Wikipedia:Centralized discussion/Citation discussion

Wikipedia citation style

Wikipedia does not have a preferred citation style. I am referring to a relatively tight definition of citation style: the data elements required according to the type of source, the syntax and sequence of those data elements, the punctuation and abbreviations used, and the text formatting.

I have not read all the back history, but what I have read suggests that editors have proposed using a single citation style, but those proposals were not adopted. As it stands now, the editor that adds the first citation to an article establishes a citation style and subsequent editors are supposed to follow that style. In practice, it's difficult to maintain a high-degree of consistency within an article due to

• complications caused by the lack of a standard style,
• editors who are unfamiliar with citing sources or unfamiliar with Wikicode,
• an evolving infrastructure,
• and other factors.

The variety of citation style options means that editors who prefer a particular style can have their way. Unfortunately, the overall result is a big mess.

• Citations are inconsistent within and across articles,
• the guidelines are confusing,
• there are too many templates,
• the template documentation is confusing and wishy-washy,
• and the guidelines and templates require more effort to maintain than they should.

Citing sources is fundamental to Wikipedia—"Encyclopedic content must be verifiable"— and the whole situation is an embarrassment.

Establishing a standard citation style will not cure all ills, but it's the best first step. Eliminating style options will simplify the guidelines and the other instructions related to citing sources, will reduce the number of templates required, will help improve in-article and across-article consistency, will reduce editing conflicts, will reduce the barriers for inexperienced editors, and in general, will reduce the time and effort devoted to non-productive work related to citing sources. The details of the citation style—punctuation, etc.—are relatively unimportant but it's critical that we reduce the clutter, improve the documentation, improve the tools, and minimize the barriers to presenting citations effectively so we can spend more time adding reliable sources and less time working around the problems caused by the lack of a standard.

Serious, reputable, professional organizations that cite sources use a single established citation style. It's past time for Wikipedia to do the same.

I propose that we establish a standard citation style. Given the prior difficulties reaching consensus, I further propose that we form a small committee to define the standard. The committee should be free of people with entrenched preferences and instead include contributors who will work collaboratively, who will compromise, and who will place the primary goal ahead of picky details. — John Cardinal (talk) 18:28, 15 January 2010 (UTC)

Comment from Ohms law

You know, I was thinking about this exact same topic just recently. It's always struck me as somewhat strange that we're so reliant upon citation styling from the academic/publishing/paper-based world here. I can definitely see looking to APA/Harvard/Chicago/whatever for guidance towards our own styling, but for us to defer to their styles here, in the ad-hoc fashion that we do, and disregard the obvious differences in platform and needs seems particularly short-sited on our part. I'm somewhat anxiously awaiting to see what sort of reaction this draws.
— V = I * R (talk to Ohms law) 19:16, 15 January 2010 (UTC)

Comment from Shakescene

¶ (edit conflict): Some things just don't lend themselves to being cited usefully in the same way as other things, and I'm not keen on a uniform citation style. In some cases, for example, the place of publication is important; in others it's either redundant, obscure or confusing. The name of a multi-national publisher that's still in business (e.g. Harper Collins or Oxford University Press) is now usually much more useful than a nominal city of publication; on the other hand, the name of a small printer/publisher/bookseller who's been defunct for centuries (e.g. William Jaggard) is less useful than just giving the city of publication.

The academic style of listing a set of works in one place, and then using the author's last name and sometimes a date in individual footnotes, works very well for some articles, but not very well in others, such as those dependent on a wide variety of books, periodicals and on-line sources. In particular, I dislike listing newspaper reporter's names in the last-first format, because it makes the citation needlessly clumsy and confusing (e.g. "Chan, Sewell, Jones, Carol M., and Smith, Jonathan W., "Bloomberg and Paterson fail to agree on budget", The New York Times, City edition, Sunday, January X, 20XX, page ZZ, (retrieved January Y, 20XX).") In some lists (e.g. List of tallest buildings in New York City), every line, following the WP:Featured list criteria, has a footnote, often to the same source; by contrast, when I was editing New York City mayoral elections, I was sorting out and fixing individual lines and candidates for each election from a different mix of the same set of sources (listed at the bottom), and footnoting every single datum would have cluttered much while clarifying or confirming little. (There are still many footnotes for some very specific points that don't apply to every election.) Sometimes, it's useful to know what day of the week, or what edition, of a newspaper is being used (e.g., a Sunday supplement, or how many days before or after a Tuesday or Thursday election); at other time's it's not. I prefer to fit my citations to the article and to the particular source without deviating radically from others' citation style.

This may seem to be a slipshod and messy-looking philosophy, but the uniformity of standard print publishers and publications is usually gained by truncating, suppressing or jumbling an item or items that are worth presenting more distinctly in some contexts.

However, there are some things that should appear in every, or nearly every citation: author, name of work and date of publication, plus the ISBN (International Standard Book Number) where available (usually after 1970) to help readers find their own copy. Where not obvious (as in Concise Oxford Dictionary), a book's publisher or place of publication should also be included. Although others find it tedious, I believe in providing the most-recent date of successful retrieval for on-line sources.—— Shakescene (talk) 20:03, 15 January 2010 (UTC)

I have to agree: some of the fields have their own citation styles and would be reluctant to give them up. Occasionally there are good reasons for these styles.

I think we'd have a hard time agreeing on a common style: I, for one, would push to minimise redundant punctuation and other formatting. I think some people would disagree with that approach. Tony (talk) 10:26, 19 January 2010 (UTC)

Data vs. Style

Just theorizing a bit here.... Seems to me that for this discussion it would be good to separate conceptually the data that goes into a citation from the style of that citation. This matters because, in a really blue-sky MediaWiki way off in the future, I can envision that you might be able to add citation data in certain ways that lets the user choose a favored style. OK, like I said, that's blue sky and I don't think there's any priority to separate citation data from its form any time soon, but I point that out as a starting point. The practical implication is that a "canonical best citation style" would need to accommodate all the data that any of the present styles accommodates, or at least any data that's deemed relevant to users of Wikipedia (and probably enough free text fields to painlessly accommodate needs that that weren't foreseen).

Lest I be getting too theoretical here, let's draw an analogy with a similar "religious" issue -- national variants of English spelling. In a theoretically optimal future Wikipedia, there's no reason why the author's locale of US, UK, AU, CA, whatever should bind the reader to read that same variant of English. (I realize that automated localization has its limits, it's more than just rendering colour as color for us Yanks, but play along with my thought exercise for a moment.) In a theoretically perfect MediaWiki/Wikipedia, a UK user should be able to add data that's rendered to me in US English, because they're adding actual data and not just words.

I don't know that we'll ever reach a future state where the citation data is freed from its format, but it seems to me that a discussion of idealized formats should at least start by recognizing embracing fluid thinking about the relationship between data and format. (I also think this is related to publishers' "truncating, suppressing or jumbling" in print -- print imposes a premium on brevity, but electronic distribution less of one. Ideally, whether you're printing hard copies of Wikipedia pages or just viewing them, you should be able to get only the level of citation detail you desire.) - Regards, PhilipR (talk) 07:18, 16 January 2010 (UTC)

Phillip: actually, you're not being all that far-fetched. citations are (and have been for a long time) structured, delimited text. for instance, APA style is period delimited: you get <authorname>. <year>. <articletitle>. <journalname>. <volumeissue>. for journals and <authorname>. <year>. <bookname>. <publisher>. for books. it should be fairly easy to specify the standard data chunks needed for any citation and then swap them around programmatically for different kinds of citation. I'm more curious about what the best format for online-citations would be - do we stick with the classic (print-media-derived) linked superscripts, or do we want to use internet technologies (using tooltips, popups, or some other non-linear modality that makes browsing through citations smoother). --Ludwigs2 09:23, 16 January 2010 (UTC)

Comment from Hoary

Wikipedia does not have a preferred citation style [ . . . viz] the data elements required according to the type of source, the syntax and sequence of those data elements, the punctuation and abbreviations used, and the text formatting.

Very true, but this mixes the important with the unimportant. As an extreme example of the latter, consider ", pp. 245–46" versus ", pp. 245–246" versus the same pair minus the "pp.". Any one of these should be immediately understandable.

Unfortunately, the overall result is a big mess.

* Citations are inconsistent within and across articles,

I am as profoundly indifferent to inconsistency of citation style across articles as I am to the inconsistency of citation style between one journal and another, or to the inconsistency of spelling (US, Oxford British, "z"-avoiding British, other) from one article to another within en:WP. Or of course to the date formatting stuff.

Citing sources is fundamental to Wikipedia—"Encyclopedic content must be verifiable"— and the whole situation is an embarrassment.

You're almost right. Often there's no source, and often a crap source is cited. But the part of the situation that you seem to be concentrating on here fails to embarrass me in the slightest.

Eliminating style options will simplify the guidelines and the other instructions related to citing sources, will reduce the number of templates required, will help improve in-article and across-article consistency, will reduce editing conflicts, will reduce the barriers for inexperienced editors, and in general, will reduce the time and effort devoted to non-productive work related to citing sources.

It seems to me that you are half right, half wrong.

The details of the citation style—punctuation, etc.—are relatively unimportant

I'm glad you said that.

Serious, reputable, professional organizations that cite sources use a single established citation style. It's past time for Wikipedia to do the same.

But you haven't explained why. I don't think that acquisition of some of the trappings of serious, reputable, professional organizations will either make Wikipedia a serious organization or can make it reputable, or anyway reputable among people with the intelligence to look beyond mere trappings. (And of course Wikipedia will remain amateur, though I wouldn't resent receipt of a wodge of back pay if the Foundation has too much moolah on its hands.) -- Hoary (talk) 09:57, 16 January 2010 (UTC)

I thought I did. At the top level, not having a standard style means we waste a lot of time and effort. The absence of a standard style makes the guidelines harder to write and that confuses editors. The absence of a standard style means there are dozens of templates when far fewer would do, and those templates have to be maintained. The absence of a standard style leads to multiple styles in the same article. — John Cardinal (talk) 14:17, 16 January 2010 (UTC)

I agree very much with this summary - there is given no other reasoning than cosmetics. And I disagree with John Cardinal's reasons given here - on the contrary implementing a single citation style would waste much more work on unimportant cosmetical issues that should rather be used for finding good reliable sources and citing them in whichever style works.·Maunus·ƛ· 14:50, 19 January 2010 (UTC)

Strong agree: From what I have seen there is entirely too much focus on style and entirely too little on content for the 'ped as a whole, not just this topic. Endless editwars and flamefests have been started over interpretations of the MOS, waring that has absolutely no effect on the actual content or readability of the article. As long as the editors focus on this sort of minutiae, we'll continue to miss the forest for the trees. Maury Markowitz (talk) 19:56, 1 February 2010 (UTC)

Maury's sentiment is fine, but the fact remains that enough editors pay attention to style to make it an issue. Indeed, at least in the USA, undergraduate university students are taught to pay careful attention to citation style; it often seems professors pay more attention to whether the sources are cited correctly than the meaning of papers submitted. Writing pleas on an obscure project page isn't going to make editors reduce their emphasis on style. --Jc3s5h (talk) 20:36, 1 February 2010 (UTC)

Preach on, brother; preach on! Jc3s5h is correct above as well, of course. the current "dont' worry about it, it's not important" attitude is just... ridiculous, really.
— V = I * R (Talk • Contribs) 22:52, 1 February 2010 (UTC)

Comment from Rannpháirtí anaithnid

As a matter of style I agree. One way in which all of the points raised above could be achieved any by others is by preferring (making mandatory?) the user of {{cite}}, {{cite web}}, {{cite journal}}, etc. If these could be combined into just one single citation template it would be very good. Not only would it help greatly to uniformly present citation information but it would encourage proper citation (how often is a raw URL given as a citation when {{cite web}} is so easy to use).

I completely agree with Ohms law. There is no need for us to conform to existing styles. In is the usefulness of the information that is important, and consistency in presentation, but we can choose our own style for our own purposes. Our own citation style might smooth a transition to a single citation template. Or, like PhilipR says, if we used a single citation template the user could choose to display citations in their own preferred format.

A bit of dreaming here: If a citation template was integrated into MediaWiki (rather than just being a template) more snazzy things could be done. For example, a user could merely enter the ISBN and page and the software could pull out the rest. Less desirably - but better than the current situation - a user could enter a lazy URL and the software could pull out the page title, format it a little and give it an access date.

What would be wrong with the following (and allowing the software:

• <ref first="John" last="Smith" title="The Book of Wisdom" publisher="Harper Collins" place="London" year="2009" isbn="1234567890-123" />

Or the following and allowing the software to do the rest:

• <ref isbn="1234567890-123" />

(In the latter case the software could 'fill in' the attributes before committing it to the server so it would not have to perform DB calls to get the reference data every time it wanted to display it.)

--rannṗáirtí anaiṫnid (coṁrá) 10:13, 16 January 2010 (UTC)

I disagree that {{cite web}} is so easy to use. Maybe I'm missing a feature somewhere but how do I insert it with a single click? The documentation could use improvement. When for example is a company listed as the author rather than the publisher? Also when an author is not given why does the placement of the date change? The current rules governing it doesn't make any sense as it is. It seems to follow most closely the APA style but then diverges in illogical ways like with the use of semicolons and the aforementioned change of date placement. It's confusing. I do like that it encourages including the whole first name rather than just an initial though and feel that the accessdate parameter in combination with URL hyperlinking is superior to the conflicting recommended methods of the established organizations. Typing out the website address after "Retrieved from" is largely unnecessary for Wikipedia purposes although if pulling it from an indirect source or archive it might still be a useful format to retain. Lambanog (talk) 10:41, 16 January 2010 (UTC)

There are problems with citation templates:

• A good article is likely to have footnotes to a lot of sources.
• Sources, even reliable sources, come in many flavors, and have many possible complexities.
• Templates designed to anticipate complexities are likely to be gruesomely large
• Citation templates are already so cumbersome that even when they are sufficiently flexible I for one never use them. I find it next to impossible to write coherent prose around them, and instead must constantly switch between the edit window and the preview, which is tiresome.

Want an example of complexity? I often use Japanese-language sources that -- either to help shopkeepers and librarians who can't read Japanese script, or merely for decoration -- have alternative titles in (quasi-) English. For each of these, I want to give the Japanese title in Japanese script (without which an OPAC is unusable), the Japanese title in roman script, and the (quasi-) English alternative title (skipped by most Japanese OPACs but sometimes usable in non-Japanese OPACs). And there could be two versions of the Japanese script: the original orthography and the modern equivalent.

And that's just one aspect of Japanese sources. There's also the matter of the names of Japanese authors of Japanese sources. Et cetera. And all this is just for Japanese.

The LSA (for example) can have a simple style sheet because its authors know (i) that there'll be one subject matter, linguistics, and (ii) that if linguistics authors need to cite something very unexpected, they'll use their own brains, perhaps aided by something like Chicago.

Aside from its unfortunate chapter on "grammar and usage", Chicago itself is an excellent book. It doesn't provide comprehensive, detailed, single, rigid prescriptions for sources in Japanese or any other script, because, as for any intelligent style manual written for intelligent people, its editors know that they can't foresee everything and that intelligent readers can use their own heads.

But you can attempt to prescribe everything in inflexible templates that cater for all conceivable sources (and remember that blogs, LP liner notes etc are acceptable for certain purposes). I think that if you succeeded at this you'd have created the equivalent of a two-thousand-page book. Me, I wouldn't be bothered either to help in its creation or to use the resulting templates. But perhaps I'm unusual, and the innocents who now clearly haven't a clue about citation will be happy to go through monster citation templates. (My own guess is that they'll exclaim "WtF?" and give up even faster than I would.) -- Hoary (talk) 10:48, 16 January 2010 (UTC)

The use of these dreadful citation templates is a cancer on WP. I advise all editors to write out the citations nice and simple, WYSIWYG, so it can be properly controlled. There was no good reason in the first place to create these templates. They should be damned. Tony (talk) 11:01, 16 January 2010 (UTC)

The sorry thing is that editors don't "write out the citations nice and simple" and where they do they don't do so in a uniform fashion. --rannṗáirtí anaiṫnid (coṁrá) 11:35, 16 January 2010 (UTC)

Also, references are different from the perspetive of data to the rest of an article in the sense that they are made up of desecrate data that has specifically defined meaning. Templates (or XML referencing like above) are useful because they break the reference into that discrete data. This has advantages. It means, for example, that the data can be reused as suggested above so that a user can choose how they want references to be displayed. A redistributor could choose to show references in another fashion (e.g. the could show an index of cited authors at the end of a series of articles). Or we could use the data in more imaginative ways. We could, for example, list all articles that have contain a specific reference.

When users "write out the citations nice and simple" the data becomes effectively dead as information for anyone except the human reading it at that specific time. And even then, far too often so little information is given (or given so inconsistently) that it is difficult for that human to comprehend either. --rannṗáirtí anaiṫnid (coṁrá) 12:00, 16 January 2010 (UTC)

I agree with Tony. Citation templates are some of the worst things to have happened to Wikipedia in terms of the writing. It's impossible to write well with them in the text, or to copy edit for flow, so articles become lists of sentences instead of flowing narratives. Added to which, they significantly slow down load time when there are lots of them—I can now tell which articles contain citation templates before I see them, because I have to sit here counting the seconds before they open. So please, if we are going to discuss a uniform citation style, let's separate that discussion entirely from the promotion of citation templates. SlimVirgin ^{TALK contribs} 15:00, 16 January 2010 (UTC)

See my response in the next section. SandyGeorgia (Talk) 15:09, 16 January 2010 (UTC)

See example in my sandbox using the code that Izno pointed to below. It separates the citation from the content making it far easier to copy edit (and to write) text regardless of whether one uses citation templates or not.

On the matter of load times, if the "template" was an extension to MediaWiki rather than a transcluded template then the over head would be imperceptible to a human. --rannṗáirtí anaiṫnid (coṁrá) 20:47, 16 January 2010 (UTC)

I agree with SlimVirgin above. I detest the cite templates—won't use 'em, won't even work with 'em—because they don't have any flexibility, they encumber the loading, and so on. I type out all my citations, and find it quite easy to do--much easier than using either the cite templates or the named refs. And I object when a bot comes through and adds named refs, too. But the named refs thing is another bag of worms (won't go into that here). I'd be in favor of mandating consistent information, and understand that some editors will always not do it, either intentionally or accidentally. Author. Title. Location: Publisher, date, ISBN (or...) place in the text. Etc. Also, someone above said that all reputable professional organizations have a style and stick to it, and so should Wikipedia: I should point out that those styles are discipline dependent: APA, MLA, AHA, etc. Pigs may fly before you will get agreement on what the style should be. The best we could possibly hope for is to get everyone to put in some basic information. I'm reviewing an article right now in which the editor used the cite templates and more than half the entries are inconsistent. And the idea of making an article a combo of styles, cite templates and free-form refs invites inconsistency. Auntieruth55 (talk) 23:23, 16 January 2010 (UTC)

I'm (pretty much intentionally) ignoring the commentary about the templates (see #Stay on track, below). The comment about the styles being discipline dependent are directly on point though. One piece of this issue, in my mind, is the fact that said discipline dependency is not at all related to Wikipedia. Wikipedia is not a newspaper, medical journal, academic paper, etc..., after all. I understand, and even appreciate, arguments such as "well, the way <insert favorite style guide here> does this is to...", but we should have our own clear cut style manual, which is comprehensive. The MOS is our own style manual, but we intentionally seem to exclude quite a bit from it, often for the (in my opinion poor) reason you stated above because people seem to think that cooperation on these issues is impossible. I think that such views are wrongheaded and damaging to the community, because ignoring issues is never a solution and inevitably makes the problem worse.
— V = I * R (talk to Ohms law) 00:14, 17 January 2010 (UTC)

Lambanog, that's some good criticism and I agree a good deal with it. There is confusion over what goes into what field like the examples you give. If the template came with a style guide I think it would help. A single "Wikipeida referencing style" could cure that. Some of the questions your raise (e.g. to initialise first names or not) can be decided by software if such a style was agreed upon (e.g. first="John" last="Murphy" can be written by the software as "Murphy, J." "J. Murphy" or even "J.M." if you liked). It's better to have more data and to trim it down afterwards to suit style.

Hoary, good points too. I'm not suggesting however to make use of such templates mandatory - just to prefer them to free-form refs. In the case of most URL references, encouraging authors to leave basic title and access dates in a uniform fashion would surely be desire to the blight of raw URL references. And in the cases of most books references the issues you raise aren't a concern. There's no way a template should be imagined to be the cure to all ills but it could capture a most uses and allow "exceptional cases" to either enter 'freeform' references or to "break" the template with exceptional cases, while providing a lead about a uniform referencing style for the 'pedia. --rannṗáirtí anaiṫnid (coṁrá) 11:33, 16 January 2010 (UTC)

(ec) I use citation templates quite frequently, but am against a guideline requiring their use. Too many editors don't like them, and too many editors can properly format a citation without using a template anyway. One of Wikipedia's biggest problems isn't the lack of citation templates; it's the lack of citations at all. Firsfron of Ronchester 04:55, 17 January 2010 (UTC)

Just as a comment to this, it is technically possible to merge all the "cite" family of templates into a single common template, with the existing spinoffs like "cite news" designed either for backwards-compatibility or to simplify the arguments specific for that reference. That actually should probably be done regardless of style format normalization in the first place. --MASEM (t) 14:23, 25 January 2010 (UTC)

Reference structure

One thing I'd like to point out (since we're on the subject) that bugs the crap out of me - can't we do something about the bass-ackwards reference structure on pages? It is such a *pain* to try to edit pages when references are specified inline (in the middle of text). references should be spelled out in the references section, and all that should appear in-text is <ref name="xxx">. I don't know why it got instituted this particular way, but it really makes no sense. --Ludwigs2 11:49, 16 January 2010 (UTC)

Yes, that would be very good. --rannṗáirtí anaiṫnid (coṁrá) 12:02, 16 January 2010 (UTC)

See WP:LDR, which should be exactly what you're looking for. dragons flight added this ability in Cite.php within the last half year or so. --Izno (talk) 19:35, 16 January 2010 (UTC)

Wow! Thanks! That is brilliant! And well done to Dragons flight. That is just what we should be doing. Anyone else, see example in my sandbox. --rannṗáirtí anaiṫnid (coṁrá) 20:47, 16 January 2010 (UTC)

Thanks.  :-) Dragons flight (talk) 10:16, 17 January 2010 (UTC)

That is very nice. I whipped up a quick template - {{Ref_cleanup}} - to put on the talk pages of some of the more godawful messes. Do you think this would work better as an article-space template? and what cleanup categories should this put articles in? --Ludwigs2 21:22, 16 January 2010 (UTC)

I think the desciription at WP:LDR could do with clarifying first. It uses the "group" attribute, which I cannot every see being used in most (any) articles. That should be taken out of the example first so people can more clearly see what it is about. --rannṗáirtí anaiṫnid (coṁrá) 21:49, 16 January 2010 (UTC)

It uses group to keep it from clashing with other examples. There is a better way to fix that, but it requires updating other examples on the page. I will take that up in the next day or so. ---— Gadget850 (Ed) ^talk 05:59, 17 January 2010 (UTC)

You should be aware that there is (or at least seems to be) a significant amount of resistance to the use of WP:LDR, currently. One of the most significant impediments to using list defined references includes the fact that Smackbot (task list · contribs) messes them up, if it does anything else on the page.
— V = I * R (talk to Ohms law) 22:15, 16 January 2010 (UTC)

The first time I looked at LDR it did have that problem. I do however have an article which is using LDRs currently at organic solar cell. --Izno (talk) 22:25, 16 January 2010 (UTC)

I'll go ahead and clarify at WP:LDR already done - does anyone know for sure if that's an ongoing problem with smackbot, or something that's already been addressed? because from the perspective of any non-bot individual LDR is a much, much better system. --Ludwigs2 05:53, 17 January 2010 (UTC)

Last I remember the issue "was being looked into" (although admittedly, now that I think about it, that was probably a couple months back). It's not just smackbot though, as there are a couple of editors who seem to be actively opposed to LDR for some reason or other. I think that you're right, and I love them, but I refuse to put the work into implementing them when there are people who will look for it to revert the changes. It just hasn't been an issue really worth pursuing, for me personally.
— V = I * R (talk to Ohms law) 07:31, 17 January 2010 (UTC)

Above, people have complained that citation templates make pages to load much slower. Please note that the use of WP:LDR also makes pages to load much slower, if you have hundreds of citations. — Miym (talk) 22:17, 17 January 2010 (UTC)

The solution to that is usually to reduce the citations to the high-value ones. Just as overlinking reduces the value of links, overciting reduces the value of citations. — Carl (CBM · talk) 11:16, 19 January 2010 (UTC)

Or (hypothesizing here) the citations could just be loaded separately and only on-demand. 75% of people probably don't look at the References. --Cybercobra (talk) 11:53, 19 January 2010 (UTC)

Or better yet, never use citation templates. Then the problem of templates slowing down the page load wouldn't happen – no templates. • Ling.Nut 11:59, 19 January 2010 (UTC)

Ling.Nut, please note that WP:LDR does not use citation templates, it is a feature of Cite.php. Nevertheless, using LDR makes page rendering much slower if you have very many references. (By the way, people should please keep this in mind when they suggest that "slow" templates could be replaced by a "fast" Cite.php implementation. Having something in Cite.php does not make it automatically fast.) — Miym (talk) 13:12, 19 January 2010 (UTC)

Of course, the fact that references were specified inline was considered a major benefit of cite.php when it was first enabled; for one thing it makes section editing much easier. Until the way references are displayed in edit mode can be changed on a per-editor basis, there's unlikely to be agreement on this issue. Christopher Parham (talk) 15:47, 19 January 2010 (UTC)

Until Wikipedia moves toward including refs at the end of articles, for example, I have made some progress in that respect by creating a user script. It temporarily moves inline refs into a different textbox, then moves them back into the text when saving. Right now, the script has not been extensively tested and only works on Firefox. I plan to improve browser support soon works on the most recent versions of all PC web browsers (IE8 requires a workaround). 23:03, 4 February 2010 (UTC) PleaseStand (talk) 12:31, 4 February 2010 (UTC)

wikEd beta is also supposed to have reference hiding. I wonder if reference hiding similar to these scripts could become a part of the MediaWiki footnote system itself? PleaseStand (talk) 06:01, 6 February 2010 (UTC)

See below: #Software solution to citation mess? PleaseStand (talk) 19:25, 20 February 2010 (UTC)

Improper sourcing

As I see it, en:WP has a sourcing problem that's more serious than anything mentioned so far in this discussion, one that's more serious than a lack of sourcing or sourcing somewhere unreliable. It's "sourcing" -- worse the more deftly it appears -- to material that does not say what the citation implies is said. Effectively, lying about the provenance of assertions. I came across a bunch of such "sourcing" just the other day within the Brazil article; here you see me look for a web page, not find it, find a mirror in web.archive.org, find that this doesn't say what it's claimed to say, and zap it, leaving a comment to prevent anybody's later reversion of my perceived vandalism.

On the rare occasion when I'm in the delightful position of both having an article all to myself and having reserves of time and effort, I do the footnotes in what probably appears a horribly laborious and highly idiosyncratic way, attempting to proof the article to some degree against later incompetence (although hardly vandalism). I summarize what's being sourced within the footnote. So:

. . . they divorced in 1963.<ref>Year of divorce: Joe Bloggs, "[http://tumtitumtitum Gossip]", ''Gossip Rag,'' 23 March 2002. Accessed blah blah.</ref>

So if somebody later adds, not from Bloggs's article but perhaps from hazy memory of a TV program seen a year previously, that this divorce was after they'd had three children, resulting in

. . . they divorced in 1963 after having three children.<ref>Year of divorce: Joe Bloggs, "[http://tumtitumtitum Gossip]", ''Gossip Rag,'' 23 March 2002. Accessed blah blah.</ref>

the reader has a warning sign. Of course, that later editor may, out of either stupidity or an urge to vandalize, instead add the fiction

. . . they divorced in 1963 because he was gay.<ref>Year of divorce and gayness: Joe Bloggs, "[http://tumtitumtitum Gossip]", ''Gossip Rag,'' 23 March 2002. Accessed blah blah.</ref>

[Sorry to be getting rather beany here.] However, the requirement for double falsification slightly increases the risk that he'll catch the attention of other, scrupulous editors.

Now, if people were to turn attention to the falsification (however motivated) of sourcing and methods to reduce the risk thereof, I'd be interested. -- Hoary (talk) 12:22, 16 January 2010 (UTC)

Yes, again! Citation should come with either quotations or (if a quotation is not suitable or possible) summaries so that a) we know what they were for and b) we know that they say what they say they say.

For example, I'm copy editing an article at the moment. There's references in there from years back. It's guess work to know what the references are for or why they were added. Are they relevant any more? Or has the reference "drifted" far away from the sentence it originally cited?

Another example, I've been in content disputes in the past. An editor throws out a references they say supports their claim (nearly always they genuinely believe it does). I know is BS, others know it's BS, but they say they've got a reference. So we've got to press them for a quotation. If we're lucky they'll give it. And we can point out where they're wrong. More than likely they won't, but refuse to remove the claim on account that "it's referenced". So we've got to trawl though libraries. If we're lucky one of us will find the supposed "reference". Most likely they won't have given a page number. So we've to read this thing end-to-end. And then what? Do we come back to the page to say that it doesn't frickin' say that? Only to have them say, yes it does - look harder. --rannṗáirtí anaiṫnid (coṁrá) 12:54, 16 January 2010 (UTC)

I'm against the use of quotations in citations because: 1) quotations don't belong in citations, and 2) even adding a quotation won't prevent vandalism or ineptitude. Firsfron of Ronchester 04:50, 17 January 2010 (UTC)

Won't stop (2), but makes it much easier to recognize. --Cybercobra (talk) 04:54, 17 January 2010 (UTC)

Only to someone checking the reference. Firsfron of Ronchester 04:57, 17 January 2010 (UTC)

It's a step in the right direction. --Cybercobra (talk) 06:36, 17 January 2010 (UTC)

I disagree. Even at FAC, a number of editors have suggested that fact-checking during the process is poor, and that Manual of Style issues are often what holds an article back from FA. If fact-checking at WP's highest level is poor, no amount of quote-adding will keep a citation error-free. Only careful editing will reduce errors of these types; the quotes have nothing to do with careful editing. Firsfron of Ronchester 08:12, 17 January 2010 (UTC)

Of course there is no silver bullet and quotes won't affect editor carefulness, but they make such continuity discrepancies easier to identify. The presence of a quote= parameter in many citation templates indicates opposing views on the matter of quotes as part of citations. --Cybercobra (talk) 08:38, 17 January 2010 (UTC)

I just checked, and the only citation template at WP:CITET which appears to use the quote= parameter is template:Cite video game. Either that page is out of date, or the quote parameter is used infrequently in citation templates. I use RefToolbar, a handy tool for adding inline cites; although it has built-in citation templates for cite web, news, book, and journal, it does not give a quote parameter for any of the four citation templates. As far as making discrepancies easier to identify, the quotes can only be useful if concerned editors are fact-checking: comparing what is cited as a source with what is in the article. I have concerns that this happens extremely infrequently, even at WP's highest levels. Firsfron of Ronchester 09:26, 17 January 2010 (UTC)

The quote parameter is in {{Citation/core}}, and is therefore usable from any template which is based on Cite core (of which, interestingly, {{Cite video game}} is not. I'm somewhat surprised that it's never been TfD'ed, honestly).
— V = I * R (talk to Ohms law) 09:54, 17 January 2010 (UTC)

Regarding {{Cite video game}}, I agree that it's a little strange to have such a template in the first place. Firsfron of Ronchester 15:10, 17 January 2010 (UTC)

I get the feeling some of you haven't tried correcting citations for country articles. You guys should try it sometime. Usually more than 100 ducks sitting in a row. Adding quotes to all of them may end up increasing the size of the article significantly. Bye bye recommended 100kb limit. Lambanog (talk) 17:42, 17 January 2010 (UTC)

Someone just reformatted the references at Yankee Stadium with a bunch of citation templates, without any perceptible difference for good or ill. (The ones whose style I liked seemed to stay that way, but the ones whose style I disliked were still unattractive.) However, it added another megabyte and a half (1,470 kB) to the article's length. See this difference. ¶ And editing others' templated citations can be a nightmare. Either the lines run together, which makes it very hard to see what's happening (especially for a new editor who may know the sources well, but not templates), or each item gets a separate line, which stretches the number of screens for editing somewhere towards the neighbourhood of countable infinity. ¶ [And I just noticed that most of these comments may be slightly off-sub-thread. Feel free to move them.] —— Shakescene (talk) 19:08, 17 January 2010 (UTC)

Wishlist: References are stored in a database separate from the article's text and are edited only through a tabular interface ala the refToolbar. Thus, no templates are ever visible to the editor. --Cybercobra (talk) 23:18, 17 January 2010 (UTC)

Would that we could... I don't really agree with the "not visible to the editor" part, but otherwise I'm right there with ya.
— V = I * R (talk to Ohms law) 07:06, 18 January 2010 (UTC)

{{Cite video game}} is used 99% of the time when sources are required for the plot/characters of a video game and not provided through other secondary sources. It probably should be moved to use the core stuff, but it is an oft-used template. --MASEM (t) 14:26, 25 January 2010 (UTC)

I adopted a similar approach to Hoary for similar reasons in Artificial intelligence. ---- CharlesGillingham (talk) 18:37, 19 January 2010 (UTC)

Stay on track

It's inevitable that a conversation about citation style will wander into other issues, but let's try to stay focused. I don't think we can make progress on other issues until we have a standard citation style. For the time being, let's avoid discussions about whether citation templates are good or bad, the problem of unsourced content, misuse of sources, etc. The question being raised here is, Should Wikipedia have a standard citation style or not? I believe it should, mostly because I think having a citation style will reduce the difficulty and the effort of making progress against some of the other issues related to citing sources. Don't worry about whether the citation style will be perfect—it won't—consider whether or not it will help solve the issues I've raised above. — John Cardinal (talk) 14:31, 16 January 2010 (UTC)

I'm glad this is being looked at, but the problem is much bigger than whether we should have a standard citation style-- the other problem is the proliferation of citing info across too many guideline pages, that are inconsistent, contradictory, and poorly written, with willy-nilly changes that aren't supported by consenus and aren't reflected in other guideline pages addressing the same issues. This is a big problem that needs to be solved. See, for example, this discussion on my talk page. Has the above discussion been posted to all of the various guideline pages where this mess occurs? SandyGeorgia (Talk) 14:54, 16 January 2010 (UTC)

Agree entirely. Some editors want to mention everything they can think of in a citation—there's a discussion going on at WP:CITE over a sentence saying you needn't add, when citing a book, whether you bought it, borrowed it, or read it on Google Books. I've seen citations where not only the newspaper is named, but the publisher of the newspaper is added too, and on and on. What needs to be communicated somehow is that all we need is enough information to help readers find the source material, and if they can't find it using the title, The Washington Post, adding the name of the newspaper's publisher isn't going to help them. But it's the nature of Wikipedia. If people can think of an additional thing to write, they want to write it, preferably inside a series of giant templates (sorry, John) that make editing the article impossible. SlimVirgin ^{TALK contribs} 15:08, 16 January 2010 (UTC)

Sandy, I think the lack of a standard citation style contributes to some of the problems you listed and diverts focus from some of the others. I have not posted notices about this discussion anywhere else, but I planned to post a notice to the WP:CITE talk page. I have to run out now, but if someone else wants to do it, go right ahead.

SlimVirgin, I think the citation style should specify the data elements that should be included for various source types and what data elements should not be included. — John Cardinal (talk) 15:32, 16 January 2010 (UTC)

A Wikipedia citation style would be the first step to resolving some of the issues mentioned - but only if the choice of that citation style is driven by the issues we want to resolve, which necessitates identifying what those issues are. The question of a citation style is more than simply asking whether we want to put the year in brackets or initialise the authors first name. The purpose of a citation style is to make it easy (even possible) for others to verify claims or assumptions that are made in a text.

On Wikipedia we have a unique set of circumstances some of which were mentioned above: citations "drifting", editors providing lazy citations, citations being mistaken (or downright false). APA (for example) can define their citation style in the way they have because of their unique circumstances. If the citation is "lazy", they reject the paper. (We can't.) The citation never "drifts" because the authors of the paper are the same. (Ours do.) Mistaken or falsified references can be captured in peer review. (Ours aren't.)

When we decide on a citation style we need to think of our unique circumstances. Otherwise it's just an exercise in putting such-and-such in italics and initialising something else. That's no good. It does no suffice the purpose of citations. We have to bear in mind that people are lazy, references "drift" across many edits, and editors often need greater evidence for statements made in that just the promise that the citation say what someone says it says. Our citation style should address those and other issues.

For that reason, I wouldn't narrow down discussion to simply deciding on a citation style, we need to discuss what we need in citations and how we need to cite works (based on the practicalities of our work here). That inevitable requires use to air current problems with citing sources. --rannṗáirtí anaiṫnid (coṁrá) 21:14, 16 January 2010 (UTC)

I just wanted to echo User:John Cardinal here. I don't see him saying that these corollary issues are unimportant, and I certainly don't think that their unimportant myself, but it would be nice to have a conversation about the actual issue at hand, here. I'd encourage, and even welcome, separate conversations about problems with improper or poor sourcing, but I think that their out of place here.
— V = I * R (talk to Ohms law) 22:21, 16 January 2010 (UTC)

In general I would oppose any effort at creating a uniform style. The negotiations toward such a style would be a massive boondoggle I suspect will produce nothing of value (why would we come to consensus on a "best style" when the rest of the scholarly world over decades has not done so?). Beyond this, being generally permissive of all standard formats helps Wikipedia attract new editors and encourages them to add as many citations as possible. The main goal of our citation guidelines is to encourage additional citations and I believe that getting picky about formats would inhibit this. I don't want people saying to new users, "Great work adding these citations! But you've used APA style footnotes. Why don't you take some time away from adding sources to learn Wikipedia's Authorized Style?" Christopher Parham (talk) 16:14, 19 January 2010 (UTC)

Why do people continue to assert that the proposal attempts to choose the "best style"? That assertion misses the point completely. I don't think there is any such thing as a "best style", and I don't think the other proponents of a standard citation style are suggesting there is. The main thrust is that allowing multiple citation styles has made a mess of the instructions and guidelines, leads to wishy-washy (or non-existent) standards for what information should be included in a citation, leads to edit conflicts, leads to a lack of consistency in and across articles, etc. And for what? So editors can have their own way regarding the choice of a comma versus a period? — John Cardinal (talk) 16:38, 19 January 2010 (UTC)

Because for your process to work we need to have a consensus of editors agree on a specific style, and maintain such agreement indefinately so that the rules aren't just changed back to permit flexibiity. Editors aren't going to concede valuable features of existing styles simply to achieve consistency (and indeed it doesn't seem you're asking them to do so, if you're portraying it as a choice of a comma versus a period) so the style chosen is going to have to incorporate every valuable feature of every existing style. That's what I mean by "best": widely agreed to be at least equal to every alternative in every situation. I think people assume you are trying to choose the best style because the alternative is that you are enthroning a worse style in place of the best, which seems very unappealing. Christopher Parham (talk) 18:09, 19 January 2010 (UTC)

Exactly - choosing a single citation style that was not agreed upon by consensus as being the most advantageous would be silly to say the least. But I also agree with John Cardinal that there is no single "best citation style" - which only means that the entire issue here is silly as it boils down to choosing an arbitrary exclusive standard simply for cosmetical reasons.·Maunus·ƛ· 18:21, 19 January 2010 (UTC)

Presenting the most important elements (for general information, to indicate the general nature of one's sources, to allow further study, and to allow independent verification) is vital. How to do so will vary with the context, as will the importance of each particular element. The illustrator's unimportant in many sources, but vital in a subject like comic books. Page numbers are essential for some purposes, and irrelevant in others. Surname-first (indeed author first) is terrific for academic or scientific books, but often leads to confusion with news articles, literature, music, or multiple authors. And so forth. But we should still insist on author, work, page number (or URL), some sort of date, and some publishing indicator where not obvious (city, publisher, web site, whatever). —— Shakescene (talk) 18:24, 19 January 2010 (UTC)

Yes. I think we already do this to some degree (Wikipedia:CITE#Citation_styles) but this should be the focus of the guidelines and to the extent we make them stricter, the "rules" should be about what to include rather than how to format things. In my view, WP:CITE has always been somewhere between the MOS and an editing guideline, and I think it does better the more of an editing guideline it becomes. Completeness and clarity of citations is the most important thing. Christopher Parham (talk) 19:45, 19 January 2010 (UTC)

Shakescene, OK, illustrator is important when citing a comic book. Where is the WP guideline that makes it clear that the illustrator must be included when citing a comic book? I don't see it anywhere. For people who use templates, {{cite comic}} doesn't have a parameter with that name, and none of the parameters that might apply (artist, cartoonist) are required. How does having a standard make this situation worse?

Who is actually confused by surname first vs. surname last, and are you sure that readers who review citations with it first never see articles with citations where surname is last?

Your examples seem like a good argument for choosing a standard, not against. — John Cardinal (talk) 21:12, 19 January 2010 (UTC)

Well said. You guys are confused... you're saying "don't do this", but every argument you make bolsters the position that this is exactly what we need to do. Christopher is also exactly correct that we need to be flexible and collectively develop something that everyone will at least acquiesce to using. If that's not the definition of consensus, I don't know what is.
— V = I * R (talk to Ohms law) 21:25, 19 January 2010 (UTC)

Comment from Ling.Nut

One style to rule them all? What a terrible, terrible idea. It has never been implemented for the reason that the idea simply sucks, and sucks horribly. To repeat the reason: different fields have different styles. editors who work in those fields should be encouraged to follow the style they are familiar with, instead of imposing a new learning curve upon them. Editors who do content review should take it upon themselves to deal with the issue of multiple styles. They are almost always folks who have already decided to make a commitment of volunteer time to Wikipedia. Editors who create content should not be burdened or concerned with this task, in order to prevent the case in which folks decide that becoming a contributor is not worth the trouble of learning a new citation style. Moreover, folks who work in a certain field are probably more comfortable reading a certain style. If there's a poll later PLEASE PLEASE PLEASE ping me so I can Very Strong Oppose.... If there's a much weaker version of this impulse, say forex, a requirement that the reference style must adhere to to that of any well-known journal in the larger or narrower field (e.g., social sciences; linguistics), then I would consider that worth discussing. [ I know folks are thinking "how would we implement this? But it in the normal flow of Wikipedia, this issue wouldn't even come up unless an article goes to FAC.] I too am peeved when folks make up their own creative styles. But... "one style to rule them all" should never be imposed in this manner upon new editors (or even established editors, for that matter).... • Ling.Nut 05:25, 17 January 2010 (UTC)

I don't get this attitude. Why should "editors who work in those fields should be encouraged to follow the style they are familiar with" at the expense of the rest of us? I mean, we don't insist, or even suggest, that the American Psychological Association or the Modern Language Associaten, for example, should adjust their style guidelines to ours or to each others (although, I wouldn't be surprised if there are individuals out there who do or would, given an opportunity...). It seems to me that most of the formatting style related issues actually arise out of the chaos caused by intentionally having no guideline. No one is right, no one is wrong, and so what happens is everyone does their own thing, causing some to be pissed at others, some to not attempt to add references at all, and some to add really poorly and problematic styled references simply out of ignorance or through being overly creative. Not that a style will be a panacea, but it'd certainly help more then just ignoring the issue.
— V = I * R (talk to Ohms law) 05:37, 17 January 2010 (UTC)

I am not as passionate about this as Ling.Nut, but I do agree that it's a bad idea to impose one style. The most important reason for this is that styles that are good for one sort of subject matter may not be as good for another. In particular, for technical material, it is often convenient to be able to use footnotes as notes, rather than as references, which means Harvard style or something like it. --Trovatore (talk) 05:39, 17 January 2010 (UTC)

So, we make that a component of our own style guide. There are many, myself included, who do that already, and the way to accomplish it is extremely easy. For the note style references you simply add a "group=<name>" parameter to the ref tag, and in the references section you add a {{|Reflist|group=<name>}} somewhere.
— V = I * R (talk to Ohms law) 05:43, 17 January 2010 (UTC)

Why is allowing different types of citation formatting equated with an "expense [to] the rest of us"? What is the "expense"? Having articles with slightly different citation styles? Heaven forbid that 3 million articles written by millions of people will have different citation styles. Who are "the rest of us"? Those who want a unified citation style? If there really were that many editors who saw the urgent need for a unified citation style, it probably would have been adapted years ago. Please don't impose a citation style on all editors; the guidelines and policies are overly restrictive as it is. This is not the time to impose further rules on editors who are already volunteering their time and efforts in good faith. Firsfron of Ronchester 07:50, 17 January 2010 (UTC)

Why does everything need to be so black and white? I'm not suggesting that we should impose anything on anyone, and I don't see John Cardinal making such a suggestion either. The costs associated with not making any choice in this matter are also addressed by John, myself, and at least one other user, above. Briefly though, we're creating conflict and confusion by refusing to make any firm recommendations/requirements in this area. The fact that we seem unwilling or unable to make a style recommendation causes conflict, reduces editor efficiency by forcing repeated discussions to occur about the same or similar subject in an ad hoc manner, and most seriously probably even prevents some of the more casual editors from adding references due to uncertainty and confusion.
— V = I * R (talk to Ohms law) 10:16, 17 January 2010 (UTC)

Well Ohms, I've agreed with you a dozen times on various occasions, but I don't agree now. The problem with your statement "I'm not suggesting that we should impose anything on anyone, and I don't see John Cardinal making such a suggestion either," is that it ignores the fact that someone else in this very discussion has: "One way in which all of the points raised above could be achieved any by others is by preferring (making mandatory?) the user of {{cite}}, {{cite web}}, {{cite journal}}, etc."

Another obvious problem is that if your preferred citation style becomes a guideline, it will be imposed at FAC (which requires that the article conforms to all style guidelines). Thus, if an editor wishes to submit his or her article to FAC, s/he will be forced to follow your proposed guideline, no matter what is stated now, when it's proposed.

Finally, the best way to stem arguing over citation styles is just not to argue over citation styles. Arguing that having different citation styles causes arguing is a cyclical paradox. :) Firsfron of Ronchester 15:37, 17 January 2010 (UTC)

Question marks and the parenthesis provide some important context... but anyway, this viewpoint is easily compared to "sticking your head in the sand" in my own view. I'm really convinced that we're collectively causing more angst and conflict then we're resolving by continuing along this tack, and I don't see much which would convince me to change that view.
— V = I * R (talk to Ohms law) 17:08, 17 January 2010 (UTC)

Quite. I agree with Ohms law's. While it is possible to make this "problem" go away by simply ignoring it, that is the definition of sticking our heads in the sand.

I was the editor who made the preference/mandatory remark around templates. There are some good points raised by other that one size does not fit all, but nine times out of ten does it really matter so much what style is used:

• Harvard: Murphy, J., 2009. Book of Wisdon. London: Publishing House Press
• APA: Murphy, J. (2009). Book of Widson. London: Publishing House Press
• Chicago: Murphy, John. (2009). Book of Widson. London: Publishing House Press
• MLA: Murphy, John. Book of Widson. London: Publishing House Press, 1993
• AMA: Murphy J. Book of Widson. London: Publishing House Press; 1993

The point is that some style should be used (where editors often don't use any) and it is preferable that a citation style should be consistent within an article and across a publication (within reason, we should not be citation Nazis). One of my reason for preferring (sorry, Ohms law!) templates is because templates can make it easier for editors who are not used to citing sources to be able to do so - and to do so in a consistent manner.

(There were some good points raised above concerning the slowness of transcluding the citation templates and around problems for copy-editing text where they appear in line. I've made a suggestion to extend Cite.php that would address those concerns.) --rannṗáirtí anaiṫnid (coṁrá) 17:30, 17 January 2010 (UTC)

Ling.Nut, I am not swayed by the argument that Wikipedia should not adopt a citation style because different fields have different citation styles. I don't think that citation styles are 100% consistent within a field—styles tend to vary by publication—but even if true form some fields, authors have to learn a lot that is specific to WP, such as wikicode, and learning a new citation style is a small barrier. If we adopt a standard style and an editor enters a citation incorrectly, it can be corrected. — John Cardinal (talk) 18:16, 17 January 2010 (UTC)

Agreed with Ling.Nut and with the same passion. We have a set of citation styles which are acceptable, and no editor should be forced to use MLA styling for a film article, nor should any editor have to use something like the hideousness that is Harvard style. Trying to arbitrarily decide that one style is somehow better than all the rest will not make the problem go away, it will just drive away dozens, if not hundreds, of the editors who actually do cite content. There are citation templates for consistentcy, and the only thing that needs to be consistent is the consistency in an article or a set of related articles (i.e. a television series, its episode list, and character list should all use the same style). Nothing else is needed. -- AnmaFinotera (talk · contribs) 18:52, 17 January 2010 (UTC)

Isn't "We have a set of citation styles which are acceptable" what we're discussing here, and what you're seemingly opposing? My impression is that creation of a citation style was the primary subject for all of this.
— V = I * R (talk to Ohms law) 20:08, 17 January 2010 (UTC)

No, this discuss is saying lets have one. I noted there are a set - as in many which are acceptable. I.E. the status quo is fine. -- AnmaFinotera (talk · contribs) 20:13, 17 January 2010 (UTC)

The thing is, the statement that "there are a set" is not at all clear to me, as demonstrated by some of the commentary above. Even Ling Nut's reply, which you said that you're supporting, seems to eschew most if not all of the current suggestions which you seem to be referring to. That's why I'm slightly confused about your statement.
— V = I * R (talk to Ohms law) 21:00, 17 January 2010 (UTC)

Technical discussion regarding Harvard Refs

Hmm? How does that get you Harvard style? --Trovatore (talk) 05:46, 17 January 2010 (UTC)

It doesn't get you Harvard style exactly, but it provides something that is much more useful/usable to Wikipedia readers, which is ultimately exactly the point of addressing this issue. What you do is that you create a section in the article's References section (preferably with a list heading (';') rather then a section heading ( '===' or more)) for the book, and then immediately after that you'd add a {{|Reflist|group=<Authors last name>}}. Then, in the article content, you'd add <ref group=<authors name>>Page X</ref> where appropriate. Doing that adds a list of references specifically for that book, to the list under the header describing that book. That way, when people click on the reference ("what's this?"), they are immediately provided with context which get's them to the reference itself ("oh, I see, this is from such-and-such book"). It's easy and clear to everyone, most especially readers, which should be why we add references to articles after all.
— V = I * R (talk to Ohms law) 07:18, 17 January 2010 (UTC)

Hm, that goes a long way to addressing what I normally despise about Harvard style in WP, which is essentially that one has to grep through the References section for the person's last name (and sometimes also year) in order to find the rest of the citation info. Although I dislike the extra complication of group=, this advantage makes it an acceptable trade-off. --Cybercobra (talk) 07:40, 17 January 2010 (UTC)

That's what {{Harvard citation}} and related templates are for — they give you links to the exact work in the references, without any need to impose footnote-style citation on all articles. --Trovatore (talk) 09:40, 17 January 2010 (UTC)

I've seen those Harvard templates, and I recommend their use as well. I recommend their use within a ref parameter using a specific group= parameter along with it however, in order to address the reader usability issues outlined above. Using the template alongside of the grouping creates nice, neat, consistent book references.
— V = I * R (talk to Ohms law) 10:02, 17 January 2010 (UTC)

Um, so what's the effect of that? The most important advantage of Harvard is that notes are clearly distinct from references; when you see a little superscript blue number, you know that's a note, and that it is not a citation. Does your recommendation preserve this advantage? --Trovatore (talk) 20:22, 17 January 2010 (UTC)

well, it depends on what your exact intent is with a reference/cite/note. If you're talking about using parenthetical referencing, The Harvard style templates are just the ticket, although their easy to do without them as well. Providing a reader with an easy and unambiguous means to get to the full references (usually with a bulleted {{cite book}} and links to/from the individual citations) then that also serves the reader in locating the source itself at the same time.

I find the term "note", along with most other terminology when it comes to these discussions (which is part and parcel to this discussion, incidentally) to be confusingly ambiguous. To me, a "note" is a bit of informational prose, usually used in order to provide additional corollary details. I get the sense that isn't what you're definition of the word is here however, so some clarification would probably help.
— V = I * R (talk to Ohms law) 20:57, 17 January 2010 (UTC)

That's exactly what I mean by note. When you're using Harvard style, then the little superscript numerals direct you to parenthetical remarks (I hope it's clear that parenthetical does not mean "in round brackets"!), prose that expands on, or provides extra precision to, the text you're reading, but which would distract from the flow of thought if rendered inline. When you see a superscript numeral, you know to click on it if you're looking for full details, whereas if you had thought it was just a citation you might have skipped it. On the other hand, if citations are what you're looking for, you can skip the superscript numerals. --Trovatore (talk) 21:03, 17 January 2010 (UTC)

←OK, understood now. Explanatory notes like this are an excellent reason to use the group parameter as well, and for the reasons which you've outlined above. If you use <ref group=name>This is an explanatory note about this content</ref> then that shows up on the page in a different reference list from all other groups, and it's distinctive on the page because it shows up as ^{[note #]} where the inline citation is displayed. incidentally, you can use anything you like in the group= parameter, and that's what shows in the inline cite/link, which allows you to tailor them to be both distinct and as unassuming as possible.
— V = I * R (talk to Ohms law) 21:34, 17 January 2010 (UTC)

This style is interesting, but still not clear to me. Do you have an article that uses it? ---- CharlesGillingham (talk) 18:48, 19 January 2010 (UTC)

If you are asking about explanatory notes using |group=, then see Arthur Rudolph. ---— Gadget850 (Ed) ^talk 00:09, 20 January 2010 (UTC)

No, I'm asking about the original post by Ohm's law, where he recommends using <ref group=<authors name>>Page X</ref> in the article body. I thought I was familiar with every citation style in use, and I haven't seen this. (Unless it's something ordinary and I'm reading him wrong; I'm coming in late here and I'm not quiet sure.) ---- CharlesGillingham (talk) 01:27, 20 January 2010 (UTC)

I'll try to find an extant example of it in a "live" page. The last time I pointed out an example though, ...certain users *ahem* went and changed the formatting style after it was offered as an example, so I'm a tad reluctant to provide a new target.
— V = I * R (talk to Ohms law) 01:49, 20 January 2010 (UTC)

Here, rather then trying to dig up an older one, I just changed Enriquillo-Plantain Garden fault zone as an example. Someone will probably change it back or to some other style, but the above is a link to a specific diff so it shouldn't matter if that happens.
— V = I * R (talk to Ohms law) 09:58, 20 January 2010 (UTC)

Getting back to {{harvnb}} and family, is there anything that can be done to address its unutterable fragility? To function correctly it requires that each usage lists the various authors |lastn=, |year= and |pp=, all spelled and ordered exactly consistently with the {{citation}}. Further the {{citation}} must eschew the use of |coauthors= and wikilinked |authorn= or the harvnb's CITEREFLast1Last2Last3YYYY won't be linked to a matching anchor. Worst of all, there will be no indication of the error until someone like Svick deliberately tests each linkage. Indeed, the {{citation}} can be completely missing without creating an error message. That has got to be fixed. LeadSongDog come howl 02:57, 10 February 2010 (UTC)

I don't think this problem can be really solved, because we can't force editors to write only correct {{harv}} citations (actually, edit filters could maybe be able to do this, but I think that's a really bad idea). But we could display error message for incorrect uses using JavaScript (templates alone can't do it) if there was consensus to do so. Svick (talk) 20:05, 10 February 2010 (UTC)

A JavaScript-based solution would be fine, I think, as the vast majority of editors run with JavaScript enabled. If I understand you correctly, the solution would check that all internal wikilinks are to labels that are actually defined, and that there are no duplicate labels. Is that easy to do? If so, I'd say go for it. Eubulides (talk) 20:17, 10 February 2010 (UTC)

That certainly sounds like a great improvement over the present circumstance. I presume that it would not be much use in "Show preview" while editing sections, but at least there would be something visible when reading the full article. Another, possibily complementary, approach might be to have a script generate inline {{fix}}-based tags to flag the problem cases found on Svick's list. LeadSongDog come howl 14:34, 11 February 2010 (UTC)

I have created a prototype of the script to show this error messages: User:Svick/HarvErrors.js. To test it, follow the instructions in the beginning of the code. This script could be later possibly enabled for everyone (if there is consensus to do so).

I have problem with placing of the error message: I think the only reasonable place is after the whole reference generated by one of the templates, but currently there is no easy way to find out, where that is. The script works fine if the {{harv*}} reference is inside <ref>, but the error message will be displaced if used directly in the text. The easiest way to fix this is to change the templates, so that the whole citation is enclosed in an <span> (this shouldn't have any side effects). Any comments, are welcome, especially what the error message should be and whether it should contain link to an explanation page. Svick (talk) 13:05, 16 February 2010 (UTC)

Also, the script currently detects only missing targets, it doesn't check for duplicate ids. Svick (talk) 13:07, 16 February 2010 (UTC)

Somehow in the above discussion I neglected to link to Template_talk:Harvard_citation#Fragile code mishandles spaces in parameters. The problem discussed there still has not been fixed despite indications to the contrary there. Subtle variations in the white spaces between the |lastn= and |year= parameter values passed to {{harv}} and those passed to {{citation}} will still cause a mismatch between the generated anchor and the internal link intended to go to that anchor. Look for underscores in the URLs and CITEREF anchors as the telltale of the problem. LeadSongDog come howl! 04:23, 25 August 2010 (UTC)

WikiProjects

This might be something usefully left to WikiProjects. In particular areas a certain citation style may be preferable, and a clear recommendation in favour of one may be reachable. Across Wikipedia? I don't see it happening. In addition, I must say that I hate citation templates with a fiery passion; and most people would consider me a geek, so I can only imagine what the average editor thinks of them. Given the ability of people (with or without AWB) and bots to impose consistency where this is agreed, I see so very little advantage in the additional complexity and wikitext, even if it is hidden from the body text (using Wikipedia:Footnotes#List-defined_references). If one day the software can hide that complexity appropriately, using templates as encoding not seen by the user unless they really want to, that would be different. Rd232 ^talk 19:15, 17 January 2010 (UTC)

Further technical refinements might be appreciated, but considering the resistance to even LDR's I'm not optimistic. I think that this discussion is a worthwhile pursuit regardless, since simply ignoring the issue obviously isn't helping anything at all. Leveraging WikiProjects is certainly an idea worth discussing, and I certainly wouldn't oppose specifically allowing interested WikiProjects to establish subject matter standards where there is wider consensus to allow them to do so. My only concern is that similar utilization of WikiProjects is... let's just say, less then satisfactory. It creates uneven coverage, at the very least.
— V = I * R (talk to Ohms law) 21:06, 17 January 2010 (UTC)

Natural citation style

It just struck me overnight, that were Wikipedia to adopt a uniform, easily-remembered, citation style, it should be what I consider natural or conversational style, what came naturally to most of us as we were reading books for fun (space, history, adventure, science, nature, fashion, sports, food, entertainment, music, cars, collecting things) before high school, grammar school, lycée or gymnasium taught us all these conflicting formal rules about footnotes and citations. That would be something like Parkinson's Law, by C. Northcote Parkinson, published in London in 1958 by John Murray, page 65 or ... published by John Murray in London in 1958 ... or (more radically) page 58 of Parkinson's Law by C. Northcote Parkinson (as published by John Murray of London in 1958). In fact this is the sort of style used for slipping an unfootnoted in-line supporting reference into a newspaper or magazine for general (non-academic) circulation. One usually thinks of the book or article first, then the author, then the details of publication (sometimes date, sometimes place, sometimes publisher).

Now I'm not advocating this as a citation style, as I just don't believe in uniform citation styles on Wikipedia. And while it's natural (just as Wikipedia uses Logical quotation that copies the text before one without adding or subtracting punctuation), and every formal style is to some degree a distortion of what you'd say in normal conversation, there are sound practical reasons why these formal, somewhat-artificial and non-intuitive conventions arose.

One is that in a given field, titles are bound to be very similar if not identical (e.g. Charles Dickens, Genetics, The War of 1812), and sometimes rather long. But, on the other hand, a single author (especially in an article about the author or her achievements) may dominate the reference list, leading to a long string of Galbraiths, Parkinsons, Schlesingers, Freuds, Churchills or Russells. And in an article covering a region's history or nature, the author's name may be unimportant or forgotten: what you want to see is a series of easily-grasped and -compared lines beginning (say) respectively, Natural History of Wiltshire, Natural History of Gloucestershire, The Botany of Devon, Trees and Flowers of Cornwall, etc.

Another reason (which I'd forgotten when first posting this) is that headlines often change between newspaper editions (and even more between print and web editions) and sometimes between books published in different places, by different publishers or at different times. Authors' names change far less frequently (although sometimes by marriage, religion, or the assumption of a noble title).

If you're going to use a formal non-intuitive convention, then no one of them has any claims to prima facie superiority over any other, except for the specific environment in which it arose. What's good for chemistry may be awful for comic books.

¶ [Now moot, since this discussion has moved here:] However, Wikipedia talk:Citing sources has over two dozen archived pages (which Mr Cardinal is forthright enough to admit he hasn't read), and this issue is certain to be a hardy perennial. If we're going to continue this discussion, we should probably unclutter WP:Village pump (proposals) by moving it to a new sub-page (of this or of WP:Village pump (policy)) or to an existing citation discussion.

—— Shakescene (talk) 20:14, 17 January 2010 (UTC)

I think a "natural citation style" would be an excellent idea myself, but it's getting way ahead of the game here. There needs to be agreement that it's even worth discussing in the first place, which it seems is what we're discussing right now.

Regardless, I wouldn't object to moving this to it's own document, and even listing it on {{cent}}. I think we need more participation in this discussion, so anything that would help that to occur would be great.
— V = I * R (talk to Ohms law) 21:26, 17 January 2010 (UTC)

Cite modifications

A lot of the reference work has to do with templates. Occasionally though people have proposals for modifying Cite to improve the referencing backend. For example, wouldn't it be great if one could pass page numbers, e.g. <ref name="bob" pages="46-52" />, or various other technical proposals. Some are plausible, and some not. However, if people have specific proposals that a lot of people can get behind, then I'll be willing to take a look at implementing them. Dragons flight (talk) 21:11, 17 January 2010 (UTC)

page number handling like that would be extremely excellent; but doesn't seem overly relevant to the "citation style" issue. Rd232 ^talk 21:35, 17 January 2010 (UTC)

You're right, it's not terribly relevant, but it is an easy example to give. I don't know if there will be any general suggestions to come out of the discussions above, but I just wanted to say that I would be happy to hear them. Dragons flight (talk) 22:06, 17 January 2010 (UTC)

Will I float my one a little clearer here rather than continuing a discussion in Bugzilla that really should be shifted over here? ;-) --rannṗáirtí anaiṫnid (coṁrá) 22:28, 17 January 2010 (UTC)

One reason that I've personally been somewhat hesitant to jump into Cite.php dev issues is that I'm not really aware of who's already working on it. I know that User:Dragons flight is (the primary?) developer, but arn't the Usability folks working on something with/alongside of Cite.php?
— V = I * R (talk to Ohms law) 21:30, 19 January 2010 (UTC)

Specific proposal

OK ... I made a proposal to extend Cite.php on Bugzilla, which Dragons flight's appears to be hinting should be discussed here. So deep breath ... here goes ...

The original question of this thread was over the "problem" of mixed-style formatting within articles and across the 'pedia. The guts of the idea I'm floating now goes back up to the "template debates" above. Some of the issues with templates raised above were that they were slow to load, cluttered up the page and forced half-baked styles. Nobody liked the idea of being forced to use a template for numerous reasons (most of which I agree with).

Before I say any more, the proposal is not force use of templates of any kind. What I am proposing here would not prohibit referencing practice as we have it now.

The proposal is that we would extend Cite.php to allow referencing like this:

Lorem ipsum dolor sit amet, consectetur adipiscing elit.<ref name="murphy" /> 
Mauris nulla erat, tincidunt et elementum quis, porta vel risus.<ref name="smith" />

=== References ===

<references style="apa">
  <ref name="murphy" first="John" last="Murphy" 
      title="Quisque Quis Orci Magna" publisher="Printing House Press" location="London" year="2009" />
  <ref name="smith">The system would not prohibit users from entering free-form text for references as normal.</ref>
</references>

(Note the example uses list-defined references but it could just as well use in-line references. Note too that this would be in-built to Cite.php so it would by-pass the slowness of transcluding the current templates, which could of course still be used.)

The "smith" reference in the above example would display as we it would now (i.e. what you write between the <ref> tags appears in the reference). So use of this system would not be prohibit manually typed references like we have now. The "murphy" reference would display as follows by default (i.e. using "harvard" style since that is what is defined as the default for the this reference section):

• Murphy, J. (2009). Book of Widson. London: Publishing House Press

The "style" in the <references> tag would be optional and could potentially be over-ridden by user preferences. So, for example, an individual user could state in their preferences that they want references displayed in MLA style instead. If so, they would see:

• Murphy, John. Quisque Quis Orci Magna. London: Publishing House Press, 2009

Another user could specify that they want references shown in the "natural" style described by Shakescene above. If so, they would see:

• Quisque Quis Orci Magna by John Murphy, published by in 2009 by Printing House Press, London.

It would also be possible to state the format to be used by specific reference, if that was thought useful also.

Furthermore, I'm only making the proposal here for clear-cut examples of books, journals and newspapers, where explicit guidance exists over how these should be referenced according to the common style guides. Again, if you didn't want your reference to be handled in the way being proposed here you could simply reference it as normal, like you do now.

Finally, this isn't a pie-in-the-sky proposal. A very basic implementation is working. And very finally, the proposal is not to encourage or prefer this kind of styling but to simply enable it as an option as see how it goes. --rannṗáirtí anaiṫnid (coṁrá) 23:36, 17 January 2010 (UTC)

(Because I'm detail-oriented) Shouldn't there be a parameter for what type of work it is? (i.e. journal article, newspaper article, book) --Cybercobra (talk) 02:35, 18 January 2010 (UTC)

It should. Explicitness is nice. It could also be implicit (i.e. depending on the attributes passed, it determines you must mean a either a book, journal, or newspaper). Support for both methods would be best IMHO. --rannṗáirtí anaiṫnid (coṁrá) 08:38, 18 January 2010 (UTC)

This is a fairly good idea, in general, I think. Now that we're on our own page here you may want to break it out into a second level heading/section. The only thing that I don't particularly like about it is "semantic", in that it utilized the otherwise somewhat inconsistent XML bracket style of markup. There's no technical reason to change that, but from a usability perspective we should prefer to avoid requiring the use of XML style markup in wikitext when possible. If the ref tag system were converted to use the {{parserfunction}} markup along side of, or instead of, then that would be good (although... the use if <ref>...</ref> may be too ingrained to change, now.
— V = I * R (talk to Ohms law) 23:44, 18 January 2010 (UTC)

For those unfamiliar with parserfunctions (e.g. me), how would that look exactly? --Cybercobra (talk) 00:05, 19 January 2010 (UTC)

Something sort of like: {{Ref: name="murphy" first="John" last="Murphy" title="Quisque Quis Orci Magna" publisher="Printing House Press" location="London" year="2009"}} See also: mw:Help:Magic words#Parser functions
— V = I * R (talk to Ohms law) 00:11, 19 January 2010 (UTC)

Er, aside from the different type of bracketing and the added colon, both of which seem rather trivial, that doesn't seem to buy you anything over the apparently already somewhat implemented pseudo-XML syntax. --Cybercobra (talk) 00:24, 19 January 2010 (UTC)

Well, as I said above, it's purely a semantic difference. There's absolutely no technical reason to prefer the use of parser function syntax, so this isn't addressed at actually "buying anything"... but then, the MediaWiki choice to use "{{" and "}}" in place of "<" and ">" or "/>" is also as arbitrary. Personally, if I had my druthers we'd use angle bracket XML like syntax instead of the more usual curly bracket syntax at all, but since the software design itself tends to use the curly bracket syntax then we should attempt to remain consistent where it is possible to do so.
— V = I * R (talk to Ohms law) 00:56, 19 January 2010 (UTC)

Two things I've got:

Re 'semantic' differences: I think the biggest reason we wouldn't want to use {{#ref:}} is because we already have the first format in wide use. Asking the developers to code for the new format might make them eat people.

One thing I'd like to see would be for us to edit the APA, MLA, so-on-and-so-forth styles. For example, instead of forcing the developers to, say, to hardcode each type of citation into the extension, for the extension to find a page (i.e. Mediawiki:Cite.php-book-MLA-(oneauthor?)). On that page will be something like:

author=last_name, first_name.
title=<u>title</u>.
location=location:
publisher=publisher,
date=date.

From this point, the extension simply takes each line and inserts the data in the reference. This allows us to control the style (both the order of the data and the various styling of the data), in the event that MLA is updated (which, you know it will be). The fault in this is that it generates overhead on our part, but at the developer's benefit (and I daresay their time is more valuable than ours). I don't know, however, if this would create more overhead as the extension is pulling from a MediaWiki pages instead of from directly inside the extension. The other issue is that something (I don't know what) needs to happen so that the extension knows how many authors a text has...

Just pillaging my mind while considering this issue. --Izno (talk) 05:47, 19 January 2010 (UTC)

Ohms law, I'm with you on the issue of semantics. I remember the first time I saw the <ref> tag I though it looked out of place but for good or for bad it's the system in play. Presumably plugging into would be relatively trivial {{Ref: ... }} (although I don't know because I have never used that syntax and never looked at what goes on behind the scenes).

Izno, one thing that Dragons flight said from the offset was that he would be against anything being hardcoded. I personally don't see the issue (the standard citation styles are well defined) but it does look like something like you say would be necessary (and I guess beneficial)

Off the top of my head another method would be to have a single page containing all of the definitions in one:

 apa:book:{$last}, |{$firstInitial}. |({$year}) |''{$title}''. |{{$location}}: |{{$publisher}}
 apa:journal:{$last}, |{$firstInitial}. |({$year}) |"{$title}". |{{$journal}}, |{{$volume}}, |{{$pages}}
 mla:book:{$last}, |{$first}. |''{$title}''. |{{$location}}: |{{$publisher}}, |({$year}) 
 mla:journal:{$last}, |{$first}. |"{$title}". |{{$journal}} |{{$volume}} |({$year}): |{{$pages}}

(The pipes ("|") indicate the splice points should data be missing and the software could be smart enough to know to remove trailing spaces and punctuation marks should the last data item be missing. If no explicit "type" attribute was passed the software could choose a "best fit".)

I don't think either way would present much difficulty.

For unusal cases (like citing computers games) we could bounce them out to an ordinary template (like exit now):

 <ref name="murphy" ... type="game" />

And in the definitons page something like:

 apa:game:{{cite game|title=$title|...}}

However bouncing many citations off like this to be handled by the on-wiki templating system would drain processing time in the same way as do the current batch of templates do now so I think it would need to be reserved for the genuinely special cases. --rannṗáirtí anaiṫnid (coṁrá) 13:28, 19 January 2010 (UTC)

It seems to me that this proposal fails to address the primary issue: it does not make it easy for an academic to drop by and make an edit with a citation. We'd like them to be able to glance at a few other examples in the article and be able to guess what they are supposed to do, or at least guess close enough that bots and more experienced editors can fix it without too much trouble. ---- CharlesGillingham (talk) 18:56, 19 January 2010 (UTC)

Charles, this is more about the technical nature of the implementation so far. Any difficulties with actually inserting refs will be handled by javascripted tools. Those tools won't care much about wether we do this with the old templates, or with new functionality in the Cite extension. Such Reference insertion tools are on the list of things that are to be developed by the Usability team, and are to be integrated in the new Beta features of the wiki editor. —TheDJ (talk • contribs) 19:32, 20 February 2010 (UTC)

Huh? With the metadata attached, there should be no issue in that regard. And in the event that they can't figure out what they're supposed to do, they can stick with a basic <ref>... --Izno (talk) 19:12, 19 January 2010 (UTC)

Charles, I think between the other stuff you above you missed all that an editor would have to do i.e.:

<ref first="John" last="Murphy" title="Quisque Quis Orci Magna" publisher="Printing House Press" location="London" year="2009" />

They could of course still cite as we do now (i.e. <ref>...</ref>) and another editor could fix it up.

(The other stuff above is the back-end stuff to put it into a standardised template according to a user's or the wiki-article authors' preferences. Mo-one would have to touch/know about that stuff.) --rannṗáirtí anaiṫnid (coṁrá) 19:31, 19 January 2010 (UTC)

Not to mention I'm sure the refToolbar could be adapted or something analogous could be made to reduce the problem to filling out form fields and then clicking a button. Also, wherever was that said to be the primary issue, CharlesGillingham? Not that I find the exact purpose of this discussion clear either... --Cybercobra (talk) 02:05, 20 January 2010 (UTC)

Sorry, I guess I was a little obtuse. I'm coming in late, I apologize.

I think the primary problem is this: editors, especially those trained in an academic field, would prefer to write their citations in a particular style. They would prefer not to have to learn a new method, especially one that requires them to write citations that look like computer code. This proposal doesn't address that issue. My feeling is that the problem with our citation format is that it is difficult to learn and use, not how the citation appears to the readers. ---- CharlesGillingham (talk) 04:51, 20 January 2010 (UTC)

There seems to be a lot of people with that opinion, but very little in the way of recommended solutions. Do you have any suggestions?
— V = I * R (talk to Ohms law) 07:16, 20 January 2010 (UTC)

Please see #Proposal for restylable Wikitext citations below. Eubulides (talk) 08:36, 20 January 2010 (UTC)

Second specific proposal

See my full proposal for full details. However, what most directly impacts the citation templates is the pages= bit described above. I am proposing that all citation templates automatically get two parameters from the software: {{REFPAGES}} for the page numbers and {{REFFIRSTUSE}} for a link to the first instance in the references section (if the ref in question is a subsequent occurrence, otherwise undefined). Then, the citation templates could automatically generate the op. cit text for multiple citations of the same source, but with different page numbers. PleaseStand (talk) 05:29, 15 February 2010 (UTC)

One size fits none

I like consistency and standardization. However, I do not think that any standardized system will work equally well for every single type of article.

As evidence, I remind editors that there's no consistency in the real world. Here's a list of four common styles heavily used in universities, and what each requires for just one little point, related to the citation formatting (the words typed above the list of sources that were used to support content in an academic paper):

• Chicago: "Center the title Bibliography about one inch from the top of the page"[1] (used by historians)
• APA: "In APA style, the alphabetical list of works cited, which appears at the end of the paper, is titled 'References.'"[2] (used by sociologists and psychologists)
• MLA: "Center the title Works Cited about one inch from the top of the page."[3] (used in humanities)
• CSE: "Center the title References (or Cited References) and then list the works you have cited in the paper; do not include other works you may have read."[4] (used by scientists)

I give this as an example because we've had to address this so many times in the last couple of years, primarily with editors whose knowledge of citation systems is limited to "well, my teacher said...", so it's easy to grab, but this is just one little point; the other differences are far more dramatic. For example, the Uniform Requirements for Manuscripts Submitted to Biomedical Journals flatly prohibits citing secondary sources. You cannot simultaneously comply with URM and Wikipedia's requirements.

If academia can't agree across disciplines on even minor points -- and can't agree, despite the efforts of many professionals, because the actual needs of different disciplines aren't identical -- then why should Wikipedia impose a one-size-fits-all solution on editors? Why should the citations system used in an article about a television show be used in an article about particle physics? Perhaps more frighteningly, which of you is volunteering to tell all the physicists that they have to use the historians' system, or the other way around? Do you really think that starting a holy war over which citation system is the One True™ System is something that helps build the encyclopedia? WhatamIdoing (talk) 02:32, 18 January 2010 (UTC)

Sorry, but One True has already been taken by John Barnes. ---— Gadget850 (Ed) ^talk 00:19, 20 January 2010 (UTC)

I did not propose that we decide which citation is the One True™ System, and I would be surprised if a committee chose an existing style unmodified. Also, while many organizations—including WP at present—may not be able to get out of their own way and settle on a standard citation style, many other organizations have done so. It's really not that hard. — John Cardinal (talk) 04:35, 18 January 2010 (UTC)

Agreed with John, here. Are we really so lacking in imagination and community talent that we can't create our own somewhat unique citation style which fits our own somewhat unique requirements and community? I tend to think that we're intelligent enough to be capable of something like this.
— V = I * R (talk to Ohms law) 07:00, 18 January 2010 (UTC)

The question is not whether we're intelligent enough; the question is whether it's a good idea. It isn't. • Ling.Nut 07:22, 18 January 2010 (UTC)

This is simply untrue. There is consistency in the real world. Indeed it is an requirement of acceptance to a publication that a paper, chapter, book, etc. would conform to the consistency of that publishing house, journal or association. But the question of differences in citation styles (for the most part) is merely a question of a style decided upon by the different publisher/societies. It is not the requirement of any discipline.

For example, the journals of the Modern Language Association of America does not initialise the author's first name and separates the year from the publishing house by a comma not a semi-colon, in contrast the journals of the American Medical Association:

• MLA: Murphy, John. Book of Widson. London: Publishing House Press, 2009
• AMA: Murphy J. Book of Widson. London: Publishing House Press; 2009

This does not reveal any dark secrets about the differences between linguistics and medicine as disciplines. It is merely a stylistic choice of those two associations. Indeed, the publications of the British Association for Applied Linguistics does something different again:

• Murphy, J. 2009, Book of Widson. Publishing House Press

This difference in style between associations/journals/publishing houses is merely stylistic. Nothing more. But, importantly, it is consistent within publications. And our publication should be no different.

That does not mean that we have to drop the way naturalists, for example, refer to individual species on the 'pedia (which is quite unique and apparently very important to them) but simply that we have a consistent style (or at least be uniform in presentation, see suggestion above).

As for the Uniform Requirements for Manuscripts Submitted to Biomedical Journals prohibition on citing secondary sources, that is akin to our prohibition on citing self-published sources or original research. It is irreverent to a discussion on citation style. --rannṗáirtí anaiṫnid (coṁrá) 09:21, 18 January 2010 (UTC)

actually, the format does reveal some obscure secrets about the differences between the disciplines: for example, books are relatively much more important in literary studies , and an exact specification of the name of the author is therefore necessary--and was essential for finding people with common last names in traditional card catalogs. DGG ( talk ) 01:35, 20 January 2010 (UTC)

Comment from ACEOREVIVED

Some years ago, I made a similar proposal here,that we should adopt the referencing style of the American Psychological Association, but it was shot down quickly. People said things like "you ought to be thankful that people cite references at all" and another comment I remember was something to the effect of "please - not the brain-dead referencing system of the American Psychological Association". The difficulty here reminds me of a fable by Aesop, where men have to build a wall to defend a city - and the carpenter favours wood,the stone mason stone, the tanner leather. People are going to be used to different referencing systems, such as Harvard or that of the American Psychological Association. I am most used to the APA referencing style (as you may have worked out, I lecture in Psychology) and I admit that I have, in the past, called for adoption of the APA referencing system, but I think that difficulty could be adopting a universal referencing style might favour those in a certain discipline. ACEOREVIVED (talk) 21:15, 18 January 2010 (UTC)

What I don't understand is the reason why people seem to get "stuck" on the idea that this sort of proposal means we should choose to bolt an existing citation system onto our own. Barring any explainations, the only explaination I can think of is either a lack of imagination or some sort of willful choice to derail the issue (for what reasons this may be I have no idea, but the occasional impression that this motivation exists for some is inescapable).

Anyway... I happen to think that this is an important issue, and I think that it's something which we could accomplish. It'll take quite a bit of time, and I'm certain that like most things here it will be imperfect at first, but if people would just set aside their skepticism about our own abilities to discuss things through for a minute and give it a chance then I know we could see this through to at least a point where the vast majority will not actively object.
— V = I * R (talk to Ohms law) 21:44, 18 January 2010 (UTC)

"...people seem to get "stuck" on the idea that this sort of proposal means we should choose to bolt an existing citation system onto our own..." Or we could bolt them all on and let the user/article authors decide which one they want to see. --rannṗáirtí anaiṫnid (coṁrá) 23:10, 18 January 2010 (UTC)

I don't know that your proposal is actually an attempt ot bolt any particular style on (APA, MLA, Harvard style, etc...). I see it instead as a good example of adapting our own style/system, based on our unique wants and needs, based partially on the styles and requirements of those similar systems.
— V = I * R (talk to Ohms law) 23:39, 18 January 2010 (UTC)

Scope of work

I do not endorse a single citation style, but just for the benefit of those who like the idea, let me point out the scope of the work.

• A citation manual must be written. APA style is described in 55 pages. The Chicago Manual of Style devotes 212 pages to citations. I see no reason the Wikipedia citation manual would be any smaller.

• Using templates and/or a cite php extention does not get you out of writing the manual. First, as Fred Brooks would probably tell you, you can't write good software, or use it, without a good specification. Second, the software will never cover every situation, so there must be the option of writing the cite manually, and there must be instructions for doing so.

• Any automation agreed upon must be written.

• For existing templates, bots would have to adapt any existing cites to the new format.

--Jc3s5h (talk) 01:16, 19 January 2010 (UTC)

So, are you saying that you don't want there to be any citation style recommendations in the MOS, or are you saying that you don't want to do the work, or...?
— V = I * R (talk to Ohms law) 02:05, 19 January 2010 (UTC)

I'm saying I don't want there to be any citation style recommendations in the MOS and I don't want to do the work. I'm also warning those contemplating doing the work what they would be in for. Also, I would regard a mandatory automated system as unacceptable suppression of sources that don't fit into the automated scheme. In other words, I would be very strongly opposed to a project that intended to do a half-assed job. --Jc3s5h (talk) 02:38, 19 January 2010 (UTC)

I think a WP style could be smaller. Even if it were as large as the largest style definition out there, that would still be a reduction: as it stands now, many WP editors need access to all the pages in the complete set of accepted styles, i.e., 55+212... If you are opposed to a half-assed job, then you must be opposed to the current status quo. What's your solution? — John Cardinal (talk) 03:31, 19 January 2010 (UTC)

As it stands, any one article can be well-done, even though the vast majority of articles are not well-done with respect to sources, let alone how they are cited. If a project puts in place a rotten citation style, no articles will be well done. Worse yet if an automated system with no escape hatch is put in place, certain ideas from reliable sources will not be allowed into the encyclopedia because their sources don't fit into the automated system. --Jc3s5h (talk) 04:53, 19 January 2010 (UTC)

So, because the guideline might be imperfect or incomplete, it's better to do nothing? Where would Wikipedia itself stand if we all took that attitude? And who's talking about anything being required to fit into some sort of automated system? You're free to think or do whatever you'd like, but I'd think that expecting the rest of us to abandon this or something like it based on a nebulous fear that "it won't be well done" is a tad bit unreasonable. You seem to have some preconceived notions about this that I'm perfectly willing to admit that I don't understand. Hopefully you can explain them so that the rest of us can address the issue(s) in a way that yourself, and those who have similar fears, would at least find acceptable enough not to actively oppose the potential guidance. Simply saying (essentially) "don't do this, I don't like it" isn't constructive at all, though.
— V = I * R (talk to Ohms law) 07:40, 19 January 2010 (UTC)

Ohms law, you and whoever you can recruit are of course free to develop any new citation system you like. But if it isn't comprehensive and well-done, it will never be designated as the Wikipedia citation style, which is the stated purpose of this whole discussion. Instead, it would just be one more system for editors to deal with it if they have to edit an article that uses it. --Jc3s5h (talk) 14:03, 19 January 2010 (UTC)

Comment from Bongomatic

Like several of the above commentators, I am relatively indifferent to the presentation style of references. However, without waiting for a Utopian blue sky future, it is easy to imagine a relatively near-future database tool that allows aggregation of references across articles—permitting various tools to determine how many and what articles reference particular sources, authors, etc. This seems like functionality very much worth enabling. So my view is that structured reference calls—regardless of display style—that separate out data elements in common formats are something worth strongly encouraging or requiring. The tools to do this are currently imperfect, but the changes to guidelines, policies, and templates (if any—ever the optimist!) resulting from this discussion should enable such functionality. ^Bongo_matic 01:43, 19 January 2010 (UTC)

I agree. Something like http://zeteo.info but built into mediawiki would be a great aid. — Carl (CBM · talk) 02:09, 19 January 2010 (UTC)

(wipes off drool) F_ck yes, that is awesome! --Cybercobra (talk) 02:35, 19 January 2010 (UTC)

Yea, I agree, but... baby steps. Considering the fact that there are people who hope that we do nothing at all (see Jc3s5h comment/reply above for the most recent example) it's probably best to keep the pie in the sky wishlisting on the backburner and concentrate on staying on track here.
— V = I * R (talk to Ohms law) 03:08, 19 January 2010 (UTC)

There are many ways to stay on track. People disagree on style—fine. Let all styles share a common data tagging methodology so that aggregation and citation checking is possible. The cite templates already in use already provide a huge amount of (unused underutilized) structured data, so this is not a pie-in-the-sky wishlist. Rather, it's a suggestion that whatever route the format / multiplicity of formats / no standardized format discussion goes, that useful functionality already partially enabled is improved rather than worsened. ^Bongo_matic 03:38, 19 January 2010 (UTC)

"Unused"? Diberri's tool, Citation bot and their kin certainly use the fact that the data is structured to help fill in the blanks. Given a correct ISBN, DOI or PMID the rest of the citation is filled in by the bot to make the source verifiable. How is that "unused"? I simply cannot understand why some editors want to use hand coding, and so unnecessarily make articles utterly vulnerable to vandals that make imperceptible changes to references. Structured data with redundancy is integral to any hope of accuracy. LeadSongDog come howl 06:10, 19 January 2010 (UTC)

Proposal by Ling.Nut

I mentioned all of these points above, but am centralizing/formalizing them here:

1. I am emphatically not OK with a one-style-fits-all approach. I say this because I believe we should cater to the readers familiar with literature in a given field, the contributors with an established style, and most especially the new contributors who have adopted a style that is prevalent in their field. What would be the reason for having one and only one style? For content review. The content review folks are already committed editors. They should cater to the needs of less committed editors.
2. However, I very emphatically am on board with a new proposal that would have wording in How to cite such as follows (I'm just banging this out now off the top of my head, and have not thought it through at all; it could and almost certainly should be modified):

Editors are free to use any method; no method is preferred over another. However, the style chosen should be identical to that of a journal that is well-known in the relevant field. If no journal is available as a standard, the style should follow any well-established format on Wikipedia. The creation or use of novel or rare styles is strongly discouraged.

• • Ling.Nut 03:49, 19 January 2010 (UTC)

Oppose - There's no reason to believe that people who read Wikipedia are familiar with the literature in a given field. I also don't think it's difficult for someone who is familiar with any citation style to comprehend a citation written in some other style. Your wording is a good example of the sort of weak guideline that I am trying to avoid. — John Cardinal (talk) 04:06, 19 January 2010 (UTC)

John, the poor wording of the proposal is utterly irrelevant, since (if you can read above) I clearly said it was off the top of my head. The key point is "nothing new, only established formats". • Ling.Nut 11:10, 19 January 2010 (UTC)

Ling.Nut, I was not objecting to the specific wording, but rather, trying to indicate that weak rules where, for example, use of novel or rare styles is strongly discouraged have led us to the current state. — John Cardinal (talk) 15:38, 19 January 2010 (UTC)

But there is reason to believe that people who edit wikipedia are familiar with the literature of a particvular field!·Maunus·ƛ· 14:33, 19 January 2010 (UTC)

I think readers are more important, and I think editors will not have any trouble with a new citation style. — John Cardinal (talk) 15:38, 19 January 2010 (UTC)

The wikipedia is written by editors. Most readers couldn't care less about citation styles, the ones that do care only care if they can locate the cited reference. But editors are the ones who will have problems and loads of extra work implementing some arbitrarily chosen standard. And editors are the ones who will leave the project or simply not join when they are subjected to so many arbitrary restrictions and requirements on editing that they can find better things to do with their time than read up on 8 MBs of MOS and citation guidelines before being able to add information. Citations have a function and imposing an arbitrary uniform standard does nothing to further that function. ·Maunus·ƛ· 18:11, 19 January 2010 (UTC)

"Identical" means that if the chosen journal is a paper only journal, the html links will have to be visible rather than just turning the title of the work blue. Wikilinks to the authors will not be acceptable. Also, the developers will have to accomodate every method of linking that any any reputable electronic journal comes up with, so that we can keep our citations identical to the way the electronic journal does citations. --Jc3s5h (talk) 05:03, 19 January 2010 (UTC)

Jc3s5h, read my reply to John. You receive the identical reply. • Ling.Nut 11:10, 19 January 2010 (UTC)

Oppose - Does not further the objective of having structured data about sources available for query / analysis etc. ^Bongo_matic 12:02, 19 January 2010 (UTC)

Come again, in English please? • Ling.Nut 12:05, 19 January 2010 (UTC)

Please see my comment above. ^Bongo_matic 13:07, 19 January 2010 (UTC)

Comment by nifboy

The purpose of citations is to give as much information as is necessary for the reader to find the source being cited, and preferably pinpoint the statement being cited as well. Currently there are a number of common problems that I've seen in my work (predominately WP:VG), mostly from an inability to shove all the necessary information into a citation template and the usual "Well how do I adapt the information I have to any citation style at all?" problem. The most prominent problems seem to deal with sources where the statement is not easily accessible and "official" citation style forces what should be very related things into multiple different citations:

• Differentiating between an original source (e.g. an .mp3, or deep within the programming of a video game), and a transcript thereof, either official or otherwise. I find the latter to be extremely helpful for accessibility reasons.
  • Similarly, an article that has been published in multiple venues of varying reliability. Cite the most reliable, but link to the most accessible if possible.
• Multi-page online articles, where one article is being cited at multiple URLs (therefore multiple citations) because multiple pages are used, or only one URL is cited without page references.
• Cases where a single source, usually the novel/game/etc under discussion, is cited dozens of times for plot information, putting quotes in the citations but repeating all the other information for every quote. See the very quote-heavy Final Fantasy IX.

I don't want to see another MOS hoop FAs have to jump through: I want to see the job of fact-checking the actual content of the article made easier, and shoehorning everything into an "official" style will probably not make things any easier. Nifboy (talk) 06:47, 19 January 2010 (UTC)

Generally, I agree. What concrete steps can we take to address this, then?
— V = I * R (talk to Ohms law) 07:43, 19 January 2010 (UTC)

Always have a freeform option? Then again, WP:IAR comes to mind. --Cybercobra (talk) 07:59, 19 January 2010 (UTC)

In short, practicality over consistency. Rather than develop a "style" and forcing all references to fall in line (frankly there isn't an appreciable difference between any of the "official" styles and the output of a citation template), we should have "best practices" that focus on getting the key pieces of information into the references, especially in strange cases that no "official" style covers. Attributing a statement to a particular author usually isn't a problem, it's verification and fact-checking that's real painful. The video games project does this through lots of quoting, which is okay so long as there is dialogue to quote. For audio/video references I'd like to see transcripts where applicable, maybe even timestamps where there isn't. Anything to help answer the question "Where did this info come from?" Nifboy (talk) 18:24, 19 January 2010 (UTC)

Yes! I can't speak for Johns Cardinal here, but I know that for myself I've always envisioned more of a "best practices" guideline then the dreaded "Tome of rules" which many seem to assume is being sought here. I should also point out that a good chunk of the work already exists, albeit in somewhat disparate locations, at for example Wikipedia:Citing sources and Wikipedia:Footnotes. We simply need to pull together the similar information, organize it a bit better, and probably extend it in some (many?) ways.
— V = I * R (talk to Ohms law) 21:41, 19 January 2010 (UTC)

"Johns Cardinal"? Is that like "Attorneys General"? <g>

I proposed a standard citation style because it seemed like having one would solve some issues directly and help other issues indirectly. I certainly didn't want to impose a tome of rules on anyone (including myself), but we impose one of the citation styles on editors now. The difference is, rather than a guideline choosing which style, the editor who adds the first citation to the page gets to choose it!

I certainly wouldn't oppose efforts to improve the current documentation, but I reviewed a lot of the pages recently and decided that the improvements would be marginal without a standard. That was actually the trigger for me making the proposal. — John Cardinal (talk) 23:13, 19 January 2010 (UTC)

The purpose of citations is more than to give the reader information to find the article: it is a very rough and approximate surrogate for indicating the likely reliability. For newspapers, it immediately gives us for familiar newspapers what we need to know about suitability as a RS. If it is a journal for example, it indicates immediately the nature of the journal, and, with experience, the expected quality. If it is a book, the name of the publisher can be immediately significant. DGG ( talk ) 01:31, 20 January 2010 (UTC)

Attribution, of sorts, is also often another large consideration.
— V = I * R (talk to Ohms law) 01:39, 20 January 2010 (UTC)

To put into writing what I think the key points are:

• Attribution of the author and publication (who said it?)
• Identifying the work being cited (what kind of source is this?)
• Pinpointing within the work the information being relied on (where can I find this info?)
• Date(s) for context (when was it said?)

So long as those key points are there and can be reasonably followed to the source I frankly don't care what "style" is followed. Nifboy (talk) 04:12, 20 January 2010 (UTC)

Comment by Maunus

Setting a fixed citation system for wikipedia is not only fixing a problem that doesn't exist, but it is likely also creating several new ones. The problem that the solution is supposed to fix is one of form - it is an attempt to create uniformity where none is needed. The purpose of citations is to make content verifiable - nothing more. Whether they are formatted in one way or the other contributes no further to this goal. I.e. multiple citation styles are at most a cosmetic problem. The problems that enforcing a uniform citation style would create are non the less immense: It would create another obstacle to attracting new contributors by once more raising the bar for participance to only those who will bother reading a large citation style guide. It would alienate those contributors who are experienced users of one of those citation systems that are not chosen. It would create a new workload for those areas of wikipedia involved in evaluating articles (FA and GA) and for the common editor. It is quite simply not a necessary or good idea.·Maunus·ƛ· 14:44, 19 January 2010 (UTC)

Additionally I would like to ask the proponents of a unified citation style to answer this: Which wikipedia policy is this proposal based in? Which of the five pillars does it work to bolster? ·Maunus·ƛ· 12:55, 23 January 2010 (UTC)

The first pillar includes this text, "Content should be verifiable with citations to reliable sources." The lack of a standard citation style makes it more difficult to meet both the spirit and the letter of that requirement. Partial sources are accomodated because there is no standard that says which elements must be included. Inconsistent, poorly-formatted citations make it difficult for readers to determine if article content is properly sourced.

If you think everything is hunky-dory with the status quo, please do a broad review of non-FA articles. — John Cardinal (talk) 14:28, 23 January 2010 (UTC)

I beg to differ. This proposal does nothing to improve the actual verifiability of articles. Whether or not all citations in an article has ISBN or publishing location does not hinder me for a second in finding the reference. If an article has sources and they are cited in a way that allows readers to identify them then that is the pillar of verifiability is met. The proposal is strictly cosmetical and as such goes directly against the spirit of Pillar 5 and 3. Wikipedia should not have strict rules or requirements outside of the five pillars. It should encourage everyone to edit and add content within the spirit of the five pillars no matter what kind of citation style they prefer for the articles they are writing. This and other kinds of Feature- or Instruction creep deter new contributors from joining and existing ones from contributing. The first step towards convincing me that ththere is is a problem would be if you could show me that non-editing readers have in fact complained about non-consistent citations in any significant way. Up untill this point this only looks like a problem because you and a couple of other editors that seem to like rules for the sake of rules, say it is. Show me the problem - don't just tell that it exist (or that I would have noticed it if I had looked better). ·Maunus·ƛ· 14:42, 23 January 2010 (UTC)

Hmm, by this argument, shouldn't the entire Manual of Style be abolished? But I'm afraid that the wrong part of the first pillar was identified. The operative part is the very first sentence of the very first pillar: "Wikipedia is an encyclopedia that incorporates elements of general and specialized encyclopedias, almanacs, and gazetteers." One of the elements of any encyclopedia is a standardized style that helps readers focus on content (the important part) and not on the format. Of course one can take standardization too far, but the pillars do not argue against all standardization. Eubulides (talk) 17:22, 23 January 2010 (UTC)

At least I would argue that the MOS should be given less importance than some people do. But no, not all standardization is bad - just the kinds that are unnecessary and as I am saying nobody has cared to show me that there is a necessity here. ·Maunus·ƛ· 18:10, 23 January 2010 (UTC)

Re: Comment by Maunus above about ISBN or publishing location. I never mentioned those particular data elements. The point is, no data elements are required. Also, I never said I like rules for the sake of rules. I've tried to describe why I think a standard citation style would help WP. Please assume good faith and please restrict your comments to the proposal and not my personal characteristics; you don't know me at all. — John Cardinal (talk) 18:05, 24 January 2010 (UTC)

It is not my intention to say any thing about you as a person. I don't believe I have said anything about you as a person, but rather about your arguments (or lack thereof), but I apologize for not wording that clearly. I am certainly not assuming anything other than good faith from your part or that of anyone else here involved, even if I find their opinions to be misguided or unfounded. What I said was that the people who propose this change seem to like rules for the sake of rules simply because they have not presented any compelling arguments that this change is anything other than rules for the sake of rules. ·Maunus·ƛ· 19:09, 24 January 2010 (UTC)

So if I write that your comments make it seem <insert insult here> that's OK? I think not, but I am moving on... — John Cardinal (talk) 21:35, 24 January 2010 (UTC)

Well, I wasn't aware that "rule loving" was considered an insult in your part of the world. In my part of the world it is considered a fact that people have varying opinions about how best to legislate, I'm sorry about that. Anyway I am still hoping that you will care to show me that there is a problem that needs to be taken care f rather than just tell it to me.·Maunus·ƛ· 06:26, 25 January 2010 (UTC)

¶ There's just an honest but fundamental disagreement: some people see uniformity as a virtue in itself (which implies lots of rules) and others don't, or rank its importance very low. (Uniform style is not one of the "elements" of other reference works that a necessarily-eclectic Wikipedia necessarily needs to imitate too closely.) There are important points about accessibility, technical limits, clarity, confusion, ambiguity, obscurity and the avoidance of needless offens/ce or possible actual harm that have to be in a style sheet, but they're a fraction of what's covered in the Manual of Style. The rest belong in some other vehicle for suggestions and guidance about common or best practice, such as a collection of essays or a periodical on the lines of the Signpost. The best and most honest of intentions to make the Manual a source of guidance or advice rather than an imperative Book of Rules or Body of Law founder on an army of 'bots, self-appointed MoS enforcers, and the MoS "compliance" required by the Featured Article and Featured List Criteria. —— Shakescene (talk) 23:37, 24 January 2010 (UTC)

Comment by Eubulides

One way to achieve a uniform style for readers that prefer it, without forcing such a style down everybody's throat, is to have a new user preference as to what style they want, and to have pages be automatically formatted to the user preference. The default style could be set on a per-page basis, by consensus of editors on that page. That way, readers who like a uniform style could see one, and most readers, who don't care, will get a style appropriate for that article. This would require MediaWiki help but shouldn't be that hard to implement efficiently. If done well, it could even accept hand-formatted citations in few well-supported formats (other formats would simply go through as-is, presumably). But at the very least it could be done via templates.

The comments about the citation templates being terribly slow are right on target. Well-sourced pages take 10 or 20 or even 30 seconds to load, mostly due to the overhead of citation template formatting: this delay is unacceptable. Also, the resulting page bloat due to metadata that 99.9% of readers don't want hurts Wikipedia as well, particularly for the majority of the Internet (the poorer majority) that is not on broadband. This metadata should be omitted except for readers who actually need it.

Eubulides (talk) 21:56, 19 January 2010 (UTC)

I have a fairly new and powerful computer, but even with that it has reached the point with articles containing lots of citation templates that I just give up before they open. They're also impossible to copy edit for flow. I've always been opposed to these things, but surely the point has been reached when their use needs to be reviewed, per WP:ACCESS if nothing else. SlimVirgin ^{TALK contribs} 22:34, 19 January 2010 (UTC)

I'd like to point out that list defined references are intended to directly address the "flow" issue, which I certainly agree can be a significant problem. If nothing else comes from this, perhaps further awareness, acceptance, and possibly even development of list defined references will occur.
— V = I * R (talk to Ohms law) 00:14, 20 January 2010 (UTC)

Surely the slowness would only happen when loading the References sections, which would be at the end of the article anyway? And References in general slow down page loading for the 70% of readers who don't care about them. --Cybercobra (talk) 02:23, 20 January 2010 (UTC)

No, the slowness occurs for all logged-in editors, before loading the References section. If you're on a high-bandwidth connection, most of the slowness occurs within the Wikipedia server that's trying to render the page. It doesn't help to have a fast browser or a good network if the Wikipedia server itself is slow, which is a major problem here. Eubulides (talk) 07:25, 20 January 2010 (UTC)

Regarding your suggest re: user preference etc., there is a proposal to do just that above. It's relatively simple. The proposal is to bypass wiki-based citation templates in nearly all cases so as to address avoid the current issues regarding slowness of loading them and addresses the copy-editing issue. --rannṗáirtí anaiṫnid (coṁrá) 23:05, 19 January 2010 (UTC)

Thanks, I'll take a look at that. Eubulides (talk) 07:25, 20 January 2010 (UTC)

PS. Please see followup proposal in #Proposal for restylable Wikitext citations below. Eubulides (talk) 08:36, 20 January 2010 (UTC)

This has been discussed before and doesn't seem to go anywhere, but canning metadata in citation templates seems to be an easy measure to improve the experience for the vast vast majority of our readers. I would also agree that getting more users on board with template-style citations would be easier if they were built into Mediawiki and thus could easily offer additional features (e.g. rendering in a style based on user preference). Useful features would at least begin to balance the increased workload and confusion they seem to carry with them. Christopher Parham (talk) 23:46, 19 January 2010 (UTC)

Do we have any actual numbers on this performance question? I for one am inclined to doubt that the metadata would even begin to approach the size of the images commonly used in large articles. Perhaps if we identify a few of the problematic slow articles we can begin some sandbox testing. Examples? LeadSongDog come howl 03:08, 20 January 2010 (UTC)

Many readers turn images off, if they are on low-bandwidth connections (this describes more than half the Internet, and is the default in most of Africa, for example). For them, the metadata are a significant part of downloading time. Please see Wikipedia talk:Citing sources/Archive 27 #Size increases due to COinS vs non-COinS bloat, which shows that for high-profile articles such as The Beatles and Michael Jackson the metadata bloats the HTML of the entire article by 9% to 19%, and that for the well-cited feature article Autism the metadata bloats the HTML by 24% to 35% (depending on whether you measure compressed or uncompressed). This is way too much overhead for a feature that hardly anybody uses. The feature should be disabled by default, and enabled for editors who need it. Eubulides (talk) 07:25, 20 January 2010 (UTC)

Ah, so the issue is the inline COinS in the rendered html, not metadata per se. That makes more sense. If the cache system permits, eliminating the COinS for IP browsers would be a reasonable option. Even simpler, change {{ISBN}}{{citation/core}} to wikilink to ISBN instead of to International Standard Book Number. For the few times it's clicked through, the redirect has a low cost, while the lengthy label on every reference has a substantial one. Similar for the links to ISSN, PMID, DOI, etc. all of which are presently to the full name when the abbreviation would be better for the purpose. This trivial change alone would shorten the COinS metadata by about 60 bytes/citation. LeadSongDog come howl 17:58, 20 January 2010 (UTC)

Comment by Circéus

I agree with Maunus above. Even when applying {{cite journal}}, editors frequently differ in how they apply it. Some, for example, include much more often than not the publisher field (which in 90% of cases is unnecessary for journals, though I've seen it used ingeniously to indicate a conference proceeding published within a journal), and medical editors insist on applying medical style with abbreviated journal and idiosyncratic format for given name (dotless and spaceless initials), on the other hand (and yes I'm being a bit mischievous there :p), some insist on adding the {{aut}} template because they cling to small-capital author formatting.

And I am completely opposed to any software-based system. One will simply move the problem: those that already refuse to use the citation, and they are quite a few (see here and subsequent edits for a good example), will simply refuse to use the ability or leave the site entirely if forced to do so. For the record, I do not consider the fr:WP Citation: namespace to be a better solution (although interesting for its bibliographic capacities, I find it even more abstruse, if that was possible, than the en: templates).Circéus (talk) 23:12, 19 January 2010 (UTC)

Sorry, what's "the fr:WP Citation: namespace"?
— V = I * R (talk to Ohms law) 00:17, 20 January 2010 (UTC)

My bad for misremembering the namespace. It is in fact called Référence:, and it is on the French Wikipedia (though not widely used AFAICT). See for example here or here. —Preceding unsigned comment added by Circeus (talk • contribs) 00:53, 20 January 2010 (UTC)

Thanks! As someone who barely reads French, haltingly at that, and who isn't terribly familiar with their standards and practices, would you mind taking a moment to explain what their actually doing there, and how it's related to us? I appreciate you taking the time.
— V = I * R (talk to Ohms law) 01:30, 20 January 2010 (UTC)

The idea seems to be that the Référence: namespace contains all the gritty elements of the bibliography, including, potentially, multiple editions, which is then linked from the article where only the barest elements then need to be provided, regardless of the article where this is done. Of course this being Wikipedia, it's not so well standardized and approaches differ significantly between editors. It seems to be most used by editors who reuses a small set of books repeatedly across many family (User:Gdgourou, who created the first of these refs, specialises in Disney material). Circéus (talk) 01:45, 20 January 2010 (UTC)

So, it basically uses transclusion, along with a namespace set aside for the specific use? That's an interesting idea really (especially since similar set-ups have crossed my mind). That seems to be a relatively low cost solution to some issues, at least. Aside form an apparent lack of buy in (which it seems that you're indicating above), are there any significant issues that people have with it's use?
— V = I * R (talk to Ohms law) 01:55, 20 January 2010 (UTC)

No, not transclusion. Look at my second link. Can you seriously imagine transcluding that?? Circéus (talk) 04:55, 20 January 2010 (UTC)

Ah... so, they just add links to the appropriate references page then? I'd probably hate that annoying extra step as a reader, let alone as an editor... I do think that using transclusion could be a possibility though, although I'm not at all sure about the space/performance impact that would have here. Using "single purpose" templates like that should make the articles much cleaner though, as long as they were located at short enough names. It'd give us tons of space to add all sorts of (non-transcluded) "metadata", and possibly some brief notes, as well.
— V = I * R (talk to Ohms law) 06:58, 20 January 2010 (UTC)

Looks like they just use links. See for example: fr:Médée (Euripide), there's a link to fr:Référence:Médée (Euripide) in fr:Médée (Euripide)#Voir aussi. Yuck.
— V = I * R (talk to Ohms law) 07:03, 20 January 2010 (UTC)

Comment by DGG

Unfortunately, there is no standard accepted citation style. Every individual academic journal has a required style, and they are most of them slightly different. In some subjects there is a fair degree of standardization, such as medicine,; in others there is a practical degree of standardization, such as the APA Manual for psychology. Our normal reference when in doubt in the Chicago Manual of Style, which explains a number of different ones, all of which are acceptable in different circumstances. Some of it is properly field dependent--a field dependent on books will have a very different preference from one dependent on articles in a small number of well-known journals. Classics, for example, requires a style using two or three letter codes for the journals, and a cryptic system of abbreviations: it works within the subject for experts, because the field deals with a finite body of work. One of the virtues of personal bibliographic software is that they permit switching references from one style to another--provided sufficient information was collected in the first place (for example, some styles require beginning and ending page numbers for articles, some insist on just the beginning page). In my experience as a librarian, when a student asked for help with this, the first question we asked was, which professor is it for, & do you have the syllabus explaining what he wants?

Wikipedia must reflect the outside world. This can be unfortunate, as here, for there is no good model: the publishers of general reference works that adopt a uniform style give references that cannot be converted into any number of specialist styles. The best hope is the adoption of standard numbering systems from fee sources when available: the PMID for example from PubMed. Interestingly, OCLC is now indexing journal articles and giving them OCLC numbers, which might be a major factor in general standardization (for an example, go to advanced search, select "accession number" as a search field, and enter 481838932) -- and I have just discovered it work in the main search box also: just enter the number. Tho progress, it makes all of our journal article formats obsolete, for that field must be added to use when available--and also makes obsolete all the journal formats in all bibliographic software. (and for good measure, it is not wholly unambiguous, for the journal as a whole also has an overall OCLC accession number, and the field for OCLC number in any existing journal format refers to that, not the article one.) DGG ( talk ) 01:28, 20 January 2010 (UTC)

Isn't "do you have the syllabus explaining what he wants?" essentially what we're asking for here (at least partially)? I think that's essentially what we were saying above, at #Comment by nifboy.
— V = I * R (talk to Ohms law) 01:37, 20 January 2010 (UTC)

Right idea, wrong methodology (comments by CG)

While I agree that a consistent citation style would have many advantages, I strongly disagree that it can be determined by a committee or that it should become part of the manual of style. I believe that one of the essential features of Wikipedia is that we try to make decisions from the bottom up, rather than the top down. The only methodology I think is in the spirit of Wikipedia is this:

1. Find an article that has bad citations, inconsistent citations or idiosyncratic citations.
2. Propose on the talk page that you would like to change the citations to a style more consistent with the rest of Wikipedia. Describe the style you intend to use.
3. Wait one month.
4. If there are no objections, change the article to a style more consistent with the rest of Wikipedia.
5. Mark the article on the talk page as "reviewed" with some standard template that carries a date. If local editors insist on an idiosyncratic style, make a note of this in the template. Could also categorize unusual choices and problems in the article so that future efforts will be able to find it. (E.g., something like: style=Chicago, citation templates=no, page number method={{rp}}, combined refs=no.)
6. Repeat (1..5) a few million times.

To help with the process, one could also:

1. Encourage others to do the same, through Wikiprojects etc.
2. Build bots that carry out aspects of this process (but do not build bots that skip steps 2 and 3).

The key point is this: we need to gather consensus from all of Wikipedia's editors. The only way to do this is one article at a time. Gradually, over several years of effort, Wikipedia's articles will begin to converge towards one or two standards.

In fact, I believe there already is a defacto standard citation method in Wikipedia: General references are only used for sources that apply to the whole article, and only if the number of inline citations to it would be ridiculous. Footnotes containing full citations are used for a source that is only used once (or twice). Footnotes are combined for identical citations. Shortened footnotes are used when multiple pages of the same source are needed. Citation templates are preferred over handwritten citations. (Examples of each these are at WP:Verification methods.) That is my gut feeling. A precise count would be useful, so my last proposal would be this:

• Prove conclusively that Wikipedia's editors prefer one style over another by simply counting the articles that use it. If more editors use it, it is the defacto standard.

This allows all of Wikipedia's editors to vote with their keyboards. The style guides should reflect whatever this kind of research discovers: they should be descriptive, rather than prescriptive.

I don't know if debates like this one will ever resolve some of these issues. I propose that instead we focus on the actual task at hand. The most realistic, inclusive and practical method to create a standard citation style is to improve individual articles. ---- CharlesGillingham (talk) 02:42, 20 January 2010 (UTC)

I'd like to note the existence of Wikipedia:WikiProject Citation cleanup. Nifboy (talk) 04:13, 20 January 2010 (UTC)

Yes, that seems like a natural place to host an effort like I've outlined above. The talk page of the project suggests that there's not a lot of activity over there. ---- CharlesGillingham (talk) 05:00, 20 January 2010 (UTC)

Top-down, bottom-up, doesn't matter. Shoving any standard down editors' throats by any means is simply wrongheaded. A foolish consistency is the hobgoblin of little minds,and all that... • Ling.Nut 06:16, 20 January 2010 (UTC)

Ling.Nut and others seem to have some preconception that those of us supportive of this proposal are looking to "Shove a standard down peoples throats", which is completely untrue (let alone a poor showing of good faith). I think that I'm fairly typical of the vast majority of editors in that I really could care less what "standard" or "best practice" we use here, I'd just like to know what it is so I don't have to guess and worry about having a confrontation with someone else over it. I'd love to "vote with my keyboard", but it's hard to do that when (to inappropriately stretch the analogy) you have no real idea of who the candidates are and what they stand for. I don't give a whit about APA, MLA, Chicago style, or Harvard style, and I'm frankly sick and tired of hearing about them. I want an easily usable style which I know will be generally supported by the community here in order to avoid most further discussion about this. Some of the opposition here is exactly correct in that this is a silly issue to argue about, so let's end the silly arguments once and for all already. Either dump all foreign styles in favor of our own custom set of styles (which I personally favor), or pick something that is appropriate for general use here.
— V = I * R (talk to Ohms law) 07:14, 20 January 2010 (UTC)

The core problem being that no matter which option is hypothetically chosen, the choice is essentially an arbitrary one (excepting patently confusing/unreasonable options). So it's hard to apply NPOV beyond the status quo of "anything that's not insane, and avoid blatant inconsistency". --Cybercobra (talk) 07:50, 20 January 2010 (UTC)

Ohms law, those who are opposed to standardization of citation styles on Wikipedia have given various (some of them valid) reasons why they are opposed to the proposal. Proposals for mandatory citation styles would indeed force editors to use the chosen citation style or styles. Despite what you and several other editors feel, several others do not feel it is "an embarrassment" to have multiple citation styles on Wikipedia, a project which covers a thousand different topics and fields. No one is accusing you or anyone else in this conversation of bad faith, but several editors do feel like a required citation style would indeed force editors to use a mandatory style, simply because we've been editing for years, and we already know that FAC and GAN reviewers often already incorrectly impose rules like, "you must use a citation template" and "you must cite in X format" (diffs can be provided upon request), despite what it says at the top of WP:CITET. Moreover, a uniform citation style guide will force editors to jump through extra hoops at FAC because FAC specifically requires that an article must "follow the style guidelines". Meaning no matter what you're proposing here, it will absolutely be made mandatory at FAC. It will be "forced" on editors, if they want a gold star on their article. You absolutely cannot state that nobody will be forced to follow the chosen citation style, because editors who don't follow that style will see their nominations at FAC and GAN fail. Firsfron of Ronchester 08:21, 20 January 2010 (UTC)

OK, for the sake of argument, let's say that no matter what we do (such as offer 2-3 clean and firm "this is how to do it" options) that you're 100% correct. What's the problem with that, and why should anyone be particularly concerned about preferring one format over another? We don't really give editors a choice in Header styles, or font styles either, and that doesn't seem to be an issue with anyone. It'll be an essentially arbitrary choice or set of choices on our part, but so what? If you don't want to use that style there will still be nobody who forces you to do so, or punishes you for not doing so. Just acquiesce by not bitching when the style is changed by others is all, and don't spend a moments thought on the issue if you'd like. And, if there's really something wrong then you'd just do what we all do all the time here, go to the document describing the guideline and seek change. The obstructionist attitude regarding this issue is honestly and frankly completely baffling to me...
— V = I * R (talk to Ohms law) 08:48, 20 January 2010 (UTC)

You're asking me to be quiet now and wait for the style guideline to be put in place, and then "seek change" to have it removed(!). I'd prefer to cut out the middle steps and not have the guideline created in the first place, thus saving everyone quite a bit of time and effort. I don't see that having different citation styles across a project as large as Wikipedia is a "problem" or an "embarrassment", and I'm not the only editor who doesn't see it. But we should wait for the "solution" anyway? That's crazy. You and several other fine editors are the ones seeking change, so you should endeavor to explain why the change is absolutely necessary (and "because we're arguing about it" is not at all a valid reason for change; I could argue for any silly proposal and then, when asked why it should be enforced, say, "because not having it causes arguments"). I'm not saying you aren't acting in good faith; I'm saying I don't see the need for the proposed guideline, because I don't see a problem here, despite your obvious attempts at explaining what you feel is a problem. Firsfron of Ronchester 09:15, 20 January 2010 (UTC)

Nah, that's not what I'm asking at all...
— V = I * R (talk to Ohms law) 10:02, 20 January 2010 (UTC)

Proposal for restylable Wikitext citations

Here's an alternative proposal for citation formatting, which builds on rannṗáirtí anaiṫnid's #Specific proposal above, but tries to be easier to use. This is a pure extension to what we have now. The idea is (1) use ordinary Wikitext in the citation and tell MediaWiki what style it is, (2) have the References section specify which style is generated by default, and (3) let logged-in editors override this default.

Here's what references would look like while you're editing them. You get to choose the format you use while you're editing. The following examples show two different editing styles, "van" for Vancouver and "apa" for APA:

<ref style=van> Kossoff EH, Freeman JM. The ketogenic diet—the physician's perspective. In: Stafstrom CE, Rho JM, editors. Epilepsy and the ketogenic diet. Humana; 2004. ISBN 1-58829-295-9. p. 53–61. </ref>

<ref style=apa> Harlow, H.F. (1983). Fundamentals for preparing psychology journal articles. ''Journal of Comparative and Physiological Psychology'', ''55'', 893–896. </ref>

If the reference section looks like this:

=== References ===

{{reflist|style=van}}

This would format all references in the Vancouver ("van") style, even if they were originally expressed in some other style. Thus the output would look something like this:

• Kossoff EH, Freeman JM. The ketogenic diet—the physician's perspective. In: Stafstrom CE, Rho JM, editors. Epilepsy and the ketogenic diet. Humana; 2004. ISBN 1-58829-295-9. p. 53–61.
• Harlow, H.F. Fundamentals for preparing psychology journal articles. Journal of Comparative and Physiological Psychology. 1983;55:893–896.

Conversely, if the reference section said "{{reflist|style=apa}}", or if the user preference is the "apa" style (regardless of what the reference section says), the output looks like this:

• Kossoff EH, Freeman JM. (2004). The ketogenic diet—the physician's perspective. In Stafstrom CE, Rho JM (Eds.), Epilepsy and the ketogenic diet (pp. 53–61). Humana. ISBN 1-58829-295-9.
• Harlow, H.F. (1983). Fundamentals for preparing psychology journal articles. Journal of Comparative and Physiological Psychology, 55, 893–896.

Sharp-eyed readers will notice that neither output, strictly speaking, is purely in either style, due to the difficulty of automating minor details like whether author initials have periods. But it's close enough so that people won't care much, and if they do care, we can improve the processor to handle initials and suchlike automatically.

Here are some assumptions that I'm making.

• Ordinary wikitext can be parsed if it's known that it uses APA style, or Vancouver style, or whatever. Obviously there will be corner cases that are hard to parse, but we can support them by using (shudder!) templates, with the goal being that the vast majority of citations can be in ordinary wikitext without templates. This is a major assumption, and we won't know whether it's true without some realistic testing.
• This can be implemented so that performance is not noticably worse than formatting citations by hand without templates.
• COinS metadata are not generated by default, but the logged-in user can ask for the metadata.
• References without the new "style=" parameter are not reformatted, so existing usage still works.

If this can be made to work, this approach has important advantages:

• Editors can edit using the style they prefer.
• Readers can read using the style they prefer.
• It's much easier to read the Wikitext than it is with templates.
• A citation can be cut from one article and pasted into another without needing to modify the citation, even if the destination article uses a different style.

Eubulides (talk) 08:36, 20 January 2010 (UTC)

The parsing involved would only exacerbate the speed concerns expressed above, and I think you underestimate the complexity involved in implementing such parsers. --Cybercobra (talk) 08:51, 20 January 2010 (UTC)

While I agree with cybercobra that the parser would be difficult to write, I think that this proposal at least addresses one of the key issues. ---- CharlesGillingham (talk) 08:58, 20 January 2010 (UTC)

Yes, the proposal is not trivial to implement, but that doesn't mean it's unimplementable. Surely there are ways to speed up the parsing, such as writing the expensive parts in C as a PHP extension (or, I hope, something less drastic than that). Eubulides (talk) 09:08, 20 January 2010 (UTC)

Previous proposals to enhance Wikipedia by extensions written in C were rejected (like better programming language for temlates, see archived discussion from VP), based on the assumption that content from Wikipedia should be easily copyable to MediaWiki instances even on hostings that don't allow such extensions. Svick (talk) 10:17, 20 January 2010 (UTC)

As you say, if a user formats their citation correctly, according to that style, the citation should be parsable. (I don't think it would not even need to be explicitily stated to be one citation style or another.) But if someone was to use a comma instead of a semi-colon, things could go awry quite easily.

The two methods are not mutually exclusive. And there's good reasons for having a parser like this. If it was kept simple, and it didn't make any assumptions, it could work well.

I would however be tempted to suggest that this kind of work be better left to a bot process i.e. allow users to type out their reference as normal, if other editors want to put it into an attribute-based citation then good for them, over time a bot process could parse recognisable "hand-written" citations into attribute-based citations.

I'm working on implementing a demo of the "attribute-based"-based proposal above. I'd be happy to work a version of your proposal above into it. Until we see these kinds of things working, it's difficult to know what to make clear decisions about (either or any of) them. —Preceding unsigned comment added by Rannpháirtí anaithnid (talk • contribs) 09:30, 20 January 2010

Your mission: To persuade me that this is not a solution in search of a problem... • Ling.Nut 09:39, 20 January 2010 (UTC)

LOL. For a good part, I think you're right. The "problem" is mainly cosmetic: it would look nicer if citations followed similar format.

While this is trivial, it is a desire that is at the heart of citation styles. The origin of citation styles is cosmetic and the difference between them is cosmetic. The various citation styles exist solely so that citations within a publication have a uniform cosmetic. The "problem" is mainly cosmetic because the subject matter is cosmetic. That doesn't mean that it doesn't matter (we look for a uniform cosmetic other aspects of the 'pedia). And it would be quite silly to disregard a desire for a uniform cosmetic when speaking about citation styles when a desire for a uniform cosmetic is their very origin and purpose.

I said "mainly" above because, whereas the citation style is purely cosmetic, the citation itself serves a practical purpose. We all know what that is: in our 'pedia it is to reference claims to secondary (and occasionally primary) sources. The practical information to be contained within a citation is quite well agreed upon. Citations are made up of discrete data. So long as we capture that data, we capture the citation. For that end, I like templates. For one thing, they can aid inexperienced editors to cite properly (by showing the discrete data they should provide, etc.). For another thing, they open the practical data within a citation to become machine readable (so that we can have niceties like author indexes, bibliographies in books, etc.).

Templates is one way we could us get around the cosmetic question too. I don't think that we will be able to agree on a common citation style. The difference between the different styles are simply (quel surprise!) stylistic: do we put the year in parenthesis or after a semi-color, at the end of the citation or after the author's name? In other publications, an editor would simply dictate the house style. In our publication, we don't have the luxury of a single editor and the choice in this case would be so arbitrary that there is nothing for established editors to agree on or for new editors to naturally assume.

The systems being proposed side-step the cosmetic choice by leaving it to the authors of a page to decide or for the reader to override. The important thing is that the sole purpose of a citation style is achieved (i.e. we have a relatively uniform cosmetic). As second win, editors who are less capable of citing sources are shown the data they need to cite properly. And, as a third win, citation data is available to be re-used in imaginative ways.

Additionally, the system being proposed avoids the pitfalls identified in the current citation templates. And it doesn't prohibit anyone from citing as they o right now.

I don't know if that will convince you. As I said, for a large part the "problem" that is being solved here is quite trivial. However, I think the "solution" comes with enough additional wins, and no costs that I can see, to make it worth the effort (on my part). I'm going to put together a demo on an external server. See how you feel then. --rannṗáirtí anaiṫnid (coṁrá) 14:01, 20 January 2010 (UTC)

When various proposals about formatting dates were made, certain solutions were absolutely unacceptable because creating different versions of a page for different readers could mess up the Wikipedia scheme. You must obtain assurance from a technical expert that your scheme would not create a performance problem. --Jc3s5h (talk) 14:24, 20 January 2010 (UTC)

Certainly. --rannṗáirtí anaiṫnid (coṁrá) 15:14, 20 January 2010 (UTC)

If this is a real performance issue, we could drop the part of the proposal about having per-user preferences about citation format. It would be very nice to have that, but it's easily detachable from the other features. Eubulides (talk) 15:47, 20 January 2010 (UTC)

Rannpháirtí anaithnid, for the demo, could you please try the following contrived example? It tests several issues that would need to be addressed in any such proposal, including special characters, wikilinks, template calls, and parameters used in one style but not another.

• {{cite web |author=John Gay, Jr. |author.= |authorlink=John Gay |work=Cheveux, "oeil" & <peigne>? |home.= |trans_home=Hair, "eye", & <comb>? |homeurl=http://blog.example.com/~phillips/hair-eye-comb/ |location=[[Walla Walla]], WA |publisher=[[Idiosyncratic Press]] |format=PDF |title=Τι να κάνει όταν κλέβεται η περούκα σας |trans-title=What to do when your wig is stolen |url=http://blog.example.com/~phillips/hair-eye-comb/stolen-wig.html |date={{circa|lk=no|2001}} |archiveurl=http://archive.example.com/stolen-wig.html |archivedate=2009-12-23 |accessdate=2009-12-24 |pages=659–61 |size=174 kB |language=Greek |quote=My goodness! Chase after the thief! |ref=Phillips2009 |separator=; }}

This currently generates the following:

• John Gay, Jr. (c. 2001). "Τι να κάνει όταν κλέβεται η περούκα σας" [What to do when your wig is stolen]. Cheveux, "oeil" & <peigne>? (in Greek). Walla Walla, WA: Idiosyncratic Press. Event occurs at 174 kB. pp. 659–61. Archived from the original (PDF) on 2009-12-23. Retrieved 2009-12-24. My goodness! Chase after the thief! {{cite web}}: Check date values in: |date= (help); Unknown parameter |homeurl= ignored (help); Unknown parameter |separator= ignored (help); Unknown parameter |trans_home= ignored (help)

but if you replace "cite web" with "vancite web", it generates Vancouver style instead:

• John Gay, Jr. [Hair, "eye", & <comb>?]. Walla Walla, WA: Idiosyncratic Press. Τι να κάνει όταν κλέβεται η περούκα σας [What to do when your wig is stolen] [PDF]; c. 2001 [archived 2009-12-23; Retrieved 2009-12-24]; p. 659–61; [174 kB]. Greek. "My goodness! Chase after the thief!"

Eubulides (talk) 00:50, 21 January 2010 (UTC)

This discussion — not merely this thread; other threads above as well — is perilously overstocked with programmers lusting after a nifty hobby project. However, there is no strong argument that any of the "improvements" that these programmers are so enamored with and are pushing so stoutly are actually needed or wanted. Restylable cites? WTF? Most people do not even glance at the cites. When I say most, I mean at least a solid 95%. Of the remaining 5% who do look at them, no one gives a sh*t if the cites have this cool restylable feature. Everyone will look at it and go "Mmmm, cool" then never use it again. Can we get a little common sense to enter the room at some point? • Ling.Nut 00:57, 21 January 2010 (UTC)

95% of readers don't edit, but that doesn't mean that we should reject all proposals for improving the editing process. Editors who don't care about citation format can simply ignore the new feature. I am not looking for a "nifty hobby project" and would rather that someone else get this stuff working. Eubulides (talk) 01:54, 21 January 2010 (UTC)

I wholeheartedly agree that change can be good, but change must offer a meaningful benefit to a large enough segment of the readers to justify the cost (in terms of download time & server load etc., but also in terms of increasing the learning curve for editors).. Here I see none. • Ling.Nut 04:03, 21 January 2010 (UTC)

Not to mention that this is likely also an impractical solution. As described, we would need both a parser and a generator for each citation style (MLA, APA, Chicago, etc.). In particular, these three citation styles vary widely and were designed to be written by humans, not computers (at least without human intervention). For example, only the first word of a title is capitalized in APA style, while in MLA style, the first and last words, as well as all others except articles, coordinating conjunctions, and prepositions with fewer than five letters. Another example is that these citation styles tend to differ slightly depending on the type of source. Journal, newspaper, or magazine? Good luck writing a computer program to figure that one out 100% accurately and with acceptable performance. Requiring the computer to generate all three citation styles is not ideal. Remember MOS:UNLINKYEARS? The consensus was that we unlink trivially-used dates in articles, and we have not seen date autoformatting since. Including every date in a special template is too much of a maintenance burden so it has not been done (and 99% of users will never register and set their date preferences). I say the same for requiring every article to include every citation style. PleaseStand (talk) 23:10, 14 February 2010 (UTC)

Proposal by Jmh649

I agree that this is a good idea. We currently have a tool by user:diberri [[5]]. Would be happy to see it applied universally. Would be great to have a bot that reformed pages using this. Consistency of formatting is important. Makes the WP look more professional IMO.Doc James (talk · contribs · email) 08:35, 21 January 2010 (UTC)

Comment by Slrubenstein

I am basically with Ling nut, Manaus, and DGG. Different academic disciplines and professions do not have different syles because they wish to be obscure - they generally develped styles that suited the kinds of writing they produced. I think there are two important points we need to bear in mind - I am not saying these are the most important points, because others have already provided so many other good reasons not to change policy. First, let's remember that what distinguishes Wikipedia is that it is the encyclopedia anyone can edit at any time. Our cittions arenot just for readers, they are for other editors, or future editors. And if someone wishes to make a real contribution to an article, they will have to do real research on the topic, and then read works by specialists in the field - who use their discipline's MOS. Let me put it another way. Why would anyone care about the system for citation/referencing? Someone who wishes merely to read an article can safely ignore all the citations. One would look at them only if one wishes to read further. And if one wants to read further on a topic, they better get used to how real researchers on that topic write and cite. Second, perhaps just an elaboration of the first: I do not think that different citation styles is the real problem at Wikipedia. The biggest problem at Wikipedia is people who research articles relying solely on the web i.e. materials that are already available to anyone else who can read Wikiedia. I have seen many people cite only those portions of books that are available through google books! An article that invites any reader also to be an editor is also telling readers that they can e more, more than just passive consumers of information. But this does invite them to do real research, and this means they should be comfortable with different citations styles. Anone with a high school degree learns a few of the most common ones, and I simply do not believe that it is beyond editors to learn others.

I see only two problem areas: where there is no one established style in a discipline (anthropology has four journals that can claim reasonably to be flagship, and they vary slightly in style), and where an article draws on research in different disciplines with different standards. Here, as with any conflict, I think the rule of thumb is the same: active editors, acting in good faith, should work out the best conventions for that article on their own. Once it is established it should be changed only if someone can come up with a compelling argument. Really, if a few editors working on an article cannot compromise on this, how on earth will they compromise on important things? I think this is a test most editors will be able to pass. Slrubenstein | Talk 21:19, 23 January 2010 (UTC)

This is basically my main point. Even I use multiple citation schemes depending on the articles I edit; some I simply use short (Author, page number) with a separate section for references (ex. Edward Drinker Cope), other times I just use a single references field for less scholarly and book-oriented subjects (Super Columbine Massacre). You can't enforce a standard style because there's no one good one. Der Wohltemperierte Fuchs^(talk) 13:16, 25 January 2010 (UTC)

Comment by N2e

I support the general thrust of the the original proposer (Cardinal) for the reasons articulated, pretty well I thought, in the proposal. The existing system is complex, difficult to learn for the novice or even moderately experienced content editor, and widely ignored by many (most) editors in any case (in that assertions are routinely added to WP articles with no sources whatsoever). As discussed in this lengthening thread by several others, I don't favor a mandatory single style, but rather the development of a preferred WP standard style that would give guidance to the vast majority of editors who are looking for some guidance and would often choose to "go with it" if such guidance were offered. I think it would be sufficient to articulate a preferred standard style, while deprecating but not forbidding, other styles. I agree with PhilipR's suggestion that we think in terms of a preferred standard that separates citation data from citation style; with the citation-related data clear, a bot could always handle style, or even, discipline-specific citation styles if there were ever to be a consensus on such a thing. N2e (talk) 01:09, 25 January 2010 (UTC)

Comment by Masem

Citation style needs to be standardized within an article. There probably should be consistency within a field (eg under specific Wikiprojects) But, I would be hard pressed to support a single common style across all of WP, as enumerated by others above. I do support identifying what are acceptable styles, and make sure templates to help generate those are well known, so that somoene doesn't try to introduce their own personal system that really doesn't work well.

But my core point is here: has anyone ever been confused about the nature of a reference due to its style that has been able to prevent them from finding this reference or at least knowing where to look for it? This presumes the citation has all the required information presented, but just not in a common form. I know from my academic days that I encountered probably a good dozen different styles and never had a problem locating the original material. It is important to have some consistency for professionalism, but because the nature and type of sources used for different articles in different fields change, some styles work better in those areas. --MASEM (t) 14:34, 25 January 2010 (UTC)

I have had problems in the past with the natural sciences habit to haphazardly abbreviate publication titles with greatly inconsistent abbreviations (despite the existence of standardized indices), but that is a really specific case since said references are usually to fairly old material and are reduced to the very barest possible form. Sometimes I see a monograph series cited as if it was a journal (or the other way around, a journal volume cited as if it were a monograph). Circéus (talk) 14:43, 25 January 2010 (UTC)

Comments by Apoc2400

On one hand, Wikipedia should be written for a general audience. Various academic fields can afford to have their own citation styles because they can expect their readers to be academics in the field, and academics can be expected to know the citation style of the field. An encyclopedia should be readable to non-experts, so we can not expect our readers to be familiar with various citation styles.

On the other hand, most readers do not care about citation style as long as it is readable. I do not think we need a consistent style, as long as all styles used are readable without actually knowing the style. Our citations should be unambiguous and somewhat verbose, writing out the whole name of journals and such.

We should all remember that people write Wikipedia articles in very different ways, and we often have problem und erstanding those who work in different ways from ourselves. Some write articles from scratch on their own, while some add smaller parts here and there. Some use books sources that support a large part of an ariticle, while others use various sources that each support just one or two sentences. None of these methods are wrong, they are simply personal preferences, or depend on what kind of subject the article is about. Some have their favorite citation style and can quickly type it out manually. Having to conform to a rigid Wikipedia house style would be cumbesome. Many other do not fully know any style or do not care. From them, using a citation template allows them to not worry about punctuation, italics and such.

In general, I think we should let each article author use the citation style they are familiar with and suits the kind of sources they use, as long as it is readable and convenient to readers. Others can improve the citations, such as expanding bare URLs, but should not change it from a style the author is familiar with. It is important to remember that a citation style that fits one kind of article may not fit an other kind. There are unclear cases. For example, making in-text citations into clickable links to the full citations is convenient to readers, but may require more complex wikicode. Article authors may sometimes have to accept that their citation style should be adapted for the web medium. --Apoc2400 (talk) 15:26, 26 January 2010 (UTC)

Comment by Semitransgenic

I would like to see WP:REFGROUP introduced as standard across all articles. Parenthetical referencing is what I would like to see the least of, for example:

The title of the work “refers both to contacts between instrumental and electronic sound groups and to contacts between self-sufficient, strongly characterized moments. In the case of four-channel loudspeaker reproduction, it also refers to contacts between various forms of spatial movement” (Stockhausen 1964, 105). The composition exists in two forms: (1) for electronic sounds alone, designated "Nr. 12" in the composer's catalog of works, and (2) for electronic sounds, piano, and percussion, designated "Nr. 12½" (Frisius 2008, 132; Heikinheimo 1972, 115; Stockhausen 1964, 104; Stockhausen 1971, 384). A further, theatrical work, Originale (Nr. 12⅔), composed in 1961, incorporates all of the second version of Kontakte (Stockhausen 1964, 107).

The sentence in bold shows what happens when multiple sources are referenced at one location, it's unacceptably cluttered for general reading purposes. Relatively, this is not such an issue with the in-line equivalent. Overall, hypertext based citations are simply more efficient. On a separate issue, relating to other cite styles, yes, new users face an uphill struggle when presented with an edit box that intermingles code and text, showing the code in a different colour would arguably make the task easier, even for experienced editors. Semitransgenic (talk) 18:13, 28 January 2010 (UTC)

When I edited this sentence yesterday, in response to Semitransgenic's demand for heavier source referencing in that article, I somehow thought this kind of remark would show up somewhere. These several separate sources could of course be tagged to the specific information in the sentence, but as the sentence conflates data from several sources, it would be unbearably complicated—something like this: The composition exists in two forms (Frisius 2008, 132; Heikinheimo 1972, 115): (1) for electronic sounds alone (Frisius 2008, 132; Heikinheimo 1972, 115), designated "Nr. 12" in the composer's catalog of works (Stockhausen 1971, 384), and (2) for electronic sounds, piano, and percussion (Frisius 2008, 132; Heikinheimo 1972, 115), designated "Nr. 12½" (Stockhausen 1971, 384). This is the reason that standards of annotation in the print medium, as well as Wikipedia guidelines, strongly encourage collection of citations at the end of a sentence or paragraph. In fact, the last three sentences of this paragraph could collect all of the references into one group at the end (the first sentence, containing a direct quote, demands a separate citation for the source). For Semitransgenic's information, there is no "inline equivalent" for these references, which are inline citations. I'm not sure what is meant by "code and text" in this context—I see only text, no markup code. Perhaps Semitransgenic can elaborate on this? I do understand that Semitransgenic does not like parenthetical referencing, which I prefer to the cluttered, ugly, distracting, multicolored appearance of hyperlinked footnote referencing. Do I correctly understand that this is a proposal to prohibit the three other means of adding inline citations presently recommended in the Wikipedia guidelines on how to present citations?—Jerome Kohl (talk) 18:54, 28 January 2010 (UTC)

Everyone standing in opposition to this (and admittedly, a couple in support) seems to have come at it with the mistaken assumption that the proposal is seeking to create new rules which would be enforced in some horrifying manner or other. I've basically given up on it, personally, since it seem that this is to nuanced of a topic for most Wikipedians to wrap their brains around. C'est la vie.
— V = I * R (talk to Ohms law) 19:04, 28 January 2010 (UTC)

In stead of assuming that people who don't agree with you are incapable of understanding your complex and nuance proposal, maybe you should check whether whether the proposal is in fact written in a way that conveys its meaning best.·Maunus·ƛ· 19:15, 28 January 2010 (UTC)

Jerome, sorry, you took the wrong end of the stick here, my comment was general, the example just happened to be from something you recently submitted, was not an attack on your methods, simply served as an example for the objection I have to the parenthetical approach, something we discussed elsewhere. By in-line equivalent I meant using ref tags for each item. The second remark was unrelated to parenthetical referencing, it was a general observation about edit boxes. Semitransgenic (talk) 19:36, 28 January 2010 (UTC)

Furthermore, a dual viewing mode for articles, activated with just a single click, would solve the clutter issue, irrespective of the citation style used: one for refs on, the other, refs off.Semitransgenic (talk) 19:41, 28 January 2010 (UTC)

Hmm. That sounds interesting. Do you mean that the reader would be able to choose whether to view the article either with footnotes or parenthetical refs? Or just that the footnotes/parenthetical references could be suppressed? Even if only just the last, this would address my annoyance at intrusive blue bracketed numerals, and would therefore represent a positive step.—Jerome Kohl (talk) 19:57, 28 January 2010 (UTC)

second scenario, suppression of refs, the article could be viewed as unreferenced prose with the click of a button, but this does imply that the refs, whatever form they take, would have to be wrapped in code, even parenthetical refs. Semitransgenic (talk) 20:15, 28 January 2010 (UTC)

The Usability Initiative is making headway in this, FYI. --Izno (talk) 22:04, 28 January 2010 (UTC)

Demo of specific proposal

Following from the specific proposal above, I have implemented a demo of the enhancements I proposed for Cite.php. These changes allow for the editors of an article to specify the style in which they wish reference to appear in - or for a user to override that preference with their own preferred style.

The essence of the enhancements combines the <ref> tag with citation data like so:

<ref name="murphy" first="John" last="Murphy" title="Quisque Quis Orci Magna" publisher="Printing House Press" location="London" year="2009" />

Citing in this manner does not prevent users from citing as they do now or from using the current batch of citaiton templates. Indeed adding a template= attribute would pass the citation data to a template like we use now along with the article's/user's preferred citation sytle.

The proposed system is greatly faster than the current batch of citation templates. To the eye it cannot be distinguished in terms of load times from from "plain text" references. (A speed test is available on the demo page.) The templates it uses can be edited without the intervention of developers through a page in the MediaWiki namespace.

The demo page is here. The demo pages are protected but feel free to create pages to test the system out. --rannṗáirtí anaiṫnid (coṁrá) 20:43, 28 January 2010 (UTC)

Thanks, this looks quite promising. I very much like the improved performance. Some quick questions:

• I assume that named refs continue to work with this new approach?
• As one can see by looking at the source for {{vcite journal}} or {{cite journal}}/{{citation/core}}, there are a lot of options for citing journals. How would these be encoded into the stylesheets illustrated in MediaWiki:Citation styles?
• Could you add a speed test for {{vcite journal}} as well?
• Is there a simple automatable way to convert citations from the old to the new format? For example, the Diberri tool currently outputs "{{cite journal |author=Tabari M, Salehi A |title=Long-term impact of municipal sewage irrigation on treated soil and black locust trees in a semi-arid suburban area of Iran |journal=J Environ Sci (China) |volume=21 |issue=10 |pages=1438–45 |year=2009 |pmid=20000000 |doi= |url=}}"; how can I use the new approach with this? For example, would it be easy to write a new template that would take the same arguments as {{cite journal}} but use this new approach?

Eubulides (talk) 21:21, 28 January 2010 (UTC)

1. The new approach would build on current approaches and wouldn't prevent anything that is currently possible.
2. The templates used in the demo are for "simple" citations. All of those other paramenters could be added. I simply kept it simple for the demo. Adding new parameters would have little or no effect on the improved performance. If you look at the stylesheet you link to you will see tokens like {{§publisher}} seperated by "¦". The values correspond to the attributes passed in the <ref> tag (e.g. <ref publisher=...>). If the attribute is present the token will be replaced with it. If it is not then the whatever appears between the two borken bars will be left out. The m:Help:ParserFunctions can be used to do more complicated things. For example there is a section that reads ¦{{#if:{{§last7}}| et al|}}¦. This means that if there is a 7th instance of "last" (which refers the the authors last name) then display "et al.". There is no tokens for {{§last8}} so the eight and subsequent author are simply ignored.
3. I can do, if you want, but I won't be able to do it tonight.
4. That would be quite trivial. Ideally the attributes that the proposed system takes should be the same as the attributes that the current system takes. That way we could (potentially) convert {{ cite journal | attribute=value }} to <ref attribute="value" />. A bot could do this if the community wanted it and tools like you mention could be updated to give one or both. --rannṗáirtí anaiṫnid (coṁrá) 21:42, 28 January 2010 (UTC)

Re number 4: I'm sure such a tool could be developed after the fact if necessary. That said, something like the 'author' parameter would be difficult to deal with, as the idea of the new system (to me) is to preserve the entirety of the data (rather than abbreviating names) while displaying the citation style of choice. It seems to me that it would be possible to take T:Cite journal and the other various cite templates (which to me are effectively Wikipedia's chosen style, in that they are the most standard style developed by Wikipedia rather than another style guide), and set up our own style within the new <ref> scheme ("style='wp'"). --Izno (talk) 22:00, 28 January 2010 (UTC)

• I was initially pessimistic about this discussion, but this has a lot of potential. Would there be a way of setting a "default" referencing style. For instance, if I am citing three urls and a book, I would be interested in doing the following:

{{reflist|style=citeweb|refs= <ref name=site1 url= title=...> <ref name=site2 url= title=...> <ref name=book1 style=citebook title= isbn=...> <ref name=site3 url= title=...> }}

I'm basically suggesting the same thing as Izno, but think that explicitly creating a "wikipedia" style would be met with opposition. WFCforLife (talk) 01:00, 30 January 2010 (UTC)

Yes, the proposed "style" attribute to the <reference> tag would effectively enable that. (The style is would use would be defined somewhere other that {{cite web}} but could be defined to mirror it. They would also load quicker than {{cite web}} does at present.) --rannṗáirtí anaiṫnid (coṁrá) 02:32, 30 January 2010 (UTC)

Strongest possible oppose to any such solution. We've already got people who are ADAMANTLY refusing anything to do with the citation templates. This just shift the problem to a more hardware, harder to "fix" issue and is guaranteed, is made mandatory to drive away many of these contributors. And if it's not mandatory, then it's just a large-scale loss of time and energy. This entire discussion has already represented way too much of a useless loss of energy. Circéus (talk) 21:03, 1 February 2010 (UTC)

Nobody is proposing to make this mandatory. And if it's voluntary, I don't see the problem. Perhaps some editors think it's a waste of time, but if others don't, and these other editors build a system that works well and is an overall improvement to Wikipedia, then that should be encouraged even if not all editors wish to use the new system. Eubulides (talk) 21:58, 1 February 2010 (UTC)

Further, this allows us to move away from templates, as [one of] the current problem(s) with citation templates is that they are very slow to load. This solution eliminates (or at least decreases its severity) that problem. It also allows us to show explicit data correlations within the text, as well as allowing us to ensure that all the references in an article are held to the same style guide. And, as Eubulides said, this won't be forced on anyone. It's a technical solution to a technical problem, which was generated as a side product of this discussion. Dare I say that the entire discussion hasn't been for naught, despite your contradictory claim? --Izno (talk) 23:15, 1 February 2010 (UTC)

Ebulides wrote "nobody is proposing to make this mandatory." I disagree. John Cardinal's introduction to this page talks about establishing a standard citation style, and talks about all the simplification that would follow; less complicated guidelines, fewer templates, etc. But none of those simplifications could happen unless "standard" means "mandatory".

I am not advocating a mandatory style; I am just stating that I believe this page is proposing a mandatory style. --Jc3s5h (talk) 23:34, 1 February 2010 (UTC)

"mandatory" seems to be a huge sticking point for many... including myself, if you think about it in the way that those who are really worried about it seem to be thinking of it. If "mandatory" means that you're "breaking the rules", then that's obviously bad. People should be able to add citations however they please. I don't see any benefit to trying to require people to add fully formed and properly formatted references to a page, since such rules it won't work anyway and they run against the grain of Wikipedia being a wiki, meaning things are easily changed and we can collectively work towards improvements incrementally. The way that I see this effort is to create a "best practice" recommendation to editors. So, you add references however you please, using whatever system you're comfortable with. Later on, others come along and format those references using our house formatting. No muss, no fuss, and nothing to argue over since nobody added references "incorrectly", and the reference formatting was changed to an acceptable standard formatting. It's certainly much better then pointless bickering over the use of templates and/or some kludge to use an outside format that happens to be some random editors favorite/most familiar style. If the "best practice" is to use an enhancement to the Cite.php extension, and such becomes a "requirement" for GA/FA, then so be it. That's not likely to affect the vast majority of editors who add original (or at least new) content, and it gives people a goal to edit towards.
— V = I * R (Talk • Contribs) 00:05, 2 February 2010 (UTC)

Agree. "Mandatory" is not good (or feasible). A single house style may have been where this discussion began but I think it has moved on from there. --rannṗáirtí anaiṫnid (coṁrá) 00:17, 2 February 2010 (UTC)

There's always hope...
— V = I * R (Talk • Contribs) 00:18, 2 February 2010 (UTC)

Following up on rannṗáirtí anaiṫnid's comments of 21:42, 28 January 2010 (UTC):

1. How would this new system handle attributes that contain quote marks, for example, the title She said "Yes"?
2. A common problem that I run into with {{vcite web}} etc. is that there is a special case for fields ending in ".", "?", or "!". Normally, fields are terminated with a period, but fields ending in ".", "?", or "!" are supposed to be output as-is, without a period following them. This notion is impossible to express with Mediawiki markup, so currently {{vcite web}} etc. have extra parameters such as |author.= that warn the template not to generate a terminating period. It would be nicer if the user could simply write "|author=John Jones Jr." rather than "|author=John Jones Jr.|author.=". Could this new system support such a thing? I realize I am asking for something that Mediawiki currently doesn't do, but I think this sort of feature will be essential if we really want citation autoformatting to work.
3. Any luck comparing the speed to that of {{vcite journal}}?
4. A user-invokable bot might be nice (I doubt whether a completely-automated one would work). Still, I expect that it'd also be nice to have a template front end to the new approach. For example, perhaps we could modify {{vcite journal}} to generate the new reference format. However, I don't see how this would work when generating citations that are not footnotes, which templates are also used for. How could that be done?

Eubulides (talk) 21:58, 1 February 2010 (UTC)

Very good questions.

1. Good catch. Normally quotation marks are escaped using \" or &quot; but that doesn't seem to work right now. I'll look into why.
2. I haven't implemented anything for such cases but it would be quite simple to catch them and avoid double punctuation marks. The script does something similar at present to prevent citations ending in commas or semi-colons should elements be left out.
3. I've added a test to compare {{vcite journal}} along the same lines as the other tests.
4. That would be possible but not something that I had in mind. At the backend of the changes to the extension is a new tag in the format of <citation ... />. That new tag actually contains the citation data. It gets automatically generated and put inside a the <ref> tag to generate the reference. It is unavailable to the user at present but could be made visible and user invokable. If so, it would do what you want it to.

By the way, I know I promised to implement a parser to convert hand-written citations to a template format. I didn't get back to you on that. Truth is, it proved to difficult. Too many things could go wrong. I hope you understand and don't feel like I let you down. --rannṗáirtí anaiṫnid (coṁrá) 00:07, 2 February 2010 (UTC)

That's OK, I understand how hard it can be to parse text. Thanks for your comments and I am encouraged by the progress so far. Some further thoughts:

1. Surely &quot; would be the way to go with escaping, as the syntax looks like HTML. Following in that line, I assume that only & and " would need escaping. But wait a minute: how would people be able to enter titles like "H<sup>2</sup> in air" and "Egyption manuscripts {{circa|200 BC}}"? Can you show an example of how that would work?
2. Thanks for looking into the trailing-"." etc. problem. That would be a real usability win. For Vancouver style, I'm mostly worried about fields that match the regular expression [.?!]['"]*$ and I don't know whether that regular expression should be hardwired or programmable.
3. The tests seem to show that this new approach is 1.2× slower than by-hand in the first benchmark (1.965 vs 1.584 secs) and 1.3× slower in the second benchmark (2.00 vs 1.531 secs). This is a huge win over the standard citation templates, which are about 6.6× slower than doing things by hand (10.396 secs vs. 1.584 secs), and is also a big (though not so big) win over the newer {{vcite book}} templates which are about 2.7× slower than by-hand citations (4.069 vs 1.531 secs). All these timings are medians of three tests, on your server. My own experience is that switching to {{vcite journal}} etc. speeds up page-load times by a factor of two in well-cited articles, and I'm hoping that the new technology will give us at least another factor of two.
4. It's very useful to generate citations that are not inside <ref> tags, and it's very common to do that in high-quality articles. See, for example, Edgar Allan Poe #References. Any replacement for the standard citation templates would have to support citations like this, so please add this to your to-do list. I'm glad to hear that this is already supported internally, and that all that needs to happen is to make it visible to users.

Eubulides (talk) 22:08, 6 February 2010 (UTC)

Re #4: It appears to me that no further alteration is needed in this regard, though I could be wrong on your intent. --Izno (talk) 23:14, 6 February 2010 (UTC)

Sorry, I don't follow the previous comment. The current demo relies on <ref>, no? So it can't do citations that are not inside <ref> tags. By the way, since my last remark I have improved the performance of {{vcite book}} etc by about 40% in a different benchmark, which suggests that (somewhat optimistically) in the benchmark mentioned above they would now be about 1.9× slower than doing it by hand, not the 2.7× slower cited above, but I have no way to test this directly. Eubulides (talk) 06:33, 8 February 2010 (UTC)

Oh, wow, I had a total disconnect there. Excuse me. --Izno (talk) 06:40, 8 February 2010 (UTC)

It occurs to me that we can have multiple groups of references on a page. Will we need to designate the style multiple times, or one time? Or both? --Izno (talk) 20:24, 10 February 2010 (UTC)

Feature request: User preference for styling citations

The following is copy/pasted from Wikipedia:Village_pump_(technical)#Feature_request:_User_preference_for_styling_citations

Suggestion for user preferences After discussing the deletion of a redundant citation tag, I considered whether or not Wikipedia/Mediawiki software could have a user preference to style citations according to different standards (e.g. MLA, APA, Chicago, etc.) Presently, sources are to be cited using {{Citation}} and similar templates such as {{Cite web}}, filling in a variety of parameters to generate a citation. An example follows:

*{{Citation |editor-last=Christoyannopoulos |editor-first=Alexandre J. M. E. |title=Religious Anarchism: New Perspectives |type=Hardback |edition=1st |date=August 1, 2009 |publisher=[[Cambridge Scholars Publishing]] |isbn=1443811327}}

will generate:

• Christoyannopoulos, Alexandre J. M. E., ed. (August 1, 2009), Religious Anarchism: New Perspectives (Hardback) (1st ed.), Cambridge Scholars Publishing, ISBN 978-1443811323

Every article on Wikipedia already has a "Cite this page" link that leads to a variation of Special:Cite (e.g. this example.) On these instances of Special:Cite, citations are given using the fields:

• Page name
• Author
• Publisher
• Date of last revision
• Date retrieved
• Permanent link
• Primary contributors
• Page Version ID

with the following styles:

• APA style
• The MLA Style Manual
• MHRA Style Guide
• The Chicago Manual of Style
• Council of Science Editors
• Bluebook
• Bluebook: Harvard JOLT style
• American Medical Association
• BibTeX (apparently a machine-readable version)

(Other styles that might be useful: A Manual for Writers of Research Papers, Theses, and Dissertations and ISO 690.)

It seems reasonable to me that Wikipedia/Mediawiki software could generate on-the-fly variations for citation styles in the same way that there is a user preference for dates. The same information is present to all users, but arranged in a way that the user chooses if logged in and with a user-defined setting. Unlike date linking in article namespace, there is no problem of overlinking, as this information is already present in the fields of {{Citation}} (or {{Cite web}}, etc.)

The bonuses to this approach are as follows:

1. No more competing citation templates. There will be no incentive to create or use any citation templates that are tied to a style.
2. Users can see citations in the manner that they prefer. Obviously, this is the goal of any user preference, but it has the added bonus of disincentivizing any bickering over citations, forking of {{Citation}}, or competing standards put forth by WikiProjects (e.g. chemistry style by WikiProject Chemistry), plus...
3. Encourage the use of {{Citation}} rather than text in articles. Presently, if a user wants to create a citation according to a style, he must type that in manually to ensure that it appears the way he wants. If this user option was available, it would encourage the use of {{Citation}} by allowing any user to see any citation how he wants.

Does this seem like a reasonable or desirable suggestion to anyone else? I wasn't sure whether I should post this here, to Meta, or to mediawiki.org, but I figured this would get me the most feedback. —Justin (koavf)❤T☮C☺M☯ 06:43, 5 February 2010 (UTC)

The proposed changes above to the software behind the <ref> tag would enable just that. See the demonstration above. --rannṗáirtí anaiṫnid (coṁrá) 08:53, 5 February 2010 (UTC)

The phrasing of the request implies the use of templates is mandatory; there is no such requirement.

Also, although date preferences are still effective for displaying system timestamps, it is extinct for articles. Jc3s5h (talk) 22:21, 6 February 2010 (UTC)

Not really extinct; you just have to do it in your user javascript. One benefit of the date unlinking is that it used to be that only linked dates got formatted. If you just reformat all' dates via javascript, this is more robust against people forgetting to link some of them. — Carl (CBM · talk) 16:14, 11 February 2010 (UTC)

What's the point? See my comment above. PleaseStand (talk) 23:13, 14 February 2010 (UTC)

Harvnb , cite templates citation template and the default editor.

The following is copy/pasted from Wikipedia:Village_pump_(technical)#Harvnb_,_cite_templates_citation_template_and_the_default_editor.

See:Harvnb discussion

There is a known anomaly whereby {{cite family}} don't automatically add the ref=harv parameter that is needed to correctly write {{harvnb}} referencing. The {{citation}} template automatically provides this. The easyCiteMain button above the editor only provides templates for the former cite family templates. This meaning that any Wiki-editor adding to an existing References/Bibliography article, who presses the obvious button will introduce faulty code.

Additionally the forms that the easyCiteMain button provides- use the field date not the field year which Harvnb requires. (It is a bit more complex than this- but in effect that is what happens.)

There are several easy solutions that come to mind.

• Add ref=harv by default, and switch the date for a year parameter are probably the quickest, though clever parsing could handling the 0.1% of twiddly bits as well.
• An extra field- year and date.
• A YN box Make Harv compatible Y/N
• Providing a {{Citation}} option on the easyCiteMain button.

What I don't want is to be told of a clever fix that I personally can do to my settings. The problem is cleaning up the litter left by good faith editors- the solution needs to be universal. Can this be discussed and passed on to ´The Powers that be.....--ClemRutter (talk) 19:33, 9 February 2010 (UTC)

Should probably wait until the discussions at User talk:SlimVirgin/templates and Wikipedia:Centralized discussion/Wikipedia Citation Style are finished. OrangeDog (τ • ε) 20:16, 9 February 2010 (UTC)

[Comment] Though small in the nature of the global discussions above it is important.

• Legacy references are corrected before they are converted
• It is a fundamental principle of DP, that style should be separated from content- if User:Bugs Bunny wants to use input forms because that is easier for him- that is correct. It is upto the software to allow the reader to decide what format he wants to see the references in. on the High Speed 1 article, part is displayed as an appendix to HMGovt publication- this will happen more if they can select how to display references
• Contributing editors have differing levels of experience and ability- acceptable reference input style should be broad enough to allow each to select one they are comfortable with.
• Naive users should be protected from inputting broken code
• Code is only as good as its documentation. What ever happens, issue locatable and correct man pages and help before loosing the code. Double Kiss principle- Keep it simple ;;;;, then go back and make it simpler.

That said- I will unwatch this page and trust you all to do the right thing, before the weekend if that is OK. --ClemRutter (talk) 10:33, 10 February 2010 (UTC)

The tool may well need changing, but making |ref=harv be the default would break too many pages, because it would cause them to generate invalid HTML that would break internal wikilinks. Please see Template talk:Citation/core #We should never render invalid HTML for details about this problem. Eubulides (talk) 17:19, 10 February 2010 (UTC)

You know, it'd be nice if the |ref=harv would actually produce Harvard references as well. the pseudo-harvard style wrech that it currently produces is just silly.
— V = I * R (Talk • Contribs) 17:47, 10 February 2010 (UTC)

What's the difference between what's displayed and what should be displayed? Is there an authoritative specification for Harvard-style references online? (One of the advantages of {{vcite journal}} etc. is that the Vancouver style does have an authoritative online specification.) Eubulides (talk) 20:17, 10 February 2010 (UTC)

For me (and I suspect for others), it isn't the "Harvard style" that matters, it's being able to see the name of the author directly in the citation, without having to look at a footnote. Any system of formatting the actual references in the references section is fine for this purpose, although it's best if the author and year come first, to make it easier to look up sources in the reference list. — Carl (CBM · talk) 15:30, 11 February 2010 (UTC)

Officially, it is called parenthetical referencing— Harvard is a colloquial name. The article on parenthetical referencing is the only one that I have seen that uses true parenthetical referencing. I think many editors confuse it with shortened footnotes, which seems to be a hybrid developed for Wikipedia. ---— Gadget850 (Ed) ^talk 15:36, 11 February 2010 (UTC)

Exactly. I mean references like in Mathematical logic#History. These are misnamed "Harvard refs" here but the "Harvard" part isn't important. — Carl (CBM · talk) 15:43, 11 February 2010 (UTC)

I've actually been looking into the details of the so-called "Vancouver Style" (it's as much a colloquial name as "Harvard style" is, I dare anybody to find actual ref formatting details in the current URMs!), and trust me, just because there's an "authoritative specification"(for ANY style, be it NLM, AMA, MLA, APA...) doesn't mean much. As far as actual publications go, they'll use whatever house format they want. For starters no non-english journals actually follwos the prohibition on accented letters, and everybody happily abbreviates journals not in Index Medicaus. It took me barely five minutes to find three medical journals using variations on Vancouver when I started my research, and hardly anybody actually uses the actual Vancouver style for newspaper or websites (if you ever try to do it you'll understand why!). Besides, Vancouver includes an explicit variant for year-second format to allow the use of parenthetical refs. Circéus (talk) 16:02, 11 February 2010 (UTC)

For the Vancouver style templates ({{vcite book}}, etc.), I've been using Vancouver style as specified in the following guide:

• Patrias K. Citing medicine: the NLM Style Guide for authors, editors, and publishers. 2nd ed. Wendling DL, technical editor. Bethesda (MD): National Library of Medicine (US); 2007 [updated 2009-10-21; Retrieved 2009-12-06].

I agree that actual publications use house format (though I'm not sure what the "prohibition on accented letters" is about, as Citing medicine allows that). In some respects the Vancouver style for web pages is better than what {{cite web}} uses, because Vancouver (rightly) focuses on nailing down the home page for the website, which is an important part of naming the source. And yes, parenthetical references are allowed with Vancouver style, just as they are in Wikipedia in general. Eubulides (talk) 21:07, 11 February 2010 (UTC)

I've been using the same source, and apparently our understandings of "Ignore diacritics, accents, and special characters in names. This rule ignores some conventions used in non-English languages to simplify rules for English-language publications." As for websites, my beef is with the hardheaded insistence on "shoving" websites in the exact same format as a book (where did you think the "In:" came from?). The result is (amongst others) a near-nonsensical requirements for a place of publication. Trust me, once you've had to format one or two website refs, you'll make damn f*cking sure you'll never have to any other again! Circéus (talk) 00:23, 12 February 2010 (UTC)

OK, thanks, I hadn't noticed that about diacritics. That's a clear error in the NLM spec, and the NLM itself doesn't agree with it (see, for example, the "Schöpf" in the title of its citation in PMID 9669133). What NLM calls "requirements" are not really requirements in practice; for example, it "requires" that any citation to (say) PMID 20069038 has to say "[Internet]" but in practice the NLM itself doesn't do that. So I agree with you that the "requirements" are near-nonsensical, but apparently that's just the NLM's style: they commonly list something as a "requirement" but then in later text say that the item is actually optional. Admittedly they could do better. Eubulides (talk) 02:23, 12 February 2010 (UTC)

The addition of the [Internet] annotation appears to have only been added to the Uniform Requirements in May 2009. It takes some time for these things to work their way through to practice, even when they do make sense. This one may take even longer ;-) LeadSongDog come howl 04:33, 12 February 2010 (UTC)

I should probably admit here that I don't particularly like "shortened"/parenthetical/Harvard style references personally. The shortness of the ref in wikitext is kind of nice, which is what I suspect that most advocates like about it, but it produces big ugly references in the actual text (what's shorter after all: "(SomeName p. ###}" or "^[1]"?), but more importantly those references are very ambiguous (without looking, tell me what "(Katz 1998, p. 774)" represents). All of that being said, they could still be used in a way by utilizing regular reference tags along with the group feature. Unfortunately, those who advocate for using shortened references don't seem willing to work with those who don't, so trying to improve references just leads to silly edit wars.
— V = I * R (Talk • Contribs) 16:06, 11 February 2010 (UTC)

Without looking at it I can tell you Katz 1998 is a mathematical history textbook. That's the point of these: once you read one or two articles in an area, you get familiar with the references, and then you can recognize them when you see them again. This system also makes it easier to tell what sort of source is being provided. If I say "Goedel's incompleteness proof (Goedel 1931) ...", it's clearly a ref to the original paper. If I say "Goedel's incompleteness proof (Smith 2004)", it's clearly a ref to a modern text. No need to flip back and forth to the footnotes. I have no objection to using some sort of automated linking system if the notes that it generates inside the text include the authors name (or some abbreviation of it) and the year. But if they just generate [1], [2], [3], then I really cannot tell what the reference is without clicking. Are you saying I can get the same output that is currently there using <ref> somehow? — Carl (CBM · talk) 16:11, 11 February 2010 (UTC)

Another way to put Ohms law's last sentence: Those who use shortened references don't like it when someone comes along to change it without actually improving an article, so trying to change the style leads to silly edit wars. Ucucha 16:14, 11 February 2010 (UTC)

You've identified the fundamental flaw with the supposed benefit yourself though: "once you read one or two articles in an area". Obviously, once your familiar with any subject area the details aren't going to be as important. Consider the fact that our audience is supposed to be "general readers" though, whom we should assume don't have any familiarity with the subject matter, and this supposed benefit is actually it's largest problem. Anyway, @Ucucha's comment/point, I just want to say that I'm attempting to convey some frustration here. This is the best conversation that I've yet to be involved in on the subject, because normally I'm simply reverted and called names for trying to make improvements. I personally don't even try to change the style of the references really, I just try to use the group feature in cite.php to direct those who would click on the reference link to more detailed information on the reference. At least we're having an actual discussion about it, here.
— V = I * R (Talk • Contribs) 19:22, 11 February 2010 (UTC)

There's no reason to assume the readers are not familiar with the area; we just don't want to require them to be, and parenthetical references are no difficulty for someone not familiar with the area. But parenthetical references have the added advantage that people who are familiar can recognize them. And they make it obvious whether a reference is just to the author of a paper:

According to Smith (1996), ...

or is to a separate source:

According to Smith, ... (Jones 1998).

No familiarity with the field is needed to tell which of those two sentences references the original paper and which does not. But when footnotes are used, they are usually stuck at the end of the sentence instead of being placed in context, and so it is impossible to distinguish these situations without clicking. — Carl (CBM · talk) 02:32, 12 February 2010 (UTC)

what I'm trying to get at though is that we can have the best of both here, with proper use of the tools which are at our disposal. This is kind of tough to discuss if we're stuck simply describing the issue to each other though, so let's use an example. I brought up Enriquillo-Plantain Garden fault zone as an example before, so let's stick with that. Granted, that article doesn't need to use such complex cites, but it was a convenient example when I brought it up before. The actual inline format could be adjusted as well, but that's what we could discuss further here, if you're so inclined.
— V = I * R (Talk • Contribs) 02:48, 12 February 2010 (UTC)

(←) That's interesting. With a little bit of dev work that could actually be a replacement for the current mishmash of templates for parenthetical referencing. It already supports spaces in the group name, so I had no trouble changing the group name to "Chen 2003". That becomes more important when there is more than one reference by the same author, or when a single author has several well-known works.

The only thing that would stop my from implementing that on my own pages is that the note number looks a lot like a page number. In particular, once you put a year into the group name, the inline notes look like [Chen 2003 1], [Chen 2003 2], etc., which looks like a page number and also just looks odd. Some sort of divider between the year and the ref number would be nice. For example, [Chen 2003 n1] and [Chen 2003 n2] would be fine, or [Chen 2003, n1].

Now we don't want to insert the n in every case; I need to think about the best way to change Cite.php to accomplish this. But I do know enough PHP that I should be able to change the code and submit a diff for the developers to review, if there is agreement about what we want the code to do. — Carl (CBM · talk) 13:19, 12 February 2010 (UTC)

Also, there needs to be a way to link to the reference without specifying a page number, if the reference is being used in general. For example, if the article says "Smith published a paper in 2003", it is typical to give a reference to that paper which will not include any page number. — Carl (CBM · talk) 13:29, 12 February 2010 (UTC)

Creating inline references that can be interpreted by a computer well enough to link to the corresponding reference list entry will differ from at least one form of parenthetical referencing, APA Style. Looking at p. 174–5 of the 6th edition, we see that a satisfactory inline citation for a paper by Kessler in 2003 could be "In 2003, Kessler's study" or "Kessler (2003) found" or "major depression (Kessler, 2003)." Also, the number of authors to include in the inline cite varies; the first cite to a work by 3 to 5 authors includes all of them, but only the first author followed by "et al." is used in subsequent cites. More would be cited if needed to distinguish works involving some of the same authors.

A consequence is that it might be necessary to go through an article and change some of the existing citations if a new source is added. If an article read "It works great (Able, Baker & Charlie, 2010)...It received award X (Able et al., 2010)" were changed to read "It was initially thought to work great (Able, Baker & Charlie, 2001) but within a few months proved unreliable (Able, Baker & Dog, 2010)...It received award X (Able et. al.,Baker & Charlie, 2010)" the change indicated by strikeouts and underlines would have to be made to an existing inline cite. Jc3s5h (talk) 14:26, 12 February 2010 (UTC)

Right. All three of these are in Mathematical logic#History, too:

1. Mathematical logic emerged in the mid-19th century as a subfield of mathematics independent of the traditional study of logic (Ferreirós 2001, p. 443).
2. Hilbert (1899) developed a complete set of axioms for geometry ...
3. From 1890 to 1905, Ernst Schröder published Vorlesungen über die Algebra der Logik in three volumes.

The third of those is pretty mush impossible to deal with via links, but the first two do OK (just linking the text between the parentheses). We may not be able to get exact APA style, or exact Harvard style, etc.; that may be some people's goal, but it isn't mine. The "et al" issue can be handled by simply choosing the right text for the link; it's not an insurmountable problem.

So when I think about what would be needed for me (personally) to use Cite.php instead of the current templates, the main thing is the numbering of the notes. This could probably be done at the same time that the pages= parameter is implemented, because the output needed for that is very similar. — Carl (CBM · talk) 16:28, 12 February 2010 (UTC)

Any style directly supported by cite.php has a special, preferred position in Wikipedia. If it isn't exactly APA, and it isn't exactly Chicago, or any other published guide, then what is it? If you are going to create a new style, the first thing to do is write the new style manual, and then write the computer code to implement it. Oh by the way, it must be possible to cite anything, because the style software can't be allowed to exclude any reliable source. Jc3s5h (talk) 18:55, 12 February 2010 (UTC)

We're not talking about generating the citations automatically here; we're talking about making the actual text of the footnote say "[Smith 2000 n1]" which vaguely resembles parenthetical referencing (e.g. [6]). The text of footnote itself would still be formatted manually or with a citation template. Presumably articles that don't use parenthetical referencing would not use this feature, and so their footnote text would say "[1]", "[2]", etc. I don't see that this makes it impossible to cite any particular reference. — Carl (CBM · talk) 20:08, 12 February 2010 (UTC)

After looking more closely at the article Enriquillo-Plantain Garden fault zone I see that each source was given a group name, and each inline cite to the article was a reference within that group. I think that would become messy in cases where many pages within a work were cited at different points in the article. It is also inconvenient for the reader, because if I'm holding the source in my hand and reading the Wiki article, I can't tell which page is being cited by looking at the body of the article, I have to go view the citation to figure out which page is being cited, and then back to the body of the article.

It looks like you are thinking of a two part group name, separated by a comma. The first part would be the group name, and the second part would be which cite this is to the group. The reader would see the group name and which cite it is, so it is a style, for which a manual should exist. Jc3s5h (talk) 20:53, 12 February 2010 (UTC)

Cite.php could certainly be updated to add better support for this sort of thing. That's something that we could, and probably should, hash out and take to Dragons Flight and/or any other developer with familiarity and access to Cite.php. I agree that the current setup is imperfect, but it's better then using plain text cludges, which incidentally also do not correctly follow any style guidelines. Also, as you can see CBM, I'm not the only person to criticize parenthetical referencing for being "messy" and/or inconvenient (incidentally though, I've been slammed more for using shorter group names with ref tags then for using the authors name, I assume because using the authors name more closely matches parenthetical style referencing).
— V = I * R (Talk • Contribs) 21:29, 12 February 2010 (UTC)

I'm sorry, but I'm going to exit this conversation now. I don't think it's going anywhere, and I have other coding to do. I am perfectly happy not using Cite.php in my articles, so I don't really have a stake to pursue here. Good luck sorting it out. — Carl (CBM · talk) 04:18, 15 February 2010 (UTC)

Oppose rhetoric

(undent) One day I'm gonna be off-wiki for a couple weeks, come back, and some self-anointed wiki-savior will have persuaded a small group of editors in a small forum to adopt One Format to Rule Them All^TM, and then it will become "Consensus". That day will be one more day that I will really have to wonder whether this project is <rant deleted> and whether I should <more deleted>. • Ling.Nut 14:29, 14 February 2010 (UTC)

That seems not only complete rhetoric, but also not-relevant to this particular thread. If you would like to continue to contribute, please do so, but do so constructively. --Izno (talk) 17:39, 14 February 2010 (UTC)

Hi Izno. Thanks for sharing. This is a page about restricting ourselves to one and only one cite format. Folks above are slaggng the parenthetical style that is employed across a tremendous swath of academic journals, disciplines etc. I trust you will now be able to connect the dots...? • Ling.Nut 03:27, 15 February 2010 (UTC)

The purpose of this page has evolved since it was originally started, obviously. If you'd like to rationally discuss parenthetical referencing then I'd personally be willing to do so, but I don't see what ranting about your personal preferences is going to accomplish.
— V = IR (Talk • Contribs) 03:49, 15 February 2010 (UTC)

I haven't ranted. And "personal preferences" is an interesting term for a family of formats employed by thousands upon tens of thousands of articles, scholars, journals, etc. I'm leaving this forum too, but with this last word: please DO NOT agree among yourselves regarding some stupid one-size-fits-all plan, and then try to act like you've attained "Consensus." The idea of having one and only one reference style is... there are no polite words for the presumptuousness. • Ling.Nut 05:29, 15 February 2010 (UTC)

Problems with citation templates

• Please see here for examples of articles made difficult to edit because of in-text clutter from citation templates—click on "show" to see what the article looks like in edit mode.

1. They're delivered in attractive packages with lots of parameters that are easy to fill in. (See the icons at the top in edit mode—click on the far-right icon to open the templates; or see this image of an article with an open blank template). There are 19 parameters for books. Editors often fill these in whether the information is needed or not, which adds to the bloat between ref tags.
2. Because they seem easy to use, they're added thoughtlessly, so that the number of in-text citations increases unnecessarily.
3. When added in-text, because of the extra characters in the templates themselves, and because of the parameters they encourage editors to fill, they significantly increase in-text clutter in edit mode. This makes articles difficult to edit, and close to impossible to edit for flow (click on "show" here for examples). As a result, the more citation templates there are in an article, the worse the writing tends to be. Articles become lists of sentences, rather than flowing narratives.
4. They slow down load time, sometimes considerably.
5. Some of them use citation styles that don't exist outside Wikipedia; see here for an example.
6. Citation styles in articles that use them are subject to the whim of whoever is editing the templates. All the styles across these articles can be changed without the authors even noticing.
7. Some or all of them use date formats that are hard to decipher and frequently misinterpreted.
8. When they're added to a References section, every entry in the list starts with "c", making it harder to scan the list in edit mode to see where to add a new citation in alphabetical order.
9. If you spot an error in a citation in read mode, it's hard to fix it in edit mode.
10. It's just as fast, probably faster, to write citations out manually once you've decided which style to use. The templates provide only an illusion of speed and ease.

Community feedback on the issue

• I strongly support moves to discourage the use of citation templates. For the life of me, I've never seen the point. How many times have I virtually given up copy-editing an article when 3/4 of the paragraph comprises template fluff? I feel for newbies and visitors, and those who are befuddled by elaborate superstructures without apparent advantage. WP has moved on since 2006, when these artefacts were introduced; we need to go plain and simple for everyone's sake. Editors need to be given back the easy control over the content and formatting of templates that comes with plain text. Tony (talk) 10:38, 25 January 2010 (UTC)
• 1 & 2 seem like non-problems. Could we have some evidence for 5?; and that's not really a template-specific problem anyway. With 7, my understanding is that many of the templates simply use the date however it is formatted by the user. 3 is a problem with templates and the lack of WYSIWYG editing generally, but would be pretty much mitigated by using WP:LDR. 4 is fixable and there is talk of a possible alternative/fix at WP:CENT/Wikipedia Citation Style. I strongly dispute 10 - if editors don't particularly care what citation format to use, they shouldn't be forced to arbitrarily choose one themselves or improperly invent a new one; additionally, they allow editors not familiar with the exacting subtleties of a given citation format to easily enter a citation in that format. 6 isn't likely to be a real problem; a shitstorm would be raised if someone unilaterally changed one of the templates to an entirely different format. Several of--Cybercobra (talk) 12:07, 25 January 2010 (UTC)

• Hi, there's a discussion about an example of 5 (using non-existent citation styles) here. The citation discussed there was:

Bowen, Phil (2008), A Gallery to Play To: The Story of the Mersey Poets, Liverpool: Liverpool University Press, ISBN 1-8463-1125-X

There are styles that use commas in footnotes, but not in a References section with last name, first name. The above seems to be a mixture of styles. This is the template:

*{{Citation | last =Bowen | first =Phil | author-link = | publication-date =2008 | title =A Gallery to Play To: The Story of the Mersey Poets | publication-place =Liverpool | publisher =[[Liverpool University Press]] | isbn =1-8463-1125-X | year =2008 }}

SlimVirgin ^{TALK contribs} 12:19, 25 January 2010 (UTC)

• Yep, that particular template does look broken then. --Cybercobra (talk) 12:28, 25 January 2010 (UTC)
• For example, the references in Annals of Mathematics appear in a single References section at the end of each article and the parts separated by commas, not periods. The Notre Dame Journal of Formal Logic uses a single references section, commas, and puts the last name before the first name. Indeed, the default BibTeX style seems to use commas rather than periods. The reason that {{citation}} puts year and last name first is that it was originally intended to go with parenthetical citations, so having year and name first is helpful to looking up references.
    Moreover: isn't the argument usually made that we don't need to keep Wikipedia style exactly in sync with any external style guide? — Carl (CBM · talk) 12:56, 25 January 2010 (UTC)
  • Indeed. In fact, despite claims to the contrary, AFAIK, very few journals actually follow every single details of any external styleguide qhen it comes to their references. they WILL vary in punctuation, name ordering, use of abbreviations or italics, full name vs. initials and so on. A good, simple example in the medical domain: in practice, no journal actually ever enforces the byzantine Citing Medicine rules for websites, reports and newspapers. Circéus (talk) 01:10, 7 February 2010 (UTC)

• We have a perfectly good compromise about reference formatting: use the style established by the first major contributor to the article. Everyone has an opinion about citation formatting, and so treating it like ENGVAR is the best solution.
    (P.S. I saw this announced on the MOS talk page. I am not sure what the structure of this page is, or why it is in user talk space.) — Carl (CBM · talk) 12:10, 25 January 2010 (UTC)
• I agree with many of the above problems, but I honestly don't think that citation templates are the problem. Instead, I believe that putting the text of a citation - whether it be a template or not - in-line with the prose is the problem. As it is, the data of the citation is within the article, and the citations are called up much later in some sort of citations/references section by another template. So this means that that prose is cluttered with citation data, but the references section boils down to a single template. This seems backwards to me. Instead, the raw-data of citations should be left to the citations/references sections, and in the prose should be some sort of small/simple template which refers to the citations later in the article (something like "<ref name="ABCD">"). In read-mode the article will look exactly the same, but in edit mode it would remove the citation data from the prose. Drewcifer (talk) 12:16, 25 January 2010 (UTC)
  • Yeah, using list-defined references would solve much of the problem imo. --Cybercobra (talk) 12:18, 25 January 2010 (UTC)
  • For an example that doesn't even need LDR, see the featured article Group (mathematics), which just uses abbreviated footnotes. Putting the full citation information into the footnote always makes the edit window look crowded, regardless of whether templates are used. — Carl (CBM · talk) 12:22, 25 January 2010 (UTC)

• The link at the top of this page to examples of articles "made difficult to edit because of in-text clutter from citation templates" show what selected articles look like in edit mode. They don't show how they would compare without citation templates. However, this has previously been done with a real example here - Wikipedia:Citing sources/Example edits for different methods – showing comparative edit mode representations for different citation methods and techniques. In agreement with Carl's point immediately above, in-line full references written freehand can also look cluttered. Of the various approaches shortened footnotes is one good way of reducing such edit clutter as they put full references in their own section, keeping them away from the body text, regardless of whether citation templates are used. --SallyScot (talk) 13:05, 25 January 2010 (UTC)
• I avoid citation templates in the articles I write, and recently explained at (perhaps excessive) length at Talk:Pseudoryzomys#Citation templates why. The templates make it more difficult for new users to add references and to achieve a consistent format (for example, check the language tags here). Ucucha 13:31, 25 January 2010 (UTC)

• I like the templates, as I don't have to waste time working out how to format the thing. Like some previous comments, I do however wish other editors would put the references in the References section, with just a label in the main text. This, of course, has nothing whatever to do with the use or non-use of templates. Jan1naD ^{(talk • contrib)} 13:53, 25 January 2010 (UTC)
  • I agree with this. I have no desire to memorize when I need to have something in quotes, bolded, italicized and where I need a comma, period, etc. I, too, have moved to LDR references since it was introduced and find it much cleaner. However, it does make section editing a bit more difficult.—NMajdan•talk 15:12, 25 January 2010 (UTC)
• Templates provide semantic information that is highly useful. We should not be moving away from making our citations machine-readable. It also helps human editors as well, as the identity of specific information in the citation is no longer dependent solely on its position within plaintext, but can be determined by parameter name in the template. Powers ^T 14:03, 25 January 2010 (UTC)
• FWIW. Some of this is driven specifically by folks attempting to delete content and insisting every sentence and every nuance in a sentence be cited no matter how illogical. Not sure there is an easy answer there but many subjects are not covered by books so a clean reference system to a bibliography seems unlikely. I agree the examples do illustarte how annoying this is and the poor writing is certainly tied to this issue. -- Banjeboi 14:15, 25 January 2010 (UTC)
• On point 1: Last time I checked that button only appear if you have enabled the refToolbar gadget. On 4: There is nothing inherent in templates that make them slow to load. It is mostly because they output a bunch of <span> tags, that is supposedly good for some metadata stuff. I would support removing those <span> tags from the standard rendering of an article. On 5: The {{Citation}} template used commas, the {{Cite XXX}} templates uses periods. I have no idea why {{Citation}} uses commas, and I never use this templates for just this reason. In general, I find the templates very useful for getting a consistent citation style and getting the italic, periods and such right. It lets me not worry about it and assume the template makers have figured out a good style. If they haven't, then the templates can be changed afterwards. They do make the wikicode much harder ot read, but I think this is more the fault of in-text references than of citation templates. I hope either some kind of list-defined references take off, or we get an edit field that collapses long templates automatically. For what it's worth, science publishing uses something similar with BibTeX, but I think the parameters are always filled in at the end, not embedded in text. --Apoc2400 (talk) 14:35, 25 January 2010 (UTC)
• I also like the consistency and ease of use offered by citation templates. They used to be a hassle for me, until I realized I could toggle their appearance on/off with a mouse click in the wiki editor. Also, any new articles I create uses list defined references. Sasata (talk) 15:20, 25 January 2010 (UTC)

• You can toggle them? What editor are you using? --Apoc2400 (talk) 15:34, 25 January 2010 (UTC)

• WikEd (maybe just the dev version, I forget) supports this and I believe it's something the Usability initative are working on. - Jarry1250 ^{[Humorous? Discuss.]} 17:01, 25 January 2010 (UTC)

• Yup, WikEd, regular version. Its the button "Toggle [REF] and [TEMPL] hiding". Sasata (talk) 17:05, 25 January 2010 (UTC)

• I'm looking at that now, but I can't see it. Can you say where it is on the page? SlimVirgin ^{TALK contribs} 17:21, 25 January 2010 (UTC)

• On the top right hand corner there's a box of 10 WikEd control buttons, and the one in question is labelled "R", at the top left. ~~

• Thanks, I found it. The problem is that it's very slow, so if you want to check a ref, you'd be waiting for it switch on and off. The interface seems a bit messy too. SlimVirgin ^{TALK contribs} 18:19, 25 January 2010 (UTC)

• I have written another script to mimic the LDR method for articles with inline refs, but the script is admittedly not complete and misses some of the ref contents (see its limitations). I agree that the ref syntax should be slightly simplified in that regard, and that we should be moving toward an LDR-like system. See my proposal for details. PleaseStand (talk) 05:15, 15 February 2010 (UTC)

• Addressing the points one by one.
  1. Unnecessary info happens even when templates aren't used.
  2. Likewise for thoughtlessly-added citations.
  3. Agree that clutter is a problem. This is ameliorated by putting citations at the end (a relatively new feature). but it's still an issue. I've proposed something in Wikipedia:Centralized discussion/Wikipedia Citation Style #Proposal for restylable Wikitext citations that would fix the clutter problem, but it'll require some MediaWiki changes.
  4. Agree about load time for {{cite journal}}, {{citation}}, etc. I have written new (still experimental) templates ({{vcite journal}}) that reduce load time considerably. In Autism it cut load time by a factor of two, for the whole article. I haven't benchmarked these new templates against doing it by hand, but I expect that the new templates fix performance problems to a great extent.
  5. The experimental templates use Vancouver system style, which is standardized; their documentation cites the standard, feature by feature.
  6. Yes, when templates change, that changes all articles. This is a sharp-edged tool: it makes articles easier to fix, as well as easier to screw up.
  7. The templates I'm aware of leave date-formatting up to the user, so this isn't an issue.
  8. The problem with alphabetization is quite minor.
  9. For most errors (misspelled author names, etc.) fixing errors is not a problem.
  10. I never thought of templates as being easier to write by hand. Obviously they take more time to write out by hand. I typically use the Diberri tool, which generates them automatically. My proposal for restylable Wikitext citations would also address this issue, by making citations easy to write by hand while still allowing for automated processing.

Eubulides (talk) 17:22, 25 January 2010 (UTC)

It's definitely faster to write this, and there's no in-text clutter: <ref>Smith, p. 1.</ref> And also surely faster to write this (because you're not having to fiddle with a template): <ref>Smith, John. ''Book Name''. Routledge, 2010, p. 1.</ref> I can't see the point of using a template, because you have to fill in the parameters anyway, so it's the same amount of typing. SlimVirgin ^{TALK contribs} 18:26, 25 January 2010 (UTC)

There may be several authors, and sometimes I want a link to Google Books. --Apoc2400 (talk) 22:27, 25 January 2010 (UTC)

It took me twenty keystrokes and mouse events to write the following citation, using the Diberri tool. This is much faster than I could have done it by hand. I could do it faster with the tool, if pressed.

Liberati A, Altman DG, Tetzlaff J et al. The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med. 2009;6(7):e1000100. doi:10.1371/journal.pmed.1000100. PMID 19621070. PMC 2707010.

Eubulides (talk) 01:58, 26 January 2010 (UTC)

I hate citation templates with a fiery passion. The limited gain from them is simply not worth it, in my view - even if you ignore the off-putting effect on newbies. Using Wikipedia:Footnotes#List-defined_references helps a bit, shifting the mess out of the body text; but it's still a mess. Exhibit A, for me: Taner Akçam. Urgh. Rd232 ^talk 23:48, 25 January 2010 (UTC)

My responses thus far are:

1. No one has presented a list a specific advantages of templates over plain text. Perhaps there are none.

1. People may be influenced by whether they mostly (i) create articles (or change the citations from plain text to template), or (ii) edit articles after the fact. Does this explain why a few people have discounted the difficulty of correcting errors in references in display mode?
2. It looks as though people are generally in favour of getting rid of the inline clutter. I suggest we need an RfC after this debate to gain consensus to discourage or even ban the use of inline template text in new articles, and to set in motion a program to convert existing articles that use it. IMO, this serious impediment to editing is unnecessary and unacceptable. I go with the by-line "short inline, long at the bottom". That is the norm for academic and administrative text Tony (talk) 23:59, 25 January 2010 (UTC)

Certainly citation templates have some advantages. However, this page is about the templates' problems, not their advantages. I both create and edit articles, and don't have any more problems correcting errors in references with templates than I do with plain wikitext. It's unlikely that any RfC would gain consensus for discouraging or eliminating citation templates, at least, not with the current Wikimedia software. A more likely way to eliminate the templates is to improve the software infrastructure so that the templates are not needed (though this will require some software-writing). Eubulides (talk) 01:58, 26 January 2010 (UTC)

I am avoiding commenting on that because I think the ENGVAR-type solution is the best here. Different people can argue personal preferences all day. Regarding an RFC, see the "see also" link below. — Carl (CBM · talk) 01:15, 26 January 2010 (UTC)

• (Hating them "with a fiery passion" is good! So do I!) I couldn't agree more with Tony. I'm surprised to learn, from CBM, that we have a "perfectly good compromise about reference formatting: use the style established by the first major contributor to the article." I wasn't aware of this principle, and it doesn't tally in the least with my own experience, which is that whenever I put an article on FAC... say one I've written in a sandbox, as I usually do, and in which I've established simple WYSIWYG references of the neat and un-wasteful shortened footnotes type ... then the first "improvement" it receives is having these short references turned into citation templates. :-( I'm also used to receiving FAC or FAR comments implying that templates are the only real, proper citations; and to getting attempts to persuade me to use a ref name instead of a short note, which would mean, with the type of sources I generally use, a thick sprinkling of probably three times as many notes... and, in short, what I encounter is an apparent conviction that my quite consistent notes are unacceptable, whether or not I've "established" them—see the argument at the Harriet Bosse FAC], for instance (where, for once, I insisted, and the original style was in fact kept—but why can that only happen at the price of a battle?) I haven't done any large-scale checking for the purpose of chiming in at this discussion—life is too short—but I've picked a few of these articles at random, and, typically, there is always a silent change from simple pre-existent footnotes into cite templates. That change happens early in the history, often while the article is still on FAC, indeed often immediately after I have begged people to leave the style the way it is. Here's an old one: [7]; a recent one (with the comment that "I've never seen that before" [8]); and here's an old article currently languishing on FAR—a dangerous place, that, where the citations were immediately templatified: [9]. (To save it from being de-featured for lack of templates ?) I haven't reverted the people who make the change; their intentions are certainly good; I don't edit war, especially not over style; and I don't want to spend my prime arguing about footnotes (something I get enough of at my RL job); but what are changes like that for? Do the templates have a purpose? All they do that I can see is make it prohibitively difficult for new users to add sources and pretty darn awkward for technology illiterates like me, too. And, CBM, is the "perfectly good compromise" something you've actually seen in the wild? Because I haven't. What I keep seeing is the predatory templates devouring the mild, doe-eyed WYSIWYG references. :-( Bishonen | talk 00:38, 26 January 2010 (UTC).
  • This sounds like another problem with FA, rather than a problem with citation templates. Both WP:CITE#HOW and WP:CITE#Wikipedia:CITE#Citation_templates_and_tools are clear about respecting the first contributor's choice. So you should just revert unilateral changes with a polite note to RTFM. If this causes problems at FA, it might give one pause to reconsider whether FA is a productive way to spend time. — Carl (CBM · talk) 01:13, 26 January 2010 (UTC)
  • I tend to agree: there are assorted annoying "ass-wipes" and "snivelling snot rags" that hang around (or watch) at FAC insisting that such things are mandated (or otherwise putting the templates in overnight as you sleep), whereas in fact they are optional – consistency being the only thing that really matter as far as citations are concerned. Ohconfucius ^¡digame! 02:34, 26 January 2010 (UTC)
  • Rather than dismissing the FA process altogether, it would be more productive to educate editors who participate in that area. Dabomb87 (talk) 03:33, 26 January 2010 (UTC)
• To paraphrase a technical writing professor I had, I don't remember citation formats; I look them up every time I need to cite something manually. It's a conditioning I have from writing too many papers for professors who wanted too many different citation styles. Thus, whenever I have to cite something on Wiki, I use a template; everything that I might reasonably need to remember to put in is there. Nifboy (talk) 01:55, 26 January 2010 (UTC)
• Personally, I do not bother with templates. I put the information in in whatever format matches the sources I use for it; This is not formal cataloging or bibliography--I can do that also, but i do not do that here. (there are places where it is needed and not used, as in wikitexts, where I , but not the people who work there, consider it essential to know exactly what copy of what version was transcribed.) We want to encourage people here to get in references for articles, and the way to encourage them is not to force them to do it any particular way. With journal articles, every discipline--sometimes every journal has its own style. I consider some of them better than others (I very much dislike the ones that do not give inclusive pagination , full author names, and full article titles, but the one sort of thing that does not matter at all is the order of elements or the punctuation in between them. I'm not even concerned too much about being consistent in an article if it slows up entering the references. DGG ( talk ) 03:34, 26 January 2010 (UTC)
• I wasn't even aware until I hit this page that it was possible to put all the citations at the end. I'll be doing that from now on, after I play with it in the sandbox a little. I think making people aware this is possible would do a lot. It may even be possible to write a bot to do this automatically for existing articles. Ray^Talk 03:42, 26 January 2010 (UTC)
• Template problem #11: Bugs. See User:Art LaPella/Citation template double period bug. Art LaPella (talk) 04:09, 26 January 2010 (UTC)
• I think the current guideline on cite templates (neither encourage nor discourage) is sound. The problem is that many editors are not aware that this is the case, and we need to aggressively discourage the idea that citation templates are required or better than hand-written ones. On the other hand, I think citation templates have a few benefits, and we should not throw the baby out with the bathwater here; moreover, I seriously doubt that editors would unconditionally agree to even discourage citation templates, let alone ban them. This issue is more important and far-reaching than date autoformatting, and we all know how much time it took to change the status quo on that issue.

  Personally, I don't mind citation templates, and use them often, but would not be bothered if I couldn't use them. I think the more successful path is to strongly encourage moderation: don't provide more information than necessary, and prefer citations at the end of paragraphs rather than after ever sentence. List-defined referencing sounds and looks like a good idea, but we shouldn't try to push it as a Wikipedia-wide standard. Dabomb87 (talk) 04:24, 26 January 2010 (UTC)

  • "prefer citations at the end of paragraphs rather than after ever sentence" But then someone reorganizes the article or adds a third source, and you've then lost the sourcing or it's insufficiently clear which sources correspond to which assertions. --Cybercobra (talk) 07:39, 26 January 2010 (UTC)
• I like the uniformity provided by templates, and intend to continue using them. I do however agree that list defined referencing should be encouraged. WFCforLife (talk), Help wikipedia. Make the pledge. 04:29, 26 January 2010 (UTC)
• Well, I've no idea how the culture I described can exist at FAC, in particular; it's not for want of telling. Look at User:SandyGeorgia tersely assuring reviewers that the entirety of the citation style requirement is that the style be consistent, and warning reviewers to stop making nominators "jump through such hoops". That the template hoop, and the other hoops adduced by many editors above, persist on FAC and FAR in spite of voices like Sandy's (of all people) suggests to me that the attitude is not confined to FAC and FAR at all—why should it be? Everything that's a problem on FAC tends rather to be a worse problem elsewhere. *I* think such a culture is born of wikignoming running berserk, and is rampant on the site. Bishonen | talk 06:00, 26 January 2010 (UTC).
  • Actually it's my understanding that FAC concentrates all these problems, partly in the nature of the process, and partly in the people it attracts. I've never gone anywhere near it. Rd232 ^talk 09:31, 26 January 2010 (UTC)

• My understanding of the problem is that

1. People learn by imitation rather than reading the guidelines. They see citation templates everywhere and assume they are required.

1. Some people obsess about enforcing rules and making all articles the same. Have you seen how those infoboxes and navigation boxes pop up in places where they are really not helpful? --Apoc2400 (talk) 11:40, 26 January 2010 (UTC)

Rd232, you mean "that little ass-wipe" and "those little Stalins"? Why thank you. More seriously, nothing is perfect, and we have uncovered on this page that the FA criteria needed to be adjusted to disabuse editors of the notions they may have picked up from the herd that citation templates are mandatory. Yes, it should have been addressed ages ago, but escaped the attention of the crowd. Tony (talk) 12:02, 26 January 2010 (UTC)

Erm, I'm sure you didn't intend to, but to the average reader it may seem like you're quoting me, when in fact I presume you must be quoting others (I never said anything like that). Rd232 ^talk 13:31, 26 January 2010 (UTC)

Incidentally, let's not get overfocussed on FAC, which affects a tiny proportion of articles. Rd232 ^talk 13:33, 26 January 2010 (UTC)

• I agree there is a case for moving the citations to the end of the page (I sometimes find myself writing a paragraph in Word to ensure good text flow and then adding the citations in the spots reserved by little numbers), but I find the citation templates useful. I think they make it easier to create and, perhaps more importantly, edit references without having to remember all the punctuation conventions and element combinations of the citation style in use. Moreover, even those well acquainted with the citation style may occasionally produce a typo or inconsistency, which the template can minimise.

  I understand that there are some problems with their use, but not all of them are specifically related to citation templates. Problems #1–3, for example, seem to reflect a more general trend of "using whatever features are available", combining the logic of "it exists therefore it must be used" with the practice of mimicking others rather than reading Wikipedia's voluminous guide- and style-book. As trite as it may sound, "better education" seems to be the only effective solution other than, well, fewer guidelines. I cannot speak about #4 and #5, but #6 only requires better policing of the templates, #7 is very relevant (the "cite x" templates which I use not only allow one to format the date as one wishes, but their documentation encourages editors to use the date style of the prose), and I consider #8 inconsequential. I have already commented on #9 and #10, though I could add that for both it is largely a matter of experience. #11 is an annoying complication, but far from insurmountable. Waltham, The Duke of 16:11, 26 January 2010 (UTC)

• Yeah, yucky. I never seem to use them, preferring the smallest amount of text that gets things done; that's usually <ref>[http://example.com Example text], date</ref>. I used to feel lazy about that, but now I feel ok. Edward Vielmetti (talk) 20:29, 26 January 2010 (UTC)
• I understand why copyeditors hate bloated citations mid-paragraph, and will change my editing habits accordingly by using list defined references. What I don't understand is how

The quick brown fox jumps over the lazy dog 3 times a day,<ref>[www.thequickbrownfoxjumpsoverthelazydog.com/lalalalal235.html Proof that they do] (12 October 2005). Accessed 27 January 2010. User:WFCforLife digital publishing]</ref> except on Sundays.

Is significantly better than

The quick brown fox jumps over the lazy dog 3 times a day,<ref>{{citeweb|url=www.thequickbrownfoxjumpsoverthelazydog.com/lalalalal235.html|title=Proof that they do|date=12 October 2005|accessdate=27 January 2010|publisher=User:WFCforLife digital publishing}}</ref> except on Sundays.

Forcing editors to care about the manual of style is a bad thing. These templates ensure that the apathetic comply with it, and that changes to the MoS can be applied to hundreds of thousands of articles instantly. WFCforLife (talk) 00:47, 27 January 2010 (UTC)

• Nearly every point outlined above seems to support:
  • Driving editors to use the relatively new feature of pushing the text of references to the reflist parameter, leavng simple <ref name="refA"\> statements in the text body. ( I would suspect that someone could come up with a bot that could automatically do this, if not add it to AWB or other tools).
  • Using standard reference templates to be able to easily (manually or with a bot) identify issues with things like dates or other formatting problems to fix or mark them to be fixed - but also not limiting editors to these as long as they use the same consistent style, which the above point would help towards (since they would be back to back in the same area of wikitext).
• Yes, it would be great if there was a better WYSIWYG editor that hid references so that you can solely focus on the language and structure (quietly moving these as you did this). But we've established that in-line referencing is a must, and standardization within an article (at least) is a must, and dumping the templates seem like both backward steps from these points. It certainly doesn't change the readability during editing aspect of references - they'll still be segments you need to scan over as you work. --MASEM (t) 13:12, 27 January 2010 (UTC)
  • I'd say wikEd is as good at dealing with this as anything: apart from the highlighting of sources in the edit window, its own previewing tool shows you the preview of the article or section without taking you to another page and shows the full in-line citations at the bottom. I've always thought that the best way to avoid mistakes is previewing, and this makes it much easier. Waltham, The Duke of 15:37, 27 January 2010 (UTC)
    • WikEd is a great tool, but also, from my experience using it, buggy and slow. I'll use it near the end of a GA/FA cleanup of an article to help hilight things, but not everyday use, nor would I recommend it for everyday use. And if we're already talking about confusion for newer editors over inline reference text, I'm not seeing introducing WikEd to them a walk in the park either. --MASEM (t) 00:07, 28 January 2010 (UTC)
    • It is certainly true that wikEd is not a solution. It doesn't work on IE, and it has issues with copy/paste. The "beta" Wikipedia editor? Even worse. Any real solution should involve the core MediaWiki software, not an unnecessarily large and buggy user script. PleaseStand (talk) 05:15, 15 February 2010 (UTC)
• There's no great alternative. We're just waiting for a WYSWYG from the usability team. At least I hope we are. I haven't heard them say much about it in a long time. They're just changing fonts, colors, and icons as far as I can tell. - Peregrine Fisher (talk) (contribs) 00:29, 28 January 2010 (UTC)

Separating the issues

There are a lot of different issues being discussed here that need to be separated. Not all of the problems are with the templates themselves, but with the way that they are sometimes (or even often) used.

1. The first substantial contributor's citation style should govern—if there is one. That is a guideline, and therefore should be followed. If the first substantial contributor's citation style matches a template format (regardless of whether the contributor used templates), it is OK to use a matching template. If no template matches the contributor's citation style, then citation templates should not be used in that article (not because they are templates, but because they produce citations in a different format). On the other hand, sometimes there is no first substantial contributor; some articles evolve in bits and pieces. If there is a predominant style and it is adequate (i.e., it provides the required information), it should be followed, template or not. Sometimes, relatively often in fact, the first substantial contributor or group of contributors is not consistent about citation style. Then you are back to square one.
2. Most articles evolve with citation chaos. In these cases, it may be more feasible (politically, if for no other reason), to impose order with templates. You tend to get fewer arguments if you use templates than if you manually edit other editors' citations.
3. Citation templates can help many editors. Personally, I don't care (although if you have an ISBN, other ID number, or URL, the Wikipedia template filling) tool can give you a good head start). But for the substantial majority of editors, you are much more likely to get citations that are both adequate and consistent by using templates than without them. In my opinion, inadequate citations is a much more widespread problem on Wikipedia as a whole (no I'm not talking about FAs) than template clutter. When templates are used, filling too few parameters is a more widespread problem than filling too many, although the latter is also a problem (one of my pet peeves is journal citations that use both the url= and doi= parameters when both point to the same web page).
4. For articles that make multiple references to multiple sources, short inline citations plus a bibliography is usually the best solution. Lots of articles have too many duplicative footnotes, sometimes using templates and sometimes using manual citations (often both in the same article). The problem here is failure to use the inline plus bibliography system, not use of citation templates. One can make a nice, consistently-styled bibliography with citation templates. (In fact, the citation templates are really styled for bibliographies rather than for footnotes, despite the way they are used.)
5. For other articles, good footnotes with full citations often suffices. Especially where an article does not have inline citations to different pages in some sources, a list of short-form footnotes followed by a list of roughly the same number of full citations in a bibliography can look silly. The inline plus bibliography system may still be best here if one expects the article to outgrow a footnote-only system. But, again, this issue has nothing to do with template use.
6. Inline template clutter is avoided by filling templates vertically. Preferably, without padding. In fact, it is easier to edit an article with vertical templates than one with dense, manual, full-citation footnotes. I often convert "horizontal" to templates to vertical when I'm lost in a sea of templated footnotes. (Hint: Even when filling vertically, it is OK to have a few related parameters—e.g., parts of one author's name, parts of a date, location and publisher—on one line.)
7. Some of the editors commenting here have a skewed view of the issue because they spend most of their time with the best articles and best editors. The rest of the encyclopedia isn't so pretty.

On balance, I think the project benefits from having citation templates, and from encouraging those editors who have difficulty with citations to use them.—Finell 04:40, 28 January 2010 (UTC)

• I think people should be able to object to and revert the conversion of manual citations to templates.
• Yes, vertical inline clutter is a lot better than horizontal inline clutter, which is unbearable for the copy-editor. But why do we have to have either? Why not short inline; long at the bottom? Tony (talk) 06:44, 28 January 2010 (UTC)
  • Again, probably worthwhile to mention we can do exactly this with the "reflist" template now, see WP:LDR. Again, a lot of problems solved if we nudges editors in that direction. --MASEM (t) 06:54, 28 January 2010 (UTC)

(edit conflict) I agree with both of Tony's points. There is no reason to convert a correctly-styled manual citation to a template; but if a template helps an editor fix an inadequate citation or one that does not match the article's citation style, I don't have a problem with that. And I said above that I agree that what you call short inline; long at the bottom as a good system; but that is a structural or layout issue, not a template issue. The discussion here is addressing the broader question of whether citation templates are a boon or a bane, with some very good editors (who get along just fine without them) leaning toward bane. I disagree with the broad condemnation, and I disagree with blaming templates for problems that have other causes.—Finell 07:15, 28 January 2010 (UTC)

Templates are still unnecessarily hard to read and edit, especially for newbies and non-geeks, even when placed at the end, WP:LDR-style. (I say that as someone who is quite happy editing HTML.) I'm sceptical too that templates encourage better citation, because of the selection bias: anyone bothering to use one is already making more effort than the average person. And I see no reason why someone imitating a good manual citation is less likely to do it right than if they're copying a template citation; if anything, with the confusing syntax, it's more likely to put them off doing it at all. The only thing citation templates achieve, I think, is enforcing uniformity of style - if users use the templates. But equally you could say manual citation achieves uniformity of style - if users imitate existing manual citations. It's just not worth the price of entry. Rd232 ^talk 09:12, 28 January 2010 (UTC)

The cite templates are not that hard if you are entering just the basic information about something - particularly if there are other examples floating in the article to work from (even moreso when bunched in LDR); the most basic fields are self-explanatory for an editor unaware of how templates work. (And if you are even thinking about becoming a serious WP editor, learning how templates work (most specifically, what transclusion is and how parameters work) is a necessity) At worst, when I see editors try to mimic the citation templates, they have at least left enough information to expand the required details. Of course, the same can be said if all the manual references were grouped in the same manner in LDR (LDR is non-specific if the ref style is manual or not). --MASEM (t) 14:24, 28 January 2010 (UTC)

To take Rd's point further, it's relatively easy to pick out formatting glitches in a citation list, since they're all packed together vertically. The main text already presents considerably greater challenges in keeping formatting (and allowable punctation and other style choices) consistent within the article. That is why I encourage all editors not to use templates.

The point about how difficult it is to correct mistakes in templated citations (No. 9) has been pooh-poohed in the discussion above. But you have to do it in whole-article, not section edit mode, using a finder, then squinting, then editing, then saving, then back to check—forget about previewing, because that will require a repetition of much of this cumbersome process again. Why is this an advantage over manual citations, which are so easy to correct and update? I'd have thought it was the clincher that swayed everyone to either link-defined references (thank you, Masem), or plain, simple and controllable manual.

I think there should be a protocol before introducing templates to an existing article—how many FA nominators have been extremely irritated to wake up and find their manual citations all converted without the politeness of asking? A talk-page request, I believe, is the very least that is required. Tony (talk) 09:29, 28 January 2010 (UTC)

Glancing over the comments, it seems that many or most editors seem to agree that adding citation templates inside the text, as opposed to in a References section, can be problematic. Is that how other people are reading this?

Tony, regarding your post above, citation templates should never be added to an article that has a consistent citation style already, unless there's consensus on talk, per CITE. Where there's disagreement, the style used by the first major contributor should be respected. Perhaps that needs to be emphasized somewhere. SlimVirgin ^{TALK contribs} 09:35, 28 January 2010 (UTC)

• I prefer templates because as a casual editor who doesn't so academic writing outside Wikipedia, I would have to look up the proper citation format in the MOS or get a style manual from the bookshelf to have anything like a consistent reference style. Templates also help me know what to look for so I'm not just using naked URL's as citations but including authors and dates as appropriate. I have never had a big problem with them cluttering up article text, even when placed in-line, but that's probably because I edit almost entirely in poorly and unsourced articles rather than FA quality stuff. Eluchil404 (talk) 11:17, 28 January 2010 (UTC)
  • Consistency of style of citation (across articles that is - and within articles it's easy to imitate existing ones) is somewhere below the bottom of my list of priorities. And you really don't need a template to remind you of what to include in a citation - Wikipedia:Citing sources is (or should be) clear enough if you need a reminder. Rd232 ^talk 12:00, 28 January 2010 (UTC)

• On the concept that the first contributor decides the template style - I'm not against this point, but I feel that if the article clearly belongs within a certain wikiprojct that has opted to use a specific ref style, that should take priority over the first-contributor presuming there's no other question of which style to use. Eg, if a biology-related project has opted to use manual Nature-style referencing and someone creates a new biology-related project with the "cite news" templates, the project preference should take priority. But again, this is only where it is clear that no other style selections could apply, though some consistency should apply. Say in a case where the article may fall into both bio and chemistry fields, each which has opted a different style, the ref style should be one of those two, as selected by the first contributor, but not some new random style. But if no style at all has been assigned, then yes, still up to the first contributor. It would be nice to aim for consistency across related articles, but not at the expense of editing wars over ref style. --MASEM (t) 14:18, 28 January 2010 (UTC)

• Returning to the point about keeping it short at the top, long at the bottom, this diff demonstrates the benefits to writers of getting these ugly templates (and other code) out of their way, while also emphasising why these templates are not a big deal when used appropriately and considerately. WFCforLife (talk) 23:21, 29 January 2010 (UTC)

• Clutter is a pain, but I like the templates. Even if the inline footnote texts are short, and the templates are only used in the References section at the bottom, the footnotes are still annoying. If I am allowed to dream, I would like a button above my edit window that enables me to toggle the display of any <ref>source</ref> or <ref name=name /> content off and on. When it's off, I don't want to see anything there, just a little red symbol like † telling me there is a citation there, so I don't accidentally delete the thing while editing the text. That way we'd be able to see the prose again. JN466 03:03, 31 January 2010 (UTC)

You cannot fight the ref templates. The reason why is that there aren't any good alternatives. You might be able to get a few more hand formatted FAs. That's about it. This page would be a lot more interesting if you started from there (reality), and then discussed what to do about it. The only thing I've heard of, is a possible WYSWYG. - Peregrine Fisher (talk) (contribs) 06:27, 31 January 2010 (UTC)

• Opposing viewpoint - Citation templates are essential to the keeping the citation formats uniform and well-constructed. As Wikipedia grows, we need more consistency, not less. Discouraging editors from using citation templates increases entropy and terrible citing. Please see User:Blargh29/Using citation templates for my thoughts. --Blargh29 (talk) 07:18, 31 January 2010 (UTC)

Well, you two come up with a way to minimise the inline clutter, because this issue ain't gonna go away. Tony (talk) 09:54, 31 January 2010 (UTC)

It's simple to minimize inline clutter: don't put complete reference information in footnotes, and there is no clutter. For example, I started recursively inseparable sets yesterday with parenthetical referencing (e.g. the article text only contains (Monk 1979, p. 111)). Abbreviated footnotes work equally well (e.g. the article text would contain <ref>Monk 1979, p. 111</ref>. I am not familiar with "list defined references" but I assume they would also work.

The cause of the inline clutter is the cluttery addition of full reference data in the middle of the paragraphs. The articles that use that system are just as cluttered when they don't use citation templates. — Carl (CBM · talk) 12:59, 31 January 2010 (UTC)

The templates cause clutter in two ways. First, the templates themselves add characters. For example:

{{cite web |first=Laura |last=Roberts |url=http://www.dailymail.co.uk/tvshowbiz/article-561161/Scheming-Heather-told-Sir-Paul-Marry-says-friend-10-000-night-ex-prostitute.html |title=Scheming Heather told Sir Paul 'Marry me or it's over', says former friend and £10,000-a-night ex-prostitute |publisher=[[The Daily Mail]] |date=2008-04-22 |accessdate=2008-07-17}}

Lots of unnecessary characters there. With one citation, it makes no difference, but multiply those extra characters by the number of citations, often over 100 in well-developed articles, and that's a lot of unnecessary words. Secondly, the presence of the parameters encourages people to fill them in. The above isn't a bad example (though there's no need for an access date for a newspaper), but I've seen absurd examples of people filling in details of the first editions of books, chapter numbers as well as page numbers, and other details that no one needs. SlimVirgin ^{TALK contribs} 13:16, 31 January 2010 (UTC)

If the reference information was not inline, it wouldn't make a difference whether more parameters are filled in, though.

The first edition stuff is actually one of the things we tend to do poorly at. If a book published in 1945 is reprinted by Dover in 2009, we shouldn't cite the book as if it was published in 2009. Unfortunately the existing citation templates make it difficult to deal with this case, and not all editors realize that one has to check the original versions of reprinted books. — Carl (CBM · talk) 13:34, 31 January 2010 (UTC)

Why would we have to check the first edition? SlimVirgin ^{TALK contribs} 13:55, 31 January 2010 (UTC)

For plain facts (the kind that could just be sourced to any textbook), it's less essential. But for direct quotes and personal opinions, knowing the year in which the person made the quote or wrote the opinion is essential to have a understanding of the context in which the quote was made. Careful research includes determining when such quotes and opinions were stated, and citing the year accurately. This is particularly important when a book is reprinted after a long delay. For example, ISBN 1-458-99622-0 is a reprint from 2009 of a book by Henri Poincaré, who died in 1912. Simply citing that book as if it was published in 2009 is sloppy. A careful citation should include both the details of the reprint (because that is what was consulted) and the original publication details of the material that was reprinted. — Carl (CBM · talk) 14:32, 31 January 2010 (UTC)

Yes, good points. SlimVirgin ^{TALK contribs} 14:35, 31 January 2010 (UTC)

• The citation template situation at present is totally unsatisfactory. The requirement for in-line citations imposes a very severe impediment to accessibility - ie the average uninitiated deterred from editing articles; add to that the horrible screen clutter that the templates themselves generate, such that the readability in edit mode is often pretty daunting. A look at any one of the templates, and one immediately sees there are an excessive number of parameters to fill. Whilst it is clear that most are optional, these large numbers of parameters often encourage editors to insert all and any parameter, regardless of whether they add any value. I'll illustrate with {{cite web}}:

• {{cite web | url = | title = | first = | last = | author = | authorlink = | date = | year = | work = | publisher = | location = | page = | pages = | at = | language = |trans-title= | format = | doi = | archiveurl = | archivedate = | accessdate = | quote = | ref = | separator = | postscript = }}

I fail to see the utility of the vast majority of the parameters, although I have come across editors who diligently fill in almost all of them. I feel that a bare minimum suffice for achieving WP:V and WP:A. It could simply be distilled down to:

{{cite web |url= |title= |authors= |date= |work= |publisher= |accessdate=}} - noting that where it is a newspaper being cited, the publisher field is probably superfluous too.

Although I would be happy to see the various templates banished to reduce screen clutter, I think on balance their disappearance would lead to chaotic citation in articles. The answer is in streamlining the templates, paring them down to essential fields. Ohconfucius ^¡digame! 14:01, 31 January 2010 (UTC)

Much of the above points are true, but are also things that we are really not taking advantage of what bots can do to help out. One bot aspect that can be done is that if a new reference using one of the two standardized CiteT sets is added to an article (something we can track), the bot could reformat the citation template to remove unused paras, take the vertical arrangement into horizontal inline, and other quiet corrections. But then, if we were to push the use of list-defined references, another bot can be made to automagically move a full ref (regardless of template or manual or malformed) into the LDR block (presuming there is wide consensus to push LDR over full inline cites). That takes a lot of the work out for the newer editors, who, at least for myself, are thankful if they just hyperlink the web reference when they add information, and achieves the goal of helping to make the prose readible. You're still looking at the list of references and standardizing them, but that's now shuffled off to its own section. --MASEM (t) 14:28, 31 January 2010 (UTC)

• While I agree with the author of the initial comments' views regarding the issue of citation templates, I see it as more helpful than harmful. Having too much information in a citation is easier for a subsequent editor to fix than trying to locate missing information in deficient citations. In the absence of a uniform, pan-wiki citation style, MLA, Harvard... uniformity takes the place of these accepted formats and that uniformity is best achieved with templates. Also, templates are easier for experienced editors to maintain than disparate citations spread throughout Wikipedia--edit the template and that editor has made many improvements, edit an individual citation and they have made just one improvement. Simply put, in spite of the fact that I have many research papers under my belt and regularly consult my MLA handbook, I greatly appreciate the Reftoolbar and feel much more capable of not only adding my own citation with ease, but also of editing those of others. Supertouch (talk) 15:49, 31 January 2010 (UTC)
• The citation templates have some annoying qualities, but they have several advantages: they deliver consistently formatted references (whether that format seems right or wrong), and they keep all the details of the sources with the relevant text, so that text can be cut and pasted easily from one article to another, and articles can be split freely (named references cause a problem but now there's a wonderfully helpful bot for that...). I should add that what I personally detest is that indirect referencing scheme where you can't just click on a number and get to the reference and click on that, but now you're supposed to get to a reference that says "so and so p. 7" and then you're supposed to look for that in a second table. This is particularly annoying when what you might want is just whichever reference is online to look up a quick explanation. Wnt (talk) 01:25, 1 February 2010 (UTC)

(edit conflict) What you call the "indirect referencing scheme", and the annoyance you describe, has nothing to do with templates (please note the heading: Separating the issues). That is the footnote plus bibliography system (Tony calls it "short inline; long at the bottom") that many of the better articles are moving to. You do point out the inconvenience of that system, but it also has some advantages that are discussed above. A variation is to use Harvard-style inline references in parentheses rather than footnotes, but the downside of this is cluttering the article's visible text.

The solution to all the problems being discussed may be a new reference system, designed from scratch for online, hyperlinked wikitext (or HTML), and incorporated in Mediawiki software v 2175.0.1.0. Ideally, the system would make it really easy for editors, with no or minimal training, to enter citations (perhaps users will enter the information in an interactive input form that only shows relevant fields and that adapts as it is filled in; perhaps the software will retrieve from a database most of the citation data if the user inputs an ISBN, DOI, or other ID number) without learning markup or citation style, would format the citation in an acceptable, consistent format, would handle multiple citation variants from the same source (e.g., different pages or section numbers in a book or article; different pieces by different authors in a book that is an anthology; different articles in the same encyclopedia or other reference work), and an easy way for online readers to see the full citation (e.g., hovering over an inline footnote call would show the full citation of the work and the specific part [e.g., page(s)] being cited inline, with a link to the source's URL if there is one). While the developers are at it, perhaps the software would provide a separate system to present explanatory notes separately from references.

But that isn't going to happen anytime soon. And, there will be a LOT of debate (really, argument) among Wikipedians over what the system should do and how the system should do it, with many Wikipedians arguing against ANY standardization of anything (You can't tell me that the author's last [or first] name will be first. You can't tell me that the citations will use commas [or periods]. You can't tell me where or in what format he date [or page or volume number] will appear.) because, It's a wiki! Still, I can dream.—Finell 02:46, 1 February 2010 (UTC)

• Comment - I think there is a difference in opinion between editors who regularly edit in contentious areas verus less contentious areas. Personally, I usually edit in political articles, where you must cite the heck out of everything, or some partisan is going to come by and muck it all up. In contrast, people who edit in less-contentious areas see less of a need for citations. I think that distinction colors the debate between pro-citation template folks versus con-citation template folk. --Blargh29 (talk) 01:43, 1 February 2010 (UTC)

Quite right, although (it seems to me) it's not a matter of subject controversy, but of the "body of evidence" used. An old, well-researched subject (perhaps still contentious) is normally backed by real printed books and naturally calls for "short inline, long in the bottom" format where templates aren't an important factor. Some red-hot controversy has nothing on it but newspaper clips and must rely on a multitude of such online pieces (template clutter). NVO (talk) 09:22, 2 February 2010 (UTC)

• I personally don't use the citation templates, although I think they have at least one recognizable advantage, which is to effect system-wide changes if and when a template is improved. But I do agree they add greatly to edit-page clutter. One possible solution would be simply to have a text editor that highlights template code in a different colour. That would make it much easier to find the actual article text you want to change. Gatoclass (talk) 07:48, 1 February 2010 (UTC)
• Something else that could be done would be to change the template so that it outputs plainer text. So for example, in the template you might have author=joeblow, but when saved it saves as just "joeblow". That might be one way to get the best of both worlds. Gatoclass (talk) 08:01, 1 February 2010 (UTC)
  • Er, what do you mean exactly? subst:-ing the templates? That partially defeats the point. --Cybercobra (talk) 08:27, 1 February 2010 (UTC)

Well, to start with, the templates could be edited to allow two-letter abbreviations for some of the most used fields (ti=, au=). What would be best would be if while editing you could access a pull-down menu to make the citations appear within the text being edited as numbers, formatted template output, or raw template input. Wnt (talk) 21:56, 1 February 2010 (UTC)

Sorry, but I thought "the point" of templates was to ensure that citations were properly formatted. It seems to me that once the fields have been filled and the edit posted, there is no more need for the code. As I understood it the complaint is that the code is clutter, so by outputting the finished string that problem would be resolved. Gatoclass (talk) 18:19, 2 February 2010 (UTC)

I think that idea has merit. The supposed advantage of templates is largely consistency; subst-ing allows that to continue. (Changes can still be made by bot/AWB that a change in the template would have produced.) The disadvantage would be that editing a citation after subst-ing would require editing the text version, or else re-entering details into the template. I think that disadvantage is outweighed by the benefit of editors not having to read code designed for machine interpretation. (I know template citations put me off making changes at all.) Rd232 ^talk 20:43, 2 February 2010 (UTC)

How do templates put you off to making changes? It's easier to edit templates, in general, since one only has to fill in parameters rather than think about punctuation. — Carl (CBM · talk) 02:31, 4 February 2010 (UTC)

Because you can do punctuation any way you see fit, and it'll work (someone can fix it if it's wrong or inconsistent). The syntax for punctuation is the most basic wikimarkup there is, so people basically know it; parameters for templates they don't know (so they may need looking up), and getting them wrong will cause problems. In addition, filling in templates requires a certain logic and is unintuitive to nongeeks in a way that adding some punctuation is not. (The whole point of wikimarkup is that it's easy and flexible and prose-like; templates break that by making it inflexible (it'll basically only work one way) and programming-like.) Rd232 ^talk 11:59, 7 February 2010 (UTC)

I'd say it's time for a break

• Report from the field: I am reasonably experienced when it comes to the citation templates, and recently came across this discussion. When an editor added some non-inline references to Carinodens, I took the opportunity to try out LDR without other templates. I had not used LDR previously. Having tried it, I can see some advantages, although I would prefer to keep citation templates as well, and would fully support a bot conversion tool (not being particularly eager to manually convert dozens of refs). I appreciate templates for forcing a consistent style, and I think that a hybrid (LDR using citation templates) would be quite satisfactory, at least for what I do. If nothing else, it would be useful for recalling references to be used in other articles, instead of trying to find them. J. Spencer (talk) 03:34, 2 February 2010 (UTC)

J Spencer, that sounds like a very reasonable view. Any ideas on how to get this hybrid written, and what specs you'd like to see? Tony (talk) 04:13, 2 February 2010 (UTC)

Well, I'm not sure that anything special has to be written; the hybrid I had in mind consists of using LDR with references in citation templates. If references using citation templates already exist, all that's necessary is to pull those references into a reflist set up for LDR, and set up the text to display the new style. I tried it with Siquisiquesuchus, which only had three references, so conversion to LDR didn't make a great deal of difference for the editing experience. However, I can see it being very convenient to have the the full references out of the text body and collected in one place, and I'm a bit embarrassed to have not known about LDR before. J. Spencer (talk) 05:24, 2 February 2010 (UTC)

These brief examples don't address scalability issues of LDR. I tried to convert a larger article (offline) to LDR, starting with the twenty longest refs bound for <ref>group=note</ref> section... it just moved all the problems elsewhere and added the complexity of manually managing the sequence of these notes. Besides (it might be a personal idiosyncrasy) for me <ref>name=name01 group=note/</ref> is hardly an improvement over a full-blown template. NVO (talk) 09:34, 2 February 2010 (UTC)

What are you referring to by "managing the sequence"? The order of references at the end? I tried switching the order in preview on an article, and didn't see any different output. J. Spencer (talk) 23:50, 2 February 2010 (UTC)

As a note, the group= parameter is optional. It is there to allow if, say, one has "footnotes" and "references" or a similar split of sources, but even there, only one group has to be tagged as such. --MASEM (t) 14:53, 2 February 2010 (UTC)

Thanks for the note about the group parameter!

I do see one particular drawback of LDR, which is that if an editor is editing by sections instead of the entire article, adding a new reference is a two-stage job: putting the full reference in the ref section, and then putting the tag in its proper place in the article (or the other way around). One can work around this by using tabbed browsing, with the references section and the intended text section both open. J. Spencer (talk) 23:50, 2 February 2010 (UTC)

Yes, I've been thinking about that lately... If changing the name of a reference used in two or more sections can be annoying, imagine what happens when you have to edit at least two sections in order to add, move or delete almost every single reference. I use tabs, but this doesn't change the fact that having to work like this clutters the history and increases the chances that some kind of accident will occur between the two edits, resulting in mismatches. Waltham, The Duke of 05:47, 3 February 2010 (UTC)

Looking at this, I don't see a huge difference between LDR and the "normal" way of referencing with named refs. (I'm always naming a ref in one section and calling on it in another anyway) Though it would be nice to have a built-in script with a friendly button near the edit window that would bundle together any identical references and name them, and then move all of the instances with full information either to the first usage in the article or to the bottom in the references section at the user's option -- during the process of editing. Wnt (talk) 01:02, 4 February 2010 (UTC)

Doing that in general would violate the WP:CITE advice to preserve the citation style already established in each article. In any case, it's not actually that hard to edit any article regardless of what referencing style is used. — Carl (CBM · talk) 02:27, 4 February 2010 (UTC)

Ah, but if the next editor has access to the same convenient button, you haven't altered the editing style. True, to avoid a waste of disk space, it would be best to ensure that the references are actually stored the same way every time and the altered order is only regenerated when you press the Edit button. Now that I think of it, the default to put the references at the beginning or at the end could then be a user preference.

I think it's unquestionably easier to create new text with the full references inline, so there's less flipping involved; and it's also unquestionably easier to read and revise text with the references tucked out of the way at the end. Wnt (talk) 19:01, 4 February 2010 (UTC)

Software solution to citation mess?

The current system of citation templates is a mess, but citation templates are not the true, underlying problem; the current way of doing footnotes is. I have created a rather imperfect user script to separate refs from the content in the way described by User:Wnt, except that the refs end up in a different text box as described by User:J. Spencer (like editing an article using LDR as two sections). This exercise has convinced me that our footnote system is currently poor, simply because of the inline refs system first used with it 20:30, 20 February 2010 (UTC):

Problems with Wikipedia's footnote system

Scripts attempting to automatically process refs need to check all occurrences for the contents, which is less than ideal—my script would be much more complex if it were to handle these correctly (see the "limitations" section of the documentation). What happens is that over time, text gets merged into articles, and citations are scattered all around the article. LDR has the disadvantage that unused citations stay in the wikitext unless specifically removed. Inline refs do not suffer from this problem, but they do make the wikitext hard to read and maintain. 20:30, 20 February 2010 (UTC) If an inline ref is unnamed, it is hard to refer to an autogenerated list of refs without resorting to the use of cryptic, autogenerated names or numbers (e.g. AutoXY-1). Both the inline and LDR methods are not ideal when it comes to merging/splitting text. We end up with op. cit. footnotes since it is undesirable to have two citation templates that are identical except for the page numbers—not exactly the cleanest way to handle citations from books (page numbers are necessary for fact-checking).

Therefore, perhaps the best way to solve the problem is make some/all of these specific changes:

Specific changes to improve Wikipedia's footnote system (initial proposal)

Software changes S1: The contents of all named refs should automatically move to the first instances in the article when a page is edited. This is already established as best practice (see the Footnotes help page—"it makes sense to choose the first"), and the software should enforce it. S2: Cite.php should allow page numbers to be passed through to the citation templates, as previously suggested in Centralized Discussion (e.g. <ref name="abc" pages="39-40"/>). This is one of the few changes to the current citation template system I would recommend. The citation templates should get two additional parameters, automatically passed in by the software: {{REFPAGES}} with the specified page numbers, and {{REFFIRSTUSE}} with a section link to the first ref with the name (defined only if this occurrence is not the first). This way, the citation template can generate the op. cit. text if necessary (to reduce duplication of information in the references section). S3: Disallow nested refs. The nesting of refs is completely unnecessary and should not be used. S4: Create and adopt better editing tools that allow for easier footnote work. It should not be unreasonable to create two new server-side editing modes (e.g. called "editmain" and "editrefs") that hide the content or the inline refs, respectively (as with my script, but for accessibility purposes and for those without capable browsers). Provide an article merge/split tool that assists with footnotes. This tool would duplicate ref contents to all instances in an article, allowing easy merges/splits while maintaining proper footnotes and sourcing. When the text is inserted into the other article, a special "merge" attribute of the ref tag should cause the normal edit page to remove contents in all instances other than the first. Also WYSIWYG improvements on existing scripts such as WP:REFTOOLS might make citation templates more usable by allowing editors to see and manipulate the actual citation text (at least source code) generated, even for citations already in articles. For example: the article Comparison of Windows and Linux is a mess: on top of the already unwieldy table markup, it has inconsistently formatted citations. Some citations use templates, while others use simple hyperlinks (which do not give enough information for a proper citation). Such tools would help in standardizing the citation style of this article as well as making the source code more readable. Policy changes P1: Require that all refs have concise names. This would help to keep footnotes maintainable. It is possible that a bot could help with the task of adding names (based on existing citation template fields), possibly with human intervention. Is it possible to block further unnamed refs from being added?

Specific changes to improve Wikipedia's footnote system (2nd version)

Note: underlined text denotes major changes from the initial proposal. Software changes S1: Removed as non-applicable S2: Cite.php should allow page numbers to be passed through to the citation templates, as previously discussed (e.g. <ref name="abc" pages="39-40"/>). This is one of the few changes to the current citation template system I would recommend. The citation templates should get two additional parameters, automatically passed in by the software: {{REFPAGES}} with the specified page numbers, and {{REFFIRSTUSE}} with a section link to the first ref with the name (defined only if this occurrence is not the first). This way, the citation template can generate the Ibid. or op. cit. text if necessary, with a link back to the first occurrence (to reduce duplication of information in the references section). S4: Create and adopt better editing tools that allow for easier footnote work. There should be a new editing mode built into Wikipedia (e.g call it editsplit) that breaks off all LDR groups into separate textboxes for easier in-browser editing. Provide an article merge/split tool that assists with footnotes. The split tool would inline ref contents to all instances in an article, allowing easy merges/splits while maintaining proper footnotes and sourcing. When the text is inserted into the other article using the corresponding merge tool, a special "merge" attribute of the ref tag should cause the tool to move the contents to the appropriate LDR group. There should be bot(s) that help remove unused refs from the LDR section if they were not used within two weeks. This should be acceptable as uncontroversial maintenance; if someone wants the contents of a removed ref, it is possible to check the history. PleaseStand (talk) 00:04, 26 February 2010 (UTC) S5 (new): Improve Wikipedia's software so that unused LDRs are cleanly hidden when an article is viewed. Note that currently, an unused ref in the references section generates a big red error message at the end. PleaseStand (talk) 00:04, 26 February 2010 (UTC) Policy changes P1: Require that all refs have concise names. This would help to keep footnotes maintainable. It is possible that a bot could help with the task of adding names (based on existing citation template fields), possibly with human intervention. Is it possible to block further unnamed refs from being added? P2 (new): Embrace list-defined references as the most intuitive refs method to update and maintain in external editors. Discourage the creation of new articles using inline refs, and encourage the conversion of existing articles to LDR. This would require implementation of P1 (above).

As you can see, only some of the improvements proposed above directly affect the citation templates themselves. Even if we were to do away with citation templates, there would still be unsolved issues in the underlying footnote system. The big problem is that inline refs were introduced without concern for readability of the wikitext, important for maintaining articles. 20:30, 20 February 2010 (UTC) We need to fix these problems if we expect articles to become more maintainable. It is only responsible that I suggest improvements that are quite possible to implement. PleaseStand (talk) 04:59, 15 February 2010 (UTC)

• Update: I saw an example of Harvard referencing at Royal Albion Hotel, nominated and chosen for DYK, and actually think that it is an elegant solution (it cleanly separates page numbers from source information) but confusing to readers in that there are two footnotes sections, both similar to LDR. Depending on the article, this could also be a maintenance nightmare, so this is why I suggest the above as an alternative. (Note that this could also be made to work with LDR if desired: the software would repeat the citation footnote as many times as necessary if the pages attribute is used.) PleaseStand (talk) 19:22, 20 February 2010 (UTC)

• Royal Albion Hotel does not use Harvard referencing, it uses short footnotes.
• While it is undesirable to have references within references, it may be desirable to have references within explanatory footnotes.
• Any solution must work reasonably well when a conventional editing program, such as Emacs or Wordpad is used, because the Wikipedia edit window is a toy software. Any editing software that discard's the user's work as easily as the Wikipedia edit window is a toy software, unfit for serious work. Jc3s5h (talk) 19:56, 20 February 2010 (UTC)

• I have posted above a revised version of my suggestion that embraces LDRs as a solution that works in all external editors. I do understand your concern (I have written article drafts in Notepad). PleaseStand (talk) 20:30, 20 February 2010 (UTC)

• As for "there should be bot(s) that help remove unused refs from the LDR section if they were not used within two weeks" the reference section may contain general references, as well as references that have been cited to support specific passages. I suppose one could distinguish general references by not giving them a name. But they shouldn't just be deleted. Jc3s5h (talk) 03:38, 26 February 2010 (UTC)

• Actually, notice that to not use a ref from the LDR section results in an error message. Therefore, unused LDR refs must be commented out. Look in my sandbox if you want to see what I mean. Therefore, such unused refs must either be commented out or deleted. Certainly it is uncontroversial maintenance to fix a page so that it does not generate cite errors – at least comment out the unused ref. The reason I suggest deleting unused refs is that 1) it is not any more problematic than the current inline system and 2) it would control the size of the references section while many or most editors who have posted on this page agree that the long, inline refs make articles hard to edit. If it is not an answer to totally remove references, perhaps the bot could automatically move the unused refs to the talk page of the article. PleaseStand (talk) 04:04, 26 February 2010 (UTC)

• Chaco Culture National Historical Park is a much better example of shortened footnotes.

• Notice that most of the sources in that article are print sources. In that case, I clearly see the advantage of shortened footnotes: keeping citations out of sight for the normal reader of Wikipedia, while serving the same purpose as Harvard refs in scholarly writing. However, as seen in Royal Albion Hotel, to successfully use shortened footnotes requires that editors keep citation templates out of the inline (citations) section, adding a link to the LDR (bibliography) section using a shorter template. Again, as mentioned above about unused refs in the LDR section, this requires maintenance (can be with the help of bots to tag articles with incorrect citation style). PleaseStand (talk) 12:21, 26 February 2010 (UTC)

• LDR does not support unused references. If a reference is defined in the list but not invoked in the content, then an error message will show, unless the defined reference is within an HTML comment.

---— Gadget850 (Ed) ^talk 04:07, 26 February 2010 (UTC)

• Now, on policy. On P1: as long as the shortened footnotes are actually short enough, it seems unnecessary to add a name to these inline refs. However, on P2: I propose a process, similar to PROD, in which articles can be converted to whichever citation style is generally agreed on throughout Wikipedia after the close of this discussion.

• For example, on March 1, 2010, an editor suggests that a particular article must be converted to shortened footnotes. He adds the following template to the top of the article: {{cite convert proposal}} and optionally leaves a reason on the talk page.
• For 10 days, a {{dated cite convert proposal|month = March|day = 1|year = 2010|time = 12:00|timestamp = 20100301120000}} tag stays on the article.

• If the conversion is uncontroversial, after the 10 days have elapsed, the tag is replaced with a new tag {{cite convert approved}}, and the conversion proceeds.
• If the conversion is controversial, the {{dated cite convert proposal}} is removed. This would be mentioned with an {{oldccpfull}} tag at the top of the article, and larger discussion is needed to convert the article.

PleaseStand (talk) 12:21, 26 February 2010 (UTC)

Suspend replacing of ref by Template:r in citations

The supporters of a newish template, {{r}}, have been replacing <ref> by {{r}} in citations of dozens articles - Special:WhatLinksHere/Template:R gives an up-to-date these articles. I object to the introduction of {{r}} without previous discussion in the relevant articles' Talk pages and in the guidelines about citations. I am also concerned that there is no proof that {{r}} will work with all of the many tools used for citation and editing, including bots. I include a statement explained these reservations at the bottom of this RFC, after the "Support", "Oppose" and "Neutral" sub-sections. You are welcome to add your own statement after this. --Philcha (talk) 17:00, 1 February 2010 (UTC)

What exactly does Template:R do?

After reading much of the discussion below, and reading Template:R/doc, I really don't understand this template's purpose or function. The documentation is far too terse for a layman like me to understand: it explains How but not What or Why (once I grasped the basics, then I'm sure the details would make perfect sense).

I know it replaces "<ref>...</ref>" with a wrap-around template "{{r|...}}", but the specific implications, applications and effects have so far escaped me.

This isn't necessarily to criticize the template itself: I just don't have enough information or understanding to come to (let alone share) a useful opinion in either direction. —— Shakescene (talk) 20:29, 2 February 2010 (UTC)

It doesn't do so much really. Where you could write <ref name=myFavBook/> to cite a named reference, the {{r}} template allows you to write {{r|myFavBook}} instead. Both styles can live together within the same article without problems, even if citing the same source. Some people (well, me at least) like {{r}} because it is (I think) simpler to remember for newer editors, and reduces the amount of markup text in an article. Others don't like it (see below) because it (a) confuses bots and tools, and (b) having two valid ways of writing the same thing confuses editors. I hope that's a non-biased summary, but I'm sure someone will correct me if I'm misleading you ;-) GyroMagician (talk) 20:54, 7 February 2010 (UTC)

The templates invokes up to five named references; thus:

<ref name="refname1" /><ref name="refname2" /><ref name="refname3" /><ref name="refname4" /><ref name="refname5" />

can be represented by:

{{r|refname1|refname2|refname3|refname4|refname5}}

The template can also include page numbers in a manner similar to {{rp}}. ---— Gadget850 (Ed) ^talk 01:48, 16 February 2010 (UTC)

The proposition

The supporters of template {{r}} should not editing the use of {{r}} into any article until:

• all relevant policies and guidelines have approved the use of {{r}}.
• all tools / bots for maintaining citations had be proved to work as well with {{r}} as with <ref>.
• use of {{r}} in an article has gained active consensus, not just apathy.

Support (Suspend replacing of ref by Template:r)

1. (proposer of this RFC) --Philcha (talk) 17:00, 1 February 2010 (UTC)
2. I had never heard of the r template and did a little digging when I saw this. To me, that seems much more difficult to use and understand than the standard <ref ... /> formatting. As long as the MOS allows differing reference styles, no one should be unilaterally changing articles. Karanacs (talk) 18:01, 1 February 2010 (UTC)
3. Templates like these should be banned entirely. They introduce unnecessary complexity to an already complex system (wikitext) making it harder for new editors to learn and making it a constant learning process for experienced editors. They make it more difficult for bot and script writers to write programs that work consistently, and especially difficult to write scripts that work on other projects without lots of changes (fr:Template:R for example does something completely different). They also make it harder to import pages from one project to another (especially if there's a name conflict with the template). Adding another layer of templates on top of what's already one of the slowest parts of page generation (parsing citation templates) can only have the effect of making it slower. Mr.Z-man 18:04, 1 February 2010 (UTC)
4. We make it hard enough for new editors to grasp the html-ish ref and /ref bits -- this will be good for template folks, and a disaster for the rest of us folks. And I see no real gain with this over simply defining r and /r to be the same as ref and /ref in point of fact. We already can catenate in ref tags if we want. Collect (talk) 18:33, 1 February 2010 (UTC)
5. Per Mr.Z-man. I suggest a subst template instead. Sole Soul (talk) 22:52, 1 February 2010 (UTC)
6. There is nothing wrong with the current system. harej 01:48, 2 February 2010 (UTC)
7. This system does not seem to have been adequately thought through, and has numerous disadvantages. Its use should be suspended while a more thorough discussion is had over its implementation. Happy‑melon 08:52, 2 February 2010 (UTC)
8. per all the above. DrKiernan (talk) 11:02, 2 February 2010 (UTC)
9. I'm not a fan of the template itself, which I find confusing and unnecessarily complex for average readers, but the issue here is not whether one approves or disapproves of this particular reference style, sees advantages or disadvantages; rather, the issue is that this template should not be implemented without discussion, as has happened recently. I support the suspension of this referencing style until further action has been taken. María (habla conmigo) 13:19, 2 February 2010 (UTC)
10. There are disadvantages mentioned, and no clear gain. Finding duplicated references can be done with "ref" references without problem: simply go to the "References" section (or whichever has the "reflist" template in it), and start checking the references, which are all packaged toguether in there regardless of the section of the article they reside in. The ^ markers help with browsing: click them from a reference, and you are sended to the place in the article where such reference was made, just as clicking the number takes you to the references section. MBelgrano (talk) 13:26, 2 February 2010 (UTC)
11. If a user as in the know as Mr Z-man says it's not bot friendly, I'll take that as a fact. WFCforLife (talk) 13:54, 2 February 2010 (UTC)
12. If it's not bot friendly, it is also likely going to screw up the ability of the Mediawiki software to tell me when an edit has removed references in my watchlist (in addition to above statements). Also, unilateral changes are never a good thing without seeking a wide consensus. --MASEM (t) 14:50, 2 February 2010 (UTC)
13. Agreed. The bot-hostile nature of this template at the moment needs to be fixed first. Even then, there is no reason at all to replace ref tags with R for no real reason. If, once the bot issue is dealt with, people wish to use R instead, that is fine, but do it going forward. Resolute 15:00, 2 February 2010 (UTC)
14. Per Mr.Z-man and others. No justification for editors to be replacing existing, standards-compliant wiki footnote markup with this template. No benefit justifies bewilderment of editors who confront this template or additional server load.—Finell 21:34, 2 February 2010 (UTC)
15. If nothing else they make searching the text by humans a pain as well. Also, this is yet another sweeping change trying to get so well-established that it's really hard to get rid of it if discussion turns adverse; it's obvious that the community needs to sign off on something that will affect the way an double-digit percentage of articles are written. Mangoe (talk) 21:38, 2 February 2010 (UTC)
16. Support. Evidently too many problems with this template. Johnbod (talk) 10:39, 3 February 2010 (UTC)
17. Per Mr.Z-man, et al. Net negative to the project, in terms of compatibility and usability. Blurpeace 02:19, 4 February 2010 (UTC)
18. Too many concerns right now. Firsfron of Ronchester 02:35, 4 February 2010 (UTC)
19. Per all of the above. Just plain wrong, IMHO, and no valid reason to use this odd template. -- AnmaFinotera (talk · contribs) 07:27, 4 February 2010 (UTC)
20. Support. Unneeded solution brought to fix a problem that didn't exist. Binksternet (talk) 19:04, 4 February 2010 (UTC)
21. As Collectonian says, this is just wrong-headed. --Malleus Fatuorum 19:16, 4 February 2010 (UTC)
22. I see absolutely no advantage to this template over using <ref> directly and actually makes things more confusing than they are. The comments that it is very bot unfriendly also doesn't help matters. This is not a valid reason to use templates and is a solution looking for a nonexistent problem. As I see it, the template is completely unnecessary and I'm very tempted to send it to WP:TfD right now. —Farix (t | c) 00:19, 6 February 2010 (UTC)
23. Per Z-Man, etc. don't break something that works fine currently. Kthanxbai! Jrtayloriv (talk) 19:36, 15 February 2010 (UTC)
24. Wikipedia markup is hard enough to parse; it's hard enough already when ref contents can appear anywhere in the article. The page1, page2... is a kludge that is not a good solution for adding page numbers to citations; any real solution should be implemented in the Wikipedia software as an integral part of a good footnotes system. I agree that as-is, Template:R is a step backward and should not be implemented without strong consensus in favor. PleaseStand (talk) 03:03, 16 February 2010 (UTC)
25. support unnecessary complication - replacing refstyles circumvents authors rights to use their preferred citation style.·Maunus·ƛ· 05:45, 16 February 2010 (UTC)
26. Support - as a latecomer to the discussion, and having read the entire discussion (which took far to long), I support the motion without further comment from me in this poll.--Kudpung (talk) 17:52, 17 February 2010 (UTC)

Oppose (Suspend replacing of ref by Template:r)

1. First, I am assuming this is not a discussion about unilaterally changing articles. That is a done deal, and we have guidelines in place that say don't do it. How to approach the situation when it occurs is a different discussion (I prefer to try to explain to the offending editor why it's a bad thing and start a discussion, to avoid arguments like this one). Wikipedia evolves and markup changes. This is a good thing. I only recently discovered {{r}} templates, and I think they're very helpful. I've used them in two articles (Malvern, Worcestershire and Malvern Water). We (the group of editors working on those pages) have decided to try the {{r}} template on a those pages, to see how page maintenance goes. We will use the template on more pages if, after some unspecified time, we feel we've demonstrated that the template works without issue and that other editors can use them. The benefit is obvious to me, in that the page content markup becomes more readable (similar to the way that BibTeX and LaTeX are separated). As a secondary benefit, on both of the above pages we discovered reference duplicates, all with minor differences (ISBN 10 vs. 13, authors names or initials, etc). Finally, a proposal to suspend the use of a template in case it may do some harm, without any demonstration, seems bizarre and thoroughly un-Wikipedian. GyroMagician (talk) 19:19, 1 February 2010 (UTC)

   I've listed at least 4 concrete ways that it can cause negative issues in my comment above. As the author of one of the scripts in question and as someone who has worked with others on projects that try to read references, I can confirm that using multiple systems will cause a significant amount of extra work, and I would be rather surprised if any program that works with references currently supports the {{r}} system. Other than the fact that it requires manually going through all the references to implement it, how does the {{r}} system help you discover duplicates? Mr.Z-man 20:39, 1 February 2010 (UTC)

   Maybe this is the wrong place to discuss it (tell me if there is somewhere better), but why does this template break your script? The reference itself has the same syntax, but is in a different place in the article (I openly admit my naivety here, and clearly I don't use your tool - but I will have a look). I'm genuinely curious.

   The {{r}} template helps find duplicates because all the references are together at the end of the article, rather than peppered through it. If some sort of systematic naming is used, the same references end up in the same place (Malvern is not a particularly good example of this, but still enough that we found duplicates). I also found it helped in checking reference consistency (do they all use cite templates, books have ISBNs, etc). GyroMagician (talk) 22:56, 1 February 2010 (UTC)

   It doesn't actually break my script, it just won't work fully with it. The script will still output the traditional tag syntax even if the rest of the article uses {{r}}. It does break things like User:MZMcBride/climax as they only look for actual <ref> tags. This is also an issue with wikEd, which treats references and templates differently. For it to work properly, {{r}} would have to be special cased as a template that's actually a reference. I'm not familiar with the wikEd code, so I don't know how difficult that would be. Basically, any program that looks at refs will see a list at the end of the article, and nothing in the article body. Note that the main text of references can be put at the end of the article regardless of what format you use, its not something special to {{r}} - see WP:LDR. Mr.Z-man 00:22, 2 February 2010 (UTC)

   Okay, I understand - thanks for explaining. So it doesn't so much break, as gracefully degrade? However, I fully understand the point about messing up bots. Is there a standardised Wikipedia process for accepting templates, or testing/retesting bot compatibility? It doesn't sound like it would take too much work to fix the problem, but it does require a (probably busy) person to sit down and do it. And thanks for separating {{r}} from WP:LDR - I was one of those people confusing the two, until you clarified.

   99.99% of template changes won't have any impact on bots at all. Most of the ones that will are things like this - where a non-template is replaced with a template (or vice versa) or cases where the existing behavior of a template is changed for some reason (which will also break the current usage of the template). Mr.Z-man 22:12, 4 February 2010 (UTC)

2. Look, this is the sort of thing that I think we should be addressing, but there are many of you all who seem to think that the current system is perfectly fine (ie.: "let people do whatever the hell they want with citations, since it's not important anyway, it's all just formatting after all") If you want to standardize either using or not using these templates then get on board with Wikipedia:Centralized discussion/Wikipedia Citation Style. Addressing citation issues piecemeal will simply create too much conflict and chaos, and personally I'd rather have the current situation where there is effectively no rules then to have some rules for some things and none for others.
   — V = I * R (Talk • Contribs) 23:31, 1 February 2010 (UTC)

   I agree that, in an ideal situation, that would be best, but we're talking about Wikipedians arguing over style issues; the chances of coming to a consensus for a full citation style guideline sometime this century are practically zero. This isn't really even a style issue, as it doesn't actually affect the way the page is rendered, only how the wikitext looks. Mr.Z-man 07:18, 2 February 2010 (UTC)

   I guess this brings us back to the question of the degree to which the use of {{r}} templates imposes technical difficulties which negate any sense of gain by using it which could not be achieved by another, already available means. It seems to me that for our purposes, that has to be the central question of significance. Wotnow (talk) 08:40, 2 February 2010 (UTC)

   I've been intentionally avoiding the "technical issues" aspect of this in order to try and avoid being confrontational at all, but now that you bring it up... I'm a huge fan of scripting personally, but I don't see this as any sort of scripting challenge. If you can't deal with a change from <ref> to {{r}}, then quite frankly you shouldn't be writing scripts. As authors and users, we are and should be beholden to the needs of manual editing, never the other way around, as is embodied in the spirit of Wikipedia:Bot policy. I'm sorry, but you're just not going to get any sympathy from me on technical grounds. More importantly for me though is the overall aspect of this. Everyone seems to think that coming up with a consistent style choice overall is impossible, and yet here we are discussing one. If we can do this, why can't we go the whole way? Also, what is the actual proposal here? If there is significant support for this are we planning on deleting the template?
   — V = I * R (Talk • Contribs) 12:47, 2 February 2010 (UTC)

   I listed several disadvantages besides the technical ones in my initial comment above. The technical issue is not just bots, its also user scripts like WP:wikEd – things that are used by human editors – as well as people trying to import Wikipedia pages into other projects (for translation, saving content about to be deleted, etc.). I'm not saying that bot writers can't make their code work with new systems, I just don't see why they should have to on the whim of a few users, given that the benefit to human editors is so trivial in most cases (saving a few characters per ref). As I said, this isn't really a style decision, as the actual output is not different, only the wikitext. I don't quite understand your logic that it should be easy to come up with a comprehensive guideline just because one part is easy. For one, {{r}} is new, its easier to decide these things before tons of editors get used to it and use it on thousands of pages. To use a real-world example, just because politicians from different parties may be able to agree on Title IV Section 2 Subsection 7 of a new law, doesn't mean they'll be able to agree on the rest just as easily. Mr.Z-man 19:57, 2 February 2010 (UTC)

   What really bothers me about this is that it's based on almost the complete antithesis of our editing policy. Editors should be free to do whatever first and ask questions later, and this is attempting to set up some sort of approval process instead. The technical aspects, frankly, see like an excuse. I believe that yourself and others are offering them in good faith, but I very much question the utility of such views. Normal editors should not ever have to consider what impact they may have on bots, if for no other reason then that such consideration requires at least a little bit of computer programming skill in order to make reasoned decisions. No, bots should always be able to deal with anything that us as human editors could possibly throw at them, or they shouldn't be running here at all. That's the primary reason for BFRA and/or code review, after all.
   — V = I * R (Talk • Contribs) 20:45, 2 February 2010 (UTC)

   Bots will not just all of a sudden stop working, they just won't work as expected, but if you're suggesting that bot operators should be constantly watching for things like this, or should anticipate them somehow (and again, its not just bots), I'm not even sure how to respond to that. If you think the issue with bots, scripts, and the AbuseFilter (as Masem referred to) is not significant, what about the others? In no particular order:

   • It adds another layer of complexity to wikitext by introducing parallel systems for doing the exact same thing.
   • It makes it more difficult to use our content in other wikis, even other Wikimedia wikis (fr.wikipedia and it.wikipedia for example, have a Template:R that does something completely different)
   • Parsing references is already slow, adding another complex template on top them might slow it down further.
   • There's not enough of a concrete benefit over the current system

   Making a bold change to one or a few articles is fine, but I would disagree that unilaterally making changes to hundreds of articles should be done without discussion. Part of bot approval is to ensure quality code, but mostly its to ensure that the task is appropriate for mass application. I don't see why manual actions across hundreds of pages should be so much different than automated ones. Mr.Z-man 22:02, 2 February 2010 (UTC)

   This discussion seems to be getting a bit long in the tooth, so I'll just briefly mention this: in my view, if your script/program can't handle changes between the use of a template or XML style elements then it's bugged. That's my only real point when it comes to the intersection of this issue and bots.
   — V = I * R (Talk • Contribs) 23:47, 3 February 2010 (UTC)

   Perhaps I'm just a terrible programmer then. Maybe you can enlighten me. Without prior any prior knowledge of the existence of templates like {{r}} or how they're designed (since it was created and introduced into articles with no announcement or discussion), how can I write a program that can identify such templates and treat them as references, in case this happens again in the future? (Short of just quadrupling overhead by recursively expanding every template on the page, since I would prefer it if my script could run well on computers that aren't brand new and may not have broadband internet) And I am trying to steer the subject away from bots (see the list in my last comment), but that requires you to do so as well. Mr.Z-man 00:09, 4 February 2010 (UTC)

   OK, subject closed then. I wasn't trying to make any statements about your personal abilities, but if you feel that I've insulted you somehow then I'm sorry.
   — V = I * R (Talk • Contribs) 00:14, 4 February 2010 (UTC)

   I'm just trying to understand your argument here. I can that understand bots should not be given priority over human editors, but I don't quite understand how programmers should also be able to anticipate the usage of such templates, which I think is what you're suggesting. Mr.Z-man 01:05, 4 February 2010 (UTC)

   The essence of the above discussion, as I interpret it is this: Ohms law says human editors shouldn't be subservient to the needs of the programming. Mr.Z-man agrees, but says the problem for programmers is that it is literally impossible to foresee every scripting variation, and given this, if there is a scripting variation that does what's needed, why not continue with that? Mr.Z-man is saying that's not the same as making human editors subservient to programming needs, but rather, recognising that if we don't need to create a programming headache, let's not do it. Wotnow (talk) 01:30, 4 February 2010 (UTC)

   Pretty much, but I also pretty much rejected the whole premise that "it's impossible to foresee every scripting variation" because... well, it isn't impossible. There are only so many characters, and wikitext XML style tags and templates are well defined. Complaining that it makes scripting harder is just lazy (there, I said it). I never intended to get into a confrontation about this, since it's really your own personal business if you want to be lazy with your scripts/programs. That's a perfectly valid stance to take, as long as you're willing to admit that you're doing it, and you're not writing anything that is remotely "mission critical", which certainly applies to any user scripts/programs here. But, if you're going to be lazy about your script/program in the design phase, don't come around later and use that to try to justify complaining that something could break your program. If nothing else, by making the complaint you've already acknowledged that you recognize the issue, so just fix your program! OK, that may require a rewrite, but seriously: who's fault is that? Look, I'm not criticizing anyone's "skilz" here, especially since I could show you some really fugly (yet working) code if it came to that, but on the other hand you'll never catch me complaining that something breaks my code (at least, I hope not). That's just... unprofessional (yes, even for a volunteer project).
   — V = I * R (Talk • Contribs) 02:31, 4 February 2010 (UTC)

   If you don't want to keep discussing this, then please just answer my question. Without prior any prior knowledge of the existence of templates like {{r}} or how they're designed (since it was created and introduced into articles with no announcement or discussion), how can I write a program that can identify such templates and treat them as something equivalent to ref tags? You don't need to provide code, just a rough idea of how your algorithm works. Mr.Z-man 03:00, 4 February 2010 (UTC)

   I don't mind discussing this one bit, I just don't want to fight about the issue (and it is a bit OT here, really). The for-sure way to deal with this sort of thing is the obvious one: load the template. If you know what a template does you could choose to specifically ignore it, or deal with it in some special manner, by hard coding some function calls or whatever, but if you don't deal with the possibilities of templates then the design of any program which deals with wikitext is going to be inherently fragile. Incidentally, I think that you mentioned text highlighting earlier, and so I'll mention that there's just no need for special recognition in that sort of application because it has effectively 0 impact on the functionality (an {{r}} may not show up in the same color as <ref>...</ref> does, but you could add a specific handler for that if it's actually important to the userbase). It all depends on what the design specs are, really. Things will obviously be different for a "real time" parser (like a text highlighter) then they will be for a general page fixer type bot (SmackBot springs to mind here). Any full bot, which is going to make changes to article text, really should take the time to load templates simply for safety's sake. If the program can change article text then taking the time to completely process the source text should really be a design requirement, and if that makes some people's system's run slow then so be it. I can pretty much guarantee that any programmatic article text parsing can still be done at more then the 8 articles/second usual bot throttle anyway (that's 125 mSec's per page, after all; although, I think that the throttle may actually be 300 ms anyway, but the docs are a bit unclear here), but even if it slows the program down slower then that... better safe then sorry.
   — V = I * R (Talk • Contribs) 03:29, 4 February 2010 (UTC)

   I think there is merit in conceding the above points. Some of us haven't given much thought to bots, other than being aware that they exist for us to learn about if the need, desire, or opportunity arises. My own thoughts are that in due course, I should learn about bots, because there are obviously benefits to be gained, once one is familiar with the Wikipedia environment. It seems to me that - issues like {{r}} aside - mostly our edits have no impact on them anyway, or in a way that makes things difficult for those developing them or scripts.

   The central goal of those of use employing the {{r}} template was to achieve list-defined references. If you look at our discussions (tedious as I concede that is), this is self-evident. This certainly applies to discussions with GyroMagician, Chienlit, and Maedin. We just thought it neat that there was a method to declutter articles of references where this proved useful.

   It just so happens that when some of us became aware of what we now know are called list-defined references, it was through examples where the {{r}} template was being used. The natural assumption from this was that list-defined references are facilitated by the use of {{r}} templates. This is true. But it transpires that it is not exclusively true. That is, {{r}} templates are sufficient, but not necessary, to group references at the bottom of the edit-window. If you look at my user page, you'll see I started to arrive at this conclusion, from a different direction. After I'd become familiar with the use of the {{r}} template, and fleshed out some of the apparent issues, I then sought the history of the method, and to know if a name already existed. It transpired there was a history, and there was a name, as noted on my user page. Of course, on finding the Signpost notification and links I noticed that the help page, uses <ref> templates.

   However, in the absence of any information to the contrary (which there wasn't until now), I assumed that the {{r}} development was an uncontroversial refinement, and proceeded accordingly. Since list-defined references were my goal, in the end, I don't care one way or another whether say, {{r}} ceases to be used, or technical issues are overcome. And I don't have any difficulty conceding that there may not be any point in overcoming technical issues given that what we want to achieve is achievable by existing means. The current discussions highlight issues I hadn't been aware of, and I have no difficulty conceding that, given that I and others can still achieve the goal of list-defined references where it is considered as useful. Wotnow (talk) 00:03, 3 February 2010 (UTC)

   For the purpose of this discussion, I will take the "oppose" stance. The comment by V = I * R also makes sense to me, in that it's better to take such discussions to a forum where some across the board consistency can be generated, rather than too many ad-hoc decisions.Wotnow (talk) 05:30, 2 February 2010 (UTC)

3. The "r" template is easy to use and improves the editability of an article. It is also completely interchangeable and intermixable with basic ref tags. I would not like to see people cluttering up watchlists by mass-replacement of ref with r, but I also would not like to be prohibited from replacing them myself one at a time where I see benefit in doing so. Looie496 (talk) 17:47, 4 February 2010 (UTC)

I have no additional argument; I agree with this statement with one exception: it is only easier to use once one has learned the system. Sincerely, SamBlob (talk) 23:41, 7 February 2010 (UTC)

Neutral (Suspend replacing of ref by Template:r)

• Change of stance. It is self-evident to anyone remotely reasonable that my own stance was leaning towards neutrality anyway, and that insofar as good faith and logic prevail, I lean towards the weight of reasonable evidence. Given that, I have struck out my original comments, and pasted my comments that are clearly of a neutral genre anyway, below, since the general context of that stands. If people are capable of showing good faith, they should show it. If People aren't capable, well that speaks for itself anyway. Those who like to gloat can gloat, those who want to get on with it, can get on with it. Wotnow (talk) 11:33, 2 February 2010 (UTC)

I should also note that I played a part in generating this kerfuffle by bringing the issues into the open for discussion in the first place. Of course I have wondered about the skeptics, and of course the skeptics have wondered about us 'converts'.

But I should note first that regardless of the way the debate to date has unfolded, all of the editors who have used the {{r}} template to date, including myself, those from whom I picked up the method, and those who picked it up from me, have used it in good faith. Looking at the discussion overall, I think it fair to say that those opposing the use of the {{r}} template, or at least demonstrating skepticism, are also acting in good faith, but from a different knowledge-base to those of us who have been using the {{r}} system. I for example, have no experience with bots and know very little about them at this time, other than that they exist, and can do some useful tasks, especially those of a repetitive nature.

I am pleased to see the style of discussion between GyroMagician and Mr.Z-man above. It is of the solution-focused kind that I have enjoyed with him and others, including Chienlit. From the discussion between GyroMagician and Mr.Z-man, I begin to see some of the concerns of Philcha, especially given Pilcha's reply to me pointing out that his opposition has not been to the concept of list-defined references, but to the use of the {{r}} template, and the technical issues arising therein.

• The same thing that attracted GyroMagician to the {{r}} template, was what attracted me and others, namely the ability to improve the editability of both the article and the references. This is evidenced in the descriptions, demonstrations, and links I've placed on my user page as the process evolved of working through the issues that we became aware of (e.g. what's a real issue, what's a percieved issue, etc). If you check the link to the Signpost of 21 September 2009, you see a link to the extensive discussion in July 2009 which preceded the successful straw-poll on implementing list-defined references. And in those discussions, you see many of the conclusions that GyroMagician, myself, and others later arrived at.

Regarding the grouping of references, I note Mr.Z-man's comment to GyroMagician above, that "the main text of references can be put at the end of the article regardless of what format you use". I have tried to demonstrate this on my user page, both in examples placed into the page, and with links to discussions, and to articles where different styles achieve the same outcome. But following Mr.Z-man's comment, prior to saving this discussion, I have placed the following reference (<ref name=Norbert1948/> ), into the section User:Wotnow#Examples of differing citation templates co-existing without issue. This is because while I have placed a range of templates in the page as a whole, which you can see for yourselves, it occurred to me that I haven't put an abbreviated 'ref' template there, despite having done so in articles I've edited. The reason for that omission is of course that prior to this discussion, the fact of the {{r}} template, and not the List-defined references technique being an issue, has not been something I or others were aware of. This discussion thus usefully brings those things to the fore. But it would be wrong to infer from enthusiastic, good-faith editing, that there is an evangelistic desire of anyone to impose their will on others, and to react to them based on that assumption, which of course is not likely to be well received. It just seemed like a good idea, that's all. Wotnow (talk) 11:33, 2 February 2010 (UTC)

• I'm only neutral by default, since no one has yet answered my question above: #What exactly does Template:R do? —— Shakescene (talk) 23:02, 5 February 2010 (UTC)

Statement by Philcha

The supporters of template {{r}} have tried to impose the use of {{r}} on Arthropod without prior discussion at Talk:Arthropod, have failed to follow WP:BRD and has insisted on restore every instance of {{r}}. In my opinion :

• Guidelines state that article layout should not changed until either there is clear consensus for the change or it is driven by a policy or guideline. I've seen no guidelines that govern change of a new type of markup for creating citations, because until recently there was one mechanism, <ref>. However, the general principle is supported by many other guidelines, including many parts of WP:MOS.
• If significant changes are introduced, WP:BRD should be followed - which means that the change(s) should be back downp until consensus is reached either way.
• Install {{r}} should be suspended until it has get the approval of all relevant guidelines and other documents (see followin paragraph), so that {{r}} should not become the default for many articles that are currently not maintained.

The introducing of {{r}} would also cause confusion:

• Inline citations by <ref> is currently the only method of approved for citing in WP:CITE and this is followed into documents such as the FA criteria. I suggest that supporters of template {{r}} should stop install {{r}} in articles until it has get the approval of all relevant guidelines and other documents that govern WP procedures, such as the FA criteria.
• <ref> is used routinely by experienced editors and taught to new editors. I see no need for editors to learn another method when <ref> has proved to work well in all situations.

I am also concerned on the impact of {{r}} on many tools:

• Citation formatters such as refTool generally produce citations with <ref> tags. These tools would not need be affected directly by {{r}}, but the article would then be have a mixture of <ref> and {{r}} citations, which could harder to see - and it is not known where relevant guidelines and other documents would approve such mixtures. ----Philcha (talk) 06:51, 2 February 2010 (UTC)
• Edit box tools such as User:Cacycle/wikEd use colour-coding and other typographic techniques to identify different types of type in the edit box - main text, comments, citations, etc. {{r}} is currently not supported by such edit box tools, so {{r}} citations will look like part f the main text, making it harder to read.
• Many bots help to maintain citations - AnomieBot has fixed many problems for me, and others can produced fully-formatted citiations from identifications like DOI. I've not seen a set of test pages to prove that all these useful bots work with {{r}} - and I think {{r}} should be suspended until it is shown that it is compatible with all these bots. --Philcha (talk) 17:00, 1 February 2010 (UTC)

• I think this is a non-issue, as all that changes is the location of the full reference - the format of the reference itself remains the same, inside <ref></ref> tags. GyroMagician (talk) 19:42, 1 February 2010 (UTC)

  GyroMagician, there are 2 issues, which statements by Chienlist and Wotnow have not clearly distinguished:

  • list-defined references, which can be implemented entirely by <ref> tags, i.e. not using {{r}} at all.
  • The use of {{r}}, 1 of whose issues is its impact on tools that are designed to work on <ref> tags. This RFC is all {{r}}, and Mr.Z-man has supported my reservations about {{r}}'s impact on tools. --Philcha (talk) 06:51, 2 February 2010 (UTC)

    • Thanks for that Philcha. My own comments point to my recognition of the first issue, which I've demonstrated - in favour of your argument - on my user page prior to completing my comment above. However, it helps to boil that down to one succinct sentence, so thankyou. The second issue is of course something us non-bot users weren't aware of, but it is obviously an issue that is not insignificant, and again worth boiling down to a succinct sentence. Regards Wotnow (talk) 08:55, 2 February 2010 (UTC)

      My comments to GyroMagician are about the impact on tools that are designed to work on <ref> tags. This RFC also needs to consider the imposition of {{r}} without prior discussion and the confusion that seems likely as a result introducing of {{r}} as an alternative technique that is not approved or documented in well-known guidelines. --Philcha (talk) 09:37, 2 February 2010 (UTC)

I don't think we need an RFC to discuss imposition of significant changes to a page without discussion - it is accepted policy that such moves are bad. But it is important to remember that often this happens when editors get carried away, in good faith, rather than as a deliberate attempt to annoy someone. I believe the particular editor who started this has apologised? Many unnecessary arguments can be saved by simply explaining that some edits need discussion to reach a consensus. I actually find {{r|}} to b much cleaner than <ref></ref>, so I can't see the confusion argument. Also, most of us learnt to write wiki-markup by copying something, which is still how most new editors learn. We do have sandboxes and a preview button (and undo if it goes badly!). Now, the technical issue (interference with bots and tools) is a useful discussion, so let's stick to that. GyroMagician (talk) 19:18, 4 February 2010 (UTC)

The confusion argument stems from having 2 ways to accomplish the same thing. New users may not understand that {{r}} is just a wrapper around ref tags. I think its more "complexity" (more to learn) than it is "confusion" (hard to learn) in this case. If we were going to mass-migrate all articles to using this system, it wouldn't be as much of an issue; the problem IMO is from maintaining them in parallel. Mr.Z-man 22:15, 4 February 2010 (UTC)

• It all comes down to the two issues you correctly identified. One, list-defined references can be otherwise achieved. Two, it poses technical issues that outweigh the benefits. Everything else is a moot point. If we erred, we erred. We can get over it. It's no big deal. Wotnow (talk) 11:33, 2 February 2010 (UTC)

Does "suspend replacing" equate to "revert usage where it has already been replaced"?

A few months ago, the article Hubert Latham was converted from the "ref" system to the "r" system. I found the new system to be more complicated, but also more orderly and, once implemented, a bit faster to use.

I took the time and effort to learn the new system, and then, today, someone just changed it back. All of that gone down the drain.

The title of this discussion makes it look like the discussion is about whether the "r" template is to be stopped for new usage. Is this the case, or does it also include the reversion of articles already converted to the "r" template back to "ref"? Is the ultimate aim the deletion of the "r" template?

Also, how does this issue differ from the replacement of "references /" with "reflist"?

Sincerely, SamBlob (talk) 00:58, 7 February 2010 (UTC)

{{reflist}} should be removed for the same reasons mentioned above. That one doesn't even reduce clutter. —Pathoschild 08:53:06, 07 February 2010 (UTC)

I would have thought this requires consultation and consensus, like any other major change. But at least the before/after versions clearly show the clutter-reduction achieved with {{r}} and {{reflist}}. BTW, nice article, and enjoyed reading that - the opening paragraph is a gem! GyroMagician (talk) 18:01, 7 February 2010 (UTC)

Personally, I think {{r}} is wholly unnecessary and the benefits do not outweigh the problems and would support deletion if it came to that. But the proposal here has a clearly defined scope in the "#The proposition" section above. Mr.Z-man 18:31, 7 February 2010 (UTC)

Template up for deletion

Wikipedia:Templates for discussion#Template:R. ---— Gadget850 (Ed) ^talk 16:31, 15 February 2010 (UTC)

Reference hiding

I've noticed that the reference section for some articles is becoming almost as long as the article itself. I think, as I'm sure most wikipedians do, that the more meaningful citations the better. However, the time might be coming (or is here already) when having an collapsed reference section might be advantageous. Perhaps the function could be added to {{reflist}} either as an option or as the default. Clicking a reference in the text would automatically expand the section. Just a thought. –droll [chat] 01:15, 9 February 2010 (UTC)

Why would it be advantageous not to be able to see (and instantly scroll to) a particular reference? Is there an incoming tax on vertical screen space? "Clicking a reference in the text" would not automatically expand the section, because that functionality is not available with the collapsible tables code we currently have. Happy‑melon 15:15, 9 February 2010 (UTC)

Such auto uncollapse functionality could easily be added though, I added such code to the Mobile version where all sections are collapsed by default. So the technical implementation is not really an issue. —TheDJ (talk • contribs) 15:50, 9 February 2010 (UTC)

I can see the advantage for mobile, but what advantage does a collapsed reference section give us for desktop readers? Does the page load faster? Collapsed boxes uncollapse when printing, so it won't save paper. Links from the article to the reference section won't work. ---— Gadget850 (Ed) ^talk 16:02, 9 February 2010 (UTC)

Actually, I'd be willing to lay money on odds that it'd cause page load times to increase (somewhat dramatically, as well), as long as browser render times are taken into account. It obviously wouldn't make much of any difference in transfer times, since it'd just be the addition of a few hundred chars at most. Those collapsing divs take a little bit of time to render though, especially around large chunks of text.
— V = I * R (Talk • Contribs) 07:04, 10 February 2010 (UTC)

Come to think of it. I once made a Javascript tool for someone with the exact same question. importScript('User:TheDJ/foldrefs.js');. —TheDJ (talk • contribs) 16:11, 9 February 2010 (UTC)

Well as to why a change might be a good idea, to me its a matter of page layout. It appears to me we should show people the meat. Even though the reference section is very important it is not why the wiki gets hits. The product is the article. The all important references would be available when wanted. –droll [chat] 06:44, 10 February 2010 (UTC)

Well, personally, I'm not a fan of the idea. First, the lifeblood of any Wikipedia article is the references, so the last thing we want to be doing is hiding them, just on a "general principles" or "philosophical" level. More importantly though, I'd turn your own argument around on you, and say that the layout would be the largest sufferer. Simpler is better when it comes to web pages. I'm not one of the purest ideologues that think that any use div or span tags is an offense against nature and God (believe me, their out there!), but the cleaner, simpler look is generally the way to go. But in the end, you're assumptions about the content being more important then the references is just plain wrong (or, rather, misplaced). I'd almost say that it's backwards actually, but not everyone feels that way either (obviously, since you don't). There are different ways to use Wikipedia, and those different users have (sometimes vastly) different needs. People sometimes have a tough time realizing that, which is probably part of the reason that AFD is such a battleground at times, for example.
— V = I * R (Talk • Contribs) 07:10, 10 February 2010 (UTC)

I'm struggling to see how this will help really. I don't think it's going to dramatically alter page load times and I see no problems with simply scrolling past reference lists if you really want to get to the bottom of a page. I can't see how hiding such important information is likely to help at all. TomBeasley (talk) 17:08, 10 February 2010 (UTC)

• This has been discussed to death before. Collapsible references fall afoul of WP:ACCESS as they can make it difficult for those using screen readers or other devices/software to view them. → ROUX ₪ 17:52, 10 February 2010 (UTC)

• By the way, I'd be very interested in hearing your views which lead to this proposal. Feel free to ping me on my talk page if you'd rather avoid the nastiness around here.
  — V = I * R (Talk • Contribs) 17:56, 10 February 2010 (UTC)

I have an old slow computer. (But my DSL connection is fast, I can download movies quickly.) For me some articles that have lots of references can take well over a minute to render. I know from testing that hiding things makes page rendering a lot faster, if the hiding is done before the page is rendered. And we know how to make it so that doesn't cause problems for users that don't have javascript.

I have manually added code in my personal /monobook.css to hide lots of things that I don't need, and that makes page load times much shorter. Actually, without those fixes I would not be able to edit Wikipedia from this computer, and I can't afford a new computer...

We have also added such code in the global javascript, for instance we hide the static edittools box below the edit window from javascript before the page is rendered, then after the page is rendered we add the javascript version of the edittools. (The edittools is the box with special characters to insert.) We did that hiding mostly to stop the page from jumping up and down during page rendering, but as a side effect it made loading of the edit pages faster for people like me. Since we do the hiding from javascript users who don't have javascript see the static edittools as usual.

So we already know how to code this. And we already know it makes page rendering much faster. So I would very much like that the references section would be collapsed by default.

Roux: I bet most "hearing" users of Wikipedia would be glad to by default not hear all the references...

--David Göthberg (talk) 20:42, 12 February 2010 (UTC)

You can't afford 80€ to upgrade your machine? That's hard. Mind laying out your specs? Maybe I can give you a tip. If you do, we better continue this at one of our userpages. Paradoctor (talk) 10:35, 20 February 2010 (UTC)

That effect is primarily from local page caching, which your browser does automatically (there are ways to turn that off on most browsers, but it's browser dependent, and the details are somewhat technical). The length of time to render a page is somewhat dependent on the Wikimedia servers as well, since template handling and the cite.php extension both require some extra server processing (and low priority processing, at that).

The main thing though, to me at least, is just that the last thing we want is to be hiding references in any fashion. I don't think that comparing this to avoiding the use of audio to read references is at all a fair or meaningful parallel to draw here, for what I believe are fairly obvious reasons. We could implement something like this, and even do it largely without performance penalties with a bit of work, but I don't think that we should, and it seems that I'm not alone.
— V = I * R (Talk • Contribs) 21:12, 12 February 2010 (UTC)

Uh? Page caching in the web browser has nothing to do with the faster page rendering I am talking about. When using CSS to hide part of the page (part of the rendered HTML) then the browser still needs to fetch exactly as many bytes, but it needs to do less rendering work. In many articles today the references are much more than half of the HTML in the page, so not having to render half of the page is much faster.

And I am well aware of when and what my browser caches. So I do my test runs in ways that are comparable and measurable. And I repeat them several times. Not like I have no hiding on first page load, then turn on hiding and load the page again (from the browser cache) and go "wow, the second page load was much faster!".

Wikipedia is now too slow to edit on old computers, unless you are a geek like me who knows how to configure away and block a lot of stuff. (The usability javascripts were the final killer, I had to adblock them to be able to continue editing since they made stuff way too slow.)

And it is probably not especially doable to read Wikipedia over a 56k modem connection anymore, unless you turn off javascript. Remember that Wikipedia is meant to be readable and editable also to users in other parts of the world, where DSL connections are not as widely available.

So we should do what we can to make Wikipedia load faster, since page load times of 30-60 seconds or more are simply not acceptable.

--David Göthberg (talk) 07:20, 13 February 2010 (UTC)

Humm... do you have stats or a reproducible test run setup that you could point me towards?
— V = I * R (Talk • Contribs) 17:17, 13 February 2010 (UTC)

What kind of world do we live in that 30-60 seconds is "unacceptable"? Really?! Put a baked potato in the microwave while you're waiting next time. What is an acceptable time and what isnt? Do we really have a "responsibility" to have Wikipedia set up for the lowest common denominator; ie- the slowest type of computer and/or internet service? Would that not keep us from innovating and improving the encyclopedia by adding the newest and "coolest" abilities to Wikipedia if we must always make sure "older" computers can still use us? At some point we must draw the line and say "ok, those with Commodore Amiga's that have been set up to use the internet... that's your own tough luck, buy a Mac if you hate PC's so much" (and no, technically a Commodore is not a PC, PC refers to IBM Clones and a Commodore is not, just like Apple but of course if you disagree take it to your own talk page not here and not mine).Camelbinky (talk) 07:59, 13 February 2010 (UTC)

• What kind of world do we live in that 30-60 seconds is "unacceptable"? -- This kind.
• Do we really have a "responsibility" to have Wikipedia set up for the lowest common denominator; ie- the slowest type of computer and/or internet service? -- No, but you're just setting up a straw man here. Nobody is talking about Commodores or Amigas. Just dialup connections, and other slow internet connections, which a lot of people still have. Try using the slow-as* internet at the cabin I live at in Idaho, and you'd really appreciate not having to wait an additional 30 seconds on top of the current 15-20. It's easy to answer no to this question, when you don't have to deal with it personally. But all the people that do will appreciate unnecessary changes with no significant benefits causing them significantly increased problems.

--Jrtayloriv (talk) 19:43, 15 February 2010 (UTC)

I'm not sure what you mean by "low priority processing" – the servers that render articles do basically nothing but render articles. What would have a higher priority? Mr.Z-man 07:52, 13 February 2010 (UTC)

How about just making a script that collapses the references, for people that care about "scrolling" to use. Everyone else, who likes the current layout, can keep it, and not have their DL times increased, etc for no good reason. sound good? Jrtayloriv (talk) 19:45, 15 February 2010 (UTC)

Personal .js or a gadget would seem to be a perfect solution for this sort of problem. I seriously doubt that it would affect actual load times at all (considering the fact that the same amount of data would need to be transmitted), but if you want the references to be collapsed or invisible for yourself, then that's perfectly fine by me.
— V = IR (Talk • Contribs) 20:48, 15 February 2010 (UTC)

I'm not sure how server-side scripting is presently implemented for WP, but clearly different sizes of thumbnail image files are being provided based on user pref settings. I see no reason in principle why a server-side script could not, in principle, choose whether to include the refs section inline or just offer a link to it, based on user settings. Is someone aware of the technicalities involved? User:LeadSongDog come howl 18:00, 17 February 2010 (UTC)

Sure for logged in users such a thing would be possible. There would be little technical difficulties. —TheDJ (talk • contribs) 16:13, 19 February 2010 (UTC)

Wonderful. I was thinking particularly of mobility and wp:accessibility issues. If a portal, a session cookie or some such could effectively say "this user or ipuser wants terse output" then our ability to deliver such could go a long way to addressing the inate conflict we've had (between the desire for comprehensive content and the desire for universal accessibility). User:LeadSongDog come howl 16:33, 19 February 2010 (UTC)

I note that specifically for mobility, Logged in users don't yet exist :D (And on m.wikipedia.org, all but the first section are collapsed by default.) —TheDJ (talk • contribs) 21:18, 19 February 2010 (UTC)

That's a good thing. Please don't edit while driving 8-0 However, viewed through a desktop browser such as the antiquated IE6 I have before me, http://en.m.wikipedia.org/wiki?search=Oxford_English_Dictionary for instance plops down the whole page, just with a different layout. It'd be ideal to be able to get just the lede and TOC, especially when paying out-of-country roaming rates or on dialup! Note that collapsing on the client side doesn't help that problem at all, it has to be on the server side. User:LeadSongDog come howl 22:02, 19 February 2010 (UTC)

IE6 is not a supported mobile device. Only iphone, android, Palm WebOS, Sony Ericsson, Opera Mini, and the game consoles are supported. (Why ? because we don't have enough devices to test.) The mobile site is not targeted at non-mobile platforms. —Preceding unsigned comment added by TheDJ (talk • contribs) 22:13, 19 February 2010 (UTC)

Sorry, I didn't mean to imply that it was, just that the mobile site was serving up the whole, flat-file article rather than collapsing it on the server. Does it do the same for mobile devices? If so, that's a lot of money we are moving from our readers' pockets to those of their wireless service providers'. But the main concern was all those dialup users on desktops and laptops. They are still a large part of the readership. If we can make their experience faster we should do so. User:LeadSongDog come howl 07:44, 20 February 2010 (UTC)

Yes, the mobile interfaces does the same. Server side collapsing of sections is a tad difficult. Not technically, technically there is no issue what so ever. There is an issue with the timespace however. Basically, if you retrieve sections (including the references) separately from the original request, you don't know if both are based on the same revision. Basically, it will introduce editconflict problems for readers. —TheDJ (talk • contribs) 14:19, 20 February 2010 (UTC)

P.S. let's make a clearer separation here. There is a difference between collapsing (which I prefer to consider to be a javascript show/hide action) and content-division (sometimes called pagination). —TheDJ (talk • contribs) 14:36, 20 February 2010 (UTC)

Wikipedia:Accessibility is no longer an issue— recent discussion at Wikipedia talk:Accessibility#MOS:COLLAPSE outdated advice? determined that collapsing and scrolling are not accessibility issues. ---— Gadget850 (Ed) ^talk 22:06, 19 February 2010 (UTC)

Template {{refbox}} was recently created for collapsed reference lists; currently unused. ---— Gadget850 (Ed) ^talk 14:36, 20 February 2010 (UTC)

Content note templates cause HTML validation errors

Templates {{ref}}, {{cref}} and {{scref}} are use to create separate references or explanatory notes in different styles. They use an HTML id for the backlink. When the same note is used multiple times, duplicate ids are created in violation of the HTML specification. For example— {{NBC Universal}} uses {{ref|2|2}} more than once; see the markup validation for the errors that propogeted to the pages where it is used. I started discussions on the talk pages, but there has been no response.

Frankly, most uses of these templates could be replaced by <ref>...</ref> with |group=. As I see it, the only need for a content note template is in list articles where the same reference is used numerous times. ---— Gadget850 (Ed) ^talk 17:09, 17 February 2010 (UTC)

<ref>...</ref> doesn't seem to work well for when you need to make a ref call from within a note - at least I've never got them to work.Nigel Ish (talk) 17:41, 17 February 2010 (UTC)

As documented at Wikipedia:Footnotes:

{{#tag:ref|reference content including ref tags|group=groupname}}

If that is too cryptic, a simple template would cover it, but it would not really do anything but insulate the #tag magic word from the user. ---— Gadget850 (Ed) ^talk 17:54, 17 February 2010 (UTC)

I fixed the problem for {{NBC Universal}}, using <ref>...</ref> with |group=, and added some cautionary notes to {{ref}} and friends. It's not at all clear to me how to fix the problem within the {{ref}} framework. Eubulides (talk) 00:36, 18 February 2010 (UTC)

Back on track

I am rather a latecomer to this discussion. I'm joining it here now because some parts of it concern action taken on two articles (mentioned in this discussion) of which I was the major contributor/creator, one which contained about 100 citations. I naturally do not own anything I write for this encyclopedia, and I therefore left it to others (in whom I generally have great confidence in their professionalism on this encyclopedia) to get on with some plans for modifying the citation system. One thing however that did perturb me was that this action took place without the required consensus being solicited, re:

Converting citation styles should not be done without first gaining consensus for the change on the article's talk page. So, tread lightly, and seek consensus first, before converting citation styles.

something which I thought the experienced editors concerned would have known about, and while one of the erticles (the one with over 100 refs) was in the middle of a GA review. (Nuff said, and I didn't fall out with the editors over these issues). So I came here to see what it was all about. I didn't expect however, to see a very different, but very noble and pertinent proposal at the top by John Cardinal.
The discussion has been turned far more into the mechanics of various templates, in particular the {{r}} thingy that I came here looking for but found John's original motion to far more interesting and to the point(s) - namely:

• Citation style
• Use of templates

before the whole thing got snarled up with the {{r}} template issue.
Let me point out one very distinct disadvantage of any system that puts the information bit of a reference someplace other than inline in the article text (because I don't think it's been mentioned in this discussion): it causes havoc when working on a series of closely related articles where a lot of text complete with embedded references is copied and pasted from the master article for use somewhere else, or where just the reference is needed for further use in another article. With a hundred or so citation templates tucked away in a different section, even if they are alphabetically listed, the editing work is doubled, the hunt for the reference takes too long, and the {{reflist}} section takes too long to load. One of the original concepts of Wiki software was the ability to edit sections only, avoiding the need to load the entire article to make one or two quick edits or even to work in detail on a section. The automatic creation of the numbered references in the reflist was a godsend. Whether I support anything in this discussion or not needs no special contribution other than copying and pasting these few comments that I have picked out as I scanned through this monumental discussion and wholeheartedly support:

• A good article is likely to have footnotes to a lot of sources.
• Sources, even reliable sources, come in many flavors, and have many possible complexities.
• Templates designed to anticipate complexities are likely to be gruesomely large
• Citation templates are already so cumbersome that even when they are sufficiently flexible I for one never use them. I find it next to impossible to write coherent prose around them, and instead must constantly switch between the edit window and the preview, which is tiresome.
• I disagree that {{cite web}} is so easy to use. Maybe I'm missing a feature somewhere but how do I insert it with a single click?
• I detest the cite templates—won't use 'em, won't even work with 'em—because they don't have any flexibility, they encumber the loading, and so on. I type out all my citations, and find it quite easy to do--much easier than using either the cite templates or the named refs.
• Citation templates are some of the worst things to have happened to Wikipedia in terms of the writing. It's impossible to write well with them in the text, or to copy edit for flow, so articles become lists of sentences instead of flowing narratives.
• The use of these dreadful citation templates is a cancer on WP. I advise all editors to write out the citations nice and simple,
• In the case of most URL references, encouraging authors to leave basic title and access dates in a uniform fashion would surely be desire to the blight of raw URL references.
• Please note that the use of WP:LDR also makes pages to load much slower, if you have hundreds of citations.

Furthermore, while he suggestion that reducing the citations to the high-value ones, just as overlinking reduces the value of links, overciting reduces the value of citations, may sound reasonable, one problem we have to contend with is that there are over-zealous editors out there, particularly GA and FA reviewers, and those on a mission to tag and flag to make them selves a massive edit count, who insist that almost every darned sentence in an article has a citation, and another problem is that I peronally know of at least one GA reviewers who freely admits that they do not think it necessary to check and verify references!

IMHO, let's get back on track and make some progress on Johns original proposal, before discussing the pros and cons of the different templates and the effects they have on bots, or the bots on the templates and the ins and outs of php, html, and js. For those that haven't looked at the top of this page in a while, it was this:

I propose that we establish a standard citation style. Given the prior difficulties reaching consensus, I further propose that we form a small committee to define the standard. The committee should be free of people with entrenched preferences and instead include contributors who will work collaboratively, who will compromise, and who will place the primary goal ahead of picky detail.

I think it's one of the most important issues (along with resolving unsourced BLPs) currently under discussion. I'm sure that John's original suggestions were aimed at establishing the way to go rather than allow this discussion to be taken over by the techies and turned into a talk about various programming languages which most of us, the regular content authors, are not really all that interested in hearing. What we want are straighforward solutions that make our editing easier, and that encourage new users to provide references.
I am one of those people who do not have entrenched preferences and whose primary goal goes ahead ahead of picky detail. I say let's get that committee started otherwise we'll still be debating this in 2012, and I'm ready and waiting with my sleeves rolled up. --Kudpung (talk) 19:27, 17 February 2010 (UTC)

As I see it, this page is a central discussion for citation issues in general and has several different discussions. You may want to move your comments to whichever discussion you were referring to. ---— Gadget850 (Ed) ^talk 23:04, 17 February 2010 (UTC)

No - I think I have summed it up adequately because like all Wikipedia 'central discussions' this one has become an unfocused tangle of huddles that would need 73 A4 pages to print it on, and it's almost impossible for a latecomer to get a proper overview. Whatever the different discussions are, it's time to stop the talking for the sake of talking, and get some action that moves us forward rather than having to spend hours reverting good faith technology to citations and their templates and discussing the ins and out of computer code.--Kudpung (talk) 08:40, 18 February 2010 (UTC)

Inline template wikitext formatting

When using inline templates which require parameters (the citation templates being the most obvious, here), in general do you prefer to keep the parameters together with each other (some people call this "collapsed"), or do you prefer to place each parameter on it's own line? 08:51, 20 February 2010 (UTC) Please note that stand alone template wikitext formatting, such as prototypical infoboxes, are outside of the scope of this discussion. For example:

Example wikitext

Single line

Multi-line

Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.<ref>{{cite web|url=http://www.example.org/|title=My Favorite Things, Part II|author=Doe, John|publisher=Open Publishing|date=30 April 2005|work=Encyclopedia of Things|accessdate=6 July 2005}}</ref> Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

Lorem ipsum dolor sit amet, consectetur adipisicing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.<ref>{{cite web |url=http://www.example.org/ |title=My Favorite Things, Part II |author=Doe, John |publisher=Open Publishing |date=30 April 2005 |work=Encyclopedia of Things |accessdate=6 July 2005}}</ref> Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

If you feel that one style is better then the other, please explain the reasons why. Try not to get stuck on the merits, or lack thereof, regarding the use of specific templates ({{Cite web}}, {{Clarify me}}, {{Coord}}, etc...). The primary goal here is to achieve a sense of preferred wikitext style among active editors. Thank you. 08:51, 20 February 2010 (UTC)

Inline template wikitext discussion

• Single line It takes up less space. I don't think either style should be officially sanctioned. --Cybercobra (talk) 09:10, 20 February 2010 (UTC)

• My current practice, an opinion only:

• I dont use {{Cites in line- I only put the cites in a bibliography, and inline a <ref name=XXXX>{{Harvnb|XX|XX|p=XX}}</ref>. This is because I produce text for articles in a limited field that refer to the same 16 texts. (Hence the importance that cite and citation ref=Harv works)
• It I am doing a quick an nasty inline, say to a few lines in a Newpapers web site, I tend to do a raw <ref>, as the cite templates are too much hassle and I lose the flow of my prose.
• While editting old text I like my citations spreads over a few lines, and will often reformat condensed text so I can analyse it more thoroughly. There sure is no reason why condensed cites cant be expanded on their firat save in the same way that a signature is.

--ClemRutter (talk) 09:35, 20 February 2010 (UTC)

• Don't police This matter is not important enough for further discussion. Or can you point at a substantial history of editwars over that one? Personally, I use inline for short and multi-line for large invocations, and follow what's there when editing others. Paradoctor (talk) 10:12, 20 February 2010 (UTC)

• No RfC needed. Why do you need to know? It doesn't strike me as important in the slightest. OrangeDog (τ • ε) 10:21, 20 February 2010 (UTC)
• Makes little difference. What we really need is some way to avoid putting the full text of refs embedded in the article text. ClemRutter's approach of using a bibliography, above, works, except when it doesn't, for the reasons he mentions.--Father Goose (talk) 10:43, 20 February 2010 (UTC)
• At least two of the above seem to be thinking in terms of citation templates only. Reading the initial statement twice, it's clear to me that all inline templates are meant, in which case single line. This is so that when one is inserted into the middle of an existing paragraph, a diff will show the actual template added. If line-breaks are involved, the paragraph effectively becomes split, and so test subsequent to the new template is not tied back to the original. --Redrose64 (talk) 12:56, 20 February 2010 (UTC)
• Single line so that diffs work, as per Redrose64. Eubulides (talk) 19:15, 20 February 2010 (UTC)
  • Say, just when is diff going to be fixed? Working around its limitations is really not the right answer. User:LeadSongDog come howl 17:34, 6 April 2010 (UTC)
• User preference: Some users prefer the compactness of the single-line style, and some prefer the easier template editing afforded by the multi-line style (I personally prefer the former). It doesn't really matter either way; while it has been noted both have merits and limits, judging which to use should be based upon user preference, in my opinion. It seems needlessly bureaucratic to have some sort of style guideline regarding this, if that is indeed the goal of this RfC. Intelligentsium 19:38, 20 February 2010 (UTC)
• Personally, I prefer WP:Ref#List-defined references, single-line. The best way to keep online sources from cluttering the page source. I don't know the intentions behind starting an RfC on this (or the weird pseudo-secrecy of insisting to keep the signature of the initiator out of it and the advertisements), but I hope you don't intend to enforce any particular style. Amalthea 20:06, 20 February 2010 (UTC)
  • This is not a discussion about referencing style. It's about whether or not to use line breaks in inline templates. --Redrose64 (talk) 20:58, 20 February 2010 (UTC)
    • Right. I should have noticed the underlined, bolded "inline". :) If I use inline citations, I very much prefer single-line. Amalthea 10:36, 21 February 2010 (UTC)
• It depends, plus who cares? I tend to use the multi-line formatting for inline refs, and single-line for most other templates, but for WP:LDR, I'd switch back to the single-line format for citation templates. I often match style with whatever's already been done. Also, I think that changing from one to the other ought to be on a list of worst ways to waste an editor's time. It simply isn't important. WhatamIdoing (talk) 08:39, 21 February 2010 (UTC)
• Matter of taste. I occasionally use the "multi line" when the accessdates need updating regularly (such as for Rotten Tomatoes, Metacritic reviews on film articles, etc). It helps in finding where the accessdate is, without the strain on my eyes. I don't agree that EVERY cite should be multi lined, as that is excessive and difficult to find the text you want to edit. It should be left up to the editor. —Mike Allen 06:47, 24 February 2010 (UTC)
• Do it in the software. I think can be taken care of in software. For example wikiEd supports "template and reference hiding". It should not be a problem to hide, expand (multi line) or contract (single line) templates in the editor software. I find multiline easier to maintain. I usually do short templates single line style. –droll [chat] 07:46, 2 April 2010 (UTC)
• Almost past caring. Multiline is right, but it takes too much energy to convince people of that fact. Source code should be legible or it isn't maintainable. See any issue of any reputable software journal in the last twenty years for examples. User:LeadSongDog come howl 17:34, 6 April 2010 (UTC)
• Single line - yes "Multiline is right" for complex templates but inline templates should be simple, and LDR fixes the most complex ones. 79.67.204.119 (talk)

COinS in citation templates

Members may be interested in discussing the utility of COinS metadata in Template:vcite journal and related templates. Martin (Smith609 – Talk) 20:00, 5 April 2010 (UTC)

Redux

It's clear from the volume and tone of discussion here and repetitive arguments elsewhere, not to mention the number of existing citation templates, that we have a problem. The current situation is clearly untenable and unhelpful, and likely a barrier to new editors.

Regardless of everyone's strongly-held personal preferences, if we were starting from scratch I contend that we would never devise the current system, and that our objectives would be something like:

• Ease of data entry
• Provision of tools for ease of data entry
• Tools accepting citations in other formats
• User preference for displayed style (including the ability to turn of display of references altogether)
• Speedy rendering
• Ease of data entry
• Standardisation of templates, to facilitate bot-related tasks
• Emission of machine-parsable metadata
• Ease of data entry

The eminently sensible proposal, above, allows for all of these, particularly ease of data entry. There is no reason why tools cannot be written, which allow even the most obstinate academics to write references in their preferred arcane style, and then render them in the editing in a standard way - Zotero already has similar abilities and could easily be adapted, for instance.

I therefore slap with a trout a number of the above commentators, who resist moves to make Wikipedia easier to edit and easier to read and ask that everyone stop bickering or finding (or inventing) obstacles and work towards a solution that satisfies everyone's needs. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 11:29, 22 August 2010 (UTC)

Time to get this show on the road. Cheers, Jack Merridew 11:54, 22 August 2010 (UTC)

Oppose for all the reasons already stated. This is the third or fourth page Jack has started this discussion on recently, and he failed to gain consensus on any of them. Citation templates add significantly to load time and citation clutter, and WP:CITE says they should not be added to individual articles against consensus. Even ignoring templates, there's no consensus to impose any central citation style on Wikipedians, for all the obvious reasons. Every attempt to do so has failed, because editors want to retain that little bit of freedom. SlimVirgin ^{talk|contribs} 16:31, 22 August 2010 (UTC)

The lengthy RfC at Talk:Halle Berry#citation style has prompted me to comment here. To complement Andy's contribution above, I believe the discussion there contains the following points:

1. The issue of list defined references is completely separate from the question of whether references are made in plain text or by using a template.
2. Citation templates organise reference data in a way that makes maintenance by automated tools easier.
3. Citation templates may make it easier for maintenance editors to keep a consistent style of references in articles.
4. Templates take a disproportionate time for the server to process; this may present an unacceptable barrier to editing, particularly in a large article with numerous templates.

If I've missed or misrepresented any points there, please correct me.

Although I'm firmly convinced that citation templates add value to an article, I can't support further encouragement of their use while the load time issue exists, especially in large articles. The enhanced tag proposal above seems to offer all the advantages of citation templates, with load times indistinguishable from plain text references. Serious consideration needs to be made of this implementation, as soon as possible. --RexxS (talk) 16:55, 22 August 2010 (UTC)

SV, you appear to be opposing something for the sake of it; or at least without reading it. I've already pointed out to you elsewhere today that I'm explicitly not proposing to "impose any central citation style"; and I say above that one of our concerns must be "speedy rendering". And it was me, not Jack, who posted the above. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:07, 22 August 2010 (UTC)

Slim, you're missing the point of the proposal; it's not about what you're opposing, at all. This cite.php extension fixes the load time issue when using templates. It supports all styles and goes so far as to allow you to set your preferred style in your user prefs. The 'clutter' would be the naming of the elements? The title="whatever" in a ref-tag, not a templte? That's about database design. This is a database driven website. My intent here is to enable accesses to the details of references by tools. I don't care about where the commas and periods go and just what gets italics. Those are presentational details. Sincerely, Jack Merridew 21:52, 22 August 2010 (UTC)

:Strongly Oppose any attempt to make templates (or particular formats of references) compulsory or preferred (with people allowed to change articles to use them against consensus and without any sort of comeback) - they add nothing to human beings reading articles and make things harder for editors. There seems to be an obsession for always finding the most complicated way of doing things and always implementing the latest technical feature without any real thought about what it adds to the encyclopedia. Nigel Ish (talk) 17:00, 22 August 2010 (UTC)

Nigel, who is trying to make "templates (or particular formats of references) compulsory"? Or proposing to " allow people to change articles … against consensus and without any sort of comeback"? You're tilting at windmills. However, where templates are used they should not be the current muddled and set of options. Templates do add value for humans, both directly (by formatting consistently and potentially reader-defined way; and indirectly, by making the citation data usable by the tools those readers use, such as aggregators and indexers. The Wikimedia Foundation have spoken recently about their intention to make more of Wikipedia machine-readable in this way. The only personal attack here is yours; I invite you to remove it. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:14, 22 August 2010 (UTC)

1)"I therefore slap with a trout a number of the above commentators, who resist moves to make Wikipedia easier to edit and easier to read and ask that everyone stop bickering or finding (or inventing) obstacles and work towards a solution that satisfies everyone's needs. Andy Mabbett (User:Pigsonthewing);" - this seems like an attack on anyone who disagrees with the proposal to me - cetainly its not very helpful.

2)"There is no reason why tools cannot be written, which allow even the most obstinate academics to write references in their preferred arcane style, and then render them in the editing in a standard way " - ie. forcing references into templates and into a particular form. Oh and changing articles without warning or discussion to use templates happens already - this will only make it worse. Nigel Ish (talk) 17:32, 22 August 2010 (UTC)

1) Wikipedia:Trout. You might actually want to understand what a personal attack is, too. 2) Writing tools which allow != "forcing". What already happens is not what is proposed. Your claim that "this will only make it worse" is utterly without foundation. HTH. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:36, 22 August 2010 (UTC)

Since my presence here is clearly unwelcome, I am striking my comments and withdrawing from the discussion.Nigel Ish (talk) 17:53, 22 August 2010 (UTC)

Slim, to be fair, Andy started the discussion on this page, not Jack. I believe the only discussion Jack has started is the RfC at Halle Berry, in line with CITE's recommendation to determine consensus on any given page. The fact that opinion is split fairly evenly on the issue of using citation templates or plain text references indicates that a broader, centralised discussion is appropriate (as was suggested several times at that RfC). I still believe guidelines like CITE need to follow practice, and should this discussion lead a novel solution to the differences between the opposing viewpoints, then we could consider modifications to CITE from common ground. I do agree that it's important to allow editors freedom to contribute in the manner they desire, and would not wish to see any imposed central citation style either. Any solution will have to allow for the presentation of references in "modified-API" or "Vancouver" or any other commonly used style on Wikipedia. Let's see how that can be done. --RexxS (talk) 17:18, 22 August 2010 (UTC)

Yes, I apologize for saying Jack started this discussion. But Rexx I think you will find that CITE does follow best practice. During any of the discussions about this over the last few years—on article talk, policy talk, FAC talk, or here—there has never been consensus to force particular citation styles or methods on editors. In fact, I think quite a few would leave if we were to try to do that, or simply ignore us. Anyway, I think I'll follow Nigel and withdraw from the discussion now, because the forest fires of the last few days have not been productive. SlimVirgin ^{talk|contribs} 18:19, 22 August 2010 (UTC)

Perhaps they haven't been productive because of the repeated "force particular citation styles or methods on editors" straw-men? Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 18:26, 22 August 2010 (UTC)

Slim, I think you've missed the gist of this proposal; you're opposing other things. Please take a fresh read of this. Sincerely, Jack Merridew 21:55, 22 August 2010 (UTC)

My main beef with this and similar proposals is that they raise the bar for any newbie and that they seem like part of strategy, intentional or no, to sneak in a universal ref standard through technical improvements (no matter what proposers themselves say, this seems to be what a lot of supports are saying). Load times and performance issues seem to me like marginal problems on the whole. I have a few questions about the points mentioned above, though:

1. How is definition of references actually considered useful? What percentage of readers, not just users, would realistically be able to define references? How many would actually know the difference between various reference standards? How many would even care...?
2. What kind of machine-reading functions do we have in mind?
3. How would the code look like for shortened notes and explanatory notes (separate or not), as in Benjamin Morrell, Norton Priory or B-movie?
4. Does the proposal assume that increased use of code is unproblematic for newbies?

Peter ^Isotalo 00:39, 23 August 2010 (UTC)

1. I'm not sure what you're asking here. In particular, I'm confused by why you think the readers will need to know the differences...?
2. Consider this: We might in the future (for instance) be able to generate a Javascript form which would take a given reference (using the <ref ... >) and allow people to input the data themselves, which is then inserted into the article. Now isn't that a sexy thought? Suddenly, Wikipedia references are easy! There are other benefits, such as (purely for example) being able to run a bot to see how many times we cite a particular author.
3. As with the citation templates, you only need to include the information you want to, but without the price of the citation templates, which is the increased load time (and extra characters). Further, you can now use whatever citation style you want (though I suspect it would be[come] good practice to define the citation style at the bottom of a given article using <references/> or T:Reflist).
4. It's actually a slight decrease, as we've removed the curly brackets which denote a template. Also, see point 2. It actually opens a pathway to being able to make editing easier. Now that's a goal I'm sure everyone would agree to be a fine one. --Izno (talk) 03:00, 23 August 2010 (UTC)

"My main beef with this and similar proposals is that they raise the bar for any newbie" How so? They can stil enter plain text, or URLs as at present. They will also be provdied with a form to fill in, like the present citation entry tool, with a field for each parameter.

"they seem like part of strategy, intentional or no, to sneak in a universal ref standard through technical improvements". So much for assuming good faith. The proposal is for a single method of internal storage templated references (Wikipedia is, after all, a database), to replace the current morass of templates. The output, seen by the reader, will be variable, either by using a switch in the template, or by reader-preference.

1. Those who care - some, clearly, do - can set their preference. Anyone else will get the local default style for the article, as at present.
2. Our references are already machine-readable. See Zotero and COinS.

Izno covers 3 & 4 well. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 10:16, 23 August 2010 (UTC)

1. To clarify: Readers of encyclopedias are generally not well-served by being offered too many options. Currently, options are fairly complicated to define and you pretty much have to be a registered user to do it. Implementing the option to switch reference formats is an extremely specific type of choice. Most readers wouldn't know Vancouver from Chicago or Oxford if it bit them in the ass (and I can hardly blame them) and most of those who did would most likely not care enough to flip a switch to change it. Would it actually be worth the time and effort?
2. Somewhat conflicting answers there. If we're already machine-readable, why include it as a benefit of this proposal?
3. I'm still not sure what this proposal would look like if I tried to write out, say, "Bamford (1973), p. 224". And does the last part of Izno's answer suggest I would have to add additional code at the bottom to "define" a standard I've already chosen? Does it mean that we'd have to add more information than just "Name (year), page" to notes? And what about comments that aren't refs?
4. I already edit without citation templates. I simply write the note exactly as I like it to appear on screen—lickety split. Is this what we consider to be more complicated to understand for the average (non-geeky) newbie?

Peter ^Isotalo 15:19, 23 August 2010 (UTC)

1. Most readers will never see the choices (just as they currently don't for date formatting or thumbnail sizes; however, they will be there for those who do need them; offering a grater degree f choice than at present.
2. It's a feature of this proposal. It's a benefit of using templates over plain text, which will hopefully occur if we lower the barriers to those editors who wish to do so.
3. You would write hat you do now, in the way you do now, if you still wish to. Hopefully, the reference entry tool will be attractive enough for you to switch to using it; but that'll be your choice.
4. As 3.

The proposal is to consolidate all the existing templates into one, easy to use, replacement; not to enforce the use of that, nor any particular style. Andy Mabbett (User:Pigsonthewing); Andy's talk; Andy's edits 17:13, 23 August 2010 (UTC)

Peter asks the pertinent question (3) about how shortened footnotes or harvnb would be coded, giving three examples. The answer is pretty much the same as now, as I understand it, with the exception that the full citation (e.g. in Bibliography in Norton Priory) would not use the {{Citation}} template, but instead use an enhanced tag:

• <citation last="Brown" first="Fraser" last2="Howard-Davis" first2="Christine" year="2008" title="Norton Priory: Monastery to Museum. Excavations 1970–87" publication-place="Lancaster" publisher="[[Oxford Archaeology|Oxford Archaeology North]]" isbn="987-0-904220-52-0" />

which is the same as <reference ... />, except that it does not generate the ^[1] but simply outputs the reference in place, in the same way that a citation template does without the <ref></ref> tags around it.

So, not much difference, it appears. However, I note that Norton Priory takes the server 10.5 secs to generate an edit preview. Replacing templates with enhanced tags produces a considerable speedup (which is greater for bigger pages). Great Fire of London is a small page with only three templates; replacing them with enhanced tags halves the time for the server to create the page. Looking at B movie, it is 109 KB in size and has no citation templates. It takes the server about 5 sec to create the page. If citation templates had been used, that might be nearer 50 sec, at a conservative estimate, but enhanced tags would not suffer from that problem.

That's not to say I'd want to make anyone use enhanced tags in B movie, but the option would be there for anyone who prefers it in a similar article, and it wouldn't prevent other editors, like Slim, from being able to contribute. Isn't that a better position than not being able to edit Canada simply because every edit preview takes at least 25 sec to load? --RexxS (talk) 18:26, 23 August 2010 (UTC)

It took me several years before I finally copied someone's "toolkit" to a page and now copy those refs/cites, as appropriate into an article. But before that I used <ref>[http://www.someplace.edu/or/other Someplace]</ref>. That usually worked. More experienced users weren't territorially happy with that, but it was verifiable and could be "touched up" by editors with sophisticated tools.

So, it seems to me that we do have an "entry level" footnoting ability. We just need to explain it to newbies. (I know, it's below in the edit page, but that alone may not be enough). Student7 (talk) 21:41, 24 August 2010 (UTC)

Just read Izno's comment about easy generation. That could be neat - pointing to a reference and (say) right clicking or something and have it generate a formatted footnote in a preselected place in the document. I'm assuming that we are talking about a substantial amount of time for somebody to come up with this easy-to-use tool. I have a suggestion on how to arrive at an interface with newbies. Student7 (talk) 10:49, 25 August 2010 (UTC)

There have been tools to simplify citation creation for a while. The RefTools gadget that you can enable in preferences, Citation templates generator, and Diberri's medical template filler are three that immediately come to mind. There's also Zotero, available as a Firefox addon, so I've little doubt that the demand for help with creating references will be met, no matter what format we use. --RexxS (talk) 14:41, 25 August 2010 (UTC)

True, the tools exist now, but none of them are exactly easy to find for the common reader/occasional editor. One is on the toolserver, one I didn't know existed (and is specific to one field), and one is a gadget buried in preferences. I was talking about being able to insert a reference at the edit screen, without leaving it or being required to use a 3rd party tool. --Izno (talk) 15:20, 25 August 2010 (UTC)

Well, the fact that the buried gadget exists shows that the concept is feasible (why isn't that turned on by default anyway?). In my experience, the common reader/occasional editor who adds a source is likely to just give something like a bare url, or the name of a book, and a wikignome will eventually come along and add more details. While it would be nice to encourage occasional editors to give more detailed refs, I don't realistically see these sort of tools as being aimed mainly at them. --RexxS (talk) 15:35, 25 August 2010 (UTC)

Tools aside, the most accessible entry-level option is a simple in-text description of the source, e.g. "In the New York Times issue of 25 August 2010, Sharon Lafranier's article Chinese Airport Drew Safety Concerns Before Crash explained that the new Lindu airport was considered by China Southern Airlines to be unfit in principle for night flights." More experienced editors can always transform that to a ref based citation, tools or no. We must remember first that anything that complicates the life of a novice editor has a long-term cost of fewer editors and second that any honest citation, however it may be formed, is a valid contribution to be encouraged. LeadSongDog come howl! 15:59, 25 August 2010 (UTC)

It seems to me that the whole discussion of how to do citations is complicated by conflicting criteria of what should be optimized. (E.g., what is easiest for new editors to create may not be easiest to read, or to verify, or modify. Using a citation/cite template, though initially somewhat intimidating, may be easier than trying to manually create a consistent citation style. Even a "a simple in-text description of the source" gets complicated when the source is cited a second time, and the editor has to decide between copying and modifying, or finding some way of referencing the first citation.) Should we identify the various desirable criteria, and how they should be balanced? - J. Johnson (JJ) (talk) 19:41, 25 August 2010 (UTC)

The novice isn't trying to reach FA-status, just make a contribution. We need to encourage that so far as we can, within the WP:V constraint. The need to do that still allows that, having verified the novice's source, the more experienced editors can fix up the citation to meet the really rather simple WP:FA criteria in whatever form the article's editors have settled on. LeadSongDog come howl! 21:46, 25 August 2010 (UTC)

Yes, but... Should our approach be that making citations as easy as possible for newbies is foremost, and making them good is secondary? Or should our goal be to have good (however defined) citations, which we then make as "easy" (accessible?) as possible for newbies? I confess that do not hold with maximal easiness as a primary goal for anything, and having done some citation cleanup I favor higher standards. But I also think that, with proper tools and a certain amount of education, even newbies can be brought around to better work. - J. Johnson (JJ) (talk) 21:10, 26 August 2010 (UTC)

Absolutely. Anything that gets them started citing their sources is helpful and should be encouraged. Once they get their feet wet, sure, we can gradually educate them further, but let's not drive away new contributors with elaborate technical requirements on how they cite their sources. Sooner or later they'll want to do better work. LeadSongDog come howl! 21:32, 26 August 2010 (UTC)

I hope we can avoid "elaborate technical requirements" in all cases. But should "making citations easier for newbies" always trump "making citations better"?

- J. Johnson (JJ) (talk) 21:57, 26 August 2010 (UTC)

That's a false dichotomy. Making citations easier for newbies does make citations better, and in no way precludes them being further improved by later, more experienced editors. LeadSongDog come howl! 19:54, 30 August 2010 (UTC)

I think we're looking in slightly different directions. For sure, it is possible for experienced editors to remedy the specific sins and omissions of newbies after the fact, which I think is how you are thinking. But having dived into some of those messes myself (and seen what rats nests some of them are) I am convinced it is more useful to fix the process and tools so that the newbies don't get into such messes in the first place.

What I am asking is: if some process is inherently buggy, and creates messes, are we better off making it easier? Or less buggy? Ideally we should do both, but what I am suggesting is that applying one criterion absolutely, without regard to other criteria, would actually be sub-optimal. - J. Johnson (JJ) (talk) 21:51, 30 August 2010 (UTC)

• Those of you in this discussion may be interested in seeing the new ref tag proposal in the village pump suggestions. Though it deals with a different aspect of citing other than copy retrieval, it is designed to be intuitively grasped by newbies. —CodeHydro 13:05, 27 August 2010 (UTC)

Should citation style be controlled at the level of the article?

Some what of a "wild blue sky" idea, as I dont' see there is currently any means or liklihood of implementation. But as a some what "theoretical" idea: would it be preferable if the citation style used in an article was specified in a way that made it applicable to all citations in article?

WP policy is that citation style should be consistent within each article. But citation style is currently controlled (if not done manually) at the level of each individual citation (by choice of template and parameters), which allows a hodge-podge of inconsistency. Now suppose we had some kind of generic citation template which collected the parameters and preprocessed them, but left output formatting to a post-process that implements what ever style is desired. A particular style (post-process) could be selected by adding a template to the article.

Would this (if we had it) be an improvement? - J. Johnson (JJ) (talk) 22:58, 22 February 2011 (UTC)

that was part of #Demo of specific proposal. Damned, Gold Hat (talk) 20:13, 16 March 2011 (UTC)

Full name, or initials?

A matter that has been vexing me: in regard of author and editor names, should we lean towards full names (as in "Smith, John"), or last name and initials ("Smith, J.")? Or (as suggested at WT:Scientific citation guidelines#Full names, or initials?), should we follow the style of the original publication, however inconsistent this may make the article? As to following the existing style in an article, what I have in mind is where no style has been established, or it is inconsistent. Suggestions? - J. Johnson (JJ) (talk) 20:51, 29 August 2011 (UTC)

Surely this depends on the differing conventions of different fields. In mathematics and many sciences, initial-only may be favoured, whereas in the humanities, full names are generally preferred. In both cases, there are bound to be exceptions with particular journals or book publishers. Other areas such as sports, cookery, or computer gaming no doubt have their own tendencies, but these will be best known to authors accustomed to writing in those fields.—Jerome Kohl (talk) 21:24, 29 August 2011 (UTC)

For as long as I can remember, WP:SCG has contained this statement: "Since Wikipedia is not a paper encyclopedia, references do not need to be as concise as they are in journals". I've always understood that to mean that where we have a piece of information available in full, there is no value whatsoever in abbreviating it. It is true that different journals may use differing formats for authors' names, but here on Wikipedia what does the reader gain if I truncate "Lambertsen, Christopher J" to "Lambertsen, C J"? The author's full name is quite often given on the actual article, (rather than when cited in a paper journal), and once I know that, why would I throw away useful information? How else would the reader know if "Belushi, J" was John or Jim? I really cannot see any value in using anything other than the full name if known, and initials until it is found. The idea that having a mixture of full names and initials is somehow 'inconsistent' is a hollow argument. If there is an argument that all author names must be reduced to initials when one first name is unknown, then an equivalent argument is that all author names must be removed if one name is not known. It should be clear that maximising information is more important than some worthless concept of 'uniformity' - we don't need to be levelling down to the lowest common denominator. --RexxS (talk) 00:35, 30 August 2011 (UTC)

If it works consistently, a printonly abbreviation would make sense. But how many editors would actually see that form to catch problems should they arise? LeadSongDog come howl! 13:52, 30 August 2011 (UTC)

I can't think of a compelling argument to not use full names, except where initials are the norm: T. S. Eliot, R. R. Tolkien, F. Murray Hamilton. ---— Gadget850 (Ed) ^talk 14:41, 30 August 2011 (UTC)

There are numerous cases where it is impossible to find the full names because the entire fields uses initials. Indian names also never list the given names. Circéus (talk) 15:51, 30 August 2011 (UTC)

Except for when the Indian names list the given names – for example:

• Patel, Dharmeshkumar N; Goel, Ashish; Agarwal, SB; Garg, Praveenkumar; Lakhani, Krishna K (2003). "Oxygen toxicity" (PDF). Journal, Indian Academy of Clinical Medicine. 4 (3): 234–7.

--RexxS (talk) 17:35, 30 August 2011 (UTC)

I vaguely recall an argument years ago that using initials in place of first names helps to conceal gender, and so might help prevent the instant rejection of female authors. Which might explain some cases I have seen where, in a given article, some author's names are full and others abbreviated.

So the position evolving here seems to be that there is no point in abbreviating a name any further than was given, and that in general we should use the form of the name preferred by the author. (I would add that expanding what an author has initialized is pointless pedantry.) Where a name is variously given full and initialized it is not wrong, and may be preferable, to use the longer form. We should perhaps require that a given author's name be given consistently in an article even if the sources use different forms. Is that generally satisfactory? - J. Johnson (JJ) (talk) 22:22, 30 August 2011 (UTC)

Why would female authors be instantly rejected, and by whom? SlimVirgin ^{TALK|CONTRIBS} 02:11, 31 August 2011 (UTC)

Good question. Apparently some folks (men?) do, and so (as I seem to vaguely recall) the argument of using initials to disguise gender. - J. Johnson (JJ) (talk) 18:10, 31 August 2011 (UTC)

I've never heard of such a thing in a modern academic journal. Would appreciate seeing a source. SlimVirgin ^{TALK|CONTRIBS} 19:52, 31 August 2011 (UTC)

What does "modern" have to do with it? Should we have different practice for citing old and new works? Converting from initials to full names, once begun, will be fraught with wp:OR potential. It should only be done where there are reliable sources for the attribution to the full name. For books, the Library of Congress "authorities" database is reliable. For pubmed indexed journals, the "ForeName" fields are often available in the xml records, though frequently these fields essentially duplicate the "Initials" fields, (possibly with added punctuation and whitespace). LeadSongDog come howl! 20:35, 31 August 2011 (UTC)

The argument I vaguely recall for using initials probably came out of the sixties, and I suspect you (SV) are younger than that. - J. Johnson (JJ) (talk) 22:57, 31 August 2011 (UTC)

[::::::::::::See this o.t.q. LeadSongDog come howl! 05:29, 1 September 2011 (UTC)

Interesting source: "Years ago, a friendly colleague told me his department was considering hiring a junior person in our field and asked me for a top candidate. After some thought I mentioned one of the best junior researchers in the field. His answer was 'But we already have a woman!'" :) SlimVirgin ^{TALK|CONTRIBS} 21:46, 1 September 2011 (UTC)

• This type of question should be asked at WT:CITE, not at CENT.
• J. Johnson has previously participated in discussions on this point, and apparently he didn't get the answer he wanted. The answer has not changed: Editors may choose any style they want including a style that spells out all names, a style that uses initials only, a style that follows the name given on each publication, or even a style that omits authors' names entirely (as, for example, legal citations routinely do). It is 100% free choice for the editors of the article. Wikipedia does not have a house style. We are not going to prescribe, or even vaguely recommend, a single answer to this question. WhatamIdoing (talk) 00:14, 31 August 2011 (UTC)

• I agree, there's no reason to recommend any particular format. SlimVirgin ^{TALK|CONTRIBS} 02:11, 31 August 2011 (UTC)

If WhatamIdoing would pay a little more attention to the original question he might see that I am inquiring "where no style has been established, or it is inconsistent". I asked -- in good faith, and without any definite position one way or the other -- whether there might be any suggestions or recommendations. (And I object to his imputation that I am fishing for any particular answer.) What I get is: "We are not going to prescribe, or even vaguely recommend, a single answer to this question." I am not acquainted with the prior history here, but if the antipathy is that strong perhaps there should be warning against even inquiring for a recommendation. - J. Johnson (JJ) (talk) 18:10, 31 August 2011 (UTC)

Find articles that cite a particular source

I want to see if there is code or a simple technique for searching all articles by their use of citation of source. For example, I wish to find articles that use Breitbart News as a resource, or what ever other sources I am interested in. Hannah Silverman (talk) 18:18, 28 January 2017 (UTC)

An interesting idea. One could simply search for, say "Breitbart News", but a straight-forward search would include all mention of same. I wonder if it is possible to search for terms within templates, but you'll need to find some expert to advise on that. Or try this: go to Advanced search and turn-off all name spaces but "Article", then search on "source: breitbart.com". I get 1,303 hits, including a lot where the citation was done without a template. Indeed, it seems to pickup on just "Breitbart", so you'll still have to determine which articles to exclude. ~ J. Johnson (JJ) (talk) 21:20, 29 January 2017 (UTC)

I think that sounds promising - Thank you. I will try it out and come up with a list worthy of review if not 100% complete. Hannah Silverman (talk) 17:40, 30 January 2017 (UTC)