Template talk:Reflist/Archive 27

colwidth obsolete
Why is it encouraged, as suggested, to remove named parameters, in this case colwidth, in favour of unnamed positional parameters? IMO that diminishes the readability of source text. I suggest to remove that point, which was introduced in the cited edit. -- Michael Bednarek (talk) 14:02, 15 February 2015 (UTC)
 * Because it is redundant. Since its inception, the columns feature was always ment to be controlled by the first unnamed parameter, but it was miscoded. Since that mistake has been fixed, it is simply redundant. I don't think 20em leaves any question about its meaning.  15:15, 15 February 2015 (UTC)
 * The only problem with "20em" being self-evident is that "2" might not be; the colwidth parameter still accepts either a number of columns or an actual width. I do understand that reading the instructions is the best course for anyone first using the parameter, of course. --User:Ceyockey ( talk to me ) 15:32, 15 February 2015 (UTC)
 * Actually, colwidth only accepts a valid width value, no fixed number of columns; only 1 allows that. Note that using a fixed number of columns is deprecated anyway in favor of dynamic columns (using column width).  16:02, 15 February 2015 (UTC)
 * I think the use of em is -- in my case -- unhelpful in the extreme. On both my old computer and my new computer it defaults to one column.  The forced two-column alternative is far better from what I have seen. My personal point of view. Epeefleche (talk) 06:17, 16 February 2015 (UTC)
 * The main problem is that you are forcing two columns on everyone's screen. So you have either a very small screen, or a very big font.  09:17, 16 February 2015 (UTC)
 * What browser and version are you using? --  Gadget850talk 12:10, 16 February 2015 (UTC)
 * Tx for the questions. My screen is 14" by 12"; 1024 x 819 pixels, 24 bit. The font is a normal one -- Times New Roman 20; not a "very big" one. I'm using Mozilla Firefox 35 on Windows 8.1, and it is up to date. Epeefleche (talk) 20:21, 16 February 2015 (UTC)


 * What do you see at Template:Reflist/testcases? --  Gadget850talk 21:08, 16 February 2015 (UTC)
 * Thanks, doctor. One col at 30em, two at 20em, four at 10em. Epeefleche (talk) 00:19, 17 February 2015 (UTC)


 * Then you should see the same in articles. If not, what arcicle? --  Gadget850talk 00:23, 17 February 2015 (UTC)


 * The problem for me, on my 14-inch-wide screen, is that refs (where we have more than 10) would look better as two columns. But 30em is typically put into articles where two columns would be better, and result in me seeing only one column. The forced-two-column approach is therefore -- for me, as a viewer on a screen over a foot wide -- the preferable one. Epeefleche (talk) 06:38, 17 February 2015 (UTC)


 * I get that behaviour (1 / 2 / 4 columns there) when I set my browser window to about 800 pixels with my default font of 16px; if I set the font to 20px, I get the same behaviour at the size you mention, 1024 pixels. Conclusion: 20px is indeed rather large. -- Michael Bednarek (talk) 06:36, 17 February 2015 (UTC)
 * Tx. It's denoted on my computer as "medium" font size. I could make it smaller than medium, but with my eyesight (which is 20-30 ... a reasonably good eyesight) medium is comfortable.  But two ref columns at that font size (which I get when 2 are forced, but not at 30em) in long ref lists are far easier to read -- for the same reason that newspapers like the NY Times have many columns. Epeefleche (talk) 06:47, 17 February 2015 (UTC)

That is helpful. Need to add a note about browser font size.

Firefox defaults to a font size of 16px: 30em gives two columns until the browser window is 11.4 inches wide. At 20px, 30em gives two columns until the browser window is 14.1 inches wide. But at 16px, 30em gives three columns on a 16 inch wide monitor. I can't test a larger monitor at the moment, but you would obviously get more columns. By setting the columns to 2, you force it to two columns for everyone.

Would it be possible to add a class so that users could add CSS to adjust the column-gap? Looks like the default is 1em, which is a bit generous. --  Gadget850talk 09:24, 17 February 2015 (UTC)


 * , users can already select  (or either   and  )to target reflists (or just   for other column templates), so there is no need for another class. 1em isn't that much though. The reflist documentation also explains how one can force two columns, see .   09:40, 17 February 2015 (UTC)


 * Which is why I defer to your CSS fu. --  Gadget850talk 09:43, 17 February 2015 (UTC)


 * I'm doing something wrong here. I can increase the gap, but I cannot decrease it unless I set it to 0. I tried 5px and other values with no effect. If I set it to 0, then the gap decreases but not to 0 (it does seem to fix the problem though). --  Gadget850talk 10:22, 17 February 2015 (UTC)


 * Seems to work on my end. You can remove ; it does not exist. You can also play with column-gap, and div col also has a gap parameter to play with. Note that with ordered lists (as used in reflist), there is always a minimum gap of 3.2em reserved for the (decimal) list markers.   10:48, 17 February 2015 (UTC)
 * was suggested when I edited my JS and added the others. And the list indent is what I was missing. --  Gadget850talk 11:02, 17 February 2015 (UTC)

Number of columns?
I was wondering, what's the criteria for deciding the number of columns to be used? Is there a rule or is it just up to the editors to decide? Illegitimate Barrister 10:23, 27 February 2015 (UTC)
 * See Template:Reflist. Common practice is to let browser software determine the number of columns. Betty Logan (talk) 10:25, 27 February 2015 (UTC)
 * Thanks. Illegitimate Barrister 10:32, 27 February 2015 (UTC)

Replacing bare references tag with reflist with no params
I often see editors replacing 18:01, 5 March 2015 (UTC)
 * Well, there's one difference -- the template is much harder to work with in VE. If there's no other reason, I won't feel guilty about replacing the template with the bare references tag when I'm working on an article (only if there are no params, of course). Mike Christie (talk - contribs -  library) 18:04, 5 March 2015 (UTC)
 * I think this may be ; see also . --  Gadget850talk 18:17, 5 March 2015 (UTC)
 * We should not be expected to change established practices in order to work around some problem with VE. Instead, VE should be fixed to conform. -- Red rose64 (talk) 20:53, 5 March 2015 (UTC)
 * I agree; reflist is a very valuable template, and I'm not asking anyone else to change their behaviour. I was just checking that it was harmless for me to change reflist to references when there are no parameters. Mike Christie (talk - contribs -  library) 21:01, 5 March 2015 (UTC)
 * User:Mike Christie removing the reflist removes the ability to simply add parameters when needed later. For instance, when the number of references pass say ten and columns may be desirable. In that natural article progression, it seems a retrograde step to undo part of that article progression. To do that only as a workaround for a bug elsewhere seems highly undesirable. Widefox ; talk 10:02, 3 May 2015 (UTC)

To reiterate: If VE has trouble handling a template used on 3.5 million pages, then the issue is with VE. --  Gadget850talk 12:19, 3 May 2015 (UTC)

Something just broke the 30em parameter
--LKAvn (talk) 12:17, 17 June 2015 (UTC)
 * Where did you see it break? What are the symptoms? -- Michael Bednarek (talk) 12:38, 17 June 2015 (UTC)

Using a bot to replace all static columns
Since it is depreciated to force the number of columns used, is it possible to have a bot replace all of them with 30em? Just curious, as I've been performing some other work on articles and am seeing a number of these cases. And each time I change it to 30em, since that is preferred. - Favre1fan93 (talk) 18:36, 13 August 2015 (UTC)
 * There may be articles where the deprecated specification of columns is a well-considered choice. -- Michael Bednarek (talk) 05:35, 14 August 2015 (UTC)

Change default behaviour to
This is the default behaviour for the majority of pages. Making it the default behaviour for the template would simplify our vast task of coding pages. Convention over configuration.

instead, I don't care.

If any page needs something different, then an explicit parameter would of course override this default behaviour. Andy Dingley (talk) 17:04, 16 November 2015 (UTC)
 * It is being considered, but at the MediaWiki software level.  17:26, 16 November 2015 (UTC)
 * Well that's an obviously stupid idea. 8-( Andy Dingley (talk) 21:39, 16 November 2015 (UTC)
 * Any further word on this? The default should definitely be 30em. Alex &#124; The &#124; Whovian  01:41, 24 December 2015 (UTC)
 * There may be articles where he omission of any parameter, relying on the current default behaviour of no columns, is a well-considered choice. -- Michael Bednarek (talk) 04:32, 24 December 2015 (UTC)
 * Yes, such as when there are few refs, and all those refs are long. -- Red rose64 (talk) 23:45, 25 December 2015 (UTC)
 * Then those few articles can have their own parameter set for a single column; however, this would benefit the majority of articles using this template. Alex &#124; The &#124; Whovian  00:55, 26 December 2015 (UTC)
 * How do you know that only a few articles use deliberately the current default behaviour? How many do you think is a few of 3,600,000 uses? -- Michael Bednarek (talk) 03:11, 26 December 2015 (UTC)
 * You use template usage to find counts, behaviors, of template users. &mdash; Cp i r al  Cpiral  04:08, 27 December 2015 (UTC)
 * That doesn't help much when assessing the default, though. If I just use when I create an article, is that because I want the default behaviour (which is currently one column), or because I only have one reference at that time (and thus might approve of changing once there are more), or because I just don't know about the available parameters, or because I just don't bother? Whereas using  is a much clearer statement that I want a one-column reflist (or possibly that I copied code from someone who did). Nikkimaria (talk) 19:20, 27 December 2015 (UTC)
 * Notelist removes most of the use cases for single-column Reflist. Also note that setting column width need not (as it tends to at present) force multiple columns for short lists. A short list can still be narrow, but a single unbroken column. Andy Dingley (talk) 20:16, 27 December 2015 (UTC)

references tag
I have been using 08:00, 16 May 2016 (UTC)
 * After experimenting, the explicitly 2-column example is easy, but forcing the 30em example to wrap was difficult. Works for 20em as a demonstration, though. User:jmcgnh/sandbox/Template_Reflist  --jmcgnh  (talk) (contribs)  08:07, 16 May 2016 (UTC)
 * The examples shown do not necessarily need to use the exact code; the examples are for illustration only.  09:16, 16 May 2016 (UTC)
 * OK, I'll go with what I've got except I won't explicitly point out the 20em vs 30em, in the spirit of for illustration only.  --jmcgnh  (talk) (contribs)  07:49, 17 May 2016 (UTC)
 * Alright, Edokter, I see what you did there. Yes, it makes it all clearer, with less discrepancy. Thanks for letting me do something that others might find useful or illustrative.  --jmcgnh  (talk) (contribs)  01:49, 18 May 2016 (UTC)
 * Of course; this is Wikipedia.  06:44, 18 May 2016 (UTC)

Open Access template not showing up inside reflist
I added a newspaper.com clipping to a citation in Gerri Major, and included the Template:Open access template at the end, per the instructions at WP:Newspapers.com. The icon that should appear (File:Open Access logo PLoS white.svg) as a result doesn't appear; hardly surprising given that this is being parsed inside Reflist. Is there a way to make that icon appear at the end of a reference, or could a true/false switch be added to make the icon appear? Mike Christie (talk - contribs - library) 14:34, 22 May 2016 (UTC)
 * Shouldn't have been place inside the  tags?
 * —Trappist the monk (talk) 14:43, 22 May 2016 (UTC)
 * D'oh. Yes.  Fixed.  Thanks. Mike Christie (talk - contribs -  library) 14:48, 22 May 2016 (UTC)

I have an idea for a change: Add the 

 template at the top of the source code.
That way, an object (like an image) in the previous section of an article will not disrupt the columns. A small obstruction in just a small part of the top of the list will cause the entire reflist to get pushed to the side. I see no reason why this shouldn't be done. FabulousFerd (talk) 21:34, 23 July 2016 (UTC)

Oh... now I see a problem. The ==References== heading won't follow this rule. So, it will stay in a position and the ref list will be much lower in case there is an obstruction. In some other wikipedias, the template automatically creates the ==References== heading, so in that case, it would work. Shame. FabulousFerd (talk) 21:38, 23 July 2016 (UTC)


 * Oppose Reflist} normally follows a   heading, and any CSS clear should happen on that element (i.e. before the heading), not after.
 * Also it's common that a Commons category will be used, and CSS floated, between the heading and the Reflist (if this is the last heading on the page, i.e. there is no  section). The added clear would thus introduce unwanted whitespace. Andy Dingley (talk) 22:09, 23 July 2016 (UTC)


 * Oppose In the present state the relist will appear on the place where it is added. With the clear-option it will be pushed to the bottom (after all sidebars or picture), leaving ugly white spaces. The Banner talk 20:18, 25 July 2016 (UTC)