Jump to content

Template talk:Reflist/Archive 7

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1←Archive 5Archive 6Archive 7Archive 8Archive 9Archive 10

Limiting to 3 or 4 columns

Some people are going around changing reflists to 5 columns using this template. 5 columns is excessive. Is there some way we could limit this template to 3 or 4 columns at the most? Kaldari (talk) 17:20, 9 August 2008 (UTC)

Yes. We can do this with some CSS per User:Anomie/Sandbox3. I had been thinking that three or four columns was too much, but have reconsidered. Articles that use pure Harvard referencing will benefit from four columns. See Calvin Coolidge#Notes— a bit of a mix here, but it illustrates the concept. --—— Gadget850 (Ed) talk - 19:19, 11 August 2008 (UTC)

Limit to 2; even three and four cause problems on some browsers. SandyGeorgia (Talk) 06:00, 12 August 2008 (UTC)

What problems? --—— Gadget850 (Ed) talk - 11:16, 12 August 2008 (UTC)
I wouldn't say problems exactly, but IE cannot cope with 2, 3 and 4 browsers, I personally think 1 is the best as everyone will see it in the same sort of way. D.M.N. (talk) 14:53, 12 August 2008 (UTC)
IE degrades gracefully, in that it always displays the references in 1 column. "Force everyone with better browsers to have to have 1 column because IE ignores column-count" is IMO extremely poor reasoning. Anomie⚔ 02:00, 13 August 2008 (UTC)

Why not just use {{#switch:{{{1|}}}||1=1|2}} for the number of columns in the template code? --NE2 06:03, 12 August 2008 (UTC)

That could be done, but it wouldn't solve the larger issues above. Anomie⚔ 02:00, 13 August 2008 (UTC)

Gadget

In the ongoing absence of reflist being made to work properly, is there any sign of a gadget in user preferences which will enable editors to force it to display a single column? DuncanHill (talk) 11:28, 12 August 2008 (UTC)

Reflist has been in this state for a long time and the peasants aren't storming the castle, so just hang on. There have been a number of complaints that various things are broken, but no details. We now have a good start on characterizing the issues. My next step is to do a recap, present the known issues and then start some proposals. In the meantime, look at Anomie's proposal above. --—— Gadget850 (Ed) talk - 11:56, 12 August 2008 (UTC)
I did, but I don't really understand it. DuncanHill (talk) 11:58, 12 August 2008 (UTC)
I don't understand it either. Ever since I started using it in place of <references/> in Australian articles, usually with no parameters, it's worked just fine. Orderinchaos 02:18, 13 August 2008 (UTC)

Usage and issues

Usage
  • Reflist is used in over 510,000 articles (88%)
  • Two columns are used in over 50,000 articles (10%)
  • Three columns or more are used in about 2,500 articles (1%)
Multiple columns
  • FireFox (13% market share): works with caveats
  • Safari: (6%): exposes a bug in Safari through 3.1.2 where the linking is mispositioned
  • Internet Explorer (73%): multiple columns does not work and is not planned to be supported by IE8
  • Caveats— where multiple columns do work, they:
    • may create widows
    • may mangle long URLs in the printable version
Font size
  • The font size should be reduced to 90%, but does not appear to work for all users, mainly those using IE
/* make the list of references look smaller */ ol.references { font-size: 100%; }
and not the .references-small entry. Changing this in user .css changes the font size.
  • Some users have visual problems with smaller fonts
User interface
  • There is no way for users to customize the look of the references section

--—— Gadget850 (Ed) talk - 14:18, 12 August 2008 (UTC)

Unless something has changed, the reflist font size is 90% and has been that for a long time. Gimmetrow 14:25, 12 August 2008 (UTC)
Reworded- font size at 90% does not work in some instances, mainly with IE. --—— Gadget850 (Ed) talk - 15:06, 12 August 2008 (UTC)
Think I figured out the problem- see above. --—— Gadget850 (Ed) talk - 20:29, 12 August 2008 (UTC)

More than two columns?

If multi-column functionality is retained in the template, I'd like to know -- is there anyone who feels the template should support more than two columns (or more than three)? (One must bear in mind that on narrow displays, three or more columns gets to be ugly. I for one find two columns adequate for nearly all cases.)--Father Goose (talk) 01:14, 15 August 2008 (UTC)

Recap and proposals

The following discussion is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.

Because of manner in which the <references /> tag is handled, the only way to support multiple columns is to use CSS3. The presents a challenge, as CSS3 is still under development and browser support is currently limited to custom implementations.

FireFox is the only browser that supports multiple columns, with some minor issues noted above. Internet Explorer still has the biggest market share at 73% for all versions— current versions do not support CSS columns, and and neither will version 8 that is in development. Safari has a bug with multiple columns that breaks linking.

Unless someone can come up with another solution, using multiple columns via CSS3 will not present the same look across browsers.

I propose that we:

  • Disable default multiple columns for all browsers.
  • Set the default font size to 100%
  • Document how a user can override font size and column setting by editing their CSS
    • Example: adding .references-small {font-size: 90%; -moz-column-count: 3;} to the user .css will show all reflists without a set column to show as three columns.
  • Look at implementing a reflist interface as a gadget.

--—— Gadget850 (Ed) talk - 15:04, 12 August 2008 (UTC)

Multiple columns are problematic for printing, scrolling, and small screens. There is no reason to use them, and they should certainly not be used here. --brion (talk) 21:48, 12 August 2008 (UTC)
The reason to use them is that they look better, in certain cases. Are the problems you mention so bad that they cancel out the advantage of improved formatting?
As for the argument that they "do not present the same look across browsers", that's true of all HTML rendering. I could only see changing it on that basis if an enhancement for one browser produced actual problems in a different browser. Like the poster above says, if it breaks rendering in Safari, disable it for Safari.--Father Goose (talk) 21:57, 12 August 2008 (UTC)
Oh, and if, as the template doc says, "Three-column lists are inaccessible to users with smaller/laptop monitors and should be avoided," then that should be enforced in code: only permit {{reflist|2}}, discard other values.--Father Goose (talk) 22:04, 12 August 2008 (UTC)
Looking at past discussions, we are never going to come to a consensus as to what this should look like— that is why the proposal is to let the user decide. We can recommend that multiple columns not be used, but if an editor makes a personal choice to show the references at six columns, then it is their problem to deal with any display issues. This way we are not forcing a decision and we might actually close out months of discussions. I welcome any better compromise or solution. --—— Gadget850 (Ed) talk - 23:01, 12 August 2008 (UTC)
An editor shouldn't be given the option to choose six columns if it looks horrendous or breaks functionality for everyone else. One of the threads SandyGeorgia lists below suggest that there are display problems (even on Firefox) with 3-column layouts, and it's established that Opera has problems with any number of columns. Further, it is claimed that three or more columns borks narrow displays (I believe it; they don't even look good on my 1600x1200 layout). So I would support changing {{reflist}} to only accept "2" as a parameter.
But what I haven't yet seen is a reason to disable two-column functionality. Yes, it could be re-enabled via user CSS or a gadget, but I don't see the need to rip out useful functionality if we can fix its specific problems instead.--Father Goose (talk) 23:54, 12 August 2008 (UTC)
Let me reflect this back at you: you would rather see the reference list according to however the reflist template was last edited rather than according to your preferences. --—— Gadget850 (Ed) talk - 01:11, 13 August 2008 (UTC)
  • If I understand you correctly -- yes, I would like editors to be able to specify one or two columns in the reflist (but not more). Ideally, I would like the ability to override their choice via ticking a box in my preferences -- though I personally would probably not tick the box.
  • Can we fix the IE font size issue to make it 90% across all browsers? It looks like, above, you've identified the source of that problem. I do think the references look better in a smaller font.
  • If having the references in columns borks printing, alter the printing-specific CSS to disable multicolumns, but leave it intact in the display CSS.
  • It sounds to me like what is needed here is a set of bugfixes and tweaking, not a near-complete removal of all of {{reflist}}'s functionality.--Father Goose (talk) 01:53, 13 August 2008 (UTC)
oopsie, I didn't mean to include two below, fixing. SandyGeorgia (Talk) 00:08, 13 August 2008 (UTC)

Reflist 2| 3|4 and above should be eliminated. Past threads documenting the problems, which are not only IE:

Since I don't speak this language, I don't know if I'm agreeing or disagreeing with the proposals here; I just want the column messiness to go away. SandyGeorgia (Talk) 23:06, 12 August 2008 (UTC)

Now that I've understood what the actual proposal is, yes, I support a default of one column, 100% font size, with users changing preferences if they want a different setup. Most of the world is on IE and isn't seeing the multi-columns anyway, and it breaks on some other browsers. SandyGeorgia (Talk) 12:59, 13 August 2008 (UTC)
  • I too support all of the proposals listed. Until a consistent-across-browsers set of code for columns, and typography-sizing for "small", can be established. (If utterly necessary, I'd request that columns be restricted to 2, and size be kept to a minimum of 92% or 95%.) -- Quiddity 01:06, 13 August 2008 (UTC)
  • I don't mind as long as it keeps working. I think too many columns (more than 3, or as FatherGoose suggests, maybe 2) is probably excessive, so if that's eliminated I really won't miss it. Orderinchaos 02:20, 13 August 2008 (UTC)
  • Extremely Strong Oppose Keep the font size as-is and implement multiple columns via CSS as I proposed above (and omit Safari, and force one-column in a print @media rule if that is such a big deal), and document how to force 100% size and/or single column using CSS and/or gadgets—in other words, do the opposite of what is proposed. If we do as this proposal suggests, we'll just end up with a WP:TFD for this template and/or people randomly removing it all across the wiki. Also, FWIW, I am not convinced by unsupported assertions of "inaccessibility" and reference to browser bugs in a bleeding-edge browser version that have been fixed in more recent releases. Anomie⚔ 02:23, 13 August 2008 (UTC)
  • "unsupported assertions of 'inaccessibility'": Would you define that, as I don't see that asserted.
  • "bleeding-edge browser version that have been fixed in more recent releases": If this refers to Safari, then I would not consider it any more bleeding edge than the FireFox 3 that I'm running; Safari 3.1.2 is the latest version, and it still has the column link position bug. Bleeding edge is using in development CSS3 to make aesthetic changes that affect about 20% of our users.
  • "just end up with a TFD for this template": Why? Reflist offers a standard way to include the references tag without mangling it as it was in the past, it offers the ability to easily use groups (not highly used, yet) and best— it adds the CSS wrapper to allow customizing the look.
--—— Gadget850 (Ed) talk - 11:18, 13 August 2008 (UTC)
  • Some people keep claiming "more than one column is inaccessible!", but the only reason I can find is because they don't like more than one column. The Safari issue seems easy to overcome, as is the long-URL printing issue if we use my proposal. Screen readers should ignore column-count just as IE does. My proposal even eliminates the possibility of someone using {{reflist|100}} which some claim is an accessibility issue.
  • No, I'm not referring to Safari: of the two Safari bugs one has not been fixed and the other is apparently fixed for some people but not others. I'm referring to a thread linked above from two weeks after Firefox 3.0.0's release, involving a bug that seems to have been fixed in 3.0.1. Also, there is quite a difference between using a new feature of CSS3 that degrades gracefully in browsers that don't support it, versus demanding that no one ever use a feature because of one quickly-fixed bug in one bleeding-edge browser.
  • Someone would eventually claim "This template does nothing for everyone that doesn't mess around with their personal CSS (which I claim only 0.000001% of people do). We should just add <references/> or <references group="foo"/> directly to the article instead." Besides needing a bot to edit 500000 articles, I can't even think of a good reason to oppose in that situation. BTW, what manglement of <references/> are you referring to that reflist fixes? I can't think how someone could screw that up beyond typing {{Unreferenced}} instead. Anomie⚔ 12:28, 13 August 2008 (UTC)
Wouldn't it be more accessible to allow each reader to make a personal decision on the aesthetics? That is, for the 20% or so who use Firefox here. The FireFox bug is fixed, so I'm not sure why you are bringing that up now. As for mangling the references tag: <references style="font-size:100%" /> or variants thereof still shows up every now and then, causing a MediaWiki error; reflist just ignores invalid parameters. --—— Gadget850 (Ed) talk - 14:17, 13 August 2008 (UTC)
I don't think anyone is arguing about whether the option for a "personal decision" should be present; in fact, my proposal above is specifically intended to enable personal decision on the multi-column issue. We're just arguing over the defaults: I'm proposing the default remain as what we currently have now (modulo support for large numbers of columns and arbitrary column-widths for technical reasons), and you're proposing changing the default on everything to "1 column, 100% font size" for IMO no good reason. As for the Firefox bug, I didn't bring it up. Anomie⚔ 17:16, 13 August 2008 (UTC)
The thing is, that whilst 2 columns are supported and in widespread use, there are going to be continual problems with edits like this where they didn't test at small screen sizes - that page looks ridiculous at 800x600, and still silly at 1024x768. For me it boils down to - "it hurts WP almost as much as it helps WP". -- Quiddity 03:06, 14 August 2008 (UTC)
Personally, I think it looks odder at larger resolutions as the infobox becomes longer than the content of the page. Anomie⚔ 04:31, 14 August 2008 (UTC)
That's a stub+infobox problem. Nothing to do with reflist. -- Quiddity 17:00, 14 August 2008 (UTC)
There are also continual problems with people inserting useless trivia, using bare URLs as references, randomly deleting sections, breaking templates, pushing POV, spamming, and inserting absolute garbage (e.g. this). What are your solutions to all these other misuses of wikitext that "hurt WP", probably more so than someone making a poor choice of column count? Anomie⚔ 04:31, 14 August 2008 (UTC)
Those problems have nothing to do with reflist. Please don't change the subject. -- Quiddity 17:00, 14 August 2008 (UTC)
I'm sorry I tried to point out that your claim that multi-column references should be completely banned because some people occasionally misuse them is ridiculous by pointing out other situations where such a draconian solution would be ridiculous. I'll try to remember to be more clear in future replies to you. Anomie⚔ 00:27, 15 August 2008 (UTC)
Most of your examples describe vandalism. Bringing them up is just muddying the water.
My claim was that this is one more reason (other than the ones explained above) as to why columns should currently be considered harmful. Please don't purposefully mischaracterize people's (my) statements. Unless you really were confused, in which case I'll try to remember to be more condescending clear in future replies to you. -- Quiddity 02:28, 15 August 2008 (UTC)
Anathem needs a {{clear}} before the "References" section (done) even when it only has one column, so you can't blame that on the columns, or even on display testing.--Father Goose (talk) 07:16, 14 August 2008 (UTC)
Adding a {{clear}} creates a huge block of whitespace. What exactly was wrong with the visual display of the article before you added it? (It looked fine to me in Firefox and Opera. I'm honestly confused). Thanks. -- Quiddity 17:00, 14 August 2008 (UTC)
The references are cramped by the navbox whether there's one column or two (but more so when there's two columns). Short articles with long navboxes are awkward no matter what you do.--Father Goose (talk) 22:14, 14 August 2008 (UTC)
I support the font-size proposal, because I have trouble reading the small fonts, with , I am told, good corrected vision. The column/non-column thing is more complex, but generally columns only help if the whole list fits comfortably on one screen, so depends on window size in both directions as well as the details of the text. Rich Farmbrough, 14:11 13 August 2008 (GMT).
Support; BTW, could someone put this on the local-watchlist header to get more people to this discussion? I've seen discussion links at the top of the watchlist before, and I think it should be done here to get more people to it. D.M.N. (talk) 14:17, 13 August 2008 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
Results
  • Disable default multiple columns for all browsers: no consensus
  • Set the default font size to 100%: no consensus
  • Document how a user can override font size and column setting by editing their CSS: no discussion

The only consensus I see is to disable multi-column support for Safari until the bug is resolved. I will make a few documentation tweaks, but everything else is status quo. --—— Gadget850 (Ed) talk - 20:24, 2 September 2008 (UTC)

Let users know there is a problem

Is it possible to make {{reflist}} generate a message on all articles where it is used with more than one column, saying something like "This article uses a multi-column reference list. This does not display correctly in most browsers. Wikipedia is designed to work only with [insert name of a browser that does work properly with it here] and suggests that anyone who uses another browser should either change or stop trying to read our articles". DuncanHill (talk) 10:31, 13 August 2008 (UTC)

Possible, yes. Desireable, absolutely not, as everything after the first sentence is completely inaccurate (Safari is not "most browsers"; the display is correct (just not quite as desired) in IE). Anomie⚔ 11:43, 13 August 2008 (UTC)
Remind me would you which browser it actually does work correctly (i.e as desired) in? DuncanHill (talk) 11:51, 13 August 2008 (UTC)
Not a good solution. We should not and cannot force users to change their browser. Organizations such as schools, libraries and companies have their systems locked down and don't allow new applications to be installed. I was at the local library last night, and their systems are running IE6. Multiple columns only works properly in FireFox, which has a market share of 13%; the Wikipedia browser usage has not been updated in years and the logs are not accessible, so actual usage is unknown. --—— Gadget850 (Ed) talk - 11:57, 13 August 2008 (UTC)
My suggestion was rather sarcastic - I agree that we should not be trying to make users adopt a particular browser, I was simply hoping that Anomie would explain why he is so vehemently opposed to having Wikipedia work as expected across as many browsers as reasonably possible. DuncanHill (talk) 12:00, 13 August 2008 (UTC)
(note aside). Pity about the browser logs being inaccessible: didn't realise that until now. I'd be very surprised if Firefox /usage/ in terms of user count for WP was anywhere near as low 13% of the total. From personal experience on 100k+ user sites, 30%+ Firefox is more typical (even 62.3% in one case) given the nature of packaged software (IE) vs. who actually /uses/ sites such as WP. Academic point, perhaps, but nonetheless... Harami2000 (talk) 09:47, 3 September 2008 (UTC)
It works as desired except in Safari, where "as desired" is "with no bugs", not necessarily "as the person who added the parameter intended". It degrades well in browsers like IE that don't support the columns. It's essentially the Safari programmers' fault, although there is the issue that it's not part of the official CSS specs, so we can't exactly say that they are doing it incorrectly. --NE2 12:26, 13 August 2008 (UTC)
I was using "as desired" to mean "as desired", not "in a different way to that which was desired". DuncanHill (talk) 12:32, 13 August 2008 (UTC)
That's not a reason to remove functionality, since the way it now displays in IE is the same as it would display with the multiple column functionality removed. --NE2 12:35, 13 August 2008 (UTC)
If by "remove" you mean take the aesthetic control away from the editor and give it to the readers, then yes, that is what I am proposing. If you look at the past discussions and this one, there is absolutely no agreement on the default font size that does apply to all browsers. My proposal is that we set it to a standard and let each individual user make their own decision. --—— Gadget850 (Ed) talk - 13:01, 13 August 2008 (UTC)
This discussion is about columns, not font size... --NE2 13:36, 13 August 2008 (UTC)
Why not allow editors and readers to share aesthetic control, instead of forcing editors to have no control whatsoever except by reinventing the current version of this template? Anomie⚔ 17:21, 13 August 2008 (UTC)
Thanks NE2, I was starting to wonder if anyone else here understood that point (despite how many times it has been told to them above). Anomie⚔ 17:21, 13 August 2008 (UTC)
I would agree with you on this:
  • if there were a way the editor could let readers see it as the editor designed; this is not possible for the 3/4 of the readers who don't use FireFox
  • if there was a consensus on the default font size (discussed multiple times)
  • if there was a consensus on the limit to the number of columns (discussed multiple times)
So far, the only consensus is to deactivate the multiple column feature for Safari until Safari is fixed. --—— Gadget850 (Ed) talk - 17:34, 13 August 2008 (UTC)
Who says that the readers have to be able to see it exactly as the editor does, as long as it's not broken? (By the way, it's the Gecko rendering engine, not Firefox, that supports it, not that that makes much of a difference.) --NE2 18:17, 13 August 2008 (UTC)

(unindent) "if there were a way the editor could let readers see it as the editor designed; this is not possible for the 3/4 of the readers who don't use FireFox" - but why should we make it impossible for the 1/4 with Firefox to see it as the editor designed? As long as it doesn't hurt IE users I really don't see the point.--Peter Andersen (talk) 20:30, 13 August 2008 (UTC)

Retrieved from "https://wiki.riteme.site/w/index.php?title=Template_talk:Reflist/Archive_7&oldid=1079008908"