This explains exactly why physical restaurant menus are so much better vs mobile site menus. If I'm viewing the menu of a restaurant on my phone, I always look in Google Maps for someone who took a picture of the menu, because it's a dense UI. Every "mobile friendly" menu site is able to show maybe 5 items on the page at once, so it takes many pages of scrolling to see everything.

They contort themselves to redefine the word density, when what they should have said is that a good interface for humans maximizes information without losing visual salience (http://www.scholarpedia.org/article/Visual_salience). That is, local density BUT ALSO clear-to-the-human-eye boundaries between information sets.

See the famous (still? hopefully?) Kadir&Brady paper "Saliency, Scale and Image Description" from 2000 for an explanation of how encapsulating information in something visibly distinct, like whitespace, increases the visual saliency of that information: https://www.robots.ox.ac.uk/~timork/Saliency/ijcv_SalScale.p...

The author doesn’t deliver on the promise made early on the essay to examine the question: Why have UIs gotten so sparse?

It’s like entire web design world decided more whitespace is better, and won’t hear anything about it. And now some desktop apps are being designed like web apps. Or take Hulu on my Roku, it will literally only show the first three lines of a four-line movie synopsis, surrounded by tons of space, and make you expand it, via multiple button presses on the remote.

I once implemented a file list view for a start-up I was working for, off of a mock from the designer, similar to what you see when you are browsing Google Drive or Dropbox in a web browser, but with only one view, a list view with very large icons. The amount of whitespace was massive; use of screen real estate extremely poor. But then, these web UIs never look like the Finder, or Outlook, do they? They could. They could feel almost as snappy, too, with some allowance for network roundtrips. The Finder is actually pretty slow these days, even on a high-end MacBook browsing local files. There are lots of pauses and stutters.

There’s an unspoken rule against labeling things, too, if you can use a row of inscrutable icons instead.

There will always be designers experimenting with taking things away. Apple has done it plenty. What if there were no scrollbars, no ports, no home button or menu bar (just swipe from the edge of the screen), no keyboard, no headphone jack. Sometimes it’s a bold direction, occasionally a clear net negative. But minimalism is a thing, it just has to be tempered.

Often, design “trends” are just trends, in my (admittedly cynical) view; there isn’t necessarily any merit at the core, or the people propagating them seem more interested in conforming to trends than asking what is good. The dynamics of fashion are easy to underestimate as an engineer.

People love to copy things. People who grow up watching action movies and become action movie directors just want to make an action movie with all the action movie tropes. That’s the main thing, not necessarily picking the things that work well in action movies and bringing in some things that just work well, period, that are original or timeless, like good directors or writers do.

Besides people loving to copy, people sometimes think following trends is why something will sell. If you’re making clothing and you aren’t hip to this year’s styles and colors, no one is going to buy your stuff, is the sentiment I presume. It can easily become overwhelmingly about conforming to trends. Designers also tend to put famous people and sources of influence on pedestals and think they will never be 1/10 of the genius of, say, the person who decided that some famous building should look like a pile of mashed potatoes, or that an Apple billboard should just be black text centered on a plain white background. In art, as in philosophy, there is so much pressure to agree that certain people are good, regardless of any objectivity or lay opinions—so much focus on status—that to even think of what you are doing as potentially-good-in-others’-eyes, you need to copy someone or some brand with high status, or somehow attain your own status, is how I think people sometimes feel.

In other words, designers of UIs might falsely think the customer cares about trends and fads, and that their work will be evaluated through a system of reference points and status divorced from actual merit, as can happen in the design world and adjacent spheres (art, fashion, etc).

I really like the distinction of "density in time".

JIRA is a really visually dense application, but it's speed, as well as the number of different screens you normally need to click on makes it feel really sparse despite the dense visuals.

Surely some of the reasons for more sparse interfaces is that on mobile:

- Peoples fingers are relatively fat and inaccurate.

- They are slower that desktop - so you'd break the load into parts

- The vertical scroll form factor and screen size limits what you can do.

- Things which are massively useful on desktop - like searching in a page or visually scanning a large doc are much harder on mobile.

Dense UIs certainly have a place. But, simple UIs are not a fad, as it seems some people in this thread see it. The goal is as simple as possible, but no simpler.

The majority of applications and websites you interact with should be simple, and a few should be complex and dense. The reason is that you aren't an expert at most applications and websites, and you want them to be simple, so you can do the thing you want to do without investing much effort. But for applications you know really well, and use all the time, you want them to be more dense, so you can get more things done with fewer steps.

Because there is no easy, cost-effective, or even feasible way to scale the same application's UI complexity smoothly from newbie to expert, the designer almost always has to try to thread a path between the two extremes. This path has to make sense for the use cases they know about, and the largest share of the users they want to serve. This is extremely hard, not extremely simple, as it may seem from an observer's position.

Our "oldschool" Windows B2B application is quite UI dense. Without looking overly busy, we've got information that can be viewed at a glance that other web-based systems use 6+ pages to contain.

I've seen users struggle to flip between many views in some SPA to figure out if things are right or not in their other system, then come to our system to correlate and looking at one or two windows they see all the same data.

I guess it's just the designers, though it seems CSS and HTML lends itself very well to information-sparse pages.

As we're transitioning to the web, due to customer demand, this is one aspect which I very strongly want to keep. We'll see how it goes.

This trend seems to be a western trend. Here is somewhere where I think we could learn from the apps in Japan and especially China.

I frequently feel for any app that I use frequently, i would prefer for it to have many options that I could use to customize its behavior. For instance, Uber

Here's a particularly bad example of lack of density:

https://investor.vanguard.com/investment-products/mutual-fun...

Related to UI but not exactly on density. Even Refilling a prescriptions from Wallgreen's seems to have become impossible from a smaller phone such mine - IPhone SE2020 - since the control to choose the pharmacy or add a zip code for searching one is not actionable, the interface automatically scrolls down and when I drag the page I can see that it's there but there's no way to access it. It appears to be some kind of React garbage optimized for larger screens but completely and utterly broken on slightly smaller ones. It's not that I have anything against the technology itself but it broke basic functionality of yesteryear that just worked. And to what avail? All seems broken and ugly these days. And this isn't even about looking under the hood and analyzing the waste this wave of technology has brought. Where are we heading to? Is everything about to get worse and worse? Who is benefitting from all this because the user isn't...

This is a really good discussion of density in different forms. I’ve always thought mobile UIs could have a density renaissance, would love to see folks questioning some assumptions of these devices - especially when the trend with LLMs is “wait a long time for a potentially incredibly wrong output” it feels like we’re going the wrong way.

This is a great article.

Presenting information is an art form. A lot of it depends on what the information is, and also, who the information is for.

One of my basic philosophies, is that the UI needs to get out of the way. This means not always using sexy little animations, everywhere (but still using them, if they also work as useful indicators of state transitions), proper contrast, minimizing overhead, like frames and controls, etc. Also, not crowding the display too much.

That said, sometimes, we need a dense display, if we have been trained for it. That Bloomberg terminal is probably fine, for many folks, because they have been trained for it, and it's a daily tool. A lot of Tufte's designs need to be presented to experienced users.

I remember the first time I looked at the train maps in the Shinagawa Station, in Tokyo. They were confusing AF. After just a couple of days, however, I had them down, and appreciated all that information.

I tried using a fancy paid Git client, once, because it was just so pretty.

After just a few minutes, though, I nuked it, wrote off the purchase, and went back to ugly old SourceTree.

When I hear of the preference of form over function for data, I think about the sinking of the SS El Faro. The captain relied on weather data from a Comercial service because they were pretty. On the "black box" one member of the bridge had said something about the weather stats and another person mocked "oh, but that's not a pretty graphic" (the captain was below sleeping). They had tried to get the captain to agree to change course many times, but he always refused telling them to come back if things got worse, unaware of how bad it already was since he was below deck. He had never bothered to notice the data his service provided was several hours old, and ended up sailing right into the eyeball of a hurricane.

One thing that really frustrated me with some designers in the software industry is that they seem to want to instinctively do very sparse designs akin to Twitter, etc... I've typically worked on engineering/scientific software where the audience want to see a lot of information and have many controls on the screen at once, often with customisation to make their workflows easier.

I love dense UIs.

Yesterday I upgraded Chrome on Windows and they replaced the folder icons in the bookmarks bar. They changed it about a year ago, but there was a flag which allowed to revert it to the "old" interface. This flag is no longer effective.

Now two folder icons (ridiculous outlines of folders) side by side take up the space of three old yellow folders, and the menu item entries are all bold and super spaced, so I need to scroll a lot.

In every Google product I first set everything to compact mode.

What is it what makes these designers think "let's make this item take up a lot of space"? Don't they think that people also want as much on screen as possible?

To me this is a dark vs. light UI discussion: compact vs. spaced.

A rather disappointing read. I was expecting an analysis explaining why there is this trend towards very sparse interfaces, or practical ways of designing interfaces that are denser in the face of design trends that are pushing all product teams to do ever more spacing out.

Instead, what I found was a reminder of the ‘laws of design’, which are certainly interesting, but which are only tangentially linked to this drift (in my opinion); and to take the most extreme example of sparse interfaces (the Bloomberg Terminal), without really any concrete elements that could help bring a little density back to our user interfaces.

...not to mention what ends the article, a lunar explanation along the lines of ‘Google's very high stock market valuation compared to Yahoo can be explained by the lack of density of its home page interface’ - really? Come on.

> Tufte's examples of graphics with a low data-ink ratio (left) and a high one (right).

Isn't it the other way around? High on the left and low on the right?

EDIT: Based on the alt text both images should be swapped.

I like the term value density in this context.

Continuing on from the Google/Yahoo example, I would be interested in the author's analysis of not just the landing page, but also the results pages. The search "value density" on google, bing, youtube, hn, chat.openai.com etc. are quite different these days.

Most apps and sites don’t have much to show for, especially institutional ones. So you’re left with a whole lot of space to fill with… something. Even if you have text, most text is pure noise, also made to fill space, that goes for images too: you just have to render stuff to convey that you’re unique, insightful, trustworthy and so on; that goes for brands and individuals. I do feel the heavy preference of users of this site towards low whitespace (they’re users, after all), and I don’t personally like it, not because it’s distasteful, but because it makes me do a lot of mistakes and confuses me a lot, so I waste some time compensating for the lack of adaptation towards my ergonomic and cognitive needs. But I won’t hate on it. I embrace it because it serves me well. The same goes for whatever style you wish to shape your content. I think we start to see design when what’s contained in it offer little to no value (considering it’s minimally usable). So I think good content is what we should primarily strive for, whatever style we package it in. It’s funny how most clients get visibly uncomfortable when you make a design that’s direct. Don’t hate on designers; most just do as they’re told, or they’re selected based on the traits the higher levels expect. It’s an illusion to think most professional designers own their work and can reasonably be blamed, if anything, blame the market’s invisible hands.

This is of course a great, well-written post, clearly summarising various important and even immutable principles that have been part of information design for over 25 years.

Too bad the vast majority of designers being paid to create UIs today not only won't read it, but wouldn't understand how to even use it. UI design today is utterly full of fail because the people doing it are so far away from the type of thinking in this post that they wouldn't know a well-designed information space if it exploded in their custard.

This article is invaluable.

We've done the trick of "short animations for delays <1sec", and "indeterminate loader for under 10sec", but one thing that's not mention is that the "determinate loader for waits between 10sec and 1min" is a huge marketing opportunity.

This is where you get to show the value of the product by listing "how much work" is getting done. Similar to how travel sites will tell you, while you're waiting for results, how many airlines they're comparing on your behalf.

Slightly OT: seeing those screenshots of Bloomberg Terminal makes me wonder why they haven't picked a better (more readable) monospace font.

Bloomberg is kind of a funny one in that on the hand it's extremely dense but the density is very local.

Lots of numbers in Bloomberg should really be a clever chart rather than a bunch of numbers. There's a reason why traders have so many screens, often so they can build their own visual equivalent of the pile of numbers e.g. it's nice to have a chart where you left it.

Speaking of numbers in finance: A problem I have using what I'll call "big tech data" tools in finance is that I often need to care very deeply about fractions of a percent whereas these tools are basically made for terabytes of sloppy data for use in a machine learning model.

I enjoyed Namecheap before the redesign. It was a dense GUI and had small icons.

I like old style Windows 95 GUIs and "portlets"

I prefer dense GUIs. I know Japan has a different design aesthetic than the rest of the world. It's similar.

>The UI was much less visually dense, but more value-dense by orders of magnitude. The results speak for themselves: Google went from a $23B valuation in 2004 to being worth over $2T today — closing in on a 100x increase. Yahoo went from being worth $125B in 2000 to being sold for $4.8B — less than 3% of its peak value.

I liked the rest of the article until this nonsense statement.

I think what is also important is "searching with your eyes" vs "searching with your fingers". If you have a lot of information to present, you can either have it hidden away or try to present as much of it as possible all at once.

In the first case the user has to search for it by clicking into submenus or scrolling in the later he can search by just moving his eyes.

I do think that searching with your eyes is often preferable. It is all around faster, especially if you realize you have been mistaken and need to search again.

I hate it if I need to scroll horizontally just because the content only occupies a third of the page in the middle with great white spaces to the left and to the right of it.

I have a wide display for a reason.

Off topic: I’m halfway through the article and can’t help but notice the relatively high number of times a word is erroneously repeated twice; wondering if it was “edited” by AI.

Looks like the author draws a lot from Edward Tufte's The Visual Display of Quantitative Information - I would recommend it to anyone interested in design, it's a very interesting read with a lot of neat visuals. Some complain that it might be a bit heavy on map related examples, but I say that's no problem since maps have pretty much the same goals as effective UX, that being organizing and presenting information to the user accessibly.

I'm dealing with this now on an accounting app I'm building that runs on both mobile and desktop. I've come to the conclusion that the mobile app and desktop app will need two very different designs. I want the mobile app to be useful for quickly checking a dashboard view and to easily enter transactions while the desktop app needs to be very dense with the choice of a compact view to reduce padding ala Gmail.

Information density in many areas are out of whack these days.

Reminds me of people taking pictures of my presentation slides at conferences where four bullet points with short phrases (say, 50 words in total) turn into a full 12 megapixel photograph.

Similarly, my entire PhD thesis written in LaTeX is a 4MB PDF file, whereas my wife's is a 700MB MS Word monstrosity. Both are mostly text, math, line plots, and tables...

Missed opportunity to use hacker news as an example

On one hand, we want the web to be accessible and weird: for anyone to make a webpage and share their thoughts.

On the other hand, your site better be perfect, even if you aren't a professional web developer. If you are a professional web developer, you probably still won't meet the bar (though you will probably use fancier tools).

It's a funny dichotomy. What is anyone supposed to do?

I love dense UIs, but in my experience they're very hard to design. You can easily make an OK design if you add a lot of space and padding, but to make a dense UI look good you need custom borders and textures, and you need a permanent vision so you can add more widgets later that won't look weird. It's sadly not a paradigm that works in many scenarios.

After talking to many users I have realized , most of them want a Single page app. They want to get their work done. They don't multiple pages, just one simple page to do everything and finish their work. Good UIs(mostly dense) try to enable it and they are loved by users.

> The UI was much less visually dense, but more value-dense by orders of magnitude. The results speak for themselves: Google went from a $23B valuation in 2004 to being worth over $2T today — closing in on a 100x increase. Yahoo went from being worth $125B in 2000 to being sold for $4.8B — less than 3% of its peak value.

Wait what??? THAT'S how you explain the differences in how their businesses fared - by the density of their UI?

I was using a Chinese website recently based on the “ant design” [^1] library / philosophy. It’s a really UI dense way of doing things and I enjoyed having everything there without having to go hunting into menus.

[1]: https://ant.design/

A good read; it captures a lot of the key points. I'd like to mention consistency -- especially for complex UIs. An expert user who has taken the time to learn the ins and outs of menus and keyboard shortcuts will RESENT you for making what seem like (and often are) superfluous changes to their workflows.

You might also want to discuss Fitt’s law. Difference between a diagram (Tufte) and a GUI is that we only look at a diagram, but a GUI we interact with, which means we need to ensure people can actually click/tap/select the element of interest. Higher density makes that harder.

I don't think it's fair to put an equation sign between a company valuation and its perceived "UI density". I believe Google would have been just as successful with a really busy home page too.

This is a good article. I always try and articulate these point by talking about how useful and quick and dense phone books where when they existed. And as for fast load times. Might I say that server side rendering may just have a reason to exist again!

The vomitous waste of screen space in UI these days feels like the digital equivalent of inflation. My phone has four times the screen area of my first smartphone yet it shows the exact same number of notification when I drag the bar down.

So is dense information good or bad? I can't work out what the conclusion is from the article, is it a case of context? I don't understand :(

High or low density, what matters to me most is how easy is for me to gather all the information I need and how fast I can act on it.

I wonder if the constraints of text-based user interfaces (TUIs) generally allow for higher information density.

The Visual Display of Quantitative Information is one of my favorite coffee table books. Props for a Tufte reference!

Surprised that this article doesn't mention Chinese UIs, which are notoriously dense. e.g.: https://i0.wp.com/technode.com/wp-content/uploads/2020/03/Co...

The example of low and high data-ink (from Tufte) is switched. I wrote the author to suggest it be fixed

Many people overlook the most obvious gain: remove unnecessary content. This includes words.

I despise the direction the Jetbrains editors are headed towards. Even with the so-called "Compact Mode", the sidebars take up double the space of the old sidebars, while containing less information contained within them and requiring more clicks than before to get to some menus. Icons everywhere, no actual label text anywhere, and the labeling option they begrudgingly added back would be gut-bustingly hilarious if it weren't so depressing, see [1] for what it looks like.

This modern trend of gigantic paddings/whitespace everywhere and abstract flat icons everywhere is horrible to me, and I don't think it even really looks better than the older interfaces they're usually replacing.

[1] https://youtrack.jetbrains.com/issue/IJPL-59808/Tool-windows...

> Actions less than 100 milliseconds apart will feel simultaneous

This is not true for some things and people. I would not call those simultaneous, rather bearable. Most users wouldn't mind it.

Input delay though is very noticeable. If keyboard or mouse have 100ms delay the user might consider that their device is doing something heavy.

And people who got used to fast software, e.g. optimized code editors or games, are even harder to please.

That Yahoo screenshot really took me back.

Density is spatial, Frequency is temporal

I wouldn’t combine them into a single metric because that combined metric is less useful

theres so much i dont know

In addition to varying over time, UI density also varies across cultures. Currently, East Asian websites tend to have higher UI density than Western websites. See, for example, Rakuten’s home page in Japan vs. its home page in the US:

https://www.rakuten.co.jp/

https://www.rakuten.com/

user-density x ui-density = constant

much enjoyed

I have 1 rule:

If you're building an app that people use for work and open every day, you should make it dense. People want to get work done, fast. They don't care about how pretty it is.

Otherwise, you should make it sparse.

Almost everything wrong with modern web UIs is arguably down to the contradiction of designing for both web and mobile at the same time. These two paradigms are not compatible. If you try to build a responsive UI that caters to both, you'll throw one or more of the groups under the bus. If you want to do both paradigms justice, you need two separate designs altogether.

Low density UIs specifically happen when you attempt to present the amount of information a 7" screen can display onto a 27" screen.

> There’s an upper limit to information density, which means you can subtract too much ink, or add too much information. The audience matters, too: A bond trader at their 4-monitor desk will have a pretty high threshold; a 2nd grader reading a textbook will have a low one.

I think this is the most important line: when taken with the axiom “design for your lowest common denominator” and the general advice given to lawyers in a jury trial “speak, explain at a 3rd grade level”

The upper limit for information density has lowered significantly for the vast majority of general users, so unless you can fix that we’re not gonna get our high density UIs back. At least not for general purpose widely distributed applications.

[dead]

100ms is much too long to feel instantaneous. Open a low latency terminal such as xterm (ideally also with a high refresh rate gaming monitor and gaming keyboard), and compare "sleep 0" with "sleep 0.05".

Set your browser to open every page in reader view by default and you never have to mind this crap again. Nor ads, nor cookie banners, nor auto-playing video, nor AI chatbot, nor newsletter popups.

Hint: You can also force reader mode by pressing CMD+Shift+R