Wikipedia talk:Citing sources/Archive 31

Should we tread on other guidelines while writing this one? While I like having an INTEXT section, here's a few changes:

• Change argues, argues, maintains to says, states, states, per WP:SAY. This change would also appear in WP:V et any al.
• Change "a helpful thing" to "helpful".
• To remove climate-change detritus, change to a vague issue where WP's throwaway implication of a majority position will not be a debate issue. E.g.: "John Smith states that these variations all constitute one species" and "his view may be held by the majority of geneticists".
• Change "fixed by writing" to "fixed by citing the statement", because a debated majority or plurality view must be cited as such.
• Change "this evening" to "on 10 November" (i.e., coded as "on {{date||dmy}}") per WP:DATED. JJB 12:49, 27 November 2010 (UTC)

Editors who watch this page may be interested in Wikipedia:Village_pump_(proposals)#Change_to_template:reflist_wiki-wide?. Please leave any comments on the village pump page. — Carl (CBM · talk) 20:29, 29 November 2010 (UTC)

Carl, what other methods are there, apart from footnotes and parenthetical referencing? [1] SlimVirgin ^{talk|contribs} 03:52, 30 November 2010 (UTC)

If you define all reference formats to be one or the other, then of course there aren't any others. But there are certainly many variations on how footnotes or parenthetical referencing are achieved. One can use cite.php or use some other system for footnotes, and can format the footnotes in many ways. One can put parenthetical references in parentheses or square braces, and can format them dozens of ways. So the examples on this page aren't (and can't be) exhaustive. We don't want to see people go around blindly changing formats just because the formats don't match the examples here.

The core policy on this page has always been that any internally consistent style is acceptable as long as it meets the needs of verifiability. It would be a complete about-face to change that to say that now only a few methods are acceptable. — Carl (CBM · talk) 04:07, 30 November 2010 (UTC)

This is what the policy said: "The two styles of inline citation used on Wikipedia are clickable footnotes (<ref> tags) and parenthetical references. Editors are free to use either method."

To suggest that they can use any method, including the above, suggests that there's something other than footnotes or parenthetical refs. But there isn't, so it's confusing as currently written. SlimVirgin ^{talk|contribs} 05:49, 30 November 2010 (UTC)

Editors could, for example, use footnotes but not use cite.php to achieve them. The problem with the wording is that it suggests a false dichotomy: that footnotes have to be done the way shown, and the parenthetical references have to be done the way shown. But the inline citations can be done in any internally consistent manner. The point of this page is not to tell editors how they must format things, only to tell them how they can format things. — Carl (CBM · talk) 13:02, 30 November 2010 (UTC)

At some point that was changed from "the two most popular styles" (correct) to "the two styles" (not)... Christopher Parham (talk) 19:20, 30 November 2010 (UTC)

[2] SV's copyedit changed "On Wikipedia, there are several different styles. The two most popular are ... " to "The two styles of inline citation on Wikipedia are ..." Editors can also use prose citations, naming the author or work in prose in sufficient detail to identify it in the list of references. Geogre seemed to prefer that, as I recall. Gimmetoo (talk) 13:16, 30 November 2010 (UTC)

Another method exists, apparently meant mainly for tables that need footnotes, but which I saw for regular footnotes. Please pardon me for forgetting what it's called or how to write it, but I had a scathing opinion of it when I found it in one article, albeit only one article. It uses one template in the article and another template in the foot of the article. If you need to add a note anywhere except the end, you have to manually renumber each subsequent note. You make the body template and the note template correspond by writing matching text into each template, which means you have to make sure no other templates of the types have the same wording. My way of dealing with it was to change it to the nontemplated <ref>Refererent.</ref> method, which change we're generally not supposed to do once a method is established for an article, but no one has reverted, complained, asked, or said anything about it, so I guess no other editor on that page minds very much. But the method exists, so I suppose it's good enough for WP:CITE as a third, fourth, or n^th method. Nick Levinson (talk) 04:25, 1 December 2010 (UTC)

You're probably thinking of the {{Ref}} family of templates (see documentation there). Those templates predated the introduction of the Cite extension which introduced support for the <Ref> tag (See WP:FOOT). Those templates are still used in some articles. Wtmitchell (talk) (earlier Boracay Bill) 23:47, 1 December 2010 (UTC)

Yes, that's it. I found the article I had edited. Thanks. Nick Levinson (talk) 02:01, 2 December 2010 (UTC)

{{Ref}}/{{Note}} creates a footnote, so it doesn't represent a counter example to SlimVirgin's observation that the only inline citation styles are footnotes or parenthetical references. ---- CharlesGillingham (talk) 19:10, 4 December 2010 (UTC)

I think we need to stick with the purely descriptive "most popular" language. It is indisputably accurate.

Nick, I've also (recently, even) seen hand-typed, non-clickable/non-ref tagged footnoted numbers. I don't recommend it—actually, I recommend against it—but it is certainly done on occasion, and if you're not hung up on having every single footnote in numeric order, it's not even very difficult to maintain as the article gets expanded. WhatamIdoing (talk) 02:39, 2 December 2010 (UTC)

Can you link to that example?

I think we should be careful not to veer too far in the direction of anything goes. We don't require per policy that people avoid spelling mistakes, but we do fix them when we see them. SlimVirgin ^{talk|contribs} 05:55, 2 December 2010 (UTC)

Manually numbered footnotes would be acceptable if there were only one author, but that is just what Wikipedia is not about; for collaboration, auto-numbering of references is vital. — Robert Greer (talk) 16:27, 2 December 2010 (UTC)

Manually numbered footntes are also footnotes, and still don't represent a counter example. ---- CharlesGillingham (talk) 19:10, 4 December 2010 (UTC)

The underlying issue here is standardization.Standardization has both positive effects and negative effects. Among the positive effects are

1. Standardization shortens learning curve for new editors. It simplifies the documentation and gives new editors clear directions.
2. Standardization helps bots and other editors to find typos and minor errors and correct them. Articles with unusual citation methods tend to be loaded with mistakes (See, for example {{wikicite}}, where I corrected over 1,000 articles with faulty wikilinks, leaving only 100 articles that used the template correctly.)

Among the negative effects are:

1. Standardization stifles innovation. New citation mothods are, by definition, non-standard. See list defined references, {{sfn}} or {{vcite book}}, all of which are "non-standard" but all of which are (arguably) "better" in some way.
2. Standardization invites edit-warring. There are editors who hate certain citation methods and other editors who love them. These editors have, in the past, caused needless strife that does not help the project. The only way to keep these editors in check is to make it clear from the start that you can't change citation methods without consensus.

I think I am just scratching the surface of this issue. Perhaps an essay would be a good place to collect these arguments.

Standardization, out in the article space, always means changing an article from an unpopular style to a popular style, and, as the guideline should point out at every turn, this is a dangerous practice that can lead to trouble and should only be carried out with caution.

(My own view is that we should eliminate (1) truly unusual citation styles (used in less than 500 articles) that are (2) damaged by neglect and (3) clearly are not an "innovation" in any sense. This includes things like unnecessary use of <cite> spans, {{harvrefcol}}, unnecessary use of {{Ref}}/{{Note}}, unnecessary use of |Ref= in {{cite *}} templates and many other, weirder things that are floating around. There are tens of thousands of articles with broken and badly formatted citations that use these methods. I think this kind of maintenance should not be controversial, and is the only kind of standardization which makes sense. Otherwise, we should only document the most popular styles and try to document innovations as soon as they are stable. I believe this is what this guideline does.)

Back on topic and with all this reasoning in the background, I think that this guideline should include the observation that there are really only two (stable, reliable, popular) inline citation styles: parenthetical referencing and footnotes. This shortens the learning the curve and is unlikely to cause edit warring. If there are editors who hate both parenthetical referencing and footnotes, we should document that. If someone can find an innovation that isn't one of these styles, we should document that. However, I don't believe there is any such innovation and I don't believe there are any such editors. ---- CharlesGillingham (talk) 19:10, 4 December 2010 (UTC)

Made a couple of minor edits that I think improve sentence flow. It's Ok if others disagree. —Preceding unsigned comment added by Clipjoint (talk • contribs) 22:29, 3 December 2010 (UTC)

MediaWiki obviously has this nice feature that a reference that is used more than once can be named, which means if you cite <ref>Miller (2000), p. 100.</ref> more than once, you can name it <ref name="Miller 2000 p. 100" />, to avoid a bunch of references that look the same. If it detects multiple identical refs, WP:AWB automatically applies this as a general fix. The question now is whether this is based on a MOS guideline, because after applying this here, User:Hegvald reverted it. Who's right? —bender235 (talk) 13:17, 8 December 2010 (UTC)

Since there is no preferred citation style, whoever edits first sets the style. Changes can be discussed per WP:BRD. Your edit summary was rather misleading. ---— Gadget850 (Ed) ^talk 14:42, 8 December 2010 (UTC)

To quote WP:FOOT, "the use of named footnotes ... is a matter for editorial judgment; some editors do repeat the entire footnote, in case rearrangement of the text removes the first note, or places it after a blank note (previously, the note had to be defined prior to use, although that is no longer the case)." Neither style is preferred; if there's disagreement about which style is more appropriate for the article, it should be discussed on the talk page. Christopher Parham (talk) 15:23, 8 December 2010 (UTC)

AWB is supposed to only add new names to references when there are existing named references. So the fact that it added them here seems to be a bug in AWB. I filed a bug report about that. — Carl (CBM · talk) 16:09, 8 December 2010 (UTC)

I would be grateful for precedents for following practices (or comment that they would not conflict established WP "official" or accepted style). These measures would make it MUCH easier to refer to WP articles in non-WP publications. (1) State, in the body of the article, the relevance of every item in the "Further reading" list to the subject of the article, and provide citation number to main bibliography unless this would become too long, in which case make secondary bibliography using {{Reflist}}. (2) Number the items in lists instead of using bullets, particularly when there are several, as in Edward Elgar. (3) Put a few (selected) references to biographies, obituaries, photograph collections, etc. at end of opening sentence in article about a person as in William Anderson, artist. I would only do this when it led to just a few references, and not when there is a Biography section covering entire life immediately after opening sentence, as for Edward Elgar Michael P. Barnett (talk) 17:51, 12 December 2010 (UTC)

I propose to add to WP:SAYWHEREYOUGOTIT, to clarify what does not need to be included in a citation. Some non-Wikipedia style guides call for information that Wikipedia does not request, and too much information may interfere with readers' use of WP.

This follows the discussion at another Talk topic. I've added microforms to the proposal.

Nick Levinson (talk) 05:25, 5 November 2010 (UTC)

That text sounds about right, although it seems a bit wordy and I'm not convinced about the last two sentences. I also wonder whether it's the best place on the page to put this information... but it is the section that editors seem to think addresses it, so perhaps it is. WhatamIdoing (talk) 20:51, 9 November 2010 (UTC)

On the last two sentences: One is for the case of citing an article in the N.Y. Times; a link to the article on the nytimes.com website should be provided, which is different than linking to a third-party site, though both may restrict access. The other is that if the only source is online then access should not be totally omitted even though only some readers will be able to access it, because a difficult-to-reach source is better than no source at all, although a more easily verifiable source remains preferable if existent. Nick Levinson (talk) 11:58, 10 November 2010 (UTC)

So in your proposal, those two sentences run:

If the publisher offers a link to the source or its abstract that does not require a payment or a third party's login for access, you should provide the URL for that link. And if the source exists only online, give the link even if access is restricted.

On the first sentence, even URLs to WP:PAYWALLed and registration-required sources are desirable in some instances. However, such links, even if free, might not be desirable in others, e.g., when the URL is redundant to the doi, or is at a notoriously unstable website (here I am naturally thinking of links to news.yahoo.com).

On the second, I'm not convinced that this is necessary in all cases. For example, many academic journals provide an online supplement to articles, such as extra images. It would be perfectly adequate to cite "Jones, Mary. 2009. Supplementary images for "<Name of Paper>." M Pressive Journal..." Furthermore, while the source may never reach hard copy, "my" link isn't necessarily "your" link. I might use Athens, and you might use ScienceDirect to reach exactly the same online publication. WhatamIdoing (talk) 00:32, 12 November 2010 (UTC)

On the first sentence:

• A link to a publisher's own website is, almost by definition, not a link to Yahoo News. If a publisher's own website is unstable, I don't think there's anything we can or should do about that. I think we should give the link anyway and let the publisher figure out whether and how to stabilize their content and their links.
• A DOI could link to several places with different users getting different destinations from the same DOI (as I understand its mechanics), so a publisher's URL and a DOI for the same source may as well coexist in the same referent. They may be redundant for some users but nonredundant for other users.
• Registration-required URLs divide into those which make registration free and relatively easy and those making that hard or costly. If I visit an institution and use their database access and then provide the resulting link in Wikipedia, it turns out you can use that link only if you have the institution's registration, and you almost certainly don't and probably can't get it. This has led to Wikipedia editors deleting my links as inaccessible to most WP users, while keeping the hardcopy bibliographic data related to the URL and the main text statement it supported.

On the second sentence: What's the alternative? If a source exists only in hardcopy, omitting the URL just because only some people can access it means not providing the main-text information it supports, because you'd be citing no source at all. It's permissible to cite hardcopy that's scarce, e.g., available only in a few libraries and only in person (if that's the most accessible you've got). So I think it should be acceptable to cite a URL that only some people can access. Better to cite an easily-accessible source the world can see, but a limited-circulation source is preferable to no source at all.

Nick Levinson (talk) 06:03, 20 November 2010 (UTC)

The discussion having come to an apparent end, I'm implementing the proposal and deleting the content of the draft, both momentarily. Thanks. Nick Levinson (talk) 07:45, 14 December 2010 (UTC)

WP:CITEHOW says:

"You should follow the style already established in an article if it has one; where there is disagreement, the style used by the first editor to use one should be respected."

Now where and when does this apply? As far as I understood it this is a rule-of-thumb for disputes in which person A changes the citation style, person B reverts it, and both end up in lengthy discussion (or worse, like edit war). In my opinion, in those cases WP:CITEHOW applies, ending the dispute by keeping the "established style". But how about person A's original modification in case of no opposition? Is this permitted (per WP:BRD), or is it prohibited (per "the established style is cemented for all eternity, no one is allowed ever to change it again")? Does person A violate WP:CITEHOW by changing the style in the first place, or is this permitted as long as he/she doesn't start an edit war, or violating WP:3RR? —bender235 (talk) 23:52, 13 December 2010 (UTC)

(This is related to the section I started below.) My opinion, based on following this page, the MOS, and arbcom decisions, is that editors should not make large numbers of stylistic changes from one MOS-approoved style to another. This is the same as the principle behind WP:ENGVAR: when there is no broad agreement that one style is preferable over another, the established style should just be left as it is.

It can be appropriate to change styles on rare occasions, such as during FA reviews, when there is clear consensus among the article's active editors that a certain style is preferable. But these situations are about specific articles that are being heavily edited for some reason. Articles that have been stable with an established style shouldn't be changed just for the sake of moving to a different optional style, and in particular editors who have never edited an article shouldn't do so just to switch from one MOS-approved style to another. — Carl (CBM · talk) 00:55, 14 December 2010 (UTC)

I agree with CBM. My opinion is that it would be disruptive for A to make a habit of changing the citation style in articles (particularly when they have not significantly contributed to the articles), regardless of whether B fails to object. People are different, and some couldn't care if another editor changed citation or other styles, while others would be highly irritated. I think editors and guidelines should bear that difference in mind, and people should not routinely change articles in ways that are essentially invisible to a reader. If A becomes interested in some article and really wants to change the citation style, I would suggest posting on the talk page and waiting at least three days. Johnuniq (talk) 01:03, 14 December 2010 (UTC)

I tell you how this would work: you find an article with 50 refs and no columns and think about implementing {{Reflist|colwidth=30em}}. But instead of implementing it right on the spot per WP:BRD, you pin a request on the talk page that certainly no one will reply to for 30 days. I hope you realize this is deadlock prone. If one has to ask each of the article's owners before making a minor edit like this, it would essentially kill Wikipedia. It is exactly how Wikipedia does not work, and the exact opposite of WP:BRD, which instead should apply here. Because, sure, there will be articles where a majority of contributors does not like the proposed style change. But why not speed up the process of finding consensus by following WP:BRD? Allow anyone to intruduce a new citation style. If it sticks, fine. If it get reverted, fine, too. Either case is better than a lengthy discussion. —bender235 (talk) 01:23, 14 December 2010 (UTC)

You could say exactly the same thing about ENGVAR; the same rationale for keeping the established style applies to that issue as to citation formatting. We don't want to have numerous editors all buzzing around changing styles on articles they have never edited, and so our guidelines recommend against making the change in the first place. — Carl (CBM · talk) 01:35, 14 December 2010 (UTC)

"We don't want to have numerous editors all buzzing around changing styles on articles they have never edited"

Then should probably start a proposal to remove one of the Wikipedia:Five Pillars. The one that says you don't own an article. —bender235 (talk) 12:04, 14 December 2010 (UTC)

This has nothing to do with ownership but with common sense. We are working with a rather heterogenous set of authors and they need to get along. "Enforcing" personal style references is not really helping with that. Common sense suggests to stay away from edits to articles which are no significant or objective improvements but are likely to annoy other editors. As you can see from this discussion the main effect of those edits seem to the latter, hence editors are advised to refrain from them.--Kmhkmh (talk) 13:01, 14 December 2010 (UTC)

But how do you know whether it annoys other editors? What if they agree that using columns is an improvement? I interpret this rule very much like User:Unomi, who said it "was never meant to exclude the introduction of change, it was to avoid a tug of war if edit warring arose." I do not enforce my personal preference by making the original edit. I would if I was re-adding it again and again after others have reverted it. But that was never that case. And adhering to WP:CITEHOW to avoid disputes, when there actually is no dispute, is just absurd. —bender235 (talk) 17:01, 14 December 2010 (UTC)

Recently, my watchlist has exploded with multiple editors who change the way footnotes are displayed on large numbers of articles, by changing <references> to {{reflist}} and/or changing the column count of the {{reflist}} template. For an example of the scale of editing I'm talking about, see these contribs and search for the word "references" in the edit summaries there. I don't want to pick on that editor, who is not the only one doing this.

It has always been the case in the past that no particular style for displaying footnotes was recommended by the style guides, and so editors of each article could choose the style they wanted. This page and the main MOS discourage editors from going around changing articles from one style to another based on personal taste.

The issue of whether the usage of {{reflist}} should be standardized has been discussed at Template talk:Reflist recently, but relatively editors follow that talk page. One outcome of the discussion was that several editors disliked that allowed wider screens to show more columns, instead of always showing 2 columns. But that change, away from a fixed column count, is one of the changes I keep seeing on my watchlist (example).

I want to post here to get a broader range of opinions. Are changes like the ones I linked above appropriate? — Carl (CBM · talk) 00:45, 14 December 2010 (UTC)

JFYI: This is directly connected to the entry above. There are three questions that need to be adressed:

1. Is replacing {{Reflist}} with {{Reflist|colwidth=30em}} actually a "change of citation style"?
2. Does BOLD, revert, discuss cover these kinds of changes?
3. see section above.

—bender235 (talk) 00:55, 14 December 2010 (UTC)

It is certainly a chagne of style: it changes how the footnotes are displayed. If it was not a change, why make it?

Historically, BRD has not been construed to cover stylistic changes, such as citation styles and WP:ENGVAR, particularly not when the changes are made to hundreds of independent articles. Instead, the rule of thumb has been to maintain the established style. — Carl (CBM · talk) 00:57, 14 December 2010 (UTC)

The term "citation style" usually refers to things like APA style or MLA style. Footnote font size or columns are not what I consider part of the "citation style".

And in my opinion, the rule-of-thumb that says "keep the established style" applies if, and only if there's a lengthy dispute (e.g. edit war) that needs to be settled. But starting the debate by being bold and changing the citation style does not, in my mind, constitute a violation of any Wikipedia rule. —bender235 (talk) 01:11, 14 December 2010 (UTC)

I agree that this "argument" is one that should be settled by vote. I am with bender235 in thinking that this kind of edit does contribute to the article. Some reference lists are just too long for there to be a long line to arduously scroll down, therefore splitting the column into 2 is the best idea. It is not a style change based on solely my opinion (if it could be called stylistic, I think it's more practical than preferential), the greater majority of people's screens are large enough to easily contain two or three columns. User:Thecheesykid, Talk to the hand...! Or my user talk page...

(Just to underscore that editors disagree about this) I, for one, prefer to see my references in one column, because it makes it easier to pick out the the author names running down the far left side. In fact, if it were possible, I would prefer that they didn't "wrap" at all. Multiple columns are just too dense. ---- CharlesGillingham (talk)

This is not about anyone's personal preferences. It is about whether you should be allowed to replace {{Reflist|2}} with {{Reflist}} (if that's what you'd like to do) w/out asking anyone for permission, or whether the original style of the article is cemented for all eternity, whether it be {{Reflist|2}}, or {{Reflist|4}}, or whatever. —bender235 (talk) 01:27, 14 December 2010 (UTC)

Optional styles that are not required or prohibited by the MOS are, almost by definition, a matter of personal preference. There is plenty of discussion to show that some editors prefer one thing and some prefer another, but none is objectively better. It appears that the only justification you have for your changes is your personal preference. — Carl (CBM · talk) 01:38, 14 December 2010 (UTC)

Optional styles that are not required or prohibited by the MOS are, almost by definition, a matter of personal preference.

Yes, they are. And since no style is explicitly prohibited or recommended, it depends on the collective preferences of the article's editors whether {{Reflist}}, or {{Reflist|2}}, or {{Reflist|colwidth=30em}} suits the article best. And how to find out the consensus? By using bold, revert, discuss. If the change sticks, its the new consensus. If it gets reverted, the old style is still consensus and should be kept. That's how Wikipedia has always worked in every aspect. —bender235 (talk) 01:50, 14 December 2010 (UTC)

What you say sounds great, but in practice (I think), we are talking about editors who focus on going from article to article (where they have never made a significant contribution) then "fixing" the references. If someone has a medium-term interest in a particular article it might be reasonable to apply WP:BRD and start by altering the references. However, that's not what we are discussing. The topic here should be whether it is appropriate for an editor to routinely change articles to their preferred style (the answer is no, it's disruptive). Johnuniq (talk) 03:28, 14 December 2010 (UTC)

I agree in such heterogenous project, cleanup actions by individual authors should concentrate on things there's an obvious consensus that a cleanup is required.--Kmhkmh (talk) 12:48, 14 December 2010 (UTC)

I edit to two columns ({{Reflist|2}}) because shorter lines are easier to read and a 3-column layout is not good on small hardware. Shorter lines would also be good for body text (we know that from newspaper and magazine design and optimum line length is relative to type size) but shorter lines in the body are harder to arrange (to do it requires adding images, infoboxes, sidebars, and other things and too many of those adds to article size, challenging some browsers). I understand fixed-width specifications can be problematic on some platforms. I don't do mass editing of volumes of articles but do it to various articles I edit for other reasons. Multi-column layouts do look bad with very short reference lists but that's uncommon and presumably future article expansion likely fixes that. I wonder whether people who read certain types of article topics are likelier to read them on large monitors or on cell phones or on old hardware (I assume lower-income nations more often use smaller monitors). So far, none of my column changes have been criticized or reverted, but I'll keep this discussion in mind. Nick Levinson (talk) 04:17, 14 December 2010 (UTC) (Corrected the template display, since it erroneously didn't show the "|2": 04:27, 14 December 2010 (UTC))

We are discussing the reference list style, as opposed to the in-text citation style or the full citation style. See Help:Reference display customization for more help. ---— Gadget850 (Ed) ^talk 05:43, 14 December 2010 (UTC)

The most important rule on Wikipedia is that everyone can edit every article. Just because you "focus" on a couple of articles doesn't mean you own them, and just because you have never contributed to the article before doesn't mean your not allowed to edit it.

That being said, I (and everyone else) have the right to modify an article, whether I'm correcting a type, adding a new infobox, or implementing a new citation template. And everyone else has to right to revert it if he/she doesn't like the change. What I'm not allowed to is to impose my change the article by repeatedly re-adding it w/out further discussion. That's the WP:BRD cycle, and I always acted this way in my past 6 years on Wikipedia. —bender235 (talk) 11:24, 14 December 2010 (UTC)

"That's the WP:BRD cycle, and I always acted this way in my past 6 years on Wikipedia." - I'd hope that this largely involved article content (which is what BRD is clearly written for), and not the sort of mass implementation of stylistic changes at issue here. Rd232 ^talk 17:33, 14 December 2010 (UTC)

Objectively, {{Reflist|2}} is the standard way of making Reference lists multi-column; on occasion {{Reflist|3}} is used when the references are very short (eg lots of references to different pages of a book). Where the hell this colwidth thing comes from, I don't know, but I'm starting to think it should simply be removed from the template. At any rate I can't see any specific circumstances where it is preferable to the status quo, and there is certainly no justification for editors going around adding it willy nilly to articles they're not seriously involved in contributing to. Rd232 ^talk 08:58, 14 December 2010 (UTC)

Objectively, {{Reflist|2}} is the standard way of making Reference lists multi-column

Is it? When and where was that consensus established? —bender235 (talk) 11:32, 14 December 2010 (UTC)

It was established over time, by practice. Colwidth as an alternative to that is a recent innovation, not widely used. Rd232 ^talk 17:31, 14 December 2010 (UTC)

ROTL So this is all based on your personal perception of some unwritten ephemeral consensus? This is ridiculous.

Suggestion: start a proposal on Template talk:Reflist to remove that "colwidth" feature. Let's see how many people agree with you. —bender235 (talk) 18:32, 14 December 2010 (UTC)

Question: how long has (a) {{reflist}} (b) reflist|2 (c) colwidth option existed? Answering that should clarify some things for you. Rd232 ^talk 18:43, 14 December 2010 (UTC)

Oh, so it's about how long a feature existed that decides whether we should prefer it over another. Then how about switching back to <references />, but that certainly existed even longer! *lol* —bender235 (talk) 21:26, 14 December 2010 (UTC)

Your question was how we know that reflist|2 is "the standard way of making Reference lists multi-column". I gave you the answer. Rd232 ^talk 21:30, 14 December 2010 (UTC)

There's a proposal to mandate a certain number of columns over a precise number of refs, and of a certain number of ems, which is going to change the text of this guideline, being discussed at Wikipedia:Village pump (proposals)#Final proposal. Tijfo098 (talk) 12:09, 20 December 2010 (UTC)

Is there a way to bundle a bunch of named references into one reference? Like <ref><ref name="nytimes" /><ref name="time" /><ref name="guardian" /></ref> but actually working. - Kollision (talk) 16:22, 24 December 2010 (UTC)

Doesn't look like it. You can nest one reference using #tag:ref (see WP:REFNEST), but multiple refs render oddly:

{{#tag:ref|<ref name="ref1">ref1</ref><ref name="ref2">ref2</ref><ref name="ref3">ref3</ref>}}

^[4]

---— Gadget850 (Ed) ^talk 17:21, 24 December 2010 (UTC)

What would the footnote look like? Like Gadget's above? Or like this?

1. ^ref1
2. ^ref2
3. ^ref3
4. ^ref1; ref2; ref3

Not sure what you had in mind.

If you just want limit the amount of markup in the wikitext, consider list defined references.

---- CharlesGillingham (talk) 19:31, 24 December 2010 (UTC)

I presumed the OP mean per WP:CITEBUNDLE. ---— Gadget850 (Ed) ^talk 19:44, 24 December 2010 (UTC)

Yeah, I was looking to "cite bundle" a few previously used references together with only one number in the reference list.

1. Last, First. "Reference 1". Work. Date. etc.
2. Last, First. "Reference 2". Work. Date. etc.
3. Last, First. "Reference 3". Work. Date. etc.
4. • Last, First. "Reference 1". Work. Date. etc.

Last, First. "Reference 2". Work. Date. etc.

Last, First. "Reference 3". Work. Date. etc.

That means having to put a <ref named="namedref1" /> inside a <ref> tag (ref tag inside a ref tag) which the software doesn't allow. - --Kollision (talk) 02:52, 25 December 2010 (UTC)

It does work if you use a group, but that means you have to use two instances of {{reflist}}:

This is a nested reference.{{#tag:ref|<ref name="ref1">ref1</ref><ref name="ref2">ref2</ref><ref name="ref3">ref3</ref>|group=note}}

;Notes
{{reflist|group=note}}

;References
{{reflist|close=1}}

This is a nested reference.^{[note 1]}

Notes

---— Gadget850 (Ed) ^talk 04:19, 25 December 2010 (UTC)

References

Ok, thanks. - Kollision (talk) 12:46, 27 December 2010 (UTC)

Consider these (arbitrarily formatted) citations:

• Random Notes (September 1999) Rolling Stone.
• CNN (26 July 2006), AI set to exceed human brain power, CNN.com, retrieved 16 October 2007 {{citation}}: |last= has generic name (help).
• NRC (1999), "Developments in Artificial Intelligence", Funding a Revolution: Government Support for Computing Research, National Academy Press, ISBN 0-309-06278-0, OCLC 246584055

In each of these cases, the source has a "corporate author"; there is no author's name. What are the recommended ways to format these citations? How should the major citation templates support these? Do the citation templates need to be changed to support these? What should a parenthetical reference or shortened footnote for these sources look like? Finally, what is the proper term for these citations? (i.e. is "corporate authors" the right term?)

I am starting a parallel discussion over at Citation/core, which hopefully will focus on the templates. Here, I hope we can discuss what APA, Chicago, etc. recommend, and reach some consensus what we should recommend. ---- CharlesGillingham (talk) 19:24, 21 December 2010 (UTC)

<CollectiveName> is the XML tag used for such authors in Pubmed. See for example PMID 21115958 (then click on "Display settings" and pick "XML", and scroll down). However, the guidance in Citing Medicine refers to them by the term "Organizations as author" here (or "Organizations as author/editor" in other sections).LeadSongDog come howl! 22:30, 21 December 2010 (UTC)

Thanks. Anyone else? I was hoping I could get a set of choices based on MLA, APA, Chicago, etc. There's really no other citation experts here any more? ---- CharlesGillingham (talk) 07:25, 28 December 2010 (UTC)

With the Bundling Citations subsection, I was a little confused on how to write the wiki markup for an intended reference. In other sections, the page gives examples of markup. Could someone please try that for this case?

One way, which numbers references within the bundle, seems to be thus: <ref>Something overall about specific references:<ol><li>One reference.</li><li>Another reference.</li></ol></ref> I guess element classes are not necessary.

I imagine we can use <ref>Something overall about specific references:<ul><li>One reference.</li><li>Another reference.</li></ul></ref> to make an unordered list with discs, but I haven't tested that.

In either case, it seems items should not be written with leading asterisks or hashes, unless we want an asterisk or hash to appear literally, and that <br /> (line break) is unnecessary.

As to the further-information template on citation overkill, I assume it should be deleted, since it's not about bundling, and it probably should be incorporated into a sentence further up the page (and then removed from the See Also list as redundant). Thoughts?

Thanks. Nick Levinson (talk) 20:20, 1 January 2011 (UTC)

I don't see the point of making bundled citations into bulleted lists, and no advantage in making the whole thing into an exercise in HTML coding. Why is the page holding this up as the only way to use a bundled citation? Just write the sentences one after another, as normal text.^[1] Or use short notes.^[2] (With short notes, you should have a bibliography/list of references at the bottom of the page. That also has the advantage of giving the reader an easy-to-access overview of the sources used.) --Hegvald (talk) 06:24, 2 January 2011 (UTC)

Stylistically, that's not always good, because putting each reference on its own line makes the list easier to read and to scan for specific items. Bulleting separates them and subnumbering separates them and also eases referring to them. At first glance, the examples you give both seem to be about the sun's size, and a fast skimmer probably wouldn't notice that they should keep reading to find out about the sun's heat. Your suggested style is more compact, so it's a useful choice, but not the only one. Perhaps your method should be added to the guideline along with coding instructions for both methods, yours being easier to code. How does that sound? Nick Levinson (talk) 06:59, 2 January 2011 (UTC)

This comment in a discussion over at WT:V caught my eye. There are a couple of points there, and I think the one re Hosting vs. Publishing is relevant here. I've dithered over this when composing some cites. I've generally ended up as identifying the "publisher" (that named field in templated cites, and the field containing that info in hand-crafted cites) as something like "WhereIfoundit.com, originally published by OriginalPublisher". I nearly ran over to the talk page for {{tl:Cite web}} to suggest a host= or similar parameter there (and Host= for {{Citation/core}}) but thought I'd ask for comments here first.

Two thoughts: (1) perhaps it's better to cast this in terms of "Republisher" rather than "Host"; (2) this concept of citing republication of previously published material might deserve some attention on the project page. Wtmitchell (talk) (earlier Boracay Bill) 23:19, 31 December 2010 (UTC)

I don't really see that a web host such as Google Books, Scribd or Project Gutenberg is really different from a library or a book store. I know of at least one book in my possession, my local library and on GB, but where it is stored has no effect on suitability as a source. ---— Gadget850 (Ed) ^talk 23:44, 31 December 2010 (UTC)

That's not really what I had in mind. I do tend to name the original publisher when I cite using info from a book I've seen on Google Books, Project Gutenberg, or the like -- even when providing a link to material requoted online by the republisher. Hmmm... I'll need to noodle about this; perhaps an example of the sort of thing that's nagging at me will come up. Wtmitchell (talk) (earlier Boracay Bill) 11:14, 1 January 2011 (UTC)

Is there a style guide that includes the host in a citation? ---— Gadget850 (Ed) ^talk 13:24, 1 January 2011 (UTC)

It's discussed at WP:SAYWHEREYOUGOTIT#Convenience links, but it might be a good idea to spell it out the format for such links here... I would suggest the following:

• "Scholar, John Q., Book Name, Reliable Publishers, NY, 1985 (scanned copy hosted on hostwebsite.com)"

Blueboar (talk) 16:46, 1 January 2011 (UTC)

Why do we need a host? Is Google Books more reliable than Project Gutenberg? If a book is hosted on multiple sites, shouldn't we include all the hosts? Should we include libraries and other places to obtain the book? ---— Gadget850 (Ed) ^talk 17:39, 1 January 2011 (UTC)

Responses: "Why do we need a host?" A reference location is important to establish a link for researchers. "Is Google Books more reliable than Project Gutenberg?" Not necessarily, but any verbatim source is considered reliable for secondary and tertiary references. "If a book is hosted on multiple sites, shouldn't we include all the hosts?" Not practical to list all sources or sites; typically in academe, the first site/source is considered primary and is all that is necessary. "Should we include libraries and other places to obtain the book?" Again, not practical but listing "via" sources is useful, although typically in "dead-tree" publishing, only one source is thus identified. FWiW Bzuk (talk) 18:13, 1 January 2011 (UTC).

Various republishers, whether online or print, may have various motives for republishing something. Google Books motive seems to be that the book was found in a reputable library. Project Gutenberg's motive seems to be that enough volunteers could be found to enter and check the text. Dover's motive seems to be the expectation that enough paper copies will be sold to make money. So the reputation and motivation of the republisher can reflect on the reliability of the source, just as the motivation and reputation of the original publisher reflects on the reliability of the source. It is not just a question of whether the republisher can be counted on to provide a faithful copy. The fact that a book was republished recently can mean that an old book is still important today. Jc3s5h (talk) 18:23, 1 January 2011 (UTC)

I definitely see your point regarding the republisher attributing reliability to older texts, but I don't believe that is really relevant for a citation. For a book, all that is necessary is to identify the publisher and ISBN, unless there is some reason to believe the republished text is not faithful to the original (in which case just don't use that source). For an online text that is republished elsewhere, I typically use the format, ex) The New York Times via Otherwebsite.com. This identifies both the original source and the current publisher, both of which are relevant here. — Huntster (t @ c) 18:45, 1 January 2011 (UTC)

The original thread was at WT:V#Ancient self-published sources. I raised the question as at the time I had Somerhill House at GAR, which it later passed. Three sources I have used are self-published, but dating from 1766, 1830 and 1832. Copies of the sources are hosted at The Weald website. I included links to the scans of individual pages of the various books purely as a convenience. Had I not done so, the references would still have been valid. As the books are copyright expired, anyone is free to republish them, but IMHO that does not make them the publisher of the book, but merely a host as mentioned above. ISBNs are a recent thing, only coming in during the late 1960s. There are a vast many books that do not have an ISBN. Mjroots (talk) 13:49, 5 January 2011 (UTC)

Should I prefer the journal name as published or the current one?TCO (talk) 22:29, 9 January 2011 (UTC)

I would like some context for the question. Do you mean: A: When you refer to an article in a journal? Or B: When writing an article about a journal?

For A, I'd say the title as it was called when the article was published. Nobody will have a problem finding it in a library catalogue, unless it it has a name shared by other periodicals, in which case some additional context may be needed. For question B, My suggestion would be to use the more recent name, unless it is far better known by an older and more long-lasting one. --Hegvald (talk) 23:32, 9 January 2011 (UTC)

(edit conflict) Presuming you are referring to the name used in a citation, then use it as published. If there is an article on the journal, then you can link it, creating redirects as needed. For example, Artillery Trends redirects to Field Artillery. If there is no article and you really feel it necessary to explain, I suppose you could do something like |journal=Pin Collector (currently Pin Monthly). ---— Gadget850 (Ed) ^talk 23:37, 9 January 2011 (UTC)

ec: I meant A, a reference citation within Painted turtle to Tanner's expedition. The publishing journal was originally called Great Basin Naturalist, but has since changed to Western Naturalist (or some such). So when you look at the pdf, or bound version, of course it is the former. When you look at the website (where I direct people as pdf is not linkable except going there), you see the journal name with new version and even a citation as new periodical name. Obviously situation will happen lots of times, so would like a general rule. Do you know how this is handled normally in academic publishing advice? (tried googling, but didn't find it exactly, but I'm sure it's the sort of thing that has been adressed. Just wondered general practice.) Oh and I did the opposite of what you say (so would need to go back and put old name in, now)TCO (talk) 23:41, 9 January 2011 (UTC)

ec: OK. I will go change it back. I think I messed this up one other place, but don't tell anyone as I have 160 refs and can't recall where I did that! TCO (talk) 23:41, 9 January 2011 (UTC)

Title as published is indeed the general rule. Shimgray | talk | 23:54, 9 January 2011 (UTC)

About a month ago, User:Hegvald and I had a debate on whether references should be named (and refered to multiple times), like I did with this edit (using WP:AWB). Hegvald disagreed (and reverted), explaining on my talk page that "real-life academic contexts outside Wikipedia" are not using that kind of a reference system. As of now, and as far as I know, there is no guideline that recommends (or prohibits) the use of named references. Therefore, I'd like to discuss this here. And finally, what about WP:NOTPAPER in this context? —bender235 (talk) 11:23, 3 January 2011 (UTC)

The established style should be kept, since named references are neither prohibited nor required. You have been told this numerous times... — Carl (CBM · talk) 12:32, 3 January 2011 (UTC)

And again, there is no such thing as an "established style". On Wikipedia, nothing is in stone. You have been told this numerous times. —bender235 (talk) 14:42, 3 January 2011 (UTC)

That includes your push for your preferred style.--Kmhkmh (talk) 14:45, 3 January 2011 (UTC)

We actually do consider that the established style for various factors: US vs UK spelling, date presentation, cite vs Harvard reference style, and the like; these are based on what the article creator, or later through local page consensus, of what to use. It is bad form to change an entire's page reference style before checking with its creator or its talk page. --MASEM (t) 15:14, 3 January 2011 (UTC)

Wikipedia is the encyclopedia anyone can edit using any style, as long as they establish the style first in the article or gain consensus to change it. ---— Gadget850 (Ed) ^talk 15:22, 3 January 2011 (UTC)

So consensus is no consensus. Or basically WP:IMADEIT, first in time, first in line, right? —bender235 (talk) 15:58, 3 January 2011 (UTC)

It's not "no consensus" but it is original author - unless of course there's something completely broken per MOS guidelines. For example, if I created a page on a British actor that only started in British works and primarily known within the UK - and used US spelling and dates, that would be changed. But stuff like references aren't affected by article type, and if the original author used a style that is completely in line with MOS on citations, it is bad form to change that without seeking the author's input firtst. --MASEM (t) 16:06, 3 January 2011 (UTC)

I understand this, as far as mere style changes are concerned, but I considered this a "fix". As does WP:AWB, because this was semi-automatic edit applying one of AWB's "general fixes", DuplicateUnnamedReferences. If that was not correct, then this feature should be removed. —bender235 (talk) 16:20, 3 January 2011 (UTC)

WP:NOTPAPER, as far as I can see, is mainly about the space restrictions of a paper encyclopaedia: no reason to exclude topics because they are no longer current or only likely to be looked up by people in some other place than the area where the [paper] encyclopaedia is marketed.

The arguments for named references seems to be pretty much that it saves space by reducing the number of footnotes. How is that needed when Wikipedia is "not paper"? --Hegvald (talk) 15:47, 3 January 2011 (UTC)

Having like a dozen references a la Miller (2000), p. 100. may be useful on paper, where there is a (partial) reference list on every page, but not on Wikipedia, where there is a single reference list at the end of the article. The question was whether this is merely a style change (i.e., a neutral change from one style to another [which, of course, should be avoided]), or rather a fix (i.e., introducing an improvement). In my honest opinion it is the latter. —bender235 (talk) 16:00, 3 January 2011 (UTC)

That is certainly a style change: the article did not have named references, and then you changed the article to use them. To make it a "fix" you would need to be changing away from a forbidden style, or towards a recommended one. The use of named references is neither recommended nor forbidden, so it is a matter of style. Our general rule is to keep the established style. — Carl (CBM · talk) 16:02, 3 January 2011 (UTC)

(edit conflict) To make it a "fix" you would need to be changing away from a forbidden style, or towards a recommended one. The use of named references is neither recommended nor forbidden.

Exactly. Therefore, as said in the original comment, I came here to start a discussion on whether named references should be recommended (to avoid duplicates). —bender235 (talk) 16:20, 3 January 2011 (UTC)

I don't really see what you mean with "on paper, where there is a (partial) reference list on every page". In paper publications, what you may find at the bottom of the page is an apparatus with footnotes, not a reference list. The list of references (bibliography) is usually found at the end of the book, or sometimes at the end of the chapter or article. Sometimes the footnotes are found at the end as well, in which case they are called "endnotes". How is this relevant in this context? Whether you use "named" references or normal, short references, they are found at the bottom of the Wikipedia page. --Hegvald (talk) 16:17, 3 January 2011 (UTC)

It's not a style change no, it's a fix. There's no reason to have 5 <ref>J. Smith (2009) "Famous Bobs of the World", Springer, ISBN 0123456789</ref> instead of one <ref name="Smith2009">J. Smith (2009) "Famous Bobs of the World", Springer, ISBN 0123456789</ref> with four <ref name="Smith2009"/>. Doing otherwise just clutters the reflist. Headbomb {talk / contribs / physics / books} 16:47, 3 January 2011 (UTC)

That is the reason to use short references, not named references. I don't think anyone has argued that one should have publisher, ISBN and all that jazz in every footnote. That is a strawman argument. --Hegvald (talk) 16:58, 3 January 2011 (UTC)

I would agree that adding ref names is not what is meant by a "style change" in the sense that it is meant. Harvard, APA, Chicago, MLA, etc. are the styles referred to as being established by authors and/or local consensus and that should not be changed. Removing duplicate references within any of the given styles is cleanup. There is no reason for a single citation to be listed more than once, rather than just referred to as many times as necessary in the article. Jim Miller ^{See me | Touch me} 17:46, 3 January 2011 (UTC)

The tradition is that all changes to the ways that references are formatted or displayed are classified as "style change"s. When a certain system is neither recommended nor discouraged by the MOS, the rule is to keep the established method. This applies both to the outward appearance of the article and to the way that appearance is achieved with underlying wikicode.

The point of this is that, because neither version is recommended, someone else could go through and change lots of articles in the other direction (getting rid of named references). The only rule that prevents that sort of back-and-forth changing is the rule to preserve the established style. — Carl (CBM · talk) 17:56, 3 January 2011 (UTC)

There is a style decision between A and B below, and A is justifiable. Among other things, the references appear in exactly the order they appear in the text, which makes them easier to follow, and B is not an obvious improvement. While I'm sure some editors don't even notice this choice, there are articles written with style A as a conscious choice, so if someone objects, best to leave it alone. Gimmetoo (talk) 18:01, 3 January 2011 (UTC)

Example A: There is no reason^{[a 1]} that anyone ^{[a 2]} needs to "remove"^{[a 3]} so-called "duplicate"^{[a 4]} references.^{[a 5]}

Example B: There is no reason^{[b 1]} that anyone ^{[b 2]} needs to "remove"^{[b 2]} so-called "duplicate"^{[b 3]} references.^{[b 1]}

A question to ask: are both styles existing styles from the published style guides (Chicago, Shrunk & White, etc.)? I agree they are different styles, both with advantages and disadvantages, and thus the type of thing that we use "first editor discretion" on, but if one of the styles is not a standard style, it is probably not our place to use it. --MASEM (t) 18:05, 3 January 2011 (UTC)

The "A" style, where no footnote number is used more than once, and footnotes/endnotes are numbered sequentially, is the standard style in print (when the number point to footnotes or endnotes). I don't know of any style guide that allows for using the same footnote/endnote number more than once, out of order, but at the same time I don't know that there is not one. — Carl (CBM · talk) 18:17, 3 January 2011 (UTC)

'B' looks like a form of Vancouver system, where refs are numbered in order of appearance, and reused. This style is used by a number of journals in science, engineering and medicine. Gimmetoo (talk) 18:23, 3 January 2011 (UTC)

So assuming both CBM and Gimmetoo are correct, these are both acceptable styles, and thus edit warring to try to convert the "first editor discretion" to a different form is intolerance. Surely discussion and consensus building to suggest a different way is proper. --MASEM (t) 18:29, 3 January 2011 (UTC)

@Gimmetoo: thanks. Do those point to footnotes/endnotes, or do they directly point to a numbered list of references? It would be nice to have an example of the former, where footnote numbers are reused, in print. — Carl (CBM · talk) 18:30, 3 January 2011 (UTC)

'B' would typically point to a numbered list of bibliographic references. Here's an example [3] from a journal, though it's different because the refs there were alphabetized and used by number in alphabetic order. Gimmetoo (talk) 18:57, 3 January 2011 (UTC)

That is the system we use in mathematics. — Carl (CBM · talk) 19:05, 3 January 2011 (UTC)

For a sequential version, I found [4] (full text [5] with subscription). Refs numbered in order of appearance in main text, and refs #10, #17 and #21 get reused early. Gimmetoo (talk) 20:25, 3 January 2011 (UTC)

"I don't know of any style guide that allows for using the same footnote/endnote number more than once, out of order."

On paper, no style guide does. Because one would have to turn pages back and forth to find a specific footnote. But then again, Wikipedia is not a paper encyclopedia. On Wikipedia, one clicks a link to see a highlighted footnote at the bottom of the page. —bender235 (talk) 18:51, 3 January 2011 (UTC)

So, where do we stand now? Does DuplicateUnnamedReferences have to be removed from the WP:AWB feature list? —bender235 (talk) 22:50, 10 January 2011 (UTC)

They are not footnotes, they are endnotes. Really think about it functionally. If someone said they were endnotes, could you prove him wrong? With footnotes, in a document, you have different bundles on different pages within an individual document. For us, our individual document is the "article". And the citations are all in a mass at the end. One section only. And it is very NORMAL to repeat numbers of an endnote. Check out any science journal or even just MS Word. The confusion comes from calling something here a "footnote", but functionally it is an endnote and we use it in that manner. Including repeated numbers.  ;-) TCO (talk) 23:46, 10 January 2011 (UTC)

Really, TCO, think about what you would call these notes on a one-page document. They're simultaneously both "footnotes" and "endnotes" too, aren't they?

Wikipedia articles are a single page with respect to the intended medium (computer screen), no matter how that single page might map to some other medium. WhatamIdoing (talk) 04:28, 11 January 2011 (UTC)

• I greatly dislike it when people come along, as they sometimes do, & condense the odd identical refs into a refname (which I never use myself). This is partly because the article is usually still being worked on & further refs may well be added at the same point, in which case the refnames will be unscrambled anyway. Let them repeat, within reason. Johnbod (talk) 00:29, 11 January 2011 (UTC)

Let's say I add a citation, then I want to link the same numbered citation again at a different point in the article, how would I do this? --Rayne117 (talk) 01:23, 10 January 2011 (UTC)

Bob,^[1] Gulliver,^[2] and Jane.^[1]

Like this. <ref name="Foobar">Foobar</ref> the first time, <ref name="Foobar"/> the other times. Headbomb {talk / contribs / physics / books} 08:16, 10 January 2011 (UTC)

Note that the name value does not have to match the citation, although it's helpful to make it recognizable, for when someone edits a page. The name is easier to write if it has no spaces, but you can use hyphens and creative capitalization, such as "theBook" or "the-book". Nick Levinson (talk) 09:58, 10 January 2011 (UTC)

See WP:REFNAME. ---— Gadget850 (Ed) ^talk 12:03, 10 January 2011 (UTC)

I would like to incorporate some more info into the "Inline citations" section as follows:

Inline citations

An inline citation is a line of text identifying a source that is added close to the material it supports, offering text-source integrity. If a word or phrase is particularly contentious, an inline citation may be added next to it within a sentence, but adding the citation to the end of this sentence is usually sufficient, so long as it is clear which source supports which part of the text. Likewise, instead of citations after each sentence, one citation can be used for group of consecutive sentences in a paragraph supported by the same source. The citation can be placed at the end of the last sentence of the group of sentences, the introductory sentence preceding a block quotation, or the topic of thesis sentence of a paragraph. Except it is recommended that citations be placed next to direct quotations or statements that are particularly startling or contrary to "common knowledge" (i.e., likely to be challenged). Note that if text citing other sources is inserted inside the text supported by one citation, citations for the existing text will have to updated. The likelihood of such a future edit occurring and how well the text will be monitored, may be considerations in deciding whether to use this method of citation.

Editors are free to use any method for inline citations; no method is recommended over any other. Two common styles of inline citation used on Wikipedia are clickable footnotes (<ref> tags) and parenthetical references.

Comments:

• I've re-written the sentences I wanted to add and put them in bold so it's easier to distinguish from the existing sentences. -Verapar (talk) 01:45, 13 January 2011 (UTC)

The sole important thing, and this is really common sense (however uncommon that may be), is that it should be clear enough where all parts of a statement or claim comes from for a reader to be able to check that statement. A proliferation of extra footnotes in the text does not help. Nor does it protect against people adding unsourced claims in disagreement with the already cited sources. Those inclined to do so will do so anyway. The guideline is fine as it is on this point. --Hegvald (talk) 17:47, 5 January 2011 (UTC)

• I added two important issues that people should consider when considering using one citation for consecutive sentences. One may be "common sense" to you, but as seen, a number of people disagree (and one even if the need for such citation) and I think it's also easy to forget. Also, if you are not able to watch the text, it can be easier to identify the "unsourced claims" if each sentence is cited or if properly sourced sentences are inserted, to continue to have all the sentences correctly cited. It is also an important reminder to update the citations for other sentences when inserting new ones, regardless of whether the text was originally your's. So it's helpful whatever the 'inclination' of future editors. Also as noted above, I also replaced the word "paragraph" with "group of consecutive sentences" as the cited text need not always be a whole paragraph. -Verapar (talk) 19:38, 10 January 2011 (UTC)

Direct quotes and startling statements aren't required to have inline citations directly after them, Verapar. Many editors do that, and many don't. What matters is that the reference is in the next footnote—at the end of the paragraph at the latest—or is in some other way clearly identified as supporting that text. And citations inside block quotes are fine. SlimVirgin ^{talk|contribs} 21:00, 5 January 2011 (UTC)

• I was first advised of the exception for direct quotes/likely to be challenged statements by User:Donald Albury and if you look above and at ::*Wikipedia_talk:Manual_of_Style_(footnotes)#final_proposal_for_.22Consecutive_sentences_with_the_same_source.22,
• Wikipedia_talk:Manual_of_Style_(footnotes)#New_subsection_for_consecutive_sentences_using_the_same_source

there are a number of people who agree inline citations should be required after them. And the exception for direct quotes is also a requirement according to these external sources:

• unc.edu page on citations in Chicago manual of style
• easybib.com page on APA style
• and I also replied to you with a good reason for having inline citation directly after "direct quotes or startling/likely to be challenged statements" in WP:FN. I've replaced the word "required" with "recommended."
• "A reason that a number of people recommended the citing of direct quotations or 'startling' phrases, is that they might be challenged and having the footnote after it makes it easier for people to verify the info before putting a ref. needed template, etc i.e. so they can be certain what source is supporting the info. As noted, it's possible someone may have inserted sentences without updating the footnotes/citations. So it may be a recommended practice for WP to have footnotes after direct quotes/'startling' phrases, although not usually required outside of WP (and yes I was already am aware of this).-Verapar (talk) 19:03, 10 January 2011 (UTC)"
• Yes I think citations inside block quotes are fine (so I changed that sentence above), but often omitted and not necessary i.e. you can have one citation for the entire quote in the sentence before the blockquote when the text inside the blockquote was from one source, as the blockquote indicates visually the text inside it is a quote. -Verapar (talk) 19:38, 10 January 2011 (UTC)

Are these requirements perhaps taken from WP:DYK, which does require its 'startling' hook to have an inline citation at the end of the sentence? WhatamIdoing (talk) 05:39, 7 January 2011 (UTC)

• Thanks for continuing to follow this discussion. I have provided my own explanation for recommending the inline citation "direct quotes or startling/likely to be challenged statements" in WP above. I've replaced the word "required" with "recommended." -Verapar (talk) 20:14, 10 January 2011 (UTC)
• The exception for direct quotes is also noted as a requirement offline according to:
• unc.edu page on citations in Chicago manual of style
• easybib.com page on APA style -Verapar (talk) 22:41, 12 January 2011 (UTC)

I think these discussions further illuminate the issues discussed in my proposal. -Verapar (talk) 19:38, 10 January 2011 (UTC)

• There is support for citations to follow immediately after direct quotes & likely to be challenged statements. This particularly due to the nature of WP, where the text can be edited/challenged by different readers. This standard aids in ensuring that sources can be readily verified for these statements.
• From the discussions, that new text can be inserted into the text supported by only one citation is clearly a factor for people in considering whether to use this method of citation over citing each sentence. So this consideration should be noted.

-Verapar (talk) 22:41, 12 January 2011 (UTC)

Editors here may be interested in the discussion at Wikipedia talk:Citing sources/example style#why_not_standardize_on_one_format.3F, an effort to impose a one-size-fits-all citation style on every article. My overall impression from the discussion (started by TCO, but expanded beyond his/her original questions) is that the goals are to ban general references, to require the use of <ref> tags (banning WP:PAREN and all other forms of WP:Inline citations), and to require the use of citation templates.

As this discussion about this guideline's contents is (oddly) happening on a subpage rather than on this page, I am providing the link here and hope that you will all join the discussion there. WhatamIdoing (talk) 04:30, 11 January 2011 (UTC)

The intention is to have a common style. We might or might not require a particular tool. Really I think the tools would be better and work better together if they were all working to produce a given style, rather than implicitly selection one themselves (practically I think just taking our base as the cite toolbar makes sense, but I don't really care as long as we pick something).

As it is now, even when contributors want to be using a common style, they face difficulties. Also, it's not about requiring tools per se. As it is now, even if I want to do everything manual (and I might, I donno, might be faster), it's not easily apparent what the styles are themselves within the templates. They are not well described as citations themselves in the manner that one would see in a guide to technical writing, how to write a research paper, or the Notice to Authors of a journal. If we had one style, we could also better describe what the style is trying to do. As it is now, templates are probably the best option, but young (and even experienced) editors fall prey to thinking of the thing as a black box and losing connection with the output, and then run into all kinds of issues when you really parse the output.

P.s. Not sure why it is on a subpage either. I might have dorked that up from the beginning. Please don't hold that against the concept which seems pretty logical, especially in a journal where people collaborate on the same articles. We talk about having consistent format in an article but almost never do. Only time I see it is when an article is controlled by one person, or after lot of pain in Featured Article process. I'm not trying to control anyone. I just want good content and format for readers. And want a tool to make producing that easier. There is so much work to do.TCO (talk) 17:29, 12 January 2011 (UTC)

• See below: "#Citing phrases for NPOV phrase-coverage tool". -Wikid77 11:15, 18 January 2011 (UTC)

18-Jan-2011: A better "tool" to help with sourcing an article would scan outside reliable sources for "key topic phrases" rather than just formatting citations. Also, at the present time (January 2011), the {Cite_web} & {Cite_book} templates need to be rewritten to use less of the MediaWiki preprocessor resources: even though an article can use 1 million tiny templates, only 550 {Cite_book} transclusions can be made. Here's the point: it is not possible to have a "List of news articles about subject XX" which has 800 citation templates, unless the list is split into multiple pages, or {Cite_web} is rewritten for better performance, such as combining the contents of {Citation/core} to allow twice as many citations, allowing 1,100 total (rather than just 550 uses of {Cite_web} or such). Meanwhile, a "tool" to help citations in articles would scan for related issues about the subject:

News phrase[a]	Person: Natalee Holloway	Person: Amanda Knox	Person: Charles Manson
Google hits for name	371,000	763,000	1,470,000
"false arrest"	2,140	1,260	4,300
"false confession"	2,750	7,090	3,850
"police hit her" (or him)	6	6,340	9
"new trial"	20,900 (5.6%)	49,700 (6.5%)	40,700 (2.7%)
[a] - Data from Google Search on 11:13, 18 January 2011 (UTC).

We need a tool which can adjust the search-engine results, such as the above from Google, perhaps scaling to omit blog counts (by some likelihood factor, not by exact numbers). Also, note the counts of subject-name, "Charles Manson" has twice the hits of "Amanda Knox" which has twice the hits for "Natalee Holloway". Then, the counts need to be adjusted as percentages, in accordance with WP:NPOV neutrality as "proportionately" reflecting the coverage in sources. The highest counts of those phrases were not for Natalee Holloway, but, instead, Charles Manson had 4,300 for "false arrest" which should be considered a significant topic for NPOV coverage of Manson (whether the arrests were later justified). Then, Amanda Knox ranked highest in "false confession" (7,090), "new trial" (49,700) as being re-tried for murder in the Italian appeals court, and "police hit her" (or "him") as 6,340 compared to just "6" webpages about Holloway and "9" for Charles Manson. Hence, for NPOV coverage of Amanda Knox, the topic "police hit her" (whether claimed or substantiated) would be a topic to cover in the intro of an article, as seemingly such a huge issue as reported, neutrally, by the tool. Also, note how a tool would not pre-censor to ban the phrase "police hit her" as being a WP:BLP vio against some police officers, because it is just data, whereas an article would need to add how police denied the allegation during court hearings, for NPOV balance. Now, compared to article citation templates, an obvious implication might be to expand citation templates to include the "phrase" concept, with phrases=xxxx to be listed/indexed as hidden text inside a source citation template, then another tool could count and report the coverage of those phrases within an article's citations. Hence, expect several citations in an article to index the key topic "police hit her" because it is so common in sources. As articles are later edited, and censored for POV-pushing, plus removing reliable sources to make topics seem unfounded, then the proportional usage of topic phrases could be seen as being slanted. Thus, 2 revisions, far apart, could be compared to see where information was censored or added to affect the relative percentages of topic phrases in an article. To help this, article citations should, internally, list the "key topic phrases" which a tool would search. We need to have numeric data to quantify the level of POV-pushing or hopefully, NPOV-balanced coverage of major topics, compared with their source citations. Currently, we have edit-counters which reveal some users editing an article "400 times" as an indication of unusual activity. More about this later. -Wikid77 11:13, 18 January 2011 (UTC)

Am curious about this: is there an obvious way of telling our readers where an offline source (map, journal, less popular books) can be found? There doesn't seem to be a "remarks" field in the standard citation templates. --Deryck C. 14:14, 18 January 2011 (UTC)

• |postscript= seems to be pretty universally implemented. I find with the Cite foo series, |postscript=. Looks best with a period and a space at the beginning and a terminal period. Fifelfoo (talk) 14:17, 18 January 2011 (UTC)

If by "where", you mean the location, then no style guide I am familiar with includes this information.

The |postscript= parameter is used to adjust the citation terminator where the terminating period may not be appropriate.

---— Gadget850 (Ed) ^talk 15:03, 18 January 2011 (UTC)

You can always add this kind of information after the citation. E.g.,

*{{cite book|last=Doe|first=John|title=Doe's book}} Available through some obscure process.

• Doe, John. Doe's book. Available through some obscure process.

--- CharlesGillingham (talk) 18:14, 18 January 2011 (UTC)

Is WP:ASL still reliable and usable? As it was added 4 years ago, so browser question is not a problem any more... What do you think? --WhiteWriter ^speaks 15:51, 12 January 2011 (UTC)

I think you're right the text needs to be updated.

There are other issues with scrolling reference lists besides old browsers. One of them is that it is no longer possible to just scan down the article to look for a reference: you have to scroll the reference list into view, and then separately scroll the reference list to show the reference that you want to see. On the other hand, there is no real limit on the vertical height of an article, and the browser already has a scroll bar, so the benefits of a scrolling reference list seem minimal. — Carl (CBM · talk) 18:07, 12 January 2011 (UTC)

This is just my two cents, but I think that Wikipedia already has too many optional bells and whistles. Scrolling lists are completely unnecessary. (All browsers have scroll bars, after all.)

Articles are supposed to be easy for anyone to edit. The wikitext of all articles should be as simple and similar as possible. A Wikipedia article isn't somebody's MySpace page. No bells, no whistles, no scrolling lists. This is feature creep. ---- CharlesGillingham (talk) 00:00, 15 January 2011 (UTC)

To better cross-check citations with text, consider just viewing an article in 2 browser windows: one window to scroll the upper text, and the 2nd window can be used to scroll through the bottom references. The same concept can be used during editing: 2 copies of the edit-window, one for preview, the other window to make changes (see essay: "WP:Advanced article editing"). I don't mind new features, in moderation: a scrolling region for a map is sometimes helpful: a user has a small map-box in an article as a viewport to see just a magnified portion of the total map, in cases where the map labels are not legible unless magnified. The user scrolls across the map. Also, I can understand an article might have several scrolling lists, stacked together, such as a multi-part list, all viewed in one window, as well as the current feature of sortable tables where the user clicks on a column to re-display the table re-sorted by column. -Wikid77 11:38, 18 January 2011 (UTC)

The best advice I ever got was to keep it simple, stupid!

Many of Wikipedia's user are not as sophisticated as some its editors.

Nor are all editors programmers — nor have many of them any reason to be.

KISS! — Robert Greer (talk) 16:30, 18 January 2011 (UTC)

• The problem with "Keep it simple" is that simplicity is in the eye of the beholder: simplicity is relative to each person's viewpoint. A person viewing a table of 275 nations, might think it is "simple" to click to sort the nations by the year they were established, even though sorting is, internally, a very complex feature. A person reading about the prosecution of Amanda Knox, in Perugia (Italy), might think it is simpler to have a short, terse article about the case, whereas another person might think including the forensic evidence is simpler, to show Guede matched DNA samples from the bedroom, body, clothing and handbag at high RFU peaks, while Amanda and her 2-week boyfriend, Sollecito (supposed as accomplices in the same room), matched only 1 small sample (each) at low RFU peaks considered inadmissable evidence in the UK or U.S. (no reliable samples of their DNA were found in the room). Sometimes, "more is less" and by writing a full, extensive article, the reader's task is simplified by having Wikipedia as their "one-stop-shop" to have all their major questions answered about a crime or other subject. We try to handle multiple people's idea of "simplicity" by having the Simple English Wikipedia to provide articles in simplified language (and typically much shorter text), while allowing large, complex articles in the full English Wikipedia, where some users think covering the whole subject simplifies their search for information. -Wikid77 23:24, 23 January 2011 (UTC)

I think your examples are a little off the point. This is "Citing sources" and here we are discussing scrolling lists in regard to reference lists (which are used primarily in articles that have parenthetical references or shortened foonotes). In the particular case of reference lists, I feel that the wikitext should be as simple and standard as possible, to help newer editors to add references without a steep learning curve. Scrolling lists make it (a little more) complicated, and as such I am against them. This is the issue I think we are discussing.

Having said that, I agree with you: in your first example a scrolling list would be okay. Your second example is pretty far off the point of scrolling lists, indeed off of the entire issue of "simplicity of wikitext". ---- CharlesGillingham (talk) 18:36, 24 January 2011 (UTC)

I'm trying to source a court decision using the {{cite court}} template. Unfortunately, the title is long. Extremely long. On the order of 3,000+ characters long. Is it OK if I truncate the title myself, ex to something like Jewel, et al v Hughes Air Corp, et al? Or do I have to use the whole really long ugly title? (see title) --Mûĸĸâĸûĸâĸû (blah?) 15:57, 19 January 2011 (UTC)

I'm a nonlawyer but I know this one. For U.S. law, conventions for citing court cases include:

• When more than one case is in the caption of one court opinion (meaning a court has written an opinion about multiple cases, not that the court has mentioned several cases within the opinion), you cite the first case in the list.
• When there are more than one party either before or after the "v." ('versus'), you name only the first party on that side of the "v." E.g., the case of Smith and Wing v. Mbutu and Jones, becomes Smith v. Mbutu, the exception being if "Smith and Wing" is actually one party, such as if it's a store named "Smith and Wing".
• Some abbreviations, but not all, may be dropped. Most commonly, "et. al." is dropped. However, "Ex rel.", like "Ex parte", is kept.
• People's first or given names are dropped. However, I don't know what's conventionally done when the given name is the person's last name, as with many Asian names, and, if the court hasn't told you, it is probably safer to give the name in full when uncertain, especially since many Asians who migrate to the U.S. reverse the order of their names thus putting the given name first, without telling you.
• People's titles are dropped unless you drop the person's name and give the short title; ditto if someone is a party as the head of a government agency, but this does not apply to anyone less than the full head. If you're writing about a lot of cases with the IRS, headed by the Commissioner of Internal Revenue, who's now one Douglas H. Shulman, you may write about Tripp v. Shulman, Tripp v. Commissioner, or Tripp v. IRS; the latter two are usually better because agency heads come and go while cases remain. Be consistent.
• Organizational indicators can be dropped unless the result is to make the name seem like it might be an individual's. If a party is General Motors, you don't need "Corp."
• Some organizational words may be abbreviated; commonly, "Railway" becomes "Ry.", but only use standard abbreviations. If you haven't seen it before in case names, probably don't abbreviate an organizational word.
• Political units get their usual abbreviations, such as "U.S.", but if the party is, say, New York, Inc., then it's not the political unit, and you'd write it as "N.Y., Inc."
• Addresses are omitted. The fact that a corporation is in a certain state doesn't much matter.
• The common middle word "versus" becomes "v." (not "vs." nonlawyers use). Beware of "v." that is part of a party's name, as might happen with some German-origin names. The most formal usage italicizes the case name except the "v."; when nearly as formal, italicize the entire case name, including the "v."
• Terms like "Defendant-Appellee" are dropped.
• Rarely, two cases have the same name after you've shortened it. If they can't be told apart by which court decided it (typically cases start in the bottom court, maybe go up, and thereafter never appear again, but occasionally a case is sent back down and a new opinion is issued by a lower court that had published another opinion earlier), number them "I", "II", and so on, as in Thesata v. Quimore I and Thesata v. Quimore II, although I'm not sure if the roman number is italicized when the "v." is not.
• Divorce and some other cases may have names like Thompson v. Thompson, which means it's not a good idea to write about what Thompson said. Instead, you'd write about Jane Thompson's position or the judgment for the plaintiff.
• Case numbers are omitted unless no publication citation is available. E.g., "No. 74-3283" is omitted because you can specify "557 F.2d 759".
• Dates are limited to years unless month and day matter to your discussion or the case is very new.
• Dates are limited to the date of the last major stage in the case when you cite multiple stages in one case. For instance: Darth v. Vader, 302 F.Supp., rev'd on other grounds, 187 F.3d 901, cert. granted, 588 U.S. 133 (2003).
• The court name is abbreviated and sometimes omitted, with the assumption that the publication being cited for the opinion or judgment indicates the court at least in part. Typically, the reporter (publication) for the highest court of the Federal government or a state covers only one court, in which case the court does not need further identification; Galahad v. Arthur, 302 U.S. 599, 99 S.Ct. 1180, 26 L.Ed.2d 263 (1970) already conveys that the U.S. Supreme Court handled the case, since U.S. Reports, the Supreme Court Reporter, and the U.S. Supreme Court Reports, Lawyers' Edition, any one of them, does not report on any other court. Since the Federal Reporter reports only on U.S. Courts of Appeals, you only need to indicate which Court of Appeals did the case, such as 306 F.3d 222 (9th Cir. 2001) or 306 F.3d 222 (CA9 2001). For Federal district courts when opinions appear in F.Supp., the jurisdiction is enough, as in "E.D.N.Y." (for U.S. District Court for the Eastern District of N.Y.) For other courts, "Supreme" becomes "Sup.", "Superior" becomes "Super.", "Appeals" or "Appellate" becomes "App.", "Civil" becomes "Civ.", "Criminal" becomes "Cr.", and so on.
• Judges' names are omitted unless you're making a point about a judge (e.g., the judge is unusually influential laterally beyond their court) or you're citing an opinion other than the opinion of the court, such as a dissent or a concurrence. If you name the judge, family name and abbreviated title in parentheses are enough, as in "(Burger, C.J.)" or "(Douglas, J., concurring)", after the rest of the case cite. "Judge" or "Justice" is "J." and "Chief Judge" or "Chief Justice" is "C.J."
• Lawyers' names are omitted unless you're making an unusual point. Identifying a lawyer as the defense counsel or the defendant's attorney is usually sufficient, if you need to refer at all.

Case names and cites above are fictional. Court names are approximate.

There's a Blue Book often used for other legal citational issues.

Perhaps this discussion should be copied into a permanent page and expanded.

Nick Levinson (talk) 02:50, 20 January 2011 (UTC)

I thought that {{reflist}} was the current template, and replaced <references/>. But someone just informed me that the opposite is true. So which is it? This policy page uses the former one. Is it out of date? Which is the right one, or the most current one? Nightscream (talk) 18:45, 6 January 2011 (UTC)

"No edit warring over it" is the only real rule. {{Reflist}} changes the font size and has features that <references /> does not (e.g., columns and groups). If anyone disputes a change in either direction, then we stick with whatever was there before. WhatamIdoing (talk) 05:37, 7 January 2011 (UTC)

{{Reflist}} does not change the font size relative to <references />, and <references /> has "groups". While <references /> does not (directly) provide column support, if an article uses reflist as {{reflist}}, then reflist isn't using the column feature and is visually equivalent to <references />. On the other hand, {{reflist}} has template overhead, so all things being equal it should be avoided unless the column feature of reflist is needed and actually being used. I think it's inappropriate for editors to replace perfectly fine instances of <references /> with {{reflist}}, certainly as the sole edit [6], but also in general. If the column features are not being used, why should any article use {{reflist}}? Gimmetoo (talk) 17:34, 7 January 2011 (UTC)

While i agree that <references /> allows grouping (used it myself) {{Reflist}} never theless seems to offer a smaller font size (it says literally so in the template description, however depending on your browser that effect might not be visible).--Kmhkmh (talk) 17:55, 7 January 2011 (UTC)

As I understand it you can use either, but you should stay away from going around and changing one template into another.--Kmhkmh (talk) 17:57, 7 January 2011 (UTC)

Both {{Reflist}} and <references /> both show a small font, as of December 21, although you may need to purge/bypass. The only thing that reflist now offers is column support. And there is now a gadget in preferences to disable small fonts. ---— Gadget850 (Ed) ^talk 18:07, 7 January 2011 (UTC)

Since both are the same, when {{Reflist}} is used for one column, then changing from one to the other should be a minor issue. As to columns, 48% of our readers still use IE or Opera which do not support columns (even IE9).[7] The 90% font size gets rounded up to 100% for a lot of IE users, although that is finally fixed in IE9. ---— Gadget850 (Ed) ^talk 18:59, 7 January 2011 (UTC)

You should refrain from meaningless edits, and since {{Reflist}} and <references /> now look exactly the same, you should not replace them with one another. Unless, of course, you add columns to a long reference list (20+ items). —bender235 (talk) 20:53, 7 January 2011 (UTC)

I don't agree that 20 refs is a "long references list". Funny you mention "meaningless edits" - like all the edits which kept changing <references /> to {[tl|reflist}} over objections for years rather than discuss a single edit that could change it everywhere? Or the people who still do it even though now it has no effect? Or the people who kept changing and still keep changing {{reflist}} to {{reflist|2}} without discussion on articles without a ton of refs? Gimmetoo (talk) 21:42, 7 January 2011 (UTC)

I'm not sure why you snapping at me. I just said that no one should replace <references /> with {{Reflist}} and vice versa, since this has no effect on the article's appearance. Changing from {{Reflist}} to {{Reflist|2}}, however, does have an effect. That is a totally different topic (and by the way, request a discussion before implementing this minor change is just ridiculous). —bender235 (talk) 23:54, 7 January 2011 (UTC)

Discussing changes to an article if all you are going to do is change the citation style is preferable. Going around and inserting your personal preferences is going to upset some people. If you do it and don't object when they revert it back, I don't see the harm however. Aaron Bowen (talk) 16:30, 28 January 2011 (UTC)

• Looks like the docs need updating, then. Template:Reflist/doc says, "This template…reduces the font size and has options for columns and groups." WhatamIdoing (talk) 23:30, 7 January 2011 (UTC)

Policy says: 'Embedded links should never be used to place external links in the body of an article, like this: "Apple, Inc. announced their latest product..."'

In Christianity articles, editors routinely cited chapter and verse (e.g., "Matthew 17:4-7") and use a template to turn that into an external link to a Bible Lookup service. It looks like this routine use of external links embedded in text is against policy. This issue has come up on Talk:Nativity of Jesus. A little guidance? Leadwind (talk) 15:49, 28 January 2011 (UTC)

FWIW, some existing discussion on this is at Wikipedia talk:Citing sources/Bible. The documentation at {{Bibleverse}}, the template in question, suggests rendering it in running text in parentheses, and that {{Biblesource}} - which links to Wikisource texts - is the intended long-term replacement. Shimgray | talk | 16:40, 28 January 2011 (UTC)

Has anyone made a Wikipedia reference style for EndNote? This would ease referencing things considerably for those of us who use EndNote (or Zotero) already. If nobody has made one, perhaps I'll find the time to make one at some point. If so, not sure how I should upload it to share with the community?

Also, I see "You can insert a link beside each citation in Wikipedia, allowing you to export the citation to a reference manager, such as EndNote". Would it be possible/ a good idea to have a bot do this automatically? — Preceding unsigned comment added by Pengortm (talk • contribs) 00:55, 9 January 2011 (UTC)

I never found one when I was looking real hard for such a tool a couple of years ago. Things might have changed; I haven't looked recently. If you find one, I hope you'll post your solution here. Cheers. N2e (talk) 22:38, 6 February 2011 (UTC)

From the WP:CITEHOW section:

You should follow the style already established in an article if it has one

I know this means if an article already has 20 refs, and you're adding #21, you should follow the established style, to have some consistency within the article. But what if one wants to change the style of all references? Does the rule above prohibit this? And if so, wouldn't this mean a certain aspect of one article can never change again? Wouldn't this put undue weight on the first editor's decision? --bender235 (talk) 02:58, 30 January 2011 (UTC)

I think if you are taking it over and doing a lot of content and taking it to GA or FA or whatever, then change the style. If you add a 21st ref, no content and then change the whole style, that's the kind of tail-chasing and water-washing away sand that makes this project a drag and writing offwiki better. Of course, if we just had a single citation style (no reason we can't, every journal manages), then all of this would be obviated, plus all the tools would support one style.TCO (talk) 03:28, 30 January 2011 (UTC)

I'm not talking about someone changing the style of the whole article on his own. I'm talking a change after consensus (for a new style) has been established on the talk page. Because right now, if you're suggesting a new citation style, people reply "no, per WP:CITE", as if WP:CITE prohibits any change at all. The thing is, if you look at an article like this, it might have an "established citation style", but it looks weird, especially with the bare URLs. Now if WP:CITEHOW meant that the original citation style has to be preserved for all eternity, it would just put an undue weight on the first editor's decision. --bender235 (talk) 11:51, 30 January 2011 (UTC)

The reason that we go with the first editor is the same here as ENGVAR: because everyone has an opinion, but none of them is favored over any other, so we need to have some easy rule to decide. The first editor rule is meant to be completely arbitrary, just as the preference of citation style is arbitrary. The point is that editors shouldn't go around changing articles to their favored citation style just for the sake of changing the citation style. If someone changes the citation style in the middle of a month-long improvement and cleanup effort, and the other editors at the article are happy, nobody else is going to complain. But if someone changes the style on 10 articles without making any content changes to the articles, that's inappropriate. — Carl (CBM · talk) 13:06, 30 January 2011 (UTC)

But why is it inappropriate, even if nobody complains in each of these 10 articles? And why are nowadays only "content changes" considered legitimate reasons to edit an article? What happend to minor edits? --bender235 (talk) 16:34, 30 January 2011 (UTC)

If a tree falls in the forest, does it make a sound? If no one (really no one) squaks, then what's the worry? However, I've got a deal for you. Will give you a C or B turtle article. You can have whatever format you WANT! Change all the old to new and roll with new for new info. However, I want you to take the article to GA. To provide work product for the reader, not just format. I'll even find a fun article. I'm having a little fun with you...but it's a serious offer. Let's rock content!TCO (talk) 19:09, 30 January 2011 (UTC)

Just like ENGVAR, the point is that everyone has their own preference. As long as no style is preferred by policy, there's no objective reason to go around changing the style. After enough long unproductive arguments about which style is "better", we've learned that the best solution is to leave the citations in the format they are already in, since everyone thinks their way is the best.

If there was consensus at the MOS page that some particular style cannot be used, of course it would be OK to change articles from that format. But since there isn't consensus about citation formatting, the only reason to go around changing it is "I think this is better", which is hardly compelling because we know from experience that there is no consensus that the changes are for the better. — Carl (CBM · talk) 20:49, 30 January 2011 (UTC)

Well, maybe in some cases there is consensus. Maybe, on an article like this one, a majority would prefer a citation style w/out bare URLs, if someone suggested one. But according to your reading of this guideline, no one is allowed to implement a new citation style, even if there was consensus. That's just irrational. I'd prefer to use WP:CITEHOW as a "applies in case of no consensus"-rule only. It should only apply when editors do not agree on one style or another. But if one implements a new style, and no one complains, it's all good. --bender235 (talk) 23:12, 30 January 2011 (UTC)

I would interpret the rule to be a bit firmer than Bender235's interpretation. If an article has a nearly-consistent acceptable citation style, it should be adhered to unless one first obtains consensus (or at least no complaints after a reasonable wait) on the talk page. Of course bare URLs is not considered an acceptable style of citation, so if most of the citations are like that , have at it. But a campaign of systematically changing citation styles of articles that already have acceptable citations with the excuse "most of the time no one complains" is not acceptable.

Actually, chances are someone will complain, just because there have been so many campaigns to change "BC" to "BCE" or "colour" to "color" or whatever that any systematic style change tends to be viewed as a direct attack on the principle of national and religious diversity. Jc3s5h (talk) 23:54, 30 January 2011 (UTC)

One has to know that by "change the citation style" I mean implementing a citation template. So if copyediting bare URLs into well-formated external links is acceptable because its an improvement, that begs the question whether transforming a hardcoded citation into a citation template is an improvement as well. Just like replacing a bare birthdate with {{birth date}} is an improvement, because it produces hidden, machine-readible microdata (check the source of "January 31, 2011" and "(2011-01-31)January 31, 2011"). For my part, I see a difference between replacing say APA style with MLA style (which is a matter of taste, like WP:ENGVAR, and what WP:CITEHOW was intended for to prohibit), and replacing any style with citation templates. I know CBM disagrees; he considers those citation templates a citation style like any other, but IMHO it is not. And I'd like to hear some opinions. --bender235 (talk) 01:37, 31 January 2011 (UTC)

What part of Citation templates are used to format citations in a consistent way. The use of citation templates is neither encouraged nor discouraged do you not understand? You have been warned! Jc3s5h (talk) 02:36, 31 January 2011 (UTC)

This bad faith is not necessary. Please be civil. ---- CharlesGillingham (talk) 06:45, 31 January 2011 (UTC)

Yeah, the more I think about, the more I agree with Jc. I said above (in another section) that changing the style would be ok with me if no one complained. But reflecting on it, I see a lot of little formatting things that became sort of standard around here less because of consensus than because some people just went around brusquely inserting what they like best. Aaron Bowen (talk) 05:36, 31 January 2011 (UTC)

Keep in mind Bender235's example: [8]. This is an article that has a citation style that is considered unacceptable. As I understand it, Bender is asking a very specific question: "given an article with bad or broken citations, does the "first editor" rule apply?" Are we forced to ignore a bad citation style on the basis of WP:CITEHOW? This is not the same thing as changing BC to BCE. ---- CharlesGillingham (talk) 06:45, 31 January 2011 (UTC)

@Aaron Brown: if anything, those citation templates are not what a single editor "likes best", but what a larger number of editors has agreed one after lengthy discussion. So in the end, it's actually the opposite of what you criticize. --bender235 (talk) 07:54, 31 January 2011 (UTC)

@Bender. Let's not debate the merits of citation templates now. Let's stick to the original question. ---- CharlesGillingham (talk) 08:11, 31 January 2011 (UTC)

The concept that citation templates are preferable has been proposed yet again and failed to achieve consensus yet again this month, this time at Wikipedia talk:Citing sources/example style#why not standardize on one format?.

I don't see where this survey was about citation templates. --bender235 (talk) 11:59, 31 January 2011 (UTC)

Well, the section right after the one I mentioned, Wikipedia talk:Citing sources/example style#Proposal: that use of citation templates be strongly encouraged is certainly about citation templates, and it failed to achieve consenus too. In any case, the "why not standardize on one format?" section is very nearly a prerequisite to considering templates superior to just typing the citation, because the templates only implement 2 styles. If we are not willing to limit ourselves to 1 style, we probably aren't willing to limit ourselves to 2 styles, and that means we are not willing to limit ourselves to citation templates. Jc3s5h (talk) 12:38, 31 January 2011 (UTC)

I think that we used to have language along the lines of not changing the style except when you had a non-cosmetic/non-personal preference purpose. For example, it's generally accepted that any editor can change from WP:General references to WP:Inline citations, as this is a clear improvement (in 99% of articles) rather than an issue of personal style. On the other hand, re-organizing the citations from Vancouver style to Chicago style, or from WP:FOOT to WP:PAREN, or from manually formatted citations to templates, is really just an issue of personal preference and usually oughtn't be done.

I think that moving from WP:ECITE to some other system is generally taken as a substantive improvement, not merely personal style (especially if the bare URLs aren't followed by proper linkrot-saving full bibliographic citations). WhatamIdoing (talk) 06:41, 1 February 2011 (UTC)

I agree that the language here could be improved. Yesterday, I noticed (via my watchlist) an editor using AWB to change articles from a one-column reference list to two column – a couple hundred in a row, all the same change, regardless of how few footnotes the article had. I don't think the current language here is direct enough in describing the general consensus against doing that sort cosmetic change en masse. — Carl (CBM · talk) 12:42, 1 February 2011 (UTC)

This needs expansion, but first: are shortened footnotes a subset of long footnotes or a separate method? ---— Gadget850 (Ed) ^talk 19:48, 9 February 2011 (UTC)

Shortened footnotes are a separate method. Each note presumes that the reader has read the corresponding entry in the bibliography. It is most useful when there are many citations to the same work, but the citations have small differences such as the page number. Jc3s5h (talk) 20:16, 9 February 2011 (UTC)

A question on the same subject. If most of the citations in the article are long citations to separate sources, but there is one or two source with lots of citations that may be better done as short citations, is it okay to combine in the "References" section the "{{Reflist|2}}" (generating the list of the short citations and the many singular long citations) with the short bullet list of the 'mother' long citations? --Rontombontom (talk) 11:13, 12 February 2011 (UTC)

I have seen that done, but it is not analogous to any citation style I have ever seen in any printed style guide. Since Wikipedia does not have any citation style guide, it is influenced by printed style guides, as adapted to take advantage of hyperlinks. Jc3s5h (talk) 14:34, 12 February 2011 (UTC)

Based on those printed style guides, can you think of an alternative to having "Notes" and "References" top-level sections in the described situation? Maybe something with sub-sections under References? --Rontombontom (talk) 14:41, 12 February 2011 (UTC)

There are many articles that use long-form footnotes for citations used only once and short form footnotes for multiple citations (to different pages) of the same work. The two techniques are not mutually exclusive, at least in Wikipedia. See Painted Turtle, for one. The editors of that article have chosen to use Notes (for explanatory notes) Citations (for footnotes containing citations) and Bibliography (for citations of books that are used multiple times). ---- CharlesGillingham (talk) 03:23, 13 February 2011 (UTC)

Thanks, I like the style in the example you linked, went with it. --Rontombontom (talk) 16:57, 13 February 2011 (UTC)

So— going by precedence, long and shortened footnotes are two separate systems that can be mixed? ---— Gadget850 (Ed) ^talk 03:29, 13 February 2011 (UTC)

Yes, it's fine to have long and short footnotes cited between ref tags, and listed under Notes or References (whatever you prefer to call it) using reflist. SlimVirgin ^{TALK|CONTRIBS} 17:45, 13 February 2011 (UTC)

Hmmm... how to document this, as we don't mix other styles. ---— Gadget850 (Ed) ^talk 17:54, 13 February 2011 (UTC)

We used to have a paragraph in here that mentioned this, but it seems to have been swallowed up over the last year. Look at this where it used to say: "Editors are free to use any method; no method is preferred over another, though the use of embedded links for inline citations is not considered best practice and is not found in featured articles. Some articles use a combination of general references, citations in footnotes and shortened notes. (See, for example, Starship Troopers, Rosa Parks or Absinthe.) Some articles use separate sections for footnotes containing citations and other footnotes. (See Augustus.)" I think this information was dropped in last fall's flurry of re-writes, a while back. You could add back in some of it. ---- CharlesGillingham (talk) 23:46, 13 February 2011 (UTC)

I have used Template:Cite doi for a journal reference I was having trouble with (ref 105 on Somerset Levels) & all seemed to have worked fine. However I'm doing final reference checks before an FAC nomination & I found although I added "accessdate=13 February 2011" into the citation this doesn't show on the article. This will get picked up at FAC and I was wondering if anyone knew how to overcome this problem?— Rod ^talk 08:35, 13 February 2011 (UTC)

|accessdate= will only display if you have specified |url=. When you cite a journal but not a URL no retrieved date should be needed (the contents of the journal paper don't vary by date, unlike a news story that may be updated over time). Rjwilmsi 20:50, 13 February 2011 (UTC)

Thanks.— Rod ^talk 21:32, 13 February 2011 (UTC)

On a related note, I think we need to put all of the ENGVAR-type information in a single section, and point a shortcut like WP:CITEVAR at it. I'm getting tired of spelling it out to editors who change long-established WP:PAREN articles to WP:FOOT, sometimes even under the mistaken belief that ref tags are absolutely required by policies. WhatamIdoing (talk) 23:02, 31 January 2011 (UTC)

Good idea. ---- CharlesGillingham (talk) 01:33, 1 February 2011 (UTC)

I've gone ahead and created that section. SlimVirgin ^{TALK|CONTRIBS} 18:31, 7 February 2011 (UTC)

Jack has reverted. [9] Perhaps he doesn't realize it's already in the guideline, just scattered in different sections. SlimVirgin ^{TALK|CONTRIBS} 00:42, 8 February 2011 (UTC)

I think the repetition of this disputed stance is unwarranted. I also think that this is an excellent example of the aggressive approach you were referring to in a section, below. The whole comparison to ENGVAR is off; Kentucky is going to be in US English; Kent is going to be in UK English. But most of what I see afoot here has no basis in anything other than WP:OWNERSHIP and first mover approaches. Before this thing holds any water, it needs to go through a community-wide RFC. Jack Merridew 00:58, 8 February 2011 (UTC)

If the article you are editing already is already using a given citation style, you should follow that style. Do not change the citation style used in an article merely for personal preference or cosmetic reasons. If you think the existing citation system is inappropriate for the specific needs of the article, gain consensus for a change on the talk page before changing it. As with issues of spelling differences, if there is disagreement about which style is best, defer to the style used by the first major contributor; if the style has never been stable, defer to the style used by the first major contributor. If you are the first major contributor to an article, then you may pick whatever style you think best for this article.

i.e. the first major contributor owns the article, no matter if later contributors out-contribute them? Martin (Smith609 – Talk) 16:46, 4 February 2011 (UTC)

Edited to follow what MOS actually says. Septentrionalis PMAnderson 03:32, 8 February 2011 (UTC)

Reverted edit by Pmanderson because MOS says no such thing, and because changing the proposal in the middle of the discussion is disruptive. Jc3s5h (talk) 12:40, 8 February 2011 (UTC)

Generally considered to be cosmetic changes

• Switching between major citation styles, e.g., switching between parenthetical and <ref> tags or between the style preferred by one academic discipline vs. another academic discipline
• Switching between citation formatting techniques, e.g., citation templates and manually formatted citations
• Changing the section heading to or from ==References==, ==Notes==, etc.

Generally considered to be helpful improvements

• Replacing bare URLs with full bibliographic citations: This is an improvement because it provides more information to the reader and fights linkrot.
• Replacing some or all general references with inline citations: This is an improvement because it provides more information to the reader and helps us maintain source-text integrity.
• In articles with several citations in incompatible styles (e.g., half the citations use ref tags and half use parenthetical citations), making all the citations use the same system: This is an improvement because it makes the article easier to read.

The above is my current proposal; I've put it in a separate subsection so that anyone can tweak it. If there are no changes or objections in a few days, then one of us should add it to the guideline (at the end?). I think that the examples are important, because it's these specific examples that editors frequently ask us about. WhatamIdoing (talk) 05:32, 4 February 2011 (UTC)

Switching between citation formatting techniques, e.g., citation templates and manually formatted citations

This is not a cosmetic change. Implementing citation templates makes citations machine-readable (by producing a hidden Z3988 object; also useful in combination with endnote.js). This is an improvement, too.

A purely cosmetic change would be switching from one citation template to another, like from Template:Cite book to Template:Citation. --bender235 (talk) 10:49, 4 February 2011 (UTC)

Bender, when not everyone agrees X is an improvement (and in particular when many perceive it has negative quality Y and assign little positive weight to feature Z). saying "X has feature Z, therefore X is an improvement" is not very persuasive. Gimmetoo (talk) 15:58, 4 February 2011 (UTC)

A) This is a discussion. No unanimity required.

B) Ok, so you consider this a cosmetic change. Let's see who else does (or does not).

C) Just because you don't understand the advantages of having machine-readable microformats on a website ("feature Z", WTF?), doesn't mean it's not an improvement.

D) What is the "negative quality Y" of citation templates? --bender235 (talk) 16:27, 4 February 2011 (UTC)

I regularly use the COinS metadata produced by citation templates (and never supplied outside of these) to save their information to my citation manager; it is also useful when locating articles in my university library, which makes it easier to find accessible versions of the references and hence verify them or expand the article on their basis.

Beside this, another benefit of using templates is that incomplete citations can be automatically expanded by bots such as User:Citation bot, reducing editor effort and improving reader experience; and links can more readily be updated and maintained.

Switching between citation templates and manually-formatted text impacts usability, not just appearance; therefore I don't think that "cosmetic" is the right term. Martin (Smith609 – Talk) 16:45, 4 February 2011 (UTC)

Adding citation templates is contentious, and quite the opposite of an improvement in the eyes of many editors. There's also no need to add them if references are already properly written. It's one of the issues CITE warns against. SlimVirgin ^{TALK|CONTRIBS} 17:37, 4 February 2011 (UTC)

Those rules of WP:CITE are not god-given. We can change them, once people realize having machine-readable citations are useful. --bender235 (talk) 19:27, 4 February 2011 (UTC)

"What is the negative quality Y of citation templates?" I'm not sure if this is serious or facetious, but they double the size of each reference, which in many cases if added en masse can significantly increase the size of an article, thereby substantially increasing load times. They clutter up editing boxes, and they are slower to fill out than using manual methods. If I asked you type out your address and phone number like this: John Doe 121 Jump Street Doeville, Florida 98845 and then asked you do put it in citation templates where in reftools or by copy and pasting the code you have to stop to change place several times (like in a job application), its obvious the manual method would be a lot faster. Aaron Bowen (talk) 18:59, 4 February 2011 (UTC)

The question was serious, but your answers are unconvincing. Additional size to the article doesn't matter, since there is no shortage of space on Wikipedia. That increased load times argument is dubious; how much is it? I mean if it increase from 1 second to 1.4, that's not the end of the world (but serious, show me some stats). "Cluttering up editing boxes" is purely a matter of taste; the same could be said about infoboxes, or any template. Yes, there are slower to fill out, but it's not like you have to enter all the code everyday; it's one edit, and it doesn't even have to be the original author of the article who implements them. Again, this is not about demanding anyone to implement templates, but just to allow it to gnomic contributors who are looking to improve Wikipedia (which I consider myself to do). --bender235 (talk) 19:27, 4 February 2011 (UTC)

But surely, Bender, you'd agree with me that "cluttering up editing boxes" is a cosmetic issue? And that removing that clutter on the grounds that it's ugly is a "cosmetic" change—a cosmetic change of exactly the sort you'd want to be discussed in advance, and done only if there is a consensus to do so, and not something that you'd want to allow a wikignome to do to your favorite articles with no warning or discussion?

There are serious objections to citation templates (which I use regularly, even thoughtlessly). The cite family is slow, and their presence makes the learning curve for newbies that much steeper. That said, I don't know why you jumped to the idea that this is a one-way street, in which citation templates are banned and manual formatting is protected, rather than the other way around. WhatamIdoing (talk) 01:20, 5 February 2011 (UTC)

The templates also encourage people to fill in unnecessary parameters. I've seen people fill in when they read a book; the name of the chairman/publisher of the New York Times; the original URL and the web archive URL (even though the latter contains the former), and access dates for both. So you end up with this long string of citation data that's lacking only a DNA sample from the author. SlimVirgin ^{TALK|CONTRIBS} 01:21, 5 February 2011 (UTC)

Size is important; many of our readers around the world still use dialup and many use microbrowsers, so any significant increase in size can make loading the page take that much longer. It's covered here but there also used to be a committee that had a lot of convincing data which suggested articles should be kept short. Unfortunately that was all deleted when the project coordinator went off the reservation and started ignoring community consensus. Size isn't the only factor, comprehensiveness is important and many of the articles I write get pretty long, but it is important to consider and its definitely a drawback of templates; a long article can be bumped up 10 or more kb if templates are added depending on how heavily referenced the article is. AaronY (talk) 09:05, 5 February 2011 (UTC)

You guys (SlimVirgin, AaronY) are confusing input with output. For my part, I'm talking strictly output here. Whether the editbox (i.e., input) looks messy or not with templates is (a) purely a matter of taste and (b) irrelevant for the output. After all, we're making Wikipedia for the readers, not the editors. With citation templates, however, the output has an additional feature (the aforementioned Z3988 object). It's like with {{birth date}}. In the edit box, February 5, 2011 might look less messy or confusing than {{birth date|2011|02|05|mf=y}}, but the latter offers an additional feature (hCard microformat, that is). The output of both ("February 5, 2011" vs. "(2011-02-05)February 5, 2011") looks exactly the same, but it's still more than a cosmetic change.

As for AaronY's argument with the article size: "10 or more kb" is way too high. For example here, the addition of citation templates added 1,162 bytes (not kB!) to the source code. It would probably be only half that much hadn't I also added all that DOI, ISBN and other missing information in that same edit. But what's even more important: that's the increase of source code size, not output size. The latter has only increased by that Z3988 object, which makes about 30-50 byte per entry, and sums up to 300-500 bytes (not kB!) in my example. I have no idea how you came up with "10 kilobyte or more". --bender235 (talk) 12:51, 5 February 2011 (UTC)

The addition of templates isn't only a question of increasing the size. There's an explanation here of why they slow down load times, often considerably. SlimVirgin ^{TALK|CONTRIBS} 13:00, 5 February 2011 (UTC)

Obviously we have reached the level of meticulous problems now. Of how much of an increase are we talking here? 10 milliseconds? Maybe 100? That's ridiculous. Compare load times of this w/out cite and this w/ cite, and tell me how much of a difference you observe. --bender235 (talk) 15:21, 5 February 2011 (UTC)

Did you read the page I linked to? Articles with lots of templates can take 30–60 seconds to load. What this means is they're not being read, because people tend not to hang around when they see a page fail to load. It also means they're less likely to be fixed up, because no one wants to wait 30 seconds for preview to load, then have to do that perhaps dozens of times in a row. The result—of both the slow load time and the visual clutter in edit mode—is that the more templates an article contains the worse the writing tends to be. SlimVirgin ^{TALK|CONTRIBS} 18:53, 5 February 2011 (UTC)

I read the page you linked to, but contrary to what you claim it compares two types of templates in speed ({{Cite book}} and {{vcite book}}), not templates and sans templates. Plus, this problem is primarly related to the length of the article in question (Israel), which is too long (it is more than twice as big as WP:SIZERULE recommends for maximum). What consumes time on this article aren't the citation templates, but the cite.php functions of referencing itself, and (in particular) the extensive use of {{harvnb}}. Even w/out citation templates, the article Israel would take ages to load.

Now, I still wonder, why didn't you compare the article versions I posted above? Because there is basically zero difference. --bender235 (talk) 20:42, 5 February 2011 (UTC)

I did compare the ones you linked to, and there is a small difference, but there are hardly any templates in it so it's not a good example. Israel is just 6,959 words readable prose size, which is quite normal; it's the 400+ citation templates that are slowing it down. Try these: Heather Mills (7,121 words; 193 templates), 2010 Haiti earthquake (8,068 words; 249 templates). SlimVirgin ^{TALK|CONTRIBS} 21:05, 5 February 2011 (UTC)

Heather Mills took less than a second to load, 2010 Haiti earthquake took about 20 seconds. Both are way over the recommended limit for article size. And neither would be loading any faster w/out citation templates (because it is not the templates that is slowing these articles down). --bender235 (talk) 22:07, 5 February 2011 (UTC)

You're lucky you got Mills to load. I'm often over 30 seconds for that one. These are standard-sized articles, and the slow load time can be traced to the templates, as numerous discussions across WP over the last two years have concluded. There's no point in ignoring the problem, because ignoring it won't help to resolve it.

I started editing Jesus myth theory around March 2010, an article that needed a lot of work, and it was hellish trying to get it to load, because someone had stuffed it full of templates. In order to be able to fix it I had to remove them, and from then on it loaded fine, so that's a very direct example I have where the only thing that changed was that the templates were taken out and replaced with manual refs. SlimVirgin ^{TALK|CONTRIBS} 22:46, 5 February 2011 (UTC)

It is not just our poor readers who must suffer the abomination that is the current citation template implementation, it is editors too. Previewing edits in an article that makes heavy use of such templates is a test of endurance. The vcite templates that Eubulides produced showed how it should be done, if one needs templates to help them at all. Manual formatting is still superior. Embedding machine-readable information in computer-generated HTML ranks as one of the most stupid ideas. Those very few people who might find such information useful should request WP produce an alternative machine-readable output format from the wiki markup, not inconvenience the 99.99% of readers who don't need it. In some fields, the citation includes a DOI or PMID or other ID that includes all the information a citation database needs, and is present even in manually formatted citations. Colin°^Talk 23:16, 5 February 2011 (UTC)

I'm somewhat tired of this debate now. For some reason, we're discussing a hardware-related problem (that might even be obsolete once the Ashburn servers are running), although this was once about style/usability. Yes, when an article grows too big, and the number of templates (all of them) has increased too much, then the rendering process slows down, and (to solve this problem) one should consider reducing the number of templates or spliting the article. But that still doesn't mean the use of templates should be avoided. And, to return to topic, that still doesn't mean implemented citation templates is a cosmetic change (and that was the question we started with). It is more than a cosmetic change, and therefore it should not be prohibited the same way we prohibit switching from British to American English. It should be allowed to implement these templates. --bender235 (talk) 23:35, 5 February 2011 (UTC)

Bender235, sometimes you have to accept that other people disagree with you and therefore you shouldn't force your preferences on them. It is widely recognised that citation templates cause difficulties for editors and for readers, as well as putting load on our servers. You are never going to get consensus for preferring them over manually formatted citations, which is what I assume your "It should be allowed to implement these templates" means. Colin°^Talk 09:40, 6 February 2011 (UTC)

← Perhaps it should be, or could be, but at the moment it's not. The argument about speed is just one of many intractable differences of opinion that will not be resolved here. This is exactly why our policy is to just leave citation styles alone. — Carl (CBM · talk) 23:43, 5 February 2011 (UTC)

• Adding structure to references *is* improvement; Bender's spot-on here. It makes the elements of the reference accessible to software tools that can then improve them further. Plain-text references are regressive, inherently prone to inconsistency, yada-yada. Citing first-mover on a 'style' is pure WP:Ownership; it's a wiki; things change; people edit. Jack Merridew 07:49, 6 February 2011 (UTC)

@CBM: There is no policy to "leave citation styles alone". There's a policy to keep the original citation style in case there is no consensus over which citation style to use. That's a big difference. The latter prevents edit wars, the first simply sets an article in stone.

@Colin: "It should be allowed to implement these templates" means literally that. One should be allowed to implement citation templates on any given article. If there's consensus to keep them, fine. If there's consensus to restore the old manually scripted refs, fine as well. But at least one should be able to implement them in the first place. And that would be prohibited by declaring it a "cosmetic change". It is not. It's an improvement. Maybe not for every article (in terms of loading time), but for most of them. --bender235 (talk) 10:53, 6 February 2011 (UTC)

After this discussion, if it wasn't already clear, it's clear that there is no consensus about switching to templates, and that the idea that they are an improvement is quite disputed. Ergo, the policy prohibits you going around making that sort of change. The idea that editors have to respond the same way at every article talk page would be silly: you know there is no consensus in favor of unilaterally switching articles to templates. — Carl (CBM · talk) 13:05, 6 February 2011 (UTC)

Consensus is not the same as unanimity. And just because some users complain about the slow loading of some pages doesn't mean citation templates don't add a useful feature. --bender235 (talk) 20:18, 6 February 2011 (UTC)

There's no policy on this, just a local group that is resisting the improvements to referencing. Jack Merridew 21:59, 6 February 2011 (UTC)

• Comment: Until editing refs is as easy as editing simple wikitext, cite templates are yet another barrier for newcomers to (a) understand wikitext, and have a sense it's not too daunting to try editing (b) get refs to work minimally, which is to say, have the correct information accessible to readers of an article. Badly formatted but complete manual refs are still understandable; broken cite templates are not. Until the ref editing complexity is properly hidden as standard, every article which uses cite templates is one less likely to attract newcomers to editing. We must not forget what made Wikipedia a success to date: its ease of editing. If it becomes an exercise in programming, as it more and more is, with the proliferation of infoboxes, complex tables, cite templates, persondata, etc it will slowly die as too many people are deterred from getting involved. Rd232 ^talk 11:34, 6 February 2011 (UTC)

You're correct, citation templates are more complex than plain text, and therefore harder to edit/understand for newbies. However, the same could be said about almost every feature of Wikipedia, from infoboxes to navboxes, and even basic wiki markup. Yet we still have all of these featues. Why do we, if newbies are confused by this? Because we do not require newbies to enter perfect wiki code. Newbies enter text on Wikipedia that needs copy-editing, so what? That's what the experienced users are there for. No Wikipedia policy requires an article creator to produce a perfectly formated article, with infobox, categories and all that stuff, and nothing less. But the guidelines recommend these features, and therefore encourage other (more experienced) users to implement them. --bender235 (talk) 20:18, 6 February 2011 (UTC)

Actually, I'm pretty sure the usability initiative has found that n00bs prefer entering ref details into a form over manually trying to format them. If they mess up, it gets fixed, just like a tpyo. And template folding is in the works, too. Jack Merridew 21:59, 6 February 2011 (UTC)

You're saying that User:Rd232's conclusion is incorrect and that there is data to prove it? ---- CharlesGillingham (talk) 23:20, 6 February 2011 (UTC)

We had a whole wiki on it: usability: (the grant has expired). There's strategy: too. There's also the wikimedia techblog^{template folding} which has talked about this. More and more of tools to facilitate entering cite templates via forms are being deployed; there are millions of instances of them. They amount to a de facto site norm. There's a proposed cite.php enhancement about somewhere that extends the ref-tag syntax to map to the cite template fields, and alleviates all the rendering speed issues, too. At its core, this is about proper information architecture; it is appropriate to structure the data; we're building a *database* that contains an encyclopedia. See: separation of presentation and content; it means that the specifics of formatting should be separate from the content. Mechanisms such as templates (or enhanced wiki-text syntax) allow the details of styling something like a reference to be encapsulated into a centralized implementation. And this is just one project, en:wp; there are hundreds of others, and I edit a lot of them: example. And they're not all encyclopedias. The whole notion of 'preserving' the plain-text format of references is regressive; We used to not *have* templates; any of them. Infoboxes used to be hard-coded html, not even wiki-table syntax; now Infobox templates are the norm. This is an ignorable guideline, one destined to fall by the wayside. Jack Merridew 00:25, 7 February 2011 (UTC)

There are discussions going on all over the place at the moment about how to attract more women editors. And the question could be expanded to how to attract more editors, men or women, outside the male 16–28(ish) demographic. I see this attempt to create a template-driven website as one of the key issues. As the site becomes more impenetrable, we will lose editors and fail to gain different kinds of new editors. SlimVirgin ^{TALK|CONTRIBS} 00:38, 7 February 2011 (UTC)

Sexism and Ageism? Jack Merridew is only 12 years old ;) I, however, am probably older than you are. So, your issue is with complexity? One of my prime concerns is the management of complexity. Free-form text refs are a sure-fire means to fail to manage the complexity of citations. References that are structured are open to improvement by tools; User:Citation bot can add a ISBN where it was omitted, WP:Reflinks can move the feckin' commas and periods to the proper side to the ref. WP:RefToolbar can expand an ISBN or URL to populate a ref; it can also diagnose problems with refs. Proper structure is key to the project's future. nb: I'm all for more women participating, here; the number of assholes in the world is not evenly divvied up between the genders (or age groups;). Jack Merridew 02:09, 7 February 2011 (UTC)

Nobody disputes that cite templates open up tech possibilities like that. The issue is whether (at the moment, without WYSIWIG-by-default) it's unambiguously or universally worth it. Is it more important to have commas and ISBNs correct, but lower probability of any given reader making the leap to editing? Rd232 ^talk 06:18, 8 February 2011 (UTC)

Quite right. Editing needs to be made easier - more like wordprocessing and less like programming. There are steps going on heading in the right direction, but until it's really achieved, any complexification of wikitext (...like cite templates...) has a very big downside in terms of deterring editors whose content competence we have great need of, but whose technical competence isn't high enough (or who are simply put off trying to learn because it looks too hard, even though it might not be). It's sad really that in 10 years of Wikipedia we couldn't manage to get the sort of WYSIWIG that eg Dreamweaver does for HTML - good enough WYSIWIG for general use, but the wikitext is accessible when needed by the more technically inclined. Rd232 ^talk 00:46, 7 February 2011 (UTC)

An edit I know of, made by a very non-technical person, was unsourced; the text remains to this day, and is still unsourced; it's quite true. No one is saying that every n00b has to use full citations. But true n00bz will balk at <ref> and get no further; a form, such as WP:RefToolbar, is accessibly to anyone with basic internet experience; it's like their Yahoo email account. They fill in some boxes and click teh 'Add citation' button. As above; it's about tools, lots of tools. It's teh internet; it's technical, both the problems, and the solutions. fyi, Dreamweaver's design view produces atrocious markup. Jack Merridew 02:09, 7 February 2011 (UTC)

The point you seem to be missing (certainly not addressing) is that whilst it's good to have editing tools that help, as long as by default people are confronted with increasingly complex wikitext, that doesn't solve the scaring-people-off problem. Only WYSIWIG-by-default can do that. In the mean time, without WYSIWIG, there is a tradeoff between using more and more complex markup because it's useful in some way, and deterring editors. As I said, the source of Wikipedia's success has been ease of editing, and that is gradually disappearing because of a bunch of complexification, and one of the factors is cite templates. The minute we have WYSIWIG-by-default, the game changes completely, but until then, we can't ignore the need to attract new editors to compensate for those leaving. In 2 years' time we might have cite templates in every article, but have fallen below a critical mass of editors, and become ever less able to maintain those articles (both superficially against vandalism and obvious problems, and more substantially for updates and improvements): that's not a possibility we can ignore because in an ideal world everyone is techy and/or the tools are available to make it easy: we don't live in that world. Rd232 ^talk 15:07, 7 February 2011 (UTC)

Be careful what you wish for. A WYSIWYG wiki-editbox would likely result in an explosion of complexity in the database. If you think things are getting all techie-goopy, now, then why would you want a wiki-word-processor? Recall that when the first WYSIWYG word processors emerged, we had an explosion of fonts and colours in all our memorandum. I served on a board of directors, once; we had a joke about the meetings: whoever has the most desktop publishing, wins. When the wiki was young, nothing was sourced; we're moving towards ever more stringent sourcing requirements and that's driving most of the complexity.

So, I said before that I'd read that WMF found in the usability studies that people preferred entering references into forms; I believe all such forms emit some sort of *template*. Sorry, I didn't bookmark it when I read it. You know that I've been tagged 'a deletionist', right? A lot of why I believe some articles should be deleted is that I don't feel we have a large enough core of editors to properly maintain the core articles that are important. Thekohser nailed that, in his comment. I believe a far greater danger to the project is the unchecked inflation of the database. Are we at 4 million, yet? 8? Cheers, Jack Merridew 00:32, 8 February 2011 (UTC)

"an explosion of fonts and colours" in DTP - yes, but these people didn't have the MoS, or a bunch of wikignomes cleaning up after them :). Seriously, that's a non-issue, not least because if we make that easy (whilst making clear what's expected) and then find we can't contain style abuses, it'll be a marker for how well we're likely to be doing on content issues... PS I'm quite happy for WYSIWIG to involve entering references in a form, and thus being recorded in the wikitext in cite template form, as long as the editor isn't exposed to the wikitext unless they want to be. Rd232 ^talk 06:12, 8 February 2011 (UTC)

There's an interesting article in The Independent today about the male domination of WP. I see the attempt to force templates and uniformity and "lots of tools" on everyone as a big part of that problem, and it's an attitude I hope we start trying to move away from. SlimVirgin ^{TALK|CONTRIBS} 14:48, 7 February 2011 (UTC)

I did not mean 'tool' in that sense ;) Your laptop is a tool. That is an interesting article, but the issue of male domination (and aggression) isn't limited to teh wiki; look around the real world. Most of wiki's problems are imported from there. Most of the domination and aggression are issues with specific cultures that have those in spades. It's the 'developed' nations that dominate the world that offer-up the most wiki-editors. I support more women editing, here; I support more editors from mellow cultures. Cheers, Jack Merridew 00:32, 8 February 2011 (UTC)

Funny article (everyone should read that). ---- CharlesGillingham (talk) 19:33, 7 February 2011 (UTC)

But I guess I should add that it doesn't really support your argument. He says Wikipedia is an expression of a male instinct to disagree. In my view, Wikipedia's refusal to agree on a "house style" for citations (and appropriate tools) is yet another example that Wikipedians are hopelessly contrarian and unable to compromise. ;) ---- CharlesGillingham (talk) 19:40, 7 February 2011 (UTC)

No, actually, it reflects the fact that we are a large collaboration, from many different traditions; many of those traditions include a standard form of citation, but they don't have the same one. The standard Wikipedian response: "Then everybody must do it MY way, and we'll be uniform" is not really an instance of the ability to compromise. Septentrionalis PMAnderson 19:52, 7 February 2011 (UTC)

Um... sort of. The contrarian says "I want to do it MY way and I won't do it like everybody else." You're confusing contrarians with fascists, who also can't compromise. As contrarians, we can't compromise, except to "agree to disagree", which is hardly compromising. I guess we'll have to disagree here. ;) ---- CharlesGillingham (talk) 22:10, 7 February 2011 (UTC)

I very strongly object to the mandatory use of citation templates, for the simple reason that there are citations ("Natalis Comes: Mythologiae siue explicationis fabularum libri decem; translated as Natale Conti’s Mythologiae, translated and annotated by John Mulryan and Steven Brown; Arizona Center for Medieval and Renaissance Studies, 2006. ISBN 978-0-86698-361-7") which will not fit a standard template, unless it have so many bells and whistles as to be less than useful to the average editor.

We can discuss the hypothetical citation-reading machine when it exists; when it does, it will probably be most useful if it is able to parse the standard format of citations without prompts and labels. Otherwise, the convenience of actual readers (even of actual editors) should not be deferred to the ease of programming a machine. Septentrionalis PMAnderson 19:48, 7 February 2011 (UTC)

I don't believe anyone is suggesting that cite templates be *mandatory*. Correct spelling, isn't mandatory. We let most anyone edit here, regardless of competence, intelligence, clue. Great strength, sure, but two-edged. <aside>know why three-edged bayonets were developed and then banned?</aside> Cheers, Jack Merridew 00:32, 8 February 2011 (UTC)

The suggestion is to make citation templates the house style, and citations without citation templates would be considered style errors. That's pretty close to mandatory. I would suggest that we send the template fans off to compose a citation manual, of comparable scope to the relevant chapters of the Chicago Manual of Style, so that we will know how to compose citations for cases that don't fit into any template, and also so that template authors will have a criteria to decide if their templates are behaving correctly. Only then should we even discuss making citation templates the house style. Jc3s5h (talk) 01:46, 8 February 2011 (UTC)

A modest proposal indeed, but also a useful one. Write when you find manual. Septentrionalis PMAnderson 02:11, 8 February 2011 (UTC)

Bender, you haven't addressed my question.

Suppose you and SlimVirgin are editing the same article. SV dislikes citation templates. She removes them. Now: do you want this page to say, "Gee, you should have discussed that first, and gone with the consensus, and left it alone if you couldn't get a consensus"? Or do you want this page to say, "Who cares if she removed all the citations templates! She declares that removing them is a significant improvement, and you should never object to a wikignome doing whatever they think improves an article."

IMO those are your realistic choices, because "You can convert to citation templates against consensus, but you can never remove them against consensus" has approximately a snowball's chance in hell of getting into this guideline, no matter how Good™ and True™ and Just™ and Right™ the cause may be. WhatamIdoing (talk) 02:18, 8 February 2011 (UTC)

Seriously, I don't know. Basically all I wanted was clarification whether "you should follow the style already established in an article" means first editor owns the article and sets the citation style in stone, or whether it means "if there are 20 refs with consistent citation style in the article, make sure that when you add #21 it looks the same." Because if it is only the latter than implementing citation templates would be an edit like any other, and therefore should follow WP:BRD. In other words: if you want to implement citation templates, do it! If you want to remove them, do it. And if someone reverts the edit, discuss. --bender235 (talk) 19:45, 8 February 2011 (UTC)

It should be only "if there are 20 refs with consistent citation style in the article, make sure that when you add #21 it looks the same"; that is what ENGVAR actually does (its wording fluctuates) - and what most such advice actually reads. First major editor should only be a fallback for articles which have never had a stable uniform style. Septentrionalis PMAnderson 20:47, 8 February 2011 (UTC)

Bender, you claim that ENGVAR principles "means the first editor owns the article". So you want the second editor to WP:OWN the article? Seriously, when those "second editors" make drive-by edits to thousands of articles of which they have no long-term interest other than establishing their own arbitrary and often incorrect style choices, generally through edit-warring with the articles' content developers, are those "second editors" not displaying their own form of ownership over the style of those thousands or more articles? I am become more of the opinion that style warriors are a disruptiive imfluence on articles, and could be blocked for WP:DISRUPTION. Gimmetoo (talk) 15:09, 9 February 2011 (UTC)

I'm coming to the same view, and in fact there's a debate touching on it at the moment on the MoS. It starts at Wikipedia_talk:MoS#Bots, but like all MoS discussions it's somewhat meandering. The debate has led to this new section in the guideline. SlimVirgin ^{TALK|CONTRIBS} 00:12, 10 February 2011 (UTC)

The date delinking RFARB had some relevant findings about using massive-scale edits to force a change to a different style. — Carl (CBM · talk) 01:41, 10 February 2011 (UTC)

@Gimmetoo: No, I don't want anyone to own the article. The "second editor's" proposal for style change should be dealt with like any other edit; which means WP:BRD: one editor changes the style, another editor might revert it, and then they discuss. If no one reverts, then everything is fine. Because, yes, there are a lot of people who love edits like this. There should simply just not be a blanket ban on these edits.

And by the way, calling this "drive-by editing" as if it was something bad is just plain stupid. There are more ways to make valuable contributions to Wikipedia than to build articles from scratch, and just because you don't devote yourself to one single article doesn't mean you're not acting in the project's interest as a whole. Threatening a ban is just uncalled for. Probably 80-90% of my 100,000 edits over the past six years have been minor edits that did not change the meaning/content, mostly to articles I've never edited before and probably will never edit again. So in your eyes that makes me a disruptive editor? --bender235 (talk) 13:49, 12 February 2011 (UTC)

Consider how this typically plays out. An article has a style A, developed as the article developed. Someone with a script changes the article to style B without discussion and without contributing to discussions. Contributors revert to style A. Script editor reverts to style B with edit summary "that's how things are done". Rinse and repeat. So yes, such style editing has some characteristics of disruption. It wouldn't surprise me that there are editors who intentially change the style of an article simply to create instability. Gimmetoo (talk) 14:42, 12 February 2011 (UTC)

The scenario you describe would be a violation of WP:BRD. That's not how I picture it. This would be more like it: an article has a style A, developed as the article developed. Someone with a script changes the article to style B without discussion and without contributing to discussions. Contributors revert to style A. So either that editor starts a discussion to seek consensus for style B, or he doesn't return to the article at all and it keeps its original style A. End of story. --bender235 (talk) 15:23, 12 February 2011 (UTC)

I would suggest that a user's behavior should be judged by the overall pattern of editing, not just one editing. Let us say the script user changes style A to style B in article U. It is reverted, just as Bender235 described at 15:23 UTC. The script user then goes on to change various consistent styles C, D, E, and F to style B in articles V, W, X, and Y. That script user is just as disruptive as if all 5 changes had been made to article U, and the script user should be banned. Jc3s5h (talk) 15:34, 12 February 2011 (UTC)

That's a good summary of the problem we're seeing with a few users lately. — Carl (CBM · talk) 15:41, 12 February 2011 (UTC)

That makes no sense at all. Just because a certain type of edit gets reverted on article doesn't mean it shouldn't be done anywhere. That's ridiculious. --bender235 (talk) 18:15, 12 February 2011 (UTC)

Once the same sort of change has been reverted on a multiple articles, doing the same thing on a dozen other articles is, indeed, disruptive. An editor who institutes his or her personal preferences on large numbers of articles after objections to those edits have been raised is demonstrating egotism rather than collegial editing. This is why our general principle in ENGVAR and WP:CITE strongly discourages changing the established style based solely on personal preference about "improving" the article. — Carl (CBM · talk) 18:47, 12 February 2011 (UTC)

@Septentrionalis: Good to see that you agree with me on this one, but a lot of people don't (see CBM's comment below). Therefore I was wondering what actually is consensus here, CBM's reading of CITEVAR, or mine. --bender235 (talk) 13:52, 12 February 2011 (UTC)

@Pmanderson: ENGVAR says two things: (1) if there are 10 uses in the same style, and you add another, follow that style. (2) If there are 10 uses in the same style, don't go changing them all to some other style just because you prefer it. WP:CITE says the same things about citation styles: follow the existing styles, and don't go around changing them based on your personal preferences. — Carl (CBM · talk) 01:41, 10 February 2011 (UTC)

I dislike citation templates. Not only do they encourage inputting of often useless data, as already remarked upon above, they potentially clutter the text in edit mode, often making it difficult to parse. OTOH, there are numerous editors who either do not know how to put in citations, or are too lazy to do so. I find it preferable to have more information than the bare url, and often fix these up by call WP:Reflinks, which usually fills in the whole shebang using {{citation}}. Note that such action often goes counter to CITEVAR, but Reflinks doesn't give users the choice, AFAIK. However, the problem overall isn't the existence of citation templates, but the tendency of many editors to overpopulate these templates by mechanically and anally inserting information merely because it's available. --Ohconfucius ^¡digame! 02:09, 10 February 2011 (UTC)

Perhaps someone with the requisite programming skill should create an alternative to WP:Reflinks which creates a citation that is free of templates, in a choice of popular styles (APA, Chicago, MLA). Jc3s5h (talk) 15:38, 12 February 2011 (UTC)

Couldn't it just subst the templates, as an option? — Carl (CBM · talk) 18:47, 12 February 2011 (UTC)

No, that puts the messy guts of the template into your article text, rather than just the output. You can, however, preview the text, copy the output, paste that in to the article as a replacement for the template, and add italics and bold formatting by hand where necessary. A definite kludge, but it amused me to do it the other day. WhatamIdoing (talk) 21:46, 17 February 2011 (UTC)

You looked at the Reflinks 2 design docs? That, template parsing, user definable per site rules, and interface improvements when my time isn't diverted to more interesting things. What is the status on mw:Extension:TemplateAdventures, single in multiple format output citation system? There is a monotary argument for speeding up template parsing [10]. — Dispenser 20:55, 12 February 2011 (UTC)

I have expanded WP:CITEVAR, but left out the example involving citation templates. That example of a "cosmetic" change said:

• # Switching between citation formatting techniques, e.g., citation templates and manually formatted citations

I still think that is a useful and valid example, as I don't think that SlimVirgin should be able to force Bender to give up citation templates in an article that has been happily using them for years, just like I don't think that Bender should be able to force SlimVirgin to accept them in an article that has happily shunned them for years.

However, IMO the most important point is all the other stuff, not the one example. Accordingly, I've added the other stuff (slightly tweaking some wording to merge with existing text, in a way that I hope is clear), and hope that works for everyone. If you want to continue the discussion about citation templates vs manual formatting, perhaps we could take that up again—but maybe not until next month, and in a new section. WhatamIdoing (talk) 22:00, 17 February 2011 (UTC)

The "A quick how-to" section says: Then add this to the end of the article:

==Notes==

{{Reflist}}

Which, historically, makes some sense -- the footnotes section is generally where quick inline citations are given in an article, but the footnotes section isn't exclusively citations, it could also include author commentary on a passage, or a tangential story which helps to explain background but doesn't really fit into the main text. For a Wikipedia article, though, this section is where the <ref> tags in an article are sent, as collated by the {{Reflist}} template. Technically, everything down at the bottom of the article is in the footnotes, the "Notes", the "External Links", it's all that stuff at the bottom of the page that's not really part of the article proper. Calling this section simply "Notes" gives it an air of impermanence, as though what's there was simply jotted down and may or may not actually have anything to do with the rest of the article. I think the section should be entitled "References" since it's where the References in an article appear. "Notes" almost implies what the discussion page is for, a place for brief remarks, marginal comments or explanations.

==References==

{{Reflist}}

Is how it should be formatted. Banaticus (talk) 19:55, 1 February 2011 (UTC)

Perhaps not. When the Notes use short-reference forms and there is a separate list with the full bibliographical data, what should that list be headed? If the notes are called "references", then should the references by called "sources"? Traditionally (in the print media), such a list was called the "Bibliography", but more recently (the last 75 years or so) they are often headed "References" or "Sources" (or some variant, such as "Publications Consulted"). On Wikipedia, these matters are very confusing, since "Bibliography" is often used in biographical articles for a list of publications by the subject of the article.—Jerome Kohl (talk) 21:24, 1 February 2011 (UTC)

Wikipedia:Verifiability basically says that everything should have some sort of citation. So, when a person is putting all of the <ref> tags in the article that they should, you're suggesting that they should post references twice, with a short version inline in the article then the "full" version in a separate section below? Shouldn't they just put the full reference in the first time? Otherwise, if anything needs to be "fixed" or if the article is reorganized and less references are used or whatever, a person would have to track down and coordinate each double reference. If a publication was consulted, but never referenced, why are we bothering to list it? And if it's referenced the first time, why list it again? I can see using things like "Smith, ibid" but isn't that taken care of by reusing a named reference? Banaticus (talk) 23:54, 1 February 2011 (UTC)

First of all Wikipedia:Verifiability says no such thing as you claim in you first sentence and second sourcing of an article does not necessarily mean inline citations. For short articles based on a few sources (and not having controversial material) it is usually good enough to simply list those at the end of the article (under references).

Also notes and references are used in slightly different manners by different authors and there is no need to mandate a particular version. Some use notes and references as described in the posting above yours, some might use it to distinguish between explaining footnotes and sourcing footnotes, some might use references for general references potentially sourcing large parts of an article and notes for sourcing only specific bits of information and to avoid listing sources under references, which are only sourcing a single fact but are otherwise of no further interest for the article's topic, etc.--Kmhkmh (talk) 00:58, 2 February 2011 (UTC)

To Banaticus: see WP:CITE#Shortened footnotes. Short notes are used when an article cites many different pages of the same book, as is common in history articles and introductory articles to academic subjects. This is Wikipedia's way of handling "Ibid". See also WP:REFNAME, which is technique which eliminates identical footnotes, whether short or long. ---- CharlesGillingham (talk) 17:31, 2 February 2011 (UTC)

The benefit to notes/references is letting you run with a series of short-form footnotes - "Smith, p. 1"; "Smith, p. 2"; "Smith, p. 7" - as well as providing more flexibility for annotating citations - "Smith, p. 19, but see also Wilson, p. 48, and despair." You the don't have to worry about ensuring that the full citation for "Smith" comes first (and doesn't get lost or duplicated), as well as avoiding having long citations making it hard to see the notes. Shimgray | talk | 23:20, 3 February 2011 (UTC)

I want to repeat Kmhkmh's comment: WP:V does not require either WP:Inline citations or WP:General references for everything. In fact, WP:V requires inline citations only in three instances:

1. direct quotations
2. a statement that has been challenged (e.g., with a {{fact}} tag) (in good faith), and
3. a statement that the person adding the matieral, using his or her best judgment, believes is WP:LIKELY to be challenged.

Also, while most editors choose to use <ref> tags, and most editors choose to use ==References== for the primary section heading, there is no requirement that they do so. They can use WP:PARENthetical citations, or manually numbered footnotes (NB that I personally think that a poor choice), or any other system that they think works for the article. They can also use any section heading that they think will make sense to the reader, such as ==Works cited== or ==Sources consulted==. We do not have a mandatory house style. WhatamIdoing (talk) 04:56, 4 February 2011 (UTC)

The section in question conflicts with WP:FNNR. ---— Gadget850 (Ed) ^talk 14:17, 4 February 2011 (UTC)

I agree with WhatamIdoing. Different citation and footnote styles are suited to different articles, according to the nature of the content and the source material. WP:FNNR best reflects established practise, which is what guidelines should aim to do. There is no entirely unambiguous terminology, and no ideal or universal solution. Clarity within each article is what matters, as always. Geometry guy 14:56, 4 February 2011 (UTC) (Is this^[1] a citation, a source, a (foot)note or a reference? How about {{citation|this}}?)

Gadget, I'm confused. How does titling the section ==Notes== conflict with FNNR? FNNR says that titling the list of citations as ==Notes== is the second-most popular choice across Wikipedia. WhatamIdoing (talk) 01:12, 5 February 2011 (UTC)

Because this project page presents Notes as the only choice. The actual title selection does vary, and depends on the article content. Examples:

• Arthur Rudolph uses Notes for explanatory notes and References for the reference list
• Chaco Culture National Historical Park uses Notes for explanatory notes, Citations for shortened footnotes and and References for the reference list
• Erie, Pennsylvania uses References and Sources, although it is unclear how Sources directly relates to the article

---— Gadget850 (Ed) ^talk 13:42, 5 February 2011 (UTC)

Those examples illustrate a variation in use but the don't illustrate a FNNR violation. Furthermore in all 3 examples it is rather obvious how you would check the veracity of the content you check the given inline citations. The sources section in Erie is probably a leftover from earlier versiona and could be transormed into inline citations (and deleted afterwards).--Kmhkmh (talk) 09:52, 7 February 2011 (UTC)

Thanks -- that's why, in a number of articles that I've seen, a separate Notes and References section isn't very helpful -- there's no clear way to see just how a given book or whatever that's cited as a reference relates to the article. If the accuracy of a given paragraph is questioned, it seems as though the only option for a new editor is to go through *all* of the listed sources until one happens to stumble across whichever source was used as a reference in that particular case. As much as I'd love to say, "Well, anyone who wants to check the veracity of a paragraph should contact one of the major article contributors", we all know that editors come and go but Wikipedia remains -- we need to lower the bar, the initial hurdle, that's necessary to add or modify information in an article as new research contributes to a greater understanding of a topic. This is why I think inline references as generated by {{tn:reflist}} should be listed in a References section, because that lends greater weight to actually linking statements made in the article to specific references, instead of simply dumping a load of sources into the end of the article as though to say, "Here's some good sources, you figure out how they were all used." Banaticus (talk) 18:29, 5 February 2011 (UTC)

I don't understand the argument there. The ability to check the correctness of the content by looking up the sources is not defined by the name of a section title, in which all those citations are listed. That you check the inline citation/footnote for a sentence/line/paragraph is an rather obvious choice and whether this turns out to be an actual source or just an explanatory note is something the reader see immediately by looking at it. so I don't see any reason for confusion, the scenario you describe above doesn't really exist in this context.

However the problem you describe has nothing to do with using notes or references as section titles, but it is a problem of not using inline citations at all. --Kmhkmh (talk) 09:41, 7 February 2011 (UTC)

Gadget, it's true that the examples on this page happen to use ==Notes== but not ==References== in that context, but I do not agree that failing to use any one of the several valid and generally accepted headings means that it conflicts with FNNR, or anything else. We don't give any examples of ==Footnotes== or ==Works cited==, and I don't see that omission as a problem, either. I don't object to changing one or more of the examples, but the current state of the page is not actually wrong. WhatamIdoing (talk) 04:03, 8 February 2011 (UTC)

Going back to the original issue, the OP is stating the References should be used instead of Notes in the example. The example is in conflict with FNNR in that it implies that the example uses the only section title. It should be noted that the section title can vary. ---— Gadget850 (Ed) ^talk 05:10, 8 February 2011 (UTC)

Based on what? Right I fail to see on what or which you base your argument. The one you've cited doesn't not support what you're claiming (at least i don't see how).--Kmhkmh (talk) 15:13, 8 February 2011 (UTC)

Gadget, I just don't get this. What do you want? For us to put a footnote on every single example that says, "Regardless of the section heading we happened to use in this example, you can use any sensible heading, exactly like we say elsewhere in this guideline. We only use one, because any experienced editor knows that if we put == Notes/References/Works cited/Sources/Bibliography/References/some other sensible section heading chosen by editors == in the example, then some poor confused person will actually paste that whole mess into an article. So whatever we used in any given example is not a binding determination that 100% of all Wikipedia articles must use that particular word or phrase for the same section: You can use whatever title you think works best for this article, okay?"

Alternatively, can you show me an example of a dispute in which anyone was actually silly enough to think that the example given here was not merely illustrative, but an unalterable, absolute requirement? WhatamIdoing (talk) 23:30, 9 February 2011 (UTC)

Banaticus, would you please look at Chaco Culture National Historical Park and tell me if you have trouble figuring out exactly which book that supports the last sentence of the first paragraph? WhatamIdoing (talk) 04:03, 8 February 2011 (UTC)

If you are referring to reference 2, the in-text citation links to the short cite Strutin 1994, p. 6, which then links to the full cite Chaco: A Cultural Legacy,. ---— Gadget850 (Ed) ^talk 05:10, 8 February 2011 (UTC)

Yes, it's perfectly clear to me, but from the above comments, I'm not sure whether it is clear to Banaticus. If it is, then perhaps s/he'll explain what the actual problem is (i.e., because if this is clear, then I have misidentified the alleged problem). WhatamIdoing (talk) 23:32, 9 February 2011 (UTC)

I strongly agree that notes are not references. Notes are separate from references, see Template:Note or articles such as Juliusz Słowacki. --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 17:23, 17 February 2011 (UTC)

There is a discussion as to whether or not to include quotes in citations discussion at MOS:talk[11]. I am linking to that rather than opening parallel discussions on both this talk page and that one. PPdd (talk) 22:34, 1 February 2011 (UTC)

The discussion is now archived at Wikipedia talk:Manual of Style/Archive 119#Style of putting Quotes in citations; it may help readers skip, or quickly check, source content. 84user (talk) 20:42, 18 February 2011 (UTC)

On the subjects of uncontroversial edits to CITE that may be controversial, is changing "Generally considered to be cosmetic changes" to "To be avoided unless there is consensus" something that should be discussed here? I think that the previous heading was correct. --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 20:08, 18 February 2011 (UTC)

While I thought the original was accurate (I wrote it), I have no objection to this change. WhatamIdoing (talk)

Actually I prefer the current formulation and I don't quite see it as purely cosmetic change either, as one version expresses a preference whereas the other does not.--Kmhkmh (talk) 21:39, 18 February 2011 (UTC)

What is the best way for me to fix a dead link that was in Aaron Schock#References? It is now behind a paywall. I quickly searched the talk archives for paywall but found nothing specific to this case.

The Chicago Tribune has described Schock's political positions to be fiscally conservative and somewhat moderate on social issues.[22]

where cite 22 was

"For Congress". Chicago Tribune. October 22, 2008

The chicagotribune URL was dead and not in the Internet Archive. I used the tribune's site search and found the now-paywalled article at http://pqasb.pqarchiver.com/chicagotribune/access/1580304701.html?FMT=ABS&FMTS=ABS:FT&type=current&date=Oct+22%2C+2008&author=Anonymous&pub=Chicago+Tribune&edition=&startpage=46&desc=For+Congress

The free abstract is just one sentence that does not support the wikipedia claim. Of course I WP:AGF and am reasonably sure the full text does support it.

The first item under Wikipedia:Citing sources#Preventing and repairing dead_links advises: "First, check the link to confirm that it is dead. The site may have been temporarily down or have changed its linking structure." Well, putting the material behind a paywall is kind of like changing "its linking structure". But should I then replace the original URL with a long paywall URL?

I was unsure so I looked at Wikipedia:Link rot#Repairing a dead link: "If you find an archived version, double-check to make sure that the material still supports the citation." Well, I tentatively decided that the paywall is a kind of "archived version" and added "|archiveurl=longpaywallurl" to the cite, which now looks like this:

"For Congress". Chicago Tribune. October 22, 2008. Archived from the original on 2011-02-18.

My concern? I feel I have somewhat misused the archiveurl parameter in {{cite news}}, and also that I have not followed the advice "If you find an archived version, double-check to make sure that the material still supports the citation." Normally I double-check such things, but I am not willing to pay money to do so.

There are likely several ways to deal with cases like this, but more advice would be welcome. -84user (talk) 22:33, 18 February 2011 (UTC)

The full text is (it's their endorsement editorial, which should be stated):

Illinois Republicans have a young phenom in Peoria: state Rep. Aaron Schock , who is running for the 18th District seat of retiring GOP Rep. Ray LaHood.

Schock won a seat on the Peoria school board when he was 19. By 23, he was president of the board and a member of the Illinois House. Now at 27, he's poised to become the nation's youngest congressman. He's running against Colleen Callahan, a former radio and TV reporter and commentator in the Peoria area.

Schock sometimes shows the signs of youth -- during the primary he called for the U.S. to sell nuclear weapons to Taiwan to persuade China to help curb Iran's nuclear ambitions. He quickly acknowledged that was a mistake. But he has built a solid legislative record. He has clear and coherent positions that reflect fiscal conservatism and a moderate approach to social issues. If the Republicans had a dozen more of him around the state, they wouldn't be wandering in the Illinois wilderness.

Callahan doesn't have the poise and clear-eyed views of her opponent. Schock is endorsed over Callahan and Green Party candidate Sheldon Schafer of Peoria, a college professor and astronomer.

Add page 46 to the citation, and remove the link. What is being cited is the Tribune itself, not any electronic edition. Since the Tribune is bankrupt, it is not surprising its archives are decaying. Septentrionalis PMAnderson 22:47, 18 February 2011 (UTC)

Ok, I searched for the text from your post above and found http://articles.chicagotribune.com/2008-10-22/news/0810210269_1_foster-republican-jim-oberweis-illinois-republicans which supports the article. I then made this edit, using the new link as an aid to readers that do not have the print version. Does that look reasonable? -84user (talk) 03:07, 19 February 2011 (UTC)

Aside from the detail that Schock was running in the 18th district, and Foster in the 14th, so Foster is off-topic. Septentrionalis PMAnderson 02:51, 20 February 2011 (UTC)

A proposed bot would change the format of dates in selected categories. For example, an article in the category "United States military" would have a date like "July 7, 1777" changed to "7 July 1777". Some dates would be protected, such as dates within single or double quotation marks (that is, if the bot can count them correctly), or dates in the title parameter of a citation template. But the APA style is allowed in Wikipedia articles, and that style calls for the names of articles in journals and periodicals to be written without quotation marks or any special typography. Any title that contained a date with a different format than the format chosen for the article would be altered, which would potentially make it more difficult to search for the article. Jc3s5h (talk) 03:07, 20 February 2011 (UTC)

New addition under discussion, "Text-source integrity" section:

It is important to remember that Wikipedia articles are usually edited by multiple editors, and change over time. Thus referencing only at the end of a paragraph composed of multiple sentences is insufficient, for the same reasons as general references are: because 1) they do not guarantee that a particular unreferenced sentence was not added later, and is not covered by the end-of-a-paragraph reference, or 2) that the end-of-a-paragraph referenced sentences was not added later as the only referenced sentence to an unreferenced paragraph.

In the following paragraph, for example, there is no grantee that the first two sentences are referenced by the end-of-the-paragraph sentence. Only the last sentence can be reasonably trusted as referenced:

 N Through the 1940s and early 1950s a wide variety of efforts were made to address the color problem.  N A number of major companies continued to work with separate color "channels" with various ways to re-combine the image.  Y RCA was included in this group; on 5 February 1940 they demonstrated a system using three conventional tubes combined to form a single image on a plate of glass, but the image was too dim to be useful.^[3]

Piotrus, I understand the concern you have, but the prescribed solution won't help. First, it leads us into stupidities. Consider this:

Education researcher Mary Jones says that there are three kinds of students. The first group is made up of students who do their homework as soon as they receive the assignments. The second group contains students who do their homework at the last possible second. The third group is comprised of students who didn't even realize that they were supposed to do the assignment.^[1]

One citation ought to be sufficient for this entire paragraph (usually placed either after the first sentence or the last). Nobody will be left wondering whether the middle sentence is supported by the same paragraph as the first sentence or the last sentence.

But even if the original editor spammed citations after each sentence, you can still have newbies (accidentally) and vandals (maliciously) destroy the source-text integrity:

Education researcher Mary Jones says that there are three kinds of students.^[1] The first group is made up of students who do their homework as soon as they receive the assignments.^[1] The second group contains students who do their homework at the last possible second.^[1] The third group is comprised of students who didn't even realize that they were supposed to do the assignment. This division was later rejected by Sam Smith, who proposed a four-part division of students: students who did their homework immediately, students who procrastinated, students who were oblivious, and students who should not have been assigned homework in the first place.^[1]

The fact is that no amount of citation by the first editor can prevent the loss of source-text integrity. It is necessarily the duty of the subsequent editors to ensure that their changes do not introduce this problem.

And, if a timely example were needed, I cleaned up an article a little while ago in which a fact-tag had come astray of the sentence it had been applied to. WhatamIdoing (talk) 00:34, 18 February 2011 (UTC)

What happens if an editor adds a sentence from a second source in to the middle of that paragraph:

Education researcher Mary Jones says that there are three kinds of students. The first group is made up of students who do their homework as soon as they receive the assignments. Independent analysis questioned whether this groups was chosen randomly.^[2] The second group contains students who do their homework at the last possible second. The third group is comprised of students who didn't even realize that they were supposed to do the assignment.^[1]

The second editor has acted properly by citing the added material, but it ends up leaving the first sentences without a clear connection to the citation.   Will Beback  talk  00:43, 18 February 2011 (UTC)

I've encountered this situation in my own sourcing "interrupting" a set of sentences to one source, and the best solution I've found is either: rewrite to move the new information out of the "block" that uses one source (your example is one that I would do that with, as it feels like an independent thought that can be added afterwards), or otherwise place a named ref tag just before the new sentence so that the existing material sourced to "1" still is obviously sourced to "1". Almost always, however, rewriting is the easier solution if the sources are two discrete things. --MASEM (t) 00:53, 18 February 2011 (UTC)

The only way to be completely sure about text-source integrity is with an annotated WP:CITEBUNDLE.---- CharlesGillingham (talk) 02:19, 18 February 2011 (UTC)

I bundle citations as much as I can nowadays at the end of paragraphs, because it reduces the clutter of footnotes, and it makes adding new sources easy:

And so on, for as many points as you add to the paragraph. If you include links to specific pages on Google books, or to the articles, the reader can click straight to the words you're relying on. You can use the same system with the Harvard ref citation templates, or with any other templates. SlimVirgin ^{TALK|CONTRIBS} 03:19, 18 February 2011 (UTC)

I think it helps to keep some perspective. It is simply not possible (or at least practical) to demand that the connection between text piece and inline citation through placing only is in such a manner that it cannot be confused at all. Citation bundling is not really fixing that either, unless you explicitly comment your sources to avoid any ambiguity. But the thing that matters here is the commenting not the bundling, commenting individual non bundled inline citation achieves the same effect. So unless we would enforce commeting of individual citations there's no way to avoid confusion completely.

I don't really see all that as much of drawback anyhow. The worst case scenario is, that somebody proof reading the article might look at one more source than he theoretically had to (big whoop). The only way to (really) check whether paragraph is properly sourced is to actually look the sources. Once you've done that you know anyhow whether the source covered the whole paragraph or merely parts of it. So if you proof read a paragraoh, you simply check all its sources/inline citations and hence it ultimately doesn't matter all that much where in the paragraph they were placed. Of course that doesn't mean that editors can distribute their inline citation randomly over a paragraph, but it means extensive discussions about "the" perfect placing are probably a waste of time.--Kmhkmh (talk) 03:53, 18 February 2011 (UTC)

There's no point in the first editor creating an annotated CITEBUNDLE for a paragraph supported by a single source: when the first editor finished, the relationship between the source and the text was perfectly clear. The burden is, and must be, on the subsequent editors to not screw up the pre-existing source-text relationship when they add new material. WhatamIdoing (talk) 17:33, 18 February 2011 (UTC)

First, let me point out that an article, nowadays, is unlikely to become GA (or higher) as long as it has any unreferenced sentences. WP:V clearly allows tagging with {{fact}} and removing (not that we do it much) any unrefeenced sentence. I thought that the text I added made it clear why this is good practice, and let me note that WP:CITE already (if unclearly) suggests this is the case; in the "General reference" paragraph. Unless each sentence is clearly referenced, using a few citations just provides reference to that sentences, and the article still suffers from "general reference" syndrome.

Second, let me address points risen.

First, the "Education researcher Mary Jones says that there are three kinds of students" example. Why are you so sure she sais this in ONE article? Maybe she published three articles, one for each group, and we cite only the last one? Unless I see a ref for every sentence, I cannot assume all three groups are described in the last article (Granted, I'll think it is likely, but not as likely if I saw the ref in every sentence and was sure of it). Further, it is possible that the sentence was first written only about the third group, and then somebody else expanded it with her other works discussing the two other groups...

You are of course right that even this system can be messed up, but referencing every sentence makes it less prone to being disrupted, as editors will get the suggestion that the content they should add should be referenced. If somebody messes things up (or vandalizes them on purpose), it is indeed the community that has to catch it and fix it. The type of error you display is actually difficult to stop and fix, but that is not relevant here - you could just as well argue that no inline citations are needed, because somebody may not add a "General reference", so why bother?

Will Beback raises an excellent point, we could add this to my original text as another potential error. See also Wikipedia:Don't hijack references.

CharlesGillingham, yes, cite bundle is useful. I've in fact not used it in the past, but now that I became aware of it, I will. That said, it is not directly relevant here (it notes why we need to clarify when multiple refs are used in one sentence, and this argument is about the clarification needed on poorer levels of referencing).

SlimVirgin, I guess citebundle may be seen as an alternative for a para referencing, but it is not as helpful, IMHO. It makes the reader have to go and look at the bundle and then try to match it up to the sentences to see which sentences in a para are referenced or not. This can be confusing if the wording is different, and the reader is not an expert in a given field. I will note that the citebundle recommends its own usage not as a replacement to referencing every sentence, but - as I noted above - as a clarification for sentences were multiple refs are used.

Kmhkmh, while there is no such thing as a perfection, referencing each sentence is more helpful than not, be it for a reader or a proofreader, as it is the quickest way for them to verify the fact they just foound (click the ref, see which it is). In such an instance they know which source to check for verification.

WhatamIdoing: I disagree. The writer should reduce any and all burdens on subsequent readers and editors. Of course, they should improve the article, not screw them up, yes - but that doesn't reduce the previous author's responsibility to facilitate their work. Wikis are all about facilitation, building on the work of others, and making that process easier.

On a final note, I'll add that nobody has presented an argument why we shouldn't reference every sentence... --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 17:59, 18 February 2011 (UTC)

PS. Before anybody comes up a straw man, I of course agree with Wikipedia:You don't need to cite that the sky is blue. --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 18:10, 18 February 2011 (UTC)

Lets look at the requirement here. We need to provide a citation for the information provided. Where is a bit of a gray area. Personally I believe if the citations for the paragraph are all inline at the end of the paragraph it should be fairly easy to determine in most cases what came from where. The exception of course would be in cases were an editor placed a direct quote or something that required an immediate citation but other than that I think at the end of the paragraph is just fine. If there is a need to place one in the middle though because there are a large number of citations being used from sentance to sentance thats also ok IMO. The key though is it differs from the ones at the end (I would say if more than 2 or 3 are used in a paragraph). I do not think we need to overburden ourselves with citations every sentance especially were the entire paragraph or a large chunk of it is from the same reference. --Kumioko (talk) 18:05, 18 February 2011 (UTC)

Piotrus, your facts seem to be incorrect. For example, my quick scan found 19 sentences—not counting the wholly citation-free lead—in today's featured article that are not followed by an inline citation. GA rules name only five types of statements that require inline citations; reviewers may not require citations after every sentence. (See this section.) There simply is no policy, guideline, or process that requires a citation after every single sentence. WhatamIdoing (talk) 18:26, 18 February 2011 (UTC)

Thank god fot that :-), not all hope seems lost.--Kmhkmh (talk) 18:34, 18 February 2011 (UTC)

Many TFAs have waited years to be featured, standards improve. That aside, as I noted above, there are obvious-knowledge sentences that does not require citations; and properly written leads that only summarize cited key claims from the article don't require cites, neither. I don't mean that literally every sentence should be referenced; rather, that all sentences that are not obvious should be (per WP:V). --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 19:40, 18 February 2011 (UTC)

But in practice people tend source more then what's required by WP:V in a strict sense, simply because the term "likely to be challenged" and/or "obvious knowledge" is rather vague not to say almost meaningless due to our very diverse readership. Let's say I write an article about some math subject, which states several mathematical facts. Now I don't need to source them, if they are obvious to who exactly? Another mathematician specialized in the same (inner math) domain, some other arbitrary mathematician, anybody with a science degree, anybody with any university degree or just highschool degree? Or you have an article on some subject of US culture now many things in it might be rather obvious to Americans, but the same things might not be obvious to British, Australian, Indian, etc readers. Though you sometimes can have something like a main target audience which you can use to decide what's obvious knowledge and what's not, in general the border here are rather vague and fluid. Due to that (and to avoid long discussions and edits conflicts down the road) many authors reference almost everything, but to avoid extreme cluttering by having one or more inline citations at every other sentence, they often prefer to provide those sources at the end of the paragraph. So we have to keep in mind that in practice we have many more inline citations in articles than only those needed for a few passages likely to be challenged.--Kmhkmh (talk) 20:32, 18 February 2011 (UTC)

Consider, for a moment, why is WP:V there. Because nature of wiki means meany authors, not all experts, and we needed to show that our material is in fact written based on sources as good or better than traditional encyclopedias. Consider also why CITE, at this point uncontroversially, notes that general references section is not enough - because wiki can be edited by anyone, anything can be change, and thus we needed to move towards inline cites. Stopping midway at paras just creates a lot of localized "general reference" problems. Why what's bad for an article is ok for a paragraph? What's so bad about saying that each sentence, unless obvious to everyone or repetitive part of the lead, should be clearly referenced? --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 20:54, 18 February 2011 (UTC)

First of all there is nothing wrong with general references for a short article, i.e. it is not bad for the article and good for the paragraph. Second the real issue here is, that the article content can be verfified with reasonable effort (not necessarily the least effort conceivabe to a particular person) and that is achieved if the sources close enough to the content they source and there are not too many sources at once or even better that their are annotated sources. It becomes only unreasonable if you have too much text and too many sources, i.e. you don't want to read all the articles sources to just to verify a paragraph, but being required to check a very few sources or to read the source annotations to see whether the paragraph is sourced or not is reasonable.

All of that has nothing to do with the reason why we source. This is an argument on how we source and which forms of sourcing are good enough. And for that matter imho paragraph sourcing is good enough.--Kmhkmh (talk) 21:27, 18 February 2011 (UTC)

@Piotr Konieczny: If the result of a review really hinges on every single sentence having an inline citation versus the paragraph having its inline sources at the end, then frankly I must say it's a rather braindead and superficial procedure which isn't really good for anything at all. to me it looks like typical form over content approach. It doesn't seem matter wether the inline citation do actually source the article, but just that it looks like it according formal criteria made up by group of wikipedians.--Kmhkmh (talk) 18:32, 18 February 2011 (UTC)

P.S.: A better criteria than single sentence sourcing might be to require annotated inline sources (bundled or not) for GA and above articles. For normal articles I'd consider than an overkill and strongly oppose it, for GA and above reviews it might a sensible criteria and at least not as braindead as just checking whether every sentence as an inline citation. with the annotated version reviewers would at least be forced to tale a slightly closer look at the sources and read at least the annotations.--Kmhkmh (talk) 18:43, 18 February 2011 (UTC)

That reviewers (and readers in general) should review sources for reliability is one thing. Making things easier on them but making it clear which source is used for each sentence is another. If a sentence is not referenced, and I cannot find the fact in the end-of-para ref, does it mean it it has been added later, the previous editor made a mistake, or I missed something in the referenced work and need to reread it again? This confusion is a waste of time that could be avoided if the sentence in question had an inline cite (in which case it becomes obvious that either a specific ref is improperly used or it needs to be reread again). --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 19:43, 18 February 2011 (UTC)

You are making it more complicated than it actually is. If you come across a fact that is not in the inline citations of the paragraph the it is not not sourced period. Whether it was an oversight by the original editor or somebody later doesn't really matter, it just matters that it is unsourced, hence your pandering above is irrelevant. The same hold for inline citations for single sentences. An inline citation might fully source a sentence or it might. If it doesn't it might the original editors error or that by someone later on and again there isn't really much of point in pandering about it. For the reviewer it mainly matters, whether the content is correct and sourced and not where a particular error/oversight originated. The latter is only of interest if there's a need for catching malicious editor, but i don't see what that has to do with a regular review. From the view of somebody proofreading the article there isn't really much of difference between single sentence or paragraph sourcing, since the proofreader checks all the sources of the paragraph anyhow.--Kmhkmh (talk) 20:13, 18 February 2011 (UTC)

You are missing the point that we are writing those articles not for ourselves or the reviewers, but for the end reader. That reader should not be required to verify the sources, he should be able to assume good faith towards us and trust us that we have done so for him (just like Britannica and such would do so). If that reader sees that a sentence has a reference, he should be able to assume we vetted it as reliable. If he does not see a reference, that he has to go further in his assumption and assume that the end-of-the-para refer is the one that has the citation. Let's face, it few people but the reviewers and very interested editors will actually bother verify the ref. If I see a sentence, I may occasionally check to see if the source is reliable, but I will usually assume in good faith that it is correct. But if I see only an end-of-para ref, I am not trusting anything above it, because I don't know if the proceeding sentences were added by the same editor, or did they go through other edits (and I am unlikely to waste my time tracking that in the edit history). To make articles as trustworthy as possible for casual readers, inline referencing of every sentence is the way to go. --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 20:26, 18 February 2011 (UTC)

Our readers are not idiots they can work with single sentence sourcing as well as paragraph sourcing. And no inline sourcing every sentence is not the way to go for our readers it just forces ridiculously cluttered sentences upon them.--Kmhkmh (talk) 21:08, 18 February 2011 (UTC)

• I admit there is an issue here, but does citing every sentence help? Plenty of sentences contain up to say four different points, & when I do cite a single sentence, it may include several different sources (a far more common practice in actual academic writing than on WP I notice). Meanwhile it looks ridiculous, & hardly anyone will actually follow such a guideline. Johnbod (talk) 18:39, 18 February 2011 (UTC)

Exactly! If somebody wants to sozrce every single sentence in his article, let him, but mandating this somewhat questionable practice that also creates a lot of clutter to other authors is a no go.--Kmhkmh (talk) 18:48, 18 February 2011 (UTC)

I don't see a problem; if a sentence contains multiple points from multiple sources, this is when the mentioned WP:CITEBUNDLE comes in place. --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 19:45, 18 February 2011 (UTC)

The point is that "one citation per sentence" is insufficient for some cases, and overkill in others. It's therefore not a useful rule for us to put forward. WhatamIdoing (talk) 21:15, 18 February 2011 (UTC)

So what is? Is one-ref-per-para a better rule of thumb? CITE notes that "general references" are bad, but is not very clear on what the recommended density of references is. --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 21:23, 18 February 2011 (UTC)

The rule of thumb is that source usually should be at least paragraph based, so that the effort for checking a prticular statement is within reason. General references are bad if the article is longer and you have references in particular if the are unspecific (like books without page numbers). There is no problem however with a rather short article that uses a few specific general references at the end.--Kmhkmh (talk) 21:35, 18 February 2011 (UTC)

Yes, that's the current policy; there should be at least one reference per paragraph, and every claim in the paragraph should be covered by that reference. Jayjg ^(talk) 21:54, 18 February 2011 (UTC)

I've suddenly gotten tired of repeating this. There is no grammar-based rule. Please go read WP:MINREF. WhatamIdoing (talk) 06:57, 19 February 2011 (UTC)

Why would that personal opinion be of any value in this discussion? That's an essay, we're discussing policy here. Current policy is that there should be at least one reference per paragraph, and every claim in the paragraph should be covered by that reference. Jayjg ^(talk) 00:57, 20 February 2011 (UTC)

Because no policy requires that. Really.

If you don't believe me, then please go try to find the policy—any policy—that says every paragraph must have at least one inline citation. If you can't, after diligent search, actually find any such claim, then I ask that you please quit spreading this myth around Wikipedia. WhatamIdoing (talk) 18:16, 21 February 2011 (UTC)

Every paragraph does need at least a ref at the end. Any article going through FAC will fail without that, and I assume it's the same at GA. CITE says: " ... adding the citation to the end of the sentence or paragraph is usually sufficient ..." I don't know whether this is enshrined anywhere in policy, but it's definitely best practice. SlimVirgin ^{TALK|CONTRIBS} 18:37, 21 February 2011 (UTC)

Policy is a complex combination of both written policy and common practice. No article can pass FA without at least a citation at the end of each paragraph, and FAs are the articles that most closely adhere to Wikipedia policy. You're not doing Wikipedia any service by "spreading" the "myth" that this is not required. Jayjg ^(talk) 19:24, 21 February 2011 (UTC)

Last time I checked, FACR wasn't a policy. And WP:GACN has always disclaimed this standard. The GA criteria require inline citations for five types of content, not for any type of grammatical structure. This means more than one inline citation per sentence in some cases, and zero inline citations per section for others.

I agree that the liberal use of inline citations is a beautiful thing, and I cannot remember the last time I wrote an article with fewer citations than paragraphs. However, this is not actually required. For better or for worse, current policy explicitly permits the exclusive use of WP:General references (assuming no BLP issues, no direct quotations, no actual challenges, and the editor doesn't expect any challenges).

Requiring the use of inline citations has been proposed and rejected repeatedly. When the community repeatedly and directly declines to enshrine something in both guidelines and policies, then that something is not the community's policy. WhatamIdoing (talk) 19:32, 21 February 2011 (UTC)

Written policy often lags behind best practice, and it's clearly best practice when you're trying to make an article decent to have a ref at the end of each para at least, in the interests of text-source integrity. The level of the paragraph is the longest we expect readers to search for the source. If articles are being promoted to GA without this, that's a pity, but are they? And where has a requirement for inline citations been proposed and rejected? SlimVirgin ^{TALK|CONTRIBS} 20:27, 21 February 2011 (UTC)

Here, for one. Didn't you notice Piotrus trying to push that into the guideline last week? Didn't you see it removed?

Do you remember TCO's efforts last month at WT:CITE/example style? "First, let's go inline cites. Most people do it, expect it. Let's just do it." And the one-size-fits-all approach was rejected with comments like these.

If inline citations were actually required for every paragraph, we'd actually come right out and say that. There's a big difference between what is required and what we want to encourage. I encourage at least one inline citation for every paragraph; I do not pretend that they are actually required. WhatamIdoing (talk) 18:37, 22 February 2011 (UTC)

He tried to add that inline citations were required after every sentence, which they're not. But they are required, and this page does imply that one's needed after every paragraph. I can't see any reason not to formalize that, because it's a very minimal requirement. If you have several paragraphs relying on one source in paragraph four, it's almost the same as supplying a general reference at the end of the article, which isn't allowed, at least not for anything well developed. SlimVirgin ^{TALK|CONTRIBS} 19:29, 22 February 2011 (UTC)

If you think that the community is going to raise its standards to require one inline citation per grammatical unit, even if said grammatical unit does not contain any (1) direct quotations, (2) information about BLPs, (3) material that has been challenged, or (4) material that is WP:LIKELY to be challenged, then feel free to propose that over at WP:V. I expect it to be received enthusiastically by a significant minority, and to once again fail to gain consensus in the end. WhatamIdoing (talk) 06:29, 23 February 2011 (UTC)

You keep referring to consensus for this failing, but it has never failed that I'm aware of. Can you say what you're referring to? SlimVirgin ^{TALK|CONTRIBS} 20:32, 23 February 2011 (UTC)

Sources of any kind are only required for quotations, challenged or likely challenged material, and some biographical information. If a paragraph has none of these things, there is no policy that requires an inline citation of any sort in that paragraph. That is intentional: WP:V is a minimal requirement, not a description of how the "best" articles might be written. The requirements of FA and GA are free to go beyond what is actually required by policy, and they often do. Pages like this guideline need to stick to what's actually required, not the FA or GA requirements. — Carl (CBM · talk) 22:47, 22 February 2011 (UTC)

User:Piotrus recently made several changes to this page that are dramatic in scope, even though there was no previous discussion that I can see (scrolled off perhaps?), no prior agreement (ditto?), and what appears to be at least some level of disagreement.

I believe these edits were made as part of a thread I am involved in. In a recent DYK entry of mine, Piotrus gave the article a ? for several reasons. One of these was his claim that in order to pass a DYK, every sentence in the article had to be referenced. User:Nyttend quickly replied that this was not the case, and I followed this up with a post on Piotrus' talk page to the same effect. Neither comment received a response.

Instead, a refimprove tag was placed in article in question, shadow mask. When I asked why he did this, he once again claimed that every sentence needs a reference. I again pointed out this was not the case, and removed the tag. Then I received a message claiming that I should check CITE. Sure enough, CITE now included text to this effect -- text he had just entered, which used text from the shadow mask article as an example!

The changes were RVed, but only with a post on my talk page suggesting material would be removed from the article wholesale. This is all very worrying.

Maury Markowitz (talk) 18:23, 18 February 2011 (UTC)

Every sentence doesn't need a citation after it. We only need references for anything challenged or likely to be challenged, and refs at the end of the paragraph are often sufficient. SlimVirgin ^{TALK|CONTRIBS} 18:40, 18 February 2011 (UTC)

(edit conflict) WP:DYK is a completely different duck. In the case of something featured on the front page, requiring that specific information to be cited is a case of necessity. That's why an article goes through a vetting and validation process in order to get promoted to the front page. By requiring this form of explicit citation, we are pretty much guaranteeing the truth of the statement. Think of it as truth in advirtising.

Of course, this only refers to the information specifically in the hook. Every sentence in the article doesn't need to be cited, but every sentence in the hook does. The whole thing you described with the refimprove tags is a little crazy, and I think it's a probably more a case of a conflict between editors than anything else.

Now, when it comes to everything not in the scope of DYK, this obviously does not apply. Consecutive statements, if non-controversial, can have the common references consolidated.

My two cents. --Mûĸĸâĸûĸâĸû _(blah?)

(edit conflict) Maury, please stop building straw man. I clearly said at DYK that more inline references is not related to DYK eligibility, and I said so again on talk. I don't understand why you are making a case out of my suggestion about how the article should be improved in the future. And a [[tl|refimprove}} tag is quite appopriate for the article (although I did not revert it removal, because I am still hoping you'll do the right thing and improve the article instead of making a storm in the teacup and arguing why it shouldn't be improved). You could've added the inline cites I asked for in half the time it took you to write all those complains on my talk page and here. --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 18:53, 18 February 2011 (UTC)

That is not the issue. The issue is that you:

edited CITE, a core document of the Wikipedia, so it matched your personal view of what it should say

that you carried out this action without any prior discussion on this talk page, although the page clearly states to do so

did so in spite of being repeatedly told that your interpretation was wrong (as evidenced here)

and did so apparently for no reason other than to win an argument with another editor

This is serious Piotrus. I would advise you not to be so glib. Maury Markowitz (talk) 19:53, 18 February 2011 (UTC)

Stop twisting the facts, will you? I self-revert my edit to CITE when I became aware of the discussion (I am still very surprised my addition was controversial). There was no dispute to win - I even passed your DYK! The only dispute is that I said the article could be further improved by having more inline cites, and you are making mountains out of molehills out of it. If you don't care to improve your article further, nobody is forcing you, but stop attacking people who are trying to improve this project. --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 20:00, 18 February 2011 (UTC)

(ec, reply to Mukka) Yes, that's true enough as a matter of common sense. When I've submitted DYK, I've added a ref in the article specifically after the hook to make sure it was easy to verify for the person checking it. Once the DYK was over, I often removed it again and placed it elsewhere. But that's an internal DYK matter, not to be extended elsewhere or added to CITE. SlimVirgin ^{TALK|CONTRIBS} 18:54, 18 February 2011 (UTC)

Is the hook a text piece separate from the article or an article quote?--Kmhkmh (talk) 18:56, 18 February 2011 (UTC)

The "hook" is the text that appears on the Main Page. It's usually a sentence or a combination of one or more sentences in the article, often slightly reworded. DYK rules require the sentence (or sentences) that is mostly closely aligned with the hook to have an inline citation, and for the DYK reviewer to verify that the cited source actually supports the sentence. WhatamIdoing (talk) 19:04, 18 February 2011 (UTC)

Some sanity at last. :-)--Kmhkmh (talk) 22:34, 19 February 2011 (UTC)

I'll put this at the bottom since I just read it: "First, let me point out that an article, nowadays, is unlikely to become GA (or higher) as long as it has any unreferenced sentences." I assume Piotr means every sentence within a paragraph as opposed to the end of the paragraph if one citation covers the entire paragraph? This is simply untrue. Period. SandyGeorgia (Talk) 08:58, 19 February 2011 (UTC)

I agree, but even if it was true, the requirements to be "GA or higher" are not the same as the policies that all articles are required to follow. This guideline is not about GA requirements, which are at WP:GA? or FA requirements, which are at WP:FA?. — Carl (CBM · talk) 22:51, 22 February 2011 (UTC)

If you have qualms about the beginning of a paragraph sourced only by a footnote at the end, go look at the sources in the footnote. If, when that is done, you don't find the content, then add {{cn}} or {{fv}}. If you do, add a sentence to the footnote explaining what is sourced where, if it seems helpful - often an entire paragraph or an entire article derives from the same two=page source. Septentrionalis PMAnderson 22:34, 18 February 2011 (UTC)

Assuming, one has access to the sources. —  HELLKNOWZ  ▎TALK 10:28, 19 February 2011 (UTC)

Indeed. With offline sources, you are really left with assuming good faith, because verification becomes very hard. --_{Piotr Konieczny aka Prokonsul Piotrus| talk} 04:18, 21 February 2011 (UTC)

Not really. That's what {{vn}} is for. If you can't get to the source, then you can ask for help. WhatamIdoing (talk) 18:39, 22 February 2011 (UTC)

Even so, WP is full of tags with backlogs of years. Finding obscure sources is not the most pressing matter. —  HELLKNOWZ  ▎TALK 20:21, 22 February 2011 (UTC)

I am against general references sections being used as citations. (see below #General references in FAs). As for the insertion of text. I assume that if that is done then the citation covers everything from the last citation or the start of a paragraph. So for example if there is a paragraph such as:

Here is my first sentence. Here is my second sentence. Here is my third sentence.^{[p 1]}

The if a new sentence is interposed between the by "Anne Other" she would do this:

Here is my first sentence.^{[p 1]} Anne Other's sentence.^{[p 2]} Here is my second sentence. Here is my third sentence.^{[p 1]}

A possibility to help remove ambiguity would be is to use a template that covers the specific text is being cited, but to date the examples I have seen change the visual look of the text which I do not think is desirable.

Another possibility, is to alter the positioning of footnotes at punctuation, {I anticipate a groan going up from some editors :-) } so that if it is to the left of a full stop (period) it means the citation is only for the information in that sentence and if it is to the right of a full stop it means from the last citation or the start of a paragraph (which ever is nearer to the citation under discussion). So in the example I gave above:

Here is my first sentence. Here is my second sentence. Here is my third sentence^{[p 3]}.

The if an Anne Other adds another sentence it becomes:

Here is my first sentence. Anne Other's sentence^{[p 2]}. Here is my second sentence. Here is my third sentence^{[p 3]}.

In which case my first and second sentences are still not cited.

At the moment it is impossible to tell if a citation at the end of a paragraph is intended to cove all the sentences in the paragraph or if it is only a citation for the last sentence. It would also mean that if there was a reference at the end of the paragraph that was to the right of the full stop that covered the whole paragraph if Anne Other added a sentence in the middle but places the citation to the left of the full stop, there is no confusion as the sentences near the start of the paragraph seem to have no citation but a check of the history would show that all that was needed to fix the problem would be to append the final citation on the paragraph to the end of the sentence directly before Anne Others addition.

-- PBS (talk) 02:40, 24 February 2011 (UTC)

It's not "impossible" if you read the source (RTFS). The idea that it should be possible to tell what is in the source without actually reading it is misguided. Readers who want to verify the article should be reading the sources anyway; readers who don't want to read the sources have little ground to worry about what they say. The punctuation thing is too fragile to be broadly adopted. Also, I always put general citations on the first sentence, not the last one. — Carl (CBM · talk) 02:48, 24 February 2011 (UTC)

I havn't see it done is it common in a subject area? -- PBS (talk) 03:05, 24 February 2011 (UTC)

I don't know if it's common, but I'm used to seeing it in science and math articles in real life, where I usually see a citation put on the thesis sentence of a paragraph, not on the last sentence. These are often citations like this [1] rather than footnotes. Since that's the way I write real papers, if I can pick the citation style for a WP article I do the same thing. — Carl (CBM · talk) 03:18, 24 February 2011 (UTC)

May a bot be approved which will make alterations that would be inappropriate within quotations and titles, if the bot protects citations that use templates, but does not protect plain citations? In particular, the APA style does not use quotation marks or any other special typography to designate article titles, and thus a bot would be unable to avoid making alterations to such a title. Jc3s5h (talk) 14:33, 20 February 2011 (UTC)

This question is inspired by Wikipedia:Bots/Requests for approval/MOSNUM Bot. I believe WP:CITE rather than WP:MOS controls citations, and WP:CITE does not give preference to templates compared to plain text for purpose of citations. Thus I believe a bot which would protect templated citations but inappropriately alter plain citations ought not to be approved. It is possible that the proponents of the MOSNUM bot might successfully argue that the change that this bot would make, a change in date format such as "20 February 2011" to "February 20, 2011", is rare enough in titles of external articles, combined with the fairly low usage of the APA style, that the level of false positives is tolerable. I personally would not favor allowing this because of the risk of edit warring between the bot and editors. Jc3s5h (talk) 14:33, 20 February 2011 (UTC)

• I wish I understood this. :) SlimVirgin ^{TALK|CONTRIBS} 16:21, 20 February 2011 (UTC)

Let me illustrate with an example. In today's online Wall Street Journal we find an article that would be cited using APA style as:

Huetteman, E. (2011, February 19). Sunday Breakfast Menu, Feb. 20. New York Times. Retrieved from http://www.nytimes.com/pages/politics/index.html

Along comes a date formatting bot. It knows that dates within citation templates, URLs, or quotation marks should not be altered, but is not designed to deal with plain citations. It is going through a category where the mdy dmy format is usually used, such as the United States Military category, so it changes the citation to read as follows:

Huetteman, E. (19 February 2011). Sunday Breakfast Menu, 20 February. New York Times. Retrieved from http://www.nytimes.com/pages/politics/index.html

The question is, should a bot that acts this way be approved? Jc3s5h (talk) 16:59, 20 February 2011 (UTC)

You meant "dmy format"; that's the simplest way to explain why the rest of the example changes mdy to dmy. Art LaPella (talk) 17:14, 20 February 2011 (UTC)

So it's going to change from mdy to dmy for no reason, for any citation not inside a citation template? SlimVirgin ^{TALK|CONTRIBS} 17:22, 20 February 2011 (UTC)

The bot that inspired this RfC would change from mdy to dmy because the article is in the category United States military, and the United States military usually uses the dmy format, so there is a reason. The question for the RfC is broader: Should any bot that knows how to avoid making undesirable changes in citation templates, but does not know how to avoid making undesirable changes to plain citations, be approved? Jc3s5h (talk) 17:37, 20 February 2011 (UTC)

Okay, I see your point now, thank you. I agree with you, the bot should know not to make bad changes to manual citations too, or shouldn't be approved. SlimVirgin ^{TALK|CONTRIBS} 17:51, 20 February 2011 (UTC)

• Bots should not edit non-templated citations if their task excludes editing of citations. I think that much is obvious. A more relevant question is if the bot is allowed to have a false positive if it is impossible to always detect the citation; such as, if it appears outside <ref></ref> and References/Bibliography/etc. sections in plain text. —  HELLKNOWZ  ▎TALK 18:59, 20 February 2011 (UTC)
• Bots should not edit, except for sharply-defined, consensus, non-controversial tasks, which require neither understanding nor judgment. This is not one.
• One example of this is the proposal (as part of this bot request) to switch dates from the format July 4, 1776 to 4 July 1776 on articles in several categories, including (as yet unspecified) subcats of Category:Military of the United States. The general intention is probably helpful; but George Washington and his subordinates (hundreds of articles) are in many of those subcats - and secondary sources do not generally use 4 July 1776 in writing of Washington; it would be an anachronism. Nobody on either side of the Atlantic used that for a century afterwards; so secondary sources don't either. (They do use the 4th of July, 1776; but the proposed bot would remove that too.) Septentrionalis PMAnderson 23:32, 20 February 2011 (UTC)