I get what he's saying, and I miss "cool" UIs too. He touched on usability and design patterns but I don't think he gave them the proper emphasis. When someone's interacting with a UI, they're...
Exemplary
I get what he's saying, and I miss "cool" UIs too. He touched on usability and design patterns but I don't think he gave them the proper emphasis. When someone's interacting with a UI, they're explicitly trying to perform a task. Personally as a frontend developer, I love adding whimsysurprise and delightcoolness to my interfaces, but never at the expense of the user's ability to perform said task. A 250ms animation may be wholly appropriate for an infrequent action, but if the user's trying to accomplish something repetitive and has to wait for that to complete every time, the coolness of the effect is going to get old very fast. Ultimately, form must follow function. The futuristic displays in sci-fi movies look great, and are usually very effective at communicating important plot information to the audience, but in most cases would be annoying (if not frustrating) to use daily. A behavior that is novel today can easily become tedious tomorrow.
Additionally, functionality in UIs needs to be discoverable. It should be apparent to users how to accomplish their goals in the shortest time possible. Something you need all the time shouldn't be buried under five layers of menus. Important controls shouldn't be hidden until you mouseover a particular region. The most pertinent information should be displayed in the most prominent places. There's a whole discipline of Information Architecture concerned with organizing and presenting data in intuitive, digestible ways. Designers are looking to onboard new users rapidly, with minimal friction, so people get stuff done sooner instead of futzing about trying to divine where certain settings are. That's a key reason why things are the way they are today.
Another hugely important consideration is accessibility. One reason all the colorful themes disappeared is because they don't meet standard color contrast guidelines for the visually impaired. Take a look at this tool and notice how quickly you fall out of WCAG compliance when you start deviating from black-on-white or white-on-black. On top of that, modern UIs need to be compatible with screen readers, keyboard input, and other assistive technologies. Usually these super-cool UIs are only cool to those who don't need special accommodations. But you can't simply turn your back on the users who do.
Now that's not to say that design is dead, that all UIs must inevitably converge on the same Platonic ur-interface. There's plenty of room for trends to come and go. Typefaces, layouts, color palettes, iconography, copywriting, etc. We've just, through decades of research and user advocacy, discovered what the constraints are. The field of possible forms of UI expression has narrowed, but there's still plenty of space for uniqueness and experimentation. Today's apps might look boring next to the menagerie of '90s Winamp skins, but they're infinitely more usable. I'd say that's an acceptable tradeoff.
One unfortunate trend that runs counter to this in desktop software is a rather strange fear of traditional menubars, resulting in either total elimination of menus (on operating systems without...
Additionally, functionality in UIs needs to be discoverable. It should be apparent to users how to accomplish their goals in the shortest time possible. Something you need all the time shouldn't be buried under five layers of menus. Important controls shouldn't be hidden until you mouseover a particular region. The most pertinent information should be displayed in the most prominent places. There's a whole discipline of Information Architecture concerned with organizing and presenting data in intuitive, digestible ways. Designers are looking to onboard new users rapidly, with minimal friction, so people get stuff done sooner instead of futzing about trying to divine where certain settings are. That's a key reason why things are the way they are today.
One unfortunate trend that runs counter to this in desktop software is a rather strange fear of traditional menubars, resulting in either total elimination of menus (on operating systems without persistent menubars) or all menu items getting shoved into a hamburger menu which by necessity either has to have a large number of nested submenus or links to modals which are overloaded and make you dig to find what you're looking for.
It's a very frustrating pattern and I wish it would fall out of popularity.
I think people are afraid of menu bars because they are often very dense and finding the specific thing you want can often be a challenge if you don't use it frequently. There's a really simple...
I think people are afraid of menu bars because they are often very dense and finding the specific thing you want can often be a challenge if you don't use it frequently.
There's a really simple solution to this problem that I wish someone, somewhere would implement. A menu bar search box. Just a simple text field to the right of the top level and as you search through it, the menu items that do not match your query disappear.
I would argue that if menu density is so high that it’s becoming a problem, the program in question is either too unfocused and needs to be split apart (unix/classic mac philosophy: do one thing...
I would argue that if menu density is so high that it’s becoming a problem, the program in question is either too unfocused and needs to be split apart (unix/classic mac philosophy: do one thing well) or is in need of a significant redesign.
There's a really simple solution to this problem that I wish someone, somewhere would implement. A menu bar search box. Just a simple text field to the right of the top level and as you search through it, the menu items that do not match your query disappear.
macOS has done something similar for over a decade now. Open the Help menu in any app and there’s a search box that not only searches the menus of the active app, but also points out where in the menu hierarchy items are. It even has a key shortcut (Cmd-Shift-/).
Some tasks are inherently very complex and require a complex user interface to accomplish. Think about all the things someone needs to do in a tool like Adobe Premiere
Some tasks are inherently very complex and require a complex user interface to accomplish. Think about all the things someone needs to do in a tool like Adobe Premiere
Of course, but that's part of the challenge of UI design. The more complex the task, the more thought and user research is required to produce a usable interface. You have to figure out exactly...
Of course, but that's part of the challenge of UI design. The more complex the task, the more thought and user research is required to produce a usable interface. You have to figure out exactly which functions to feature prominently, which to tuck away, and to what degree the user can reconfigure the UI to meet their needs.
The problem is that increasingly, none of that is happening and usability is barely considered at all (resulting in numerous, difficult-to-nagivate menus and the like) or are egregiously oversimplified in service of visual appeal.
Personally I actually like menubars, at least when they're done right. Not because they are a usability boon, but because it usually acts as a quick-reference to keyboard shortcuts for the stuff I...
Personally I actually like menubars, at least when they're done right. Not because they are a usability boon, but because it usually acts as a quick-reference to keyboard shortcuts for the stuff I need to do.
Though I can't say it's really related, I notice that they disappeared around the same time as pretty much every "classic" UI widget started to fade away or be drammatically simplified, and I find those much harder to use because they get rid of the shortcuts that made them fast to use in the past. Dropdown menus are a big one; with classical UI design, you could start typing and it would bring the selection to the first matching entry, so it was very fast to select items when there were many options to choose from. I've seen a few applications that override the default browser UI and attempt to reimplement this feature, but they're usually rather clumsily implemented so using the keyboard is a second-class citizen still. Though to their credit, they usually have the quick select act as a true search that only shows items that have the string you entered and don't have a time-limit to keep typing before it resets. In any case, I find it frustrating that the simple act of filling out a form on the web has gone from something I could accomplish without needing to even move my wrist to something is a marked downgrade to me.
Perhaps the most jarring example I have seen of a UI change is when Inkscape changed their gradient UI from a panel to a bunch of tiny handles overlayed on top of the image. While I wouldn't consider either design to be particularly intuitive, I found the redesign to be extremely alienating because there was very little text to explain what was happening. Reading the documentation didn't help either, because it's difficult to describe exactly how it works and how to accomplish what you are trying to do. I had to look up a video to explain it, and that's one of my least favorite ways to learn. To the developers' credit, they have ironed out a lot of the pain points in subsequent releases, so the current implementation is arguably the best.
I suppose the thing that I have a bone to pick at is trendy design patterns. In particular, I've hated Google's Material UI design language since the day it came out; things that use it have the tendency to be either full of extraneous and gaudy widgets or are extremely bloated web applications that run slowly. I don't think I've ever had a point where I had to log into Google Ads or Google Analytics without getting frustrated. I think striving for usability is a great thing, because being inclusive and letting more people use software that will improve their lives is a good thing, but the problem I have with that idea is that in practice tons of designers are just copying patterns without actually knowing why they're doing it and so there's always someone who is being thrown some sort of impediment.
I love this feature, and I'm amazed this is not standard on every operating system. It's super handy for those applications which don't have a very well organized menubar. cough cough Adobe
I love this feature, and I'm amazed this is not standard on every operating system. It's super handy for those applications which don't have a very well organized menubar. cough cough Adobe
Apple made an extremely wise move when they settled on a universal menu bar provided by the system instead of giving individual app developers the freedom to implement their own menuing systems.
Apple made an extremely wise move when they settled on a universal menu bar provided by the system instead of giving individual app developers the freedom to implement their own menuing systems.
I'd love to see it on other OSes too, but that's largely infeasible thanks to menubars being owned by programs rather than the OS elsewhere. Chances are, on Linux and especially Windows desktops,...
I'd love to see it on other OSes too, but that's largely infeasible thanks to menubars being owned by programs rather than the OS elsewhere.
Chances are, on Linux and especially Windows desktops, if you have 5 programs with menubars open, they all implement their menubars differently. One uses win32 menus, one Qt menus, one UWP menus, one has its menu implementation in HTML/JS/CSS (electron), and the final one has a custom menu implementation as part of its immediate mode UI toolkit. The only way menu search could appear in all is if the idea becomes accepted as a baseline desktop software feature, with each program bringing its own implementation.
The reason macOS can do it is because there, the menubar belongs to the OS and is perfectly consistent regardless of how the app in question is written or what technologies it's using.
In my opinion, menubars should be OS-owned even on Windows-style desktops without a persistent menubar. It's too important to entrust to programs and is a feature that the user should be able to consistently use.
In the video the speaker talked about having designs being standardized to make them easier to use. And that really resonated with me because for the most part, before the mid 2000s UIs were...
In the video the speaker talked about having designs being standardized to make them easier to use. And that really resonated with me because for the most part, before the mid 2000s UIs were already essentially standardized. Windows had their one API everyone used, and then they had .NET which basically gave the same set of widgets but with a little refinement. And of course with Apple there were usually only one choice you a developer should make for UI, possibly with a legacy option that was being depreciated. While Aqua and Carbon looked different from MacOS Classic, they still functioned essentially the same. While there were weird and wacky UIs out there, they were usually well thought out, and were typically used for simpler interfaces.
If I had to put all the blame on one project, I would blame Microsoft for Office 2007's Ribbon interface. That was when Microsoft became manic about their interfaces; Windows from then on started having some apps still using GDI, some still having the traditional menubars, and some that used the ribbon. Shortly after, we'd get Windows 8 and the Metro UI, and that's roughly when all hell broke loose across software from everyone else.
The problem the Ribbon attempted to solve was real. Even before Office 2007, Office had collapsing menus meant to hide the seldom-used items to reduce clutter. They attempted to provide a single...
The problem the Ribbon attempted to solve was real. Even before Office 2007, Office had collapsing menus meant to hide the seldom-used items to reduce clutter. They attempted to provide a single place for what pop-up contextual toolbars previously attempted to handle: there's no point having your WordArt toolbar taking up space with gray icons when the document has no WordArt.
I'll let others discuss what a better solution would have been, but "things were fine before" is wishful thinking.
I’m not saying things were better before. I am saying that the ribbon was the start of conflicting design patterns that make computers as a whole harder to use.
I’m not saying things were better before. I am saying that the ribbon was the start of conflicting design patterns that make computers as a whole harder to use.
This is the standard across a lot of applications these days, but at the same time, is criminally underused because a large portion of the adult population still thinks everything should be in...
A menu bar search box.
This is the standard across a lot of applications these days, but at the same time, is criminally underused because a large portion of the adult population still thinks everything should be in folders and menus.
I sincerely believe that in a generation or two "folders/menus" will be a lot more like tag systems.
I've had this pointed out, and I feel I don't love the existing implementations. Since the solution I described filters the menu bar itself rather than displaying a list of search results, it also...
I've had this pointed out, and I feel I don't love the existing implementations.
Since the solution I described filters the menu bar itself rather than displaying a list of search results, it also has the side effect of slowly teaching you where these things are so you likely don't keep using the search box as a crutch.
I see menu's as backwards at this point. When coding or even navigating the screen, being able to push a command (Ctrl + shift + P for example in vs code) and type a few letters to get the exact...
I see menu's as backwards at this point.
When coding or even navigating the screen, being able to push a command (Ctrl + shift + P for example in vs code) and type a few letters to get the exact command I want it always going to be faster, and easier to dig through, then navigating a menu.
Having things available only through a command palette or search function eliminates discoverability. You can't use it to get to a command that you don't know the name of or don't even know exists.
Having things available only through a command palette or search function eliminates discoverability. You can't use it to get to a command that you don't know the name of or don't even know exists.
The two are complimentary, in my opinion. Command palettes are great for experienced users, but for novice they might be a little too much like “magic”. For those users, menus are nice since...
The two are complimentary, in my opinion.
Command palettes are great for experienced users, but for novice they might be a little too much like “magic”. For those users, menus are nice since they’re a straightforward, highly standardized overview of what an app is capable of.
The former default desktop environment of Ubuntu Linux, called Unity, combined the two. It had a system-wide command palette it referred to as the HUD which pulled its contents from the menus of the focused app. This worked very well, letting power users interact with menus via the keyboard exclusively (even apps that hadn’t implemented their own palettes) while everybody else could mouse through menus like they always had.
Unfortunately, Ubuntu dumped Unity for a lightly customized variant of GNOME a while back, which lacks this feature and while Unity still exists as an independent project, it doesn’t receive much attention any more and the bits that allowed it to pull menus from apps have become flaky.
This is fairly ideally-implemented in ServiceNow, imo. They have a menu search that operates as you describe it, and an identical favorites menu bar (togglable) that collects things you pin to it,...
This is fairly ideally-implemented in ServiceNow, imo. They have a menu search that operates as you describe it, and an identical favorites menu bar (togglable) that collects things you pin to it, with identical functionality, as well as variants of those default pages with filters/parameters preloaded in them. There is additionally a History bar that uses the same structure and collects a list of all pages visited that session. Absolute peak usability.
A menu bar search has been present in the Unity (HUD search) and KDE Plasma desktop environments for years. They're very convenient! Additionally there is a third-party plugin for KRunner (system...
There's a really simple solution to this problem that I wish someone, somewhere would implement. A menu bar search box. Just a simple text field to the right of the top level and as you search through it, the menu items that do not match your query disappear.
A menu bar search has been present in the Unity (HUD search) and KDE Plasma desktop environments for years. They're very convenient!
Additionally there is a third-party plugin for KRunner (system search bar for KDE Plasma) called KRunner-appmenu, which makes menu entries of the active window searchable even without a global menu bar.
Couldn't agree more. Speaking as a user, I have long since gone past the point where it's very annoying to need to deal with every Johnny-come-lately designer's desire to reinvent and "put their...
infinitely more usable. I'd say that's an acceptable tradeoff
Couldn't agree more.
Speaking as a user, I have long since gone past the point where it's very annoying to need to deal with every Johnny-come-lately designer's desire to reinvent and "put their stamp" on each UI. There's something to be said for standardization, as well as functionality.
I'm reminded of DVDs and specifically DVD menus. Back in the day, a handful of DVD designers got it in their heads to be "cute and clever" with the menu interfaces.
Now, on its surface, that kind of thing is okay. As long as the menus come right up without delay, and work when you navigate and push buttons without delay. Not all of them did.
The one that's always stuck in my mind was the Terminator 2 disc. That designer, whose name has long since been lost from my mind, decided to do some "clever and cute" things. He did similar stuff with other DVDs he worked on, but the T2 disc is the one that I particularly remember.
One of his notions was you had to "explore" the interface. He hid content on the disc behind his programmed exploration paths. And, mind you, nothing on the interface told or encouraged you to start randomly poking buttons and trying to scroll the on-screen cursor in strange ways.
There was some sequence of buttons you had to push to enable you to then be eligible to input the date of Judgement Day (082997) that would abruptly open up a new menu option. The only thing that was visually indicated was when you'd completed that entire sequence, and the hidden option finally appeared.
Prior to that, you had to be lucky, or have read up online from someone else who'd figured it out. And this was when the internet was small. I've always just assumed some friends of his were shown the "secret", and it spread like degrees of Bacon until someone put it on a forum, and other forums repeated it, and eventually it became "somewhat known" amongst enthusiasts. But random ordinary users? No chance.
Only if you did that stuff could you watch the director's cut version that came on the disc. And you had to do it ... every ... single ... time you loaded the disc into the player. Had friends over and wanted to see it? Hang on while I find my notes, and puzzle through ... ooops hit the wrong thing at the wrong time, gotta start over. Hang on though guys, I'll get it.
I just don't feel "creativity" should be a key quality of an interface. Any interface. Web browser, software, hardware, vehicle controls, light switches, device controls, anything. An interface should be clear, obvious, and responsive. If all those are covered, sure amuse yourself and the consumer with colors and logos and even a little art in the blank space if there is any. But before that, someone shouldn't have to "figure out" how to use it, or fight with it to use it..
Though, as a comedic aside, Kahn in Wrath of Kahn after Kirk remotely lowers Reliant's shields, scanning frantically across the controls that aren't labeled ... very occasionally dense interface design can save the day. "Override! Where's the override?"
This reminds me of the 'easter egg' trend in DVD menus which could be fun but was also sort of annoying if you had a small TV or weren't really great at seeing a teeny tiny object in the...
This reminds me of the 'easter egg' trend in DVD menus which could be fun but was also sort of annoying if you had a small TV or weren't really great at seeing a teeny tiny object in the background suddenly light up. Some DVDs that I had that did this hid the commentary or bonus features this way.
The best/worst offender (abuser) of this in my experience was the Aqua Teen Hunger Force DVDs. The easter eggs became increasingly more ridiculous until on Season 4, when you hit the 'play all' button, the screen would divide into tiny squares with each one playing a single episode from the season. Hilarious, but if you actually wanted to watch every episode, you had to select one to watch, complete it, return to the menu, select the next one, etc.
Often, what I see is that directly conflicts with because they were so focused at removing distractions ("The most pertinent information should be displayed in the most prominent places") that the...
Often, what I see is that
Designers are looking to onboard new users rapidly, with minimal friction, so people get stuff done sooner instead of futzing about trying to divine where certain settings are. That's a key reason why things are the way they are today.
directly conflicts with
UIs needs to be discoverable. It should be apparent to users how to accomplish their goals in the shortest time possible. Something you need all the time shouldn't be buried under five layers of menus. Important controls shouldn't be hidden until you mouseover a particular region.
because they were so focused at removing distractions ("The most pertinent information should be displayed in the most prominent places") that the software ends up being optimized for the occasional user and frustrating to someone who uses it daily. Where's my ctrl+shift+F12 fast menu?
I did not like this video at all. It is, funnily enough, full of the things that led us to simpler minimal designs in the first place. The awkward shoe in of his paid behind the scenes content and...
I did not like this video at all. It is, funnily enough, full of the things that led us to simpler minimal designs in the first place. The awkward shoe in of his paid behind the scenes content and the out of place segue to a different video at the end just to mention two. It’s clear from his opinions and his style: this guy is a commercial designer with very little differentiation or unique personality. He’s what happens when every website looks the same.
The big issue to me is that the core thesis is only correct if you have a very narrow view of design history. Contrary to popular belief, text mode and especially skeuomorphism did not “die”. When google switched from their Holo design in Android 3 and 4 to Material with 5, Material Design was unapologetically skeuomorphic. Hell, it was in the name. Material Design. Seriously, everything looked like paper, cardboard, it was a delight, and not at all “flat” despite what the video is claiming.
No, he’s completely missing a big point: Usability and design patterns didn’t drive the UI style changes of the 10s, it was a shift to accessibility, something he doesn’t mention at all. It turns out the human brain understands a list of items perfectly well and doesn’t need delightful and hard-to-implement pixel perfect recreations of real life, and making things black and white aids in vision accessibility as well. The background of the notes app didn’t change because the collective creative spirits of Apple’s designers were killed with bug spray, it was because there is just a “correct” way of approaching accessibility. And lo and behold, the current new hotness in design is “Neumorphism”, a design philosophy that rejects contrast and differentiation and is much worse for visually impaired users as a result.
And I certainly reject the notion that this shift made fun designs extinct. No, that’s just untrue, and saying a modern building is “boring” because it doesn’t look like a Venetian townhouse is an opinion that’s not only wrong in the worst possible way but presented with such a ridiculous amount of misplaced confidence it makes me wonder if this designer slept though most of their art history classes. No, modern designs aren’t boring. Go look at the icon for Mail.app on macOS and find the address of Apple Park gently imprinted on it so tiny you can only see if you zoom in to amounts you’d never see in a UI. Double click your iPhones side button and rock the phone back and forth and see the cards shimmer. Hover over the emoji button in slack and have it wink at you. It’s still there, and just because your Spotify player doesn’t whip the llama’s ass doesn’t mean cute and unnecessary details are dead. In fact, they were never even sick. You just had to agree to meet the designers in their style.
I guess I’m not surprised. Surface level critiques with barely any substance behind it, mockery of years of art and design history with the simple justification of “boring” and no further elaboration, and missing and confidently wrong information. I guess you see yourself in the art you consume most of all.
I think his standout frustrating example for me was Space Shuttle vs Space X. That's not a UI issue. That's the difference between "Professionals need to be able to do extremely specific things...
I think his standout frustrating example for me was Space Shuttle vs Space X.
That's not a UI issue. That's the difference between "Professionals need to be able to do extremely specific things quickly" and "this is a payload delivery craft". It's like comparing a modern widebody AirBus to an Ercoupe or the Enterprise to one of it's shuttles.
It's frustrating that so much UI is seemingly focused solely on new user onboarding. That's important, especially for a tech startup, but after a task becomes sufficiently complex you need some depth to your interface, and very very few companies even bother because they're too busy redesigning the look of their shallow one.
Not a huge fan of this video for reasons which were explored in more detail than those who commented before me, but commenting from the perspective of someone who had to design control system UIs...
Not a huge fan of this video for reasons which were explored in more detail than those who commented before me, but commenting from the perspective of someone who had to design control system UIs alongside more standard, run of the mill "form" UIs, this critique doesn't seem incredibly valid to me.
For instance, when talking about Venetian houses, he fails to mention that "modernist" architecture is flat not because of its purported laser focus on practicality, but rather a deliberate design choice. A lot of elements in modern architecture are functionless, whilst a lot of elements in Venetian houses were practical at the time they were built, and these functional elements became redundant once new technologies appeared.
Furthermore, I don't see Rivian's attempt at a flashy touchscreen interface in a vehicle as a positive thing. Cars should have straightforward controls, which are easily distinguishable at a cursory glance, and a minimum of visual bloat around said UI elements.
My favourite UIs to use are the most boring, flat UIs ever imagined, with lots of contrast, clear indications as to what each button does, with actionable elements a literal touch away. I adore how Tildes was designed because everything pops out, you know when you can interact with an element because its colour has strong contrast with the background, and space is used very intelligently without significant differences between the mobile/desktop experiences.
I disable animations whenever possible, and I abhor newer "experimental" websites that try to be cute and unique with horrible, unintuitive design that is only aesthetically pleasing. I've encountered websites that download literal tens of megabytes of data for no good reason, hijack the scrolling mechanism, have features/information hidden behind specific actions, so on, so forth.
The video mentioned Rivian and showed some stylized CAD renders of the vehicle. As I recall that didn’t even appear to be an interface, it could’ve been from a TV commercial. For that reason I...
Furthermore, I don't see Rivian's attempt at a flashy touchscreen interface in a vehicle as a positive thing. Cars should have straightforward controls, which are easily distinguishable at a cursory glance, and a minimum of visual bloat around said UI elements.
The video mentioned Rivian and showed some stylized CAD renders of the vehicle. As I recall that didn’t even appear to be an interface, it could’ve been from a TV commercial. For that reason I didn’t really understand the point he was making by bringing it up. But I’m not familiar with Rivian products; maybe that actually is what the touchscreen UI looks like.
So let’s talk about touchscreens in cars.
Old cars had physical push-buttons, click-knobs, sliding switches, very tactile controls that could be manipulated using touch alone (with a bit of muscle memory). Touchscreens require you to look at them, because the interactive elements feel exactly the same as the static ones. There’s literally no way for a driver to use a car’s touchscreen without looking away from the road. This is a major safety issue.
As far as I’m concerned, the move to touchscreen controls is a cost-cutting devolution. I hope it’s just a brief fad and automakers will start building ergonomic, analog components again. Which isn’t to say that touchscreens in general should be left out of cars — they are a great part of nonessential infotainment functions. I love CarPlay and I think it’s an excellent way to embed smartphone features into a vehicle. But critical car operations like climate control, lights, wipers, etc. should not be located there, where adjustments could mean life or death.
I'm not sure if this is the same type of modernism, but there was this one house without eaves - eaves are a functional feature that overhang walls of the house to protect it from midday sun,...
For instance, when talking about Venetian houses, he fails to mention that "modernist" architecture is flat not because of its purported laser focus on practicality, but rather a deliberate design choice. A lot of elements in modern architecture are functionless, whilst a lot of elements in Venetian houses were practical at the time they were built, and these functional elements became redundant once new technologies appeared.
I'm not sure if this is the same type of modernism, but there was this one house without eaves - eaves are a functional feature that overhang walls of the house to protect it from midday sun, thereby keeping it cooler in summer (winter's midday sun is at a more oblique angle and is less affected by eaves).
So, removing the eaves might look like a function-over-form design choice but it was actually precisely the opposite.
I think something he overlooks is hardware. Over this time period, I am going to assume there's been an overall trend of consolidating software to singular standards. I think it's also a bit...
I think something he overlooks is hardware. Over this time period, I am going to assume there's been an overall trend of consolidating software to singular standards. I think it's also a bit easier to assert that it's been definitely done to hardware. Your UI has to make sense for a smartphone. It might even be the same UI as the desktop version (if I understand React Native well). When I think about earlier cell phones (like the Nokia banana phone), you could even question while all screens are rectangular instead of like, circles.
I think it's interesting he also talks about buildings. In New York, we lost the ability to make old beautiful buildings. As far as I understand, it was more expensive than not adding detail. Then it became increasingly expensive as less and less people knew the art of producing such works.
I get what he's saying, and I miss "cool" UIs too. He touched on usability and design patterns but I don't think he gave them the proper emphasis. When someone's interacting with a UI, they're explicitly trying to perform a task. Personally as a frontend developer, I love adding
whimsysurprise and delightcoolness to my interfaces, but never at the expense of the user's ability to perform said task. A 250ms animation may be wholly appropriate for an infrequent action, but if the user's trying to accomplish something repetitive and has to wait for that to complete every time, the coolness of the effect is going to get old very fast. Ultimately, form must follow function. The futuristic displays in sci-fi movies look great, and are usually very effective at communicating important plot information to the audience, but in most cases would be annoying (if not frustrating) to use daily. A behavior that is novel today can easily become tedious tomorrow.Additionally, functionality in UIs needs to be discoverable. It should be apparent to users how to accomplish their goals in the shortest time possible. Something you need all the time shouldn't be buried under five layers of menus. Important controls shouldn't be hidden until you mouseover a particular region. The most pertinent information should be displayed in the most prominent places. There's a whole discipline of Information Architecture concerned with organizing and presenting data in intuitive, digestible ways. Designers are looking to onboard new users rapidly, with minimal friction, so people get stuff done sooner instead of futzing about trying to divine where certain settings are. That's a key reason why things are the way they are today.
Another hugely important consideration is accessibility. One reason all the colorful themes disappeared is because they don't meet standard color contrast guidelines for the visually impaired. Take a look at this tool and notice how quickly you fall out of WCAG compliance when you start deviating from black-on-white or white-on-black. On top of that, modern UIs need to be compatible with screen readers, keyboard input, and other assistive technologies. Usually these super-cool UIs are only cool to those who don't need special accommodations. But you can't simply turn your back on the users who do.
Now that's not to say that design is dead, that all UIs must inevitably converge on the same Platonic ur-interface. There's plenty of room for trends to come and go. Typefaces, layouts, color palettes, iconography, copywriting, etc. We've just, through decades of research and user advocacy, discovered what the constraints are. The field of possible forms of UI expression has narrowed, but there's still plenty of space for uniqueness and experimentation. Today's apps might look boring next to the menagerie of '90s Winamp skins, but they're infinitely more usable. I'd say that's an acceptable tradeoff.
One unfortunate trend that runs counter to this in desktop software is a rather strange fear of traditional menubars, resulting in either total elimination of menus (on operating systems without persistent menubars) or all menu items getting shoved into a hamburger menu which by necessity either has to have a large number of nested submenus or links to modals which are overloaded and make you dig to find what you're looking for.
It's a very frustrating pattern and I wish it would fall out of popularity.
I think people are afraid of menu bars because they are often very dense and finding the specific thing you want can often be a challenge if you don't use it frequently.
There's a really simple solution to this problem that I wish someone, somewhere would implement. A menu bar search box. Just a simple text field to the right of the top level and as you search through it, the menu items that do not match your query disappear.
I would argue that if menu density is so high that it’s becoming a problem, the program in question is either too unfocused and needs to be split apart (unix/classic mac philosophy: do one thing well) or is in need of a significant redesign.
macOS has done something similar for over a decade now. Open the Help menu in any app and there’s a search box that not only searches the menus of the active app, but also points out where in the menu hierarchy items are. It even has a key shortcut (Cmd-Shift-/).
Some tasks are inherently very complex and require a complex user interface to accomplish. Think about all the things someone needs to do in a tool like Adobe Premiere
Of course, but that's part of the challenge of UI design. The more complex the task, the more thought and user research is required to produce a usable interface. You have to figure out exactly which functions to feature prominently, which to tuck away, and to what degree the user can reconfigure the UI to meet their needs.
The problem is that increasingly, none of that is happening and usability is barely considered at all (resulting in numerous, difficult-to-nagivate menus and the like) or are egregiously oversimplified in service of visual appeal.
Personally I actually like menubars, at least when they're done right. Not because they are a usability boon, but because it usually acts as a quick-reference to keyboard shortcuts for the stuff I need to do.
Though I can't say it's really related, I notice that they disappeared around the same time as pretty much every "classic" UI widget started to fade away or be drammatically simplified, and I find those much harder to use because they get rid of the shortcuts that made them fast to use in the past. Dropdown menus are a big one; with classical UI design, you could start typing and it would bring the selection to the first matching entry, so it was very fast to select items when there were many options to choose from. I've seen a few applications that override the default browser UI and attempt to reimplement this feature, but they're usually rather clumsily implemented so using the keyboard is a second-class citizen still. Though to their credit, they usually have the quick select act as a true search that only shows items that have the string you entered and don't have a time-limit to keep typing before it resets. In any case, I find it frustrating that the simple act of filling out a form on the web has gone from something I could accomplish without needing to even move my wrist to something is a marked downgrade to me.
Perhaps the most jarring example I have seen of a UI change is when Inkscape changed their gradient UI from a panel to a bunch of tiny handles overlayed on top of the image. While I wouldn't consider either design to be particularly intuitive, I found the redesign to be extremely alienating because there was very little text to explain what was happening. Reading the documentation didn't help either, because it's difficult to describe exactly how it works and how to accomplish what you are trying to do. I had to look up a video to explain it, and that's one of my least favorite ways to learn. To the developers' credit, they have ironed out a lot of the pain points in subsequent releases, so the current implementation is arguably the best.
I suppose the thing that I have a bone to pick at is trendy design patterns. In particular, I've hated Google's Material UI design language since the day it came out; things that use it have the tendency to be either full of extraneous and gaudy widgets or are extremely bloated web applications that run slowly. I don't think I've ever had a point where I had to log into Google Ads or Google Analytics without getting frustrated. I think striving for usability is a great thing, because being inclusive and letting more people use software that will improve their lives is a good thing, but the problem I have with that idea is that in practice tons of designers are just copying patterns without actually knowing why they're doing it and so there's always someone who is being thrown some sort of impediment.
I love this feature, and I'm amazed this is not standard on every operating system. It's super handy for those applications which don't have a very well organized menubar. cough cough Adobe
Apple made an extremely wise move when they settled on a universal menu bar provided by the system instead of giving individual app developers the freedom to implement their own menuing systems.
I'd love to see it on other OSes too, but that's largely infeasible thanks to menubars being owned by programs rather than the OS elsewhere.
Chances are, on Linux and especially Windows desktops, if you have 5 programs with menubars open, they all implement their menubars differently. One uses win32 menus, one Qt menus, one UWP menus, one has its menu implementation in HTML/JS/CSS (electron), and the final one has a custom menu implementation as part of its immediate mode UI toolkit. The only way menu search could appear in all is if the idea becomes accepted as a baseline desktop software feature, with each program bringing its own implementation.
The reason macOS can do it is because there, the menubar belongs to the OS and is perfectly consistent regardless of how the app in question is written or what technologies it's using.
In my opinion, menubars should be OS-owned even on Windows-style desktops without a persistent menubar. It's too important to entrust to programs and is a feature that the user should be able to consistently use.
In the video the speaker talked about having designs being standardized to make them easier to use. And that really resonated with me because for the most part, before the mid 2000s UIs were already essentially standardized. Windows had their one API everyone used, and then they had .NET which basically gave the same set of widgets but with a little refinement. And of course with Apple there were usually only one choice you a developer should make for UI, possibly with a legacy option that was being depreciated. While Aqua and Carbon looked different from MacOS Classic, they still functioned essentially the same. While there were weird and wacky UIs out there, they were usually well thought out, and were typically used for simpler interfaces.
If I had to put all the blame on one project, I would blame Microsoft for Office 2007's Ribbon interface. That was when Microsoft became manic about their interfaces; Windows from then on started having some apps still using GDI, some still having the traditional menubars, and some that used the ribbon. Shortly after, we'd get Windows 8 and the Metro UI, and that's roughly when all hell broke loose across software from everyone else.
The problem the Ribbon attempted to solve was real. Even before Office 2007, Office had collapsing menus meant to hide the seldom-used items to reduce clutter. They attempted to provide a single place for what pop-up contextual toolbars previously attempted to handle: there's no point having your WordArt toolbar taking up space with gray icons when the document has no WordArt.
I'll let others discuss what a better solution would have been, but "things were fine before" is wishful thinking.
I’m not saying things were better before. I am saying that the ribbon was the start of conflicting design patterns that make computers as a whole harder to use.
This is the standard across a lot of applications these days, but at the same time, is criminally underused because a large portion of the adult population still thinks everything should be in folders and menus.
I sincerely believe that in a generation or two "folders/menus" will be a lot more like tag systems.
I've had this pointed out, and I feel I don't love the existing implementations.
Since the solution I described filters the menu bar itself rather than displaying a list of search results, it also has the side effect of slowly teaching you where these things are so you likely don't keep using the search box as a crutch.
I see menu's as backwards at this point.
When coding or even navigating the screen, being able to push a command (Ctrl + shift + P for example in vs code) and type a few letters to get the exact command I want it always going to be faster, and easier to dig through, then navigating a menu.
Having things available only through a command palette or search function eliminates discoverability. You can't use it to get to a command that you don't know the name of or don't even know exists.
As opposed to hopelessly wandering hundreds of menus and ribbons?
There’s also help documentation which is way better for discovery
The two are complimentary, in my opinion.
Command palettes are great for experienced users, but for novice they might be a little too much like “magic”. For those users, menus are nice since they’re a straightforward, highly standardized overview of what an app is capable of.
The former default desktop environment of Ubuntu Linux, called Unity, combined the two. It had a system-wide command palette it referred to as the HUD which pulled its contents from the menus of the focused app. This worked very well, letting power users interact with menus via the keyboard exclusively (even apps that hadn’t implemented their own palettes) while everybody else could mouse through menus like they always had.
Unfortunately, Ubuntu dumped Unity for a lightly customized variant of GNOME a while back, which lacks this feature and while Unity still exists as an independent project, it doesn’t receive much attention any more and the bits that allowed it to pull menus from apps have become flaky.
This is fairly ideally-implemented in ServiceNow, imo. They have a menu search that operates as you describe it, and an identical favorites menu bar (togglable) that collects things you pin to it, with identical functionality, as well as variants of those default pages with filters/parameters preloaded in them. There is additionally a History bar that uses the same structure and collects a list of all pages visited that session. Absolute peak usability.
A menu bar search has been present in the Unity (HUD search) and KDE Plasma desktop environments for years. They're very convenient!
Additionally there is a third-party plugin for KRunner (system search bar for KDE Plasma) called KRunner-appmenu, which makes menu entries of the active window searchable even without a global menu bar.
Couldn't agree more.
Speaking as a user, I have long since gone past the point where it's very annoying to need to deal with every Johnny-come-lately designer's desire to reinvent and "put their stamp" on each UI. There's something to be said for standardization, as well as functionality.
I'm reminded of DVDs and specifically DVD menus. Back in the day, a handful of DVD designers got it in their heads to be "cute and clever" with the menu interfaces.
Now, on its surface, that kind of thing is okay. As long as the menus come right up without delay, and work when you navigate and push buttons without delay. Not all of them did.
The one that's always stuck in my mind was the Terminator 2 disc. That designer, whose name has long since been lost from my mind, decided to do some "clever and cute" things. He did similar stuff with other DVDs he worked on, but the T2 disc is the one that I particularly remember.
One of his notions was you had to "explore" the interface. He hid content on the disc behind his programmed exploration paths. And, mind you, nothing on the interface told or encouraged you to start randomly poking buttons and trying to scroll the on-screen cursor in strange ways.
There was some sequence of buttons you had to push to enable you to then be eligible to input the date of Judgement Day (082997) that would abruptly open up a new menu option. The only thing that was visually indicated was when you'd completed that entire sequence, and the hidden option finally appeared.
Prior to that, you had to be lucky, or have read up online from someone else who'd figured it out. And this was when the internet was small. I've always just assumed some friends of his were shown the "secret", and it spread like degrees of Bacon until someone put it on a forum, and other forums repeated it, and eventually it became "somewhat known" amongst enthusiasts. But random ordinary users? No chance.
Only if you did that stuff could you watch the director's cut version that came on the disc. And you had to do it ... every ... single ... time you loaded the disc into the player. Had friends over and wanted to see it? Hang on while I find my notes, and puzzle through ... ooops hit the wrong thing at the wrong time, gotta start over. Hang on though guys, I'll get it.
I just don't feel "creativity" should be a key quality of an interface. Any interface. Web browser, software, hardware, vehicle controls, light switches, device controls, anything. An interface should be clear, obvious, and responsive. If all those are covered, sure amuse yourself and the consumer with colors and logos and even a little art in the blank space if there is any. But before that, someone shouldn't have to "figure out" how to use it, or fight with it to use it..
Though, as a comedic aside, Kahn in Wrath of Kahn after Kirk remotely lowers Reliant's shields, scanning frantically across the controls that aren't labeled ... very occasionally dense interface design can save the day. "Override! Where's the override?"
This reminds me of the 'easter egg' trend in DVD menus which could be fun but was also sort of annoying if you had a small TV or weren't really great at seeing a teeny tiny object in the background suddenly light up. Some DVDs that I had that did this hid the commentary or bonus features this way.
The best/worst offender (abuser) of this in my experience was the Aqua Teen Hunger Force DVDs. The easter eggs became increasingly more ridiculous until on Season 4, when you hit the 'play all' button, the screen would divide into tiny squares with each one playing a single episode from the season. Hilarious, but if you actually wanted to watch every episode, you had to select one to watch, complete it, return to the menu, select the next one, etc.
Joke could've been maintained with two buttons: "play all (simultaneous)" [default] and "play all (sequential)"
Often, what I see is that
directly conflicts with
because they were so focused at removing distractions ("The most pertinent information should be displayed in the most prominent places") that the software ends up being optimized for the occasional user and frustrating to someone who uses it daily. Where's my ctrl+shift+F12 fast menu?
I did not like this video at all. It is, funnily enough, full of the things that led us to simpler minimal designs in the first place. The awkward shoe in of his paid behind the scenes content and the out of place segue to a different video at the end just to mention two. It’s clear from his opinions and his style: this guy is a commercial designer with very little differentiation or unique personality. He’s what happens when every website looks the same.
The big issue to me is that the core thesis is only correct if you have a very narrow view of design history. Contrary to popular belief, text mode and especially skeuomorphism did not “die”. When google switched from their Holo design in Android 3 and 4 to Material with 5, Material Design was unapologetically skeuomorphic. Hell, it was in the name. Material Design. Seriously, everything looked like paper, cardboard, it was a delight, and not at all “flat” despite what the video is claiming.
No, he’s completely missing a big point: Usability and design patterns didn’t drive the UI style changes of the 10s, it was a shift to accessibility, something he doesn’t mention at all. It turns out the human brain understands a list of items perfectly well and doesn’t need delightful and hard-to-implement pixel perfect recreations of real life, and making things black and white aids in vision accessibility as well. The background of the notes app didn’t change because the collective creative spirits of Apple’s designers were killed with bug spray, it was because there is just a “correct” way of approaching accessibility. And lo and behold, the current new hotness in design is “Neumorphism”, a design philosophy that rejects contrast and differentiation and is much worse for visually impaired users as a result.
And I certainly reject the notion that this shift made fun designs extinct. No, that’s just untrue, and saying a modern building is “boring” because it doesn’t look like a Venetian townhouse is an opinion that’s not only wrong in the worst possible way but presented with such a ridiculous amount of misplaced confidence it makes me wonder if this designer slept though most of their art history classes. No, modern designs aren’t boring. Go look at the icon for Mail.app on macOS and find the address of Apple Park gently imprinted on it so tiny you can only see if you zoom in to amounts you’d never see in a UI. Double click your iPhones side button and rock the phone back and forth and see the cards shimmer. Hover over the emoji button in slack and have it wink at you. It’s still there, and just because your Spotify player doesn’t whip the llama’s ass doesn’t mean cute and unnecessary details are dead. In fact, they were never even sick. You just had to agree to meet the designers in their style.
I guess I’m not surprised. Surface level critiques with barely any substance behind it, mockery of years of art and design history with the simple justification of “boring” and no further elaboration, and missing and confidently wrong information. I guess you see yourself in the art you consume most of all.
I think his standout frustrating example for me was Space Shuttle vs Space X.
That's not a UI issue. That's the difference between "Professionals need to be able to do extremely specific things quickly" and "this is a payload delivery craft". It's like comparing a modern widebody AirBus to an Ercoupe or the Enterprise to one of it's shuttles.
It's frustrating that so much UI is seemingly focused solely on new user onboarding. That's important, especially for a tech startup, but after a task becomes sufficiently complex you need some depth to your interface, and very very few companies even bother because they're too busy redesigning the look of their shallow one.
Not a huge fan of this video for reasons which were explored in more detail than those who commented before me, but commenting from the perspective of someone who had to design control system UIs alongside more standard, run of the mill "form" UIs, this critique doesn't seem incredibly valid to me.
For instance, when talking about Venetian houses, he fails to mention that "modernist" architecture is flat not because of its purported laser focus on practicality, but rather a deliberate design choice. A lot of elements in modern architecture are functionless, whilst a lot of elements in Venetian houses were practical at the time they were built, and these functional elements became redundant once new technologies appeared.
Furthermore, I don't see Rivian's attempt at a flashy touchscreen interface in a vehicle as a positive thing. Cars should have straightforward controls, which are easily distinguishable at a cursory glance, and a minimum of visual bloat around said UI elements.
My favourite UIs to use are the most boring, flat UIs ever imagined, with lots of contrast, clear indications as to what each button does, with actionable elements a literal touch away. I adore how Tildes was designed because everything pops out, you know when you can interact with an element because its colour has strong contrast with the background, and space is used very intelligently without significant differences between the mobile/desktop experiences.
I disable animations whenever possible, and I abhor newer "experimental" websites that try to be cute and unique with horrible, unintuitive design that is only aesthetically pleasing. I've encountered websites that download literal tens of megabytes of data for no good reason, hijack the scrolling mechanism, have features/information hidden behind specific actions, so on, so forth.
The video mentioned Rivian and showed some stylized CAD renders of the vehicle. As I recall that didn’t even appear to be an interface, it could’ve been from a TV commercial. For that reason I didn’t really understand the point he was making by bringing it up. But I’m not familiar with Rivian products; maybe that actually is what the touchscreen UI looks like.
So let’s talk about touchscreens in cars.
Old cars had physical push-buttons, click-knobs, sliding switches, very tactile controls that could be manipulated using touch alone (with a bit of muscle memory). Touchscreens require you to look at them, because the interactive elements feel exactly the same as the static ones. There’s literally no way for a driver to use a car’s touchscreen without looking away from the road. This is a major safety issue.
As far as I’m concerned, the move to touchscreen controls is a cost-cutting devolution. I hope it’s just a brief fad and automakers will start building ergonomic, analog components again. Which isn’t to say that touchscreens in general should be left out of cars — they are a great part of nonessential infotainment functions. I love CarPlay and I think it’s an excellent way to embed smartphone features into a vehicle. But critical car operations like climate control, lights, wipers, etc. should not be located there, where adjustments could mean life or death.
I'm not sure if this is the same type of modernism, but there was this one house without eaves - eaves are a functional feature that overhang walls of the house to protect it from midday sun, thereby keeping it cooler in summer (winter's midday sun is at a more oblique angle and is less affected by eaves).
So, removing the eaves might look like a function-over-form design choice but it was actually precisely the opposite.
I think something he overlooks is hardware. Over this time period, I am going to assume there's been an overall trend of consolidating software to singular standards. I think it's also a bit easier to assert that it's been definitely done to hardware. Your UI has to make sense for a smartphone. It might even be the same UI as the desktop version (if I understand React Native well). When I think about earlier cell phones (like the Nokia banana phone), you could even question while all screens are rectangular instead of like, circles.
I think it's interesting he also talks about buildings. In New York, we lost the ability to make old beautiful buildings. As far as I understand, it was more expensive than not adding detail. Then it became increasingly expensive as less and less people knew the art of producing such works.