The top comment from yesterday's HN submission mentions how you can configure Wikipedia (or any MediaWiki) page to use whatever theme you like, including all the previous themes, which are still actively maintained: https://news.ycombinator.com/item?id=34431533
The navigation frame on the left is absolutely massive. At first I thought you linked the mobile view, but no, it really is wasting a quarter of the width of the page (on my monitor) on white space.
The main feature of the new design is that the width of the article body is now limited, rather than unlimited as previously, under the idea that overlong lines reduce the readability of text. With limited text width, removing the navigation sidebar would not decrease the amount of whitespace. If you would prefer to have unlimited width, you can configure this behavior in the new theme or select a different theme altogether.
They tried this already in 2014 as part of the Typography Refresh™ package, but reverted it rather quickly after a backlash. Likewise line height changes and not-quite-black black text. Only the relatively benign font changes stayed.
Only if you create an account and log in, right? I've gotten along without having a Wikipedia account for the last 20 years, I guess now I have to make one. If that was a product goal, it was a very sneaky success.
As mentioned in the above HN link, you don't need to log in. If you're annoyed by the logged-out approach of appending a query parameter to the URL, other people in sibling threads shared Greasemonkey scripts (or whatever they're called these days) for doing so automatically.
I'm pretty sure that "getting the same four Hacker News users, who feel personally wronged every time a website becomes more readable, to sign up" is nowhere near the top of their list of priorities.
Grrr, I hate readability! Just the idea of proper line lengths makes me angry.
But I think everybody has to create an account, not just us salty grandpas. Like I said, if it was a product goal, it's a sneaky way to do it. Not saying that is the case.
It's also the perfect width to add a friends list or sidebar full of skeuomorphic widgets or live updating feed of edits to the site. There's no reason to assume Wikipedia would introduce advertising (they're not a corporation, they're not motivated by profit, they have a healthy chunk of cash in the bank). Of all the organizations I'd expect to give in to greed, Wikimedia doesn't crack the top 100.
People keep making that claim, and yet when I ask for example of lying, I never seem to get anything.
Just asking for money while you already have money is not inherently lying. You might not like it, but lying involves making untrue statements. "We need your money" is not untrue, even if you already have a lot of money. "We need your money or else we'll shut down" would be untrue, but all of the pitches I've seen don't include anything like that.
No, it sounds like running one of the most popular websites on the Internet requires well-paid employees, benefits, and infrastructure investment. 300 employees is two orders of magnitude fewer employees than Microsoft laid off this week. 300 employees to run one of the most trafficked and useful websites ever created is a strikingly small number.
Yes, I want wikipedia to have redundancy. Yes, I want Wikipedia to be resilient against incidents. Yes, I want them to have lawyers to respond to legal issues (both founded and malicious). Yes, I want them to invest in new projects that may or may not pan out into useful services like the ones I already enjoy (it's hard to appreciate how many Wikimedia offers).
If you told me that they spend $100M each year to keep the site online and running well, I'd believe you. A third of that is surprising because it's an astoundingly small number given the sheer amount of value that they create with it, especially for me personally.
You have a wildly distorted view of what it takes to keep the lights on for a heavily trafficked website. All of the people you mentioned might be volunteers, but they only volunteer because the platform exists and works well. And it continues to exist and continues to work well because people are paid to do it.
And if you disagree with that, then where are all the top 1000 websites that are run by "a few" people?
What?
How much work do you think it is to run a HTTP server that serves static content in a datacenter? I've ran moderately popular websites before and it's zero work most of the time.
In 2005, Wikipedia ran on $5000 with one employee and over a billion views just fine. People did, in fact, contribute to it despite money not being thrown into mud.
> And if you disagree with that, then where are all the top 1000 websites that are run by "a few" people?
Where are all the static HTTP websites that are ran by dozens of people? At most SaaS companies, keeping the lights on for the HTTP server is the least work you do.
Your view of the world seems strongly distorted by zealous belief in FAANG overstaffing.
> run a HTTP server that serves static content in a datacenter
It's not static content!
> I've ran moderately popular websites before and it's zero work most of the time.
I run a service that gets ten million requests on a bad day. There's almost always work to be done.
> Where are all the static HTTP websites that are ran by dozens of people?
Most static websites don't allow anyone to sign up for an account and mutate the content.
> At most SaaS companies, keeping the lights on for the HTTP server is the least work you do.
Most SaaS companies aren't serving a multi-terabyte mutable dataset to billions of people a month. They also aren't dealing with tens of millions of pages of user-generated content. Or a huge volume of legal requests because of all of their user-generated content.
> zealous belief in FAANG overstaffing.
Which FAANG company has only 300 employees?
Seriously, 23 billion page views per day. That's like if each of their employees oversaw a service that got 76M page views per day. That's a quarter million page views (not requests!) Per second. Half a billion pages on Wikipedia alone: that's a few million pages per employee.
yeah I use Wikipedia zoomed in and it always uses the entire width of my screen, new design is not an exception. Having text as a max-width column is good, otherwise it's harder to read.
By making text flow to 100% of the width you can easily resize the browser window for everyone's perfect reading width.
I usually run my browser tiled to 50% of the screen, the new design wastes a lot of space at that size (you have to shrink to really narrow width before it enters 'mobile mode' and uses 100% of the width again - I mean it's pretty insane, you can see more text at say 30% width, but at 40%-60% it pops out that useless white space padding on the left leading to a narrower effective text width)
Yeah, I just tried it out and it looks pretty good! (If I resize my window to be 1/3 the width of my monitor, that is.)
I totally agree with you: I like to use my browser at ~50% of screen width on a 27" screen, so it looks terrible there because of all the wasted space. If they made it work at 50% width like it does at 30% width, I'd have no complaints.
There isn't wasted space. The left supports the table of contents, which can get quite long and layered. Plus it also supports the expanding top menu button. The right side can be expanded on by clicking the square button at the bottom if you're not a fan of the "narrow" reading mode.
Maybe we aren't seeing the same things, To the left of the table of contents there is a lot of wasted space still. and to the right there is even more wasted space.
Navigating is also slow, it takes a second to scroll down, while before it was instantaneous.
The button is at the bottom of the screen to the right. Visit an example page [1], and it should be there at the very bottom right corner. It is set to only appear on wide screens (1600px+), where the "wasted space" can actually be utilized.
Seems being directly linked to pages doesn't show the new design sometimes if you aren't logged in. But I'll provide the page where they talk about the new theme itself and the expand button [2].
I wasn't expecting any controls to be outside the main grey area, so it's a bit weird the content width button is placed where it is. Also, I wonder if defaulting to wide mode when a wide monitor is detected is a better choice.
Either way, I don't have a strong opinion on whether its an improvement or regressive. Pretty much everything nowadays seems to be converging on a narrow column of text.
That's completely false. Earlier, when first viewing a Wikipedia page today, I experienced the same (very) mild panic that typically accompanies visiting the mobile page by accident. A regression in the visual appeal of the layout, in its readability at a glance, and in the efficiency of its use of space. Earlier design could be considered "iconic", so this change is especially detrimental.
I also thought I was served the mobile site at first, but after using the new site a bit I actually quite like it. One of my longstanding complaints with the old design was that using the full width of the screen for text is very hard to read on a modern monitor, so I'm glad to see they narrowed it down a bit. I also really like the floating table of contents.
I had a similar reaction. I was so surprised that I spontaneously pulled back from my chair when I stumbled upon the redesigned site earlier. It was a "wtf is this??" moment.
I don't know if it's just that I'm not used to the new look, but I honestly hate the pillarboxing to fit the new maximum content width.
Clickbait headlines get justifiably trashed on HN, but this is an example of a headline that is creative and playful while also communicating the essential information:
- First word says exactly what the article is about: Wikipedia.
- First two words give more precision and the news itself: "Wikipedia's Redesign."
- First sentence communicates it was a minor redesign.
- Second sentence communicates it was intentional, and there will be rationale shared.
This type of headline writing should be expected, but sadly it's the exception.
I've been using the "Modern for Wikipedia" chrome extension for a while and it seems like they Wikipedia team has borrowed a lot of ideas from the extension. I still prefer the extension since you can change the theme away from bright white. Great changes overall.
I have been using the Timeless skin [0] for quite some time now and the main reason I switched was that the content would be a bit more condensed towards the middle of the page. This helps a lot on larger monitors and I can see why this new theme also does something similar. However, the large amount of whitespace on the sides is just wasted and in combination with the fact that more interaction is required to do basic things (like changing the language of a page) is something that makes this new theme not "barely noticeable".
Having used it briefly, I'm not a fan of the redesign (which I think the headline has wrong - I think it's very noticeable) as it is currently.
I understand the goal is to limit the line width of articles to make it more comfortable, and I don't disagree with that, but the content feels poorly framed now, just floating in the abyss of 'empty space'. The sticky side nav (which I think is a great idea!) feels unconnected from the article itself.
I think the basis is good, but final execution is lacking. I think with a few relatively minor tweaks it could be even better.
Edit: I think the page looks better with the hamburger menu open, with the different background give the article a bit more structure to sit against.
So many knee-jerk reactions by people disliking the change, so I want to give another view: I really like the new design.
- When reading an article, I want to have absolutely minimal visual noise. Removing the old sidebar helps.
- The smaller column width helps with readability, because it's easier to scan back to the start of the next line.
- It's great to have the table of contents always visible, to quickly jump around the some of the miles-long articles typical for Wikipedia.
- I don't need anything else on the page - so I don't have any complaints about "inefficient use of space". On the contrary - more whitespace helps me focus on the article.
- Unlike with Wikiwand, there are no distracting animations.
I strongly dislike that the table of contents have folded entries. I now have to click to get the link to a subheading.
In general, for quite a few things I need extra clicks to get to it.
Also, as I mostly print articles[1], I had a special printing layout set up to print Wikipedia articles. I need to come up with a new layout (reader view for Wikipedia used to be terrible).
Clearly designed by someone who only uses one language.
The one most used element for me was the side bar, where I could hover over the language to see the translation of some term. Now it's hidden behind a click, and a looot of scrolling. Just try to reach Svenska.
I can work around the style changes easily with custom CSS and stylus, but I'm not impressed by that change.
Thank goodness they are not as hung up on having a button as it initially appears:
"We are currently considering adding single-click functionality for the most-frequently used languages."
Sadly, while that means that the overall experience will be improved, it will solve exactly nothing for those who are unlucky enough to use a less-frequently-used language.
I think they might mean most-frequently used of that user (I think the "suggested languages" already are supposed to work like that.) But that presumably requires an account.
It has already taken multiple clicks for users wanting to read English Wikipedia in a language that wasn't one of the nine in the sidebar. In fact, clicking the "287 more" link at the bottom of the left sidebar in the old design brought up the exact same box that clicking the "297 languages" link at the top of the new design does, with the same list of languages in the same order.
Alternatively, one could create an account and set a default language, or visit a wikipedia site dedicated to that language.
I've always seen all the language links in the sidebar without clicking anything. I can see how, if you have had a different experience, the new arrangement might seem less of a disruption.
Weirdly for me, it seems to have been rolled out inconsistantly - maybe it's in progress?
For me right now, the Emacs page is new, while the GNU Emacs page is old, for instance:
It *really* doesn't degrade gracefully under a small amount of text scaling. I don't think they fully tested this redesign before deploying it; I don't think my configuration's an obscure edge case.
My complaints:
* Very narrow width for the main article body, where sidebars are present
* Far too large margins; incorrect scaling
* Jarring left/right asymmetry
* Undesirable scrollbars on internal elements (left Table of Contents); hinders navigation
* Space above and below left ToC is wasted and ugly
(To be transparent, I *did* cherry-pick this specific page so that I could show multiple issues in a single picture).
It's probably possible to find screen widths at which comparisons favor the old skin, but all I see is improvement. From old to new, minimum or maximum width, more content is visible, and line lengths are more readable. Kudos on the new design.
If you scroll down past the lede and the table of contents the old design has more information. It's just that giant square of whitespace next to the ToC that is holding back the old one.
On longer articles, you do eventually get to a point where overly-long line widths mean more content fits onscreen at once, but at the cost of readability.
Long line lengths are objectively bad for readability[0] and information retention[1], even among people who claim to prefer them. That said, I believe long line lengths are faster.
I can't help but think Wikipedia would be a better product if it was still run on a shoestring budget.
The cash overhang of the Wikimedia Foundation means that the internal incentives are to overhire, and for those overhired employees to make work for themselves.
Maybe I'm misinformed, but I thought most of that cash went to fund stuff that had no relation with the Wikipedia site at all, not for make-work employees.
Arguably both random projects and make-work employees. In the past 6 years the number of employees has increased by 150% from 280 to 700, revenue (donations) increased from ~80M to ~160M, while the number of monthly visitors has increased only ~10%.
Regardless, it is extremely hard to see how they could be providing additional value for their rapidly growing expenditures.
78 comments
[ 2.9 ms ] story [ 168 ms ] threadThe top comment from yesterday's HN submission mentions how you can configure Wikipedia (or any MediaWiki) page to use whatever theme you like, including all the previous themes, which are still actively maintained: https://news.ycombinator.com/item?id=34431533
The navigation frame on the left is absolutely massive. At first I thought you linked the mobile view, but no, it really is wasting a quarter of the width of the page (on my monitor) on white space.
Today the only reminder of it (a rather cryptic one) seems to be https://en.wikipedia.org/wiki/Wikipedia:WikiSpeak#Typography...
But I think everybody has to create an account, not just us salty grandpas. Like I said, if it was a product goal, it's a sneaky way to do it. Not saying that is the case.
https://en.wikipedia.org/wiki/User:Guy_Macon/Wikipedia_has_C...
Just asking for money while you already have money is not inherently lying. You might not like it, but lying involves making untrue statements. "We need your money" is not untrue, even if you already have a lot of money. "We need your money or else we'll shut down" would be untrue, but all of the pitches I've seen don't include anything like that.
> Sounds a lot like cancer, doesn't it?
No, it sounds like running one of the most popular websites on the Internet requires well-paid employees, benefits, and infrastructure investment. 300 employees is two orders of magnitude fewer employees than Microsoft laid off this week. 300 employees to run one of the most trafficked and useful websites ever created is a strikingly small number.
Yes, I want wikipedia to have redundancy. Yes, I want Wikipedia to be resilient against incidents. Yes, I want them to have lawyers to respond to legal issues (both founded and malicious). Yes, I want them to invest in new projects that may or may not pan out into useful services like the ones I already enjoy (it's hard to appreciate how many Wikimedia offers).
If you told me that they spend $100M each year to keep the site online and running well, I'd believe you. A third of that is surprising because it's an astoundingly small number given the sheer amount of value that they create with it, especially for me personally.
Not a single cent goes to the people who actually add value to Wikipedia, i.e. the contributors and admins. They are all volunteers.
The only work that Wikimedia does is keeping the servers on, which is not work for many people, certainly not more than a few.
And if you disagree with that, then where are all the top 1000 websites that are run by "a few" people?
In 2005, Wikipedia ran on $5000 with one employee and over a billion views just fine. People did, in fact, contribute to it despite money not being thrown into mud.
> And if you disagree with that, then where are all the top 1000 websites that are run by "a few" people?
Where are all the static HTTP websites that are ran by dozens of people? At most SaaS companies, keeping the lights on for the HTTP server is the least work you do.
Your view of the world seems strongly distorted by zealous belief in FAANG overstaffing.
It's not static content!
> I've ran moderately popular websites before and it's zero work most of the time.
I run a service that gets ten million requests on a bad day. There's almost always work to be done.
> Where are all the static HTTP websites that are ran by dozens of people?
Most static websites don't allow anyone to sign up for an account and mutate the content.
> At most SaaS companies, keeping the lights on for the HTTP server is the least work you do.
Most SaaS companies aren't serving a multi-terabyte mutable dataset to billions of people a month. They also aren't dealing with tens of millions of pages of user-generated content. Or a huge volume of legal requests because of all of their user-generated content.
> zealous belief in FAANG overstaffing.
Which FAANG company has only 300 employees?
Seriously, 23 billion page views per day. That's like if each of their employees oversaw a service that got 76M page views per day. That's a quarter million page views (not requests!) Per second. Half a billion pages on Wikipedia alone: that's a few million pages per employee.
I've customized it to be just how I like using their settings.
By making text flow to 100% of the width you can easily resize the browser window for everyone's perfect reading width.
I usually run my browser tiled to 50% of the screen, the new design wastes a lot of space at that size (you have to shrink to really narrow width before it enters 'mobile mode' and uses 100% of the width again - I mean it's pretty insane, you can see more text at say 30% width, but at 40%-60% it pops out that useless white space padding on the left leading to a narrower effective text width)
I totally agree with you: I like to use my browser at ~50% of screen width on a 27" screen, so it looks terrible there because of all the wasted space. If they made it work at 50% width like it does at 30% width, I'd have no complaints.
Navigating is also slow, it takes a second to scroll down, while before it was instantaneous.
This might look ok on a smaller monitor, but on 4K it is very sparse.
Seems being directly linked to pages doesn't show the new design sometimes if you aren't logged in. But I'll provide the page where they talk about the new theme itself and the expand button [2].
[1] https://en.wikipedia.org/wiki/2023_Antiguan_general_election
[2] https://en.wikipedia.org/wiki/Wikipedia:Vector_2022#How_to_e...
Either way, I don't have a strong opinion on whether its an improvement or regressive. Pretty much everything nowadays seems to be converging on a narrow column of text.
https://wikimediafoundation.org/wikipedia-desktop/
I don't know if it's just that I'm not used to the new look, but I honestly hate the pillarboxing to fit the new maximum content width.
I mean, I'm old enough to remember when that earlier design was brand-new, and its predecessor was just as iconic, IMO. :P
- First word says exactly what the article is about: Wikipedia.
- First two words give more precision and the news itself: "Wikipedia's Redesign."
- First sentence communicates it was a minor redesign.
- Second sentence communicates it was intentional, and there will be rationale shared.
This type of headline writing should be expected, but sadly it's the exception.
[0]: https://en.wikipedia.org/w/index.php?title=Main_Page&useskin...
I understand the goal is to limit the line width of articles to make it more comfortable, and I don't disagree with that, but the content feels poorly framed now, just floating in the abyss of 'empty space'. The sticky side nav (which I think is a great idea!) feels unconnected from the article itself.
I think the basis is good, but final execution is lacking. I think with a few relatively minor tweaks it could be even better.
Edit: I think the page looks better with the hamburger menu open, with the different background give the article a bit more structure to sit against.
- When reading an article, I want to have absolutely minimal visual noise. Removing the old sidebar helps.
- The smaller column width helps with readability, because it's easier to scan back to the start of the next line.
- It's great to have the table of contents always visible, to quickly jump around the some of the miles-long articles typical for Wikipedia.
- I don't need anything else on the page - so I don't have any complaints about "inefficient use of space". On the contrary - more whitespace helps me focus on the article.
- Unlike with Wikiwand, there are no distracting animations.
In general, for quite a few things I need extra clicks to get to it.
Also, as I mostly print articles[1], I had a special printing layout set up to print Wikipedia articles. I need to come up with a new layout (reader view for Wikipedia used to be terrible).
[1] https://blog.nawaz.org/posts/2021/Dec/consuming-articles-off...
The one most used element for me was the side bar, where I could hover over the language to see the translation of some term. Now it's hidden behind a click, and a looot of scrolling. Just try to reach Svenska.
I can work around the style changes easily with custom CSS and stylus, but I'm not impressed by that change.
"We are currently considering adding single-click functionality for the most-frequently used languages."
Sadly, while that means that the overall experience will be improved, it will solve exactly nothing for those who are unlucky enough to use a less-frequently-used language.
Alternatively, one could create an account and set a default language, or visit a wikipedia site dedicated to that language.
https://en.wikipedia.org vs https://zh.wikipedia.org for example.
Once you've switched, further links within the site should keep you in that language.
You see something different from this? https://ibb.co/pKbgw6w
Are you logged in, perhaps?
This is the first link: https://paste.pics/dca5897542c8fb939608262307aeafed
Many thanks for the query workaround!
https://en.wikipedia.org/wiki/Emacs
https://en.wikipedia.org/wiki/GNU_Emacs
https://i.ibb.co/RzNWLjY/before.webp
https://i.ibb.co/B2W3Bt6/after.webp
It *really* doesn't degrade gracefully under a small amount of text scaling. I don't think they fully tested this redesign before deploying it; I don't think my configuration's an obscure edge case.
My complaints:
* Very narrow width for the main article body, where sidebars are present
* Far too large margins; incorrect scaling
* Jarring left/right asymmetry
* Undesirable scrollbars on internal elements (left Table of Contents); hinders navigation
* Space above and below left ToC is wasted and ugly
(To be transparent, I *did* cherry-pick this specific page so that I could show multiple issues in a single picture).
New skin, same width: https://ibb.co/PgxF8hf
Old skin, maximum width: https://ibb.co/WsptkCY
New Skin, same width: https://ibb.co/x8NmgZq
It's probably possible to find screen widths at which comparisons favor the old skin, but all I see is improvement. From old to new, minimum or maximum width, more content is visible, and line lengths are more readable. Kudos on the new design.
Long line lengths are objectively bad for readability[0] and information retention[1], even among people who claim to prefer them. That said, I believe long line lengths are faster.
0. https://baymard.com/blog/line-length-readability
1. https://uxmovement.com/content/how-margins-and-line-lengths-...
The cash overhang of the Wikimedia Foundation means that the internal incentives are to overhire, and for those overhired employees to make work for themselves.
Regardless, it is extremely hard to see how they could be providing additional value for their rapidly growing expenditures.