Wikipedia talk:Manual of Style/Archive 46

Archive 40

←

Archive 44

Archive 45

Archive 46

Archive 47

Archive 48

→

Archive 50

Changes to Cite. (continued)

This discussion "Changes to Cite." continues from Archive 45.

Real arrows

Would anyone object to changing the semantically-meaningless carats (^) to real up-arrows (↑) in the current version of Cite, and still leave open the possibility of changing the format according to the resolution of the discussion above? —Michael Z. 2006-02-23 15:45 Z

I think that was chosen because not all browsers are Unicode enabled and so it would show as a blank box for some. Rmhermen 18:11, 23 February 2006 (UTC)

Doesn't using character entities (↑ to ↑) usually solves that? Circeus 18:14, 23 February 2006 (UTC)

I think boxes or even question marks wouldn't be much worse than the circumflex. I don't think my first thought is “upwards” when I see one.

AFAICS, using character entities wouldn't make much of a difference. Another problem is that some users may simply not have a font on their system that contains arrows.—Wikipeditor

The up-arrow displays correctly on an out-of-the-box Mac OS X, Windows XP, and Windows 98 system. In Lynx text-only browser (in a Mac terminal window) it displays correctly when Lynx is in UTF-8 mode, and gets converted to an ASCII dash-carat (-^) when in 8-bit text mode. Does anyone know of a configuration where the up-arrow fails? —Michael Z. 2006-02-23 19:01 Z

All wikipedia pages have <meta http-equiv="Content-Type" content="text/html; charset=utf-8" /> at the top. The only problem I've ever heard dealing with browser/unicode compatibility is IE.. but I think you have to screw something up to get it there. Instructions on how to fix that error here. Also, on something way more prevalent: page histories, the → symbol is displayed whenever a specific section is edited. Thus, Unicode browser compatibility is not an issue. I say go ahead. drumguy⁸⁸⁰⁰ - speak? 21:30, 23 February 2006 (UTC)

I support using real arrows. I don't know of many systems these days that would not be able to display them correctly. Kaldari 02:01, 24 February 2006 (UTC)

I used ^ originally because that's what the Note template was using, but real arrows of course are a much better idea. I've changed the default in the software, this'll presumably have to be changed manually on enwiki. —Ævar Arnfjörð Bjarmason 23:40, 25 February 2006 (UTC)

Cool. I see that this is now active. Dv82matt 05:47, 27 February 2006 (UTC)

Just a thought. What about using a small supercript 'a' we already use letters if there is more than one backlink. Why not keep it consistent? --Dv82matt 01:27, 24 February 2006 (UTC)

An example of how it would look:

1. ^{a b} Marx, J. L. New disease baffles medical community. Science. 1982;217(4560):618-621.

2. ^a New York Times. January 2010 issue, page 16

3. ^a UNAIDS

Dv82matt 01:37, 24 February 2006 (UTC)

I don't think this is ideal:

• Visually, a list looks like a list because it has a column of labels (numbers or bullets), followed by neatly aligned text. The back-links break the neat alignment of text and they look like secondary labels for the notes, but they are not.
• Functionally, each number is a label helping visually identify the note; the main purpose is to help find the note from its reference in the text. The back-link is a secondary function, subordinate to the text of one note, and it often will not be used by readers who simply click their browser's back button.
• Superscripts are used by typographers to pull note references out of the flow of text, and not for note labels. The usage here seems a bit odd.
• There is no need for numbering (lettering) an item a, when there is no b. For example, in standard Harvard referencing you only see (Smith 2005a) when there is also a 2005b.

Here's an example with back-links after the notes. Since they don't align with each other in a column, they don't present themselves as a major defining element of the list, but each as a separate element subordinate to its note. I've used non-breaking spaces to make each one a bigger mouse target.

1. Great, El Borbah The (2006). New-fangled back-links. Newtown: Innovation. ↑ 
2. Verne, Jules (2006). Citations start with authors name: just like in virtually all publications. City of the Future: Long-named Publisher for Line-Wrapping. ↑ 
3. Who, Joe (2006). Back-links are visually subordinate to the note's content. Newtown: Innovation. ↑a ↑b

—Michael Z. 2006-02-24 16:49 Z

I do like the look. Another reason for the arrow to be at the back of the note rather than the front is to prevent an ugly column of arrows from forming when there are multiple notes.

I still think it's more intuitive to have the backlink at the front but maybe thats just me. Anyway since linking to the number itself is problematic your alternative seems reasonable. I say we go for it. --Dv82matt 17:59, 24 February 2006 (UTC)

There is a case to be made for Dv82's caret-free references with backlinks residing in alphabets (they look good!). However, the problem with it IMO is the use of the alphabets where they aren't actually indicated; ie. alphabets are properly placed next to a reference when there is more than one citation of that reference in the text (Michael is correct to allude to the Harvard style conventions). The more I think about it the more I'm persuaded that Michael's solution is an excellent way to go: it's functional, it's consistent with professional citation conventions, and it's not hard on the eye. ENCEPHALON 12:42, 25 February 2006 (UTC)

Backlinks after the note - straw poll

Do you support or oppose formatting backlinks in the style Michael has proposed?

Example (copied from above):

1. Great, El Borbah The (2006). New-fangled back-links. Newtown: Innovation. ↑ 
2. Verne, Jules (2006). Citations start with authors name: just like in virtually all publications. City of the Future: Long-named Publisher for Line-Wrapping. ↑ 
3. Who, Joe (2006). Back-links are visually subordinate to the note's content. Newtown: Innovation. ↑a ↑b

• Support - Dv82matt 05:47, 27 February 2006 (UTC)
• Support - I'm not a huge fan of real arrows (I don't like the fact that they start from below the baseline, that's all), but I do prefer the arrow after the note. PizzaMargherita 06:57, 27 February 2006 (UTC)

  PM, that seems to be the design of a particular font; my arrows sit right on the baseline (using Apple's Lucida Grande font in my user style sheet). —Michael Z. 2006-02-28 18:20 Z

  Ah, of course, thanks for the clarification. (Btw, I use Firefox with no tweaks.) I reiterate my support for this, it's much more natural, is it not? I mean, you go to the footnote, you read it, and at the end of it (which could be many lines further down) you find the "up" link right there where it makes most sense. Oh well. PizzaMargherita 22:45, 28 February 2006 (UTC)

• Oppose - Although I don't feel strongly either way. I like being able to quickly jump up and down in the article between references and notes and doing this takes a longer time when I can't expect all the backlinks to be positioned at the same distance from the left margin.

  Use your browser back button, or its associated keyboard shortcut. —Michael Z. 2006-02-28 18:20 Z

• Oppose - I don't see the point, and prefer the clean look of the current system. In my experience, moving the arrows back would only make them harder to find, as each reference will contain other linked items. Try telling an inexperienced reader to click on the arrow, somewhere in the reference. -- Ec5618 10:28, 28 February 2006 (UTC)

  After clicking a footnote reference, most novice Web users may not even realize they are still on the same page. They would just click the browser back button without thinking. The back-links are just an additional nicety. Personally, I don't think the current system looks very clean at all, especially when a/b notes break up the list's left-alignment. —Michael Z. 2006-02-28 18:20 Z

• Support —Michael Z. 2006-02-28 18:20 Z
• Oppose. Too messy. Mark^{your words} 22:50, 28 February 2006 (UTC)
• Support. Much neater than the current set up. Regardless of the placement of the arrow, however, I think the more important issue is that it is an arrow (and not the non-intuitive caret). Cite.php currently displays an arrow, apparently a change made by someone other than Ævar. Now if only those brackets around the supersripted numbers are removed... ;-) —Encephalon 04:48, 3 March 2006 (UTC)

Another Option

What about using a small superscript arrow ^↑? It could look something like this:

References

1.  ^{a b} Marx, J. L. New disease baffles medical community. Science. 1982;217(4560):618-621.

2. ^↑ New York Times. January 2010 issue, page 16

3. ^↑ UNAIDS

Just an idea. Any opinions? Thoughts? --Dv82matt 11:24, 3 March 2006 (UTC)

Status

Proposed: No carats for backlinks. Just link from the number, or if there is more than one backlink, link from the letters.

I've implemented this and deployed it, you can see a demo of it at Wikipedia:Sandbox/Ævar. Ævar Arnfjörð Bjarmason 11:34, 24 February 2006 (UTC)

Subreferences

I've got a completely different idea. The context of each reference should be centralized in the references section, and the context of each particular fact should be near that fact (this fact is on page 34, etc.) Then, when the references are generated, the "fact references" are a sublist of each list element:

References

I've written it up (with a more detailed example) here: m:Talk:Cite/Cite.php#A_different_idea. Please comment. — Omegatron 17:45, 23 March 2006 (UTC)

Maybe I'm missing something, but won't that make the references completely useless when the article is printed, particularly if certain sources are heavily used throughout? Kirill Lokshin 17:53, 23 March 2006 (UTC)

Oh, do you mean because there's no connection between the bullets in the references section and the numbers in the article? Good point. The current references system doesn't handle this, either. Maybe they should be 1.1, 1.2, or 1a, 1b, etc.

I didn't really think through that aspect in detail. The meat of my proposal is to centralize the references in the actual references section, instead of sprinkling them throughout the article, as explained on meta. — Omegatron 22:37, 23 March 2006 (UTC)

It's a clever idea, but I'm not sure if it would be possible to implement it without breaking the simpler one ref tag → one footnote number uses of the current implementation. More problematically, it turns the ref tag from a mere formatting tool to produce a footnote into a more sophisticated piece of metadata, which would cause problems for those uses of footnotes that contain things other than simple citation of sources. Kirill Lokshin 01:42, 24 March 2006 (UTC)

1. It is currently possible to use multiple ref tags to link to the same footnote. The text inside of the first one is used in the references section. Text inside of each subsequent instance is ignored, so I don't know what purpose it is currently being used for, if at all. If this were implemented, the text inside each subsequent ref tag would just show up as subreferences of the first.
2. What uses are there for footnotes besides citation? Why would this break them compared to the current system? — Omegatron 02:40, 24 March 2006 (UTC)

Explanatory footnotes also exist ;-)

And I'm not sure whether this approach would break them or not; would it change how ref tags with no "name" attribute are treated, or would those still work as plain numbered footnotes? Kirill Lokshin 02:46, 24 March 2006 (UTC)

I thought we weren't supposed to use those.

It would depend how it is implemented, of course.  :-) Going on just the above description, unnamed notes would work the same as they currently do. On the meta page, I expanded it even more to suggest that an unnamed ref tag would be automatically named (first word of the content or whatever) and moved into the references section on save. Same for the first instance of a named tag that doesn't already exist in the references section. Explanatory notes would still work the same way with this, but would no longer be next to their original position in the source, which may be suboptimal. — Omegatron 03:01, 24 March 2006 (UTC)

I think auto-moving tags and whatnot should be avoided. It should still be possible to use cite.php to create a proper list of numbered footnotes (see this article, for example) without having them converted to this new format.

On another point, this wouldn't really allow a single footnote that referred to multiple sources; that would still have to be done manually. Kirill Lokshin 03:26, 24 March 2006 (UTC)

I think auto-moving tags and whatnot should be avoided.

Hmm... The basic premise here is that the reference details belong in the references section of the source code, and not scattered throughout the article. If this were used, I don't see why there would be any reason to allow some references to be inline. It defeats the whole purpose.

Because editors should be able to make that decision on a per-article basis. Kirill Lokshin 12:52, 24 March 2006 (UTC)

But there's absolutely no reason for it to be different from article to article. — Omegatron 13:50, 24 March 2006 (UTC)

(see this article, for example)

I don't understand. I didn't even have something like that in mind when I thought of this, but it's a better application for this style of notes than anything I could have come up with; a perfect fit. Why wouldn't we want that converted to the proposed format? (There are even comments in the source code trying to explain to the editor what each "ibid" reference is referring to! <ref><!--Norwich, ''History of Venice'', 394.-->Ibid., 394–395.</ref>)

But again, you've now gone from allowing a new (easier to use, in your view) citation style to breaking other commonly accepted styles (Chicago Manual of Style, in this case). We may not require correct footnoting, but it should still be possible to create it; the style you've proposed might be clever, but it isn't accepted by regular style guides. Kirill Lokshin 12:52, 24 March 2006 (UTC)

"Correct" is subjective. Those style guides are for paper, which Wikipedia is not.

Besides, as long as the semantic information is in the article source (and it would be with this proposal), you could have a user preference to render the references section however you want. Heirarchical, ordered list, unorganized dump of ibids, or whatever you like (enough to program into the extension). — Omegatron 13:50, 24 March 2006 (UTC)

this wouldn't really allow a single footnote that referred to multiple sources

I don't understand. Why would you want to create a single citation for two sources? — Omegatron 05:42, 24 March 2006 (UTC)

To reiterate: footnotes (the typographical thing) are different from source metadata (the conceptual thing). According to any formal style guide, having two footnotes in the same place is wrong; the accepted way of doing so is to have a single footnote that includes both sources for the fact you're citing. Kirill Lokshin 12:52, 24 March 2006 (UTC)

Why? — Omegatron 13:50, 24 March 2006 (UTC)

Because Wikipedia, although not paper, should still follow the proper rules of style for formal English writing. To wit, from the Chicago Manual of Style, 15th edition, page 601: "A note that applies to more than one location should be cross-referenced; a note number cannot reappear out of sequence" (16.33) and "Using more than one note reference at a single location (such as ^{5, 6}) should be rigorously avoided" (16.34). Your proposed change would force everyone—even those editors who would ordinarily not do so—to violate these rules. 14:08, 24 March 2006 (UTC)

But why does that rule exist? What purpose does it serve? Style guides exist for a reason; to serve the reader. The reader should not suffer for the sake of prescriptivism. If the purpose behind the rule is good, and it's absolutely necessary for some reason that I can't think of, then I don't know of a good solution for merging multiple citations within this proposal. If the original purpose of the rule is due to a limitation of print, we'll violate the rule to improve the reader experience. (Other style guides don't even use numbers.)

It wouldn't be any worse than the current system; we currently use multiple note references for a single fact and I haven't seen any complaints. I think the drawbacks are negligible compared to the benefits this would bring. — Omegatron 14:53, 24 March 2006 (UTC)

The obvious reason is that, were we to need a number of citations for a single point, we'd wind up with a horrible little list of numbers^{4a 7b 8 19 22} that would be far more intrusive than a single, combined footnote.¹² In addition, we could no longer have a footnote that, say, compares the information given by two conflicting sources, as the new style would force us to split it apart.

(More generally, the better style guides tend to have good reasons for doing things a particular way, even if those reasons are not always immediately obvious.) Kirill Lokshin 14:59, 24 March 2006 (UTC)

Yes, I know they do. I was asking what the reasoning behind this was.

I agree that avoiding clutter would be good (that is the purpose of this proposal, after all), but I can't think of a way to combine references with this proposed system. Multiple superscripts really don't look that bad to me, are currently in use on the Wikipedia, and aren't very common, anyway.

As for comparing multiple references, which is also rare, they could just have "(compare with...)" and vice versa, with a link to each other?

References

I agree that it's not perfect, but I think the benefits of this system would outweight these disadvantages. — Omegatron 00:48, 25 March 2006 (UTC)

I don't agree. You're trying to deal with a fairly limited problem—people putting overly large templates directly into the footnote call—that can be dealt with by other, non-technical means; your solution destroys the flexibility of the existing footnote tool in favor of a straightjacket solution that, as you admit, doesn't work as well as the current one in a number of cases (and which forces references into a style that nobody actually uses, to boot).

I would have little concern if this were just added as an additional feature to cite.php, and the support for raw footnotes (using unnamed ref tags, for instance) were retained; but I strongly object to breaking the flexibility of the current implementation, particularly via the use of automatic post-edit conversion, to cater to a small minority of people that can't—or won't—learn proper footnote style. Kirill Lokshin 02:17, 25 March 2006 (UTC)

No, I admit that it has one flaw in the context of "proper" footnote style according to one particular style guide, but is otherwise clearly superior to the current system, which has many flaws.

I would object to making this an "add-on" feature, as that would encourage even more inconsistency between articles. — Omegatron 17:54, 25 March 2006 (UTC)

Let's recap, then. Your proposed system has a number of flaws, as I see it:

1. It prevents combining footnotes, forcing multiple note numbers at a single point.
2. It prevents the order of the notes in the article from matching the order of the notes in the notes section.
3. It produces a combined notes/references section in a style that is used nowhere else, which is confusing for the reader.
4. It makes changes to the article text that are beyond the control of the editors apparently making them.
5. It makes little allowance for discursive notes, which now wind up in a "References" section for no apparent reason.
6. It will break existing articles that use the cite.php extension.

It's only advantage, meanwhile, is that it forces the bulk of the citation information to the bottom of the article. This is not usually a problem if the footnotes are done properly in the first place; there is no substantial difference between <ref>Doe, ''My Book'', 57.</ref> and <ref name="Doe">p. 57.</ref>. So why exactly is it necessary to force an inflexible, non-standard style on everyone? Kirill Lokshin 20:08, 25 March 2006 (UTC)

No, no, no.

1. It does not prevent combining footnotes. Just enter them like you currently would, but you'll need a unique name tag for each (serving the same function as the inline comments in the example article, so it's not even any more work than doing it that way. Or you could just enter them without a unique name and they would be auto-named for you on save.)
2. No. Another benefit of putting the reference content in the References section is that you can order the references however you want. With the current system, you are limited to the order they appear in the article.
3. Debatable. We think the subreference style would be superior for the reader, but apparently others disagree. Regardless, editors could still use a separate Notes and References section if they wanted to, and do CMS style notes, by giving each reference a unique name.
4. What? This doesn't affect the article text.
5. Yes. But that's where the footnotes visually appear when the page is rendered, so it's confusing to newcomers that they can't be edited by editing the References section. It's also frustrating for regulars as we have to search through the article wikicode trying to find the footnote text so we can edit it instead of just editing the References section where it appears. You claim that moving the reference source code on save would be confusing, but I maintain that using source code in one section to create rendered content in another section is much more confusing, and the current ref tag system is the only place something like this occurs on the Wikipedia. Getting rid of this bizarre functionality is the main reason for this proposal.
6. It will be backwards compatible with the current system. That's why it uses the same ref and references tags (but uses <references></references> instead of <references/>).

I also want it to address the other points in the list of concerns I raised at m:Talk:Cite/Cite.php#A_different_idea. — Omegatron 16:45, 26 March 2006 (UTC)

1. But if each note has its own unique tag, how would the sub-reference functionality work? The tag (similar to the old {{ref}}/{{note}} system) wouldn't necessarily be identical for different citations from the same source.
2. Perhaps I wasn't clear here. If you proposal is limited to moving the content of references to the reference section, this may be true; but if it also includes the sub-references part, then the notes at the bottom will be interleaved with the reference works, causing them to have a different order than the note numbers in the article text.

As far as your other points: as I said at the outset, I have no qualms with this so long as the existing functionality can be replicated. Kirill Lokshin 16:56, 26 March 2006 (UTC)

I will make some more detailed examples of what I am envisioning when I have time. What I am imagining would certainly need tweaking and discussion about details before going live. The primary proposal is to move the reference content into the references section, where it appears in the rendered page. — Omegatron 06:05, 31 March 2006 (UTC)

Much too complicated

There is really no need to make this quite so complicated. Having the system actually move text around is not something to which wiki-editors are accustomed. In any case, having the citation next to the text to which it refers is probably a good thing.

The most that is necessary with respect to sub-references is something like an extra attribute on the <ref> tag which will add text identifying a section of the text identified in the main reference. So something like <ref name="foo" section="p. 57"/> or alternatively <refsection name="foo">p. 57</refsection> might result in something like:

1. ↑ whatever text was in the foo reference

   ^a p.57

Whilst the actual syntax is open to argument there is no need for anything more complex than this.

Wuth respect to arguments over whether we follow the Chicago Manual of Style or whatever manifold other styles are available, I suggest that we would be better off synthesising a single consistent citation style of our own rather than succumbing to endless style-wars. Ending the foolish forking of templates to deal with "Harvard referencing" would be a good start.

HTH HAND —Phil | Talk 00:26, 26 March 2006 (UTC)

• My chief proposal here is not the subreferences, but moving the reference source code into the References section where they appear in the rendered page. I just think subreferences would be a logical extension of that.
• The "moving text around" is just supposed to be a shortcut method for entering references, like the [[pipe trick|]] or ~~~~ signatures, which are also modified on save. Then you could enter references while section editing, but still edit them later by editing the References section, where they will appear when the article is rendered. Having source code in one section generate content in another section is not something to which wiki-editors are accustomed, either.
• My proposed tags aren't any more complicated than your example <ref name="foo">p. 57</ref> and make sense in the context of the current style, which ignores such text. Not any more complicated than the style used in the example article, either. The details of functionality and syntax will need to be tweaked and discussed, of course. — Omegatron 16:45, 26 March 2006 (UTC)