Wikipedia talk:Citing sources

Bundling Considered Harmful

I'd like to point out that the idea of reference bundling, while not terrible per se, is being "blindly" applied in certain articles without regard to whether it's really necessary (not every instance of 2 refs in a row calls for bundling), and with no apparent consideration to what the ref list looks like after it's done (see the first example here—I'm not citing a specific "real" example so as not to call out a particular editor). As with anything, editors still need to think about what they're doing, and whether their changes benefit readers. - dcljr (talk) 05:06, 3 November 2015 (UTC)

Particular cases should be discussed on the Talk pages of the articles concerned. This page would be more suitable for discussing "bundling" generally. Anyone interested? ~ J. Johnson (JJ) (talk) 00:23, 4 November 2015 (UTC)

Mildly differ with both the above. That is, we should have a specific example to see what the problem is. It is the edit not the editor, in question. On bundling in general: it is very useful. Especially when the footnote supports more than one sentence. I strongly prefer titled bundles (a descriptive name at the top like footnotes 1, 7 and 13 here) instead of naked bundles (no indication of what the cites have in common, like footnote 29 here). Titled bundles are especially needful on Wikipedia since text gets shifted, footnotes get orphaned. Without the title you can't be sure all the cites in the bundle still support a revised sentence or paragraph. Now, editors with habits leftover from print publications, books and articles, don't see the need for titles on bundled footnotes. The habitual assumption is the text over the footnote suffices. But Wikipedia is a different animal: footnotes on Wikipedia don't stay glued to their text. In sum: yes to bundling, and preferably titled bundles.ElijahBosley ^{(talk ☞)} 13:42, 4 November 2015 (UTC)

OK, well, what do people think about this way of bundling? (Yes it's about the edit and not the editor, but this editor has changed multiple articles in the same way.) Compare the reflist before and after the changes (in particular, see "after" notes 9, 15, 22…). I suggested a different way (yes, I know it was a naked bundle) and pointed out it doesn't work with named references. The other editor has tried doing things differently, apparently with mixed results. If someone more familiar with bundling could weigh in on this, that would be great. - dcljr (talk) 03:13, 5 November 2015 (UTC)

Competing formats for cites? Is that the issue? Rather than bundle/not bundle? I note that with naked bundles, using quotes in the cite is wonderfully helpful as in footnote 2 in the example given above. The second cite in the footnote offers the quote: "R is also the name of a popular programming language used by a growing number of data analysts inside corporations and academia. It is becoming their lingua franca ..." That tells the reader what the footnote buttresses, in a way that a naked bundle would not.ElijahBosley ^{(talk ☞)} 14:49, 5 November 2015 (UTC)

The solution you want is shortened footnotes. {{sfnm}} supports the inclusion of multiple sources in a single footnote. This style allows multiple authors and exact page numbers all rolled into one superscripted number. That same source can be reused, yet with different page numbers. The problem is that all the citations would need to be converted to the shortened footnote format, which is uncommon and would confuse most editors. So yes, there is a better solution than using {{refn | to bundle citations, however since that solution looks like {{sfnm | 1a1=Phillips-Fein | 1y=2009 | 1p=115 | 2a1=Hamowy | 2y=2008 | 2p=217 | 3a1=Perelman | 3y=2007 | 3p=64 | 4a1=Schneider | 4y=2009 | 4p=47 | 5a1=Mirowski | 5a2=Plehwe | 5y=2009 | 5p=285 | 6a1=Olson | 6y=2009 | 7a1=Lichtman | 7y=2008 | 7p=160 | quote5="… going so far as to help Mises publish his Magnum Opus Human Action …"}} , that solution also has baggage. Abel (talk) 18:27, 5 November 2015 (UTC)

Yes, EB, the issue is how bundling is done rather than whether or not to bundle (although personally, I think bundling is a solution in search of a problem, but whatever...) OTOH, the issue is not simply "competing formats"; I think the {refn} way of doing it is actually fundamentally confusing to readers, since it is unlike any other common method of footnoting in use around here. (IOW, the objection is to a footnote containing merely a list of bracketed numbers, which are references to other footnotes — that's just crazy [and these numbers are also shown in the tooltip when hovering on the note number, which is even worse]). The shortened notes approach, if bundled as above, would be confusing to more editors, but I care a little less about that (unless, of course, it leads to many bad edits). Finally, the objection to naked bundles seems to me to be merely an aesthetic one (it doesn't bother me, but I would not object if someone wants to take the time to add "titles" to such bundles). - dcljr (talk) 20:22, 5 November 2015 (UTC)

Again, the solution you want is shortened footnotes, for example: "... and individual lectures." Abel (talk) 03:38, 6 November 2015 (UTC)

Notes

References

• Phillips-Fein, Kim (2009). Invisible Hands: The Making of the Conservative Movement from the New Deal to Reagan. New York: W. W. Norton. p. ii, 27, 52, 60, 86, 101, 115, 116, 124, 149, 167, 265, 270, 285, 286. ISBN 978-0-393-05930-4. {{cite book}}: Invalid |ref=harv (help)
• Hamowy, Ronald, ed. (2008). The Encyclopedia of Libertarianism. Thousand Oaks, CA: SAGE Publications, Cato Institute. pp. 62, 217, 221, 335, 416, 417. ISBN 978-1-4129-6580-4. LCCN 2008009151. OCLC 750831024 http://books.google.com/books?id=yxNgXs3TkJYC. {{cite encyclopedia}}: Invalid |ref=harv (help); Missing or empty |title= (help)
• Perelman, Michael (2007). The Confiscation of American Prosperity from Right-Wing Extremism and Economic Ideology to the Next Great Depression. New York: Palgrave Macmillan. p. 64. ISBN 978-0-230-60046-1. {{cite book}}: Invalid |ref=harv (help)
• Schneider, Gregory L (2009). The Conservative Century: From Reaction to Revolution. Lanham, MD: Rowman & Littlefield. p. 47. ISBN 978-0-7425-4284-6. {{cite book}}: Invalid |ref=harv (help)
• Mirowski, Philip; Plehwe, Dieter (2009). The Road from Mont Pèlerin: The Making of the Neoliberal Thought Collective. Cambridge, MA: Harvard University Press. pp. 15, 19, 21, 53, 156, 190, 196, 243, 281, 284, 293, 387, 397, 410. ISBN 978-0-674-03318-4. {{cite book}}: Invalid |ref=harv (help)
• Olson, Wayne (September 28, 2009). "An Inside Look at the Foundation for Economic Education FEE" (Interview). Interviewed by Pete Eyre. {{cite interview}}: |access-date= requires |url= (help); Invalid |ref=harv (help); Unknown parameter |city= ignored (|location= suggested) (help); Unknown parameter |program= ignored (help)
• Lichtman, Allan J (2008). White Protestant Nation: The Rise of the American Conservative Movement. New York: Grove Press. pp. 160, 171, 173, 206. ISBN 978-0-8021-4420-1. {{cite book}}: Invalid |ref=harv (help)

Okay, someone please clarify for me: what exactly do we mean by "bundle" and "bundling"? Is a string of footnote links in the text (e.g.: ^[1][7][13]) itself a bundle? Is "bundling" putting all of the same citations into a single note? And (EB?) what are naked and titled bundles? ~ J. Johnson (JJ) (talk) 23:56, 5 November 2015 (UTC)

Yes: bundling lists several cites in one foonote, usually with bullet points. Instead of a separate footnote for each cite which leaves text looking something like this^1234567. Examples of bundled footnotes are footnotes 1, 7 and 13 here. Each has a descriptive name at the top. That is the title. And while I am at it let me give a pat on the back for the excellent answer by Abel above.ElijahBosley ^{(talk ☞)} 13:24, 6 November 2015 (UTC)

Thank you. I see several variations here:

• the basic bundling of multiple citations,
• bundling in a list format (versus "serial", see below),
• a list format with or without bullets, and
• a list format with or without titles.

There is also a very basic variation of whether what is bundled are the full citations (as seen in the example fn. 1) or the short cites (aka shortened footnotes). Here is an example of a serial (non-list) format, which also uses short cites:

• Lictman 2008; Phillips-Fein 2009; Mirowski & Plehwe 2009.

So there is basic bundling, and there are a number of variations. All of which merit discussion. ~ J. Johnson (JJ) (talk) 23:48, 6 November 2015 (UTC)

Yes, there are many ways to bundle citations. The rub comes from none of those variations being wrong. All of those methods are equally correct. All editors are free to prefer one method over the others, but there is no policy mandating one method be used beyond respecting the method already in place if that situation exists. Abel (talk) 01:36, 7 November 2015 (UTC)

I don't know about "equally correct", or even useful, but I'll allow that none is mandated or prohibited. Yet the contrast of these styles is so jarring that concurrence towards a more standard style would seem for the better. At the very least editors could be more conscious of what they are doing, and why, rather than sticking with whatever they have always done solely because they have never considered the alternatives. ~ J. Johnson (JJ) (talk) 23:20, 7 November 2015 (UTC)

Simply wrapping existing references in a {refn} template is wrong, and should be prohibited discouraged in this guideline. - dcljr (talk) 01:57, 8 November 2015 (UTC)

We understand that you do not like how that one style looks, however, that does not change the fact that all of those styles are valid choices. I agreed with JJ that the style options are radically different and having all of Wikipedia pick one style would be a large improvement, however, that does not seem likely to happen anytime soon. You might as well accept that regardless of how much you dislike how that one style looks, there is currently nothing wrong with that or any other method.

"While citations should aim to provide the information listed above, Wikipedia does not have a single house style, though citations within any given article should follow a consistent style."

So yes, the shortened footnotes example above does everything you want and looks the way you like, but it is not okay to just go around changing other methods to that method as it is no more valid than the existing method regardless of how much you prefer how it looks.

If and only if you gain consensus for the change on the article's talk page may you then change the citation style to the method that you prefer, for that one article. So if that one style method bothers you that much, just start a new topic on the talk page and once you have consensus, convert the article to shortened footnotes. Abel (talk) 16:46, 9 November 2015 (UTC)

Hear, hear. Though I should clarify that what I favor is not picking one style, but moving towards a more standardized usage with fewer bizarre variations. At the very least we should be more aware of the advantages and disadvantages of different "styles". ~ J. Johnson (JJ) (talk) 23:42, 9 November 2015 (UTC)

Interesting viewpoint you express there, Abel, given that you are going around changing consecutive refs in articles to {refn} bundles with no prior seeking of consensus. In any case, you'll notice if you look through my edits that I have not been changing the referencing style in any articles, so I'm not sure who your latest remarks are actually directed towards. More importantly, though, I must push back on the idea that somehow the {refn}-bundled "style" is an equally "valid" one. Where in this guideline or, indeed, anywhere else in Wikipedia's policy or guideline pages, does it mention that simply wrapping multiple consecutive refs in a {refn} template is an acceptable "style"? - dcljr (talk) 09:05, 10 November 2015 (UTC)

The refn bundles maintain the existing style rather than changes the style, which would require consensus. Abel (talk) 16:20, 10 November 2015 (UTC)

I reject the idea that bundling (generally) is invalid, or harmful. Still, "style" tends to be applied broadly. And bundling, especially when done in a list format, is not a minor change. Where anyone objects I think it is quite inadvisable to proceed without obtaining consensus. ~ J. Johnson (JJ) (talk) 00:31, 11 November 2015 (UTC)

Having seen people wholesale delete valid sources just because they find^{[1][2][3][4][5][6][7]} unsightly, I have to agree. Bundling is the only thing that I have seen that makes those people happy. Is bundling still a giant mess with a whole host of its own problems? Absolutely. That does not change the fact that bundling allows people to maintain verifiability while appeasing people who demand readability. Ideal solution? No. Best we have at the moment? Yes. Seems like a "never let the perfect be the enemy of the good" kind of situation. Abel (talk) 14:02, 11 November 2015 (UTC)

I think I generally agree with you. While allowing that some editors think a long line of note-links is quaint, I think bundling in some form is beneficial. (Details obviously TBD.) But there is no advantage to riling up others with mass changes they are not ready to accept. It would be better to sort out the objections, and what a generally acceptable form would be. ~ J. Johnson (JJ) (talk) 23:18, 11 November 2015 (UTC)

What you describe is the current policy, stick to whatever style exists until someone asks for a change on the talk page. Abel (talk) 01:00, 12 November 2015 (UTC)

Just learned that you can link directly to the References sections so for anyone placing a request on a talk page the following might be helpful: "Please chime in with your preference for the current or the proposed formatting."

Please chime in with your preference for the [[Special:Diff/690200494#References|current]] or the [[Special:Diff/690200296#Notes|proposed]] formatting.

Abel (talk) 01:13, 12 November 2015 (UTC)

Continuing

Am I crazy or has no one actually directly addressed the objections I have raised about merely wrapping multiple refs in a {refn} template? (Note that I am not talking about naked vs. titled bundles, shortened refs, changing "citation styles", or whether bundling should be banned in all its forms. Other commenters have raised these issues, not me.)

The objections are:

1. it results in odd, potentially confusing footnotes (see footnote 9 in the References section here),
2. it defeats the usefulness of "mouseover" checks of the citations (e.g., mouseover the ^[9] in the article text here), and
3. it does not follow any example of bundling given here or on any other guideline or policy page that I know of.

It's fine to discuss better ways of doing things, but I keep objecting to a specific thing and Abel keeps replying in a way that seems to avoid addressing the specific thing I am objecting to. So, Abel, do you still think that the way you changed the refs in the first link in this comment is a valid, acceptable way of bundling refs, and if so how do you respond to my objections listed above? (And if anyone else has responses to these specific objections, of course, I am interested in hearing them.) - dcljr (talk) 06:26, 16 November 2015 (UTC)

We all understand that you do not like how some bundled citations look. The part that you leave out of the example was how, to humor you, I went to the trouble of changing the bundled citations to the bulleted format that you prefer, yet could not be bothered to do yourself. That does not change how everyone agrees that there is no one correct method for bundling citations aside from maintaining the current citation style until editors agree to a new style on a talk page. Abel (talk) 21:46, 16 November 2015 (UTC)

I "could not be bothered to do [it myself]" because (1) I literally did not have time to spend on it, and (2) it would be undoing changes you had just made. If an editor makes problematic edits to an article, they really should take the responsibility to fix the problems they have introduced. And might I point out that you still have not addressed any of my objections in any substantive way. - dcljr (talk) 06:22, 19 November 2015 (UTC)

Bulleting the bundled citations did not undo anything, which is the point. All it did was change the appearance of the citations. Abel (talk) 17:42, 19 November 2015 (UTC)

Both of you, cool it. If you two want to sort out some "problematical edits to an article", please do so at the appropriate Talk page. Or if you have some interaction issues, try each others' Talk page. (Though I would suggest finding a moderator, lest you just wind each other up.) Abel, I gather dcljr objects to your bulleting the bundled citation. If that was done just to illustrate some point, would you object to undoing that, at least for the duration? ~ J. Johnson (JJ) (talk) 23:58, 19 November 2015 (UTC)

The opposite is true. The bulleting was to appease dcljr, yet was not enough. dcljr does not like how [4][5] looks when inside [3]. The solution to this is {sfn} and {sfnm} which bundles citations as a part of how the template works, so the unwanted appearance cannot happen. Abel (talk) 01:44, 20 November 2015 (UTC)

A solution for what? So far all I see as any kind of problem is the suboptimal interaction between two editors. So I'll put to you the same question I put to dcljr: what is your issue here? Can you two come up with a mutual statement of what problem(s?) or issue we should be addressing here? ~ J. Johnson (JJ) (talk) 23:27, 20 November 2015 (UTC)

I do not have an issue. I have never claimed to have an issue. The sfnm template is one solution to how dcljr does not like what you called "a ref within a ref" and I called "how [4][5] looks when inside [3]". Your "a ref within a ref" label makes more sense as it is more clear. I think dcljr called that a "naked citation" or something. Abel (talk) 02:10, 21 November 2015 (UTC)

(interpolated comment) No, someone else brought up naked bundles. In my "20:22, 5 November 2015 (UTC)" comment I called it "a footnote containing merely a list of bracketed numbers, which are references to other footnotes". More importantly, though, I cited and linked to specific examples of what I'm talking about in my comment of "03:13, 5 November 2015 (UTC)" and "06:26, 16 November 2015" (more about which in my comment below, posted at the same time as this one). - dcljr (talk) 12:14, 6 December 2015 (UTC)

Okay, I am glad to hear that. It seems to me that the two of you were more into "who did what" than the general merits and applicability of bundling. It might be useful to have an article in hand as an example, and even your specific edits as specific instances. But hopefully that is past, and we can discuss bundling without fighting about it.

You mentioned a solution. Should we look at whatever problem or issue you think it solved? ~ J. Johnson (JJ) (talk) 23:31, 21 November 2015 (UTC)

Happy to supply one. So, due to the nature of how the sfnm template works you cannot get a ref within a ref. Abel (talk) 06:53, 22 November 2015 (UTC)

The exact same citations in sfnm format look like this. Abel (talk) 06:53, 22 November 2015 (UTC)

Notes

References

1. Lopez, Mary Stachyra (2014-06-16). Centreville and Chantilly. Arcadia Publishing. ISBN 978-1-4671-2023-4. {{cite book}}: Invalid |ref=harv (help)
2. Brent, Chester Horton (1946). Descendants of Col. Giles Brent, Capt George Brent and Robert Brent, Gentlemen. Priv. print. by the Tuttle Pub. Co. {{cite book}}: Invalid |ref=harv (help)

But {{sfnm}} is not intended to put a ref within a ref. (Nor, for that matter, does talkquote.) It creates a footnote (implicitly within ref tags) containing short cites, which link to full citations provided elsewhere. With {{refn}} you are putting the full citation into the footnote (ref in a ref). Same citations, but totally different approaches. ~ J. Johnson (JJ) (talk) 00:33, 25 November 2015 (UTC)

---

Sure, let's address your objections. But appears you are not concerned about bundling as such, but this particular form of bundling "multiple refs in a {refn} template", where the footnote contains links to other footnotes. That seems excessively clunky. Does anyone have any points in support of such an arrangement? ~ J. Johnson (JJ) (talk) 23:47, 16 November 2015 (UTC)

Yes, I thought I had made it abundantly clear that my objections concerned this particular way of bundling. (In fact, I have said it five times.) Does anyone else have any useful commentary in favor of or against this particular method of bundling, especially in light of the 3 objections about it raised above? - dcljr (talk) 06:29, 19 November 2015 (UTC) [Thank you, JJ, for your specific comment about it. - dcljr (talk) 06:31, 19 November 2015 (UTC)]

I am not really interested in how many times you have previously discussed something, and certainly not in debating it. And, no, your objections are not "abundantly clear" when you mix them up with complaints about others' edits, and are not adequately specific. E.g., from Abel's comments I would think the issue is about bulleting. So tell me (I'm new here): Is that your issue? Or is your issue just "refs in a ref"? ~ J. Johnson (JJ) (talk) 23:59, 19 November 2015 (UTC)

(interpolated comment) How can you ask that when I just answered your Nov 16th question about what I was objecting to? (And BTW, the "times" I have "previously discussed" this are all on this very page, in this very dicsussion — in which you have been involved from the beginning.) - dcljr (talk) 12:14, 6 December 2015 (UTC)

@Dcljr: I am interested in hearing from you. Can you identify the issue that concerns you? ~ J. Johnson (JJ) (talk) 00:29, 25 November 2015 (UTC)

@J. Johnson: Look, I hate to say this, but I have a hard time believing you have actually read my comments and followed the links I have provided. If you had, you would have seen from the very beginning that my "issue" is having "refs in a ref", as you say. (And BTW, my objection could not be about bulleting, since I cited — in my comment of 03:13, 5 November 2015 (UTC) — one of my own edits where I used bulleting. This is why I have been getting progressively more frustrated here, because every time I cite specific examples of what I'm talking about, the replies either talk about completely different issues or — in your case, JJ — ask for an explanation, example, or answer to a question, when I have just provided same.) So, before we go any further, please reread my comment of 03:13, 5 November 2015 (UTC), which begins "OK, well, what do people think about", paying special attention to the first 3 links; then read my comment of 06:26, 16 November 2015 (UTC), which begins "Am I crazy", paying special attention to the 1st, 3rd and 4th links. That is where you will find my explanations of and specific examples of what I am objecting to. - dcljr (talk) 12:14, 6 December 2015 (UTC)

Yes, I read your comments, and followed your links. It made my head spin then, and all your underlining and all your bolding does not make anything clearer. Please pause a moment and listen: bitching about how many times you have said something (your prior comment, on the 19th) is not an explanation, nor any kind of "points in support" (as I requested). Likewise, your comment of 03:13 5 Nov. is mostly waving your hand around without any definite statement of what we should make of what you point to. Your following comment at 20:22 is more substantive, but you flailed away at so much it still makes my head spin. And now, when I asked a simple question ("Can you identify the issue that concerns you?"), which could be answered very simply (as simply as "yes" or "no"), you tell me to scroll back to a previous comment and find the 1st, 3rd and 4th links (does that include the note-link?), which don't really explain, and even suggest that you object to the R programming language. While you may think you were abundantly clear, sorry, you were not. Your frustration arises because your explanations, etc., are not clear and definite. Allow me to help with this.

What would help here is if we could lose all this prior history and start afresh with a simple statement of what issue (or issues) you have in regard of bundling. Even if it is incomplete, fine, we can incrementally improve it to a point where a simpleton like me can understand what you object to. Then we can work on why. Okay? ~ J. Johnson (JJ) 23:31, 7 December 2015 (UTC)

It appears we are restarting at your new section, #Bundling revisited (below). ~ J. Johnson (JJ) (talk) 22:12, 9 December 2015 (UTC)

Discussion about generalizing |editor= to support other roles

See Help talk:Citation Style 1#Contribution rather than Others (I can't make the section link work properly here!)

The idea is to deprecate the |editor= set of parameters and deprecate using |others= for other types of contributor, such as prefacers, translators, illustrators etc. The |author= family of parameters plus a new |role= parameter would be used instead. All input welcome. Aymatth2 (talk) 17:14, 4 November 2015 (UTC)

See Help talk:Citation Style 1/Archive 9#contribution= rather than others=. ~ J. Johnson (JJ) (talk) 23:39, 7 November 2015 (UTC)

Typos in web titles

Are we meant to correct typos in the titles of web-based sources, or leave them exactly as they are? (besides changing ` to '). I've always thought it's the latter, for absolute accuracy; likewise not changing hyphens to dashes and vice versa. Mac Dreamstate (talk) 16:07, 12 November 2015 (UTC)

It is normal to adjust titles to follow the conventions followed by the publication (Wikipedia) that is mentioning some other publication. So if a book-length source writes its title as ONE FOR ALL AND ALL FOR ONE we would write it One for All and All for One in the body of an article. If we were putting it in a citation, we would follow whatever citation style was being used in the article. For example, if the article used APA style we would write One for all and all for one. The citation templates don't have any rule about how to write titles, so do whatever the previous editors of the article have done. Jc3s5h (talk) 16:42, 12 November 2015 (UTC)

How about this ref, which has a typo in the title: "W.A.S.P. Announces 'The Crimson Idol' 15th-Anniverary Tour". I've been under the impression that such mistakes are to be left as is, since they are not being used in prose. Granted, it's a glaring typo, but that is how the article authors decided to leave it—for the sake of accuracy, surely we should reproduce it as such? Mac Dreamstate (talk) 16:51, 12 November 2015 (UTC)

The idea underlying the metadata included in citations is to make it possible for readers to verify the information in them. It becomes especially important that the data be exactly as given in the source in the event that the link goes dead due to the source page being moved or deleted, since otherwise Google searches may not find the correct result. So yes, when it comes to article titles (chapter titles, other short work titles, etc.) leave non-punctuation errors as found in the source. If there's an error in a source's title that really bothers you, you can always add a [sic]. —GrammarFascist _contribs^talk 21:15, 17 November 2015 (UTC)

Should CS1 errors be restored?

Re https://en.wikipedia.org/w/index.php?title=LMS_6399_Fury&curid=2024630&diff=693616758&oldid=693616168

Is it better to use |title=none and remove the error (a magazine cite, where the magazine title is present), or is this "a cop out" ?

Thoughts? Andy Dingley (talk) 19:09, 3 December 2015 (UTC)

Until the error can actually be corrected, the error should be emitted by the template. --Izno (talk) 20:30, 3 December 2015 (UTC)

What's the "error" though? It's likely that (given the context here) there was never any "title" used in the original source (or else it would have been added originally). It's not unusual for magazine sources to offer a publication title and a page number, but no more. Nor is there much credible likelihood of this "title" getting added. Andy Dingley (talk) 21:00, 3 December 2015 (UTC)

The article title should be inserted. Since articles in that publication do have titles, it is incorrect to state that they do not have one. DrKay (talk) 21:16, 3 December 2015 (UTC)

Yes, the article should be inserted. Andy, your section title is misleading: this is not a matter of restoring an error, but of restoring the message of an error. The error being the lack of a title. I don't know where it is "not unusual" to not provide the title of article (aside from journals), but if there is a nest of such mis-usage on WP it ought to be cleaned out. ~ J. Johnson (JJ) (talk) 22:03, 3 December 2015 (UTC)

Flight is just one news-heavy journal where this is a regular issue. Large articles have a title, but lots of the most useful content is in a section (which might have a section title) but the individual items are just separated by bullets. There is no title here at the article level. We should not pretend that there is. Andy Dingley (talk) 22:13, 3 December 2015 (UTC)

Then use the |department= parameter. If the very specific item you cite does not itself have a title use the title of the next larger division that does. (BTW: We are not "pretending" anything. If you cite something from an article you cite its title. If you cite something from a regular department or section you cite its title.) ~ J. Johnson (JJ) (talk) 23:25, 3 December 2015 (UTC)

Using |department= is all very well, but it won't stop the CS1 error. The way to stop that, when |title= really is irrelevant, is to use |title=none. This is the deliberate behaviour built into the template. Yet that's getting reverted. Andy Dingley (talk) 00:56, 4 December 2015 (UTC)

The |title=none functionality was added to support certain types of academic journal citations that list only the barest minimum of information. The original discussion is here.

—Trappist the monk (talk) 01:46, 4 December 2015 (UTC)

And right at the start of the discussion that Trappist has linked to is his statement: "I think that all CS1/2 citations must have |title=." So there it is: as "title" in this context is presumed to that of an article, the implication is that only articles can be cited. A sentiment echoed by @Redrose64: where he deleted the 'title=none' with the comment "[w]hat we need is the title of the article". But perhaps he would accept citation of a department ("a regular department, column, or section within the periodical or journal").

Andy, try using 'title=none' along with 'department=', and let's see what happens. ~ J. Johnson (JJ) (talk) 23:12, 4 December 2015 (UTC)

The error message serves to indicate those refs which do not (as yet) have a non-blank |title= parameter. What I want to see is an actual article title being added by people with access to the magazines concerned. What I do not want to see is people adding a whole bunch of |title=none merely to clear that error message (as here): that is what I mean by "a cop-out", since it does not help anybody, and gives the misleading impression that the articles have no title. A ref like "The Railway Magazine. London. 1929. {{cite journal}}: Missing or empty |title= (help)" is scanty enough as it is: besides the absence of a title, we don't know which page, or even which month; so to verify that ref, somebody needs to search through twelve issues (about 84 pages per issue, excluding the advertising pages at front and back of each issue, total about 1008). If we have an article title, then somebody with an index for the volume (included at the back of the June and December issues) can at least find out the month and page. --Redrose64 (talk) 09:58, 5 December 2015 (UTC)

I agree that generally there should be a title of some sort, and that "none" should (generally) not be used to "clear the error message", and certainly not to evade the duty of providing full bibliographic details. The problem is that in this context "title" is implicitly (and in your comments, explicitly) that of an article, and the case Andy presents is where the material is taken from a department. You "want to see is an actual article title", but that is not what is being cited. Such cases are properly handled using |department=, except that the template still demands a |title=. As Trappist is unwilling to alter that behavior, we need a working understanding that when |department= is specified it is okay to "clear the error message" using |title=none. ~ J. Johnson (JJ) (talk) 20:54, 5 December 2015 (UTC)

Sorry for the ignorant question, but what is a "department" exactly? Is that like "Bagehot" in The Economist or "Mathematical Games" in Scientific American? Or what, exactly? If others are as ignorant as I, that could be making the discussion difficult. --Trovatore (talk) 21:15, 5 December 2015 (UTC)

Yes. As defined at Template:Cite journal#Periodical: 'department: Title of a regular department, column, or section within the periodical or journal. Examples include "Communication", "Editorial", "Letter to the Editor", and "Review".' Also "Bagehot" and "Mathematical Games". Conceivably anything in a periodical that is not an article. ~ J. Johnson (JJ) (talk) 22:47, 5 December 2015 (UTC)

How do you know "that is not what is being cited"? Do you have the relevant issue of The Railway Magazine? If so, please add the month and page, at the very least; if not, please do not make statements that you cannot substantiate. --Redrose64 (talk) 00:20, 6 December 2015 (UTC)

In the specific edit presented there is no way of telling what was cited, whether in the nature of an article or not-an-article, because the citation is incomplete. Which is what I initially chided Andy about. But please note his comment of 22:13 3 Dec. specifying where "useful content is in a section" (emphasis added). That is to say, not-an-article, for which |department= is appropriate. (Andy: would you care to provide us with a suitable example of a citation that uses |department=?) ~ J. Johnson (JJ) (talk) 21:59, 7 December 2015 (UTC)

Bundling revisited

Starting over. Having unexpectedly had quite a bit of free time on my hands over the last two days, I have composed the following description of exactly what bundling is, how it can be done, and how it should not be done. I hope this is clear to everyone.

What is "bundling" in the context of references?

Bundling references is taking something like this:

In this first example, the references are not bundled.

References

and turning it into something like this:

In this second example, the references are bundled.

References

(Yes, I know these references are ridiculous.)

What are some other ways of bundling?

The last example used a bulleted list and "labels". Variations include not using a list:

In this third example, the bundled references take the form of a paragraph.

References

or not using labels, which results in a so-called "naked bundle":

In this fourth example, the references are listed in a naked bundle.

References

Whether to use a bullet point for the first citation in a naked bundle is a matter of taste, but omitting it might offend some editors (and it may — or may not — be confusing to people using screen readers).

In this fifth example, the references are listed in a naked bundle but the first citation lacks a bullet point.

References

Of course, the list format can be retained without using bullet points:

In this sixth example, the references are listed in a naked bundle without any bullet points.

References

but that can get confusing when the individual citations are long enough to be "wrapped" into multiple lines.

What is the purpose of bundling?

As far as I can tell, the two "legitimate" purposes of bundling are:

1. To avoid having a series of (bracketed) footnote numbers all in a row in the article text.
2. To clarify which source supports which piece of information.

Is bundling necessary?

It depends on how important you think it is to accomplish these goals. Note that purpose #2 can usually be accomplished by simply placing the references in the right places in the text:

In this seventh example, the references are not bundled.

References

Granted, some people dislike having multiple footnotes spread across a sentence just as much as having several in row at the end — and to be fair, it is sometimes difficult to choose the "right places" for the various references, given that each source can support multiple facts.

Personally, I don't place much value on purpose #1, so merely doing that for it's own sake seems useless to me. Purpose #2 is much more important, but like I said, it can often be accomplished without bundling. In my opinion, bundling should only be used if it is not clear which of multiple sources support which fact and it is not possible to clarify this by proper placement of footnotes. (Obviously, such a case would not be fixed by a naked bundle, so only a "labeled" bundle would be appropriate.)

Note that the second and third examples fulfill both purposes, the fourth through sixth only #1.

How should bundling not be done?

One should not simply wrap a {{refn}} template around the existing series of references without removing the <ref> tags (and without formatting the references in some manner shown above), because then we get this:

In this eighth example, the references are bundled by simply wrapping them in a {{refn}} template.

References

Why do I have a problem with this way of doing it?

1. While it avoids having a series of footnote numbers in a row in the article text (purpose #1), it merely pushes this problem to another place on the page; now there's a series of footnote numbers in a row in the list of references. And while it is not impossible to figure out that this means the information at footnote 4 is supported by the sources listed in footnotes 1 through 3, I believe this places an "unnecessary cognitive burden" on readers (it doesn't match what references typically look like in Wikipedia nor in any printed source I know of).
2. Obviously it doesn't clarify which source supports which piece of information (purpose #2).
3. The footnote numbers (1 through 3) for the three bundled references do not appear anywhere in the article text, which again is unexpected and might confuse some readers.
4. The "mouseover" mechanism (which is not enabled in these examples because I'm using {{reflist-talk}}), whereby a reader can "hover" over a footnote number in the article text and see a "tooltip" containing the citation, merely shows "^[1][2][3]" in the tooltip, which is very odd and definitely unexpected (and hovering over the 1, 2, or 3 in the tooltip doesn't bring up any additional information). You can see an example of this "in the wild" by hovering over footnote 9 in the lead section of this revision of the "R (programming language)" article.
5. The "return-link" mechanism (which is likely to not work too well here because everything is so close together in the example), whereby a reader can follow a link back to the article text from a note in the references, does not work for footnotes 1 through 3: clicking on the "^" in front of footnote 2 in the reference list, for example, would bring the reader to footnote 4 in the references section, not to the place in the article text that the source in footnote 2 is related to. Again, you can see this effect "in the wild" in the "References" section of the same article revision: the "return links" on references 5, 6, 7a and 8 all lead the reader to reference 9 instead of to the article's lead section.

Conclusion

References should not be bundled by merely wrapping an existing series of references in a {{refn}} template, as done in this edit (for example).

Does anyone take issue with any of this? Can we agree that the eighth example shows what should not be done when bundling references? - dcljr (talk) 02:39, 9 December 2015 (UTC)