Wikipedia talk:Manual of Style/Images

Fixing images below the default size

This used to be deprecated in the MOS, and it certainly should be (sorry, can't provide a link - I'd be grateful if anyone can). I don't want to re-open the vexed issue of fixing at higher than the default 220px, which we currently deprecate, but like many people I routinely do this, at least for main images in the lead. The case against smaller-than-default images seems much simpler - is there ever a good reason for doing this, for images with a typical aspect ratio? I can't think of one, and have for years removed all examples of "120px" etc that I see, & I don't remember anyone ever complaining. There is an exception needed for images eg 10 times taller than they are wide, but I think the existing text covers that fine. However it gives the clear impression that too small images are fine with the MOS.

Proposals

• A) At the moment we say: "As a general rule, images should not be set to a larger fixed size than the 220px default (users can adjust this in their preferences). If an exception to the general rule is warranted, forcing an image size to be either larger or smaller than the 220px default is done by placing a parameter in the image coding."

I propose changing this to "As a general rule, images should not be set to a larger or smaller fixed size..." (new text in bold). Any objections?

• B) I'd also like to add something specifying that this applies to multiple images, which seem (unfortunately in my view) to be fashionable at the moment. So at the end of the list of bullet points, I'd like to add:

"* Multiple image templates should not be be over-used, and each image should appear at at least the default image size."

Please comment on these below, specifying A & B. Thanks Johnbod (talk) 21:57, 12 June 2015 (UTC)

Discussion

• Fixed image sizes should be avoided in general, but there are cases where they're appropriate, such as in tables. Exceptions should be exceptional, though, and I think they are covered by the "as a general rule" wording—so, basically, I agree with what you've presented. Curly Turkey ¡gobble! 22:24, 12 June 2015 (UTC)
• I've seen fixed images done on infobox images all the time. I agree that outside tables (which infoboxes technically are) fixing the size should be avoided. --MASEM (t) 22:43, 12 June 2015 (UTC)

  Images in infoboxes count as lead section images, which have always been permitted to be larger - up to 300px wide. As regards images in actual tables, see Rolling stock of the Bluebell Railway#Steam locomotives, where the images are 200px wide, including the upright ones - which makes some table rows excessively tall; compare List of Great Central Railway locomotives and rolling stock#Mainline steam locomotives (175px) or Rolling stock of the Kent & East Sussex Railway (heritage) (150px). --Redrose64 (talk) 22:53, 12 June 2015 (UTC)

Can we get away from the usual issue of whether fixed sizes are good, bad, or downright evil, to address the question of whether this page should continue to use language that implies that images fixed small are better than images fixed large? I agree table images are another exception. Johnbod (talk) 03:23, 13 June 2015 (UTC)

• Oppose. I would oppose both as instruction creep, especially the multiple-image template suggestion, not because I like small images (the opposite is true), but because these decisions should be left to the people writing the page, not imposed centrally. Editors forget that the MoS is just a guideline, and go around trying to force it on articles in which they otherwise have no involvement. Every additional rule creates another weapon. This makes the MoS strongly disliked (e.g. see the recent discussions about creating a central style board), which is unfortunate because it's a very helpful document for style advice. Sarah (SV) ^(talk) 00:28, 13 June 2015 (UTC)

As my comment above. Johnbod (talk) 03:23, 13 June 2015 (UTC)

• Comment. This discussion has caused me to discover relative sizing (|upright=) which I didn't know about before but seems preferable to absolute sizing in almost all cases. I had thought that gif animations and bitmap images smaller than the default size needed absolute sizes to allow the animation to work and prevent being resized to larger than the resolution of the image, respectively, but if that was ever true it doesn't seem to be any more. However, there doesn't seem to be a way to use |upright= within {{multiple image}}, and there are probably other cases where absolute sizing is still important, so I wouldn't want to see a blanket prohibition. On the other hand, the same reasons that larger-than-default absolute sizes are bad make smaller-than-default sizes bad as well, so expanding the recommendation about fixed sizes to include smaller-than-default ones seems harmless. If we're going to make this change, it would be simpler to say simply that "as a general rule images should not be used with fixed sizes". The part about whether the size is larger or smaller than the default is a red herring and should be left out; why is using a fixed size equal to the default any better of an idea than the other two cases?—David Eppstein (talk) 00:38, 13 June 2015 (UTC)
  • I agree with Mr Epstein—the wording should be more like "As a general rule, images should not be set to a larger or smaller fixed size..." (I think that addresses SV's concerns about instruction creep as well) Curly Turkey ¡gobble! 03:45, 13 June 2015 (UTC)

I have used smaller-than-default-size images on occasions, when the infobox equates or exceeds in length the text on the left. I figure that a smaller image facing that long infobox will be less offensive to the anti-sandwitching purists who believe that no images should ever face a sacrosanct infobox. The only other solution is to place the image below the infobox, and out of view --Lubiesque (talk) 12:07, 10 November 2015 (UTC)

{{Images}}

template:Images has been proposed for deletion, see Wikipedia:Templates_for_discussion/Log/2015_October_12#Template:Images -- 70.51.44.60 (talk) 03:45, 13 October 2015 (UTC)

Left placement - inaccurate guidance

In most cases, images should be right justified on pages, which is the default placement.

To my ears, "most cases" means well upwards of 50%, probably at least 75%. If this statement reflects community consensus, why do so many good articles (GA) use both left and right placement? Many of those that don't only have one or two images. The above statement is closely followed by:

Multiple images in the same article can be staggered right-and-left.

In other words, the guidance here is: If you think x is better, do it. If not, don't do it. This is not guidance, and the effect would be the same if you removed both statements. Clearly, the community likes the second statement more than the first, and the guidance should be modified to reflect community consensus, regardless of how we feel about the issue individually (this is not a request for opinions about left-right placement). ―Mandruss ☎ 04:49, 14 October 2015 (UTC)

Mandruss, did you become aware of this matter because it is addressed on my talk page? On my talk page, I directed Hike The Monicas here to this talk page to discuss this topic. If he is reading this now, he should not continue that discussion at my talk page, especially since it got off track.

As for your "[W]hy do so many good articles (GA) use both left and right placement?" question, many WP:Good and WP:Featured articles do not comply with each and every guideline, especially when editors go in and change the articles to be a certain way well after the article reached its WP:Good or WP:Featured status. For this case, what proof is there that "the community likes the second statement more than the first"? And why would you start this section and then state "this is not a request for opinions about the issue"? People are obviously going to weigh in on this matter here at this talk page if they want to. Flyer22 (talk) 05:02, 14 October 2015 (UTC)

It used to be the case that articles with all images on the right would be criticised at FA, & pressured to change. I'm not sure this still applies. Now screen sizes are so varied, the ideal image placement is in rather a mess. Johnbod (talk) 12:31, 14 October 2015 (UTC)

• I knew of nothing about this on your talk page. Sometimes people think about the same issues independently from each other.
• Your point about changes after GA approval is valid, although there must be a ton of that going on. I guess there's no way to really know, short of polling a large number of GA reviewers.
• Maybe an RfC is needed.
• Being clear about my intent here is not an attempt to dictate what is discussed.
• Your tone, frankly, seems a bit confrontational, accusatory, and un-Wikipedian. I came in good faith with an issue that I feel is important, not to get involved in a fight. ―Mandruss ☎ 05:23, 14 October 2015 (UTC)

Apparently, my tone these days naturally comes off as "a bit confrontational, accusatory, and un-Wikipedian" to certain editors; that was not my intention in this case. In my experience, though, from what I see on this site day in and out, confrontational and accusatory are very Wikipedian. Flyer22 (talk) 08:01, 14 October 2015 (UTC)

The guidance is to place images on the right, unless there is a reason to place them on the left, such as to stagger images, to avoid stacked images, or to have people face the text, and so on. DrKay (talk) 10:59, 14 October 2015 (UTC)

If that's the intended guidance, it's far from clear. Also, many editors believe that judicious use of left-right placement is more visually attractive, and they would consider that a reason. If the guidance excludes that reason, it should say so. ―Mandruss ☎ 12:13, 14 October 2015 (UTC)

It could be sharpened up. The guidance probably has articles with few images mainly in mind. In particular it might say that the lead image should normally be on the right, which is a very strong convention. Johnbod (talk) 12:31, 14 October 2015 (UTC)

I believe that "most" is meant to mean "more than 50%", but not necessarily much more than 50%. For an article with a few images and a lot of text, then having three on the right and one on the left (Mandruss's example of 75%) would be reasonable. However, it's necessary to take all the facts and circumstances into mind, e.g., the presence of an abnormally long infobox.

Is there a difficult dispute going on here? This seems like something that is usually resolved by editors talking it through. And, since someone mention GAs above, then I note that compliance with this guideline is not on the Good article criteria. WhatamIdoing (talk) 23:41, 14 October 2015 (UTC)

Good point as to GA criteria, and I retract that argument. I was under the mistaken impression that GA represents a de facto community consensus on this issue. Sure, we can go with local consensus on this, which will be unknown to or ignored by many, and will endlessly ping-pong back and forth as the local mix of editors changes. I was hoping to avoid all that as unnecessary. Regardless, the current guidance here does not facilitate a local consensus as it can be used to support either position with pretty much equal strength. Like I said, guidance that nets out as, "Do what you think is best" is not useful guidance and only complicates matters and contributes to instruction creep. ―Mandruss ☎ 00:09, 15 October 2015 (UTC)

Band timeline images

A discussion is happening at Wikipedia talk:WikiProject Musicians#Create Member Section/Timeline Standards related to standards around generated timelines. The suggestions seem to violate the suggested size guidelines. It would probably be best if interested parties could comment to either support the 800 pixel width suggestions or give reasons against them. Walter Görlitz (talk) 05:43, 10 November 2015 (UTC)

Parenthesis

Are image captions supposed to be in parenthesis? Is this the proper caption "(c.1908)" or is this the proper caption "Tiffany circa 1908".

I'd put "Tiffany c. 1908" myself. Having it all in parentheses is odd. Johnbod (talk) 12:35, 7 December 2015 (UTC)

"Articles about ethnic groups or similarly large human populations should not be illustrated by a gallery of images of group members"

Sandstein added the following to the page: "Articles about ethnic groups or similarly large human populations should not be illustrated by a gallery of images of group members, because selecting them is normally original research, and often contentious." It was added per the WP:Consensus formed in this discussion. But that close only concerns ethnic/"race" matters, not all topics about large human populations. This is why I stated at Talk:African Americans, "I'm not sure how the WP:Consensus from that discussion will hold up, given that we still have such galleries at the Man and Woman articles, etc., which are just as subjective, but it's the WP:Consensus for now." Sandstein's wording here at the guideline is broader than the aforementioned close. Furthermore, I don't see how the ethnic/"racial" images are normally original research; in some cases they are, but they are most often based on what WP:Reliable sources state. For example, the inclusion of Mariah Carey as African American, which resulted in a big dispute at Talk:African Americans, is based on sources; this is also made clear by this RfC at Talk:Mariah Carey.

I'll alert Wikipedia talk:Manual of Style to this matter for wider input. Flyer22 Reborn (talk) 21:56, 4 January 2016 (UTC)

My impression from the RfC was that while the question posed was about ethnic groups, the arguments advanced were such that they applied to all large human populations (OR, contentiousness, and the question of whether a select few individuals can visually represent a large group). But if others don't see it that way, I'd have no problem with striking " or similarly large human populations" here.  Sandstein  22:02, 4 January 2016 (UTC)

A possibly useful (local) consensus was reached in this discussion at Hungarians around this issue. Basically, when individuals are specifically discussed and their relevance described, portraits complement the text. But an arbitrary gallery of individuals used simply to represent the group as a whole adds little value and invariably breeds conflict over which individuals should be included. Probably won't defuse all disputes around this issue, but this approach puts the onus on those who insist on including an image of a given person to actually develop the article to include that person's relevance. Laszlo Panaflex (talk) 22:08, 4 January 2016 (UTC)

• Support a broad application of this principle. These "galleries of notables" are not useful and do nothing but generate strife of multiple sorts: Whether a give person qualifies, whether they're exemplary, whether a mixed-background person is having one aspect of their background overemphasized (the central issue in the Carey debate, and one that is not resolved despite suggestions above that it is), who gets to be included, whether the mix is balanced (modern and historical, arts and sciences, political and non-political, male and female, etc., etc.), how many to include, and so on. It's also lead to really excessive treatments (some of the historical versions of the article White people were pretty crazy). These infobox galleries really don't do anything useful for the reader. More often than not they simply distract, by injecting "why is that person included here?" skepticism into their reading experience. I agree with the above that inclusion of people in the text is more helpful, where the context requires a sourced justification for their inclusion. The purpose of infoboxes is to summarize the gist of articles, not to serve a back door for the insertion of OR and trivia. In the end, this is very much like flag icons, and it shares many of the same rationales against inclusion, especially PoV-pushing overemphasis and a propensity to generate frequent conflict that has little to do with making the encyclopedia better.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  12:28, 5 January 2016 (UTC)
• Support the general principle. The wording could be adjusted to more clearly reflect the idea that it is the "galleries of notables" that are inappropriate; individual images of representative (non-notable) samples from a population are ok. —David Eppstein (talk) 18:46, 5 January 2016 (UTC)

Sandstein, Laszlo Panaflex, SMcCandlish and David Eppstein, my thing is: The current wording is overly broad when the consensus discussion was not that broad, no matter that Sandstein states that he interpreted it that way. The word similarly can be interpreted in different ways. I noted that "I don't see how the ethnic/'racial' images are normally original research; in some cases they are, but they are most often based on what WP:Reliable sources state." This means that I don't like the implication in the current wording that it's usually WP:OR that's the issue; that's not the usual issue. If a source exists calling the person a certain ethnicity/"race," then that's not WP:OR. The usual issue is people preferring certain celebrities/notables to other celebrities/notables and thinking that one person is more deserving of inclusion than the other, which includes feeling that one person looks more white or black, etc. than the other person. It has more to do with POV than WP:OR. And this does not only happen with notables. Stating that this gallery matter should only apply to notables will not stop the same POV-pushing from happening to no-names being included in such a gallery. Furthermore, I don't see how one can regulate this. For example, at the White people article, images of notables are still currently there, even after I left this note. If people see that one article (like White people) gets to have such images, they will feel that the other article (for example, Black people) should as well. If it's a gallery of no-names, someone will eventually add a notable person, and then someone will add another and another notable person. And I reiterate that this same type of POV-pushing has happened at the Man and Woman articles. In the case of the Woman article, see this discussion and keep scrolling down; the idea of including a trans woman is constantly at debate at that article's talk page. I don't see that these image matters will be regulated too successfully, if successfully at all, especially if it is applied to more than just ethnicity and "race." And another important aspect is that if an editor feels they are only allowed to add one lead image in cases such as these, that one lead image will become a source of contention as not being representative enough. The desire people have for lead images, for those images to be representative, and to include notables instead of no-names, will likely overrule this guideline. It will be another guideline that barely anyone follows. Flyer22 Reborn (talk) 22:13, 5 January 2016 (UTC)

True, the original discussion was not that broad; that's why we're having another discussion now about applying the same principles more broadly. As for OR, I'm not so much worried about that (as we can generally find sources attributing membership of people in these groups) and more worried about two other issues: (1) the fact that the notable people of a group are not representative of that group, almost by definition, since they're the ones that stand out, and (2) the often-contentious issue of who to include in a space that is generally too small to list all notable members and where many editors are likely to have political reasons for pushing to include some specific people and exclude others. —David Eppstein (talk) 22:22, 5 January 2016 (UTC)

(edit conflict) The fact that the RfC wasn't as broad as what's contemplated now is why we're having the discussion – to see if there's consensus to extend it. The fact that the consensus might not extent that far isn't an argument against the proposition, it's a precondition of it. I can attest to quite a number of NOR issues being raised in debates about this stuff, though I agree NPOV is more commonly implicated. Both being implicated is significant, and even if only one of them was, it would still be reason enough to act against this misuse of infoboxes. If people think that some kind of "gallery of qualifying faces" is needed, they can use the Template:Ntag feature, in the main body of the article, and provide source citations at each entry in it, to forestall disputes and confusion. I entirely agree with you that limiting these infobox galleries to "notables" would not fix the issue. How would we be including non-notables anyway, if there aren't any articles about them to link to; just pasting in random faces from Commons pictures? I'm skeptical that inclusion of non-notables is a big issue, though I have not trawled every single discussion about these things, so I may have missed something. I agree that there will always be debates about images included in such articles. If they're kept out of a tight-grid infobox, and required to be sourced in the article, this greatly reduces the potential for POV, OR, and related disputes about these pictures. For example, the obvious way to deal with including a transwoman at the Woman article is to have a section for that (a short one linking to the main article on that, per WP:SUMMARY) with one or more photos that pertain, in that section. No issue of "should there be a TG woman in the infobox" every arises in such a case. It's all a WP:COMMONSENSE matter to me: If people are fighting about some style matter that is not actually important to the encyclopedia, look for a way around it by removing what they're fighting about since it's not essential – like taking a toy away from two children who can't stop squabbling over who gets to play with it. A potential solution for an article like African Americans is to rotate who appears there. We could probably set up a script that auto-rotated them from a pool of images, though that might require a Lua module. Or just nominate a new image of the month, or whatever. For such articles, having one male and one female for representativeness purposes should work, but we don't need a PoV-pushing and OR-infested large collection of images there being included for dubious reasons.  — SMcCandlish ☺ ☏ ¢ ≽^ʌⱷ҅_ᴥⱷ^ʌ≼  22:34, 5 January 2016 (UTC)

(ec) There's always the possibility that guidelines are not followed. But at least in the cases covered by the RfC there was the active desire of many editors to get rid of these galleries because of the problems associated with them, so I think it's worthwile to note the principle here. That doesn't change that case-by-case dicussions may need to be had for many pages. Regarding OR, I too read the discussion as highlighting the OR problem of how to decide who to include, rather than whether certain individuals belong to the group, which is more often verifiable.  Sandstein  22:35, 5 January 2016 (UTC)

Per what I stated above, I just don't see this guideline working. There are guidelines that are commonly followed, and then there are the ones commonly ignored. If this guideline were to work, I agree that it would obviously result in a lot less arguments. As for the case-by-case basis matter, that will turn into a "Since that article has a gallery for the lead, this one can too." argument. Again, see my White people article vs. Black people article example above; that aspect will become a reality if the White people article is allowed to have images. As for original research, the WP:OR policy is based on the WP:Verifiability policy; I don't interpret it as being any broader than that. Regardless, per what I stated above, I think that the wording you added needs improvement. Flyer22 Reborn (talk) 00:25, 6 January 2016 (UTC)

What will almost always be OR is the selection of which images are included and the resulting implication that these are the most notable. Very unlikely to find a RS that says "these are the 16 most notable Xs." So the resulting selection is necessarily OR and generally POV as well. Laszlo Panaflex (talk) 00:31, 6 January 2016 (UTC)

Laszlo Panaflex, I see what you mean, but I don't interpret the WP:OR policy as policing infobox images in that way, except if an editor were to state, "These images should be in the infobox because they are the most notable people according to reliable sources.", or something similar to that. Otherwise, we freely select infobox images, and, per WP:Pertinence, have more leeway with them. Of course, no source is going to state that we need to have a certain number of people in a Wikipedia infobox. In the case of people, a person generally should not be included as white, black, etc. unless a reliable source validates that inclusion. I state "generally" because there is the case of no-names, who are not likely to have sources verifying their ethnicity or "race," and some people will argue that a person is obviously white or black, etc.; I also state "generally" because stock photos have been discussed at WP:Med before, and some editors still are not sure if we should have an image of a person looking like they are suffering from a certain medical condition if they aren't; see Wikipedia talk:WikiProject Medicine/Archive 67#Wikipedia to promote stock health photography. Flyer22 Reborn (talk) 00:54, 6 January 2016 (UTC)

• Oppose removal If you have issue with one of the images, get consensus to change that one image. Ethnic looks are diverse, one image does not do the trick. Unless you are going to devise some ethnic test for Wikipedia, like South Africa's "one drop of blood" law or Germany's "1/8 Jewish or less" law, people have the right to self identify for their ethnicity and their sexual orientation. --Richard Arthur Norton (1958- ) (talk) 22:49, 5 January 2016 (UTC)

Richard Arthur Norton (1958- ), your comment seems to be arguing against Sandstein's addition. If so, I think that that using "Oppose" for your comment is better than using "Keep." Also take note that I did not have a problem with the result of the aforementioned discussion (other than it seeming to assert that WP:OR is the main issue). My issues with the matter are noted above. Flyer22 Reborn (talk) 00:25, 6 January 2016 (UTC)

• I would think that the advice to avoid including any photo of a person of race/ethnicity/gender/whatever that is identifiable in these galleries should be avoided for reasons stated above (editors pushing to include a specific person of note), but I'm not sure if outright removal is needed. If you eliminate the use of notable persons, you are left with images to select from that are 1) free and 2) verified to be people that are part of that group, and then the choice should then become a matter of presentation of what are the best images of this combined set. Which while still a potential for editorial dispute will far less be a problem as when notable persons are involved. Obviously, in considering race and ethnicity, editors should strive to cover both genders and all ages with the available pictures, though there's no need to force a specific balance (50/50 male/female, 25/50/25 young/adult/senior etc) just as long as it looks good. --MASEM (t) 00:54, 6 January 2016 (UTC)