discussion as it pertains highly to citations. --Izno (talk) 20:53, 27 March 2019 (UTC)

xkcd at it again

I feel that this xkcd is an entertaining commentary and example on which statements are controversial enough to need (inline) sourcing ;) Nosebagbear (talk) 00:00, 28 March 2019 (UTC)

Positioning footnote when attributing a whole section/list/table to one source

Example: The Front Page#Cast (permalink, in case someone fixes it). We have a long list of actors and their roles, and every entry ends with a footnote linking to the same citation. This looks ridiculous, right? But how would you fix it?

1. Only keep the footnote on the last item in the list
2. Remove all the footnotes for the list items, and add a footnote to the section heading (as is done at Wicked (musical)#Musical numbers[38])
3. Remove all the footnotes for the list items, add the word "source" immediately after the list, followed by a footnote.
4. Remove all the footnotes for the list items. Add some text before the list saying something like "The following is the cast, as listed on Playbill.com" plus a footnote.
5. Remove all the footnotes for the list items. Add a footnote (without text) on a separate line between the section heading and the list, as in 2010_US_Open_Series#ATP.
6. Like 5, but put the bare footnote after the list.
7. Don't fix it, it's fine as is

I'm honestly not in love with any of these options. 4 is a nice option if there's some text that would naturally introduce the list/table, but in this case I don't think the added text is very helpful to the reader - it's just a hook on which to hang a footnote.

Semi-related archived discussions: How to cite tabular data, Question of placement, How to apply a single cite to an entire section?

(Also, maybe it's too WP:CREEPy, but I wonder whether this should be documented somewhere, since it seems like a common scenario) Colin M (talk) 00:05, 31 March 2019 (UTC)

I prefer #4 for lists like this. Tables can be cited in their <caption> element if the citation is true for every row (|+ ABC<ref>); however, in the table case I find things go out of date commonly as the tables increase in size due to age. Sometimes if there are few refs keeping them in the caption makes sense, but sometimes you get up to 10 refs or so and suddenly the cite-each-row option looks attractive. --Izno (talk) 16:26, 31 March 2019 (UTC)

Where to find help for a full cite check?

Is there a place to request cite checks? I've run into such numerous sourcing problems at SNC-Lavalin affair (WP:INTEGRITY, WP:WEIGHT, WP:SYNTH, WP:OR, WP:PRIMARY). Thus, I've been trying to perform a full source check there, but I just don't have the free time to do it in a timely manner. A number of editors have been editwarring to remove the {{Cite check}} template there before the check is complete. Curly "JFC" Turkey 🍁 ¡gobble! 00:39, 17 April 2019 (UTC)

What do you mean by "trying to perform a full cite check"? Possibly (your phrase can be parsed in different ways) you want to "check the full citations"? Well, do you want to check that the full citations are complete and in proper form? Or do you want to verifiy that the source identified by the citation actually supports the material for which the source is cited?

Or, seeing as this seems to be about use of the {{cite check}} template, perhaps you want to do a "full" cite check? Sorry, I don't know what a "full" cite check is. That template is for alerting readers that "citations in an article or section may be inappropriate or misinterpreted"; I don't see any "full" about that. If you want to "fully check" the citations, fine, but you don't need to "alert" the readers that you suspect problems that you have not had time to substantiate. That in that article there are no {{verify source}} tags, let alone any {{failed verification}} tags, strongly suggests there is no basis for the cite check template.

You say a number of editors are edit-warring. That looks like you on one side, and four editors on the other side. As it really looks like you are here asking for some kind of tool to justify an alert for problems you can't find where other editors object, I think there is nothing further to discuss here (unless it be the inadequacies of the {{cite check}} template). I will have additional comments for you at Talk:SNC-Lavalin_affair. ♦ J. Johnson (JJ) (talk) 20:48, 17 April 2019 (UTC)

J. Johnson—I don't "suspect" problems—I've found numerous, and corrected a number of them myself, either by fixing the text, or fixing or adding citations. There are no {{verify source}} tags because I fix what I've found. This is why the full source check is so time consuming, and why I'm asking for help. As you're not offering to help or improve the article in any way, you really shouldn't be engaging in an editwar there. Please restore the tag. Curly "JFC" Turkey 🍁 ¡gobble! 01:18, 18 April 2019 (UTC)

To answer the original question as I understand it, there is no such place. Mike Peel has an RFC about having an explicit separate place to do a source review. The closest thing today would be a topical WikiProject. Otherwise, I guess you could try the help desk or VPM. --Izno (talk) 22:20, 17 April 2019 (UTC)

Woops, wrong Mike. Mike Christie. --Izno (talk) 22:21, 17 April 2019 (UTC)

The RfC failed; it would have established a separate source review step as part of FAC. Mike Christie (talk - contribs - library) 23:44, 17 April 2019 (UTC)

Mike Christie—that's really unfortunate. The SNC-Lavalin affair article is riddled with sourcing issues, and a number of POV-pushers keep removing the {{Cite check}} template before the citation check is even complete. Now J. Johnson has joined in the editwarring to have it removed. It's exasperating that these issues are simply allowed to fester.

If you want a cited sourced checked against the article's content to assure that it actually sources the article content, you can ask at WP projects maintaining the article in question or you can try to request a digital copy of the source in question at WP:RX to do the check yourself.--Kmhkmh (talk) 22:31, 17 April 2019 (UTC)

Kmhkmh: I have, but the POV-pushers have been successful in driving every other objector off but me. The sourcing issues are severe and widespread, and certain editors there are doing everything they can to hide them—repeated removal of the template being but one tactic. Curly "JFC" Turkey 🍁 ¡gobble! 00:54, 18 April 2019 (UTC)

Curly Turkey: I am not engaged in an edit-war, and my removal of the {cite check} at SNC-Lavalin affair is a matter for the Talk page there. As to any matters which might be appropriately discussed here: you asked, "Is there a place to request cite checks?" And you have received an answer: No, there is no such place. I remind you that this talk page is for discussing improvements to the Citing sources page. You have not raised any issue regarding the Citing sources page. You are quite simply looking for resources to use in a dispute elsewhere; that is off-topic. Unless you have further comments pertinent to this page it appears this discussion is done. ♦ J. Johnson (JJ) (talk) 19:11, 18 April 2019 (UTC)

Repeat references

In the last day, I have discussed the issue of Straight vs Curly quote marks in regard to Ref Name at Help talk:Cite errors and at Village Pump. I followed up with edits adding instructions at Help:Cite errors and at Help:Cite errors/Cite error references no text. I propose adding a similar instruction as the last paragraph in the Repeated Citations section of this Project page, namely:

Use straight quotation marks " " to enclose the reference name (size exaggerated). Do not use curly quotation marks “ ”. Curlies will simply be treated as another character, not as quotation marks. If one style of quotation marks is used when first naming the reference, but the other style is used in a reference repeat, an error message will be seen on the published page.

The section Help:Footnotes#Footnotes: using a source more than once already has adequate instructions about the Straight vs Curly issue, so no added instructions are needed there. DonFB (talk) 05:47, 24 April 2019 (UTC)

Archiving a source without original source

I am trying to switch a source on the Del Barber article to an archived source as NewCanadianMusic.ca does not exist anymore. How can one use the archived version without the original? I wouldn't be switching over to the archived version if the site still existed. Mr. C.C.^{Hey yo!}_{I didn't do it!} 03:29, 6 May 2019 (UTC)

I took a crack at it by adding {{dead link|date=May 2019}}. A bot will hopefully come along and attempt to replace the link with an archived copy fairly soon. —BarrelProof (talk) 21:24, 6 May 2019 (UTC)

@BarrelProof: Thanks. Mr. C.C.^{Hey yo!}_{I didn't do it!} 00:19, 7 May 2019 (UTC)

RfC regarding italicization of the names of websites in citations and references

There is a request for comment about the italicization of the names of websites in citations and references at Help talk:Citation Style 1/Archive 72#Italics of websites in citations and references – request for comment. Please contribute. SchreiberBike | ⌨  04:53, 18 May 2019 (UTC)

Citation template of jpg and other images

There is a url used as an inline citation (an image of a grave stone).

• http://mapage.noos.fr/xgen/roglo/ga.lestevenon.jpg

I would like to place it in a citation with a website, title, etc. Is there a specific template for dealing with such things? PBS (talk) 11:56, 22 May 2019 (UTC)

Perhaps {{cite sign}}.

—Trappist the monk (talk) 12:57, 22 May 2019 (UTC)

reference with the text, or in the References section?

Not that I am disagreeing, but an editor moved a reference to the References section. Specifically, <ref name="xxx"/> in the text, with the full <ref name="xxx">...</ref> in References. Personally, I think this is fine, as it often makes it easier to edit references, but rarely see it done. It seems that the visual result is the same in both cases. Gah4 (talk) 20:06, 29 April 2019 (UTC)

@Gah4: WP:LDR is an acceptable citation style, but the original citation style of an article should not be changed unless there is consensus to do so. – Finnusertop (talk ⋅ contribs) 20:44, 29 April 2019 (UTC)

A common disputation is whether changes that make no difference in the display of the text are subject to CITEVAR. The reason you don't see things done in certain ways is often because there's herd mentality of doing things just like everyone else does. ♦ J. Johnson (JJ) (talk) 22:27, 29 April 2019 (UTC)

I have made this sort of change without prior discussion before myself, but I think if there is any objection then you need to stop and discuss rather than keep pushing for it in the article itself. There are pluses and minuses to placing references at the end like this. On the one hand, it unclutters the wikitext of the article so you can actually find the text you're trying to change. On the other hand, it makes section-by-section editing much harder, because the references you need for the section aren't there. On the whole I prefer putting them at the end but I can see how reasonable editors would disagree. I think there is also scope for disagreement on whether this is covered by CITEVAR, too (the actual question here) so that's why I tend to take an intermediate position: try it only once and then back off if anyone objects, more quickly than I might back off for other kinds of changes. —David Eppstein (talk) 00:08, 30 April 2019 (UTC)

That's how I view WP:BRD: feel free to try something, but back-off if anyone objects. ♦ J. Johnson (JJ) (talk) 23:30, 30 April 2019 (UTC)

The article I was asking about is pretty small (one paragraph) and only one reference. (Probably should be called a stub.) I didn't actually know how to put references in the References section before, and now that I do know, I might be more likely to do it. Gah4 (talk) 07:12, 1 May 2019 (UTC)

You might also be interested in seeing WP:Basic citation concepts. ♦ J. Johnson (JJ) (talk) 21:53, 1 May 2019 (UTC)

Somewhat off topic, but perhaps useful:

I have quite a bit of experience with this. Years ago I sought and received a local consensus to convert an entire article to LDR while it was still being actively developed. Once I completed that, I found that every last editor still did citations the other way, for any of a number of reasons. So I became the unintended LDR-keeper at the article, converting each new citation to LDR and cleaning up the errors resulting from editors' lack of understanding of LDR. Despite seeing those edits in the page history and the effect in the wikitext, even the regulars at the article still did citations the other way. It's a chicken-and-egg problem: editors won't understand LDR unless it's widely used, and it won't be widely used unless editors understand it. I don't think LDR will ever be really viable absent an editor willing to make such a time-consuming commitment as LDR-keeper. Anyway, as suggested above, it's unclear whether LDR is a net benefit when all factors are considered. ―Mandruss ☎ 22:29, 1 May 2019 (UTC)

We should talk about LDR someday. When we can find a few "round toits"? :-) ♦ J. Johnson (JJ) (talk) 21:06, 2 May 2019 (UTC)

I thought WP:CITEVAR as it currently stands is pretty clear. In the "To be avoided" section:

“ When an article is already consistent, avoid [...] changing where the references are defined, e.g. moving reference definitions in the reflist to the prose, or moving reference definitions from the prose into the reflist. ”

So I thought this was against guidelines? If it's ok, then I owe a fellow editor a big apology. —hike395 (talk) 21:27, 22 May 2019 (UTC)

Keep in mind that WP:CITEVAR is part of a guide line, not a third rail, and that it explicitly refers to making edits per personal preference and without consensus. But then, there have been numerous disputes where some editors maintain that it applies only to the displayed style, and does not apply to trivial "under the hood" details like, oh, template formatting style. I'd say that while "asking first" for any possibly contentious edit isn't required, anyone editing boldly (perhaps brashly?) should not take offense when there is push-back. ♦ J. Johnson (JJ) (talk) 23:23, 22 May 2019 (UTC)

I think, as a rule of thumb, CITEVAR is best understood as: "Lots of ediitors care a lot about various aspects of citations. Please respect that and be doubly careful about seeking consensus and to avoid making changes that aren't absolutely necessary. There have been huge conflicts and edit wars about this. Which is dumb. So please bend over backwards to avoid creating such in future." Switching to or from LDR is thus clearly what CITEVAR is trying to address: not the technicalities or technical details, but the "avoid creating needless conflicts" sense. Or put another way; to argue whether a particular change to citations is covered by CITEVAR is effectively missing its main point. --Xover (talk) 05:57, 23 May 2019 (UTC)

Well said. But some editors will insist that "there's no rule against it", or complain of WP:OWNership, or some such. ♦ J. Johnson (JJ) (talk) 18:53, 23 May 2019 (UTC)

When no (Google books) preview available for cited book

Hello - Is there a Wikipedia protocol or 'best practice' concerning Google books citations that lead to a book without a preview or search function, such as this one or this one? If not, some insightful advice would be welcome. Thanks, cheers. TP ✎ ✓ 12:28, 26 May 2019 (UTC)

I would just remove it - it's not much better than just linking to an Amazon page for the book (minus the obvious promotionalism). Primefac (talk) 13:07, 26 May 2019 (UTC)

For the avoidance of doubt, you might want to clarify that you'd remove the link from the citation, not remove the whole citation itself. Provided, of course, the citation otherwise contains all the necessary bibliographic details. Personally, I tend not to include google books links at all, unless the source is accessible there in full: reliance on previews and snippets encourages not taking account of the context, and so is bad for article quality. And a caveat: whether a source is accessible on google books, and how much of it is accessible, probably varies with time and place: what you're allowed to see on your computer today might not be the same as what another user in a different country will be able to see next month. – Uanfala (talk) 13:23, 26 May 2019 (UTC)

Thanks for your replies. But isn't just citing x page in y book, without means of verification, a bit of an open door to a contributor 'sourcing' dubious claims? In the article I'm presently working on, some have been kind enough to provide quotes from some inaccessible material, but even this is not always verifiable. TP ✎ ✓ 15:20, 26 May 2019 (UTC)

Citing x page in y book is the means of verification. See WP:SOURCEACCESS. – Uanfala (talk) 15:25, 26 May 2019 (UTC)

So to verify that citation, one would have to actually find the book in print (university, library, bookstore)? Again, that seems to be an open door to 'itsintherejusttakemywordforit-ism'.

But come think of it, I've had to resort to that for some rarer subjects (an article about a train that once ran a ring around Paris that stopped running in the 1930s, for example - only a few out-of-print books and archive records were available)... but what about citations where more modern/mainstream publications are available, around controversial topics, or on doubtful claims are concerned? One can claim (or 'interpret') just about anything if the source work cited is not available. I would tend to doubt such sources for extraordinary/dubious claims, but would appreciate some other thoughts/experiences about that. TP ✎ ✓ 16:45, 26 May 2019 (UTC)

Of course you have "resort" to going to a library to look at a book or journal that's not on the internet. We not only can't, but absolutely shouldn't, try to build an encyclopedia based only on online sources. Peter coxhead (talk) 18:43, 26 May 2019 (UTC)

@ThePromenader: This is bit tricky. In general there is no reason to link google books if no preview is available. However due to the weird (partially personalized) inner workings of google algorithm, the fact that you can't see it (now), doesn't mean that other people can't see it or that you or other won't be able to see it in the future. Also sometimes while one edition of a book might become inaccessible at Google books another edition might still be accessible. So before deleting one should check that the content is not accessible in other editions.--Kmhkmh (talk) 19:54, 26 May 2019 (UTC)

I don't see how one is supposed to determine that, given that accessibility may depend on factors like what country one is trying to access the link from. So it's safest just not to delete these links. —David Eppstein (talk) 20:17, 26 May 2019 (UTC)

I agree but and i sort of want to hint atthat. However afaik there is no consensus on that, some people procipally delete links that are or seems dysfunctional (and some don't like to see Google linked at all). Bare of consensus or clear regulation editors will simply have to use "good judgement".--Kmhkmh (talk) 00:40, 29 May 2019 (UTC)

It seems to me that a Google preview or snippet of a book page, to which some text of a WP article is presumably referenced, is better than not having such preview. Yes, having the entire book online would be ideal, but failing that, why not prefer a snippet; why opt to avoid a snippet, which could show, explicitly, the basis for Verification? DonFB (talk) 00:03, 27 May 2019 (UTC)

Google snippets can sometimes be useful provided care is taken to ensure that they're not something that could be contradicted by unseen wording outside the snippet: if there's an issue with snippet or preview, add |quote= to the reference template and quote the words you're working from, that will be helpful for anyone else searching for the relevant part of the source. Better, when possible, to look at the original book – full text online is helpful, but subject to transcription errors. . . dave souza, talk 09:44, 27 May 2019 (UTC)

From a reader-verification perspective, I see sense in asking (the contributing author) for a quote if the original work is not accessible (as a minimum courtesy)... and that might dissuade the temptation to cite 'interpreted' sources (or partake in WP:SYNTH), also. Might this even be Wikipedia protocol material? TP ✎ ✓ 10:07, 27 May 2019 (UTC)

I completely disagree. The editor or author is only required to give an exact (offline) reference and nothing more - period! Such "courtesy quotes" just bloat references and are a dangerous simplification as you can verify something properly only if you read the context of that quote as well, so the surrounding text in the source.

There is however another sort of courtesy that is if another editor has reason to distrust the reference or has some other question, then he may request a quote or private copy. For that he can ask the author who inserted the source or ask at WP:RX.--Kmhkmh (talk) 00:52, 29 May 2019 (UTC)

The primary goal of a citation is to tell the reader where they can find the source. For a book, the title, author, page number, and perhaps publisher is what you need. If you've got an ISBN, that's a great convenience, because it makes searching easier. And, likewise, if you've got an on-line version, a URL to that is an even greater convenience. But, a URL that just points to a google/amazon/whatever index page, doesn't add anything useful. -- RoySmith (talk) 01:05, 29 May 2019 (UTC)

How to handle reliable sources that provide conflicting info?

I'm currently working on adding cites to a page that has a banner declaring the need for such work. (Sorry, I don't know the technical terms Wiki uses for this stuff.) I currently have 2 sources that state one thing and 2 that state something completely different - specifically the date of formation of a band.

The issue: The band formed with 4 members for a short time, disbanded and at some point reformed with some same, some different members relative to the initial group. This occurred over the fall of one year into the spring of the next year. The reformed group with the somewhat different members is the one that was signed to Columbia Records. Like I mentioned I currently have 2 sources that support one date, 2 that support a different date.

All four source are publication/interview based (Rolling Stone & Boston Globe, actually) and web based with editorial oversight. In reading over and comparing info I think I can come to a reasonable conclusion on the matter, but it's a bit tricky in places. Is it best to present my conclusion and use all four sources to cite that conclusion? Provide a brief mention of the conflict, mentioning each date with cites that support each date? I'm still acquiring sources also - I have about 12 so far. Thanks for any help offered. THX1136 (talk) 03:01, 27 May 2019 (UTC)

Don't add your interpretation, that is WP:OR. From what you've said I'd base things on your second paragraph: "The band initially formed in 2000. It disbanded the following year and reformed with a partial change of members."Martin of Sheffield (talk) 09:26, 27 May 2019 (UTC)

References

Thanks for your input, Martin. Seems like a good path forward.THX1136 (talk) 13:42, 27 May 2019 (UTC)

• I faced this problem in Margaret Sibella Brown. I just cited all the sources and added a note explaining how the sources differed and why I chose the one I did to believe. That way, people can still go look at the original sources and draw their own conclusions. -- RoySmith (talk) 15:16, 4 June 2019 (UTC)

How to cite an English translation?

In Draft:History of the models of the solar system, Kaulins, Andis (2005). "Die Himmelsscheibe von Nebra: Beweisführung und Deutung" is a paper cited in the original German. There's also an English translation available. What's the best way to provide URLs for both the original and the translation? I shoved the english URL into the Translated title field, but that's not really right. -- RoySmith (talk) 15:22, 4 June 2019 (UTC)

Provide the one you read per WP:SAYWHEREYOUGOTIT. If the other is generally interesting, you can place it in further reading, or if you read both, use a separate reference for each, or if you're really certain you want them in context of each other, two templates in the same ref statement. --Izno (talk) 15:54, 4 June 2019 (UTC)

(edit conflict)

Yeah, don't do that; it doesn't work for you and it doesn't work for readers. You are trying to use {{cite journal}} for something that it was not designed to do: cite two sources with a single template.

To render the translated title, Module:Citation/CS1 adds [ at the start of the value assigned to |trans-title= and it adds ] at the end; these are the brackets that are used to identify the translated title in the rendered citation. But, in this case, MediaWiki thinks that the trailing ] is part of the url; ancientworldblog.blogspot.com doesn't like that.

The content of |trans-title= is not made part of the template's metadata so readers who consume your citation via the metadata will not know that there is an English translation.

Were it me, I would cite both sources separately, perhaps in the same <ref>...</ref> tag:

{{Cite journal |last=Kaulins |first=Andis |date=2005 |title=Die Himmelsscheibe von Nebra: Beweisführung und Deutung |trans-title=The Sky Disk of Nebra: Evidence and Interpretation|url=https://www.efodon.de/html/archiv/vorgeschichte/kaulins/2005-SY6845%20Kaulins_Nebra.pdf|journal=Efodon-Synesis|language=German|issue=2|pages=45-51}}

Kaulins, Andis (2005). "Die Himmelsscheibe von Nebra: Beweisführung und Deutung" [The Sky Disk of Nebra: Evidence and Interpretation] (PDF). Efodon-Synesis (in German) (2): 45–51.

{{cite web |last=Kaulins |first=Andis |date=25 June 2005 |title=The Sky Disk of Nebra: Evidence and Interpretation |url=https://ancientworldblog.blogspot.com/2005/06/nebra-sky-disk.htm |website=Ancient World Blog}}

Kaulins, Andis (25 June 2005). "The Sky Disk of Nebra: Evidence and Interpretation". Ancient World Blog.

—Trappist the monk (talk) 16:19, 4 June 2019 (UTC)

Guidance on self-citing sources?

I'm sure there's something written up about this, but I haven't found it. I see a lot of "Song 'Foo' is played during the climax of film 'Bar' {Citation needed}" (or "historical character 'John Smith' appears in book 'Tam'" {Citation needed}) out there. It seems to me that "Bar" is its own best reference to what it contains. Is there guidance for when and when not to add citations here? - Immigrant laborer (talk) 18:30, 15 June 2019 (UTC)

@Immigrant laborer: A piece of media—book, movie, song, etc.—is a permissible primary source for itself. So long as you are only citing it for non-controversial factual information about itself it is a perfectly reliable source, and may often be the best and preferable source for that information. For any information that is controversial (e.g. did a given artist collaborate on a given song) or any interpretive statement you will need to cite a secondary source.

There is, sadly, no good guidance on this in any policy that I have found. The closest to address it directly is MOS:PLOTSOURCE, which has a fairly narrow focus. The rest is just a logical consequence of WP:PSTS (part of WP:NOR) and things like WP:ABOUTSELF (part of WP:V) and WP:SELFSOURCE and WP:RSPRIMARY (both part of WP:RS). But discussions at WP:RS/N have repeatedly found primary sources to be reliable sources for non-controversial factual information about themselves, so this isn't really a controversial issue in general, even though any given application of this principle may well be so. --Xover (talk) 05:15, 16 June 2019 (UTC)

Personally, I've long held that when it comes to most citations there's little practical difference between primary, secondary, and tertiary sources: no matter what the "kind" of source, we can only use the source for what it says and cannot make our own interpretation. There's nothing magic about a "secondary" source that changes any of that, it seems to me that people just confuse the heuristic that a secondary source is more likely to contain the kinds of interpretation we need for an encyclopedia article with some intrinsic property of being "secondary" (even WP:RSPRIMARY falls victim to this). Or are trying to use WP:N-like language to deal with the lack of a policy about "relevance to the topic" when someone comes along trying to cast aspersions in an article.

People seem to also sometimes confuse "secondary" with "independent" or "reliable". When something is controversial we want it filtered through some authority on the subject that is still considered independent. This particularly can break down when we're citing a statement like "Authority X says Y", as the citation may well use some publication of X as a (primary) source for that statement.

Also of note is WP:BLPPRIMARY, which forbids using primary sources for BLPs. In glancing through past discussions, it seems this is a combination of avoiding linking to documents that would be useful for doxxing, the "no 'relevance' policy" thing, and concern over casting of aspersions by using sources that were later proved erroneous. Anomie⚔ 12:32, 16 June 2019 (UTC)

Thank you both! - Immigrant laborer (talk) 13:01, 16 June 2019 (UTC)

@Anomie:WP:BLPPRIMARY does not forbid the use of primary sources in biographies of living persons, it says "Exercise extreme caution in using primary sources." So there would be no problem using the dust jacket of a book by the subject of a biographical article, published by a reliable publisher, as a source for an uncontroversial statement about the author. Jc3s5h (talk) 14:43, 16 June 2019 (UTC)

Question and suggestion about attribution

I've read Wikipedia:Citing sources#In-text attribution, but there isn't any mention of attribution at the end of articles in the references (i.e., "Attribution [reference]"), nor can I find it elsewhere. (It doesn't help that every Wikipedia page ends with the notice "Text is available under the Creative Commons Attribution-ShareAlike License ..." (emphasis added).) Where is it mentioned? Adding a link to Wikipedia:Attribution, would, IMHO, also be good. —DocWatson42 (talk) 05:59, 25 June 2019 (UTC)

WP:INTEXT refers to including the source inside a sentence. (John Smith asserts that the Earth is flat. instead of The earth is flat.) WP:Attribution (a failed proposal) refers to WP:V/WP:NOR. (This is a statement cited to a reliable source. instead of This is a statement that cannot supported by a reliable source.) WP:FREECOPYING discussed attribution for prose copied from a freely licensed source. (This is a sentence that I copied from a public domain work.) — JJMC89 (T·C) 01:28, 26 June 2019 (UTC)

References

Citing tweets that contain emojis

I made a post earlier in Template talk:Cite tweet. I have a feeling there are more editors watching this page, so I'm cross-posting it here in hopes of getting more perspectives/opinions:

Some tweets may contain emojis. I ran into this problem earlier when using the template to cite this tweet. Should emojis be included in [the title parameter]? On one hand, it could be argued that including emojis is important for keeping the full source quote intact (and for contextualization). On the other hand, emojis in the citations could be distracting or viewed as unencyclopedic or unprofessional. I think certain emojis, like the US Flag one in the previous example, don't render properly in Wikipedia. What is the consensus on including/excluding emojis?

Bobbychan193 (talk) 00:12, 11 August 2019 (UTC)

Editors should consider responding there rather than here, as the original question was posed there (per WP:MULTI). --Izno (talk) 15:30, 11 August 2019 (UTC)