So very much effort is spent on making websites 'pixel perfect' conforming to some marketing person's designs.
Human Interaction Engineers should be dictating this instead; with a focus on making information easy to parse, for humans and machines. For placing action widgets in locations that reduce errors.
I do not use Google Meet, but do those buttons at least now stay visible and not only appear if you know where to mouse over? Very relevant comment from a few months ago: https://news.ycombinator.com/item?id=24965293
Google Meet is an interesting example. The mobile apps (Meet, and for some reason, within Gmail) have the End Call button on the left, which is more reasonable I guess, but makes me tap the wrong button on desktop all the time.
Shouldn't the submit button be obviously on the right? It's an operation that goes forward in the procedure, in time, like the x axis the process should point right.
Do all web forms have a left aligned button by default? I just realized they do on HN but if you'd asked me 5 minutes ago I'd have told you it was on the right.
No, it depends where you are aligning things among other considerations. But if your form follows the typical "F" pattern, then putting the button where the user's eyes are is the logical step.
Not necessarily. Having the submit button first means it's the first action the user sees since most people (at least Westerners) read left-to-right. However, your reasoning with having the order the opposite with the submit button last is spot-on.
No specific advantage has actually been attributed to either choice of order. What matters is (1) keeping the order consistent throughout an application, and (2) following the provided style guide for the platform you're developing on.
(Also, the order may need to be changed if the primary action is destructive, such as a "Reset" button.)
> (1) keeping the order consistent throughout an application
Not even just an application. The whole phone needs to be consistent. Every app in it. Otherwise it's just repeatedly discovering and forgetting how to use each and every app's unique interpretation of what their average user wants and what their average user thinks is intuitive.
I mostly agree, and that is why I mentioned that you should follow the general design guidelines for your respective platform. However, when creating a web app, general guidelines are more "implied."
The biggest reason why consistency is more important than following platform guidelines is for cross-platform apps available on multiple devices that have different platform design guidelines. It's obviously not feasible (and I would argue not user-friendly either) to switch the order of action controls for the same app on different devices, especially when the app is available via a web interface as well as native.
On the left it's more linear, rather than tucked away in the far-right corner. It seems more natural to me.
For example I would think this makes much more sense:
Name [arp242 ]
Email [arp242@example.com ]
I have a HN account []
[Submit]
To:
Name [arp242 ]
Email [arp242@example.com ]
I have a HN account []
[Submit]
In the first example the "Submit" is aligned with what you're filling in: Name, email, that checkbox, etc. Your eye will naturally fall to the "submit" because that's the next in the series.
This also works much better if you put the labels above the inputs:
Name
[arp242 ]
Email
[arp242@example.com ]
[] I have a HN account
[Submit]
[Submit]
If you click (or tab) through every one of the inputs one-by-one then you'll end up on the "Submit" if it's on the left, but you need to jump to the right if it's placed there.
And people read things from left-to-right; I find left-alignment almost always more natural; this is why most navigation sidebars are also on the left (which is often also inverted in websites using right-to-left scripts).
It's even worse if you also place a reset button; I would imagine more than a few people in a hurry will accidentally reset forms if it's placed on the left and has equal prominence to submit. Aas the article already mentioned, you probably shouldn't have a reset button at all, and most forms these days don't.
Anyway, I think "it goes forward in the procedure, in time" is overthinking these things too much in too abstract terms. Just put things where people's eyes and mouse cursor will naturally go for these kind of things will get you a long way.
All of that being said, consistency is also important, so if there's a system/platform where things are consistently on the right then sticking to that is usually more important.
Having actions on the right is more natural - this is reinforced by the UX of mobile devices we all use every day. Tap and we flow forward, to go backwards, we go left.
Like spoken languages, the language of design changes over time, and what used to be normal can be quickly outdated (like the reset button on forms).
A desktop computer is not a mobile device though; there are different sensibilities involved because pretty much everything is different. A lot of UX regressions from the last years come from this mistake.
What is more "natural" is a pointless discussion anyway IMO; I regret phrasing it like that and I wish I could edit it. As I said, the key is to put things where people's eyes and mouse cursor will go, and that's rarely a "jump" to an entirely different place on the screen (on mobile this problem exists less because the screens are small). While it doesn't matter too much on narrow forms (less of a jump), on wide forms it's a bigger issue (or if it's placed to the right of the form inputs take).
Having been conditioned on Windows since my childhood (like nearly every person out there), I would be very weirded out if the Submit button was on the left on a desktop.
Whichever is chosen, just make it visually distinct from the other buttons, and be consistent, and I'll be content (if not actually happy, depending on which one matches my opinion).
I'm surprised to see so many people insisting that the submit button should be on the left. Next time you get a dialog box (or equivalent) on a platform that has user interface guidelines, pay close attention to the button placements. You'll find a preference for right-aligned buttons.
Windows: "Right-align commit buttons in a single row across the bottom of the dialog box, but above the footnote area. Do this even if there is a single commit button (such as OK)."
Mac: "Any buttons in the bottom right of a dialog should dismiss the dialog. An action button, which initiates the dialog’s primary action, should be farthest to the right." (Also noteworthy: "Separate destructive buttons from nondestructive buttons.")
Entirely agree - left to right reading direction, the last thing you want to do is submit the form.
Having it on the left, imo, people might look there first but continue to scan to see what the other control is.
When you finish reading a sentence of text in a paragraph, you certainly have the last word of it stick out, how is this different?
It's a leftover from windowed UI design. In an OS level window requestor, the "affirmative" button is traditionally on the left, and the "negative" button is on the right. Progressive UIs were not common, except in wizards.
When forms were implemented, it was important to keep the behavior consistent with what people experienced in their OS.
Now that we have touch phones, the situation has changed, but the best practices have not.
The notion is probably "first/common choice on the left." However, I get what you are saying about the go-forward context but that can change based on your language RTL vs LTR.
3: It's obvious to me <observation that might require skill in domain>.
And then there's no appeal to an expert in the field, or observed behavior.
It's fair that sometimes negative effects are obvious, and if the writer was observing the deleterious effects of the button placement, I could see where they're coming from.
Totally agree. In fact, it is one of my biggest issues moving from Linux to Mac (which is supposed to be oh so good about UX/UI). Fortunately there is a hack where you can map the Command+Q to something less malign like "invert the colors on the screen" so I won't accidentally kill the whole app when intending to only close a small part of it. I just wanted to disable it, but well at least inverting colors is reversible.
Having a dedicated keyboard shortcut for Quit made a lot more sense when it was something you’d often do as much as or more than closing documents, sometimes dozens of times during a single project. (Open Illustrator, copy, quit Illustrator, open Word, paste, quit Word, open Illustrator...) It’s less helpful nowadays when you can just leave applications lying around until next time you need them.
Not only should the "END CALL" icon/button be more removed from the other more common actions like muting yourself, there should be a small "End call" label underneath or above the icon as well. Seeing what the buttons do without needing to guess or hover is important.
The only way to get good design is to hire good designers. Money doesn't buy good design. This UI was shipped by Google. They have money.
The Google Meet example is utterly crazy, but for Submit/Reset I think I see how it happens. You're thinking through the user's flow, and at the end of that flow you have Submit and Reset or Cancel. You think about them at the same point in the workflow, so you by default put them together. I see the same thing with Approve/Reject right next to each other in internal review tools.
And on the point about why the form has a reset button at all, you see the option in the spec (again, right next to submit) and figure it must be there for a reason, so I guess I should include it? Everyone else is, it must have some purpose...
Credit to the author for admitting he doesn't really know UX. And he does have a point about the importance of differentiating destructive actions from non-destructive ones.
Determining which those are can be really hard depending on how your users use your app. And won't always be universal. The example of the "commonly used" closed window versus quit app... the destructiveness of the former is equivalent.
When one is a common action and the other causes more pain when performed accidentally, it's been more helpful to add an option to ask if that was the intended action than to retrain now-generations of users on keyboard commands. Cmd+Q/Cmd+W have been used this way since I was in diapers.
The destructiveness of "close tab" is not equivalent. Control-shift-T will instantly undo the close-tab action. And there is a menu item that lists "recently closed tabs".
Try finding the menu item that brings back the app once you've quit.
One of the most annoying things about Firefox is its limited reopen close tab implementation: I’ve gotten really used to Safari’s behavior here and is really irritates me when I can’t reopen the last tab because it triggers whatever Firefox uses to reset the history.
Speaking of Safari having good UX around undoing state changes that I miss in other browsers: why can a website put a navigation in another tab, but I can’t press back to get out of it in any other browser? (Safari’s implementation isn’t perfect and I can think of ways to improve it, but a simple “yep I didn’t want to go here” should be just as effective if the website specifies `target` as when it doesn’t.)
> The destructiveness of "close tab" is not equivalent. Control-shift-T will instantly undo the close-tab action.
And every browser has a way to reload tabs from the previous session. Both are equally destructive if the tab/s had state you were not finished with and don’t preserve that state when reopened. At least quit can be guarded with an “are you sure?” prompt out of the box.
> Try finding the menu item that brings back the app once you've quit.
The um... Dock or Taskbar or Start Menu or Spotlight or every other thing people use to launch apps every day?
In Google’s defense the size & spacing on those is way beyond the minimum recommended for touch targets and that placement keeps them in easy thumb reach. I would assume that’s why this shipped.
Slight rant, but I feel like phone interface design keeps getting worse rather than better.
We don't have to go back to skeuomorphism of the old iphones, but whenever I use my phone I have no idea what's tappable and what's not. There's even this trend now of not making it obvious what's even a text box or what's not. It /looks/ nice, but it's infuriating to use. And then you have the weird gestures that are completely undiscoverable. Like pulling down from the top right to get the utility menu thing. I mean, yeah, I know it's there, but I never would have actually discovered that on my own. Also now there's this trend of just hiding everything to make an app look minimalistic and simple even though it's not. So I have no idea what it can even do when I look at it. And I /still/ have no idea how to properly line up apps side by side on my iPad. I mean I kind of do, but I constantly forget, and worse, once I do have them lined up, it's hard to get rid of them. It's all insanely undiscoverable.
I wish UX designers would realize it's not all about being pretty, you have to actually give people an idea of what things actually DO. All the clunky old interfaces with the bevelled buttons and huge scroll bars and stuff might have been ugly as sin, but at least they weren't a constant confusion.
I often feel like I’m the only person who likes the sort of extreme skeuomorphism of Apple circa 2007: it wasn’t perfect, but it didn’t have the soulless corporate feeling of flat UI styles.
Skeuomorphism has the problem that everything looks like a photo - it doesn't advertise what's usable clearly either.
The problem with flat UIs is that they're also in the middle of abandoning any conventions on what's interactable as well.
Ironically Windows circa 3.1 and definitely by 95 had this nailed: is it greyed out? 3D? You can interact with it. Not 3D? You can't. 3D but greyed out? It is contextually disabled.
Simple and clear at a glance. What that interface got wrong was the MDI motif - multiple document interface never really worked as well as Microsoft wanted, although if they'd made the leap of making it tiling by default they would've got there.
You definitely are not alone - I still love the old school Aqua UI. I just switched to Linux after almost two decades on Macs, and the primary reason I chose Elementary OS and Window Maker (I use both) is their somehow retro look and feel.
The worst thing is the new dialer. It's bullshit. It took me an hour to figure out how to paste a phone number into the damn phone. Turns out, you hold your thumb over an invisible UI element and then the paste popup appears miraculously. I nearly threw the damn thing across the room when I figured it out. Throw in a box and remove that headache, although it won't be nearly as sexy I'm sure.
I think this is partly why older folks often say something along the lines of "you young people are so /good/ at technology". We've had so much exposure that we've built up a mental model of how even non-intuitive / hard to discover things should work. Sometimes when I update my phone after a number of years or try and operate someone else's phone when they ask for me to "fix this problem" it's hard because of that exact problem, I don't know where or how to access the thing I want.
But it's not all bad; you can't put heaps of buttons on a mobile interface because of the limited screen size and lack of precision with pressing them, so some amount of 'magic' and hiding is necessary imo.
I thought about that, why was I good at technology relative to my parents? It was because I had no job, and little responsibility, so I had the free time to go through every single setting in the control panel and see what it did, or every single setting in my phone and poke around, and it was only after spending that time that I actually became 'good at technology.'
In contrast to my father, who gets off work and tries to do something with his computer and has an issue. It's simply faster for him to call me from my room and have me, whose invested tens of hours poking through all the pokable things already to solve his issue. Simply put, between working, commuting, being an adult, etc, my father had no real time to invest this time in pure discovery, when what little precious time he had as an adult had to be divied up in the most valuable way.
I see that now as an adult, free time is precious. You don't have the time to learn like you did as a kid, when you could just throw 8 hours at something. I barely find the time to play my guitar for a half hour a day between working and being exhausted after the working day. I can't imagine having to learn how to use a computer, at this age. I simply have no free time to invest in such with all the other things life throws at you to prioritize, and what free time I do have I'm mentally tapped at that point, and I fully expect in a few decades at this rate to become a technology dinosaur just like my parents and grandparents were.
It feels like at the same time they keep hiding functionality the add some very in your face ways of telling you about the functionality or soliciting feedback. "Hey we added a new feature", "Did you find this screen helpful?", "want to make the most out of this app?". Instead of thinking about UX rules and applying them I'm constantly in some sort of AB test or survey when I'm trying to get stuff done.
A couple of examples getting in my way at the moment, the android (or nokia) phone app added a full screen "call your favorite contacts with just one tap" image to the favorites screen. I could do this until they added the obnoxious message, now I have to scroll down to even see them. The other would be netflix constantly AB testing me on whether to show the next episode button or jump back to the home screen, just make a decision, the constantly shifting interface is worse than either option.
> Slight rant, but I feel like phone interface design keeps getting worse rather than better.
In my opinion, my interfaces are getting worse rather than better. I feel like a few years ago, a lot of the major companies hit a tipping point with UI optimization and realized a LOT of what they were doing was unnecessary (and perhaps even costing the business money). Anything that didn't clearly serve a purpose got removed - streamlining for a few core usecases.
The best phone UX experiences I've ever had was swiping/multitasking on BlackBerry 10 OS. I wish Android liberated it.
- Swipe from bottom to wake.
- Swipe from bottom whilst inside an app, peaks all open apps, once let go it fully minimizes the app. During a peek you also see the number of unread notifications on the left.
- Swipe from left to go back.
- Swipe from bottom, then to the right to access the BlackBerry hub, which aggregates ALL emails/IMs/notifications in a single list (can be customized into groups).
- Swipe from bottom left corner towards center hides on-screen-keyboard.
- Top swipe displays app specific option/help/misc links.
- 2-finger swipe from top revealed quick settings (wifi, flashlight, etc). No notifications here (they're in the hub)!
- Not swiping, but it had a clean, single unified location for all app notification/permissions/etc which feels much easier than Android.
I feel if BlackBerry released the Q10 2-3 years earlier we would have had a totally different phone ecosystem today. I still miss it.
It doesn’t sound like most of those gestures are particularly discoverable though, which is sort of what the parent comment is complaining about when highlighting widgets that don’t look like widgets. There are lifesaving bits of UI in iOS for example that I didn’t know about for years, like holding the space bar to get directional control of the cursor. I like that it exists but I don’t feel that good UI can be something you not only have to be told how to use, but didn’t even know it existed.
A tutorial on first use is probably not enough to have most users remember the gestures or even just which options there are. Maybe spaced repetition would be a better approach.
There were really only two gestures you had to learn: swipe up to go home, and down for settings, and these are fundamental to using the device (so very hard to forget). Pretty much everything else came naturally.
On the home screen, swiping down gave you the system quick settings, and in an app it gave you the app's settings, so the two-finger gesture was just a power-user shortcut to the system quick settings while in an app.
Swipe left to go back was very discoverable, and worked differently than iOS or Android, because the entire OS was built around the idea of stacks of pages. Swiping left was just pulling a page off the stack (fluidly animated and cancellable, and you could start anywhere on the screen). There was also a back button on the left-side of the toolbar that you could tap to do the same thing.
Swipe up and to the right to go to the notifications Hub was also discoverable. The Hub was the left-most page on the home screen, so "up-right" simply combined the gestures into one fluid gesture; it was also totally optional.
They were discoverable, though. It boiled down to "swipe up from the bottom to go home, swipe down from the top for settings", and at first start-up it forced you to perform these two actions before setting you free.
Swipe left to go back worked anywhere (you didn't have to start from the edge of the screen) and dynamically showed the page being pulled off the stack, so it was super discoverable. All of the swipe gestures worked this way—"peek" was core to the interaction model because it made users feel in control, and let them cancel an action by just dragging back to where they started.
The two-finger swipe from top was only to bring the system quick settings while in an app (the two-finger gesture was also only added late in the OS's life, at the behest of power users); the quick settings were available on the home screen with a single swipe. While in an app, a single swipe brings up the app's settings.
The only non-discoverable gesture was swiping the corner of the keyboard to dismiss it, and it was also totally useless. The advertised way to dismiss the keyboard was by long-pressing the space bar (there was an icon showing this).
I suppose it was this bottom left swipe I was really picturing, and I will admit my experience of iOS colours my enthusiasm for invisible bits of UI in general. I can accept it's possible to have a small pool of gestures and still allow users to build a strong, consistent mental model of how to interact with a device.
> The best phone UX experiences I've ever had was swiping/multitasking on BlackBerry 10 OS
It's a pity more haven't experienced the useful aspects of BB10's interface (10.3.2 IIRC was when it received a much appreciated UI appearance update, fwiw).
I've tried Android, iOS, Windows Phone and BB10 is still my favorite. Windows Phone in particular was surprising in how unintuitive various of the gestures were in comparison.
As a side note, on their phones with a physical keyboard (which featured touch detection on the surface of the entire key array) text interactions were much improved. Finessing text selections was particularly easy compared to using the touch-screen. Double-tapping the keys (without depressing them) brings up the loupe and from there one can hold down Shift while gliding around the keys to adjust the selection, much like a laptop with a touchpad.
Black and white low resolution interfaces were great exactly for this reason.
512 x 342 x 1 bit color created some pretty fantastic usability.
I've wanted a modern consistent, monochrome interface for a while as a general computing interface.
I've been looking at the e-ink devices for inspiration.
Forced contrast, forced visibility, it all has to be apparent. I've been hacking lua for this holy grail for about 6 years now, before that in perl, then in C since around 2001 or so.
I've got foot pedals, midi controllers I use for general computing, lots of little hacks. Multiple 4k monitors in portrait mode, lots of hacking with arduino sensor packs. Still not there yet.
Interfacing is the current limitation in computing. I've got 128 cores, hundreds of gb of ram, and no good way to use it other than the current paradigms. There's gotta be something better
Symbols on fuckin' telephones. I still cannot reliably answer telephones; growing up in a world where picking up the handset was to close that circuit and answer the call, the need to pick up the handset and then operate fucking buttons is still such a pain. I get it, there has to be an operation, because the phone could be in any position and any orientation at the moment of the phone call, being used for something else. But still.
These things have huge resolution now and the user can pick their language; would it really kill them to write words over the little icons? If in English, perhaps "ANSWER", "HANG UP", "MUTE" and so on, but the magic of words is that it doesn't have to be those exact words ("HANG UP" and "END CALL" are completely different sets of words yet in the context of a phone call, mean the same thing to almost everyone - magic).
Words. They carry so much information. So much. I know, people want to find some magic picture that carries full meaning to all cultures, but there simply ain't no such picture and there never will be; there's no shame in using words. Please. Use words. I'd even happily take them in a language I don't even speak, so long as it used an alphabet I could read (or even not - I can read some common software related words in Japanese simply through having sounded them out a few times). For me at least, words are easy; the ever-changing mist of icon-style-du-jour is not.
Along these lines, I admit with shame that lately I’ve been switching to airplane mode before looking up or modifying a contact, because it’s not obvious to me which combination of symbol-labelled buttons will do this without initiating a phone call. I feel old.
And don't get me started on trying to click on a missed call to either find out more about the number, access the voice mail or simply get rid of the "new notification" dot. I more often than not start calling the number although that's the last thing I want to do.
A few weeks after I got my first touchscreen phone, I had to google for how to answer it.
I had assumed you'd just tap the "answer" button, but that failed more often than not. It never would have occurred to me to swipe a minimum of 3cm, starting with the answer button, and I must assume this knowledge has spread to users by osmosis rather than discovery.
The accept button animates on most phones in the way you need to move it. But it's true that I've seen many first timers get confused, probably because of the inherent pressure that accompanies a phone call.
I'm in the "text or GTFO" team, and the opposite is an especially annoying trend in desktop webapps. My bank rewrote their UI to icons-only a while ago and it's complete shitshow. I randomly click around to go to the screen I want because icons are non-descriptive at all.
Having said that, text has an annoying feature of needing to be localized, and localization of button texts is tricky because you want a very short string, in order to not overflow, because you have very small screen estate on mobile.
(It doesn't also help that some native built-in components have APIs that show the icons only, without text)
This. It is confusin for me as someone who grew up with internet. We have a huge screen, use it. Geez. Forget about my parents and older relatives trying to figure out how to hangup. I miss the green and red buttons on Nokias..
There is no creativity any more. Everybody is too busy following apple
What I definitely don't want is four almost identical icons with nothing to tell me what any of them do, or swipe gestures that are impossible to discover, or (the worst) swipe gestures that expose icons.
And then, the absolute worst of all, the useless "upgrade" that offers zero additional functionality, but changes the location or design of all the icons of an app that you've already learned.
My favorite was UI of a smartphone where a Hangup button when pressed will end a call and will display a contact screen of the other party with a Call button exactly at the position where the Hangup button was. So when you are about to end the call and the other side hangs up first, at the moment you pressed the Hang up button, it has turned into Call button and instead of hanging up you actually start to call them back.
I'm on stock Android using the stock Google made Clock app for my alarms and I still sometimes need to double-take and make sure I'm not snoozing instead of turning my alarm off.
This[1] is what it looks like when the alarm goes off. (Swipe left to the "zzz" snooze icon or swipe right to the off icon). I can't always rely on muscle memory because I am not always consistent in which direction I put my phone down on the nightstand. This is not the best UI for a person who has just been woken out of sleep and still groggy.
There is also plenty of room to replace those three icons with "Snooze" and "Stop Alarm" buttons. Plenty of room even for multiple snooze buttons of varying duration. There is a downside to the button approach, though. I have one app that uses buttons to snooze or stop the alarm and I almost always touch some random one when I'm pulling the phone out of my pocket. That's why I like some apps that use buttons that have to be held down for a second or two to register for certain actions.
Phone interfaces are a minefield for me for some reason. It's too easy to accidentally push a button you didn't realize was a button when you were trying to scroll down the screen but pressed it too hard when you did so. I run into lots of pitfalls like this. I grew up in the era where the OS interface was a "READY" prompt, so I've seen most of the paradigms out there over the years. Nothing has annoyed me as much in recent memory as having a touch-sensitive interface on a slick, compact phone with rounded edges that I desperately don't want to drop.
I wonder if part of this is a business problem, not a UX problem:
- most of those apps with the hidden interfaces have no business being on a pocket computer in the first place
If we stopped playing the game of maximising engagement, and returned to "practical usefulness" as the primary driver for application design, we'd probably make better apps.
It's for this reason that I only ever open my investment account on tablet or desktop. I'm much more comfortable with the full-size dashboard and more visible information architecture.
There is a solution, but at this point, I think no vendor has the guts to apply it.
The solution is:
1) Publish Human Interface Guidelines that detail a rich set of standard gestures, how various tappable elements MUST be marked, and how they SHOULD be arranged.
2) Publish abridged HIG for end-users as a part of user manual for the device/platform. Aim for closed-world reasoning, i.e. the user must be able to build a mental model of, "if I can't see this functionality here, here or here, it does not exist", and not "it may be hidden somewhere else".
3) Tell app developers to stick to the HIG or GTFO.
I know, wishful thinking.
The decay started long, long ago. The other day, someone commented on HN with a link to a piece of old Microsoft WinAPI documentation, I think Windows 95 era or older, where there was a side note on window styling and "escape hatches" that unfortunately had to be built in, because marketers are marketers and desperately want to fuck up usability to put branding on things. Back then, platforms already gave away too much control over styling to software vendors (where originally this control resided with end-user).
You know, by the Win 95 era the GUI toolkits would take descriptive code informing what kinds of interactions you can have, and automatically add the correct markings and widgets.
It's interesting that nowadays, with all the work gone into sandboxes and frameworks, nobody seems able to do that.
Nobody cares to do that. Declarative UIs, where you describe what is being represented and what can be done with it, and the framework styles it up for you - these are desirable by developers and users, but not by people holding the money to pay for development.
Another annoying trend is buttons that look like text boxes. Oh. A white rectangle with a gray border? I guess I’m supposed to type there. Nope. It’s a button cleverly disguised as an input.
I haven't seen iOS in a while, but it's sad to hear that it is going downhill. It seems to me that there's a bit of an echo chamber in phone UI: it's done, and reviewed, by heavy users, who already know how it used to work and take a lot of glitches in stride. Maybe I am trying too hard to explain why everything is two steps away from being good.
I recently got an Android 10 phone, from a vendor I had never heard of: Ulefone.
I was coming from a Sony Xperia compact running Android 8 - roughly the same hardware, except not rugged. I presume Sony polished Android a little bit more than Ulefone did, but even taking all that into account, Android 10 was full of shockingly bad UI "decisions".
First: the Do Not Disturb icon in the pull-down thingy is a but when enabled a appears in the top bar.
Second: Pulling down the pull-down thingy once displays a row of 6 icons with no labels. Pulling twice displays 5 columns of icons with labels. It takes extra planning to rearrange the widgets in 5-column mode such that I get the 6 I want in the first row but also a logical grouping in 5 columns.
Maybe these are Ulefone-isms, but whenever I trip over them I imagine how hard Steve Jobs would have fired someone who put this stuff in front of him.
Just yesterday, I wanted to make a normal voice call to someone I usually contact via WhatsApp. When on their contact page, I couldn't tell which icon would make a WhatsApp call and which would make a PSTN call. The WhatsApp phonecall option appeared with the phone's "phone" icon, and the PSTN option had no icon at all. Both had text saying "voice call".
Oh, and don't even get me started about my Android TV. FFS.
>And then you have the weird gestures that are completely undiscoverable. [...] I mean, yeah, I know it's there, but I never would have actually discovered that on my own
I noticed a lot of UI problems like this when my parents changed from Windows Phone to Android and I had to help them. There are a lot of small actions that make sense if you used Android in the past (since they've been slowly introduced) but are completely bonkers to anyone picking it up now:
- To reject a call, swipe the "Accept call" icon down (this one is particularly horrible);
- To enable Wi-Fi, bluetooth, etc. swipe from the top;
- To dismiss a notification, swipe it to the side.
I agree and think that a good chunk of the apps I use on an every day basis slowly get worse UI over time.
Spotify on Android is pretty awful at this point. Now I can't tap and hold to get to the context menu anymore, now I have to use the triple dot button. Why remove that? Usability is also pretty terrible. Why can't I load the list view of an album I have saved? What the hell. Same with entire playlists that I have saved and downloaded.
Snapchat (I wish I didn't have to use it, but it's the main for of communication a lot of my peers use) keep changing things constantly when nothing was every broken. Over the past three years, they've changed the order and number of tabs they have at least four times.
Google Photos on Android is also annoying separated IMO. Give me an order by date and an order by album/folder. Please just load the folder structure. I assume they do things the way they do to maximize use of their cloud storage, so it is likely intentionally awful for local images.
The official Reddit app is absolutely awful and slow. Reddit Is Fun (third party app) is, and always will be, my favorite Reddit experience. Clean and fast.
That was kind of ranty, but I just wish UI's were simpler on mobile.
I appreciate the OP's point. In fact, I accidentally hung up a video conference literally today due to pretty much this exact issue (no, it wasn't Google Meet). I even appreciate the colorful language and style of writing. Sometimes is just feels good to get one's frustrations out. However, allow me to defend the UI designers.
The thing the author doesn't seem to realize / acknowledge is that UI/UX design is about balancing enormous numbers of competing constraints and concerns. It's a human problem, and like most problems of this genre there is no "best" answer, only different sets of weights for different, often competing, concerns. Simplicity and ease of use are good things, but they are almost always at odds with flexibility and power...also good things. Do you make things bigger with more space so they're easier to see, or do you make them smaller so you can fit more things on the page? Do you use color so you can communicate more and catch the eye more quickly, or do you avoid that so that your app is friendly to colorblind users? (Yes, I know there are color schemes that can achieve both to a decent degree.)
These kinds of tradeoffs are lurking almost everywhere you look in UI/UX design, but this kind of nuance seems to be lost on the author. He only seems to see his set of priorities for a UI. Yes, there are plenty of cases where one thing is pretty objectively worse than another, but usually it's more subtle than that. I'm way more impressed with someone who can talk intelligently about the tradeoffs than I am with someone who fixates on something that very well might have been traded off and rant about it.
There’s also the issue of trying to balance what you know to be good vs the desires of the person actually paying you. Fighting this directly doesn’t get you very far (you’ll most likely be replaced by someone more pliable) and some people cannot be convinced of some things.
Everyone and their mother has an opinion on what they think is good design. I am thankful that as a developer I am not having to fight over pixel pushing.
Zoom solves this problem by.... adding a confirm popup button. As far as I can tell they're the only meeting app that does that, I find it almost annoying most of the time but very useful when I hit it by accident.
I've often wondered why this kind of confirmation couldn't be conditional on the confidence that the click / touch was intentional?
As an example, my favourite pet peeve was using Visual Studio with old-school, upfront locking source control (like TFS), and then accidentally drag and dropping a file or folder due to lag in remote desktop, or a failing mouse which sent two click events in 1 ms or something. VS duly pre-emptively locks the 10k files in the folder you just dragged, and begins a 5 minute operation you'll have to somehow undo later, even though it should be fairly obvious from the click events that it was non-intentional.
Going back to the meeting example, surely solid, accurate taps in the center of the hang-up icon could be taken as intentional, but a kind of glancing, less accurate one needing confirmation?
That type of approach is the way forward in UI. Maybe if we call it something silly like MLUX it will catch on sooner.
Another instance of the same principle is that if an unexpected button/element appears and I click it in <30ms or whatever the fastest possible read+react time is, the click isn’t intentional and should be ignored or confirmed. This should scale over time based on user familiarity and speed.
This culminates at:
> as the technology became more sophisticated the controls were made touch-sensitive - you merely had to brush the panels with your fingers; now all you had to do was wave your hand in the general direction of the components and hope. [HHGTG]
That popup annoys me, because I never hit the button to leave a call by accident. It used to annoy me even more because there was a time last year when it was impossible to leave a call by keyboard only on Windows, because all methods of normally closing the window would trigger this popup, but you couldn’t submit that popup by keyboard. Now at least I can Alt+Q Enter or something like that.
(I confess that the first time I encountered this I opened Task Manager and found and killed a Zoom.exe process all by keyboard, just on principle. That was when I discovered that Zoom has two processes so that the main one can restart the call if the call process crashes!)
But you know one potential factor for my never hitting it by accident? They use text labels (“Leave Call” / “End Call”) rather than an icon. Much easier to get right.
It feels like one of those things where the context of the meeting/call should be taken in: was the scheduled time of this call for an hour and you've left 10 minutes in, maybe a confirmation dialog is more useful than when you've decided to leave within a few minutes of the finish time. Doesn't cover all the edge cases etc., but just observing that "the right UX" is super contextual and not a one size fits all thing.
Whether taking that context into account is a good idea is very debateable.
The more contextually 'intelligent' a system is, the harder it is for the user to model it, and ease of user modelling is often more important than reducing the number of interactions need to complete a task.
In this particular case, it wouldn't be possible to know if your keyboard sequence that quits a call would feature an unnecessary enter at the end or not.
Doesn't fix the problem of accidentally turning your video on when you meant to shut the call off. I've taken plenty of calls in places where it'd be more professional to keep the video off (especially when outside home)
I agree with you about the tradeoffs and the importance of looking at the problem from all angles, but I would argues that this is clearly the case wher aesthetics won over usability, which should _never_ happen. It should be easy to visually separate the "hang up" button from other, much less destructive and more commonly uses buttons (especially "mute"), for example with some small amount of separating space. The only reason this is not done is that it "wouldn't look balanced" and designers would be unhappy.
In my experience designers are the worst people to hire for UX because they often sacrifice usability for aesthetics. Programmers fare a bit better (usable UI usually doesn't conflict with the code quality), still not perfect though. Casual users are probably best, with some education in usability of course.
Completely agree. And the thing I love about your comment is that even though you've found solid argument for why the OP's criticism was right, you articulate the tradeoff and explain why you think it was the wrong one. This is a dramatically more effective and compelling approach IMO.
If you happen to do design work and are looking for a job, I'd love to chat and see if there are any possibilities for collaboration. If you're interested, feel free to drop me a line at my username at google's mail service.
> It's a human problem, and like most problems of this genre there is no "best" answer
For this specific problem (Google Meet UX/UI) sure anyone agree with: don't put a destructive action that requires no confirmation (leaving the meeting) next to a common action (mute/unmute yourself). If designers don't get that right, sorry but they are not competent designers. It's like a programmer that, in order to "balance enormous numbers of competing constraints and concerns" decides to not escape user-provided HTML in the frontend. Well, that programmer is not a competent one.
Another classic case of this is your car's "key fob" buttons. You usually have three functions on three similarly sized buttons right next to each other: lock doors, unlock doors, and... SOUND THE DEAFENING ALARM. Really? Can anyone spot the one that doesn't belong next to the others?
On my key fob the alarm button is small, red, and inset into the side, and it requires a lot more force to push. I've never pushed it accidentally. On the face are Lock/Open trunk/Unlock, but it's not obvious that you have to hold the trunk button for two seconds for it to work (this is probably because you have to manually re-latch the trunk if you unlock it).
> The thing the author doesn't seem to realize / acknowledge is that UI/UX design is about balancing enormous numbers of competing constraints and concerns. It's a human problem, and like most problems of this genre there is no "best" answer, only different sets of weights for different, often competing, concerns.
To underline this point, I think in general design suffers from a lot of bikeshedding at tech companies. There are likely many designs tucked away in discarded files that addressed this specific pain point, too, but were discarded at the request of some PM, or manager, or director. Then there's the process of user testing and experiment design that is used to validate UI/UX of products like this. This design may have actually tested well even if it wasn't the team's favorite... I've worked places where that data is used to override a designer's opinion.
Now apart from all that of course, Occam's razor probably applies as well I guess... perhaps this bit of UI was just poorly designed. But I see a lot of chatter on HN regularly about design being superfluous, subversive, unintuitive, "bad" when really I think many (most?) designers are unhappy with the designs that ship out as well.
I think the trade off here is that if you don’t put the end meeting button there, users will get frustrated because they can’t find the end meeting button.
This is a classic challenge in interface design: this confusion pops up every time you have a button which both toggles state AND represents the current state of the world.
It's rare to see it solved cleanly. I try to avoid interface elements that attempt to combine these two roles entirely.
I quite like the way the Twitter "Follow" button works: when you click it the text changes to "Following", which I think is just clear enough, but only because it's a button that every Twitter uses frequently enough that they are likely to remember how it works.
That's why I actually like the "split pill" style toggle for a binary choice like that, where one side is clearly pressed and engaged, and the other side is clearly up and not engaged and waiting for a user to click it.
There’s a fine line between valid complaints about widespread problems in user interface design, and making a frustrating mistake a few times and blaming this on “obviously bad” user interface design. The complaint about two keyboard shortcuts being nearby (on the author’s particular keyboard) is pretty darn close to that line. There’s nothing semantic about the physical locations of two keys on a keyboard like there (arguably) is for a submit and reset button on a user interface.
On the author’s particular keyboard (Q and W are next to each other)? I think you mean on 98+% of keyboards in the US. It’s not like it’s some custom-assembled mechanical keyboard from a Kickstarter or group buy.
I just had an idea that might explain why this happens.
Firstly, I've observed a phenomenon when I talk to a programmer about a problem I want to solve. They might get excited, learn a bit of the problem domain and then go off and build something that kind of solves part of my problem. But then I'm stuck in a endless cycle of explaining the rest of the problem, to someone who isn't really interested, then waiting for a new iteration and then evaluating how it doesn't completely the solve the problem until the programmer gets bored and goes off to find something new and exciting to do with computers. Programming the computerwas always the end for them. They don't actually care about the problem. I care about the problem. So for me computers are a tool to solve my problem. I might as well write the program myself because it's less work to become a mediocre programmer than to intimately understand the problem.
My theory is that something like this happens in interface design. There are designers who love to create beautiful designs, and we love that, but beauty is the end of it for them. They don't actually care about the problem that the program (website, app, whatever) is supposed to solve. If they were stuck in a job filling out forms all day they would quickly learn the lesson to put the reset button out of the way. If they were intimately familiar with the problem, the better design would be obvious. And so it is that people with no "design skills" can point out the obvious mistakes of the designers.
It's also true that good design is hard. Starting with a blank sheet and creating something, let alone something good, is daunting. Perhaps the hardest part is having the humility to admit, "I don't understand the problem sufficiently" and the empathy to care about the problem enough to learn it well.
I think partly it is because UX is measurable (“A is less usable than B and thusly bad”), whereas aesthetics are not (“different tastes”). A kind of means to escape personal emotional accountability.
The problem is also what we do with the ability to measure. Typically, we're making easy things intuitive, and difficult things impossible. This is exactly the wrong way to go about things, if you care about delivering value to users[0]. We should be making the easy things easy, hard things possible, and forget about the whole intuitiveness thing.
There's this widespread belief now that software is only good if a user who never saw it before can become proficient in it in seconds to minutes. I think this is one of the most devastating, dangerous ideas in computing. The only way you can achieve a learning curve like this is by removing almost all functionality from software - make it so dumb that it really takes only a minute to figure it out entirely. Sadly, this is what we see in mobile and web applications these days.
What worries me here is that we've conditioned everyone to assume software is immediately and fully discoverable. Nobody is expected to read the manual these days, and so manuals are not provided, and since manuals are not provided, any feature that cannot be made apparent without explaining it in the manual goes away.
(Even with kitchen appliances, the situation isn't that bad. When a person sees a particular appliance for the first time, they do read the manual, or get someone to show them how to operate it. Maybe it comes with the fact that buying appliances is expensive and overall a hassle, whereas software is too easy to procure?)
--
[0] - I highlight that condition, because it's my belief that most software vendors don't care about delivering value to users. They care about making money off users, and there are many cheaper ways to do that than creating a truly useful and ergonomic product.
I *try* to make my software initially intuitive and gradually discoverable. I think a gentle learning curve is better than a completely flat one.
I thought quite a bit recently about the problem that you are talking about (diminishing end user value in software). I think the source is that we have a lot of devs and designers whose only experience is designing for maximized conversion rate or maximized engagement. As opposed to professional software.
They have no habit of designing for value.
And *that’s the culture*. Go to any awwwards gallery - all of the websites or webapps mentioned there are essentially ads with minimal content.
Which isn’t too bad, since it leaves a market untapped through the collective arrogance of the incumbents...
> The problem is also what we do with the ability to measure. Typically, we're making easy things intuitive, and difficult things impossible. This is exactly the wrong way to go about things, if you care about delivering value to users.
Right. We're prioritizing learnability over usability.
> We should be making the easy things easy, hard things possible, and forget about the whole intuitiveness thing.
Whoa there... Another way of thinking about 'intuitiveness' is in terms of affordances. We mustn't throw the baby out with the bathwater. While it is too much to ask that every function should be obvious upon first seeing the UI, it is not too much too ask that every function should at least be obvious in retrospect after trying to use it or having it demonstrated, and of course leveraging affordances and interaction patterns the user is likely familiar with from elsewhere should be given priority.
> Whoa there... Another way of thinking about 'intuitiveness' is in terms of affordances.
Yes, of course. Thanks for bringing this up. I apologize, I went a bit too far there - what I meant was just "intuitiveness" in the sense of expecting people to be immediately able to work well with something they see for the very first time, with no explicit learning or training.
I also don't mean to ignore familiarity with UIs in general - yes, unless you have a good reason, it's a good idea to copy design elements users are well familiar with (if they're not completely insane, or dark patterns). This matters particularly on mobile and desktop. On the Web, everyone is used to websites looking different from each other, but there are still higher-level patterns (like footer with company info, "contact" link somewhere on the site, site logo redirecting to home, etc.).
I totally agree about affordances, and mental handles in general. "Obvious in retrospect" is a great way of putting it - once you know a feature exists, or used it briefly, it should be easy to find it again. Once you familiarize yourself with a bunch of features, it should be obvious where to find them, because they should fit a consistent mental model. In a way, it's the job of the UI - to let the user learn the correct mental model of the application, and how it manipulates underlying resources.
For that to happen though, you as a software team need at least to a) have a consistent mental model yourself, and b) design both "backend" and UI around that model. I think this is one part where we fail frequently, but unintentionally - the developers and the designers don't spend enough time ensuring they have a shared mental model. When this happens, you have UI that may be consistent with itself, but feels off when used, and every now and then you see surprising behavior or incomprehensible error messages - that's the "backend" model leaking out.
This is exactly right. I also see this from programmer side-projects. The programmer finds a new cool idea, makes a MVP, then gets bored and moves on. Github is a holding yard for such things.
Nothing wrong with writing simple side-projects and moving on, that's an excellent way to learn. Not everyone can afford to donate a bunch of their time in maintaining a project for free, especially considering it's often such a thankless task.
The fact that the code is available means someone can fork it and make their own changes and improvements.
It’s not necessarily about getting bored. The research/experiment process can be the goal in itself. Not everything is about building a finite product.
This is the kind of thing project-based courses ought to train programmers to expect: Yeah, you're a legitimately good programmer, but programming mostly comes down to solving other peoples' problems, so you can't rest on your laurels. There's a whole, wide world out there of companies that need bespoke software, and most of the actual effort in writing that code is ensuring it properly enforces the business logic of a business you know nothing about.
Therefore, in this project you'll be helping a professor from some non-CS department write educational software. You have to balance accessibility with pedagogical accuracy, as guided by someone who doesn't secretly know that some specific algorithm is the magic key to solving the problem.
There's a 2nd order problem, where the stake holder doesn't actually know the requirements until they see a sample implemented. "I'll know it when i see it" style. This is often the case, with bespoke software.
If the stakeholder has such a clear vision of their end result, they probably won't need you to help them implement it!
I partially agree that designers love to create beautiful designs and they don't care about the products.
This is giving a false complacency to what outta be a tight intergration between design + developers with an oversight from the product owner. It is simply inexcusable. We're actively creating a culture of disregard/complacency/ignorance by not shedding light on this as a huge problem in UX/UI design.
Your comment makes me think of the distinction Simon Wardley[1] makes between Pioneers, Settlers and City Planners. The former are the ones who get excited by making something new (and not solving the problem), the latter are the ones who get excited by making sure things keep running flawlessly.
I do believe there are programmers for every category though.
Kirk McKusick had a similar metaphor around road building, with some people hacking a path through virgin jungle with a machete, some bulldozing the road, some paving it, some adding lamp posts, and some painting the lamp posts.
As someone who both designs and programs for the better part of his life, nobody is really safe from that phenomena. If programming and designing is communication this is akin to go "let me say something", sprouting an interesting point, leaving the inplied punchline hanging and exiting the room while ignoring the reaction. If it happens once, okay — if it happens all the time, maybe you should do something about it.
The thing is: both design and programming are more often than not messy when things get real, but every programmer wants to create clean and beautiful code and every designer wants to create aesthetical designs. Many intuitively perceive this "beauty" to be in tension with the complexity of the problem that needs solving. The goal of programming and design is however not to create beauty, but to solve problems and do so beautifully and clearly communicate these solutions, because it doesn't matter how beautiful your code/design is if nobody understands it. Beauty is the cherry on top that you get to achieve once the cake below is done, tastes well and puts a smile on the faces of all the party guests. If you are good you might already plan the cherry into the shape of the cake at the very beginning — but confusing the cherry for the cake IMO means you are either not there yet in terms of your profession or the thing is a toy project to try things out specifically — which is totally acceptable if everybody involved agrees and totally egoistic and shitty if they don't. Don't agree to serious projects you are not willing to commit to once the initial interest fades.
Really good designers are about integrating both usabilty and aesthetics in an iterative process, and do so in such a way the problem is solved, and solved beautifully, all while reducing the cognitive overhead needed by users.
And this cognitive overhead is really what should drive us. The work of programmers and designers is so beautiful/dangerous because it multiplies to a thousand people for a thousand hours. Our decision affect people on a daily basis. Removing a papercut might seem like nothing, but if it avoids irritating even ten other people who use that thing 10 times a day, isn't it the obvious thing to not half-ass on, not to toy around with?
Btw. Design that is only about aesthetics is something that I call "styling" maybe we need a similar word for programmers who toy around?
There's a research field called Semiotics of Human-Computer Interaction studying user interfaces from that angle. Developers usually have a technical background from maths or engineering, and often are not aware of the importance of linguistics in our field.
Both programming languages and GUIs are languages (artificial, sure, but linguistics also study those) which are used to write expressions that can be read by humans. Semiotics, the study of signs and their meaning, provide methods to analyze how users make sense of the software artifacts (products and tools) delivered to them.
One researcher called the user interface a designer's deputy, i.e. a messenger that conveys in its entirety a message that the programmers wants to communicate to the users.
This deputy must stand on its own, since it's the only information available to the user.
Users then perform sense-making on the symbols in the interface, to infer the meanings of all elements [1]. Notice that the user can't see what each interface element actually does, since they don't have access to the code; they can only guess what it does from the available symbols. If the symbols lead the user to infer an incorrect meaning, communication breaks.
I sympathize with what you’re saying, though I would like to point one thing regarding your first part: writing software is sometimes a way to wrap your mind around the problem and think about it. When trying to solve something new I like to start writing code early on because that gives me a canvas on which I can start drafting ideas. It’s more concrete than a whiteboard but can still be very abstract and flexible.
That’s how I identify what I understand and what I don’t.
It’s always been notable to me that the thing that killed Flash wasn’t the wars with the usability community 20-odd years ago, it was the iPhone. People dumped these garish interfaces because a beautiful device came along that refused to run them, not because they were in many or perhaps most cases terrible for the task at hand.
I think a lot of the methodological knowledge built by usability people has probably been ignored in favour of a metric-driven approach these days. That approach is very good at identifying bottlenecks in existing workflows or deciding between two ideas, but ultimately lacks the empathy you describe. You can only really build that by engaging directly with users, and watching them suffer at the hands of your creations.
I've always felt it a bit weird that iPhones quickly became the norm for all designers I've worked with, them hailing it as a so great to use. While I'm stuck not knowing how to find stuff, it's always hidden behind some undiscoverable swipe gesture. The home button used to have 10+ or so different functions based on context. They way you touched the screen (click, long press, 3d touch or whatnot) mattered.
Do normal users know all this? Or do they just only use the 10% easy functionality of the phone and is happy with that? Maybe articles like "17 Things You Didn't Know Your iPhone's Home Button Could Do" is a sign usability has been sacrificed in order to have a clean&neat design.
I have a half-baked still-gooey thought about this, but it's really around the time of the iphone (and the take off of the smartphone) that discoverability got thrown in the trash. Search bars instead of menus became the primary means of interaction, and ironically you can only find something with the search bar if you already know what you are looking for.
Of course menus don't translate well to touch devices, and menu systems can quickly become unorganized junk drawers for functionality, but it feels like discoverability was never fully solved on touch devices.
As an Android user it's also a problem here. The settings menu got a Search feature a few releases ago and its a sad concession that people have no idea where to find things in the Settings menu.
And it also has the same problem you describe - you don't know how to find something unless you know what it's called. And even if you do find something, there's no breadcrumb trail to learn how to find it without using search. You have to look at what's on the current page and like some kind of reverse engineering sherlock, think about where a designer could possibly have put this page. It's insane.
Actually, having 10% common functionality easily available for unsophisticated users, and the rest hidden but fast to access, is good UI.
Sure, discoverability is good to have, but it is not a failure that you need to be told about it as long as it only contains features useful to power users.
Requiring experts to undergo some training to make the most out of their devices is acceptable UI if the final interactions are optimized for their use; double so if it doesn't interfere with usage by untrained users.
Yes, but I just feel that I have to help my parents with the most menial of tasks, but even then I sometimes cannot figure it out without googling.
I wonder if it's kinda a stockholm syndrome or if it caters to one's ego somehow. After discovering a feature, one feels smarter / more connected to the device, compared to if it was actually easier to use from the get go.
Examples of such tasks? I gave my 65yo mom an Iphone a few years back and there were literally zero issues after the quick intro on how to launch apps etc
I wish my 71yo dad was as proficient as your mom. He doesn't remember anything he doesn't do regularly.
For example, he texts frequently but doesn't know how to check his email. Because I am the tech support child, he reaches out to me for iPhone help. However, since I am not an iPhone user, I always struggle and sometimes fail to solve his problem.
Perhaps this is a sign I have not dedicated enough time to learning this new OS in my family ecosystem, but I do feel it is the least-similar to the other OS's I have used. It's like if my dad was learning programming and asked if I'd help him with his Lisp program. No thanks.
Though maybe I am just a luddite and the universe is written in iOS.
Yeah. You are supposed to say "dumb, little ole me, look how smart this this thing is, hiding all this cool stuff...isn't that neat!" It fits well with the kind people our society churns out. Victims that blame themselves.
I have never understood the hype about the iPhone interface. I had an early Android and I loved it. Then, for a year or two, I had an iPhone 4 which I got from work. I never learned to love it. It always felt a bit obtuse and like it didn't quite do what I wanted. I felt like the interface was sacrificing usability for a superficial sense of elegance and efficiency for smoothness of animations.
When that phone stopped working I got a decent Android phone and it felt so much more comfortable.
Interesting. I've worked in design orgs since the iPhone came out and in my experience it's always hovered around 50/50. Maybe it's an East Coast, West Coast thing?
I will say that when it first came out the direct manipulation of the iPhone UI was a qualitatively different experience compared to any other touchscreen for a very long time. Back then most Android reviews included caveats like "stuttery", "janky", or "low resolution", and these had the cumulative effect of spoiling the illusion of direct manipulation.
Today the gap is much smaller and easier to ignore, and the iPhone has added lots of hidden affordances like swipes and double, triple, and force taps to cater to expert users at the expense of novices.
How is it as the expense of novices? They were fine without it before and they’re fine without knowing about it now. But if they do end up wanting to do some of that advanced stuff, it’s one Google or friend-suggestion away.
It's a persuasive argument, and one I've used myself.
One danger is if the affordance is triggered accidentally. For example, my kids like pushing buttons and swiping when I wouldn't think to do those things and accidentally switch apps or enter into guided access modes without meaning to. In my case I occasionally trigger things like sticky keys unintentionally.
Another danger is if application developers start assuming that they can rely on users knowing hidden affordances and use swiping, double-tapping, force pressing, right-clicking, etc in the core application workflow. Not an Android iOS example, but last night one of my kids started playing Stardew Valley as their first non-tablet game. It has a steep learning curve since it relies on multiple different keyboard keys and on differentiating between right and left-click, whereas I wouldn't consider right-clicking to be a novice skill.
On Windows anyway, I think most users are very familiar with the right-click context menu. A lot of people don't even know the keyboard shortcuts for cut/copy/paste because they just right-click and there it is.
For complete novices, the first thing they do is click the mouse, and since there are only two buttons it doesn't take long to figure out that left-click is the primary button, and right-click brings up a menu with handy options (practically everything has a context menu in Windows). But with the Web, where context menus are rare, I wouldn't be surprised if it's not as well-known anymore.
I think you may be right about the web. It is not my experience that novices think to look in right-click context menus on the web, with the exception of text operations like copying text that you mention, and sometimes link actions. (I'm basing this off dozens of usability studies on web-based applications, it may be different for older windows applications.)
I have theories -- maybe it's because most web apps do not bother with context menus (outside of text operations), maybe it's because more people have developed their mental model from touchscreen devices where context menus are less commonly used, maybe more people are on laptops with trackpads that do not make right-clicking as obvious, maybe because there's no visual affordance indicating which onscreen elements have a useful context menu and which have the standard webpage context menu so it's a guessing game. I don't honestly know. But I am confident that if you place key functionality in a context menu of a web-based app that most novice users will not discover it on their own. As per Jakob Nielsen: "...be warned: less skilled users rarely use these [context] menus." [1]
Once end-users learn how to use your app (and therefore are no longer novices) then they seem to have no trouble remembering and using context menus, so it's a great way to expose expert affordances.
In the case of my kids I can also say that the Stardew Valley user interface (such as keeping right-click and left-click actions straight) has been the most difficult part of the game so far. And it doesn't seem like it was particularly necessary distinction to make -- there seem to be few cases where both right- and left-click actions are equally appropriate.
> [...] but ultimately lacks the empathy you describe. You can only really build that by engaging directly with users, and watching them suffer at the hands of your creations.
While that's the only way to build empathy, you should be aware (and beware) that it is also a good way to build contempt (as counterproductive as that often is). Many of the "dark patterns" probably originated that way.
Could you explain more? Most dark patterns seem to me to be metrics-driven tricks to maximise people accepting GDPR or spam opt-ins, or making additional unintended purchases etc. I don't believe they come from any sort of process of watching users on a task and talking to them. If you're saying that understanding your users allows you to better be hostile towards them, then yes, I agree, but that's bad user interface design by definition.
I mean I hat I've observed developers and designers who, when forced to watch users struggle, have the reaction "damn, they're stupid" and further that dark patterns are sometimes motivated by "I bet a lot of people are stupid enough to..."
This is why in so many problem domains, especially outside technical problems for IT people, proprietary software is the only option, or the only complete option. The best way to ensure people pay close attention to those last areas of fit and finish, and make sure they get attention from people with the different sets of skills needed to do a good job in all of those areas, is to pay them. That means you need a really solid revenue stream, and initially need enough capital up front to get the project going.
I know, there are open source business models which work for some companies and that's true but generally in the sorts of technical problems for IT people. Compared to the IT industry as a whole those are edge cases. They seem more important to us embedded in the IT industry, but compared to the overall multi-trillion dollar global IT industry they're peanuts.
This is why most linux distros have interfaces that are missing minor features and can be frustrating for users who don't have command line experience. For example, why can't I easily make a shortcut in ubuntu? Or how about the fact that you can't get full file paths shown in the default file manager, I often know what directory I want to get to but have to figure out what maze of folders I have to go through to get there.
> Or how about the fact that you can't get full file paths shown in the default file manager, I often know what directory I want to get to but have to figure out what maze of files I have to go through to get there.
Windows also has a similar problem: The URL might not actually have the "true" path. It's so infuriating each time I try to get the path to my Downloads folder on Windows, and end up getting "This PC\Downloads". Please. I'm a grown up. Give me the real damn path...
Seriously, Linux has a lot of UX issues, but hiding the full path of files is a problem that plagues literally every single OS file manager. It seems that at some point, all the major OS makers agreed that users are too stupid to understand a file hierarchy, and came up with tricks to hide it.
I find this annoying as well, but it only happens when you use quick access shortcuts. And then once you're in a sub folder it'll display the full path in the title bar provided you turn on "Display the full path in the title bar" in the Folder Options.
Navigating directly to c:\Users\MyUser\Downloads will also show the full path in the title bar.
It seems some of these shortcuts have "special" behaviours in explorer.
According to https://superuser.com/questions/1362386/always-show-absolute... Windows does not mess up the path (yet) if it is a UNC path of the format \\Computer-Name\Users\MyUser so you can pin these folders to Quick Access to benefit from bookmarking folders while keeping the ability to copy folder paths easily.
> It seems that at some point, all the major OS makers agreed that users are too stupid to understand a file hierarchy, and came up with tricks to hide it.
And this is why I find myself mostly in the text shell.
Were I to go graphical the minimum I want is a simple address bar that I can override the current path with a specified path at will.
Maybe a tree view like in Windows 95 Explorer just to pretty things up. Some of the modern comforts like thumbnail images would also be welcome.
I wish they would stop insulting their users who have taken the time to understand the underlying technicalities of a file system. All this should be user-configurable in a good file manager. SpaceFM and PcmanFM come closest to what I've specified.
Hm. Cater to the 1% who have shown that they're able to learn complex concepts easily, or support the 99% who would rather not think about internals and use a computer as a simple tool.
That's not a difficult question. The only time it makes sense to expose the internals is if you build a specialist tool for people who like internals exposed.
This means that no, it shouldn't be "user configurable" either. You don't want a combinatorial explosion of states in your UI for a tiny set of users in the general purpose case, and you don't want to create a honking fat tool for specialists who'll only use the specialist path.
The computer industry is discovering what the mechanical tools industry has known for a while: You can build a general purpose tool that's simple & straightforward, or you can make a specialist tool.
Pretty much any specialist will be slightly unhappy with the specialist tool and modify it to their own requirements, but you can't build a specialist tool that makes even close to all specialists happy. The specialists will also continue to use the generalist tool if it gets the job done, but they'll complain every step of the way.
Being able to see the path in Finder doesn't seem like something that only 1% of users would want, though. I've seen "normal people" get very frustrated because they couldn't figure out where they saved a file. They saved it in a folder called "Documents", but they then tried to find it in a different folder also called "Documents". Hardly seems problematic to show the path to the folder.
Seems unlikely, given that in MacOS you could e.g.:
- Drag and drop the file into a terminal to insert the path
- Copy and paste the file into the terminal, which inserts the path
- Press Command+I to show a popover describing the path to the file
- Enable the "show path bar" option in the Finder's "view" menu, which will show a nice graphical representation of the path to the current location at the bottom of the window (which lets you copy the POSIX path to each item)
- For more advanced users, run the command `defaults write com.apple.finder _FXShowPosixPathInTitle -bool YES` to show the full path to the current folder in each Finder window title
Also in the same ballpark, you can right-click on the window title and see a dropdown with the parent directories. Which primarily lets you navigate to them.
Ah, and also if you alt-rightclick on an item, you can copy its path to the clipboard. Or you can open the ‘Edit’ menu with alt pressed and copy the path to the current folder. Apparently there's even a shortcut for this.
MacOS has been a pain for me with this; in some of the later revisions it's always a pain to get around to the right place. I don't spend time learning Mac in depth; I'm mostly a Unix person, using a Mac as just a terminal to the stuff I actually deal with (which mostly is web based or run on a datacenter somewhere.)
It's certainly fine if you work a lot locally with the Mac; but for me, the cost-effective thing has been to fumble around each time (since I only do this a few times a year) rather than memorize handling that I use so seldom.
The parent comment was complaining about that being impossible in Linux before the conversation switched to Mac. Which is it?
Also, nothing to memorize about copy & paste or drag & drop. I would say you’re used to Linux GUIs not doing what you expect so the obvious solutions aren’t even considered.
> Also, nothing to memorize about copy & paste or drag & drop.
To a terminal. I thought Linux was the one where you needed to open a terminal to do basic things, yet from my experience it's usually MacOS that needs a terminal for such basic things as turning off mouse acceleration.
Where did Linux enter into it? I just don't typically use GUIs much, and I use the Mac GUI much more than I use Linux GUIs. But I've used command lines since I started with computers (before GUIs were common.)
In macos I don't have to edit random .conf files in system/application directories just to make my wifi work correctly. The graphical UI is configurable enough that I rarely need to modify anything outside of my normal user directories.
Where as in ubuntu the forums/stack overflow usually say "go to this file at /sys/whatever and write in this". Hence the need for full file paths.
Right, and I mean there's no way to discover that hotkey if you don't already know it, which is why requiring it for a common UI function is bad design.
Thank you for posting this, a thousand times. It really clearly illuminates precisely why in open source the interests of the user and the developer are fundamentally differently aligned than in commercial software. Not miss-aligned necessarily, but differently aligned.
In open source the developer has the power, they get to decide what they work on and nobody can tell them otherwise. The best a user can do is post a begging letter in a bug tracker. It doesn't matter what the user thinks, even a majority of users, all they can do is ask.
In commercial software it's the user putting bread on the developer's table. The user feeds and clothes the developer's children, and/or pays for their supply of mountain dew. If the user wants something, bye and large they get it, or at least they have a pretty solid chance of it more often than not. A vote for a feature speaks a lot more convincingly when it's backed up by a wallet.
There's nothing wrong with open source or free and libre software. It's great, I love it, but a lot of it's proponents seem to think proprietary software is some sort of crime and genuinely don't understand why proprietary software dominates so completely in so many domains outside of IT infrastructure.
Good luck convincing a commercial software developer to apply the changes you want. It's rare that a company, commercial or not, changes software because of what customers want (marketing department koolaid doesn't apply).
Developers know best, even when they don't, because they're the ones doing the actual work. But in the FOSS ecosystem you have: i) the freedom to do the changes yourself ii) the freedom to offer help with design, documentation, ideas iii) a public bug tracking and issues manager
Let me know when you can do this commercially.
On a tangent: it is indeed a crime to use commercial software when there are libre alternatives; every use moves the needle one tick deeper towards Eternal September
> Developers know best, even when they don't, because they're the ones doing the actual work.
For me, a large part of the fun of being a developer is enabling others to do stuff they otherwise couldn't. As such we absolutely entertain feature requests and similar, and implement a lot of them.
When sales come back from a sales presentation laughing and telling of jaws hitting the floor, it's almost always due to features that started as a suggestion from one of our users.
Very often though the feature requests are trying to solve XY problems. Often there's a better route to achieving what the user wants, which almost always is some way of avoiding redundant work or other workflow simplifications.
Us devs often do know best when it comes to edge cases and limitations, and about other use-cases that this particular user haven't considered.
However most requests are born from something real, so we will usually inquire what the user is after, in an effort to determine the impact and alternate routes. I might even contact other customers who I know use that module or have a similar work flow and ask them what they think.
And based on that implement changes that make the program better not just for that user but for all our customers.
If you don't have the skills to make the open-source changes you want yourself, but you have the money to make it happen, there is nothing stopping you hiring a developer to build the patch that you need. I don't know of a reliable way to do that with proprietary software.
I've worked for two ISVs who's business model was partly based on exactly this. Their customers would directly fund the development of custom features they needed. The first company developed cellular radio network planning software, the other developed business middleware.
> there is nothing stopping you hiring a developer to build the patch that you need
unless it's accepted upstream, you've got an ongoing maintenance problem on your hands. getting an idea in isn't always just about time/resources/money. if your idea doesn't fit their 'vision', it won't be accepted, regardless of how much you have funded your feature. do you now take on maintaining a fork? sometimes the answer might be 'yes', but I suspect in most cases it's going to be 'no'.
And the decade-plus-old bugs I've filed on Google's various tools say what?
(I've given up either reporting bugs or, where at all possible, using their software, as it's abundantly clear my interests and theirs are not in the least aligned.)
How much did you pay Google? Thanks for lending support to my argument that paying for software is the best way to ensure the user and developer's interests are aligned.
Some people pay for Google apps on a custom domain ($12/month exactly), but I don't expect Google to even answer the phone since that's small change for you.
At a certain scale (much smaller than Google's, probably already at about 100 customers) it's impossible to please _all_ users of your software, so you try to please the majority. And whatever you change, there will always be that 'one guy' whose workflow will break.
You were specifically contrasting open-source and proprietary (commercial) software.
The Google model dominates the proprietary world presently, and even long-term shrinkwrap / clickwrap vendors such as Microsoft are shifting in whole or part to advertising-supported software.
What I pay for Google software is indirect, but given a roughly $100 billion global spend on online advertising, allocated largely among the world's richest 1 billion people, that amounts to about $100/year for the privilege of tools which frustrate rather than delight me.
As I've described in "The Tyranny of the Minimum Viable User", odds are strong that mass-market software of any stripe, including proprietary whether paid, subscription, or advertising-supported, will fail to address power-user / elite-user interests:
I think the things you are complaining about are more personal preferences and being accustomed to Mac / Windows system.
> why can't I easily make a shortcut in ubuntu?
> how about the fact that you can't get full file paths shown in the default file manager,
Is it any worse than needing to find the setting to display file extension in windows?
I am on Cinnamon desktop as it's a bit old school and predictable (for people who have been around since windows 98). I right click a file and I have an option to create a link. And my default file manager shows the full file path in the address bar . So I don't think it's a Linux problem rather than a Gnome 3 problem. But I am sure plenty of people don't mind Gnome 3 (everyone else at my work used the standard Ubuntu desktop.
There's a third category: "enterprise" software, where the people doing the purchasing are completely disconnected from the people using it and the developers. This results in appalling UI disasters like Lotus Notes.
There used to be meme among my friends that we’d express how much a piece of enterprise software sucked by how expensive we thought it cost. For example, “Wow, this sucks. This must cost $500k a year.”
It was scary how close we got to some published prices.
What do you mean by "proprietary" in this context? I believe that the availability of the software's source code to its users is unrelated to the fact of whether the actual development of such software is paid or not.
Moreover, free and open source software (free as in "freedom", not necessarily free as in "free beer"), makes it generally simpler for the regular users to provide input regarding the features they need changed. So, in a way, it also helps to achieve the goal you describe:
> The best way to ensure people pay close attention to those last areas of fit and finish...
> simpler for the regular users to provide input regarding the features they need changed
Don't get me wrong, a good interface needs to take user feedback into account. But trying to accommodate the union of all features needed by all users is a recipe for madness. Somewhere you need an engineer, PM, designer, CxO, or someone who can make judgement calls and decide which user needs are more important than others.
>I believe that the availability of the software's source code to its users is unrelated to the fact of whether the actual development of such software is paid or not.
This is true, and I've seen that happen too, one of the companies I worked with would sell source licenses to customers, and as I understand it this was very common in the mainframe business going way back. The software vendor retained rights to sell and distribute the software though, that's what I mean by proprietary.
I've seen these arguments for open source and libre software before many times, but there's a huge discontinuity between the theory and what actually happens in reality. In the real world there are tens of thousands of small and large software houses producing niche software for diverse use cases for businesses all over the world. Hundreds of niche engineering design, test and optimisation applications, B2B services, audio and video tools, chemical engineering tools, automation and industrial control systems, booking and billing systems, here are an almost infinite variety. Most of them are only known to people actually in these niche specialisms. In that world customers paying for customisations is stock in trade, it's entirely normal. In fact the company I'm at right now is paying the vendor for customisations to our incident and change management ticketing system.
In comparison open source, outside nerdy IT oriented tech projects, might as well not exist. It's minuscule. Barely even a footnote.
> proprietary software is the only option, or the only complete option.
If only that were true. I've watched a completely non-IT person work with both vanilla Gnome 3 and Windows 10; they found the former far more intuitive and unobtrusive (not looking at the individual apps, just the basic desktop UI). Proprietary design-by-committee doesn't necessary make for better solutions: hence the numerous complaints recently from people (often with accessibility issues) trying to book vaccination slots on government websites.
I think you're missing a few steps in your development process, or at least one. Normally it goes like this: problem->solution->implementation->testing. Each step has a role associated with this: user->analyst->coder->tester.
If you go straight from problem to coder, you shouldn't be surprised that you don't have a proper solution, because nobody really made a proper solution. That's like building a house without an architect. Hire an analyst.
If I know that the user changes what he originally wanted, damn right I give the thing to the user instead of wasting effort on perfection and edge cases that get thrown out.
I don't know. The projects I got involved in that had an analyst involved tended to have reams of documentation for the obvious features, completely missed the non-obvious features and was wildly wrong in parts.
This is what agile was created to solve. It was an acceptance of the reality and an attempt to live with that rather than trying to bend it to your will.
> I think you're missing a few steps in your development process, or at least one. Normally it goes like this: problem->solution->implementation->testing.
There should actually be testing associated with every step:
Is X a real problem worth solving (severity or cost x frequency)?
Would Y be an appropriate solution (test a mockup, prototype, or stub UI)?
Does the Z implementation work (as in, not just does the problem get solved when Z is used, but does Z actually get used to solve the problem)?
This sounds like a flawed development cycle. Where's the end-user testing?
In defence of the developers/programmers, many people are really bad at explaining what they want, not least because it's often not what they actually need.
If a designer is designing something that looks good but doesn't work well (i.e., solve the problem), they are bad at usability/UX. They should be getting user feedback on their design before it's implemented - that's what user-centred design is about.
There's often a gap when it comes to understanding requirements - between the end-user and the developer, between the designer and developer, etc. Requirements gathering is actually a specialised skill and one of the key duties of a business analyst.
I agree and I think your ideas form a good model to explain so many issues we've been facing more and more often.
Some days ago I wrote here in HN about the change my stock broker made on their default home broker. It immediately becomes obvious that the people responsible for that design have never performed anything nontrivial with stocks. They created a symmetrical, colorful platform. The symmetry is enforced, therefore you don't have the flexibility of setting up your quote-boxes any longer (perhaps because that would break the artist's concept). Just to mention one of many issues the new design brought.
The same applies to the Google Meet example in the article. Functionally, you'd never want those buttons presented that way. But placing them like that makes them look nice and symmetrical and Gestalt-related stuff, so that's the way to go and deal with it, dear users.
> Perhaps the hardest part is having the humility to admit, "I don't understand the problem sufficiently" and the empathy to care about the problem enough to learn it well
I agree, but don't see that happening, not in the short term at least. I had an argument with a designer some time ago about how much longer the project was going to take and how her ideas were actually substracting value from a user perspective. But she ended the discussion with a "this project has my signature, my reputation is in there". And that's what my criticism against current UX trends is all about. Artist's concept trumps user needs, project maintainability and everything. They are investing all the resources on graphic design matters and this is obstructing the whole field.
Some UX folks argue that designers like those "aren't true UX designers". I agree, but those "false" designers seems to be outnumbering the "true" ones. The latter may eventually have to found a new discipline.
There's a certain corollary effect to what you describe that perhaps applies to "modern" interfaces (both web and native desktop/mobile).
Concurrency is excellent as an idea for systems design but horrible for user interfaces. Nothing seems to provide a stable interface and some concurrent process considers it it's privilege to suddenly modify a list of things I'm choosing from .. resulting in the items shifting just a few milliseconds before I click or tap my choice, which results in the wrong choice. This shift happens not only in vertical lists, but also tab-bar style buttons.
To be precise concurrency isn't to blame for it, but it's more like laziness. The interface elements should be locked in place if the system detects that I'm about to select something. Whatever else is waiting to show up can wait, because I obviously didn't need to know about them a few milliseconds earlier.
> items shifting ... before I click ... results in the wrong choice
Or a window from a different app pops up (the app took a while to start), steals the Enter key press, interprets Enter as "Yes do [something]", and then does it (but I didn't want that!).
Why can't OS windows be click & mouse disabled for 2 seconds, after they open
If the product manager dictates UI and UX design then they are the UI and UX designer and their title is meaningless.
But what I think you really mean is that people are often in a blame culture that strategically puts up responsibility defenses which distracts them from producing a good result and focuses them on red taping to cover their asses. It is quite possible in these organizational structures to produce absolute shit and tick all the boxes and get everything approved and make sure no one in the position of actually doing stuff gets blamed for anything.
Not saying that it's their personal fault that certain companies are that way, but it is possible to have some self respect and march back up to the manager and say hey this is shit and we should do a better design rather than waste time on this (ok, yes in a more articulate and respectful way) - if you think that risks your job then it's probably going to be more fruitful working somewhere else anyway. Disclaimer: yes I know real life has other restrictions that means not everyone can do this.
I think it's simpler than that, in this case. What OP is describing is essentially a *control panel* (a small interface of grouped inputs for common tasks). This is an artifact of engineering design, where physical space and materials are usually at a premium, both in construction and usage. Once you've designed and built your controls, that's it; there's no spiriting a secondary interface out of thin air.
However, with dynamic display-based interfaces, you can do just that. Much as I tend to loathe Apple's design, the way they handle iPad shutdowns is quite good: first, you press a button; then, the entire display changes to focus on this task. If you want to shutdown, you then have to perform an entirely different gesture to confirm. This layering of visual feedback and input types escapes the control panel paradigm and correctly takes advantage of the freedom that act lends to better communicate with the user.
Unfortunately, doing so often means eschewing standards and best practices to find a solution that should work better for users. That is extremely difficult to get right; it's not very surprising that designers would purposely decide to use a flawed but known model instead.
We need to get it out of our heads that designers only do things because they're visually satisfying. They are making purposeful and critical decisions to meet a design objective.
> There are designers who love to create beautiful designs, and we love that, but beauty is the end of it for them. They don't actually care about the problem that the program (website, app, whatever) is supposed to solve.
They're essentially aestheticians, not designers. Yep, I hate them too.
You literally described my job, I bridge the gap between design, code and make sure there is a thorough implementation of both, I have skills that overlap in those areas and so I can advise clients properly for any gaps that might have arisen. My bible is About Face : Interaction Design.
Speaking as a programmer who only occasionally interacts with design and considers himself largely terrible at it, reading "About Face" is probably the single book that helped the most to make me (slightly) less terrible.
It is always good to hear this! Especially the part about smart products and posture is very good to know by heart. So many apps still just botch their posture or completely forget the users intent and context (if they even consider it!)
Haha, yes! And that book is pretty ancient. I think the first edition dealt exclusively with building Windows desktop apps because the web wasn't ready yet. It lays out some of the blindingly obvious UX axioms that a lot of devs just don't bother to think about. One of them is "hide the ejection lever" (a metaphor for the ejection lever on a fighter jet) for exactly the kind of thing MJD is discussing. Have those irreversible, destructive controls be present and findable, but really hard to click by accident.
I think my favorite lesson was their term "implementation-driven design" that I still see pretty frequently. It's where you build your UX to match your system architecture and is the opposite of "user-driven design". Like building forms that are just one-to-one with the database tables. It's why I cringe every time I see devs on hacker news saying they don't need managers or designers to build products because you'll end up with an implementation-driven design more often than not.
I am beginning to think that hiring for enthusiastic programmers might be a mistake. Get on old cynical bastard like me. I just want to make my own life easy, and that means writing the easiest to maintain code that I can. Not trying out some new tech that has promised to solve all my problems.
I think at one time there were a number of designers who transitioned from the print/magazine/brochure world to the web. They tended to prize "pixel perfection" and aesthetics over UX and functionality. This way of thinking is fine for making static landing and marketing pages, but is very counter-productive when you are building user-friendly, interactive and responsive applications. This is opposed to people who are actual UX experts who know how to trade off asethetics with usability.
There certainly were, I worked with a bunch of them. They were all talented, came from print and shifted to web + print. My job was to turn their gigantic Photoshop files into hand-crafted HTML, usually tables to layout carefully cropped and compressed JPEGs, and form elements :)
This mirrors my experience as a UX designer coming from the usability side.
Many graphic design refugees are interested in learning about usability, user research, heuristics, Fitts' Law, GOMS/KLM, etc. But it takes time and energy. I've found that some small companies with no established UX team start out gravitating to shiny portfolio examples and end up putting visual design (VX) folks in charge of interaction design which sometimes goes poorly e.g. https://medium.com/intercom-inside/the-dribbblisation-of-des...
"Too many designers are designing to impress their peers rather than address real business problems." This has its parallel with the developer world, where tech choices are made to impress the peer group (including those in other companies likely to be hiring you) rather than address the business problem at hand.
I think the problems with creating a working solution run both ways. I’ve been involved in many design sessions where the subject experts didn’t understand how to make their envisaged product useful to their audience. It’s a common problem in programming as well, you can write code that performs well and is correct but the interface is such a mess that it’s completely unusable by anyone else. If someone is designing for beauty alone they’re caught in the same trap, the composition will work well by itself but its purpose will be unintelligible. Producing a good design requires working with all parties involved including the end users.
Architects - of the original building sense have a similiar misplaced emphasis problem being more about appearances and "advancing the field (read: novelty)" than functionality. Brutalism was an infamous example for making downright needlessly depressing buildings to live and work in that don't even save on maintenance in spite of eschewing ornamentation. Other fun foibles include putting the duct work on the outside of the building, the mold trap that is Fallingwater, and accidentally having a skyscraper melt a very fancy car with its glare before they had to sandblast it away.
It is kind of a tautology to say it but this hints at it being a human organizational social issue behind the pathology if it keeps cropping up.
I agree that good design is hard and often involves substantial work beyond just making something on the sheet.
You're describing me and my ''kind'' of programmers very well.
I try to make myself interested in your problems. However, most of them seem really ''otherworldy'' to me or are in a domain I don't care about about. I became a programmer because I like the art of the trade, and I keep myself interested by writing interesting code.
I like that theory. It may explain why the Google Workspace (/G Suite/Google Apps) icons are now so indistinguishable. If one looks nice, why not make the others look just as nice?
I think it's important to note that product lead, system architect, and UX architect are distinct roles, each with distinct responsibilities, for a reason.
It's not my job as a coder to care about your problem past what has been described to me. It is your job, as product lead, to care about the problem and design a sufficient information architecture, subdomain map, and other documentation to model your problem.
This goes both ways. It is not your job to care about my problems past how it impacts the product. I don't expect you to know or care about how we design the software or how it is implemented; whether we use snake or camelCase, if we use openssl or pgp, if we choose mysql or postgresql.
I just think it's important to outline that much of your comment can be applied in the inverse, and saying that "It's less work to become a mediocre programmer than to intimately understand the problem", is the same as saying "It's less work to become a mediocre product analyst than it is to intimately understand how to code".
I agree with your last point, it's important for every department to have humility and empathy for the problems of their peers. But, it's not a problem to be figured out by designers and coders; it's instead an issue that extends further and requires effort from everyone involved, and at every level, to support.
I agree here. I think a big part of the problem, even in big companies, is that those roles blur too much.
In particular, it is frequently the case where a product lead says something like “we need a way to...” and leaves the implementation open to discussion to all.
This often leads to a programmer, who is more deeply concerned with how that function will operate on the data, coming up with a “how about a button that...” and/or just implementing the button as a suggestion.
And then the UX architect and product lead, knowing that they don’t want to piss off this coder for fear of future pushback on their ideas, just caves and says “fine”.
That’s the most common scenario I’ve seen at larger companies. That, and having woefully inexperienced UX and UI designers in the first place.
The single most important lesson I learned in any of my software engineering courses was to listen first, build later. And keep asking questions until you understand the full requirements of the project! So much time and effort is wasted when programmers build first and ask questions later, and the whole process is needlessly frustrating for everyone involved.
You're spot on re: humility and empathy. I see so much hubris in programmers/computer people thinking they/we can fully understand the world's problems, much less solve them.
This is a pretty lazy explanation of a problem, leaning on somewhat disparaging stereotypes like "nerds just want to nerd out" without even giving specific examples or looking at deeper reasons this might be happening.
For starters, what context have you even experienced this in? From your description it sounds like both you and the programmer (or designer, as the case may be) are just doing this as a side project. Of course there's not going to be the incentive to follow-up and do the hard work of finding actual product-market fit if it was supposed to just be a fun learning experience from the beginning. Did you ever communicate with them what your expectations are, what the end goal is, how much of a time commitment you expect, and was the other person on the same page and equally invested in it? Also, why are you expecting the programmer to both figure out the nuances of the product and implement it? What is even your role in this, what do you bring to the table?
It sounds to me like you did not communicate well or did not set expectations properly about this project, and now you're blaming the other person for doing a bad/lazy job.
> In particular, don't put "close this window" on control-W and "quit this application" on control-Q. I'm looking at you, Firefox.
Oh man, I actually remapped quit on Firefox because this bit me too many times and Firefox kept removing ways to confirm quit. I will die on the hill that "confirm quit" is the correct behavior.
(I have, in fact, died on this hill enough times I may be outing my HN burner with this comment.)
My favourite is editing crontab (the scheduler on Linux). It's -e to edit the file, -r to delete all content from it with no confirmation dialogue or anything. I fell for that twice already.
I can't imagine how often you need to delete everything from a file without using a text editor. I'm sure it's not that common that you need to give it a one-letter flag (instead of --remove or --delete) and put it right next to the fucking edit flag.
Or no flag at tall to replace it entirely with the stdin contents. How many commands are there that completely destroy all the data they can access if you call them without parameters?
I really wonder if there is any script on any distro that uses this behavior. Most distros did go into the sane "let's split the crontab into many files" option.
> In particular, don't put "close this window" on control-W
Right — Control-W should always¹ be ‘erase word’. Windows/IBM really mucked things up by stealing Control for GUI operations (among other things like conflating Tab with Next-Field because they're the same thing on punch cards).
¹ Okay, ‘end of transmission block’ is also acceptable.
UI is designed to generate profit. Sometimes that means contributing to an optimal user experience, but sometimes it doesn't. Sometimes how it looks in marketing materials is more important to the bottom line than whether or not the user hates it after they're already locked into the ecosystem.
There's an aspect of UX design that is obvious to users and not obvious to programmers:
Not all UI elements should have the same visual weight
If you have 3 buttons, probably one of those buttons will be pressed 80% of the time, and the other buttons pressed less than 10% of the time. How do you style those buttons?
As a programmer, we like to think of the three buttons symmetrically. We want all the buttons to look the same and behave the same, because then they're easier to style and easier to reason about. Our instinct is to make a button class and then place all the buttons next to each other in a nice neat table.
To a user, the three buttons are different, and it should be obvious which one is the button you're expected to press most of the time. From the user's perspective, "Submit form" isn't really the same type of UI element as "reset form". The submit button should be big, bold, colorful and obvious. My eyes should naturally settle on it. The reset form button (if it exists) should be small and non-obvious. Its an advanced feature. It should be out of the way and most people should never notice it.
My email compose window has this problem. It has 3 buttons - "Send", "Save Draft" and "Discard". When I'm writing an email, I'm not choosing between 3 equivalent options. I hit send about 80% of the time I type an email. Once I've written an email, if I visually hunt for the obvious button on the screen, my eyes should naturally settle on "Send". Styling should make that obvious. But no - there are 3 buttons I have to choose from. They're all next to each other, and they're all styled in an identical manner. The interface makes users actively hunt for the "Send" action. This is bad UX.
One day my Android phone updated and the interface for accepting/ignoring calls changed. A circle appears and you swipe up to answer and swipe down to ignore. I get a call, press on the circle, and swipe up, nothing happens. I press on the circle swipe down, nothing. Turns out the interface depends on the distance swiped up or down, and has nothing to do with the circle. By starting on the circle I didn't have enough runway to trigger the gesture. I love puzzles and felt pretty smart after solving this one after only the fourth missed call.
Good one, though I surprisingly rarely use the enter key on my phone. Most text replies are single line and I send them with a dedicated button in the app.
What really gets me is the touch bar on macbooks where I touch it by mistake when typing numbers, sending all my windows dancing around. Drives me up the wall. Leave more space there, Apple! Or just get rid of that stupid touch screen gimmick.
I feel this way whenever I watch YouTube streams with live chat. It’s unwatchable. Unlike Twitch, the chat is an overlay on the video (not sure if that’s true everywhere or not). I’m positive some designer made a really beautiful looking still image without thinking of what it might be like to follow a chat (or watch a video) in this way.
I’m quite positive this is a byproduct of the product->design->engineering waterfall, because while we say we are agile, nothing really gets iterated on once an engineer starts coding. If it isn’t caught in the wireframe phase, it’ll take a redesign to address.
510 comments
[ 2.2 ms ] story [ 355 ms ] threadHuman Interaction Engineers should be dictating this instead; with a focus on making information easy to parse, for humans and machines. For placing action widgets in locations that reduce errors.
https://neil.computer/notes/oh-sorry-i-was-on-mute/
Example: https://pixfeeds.com/images/technology/phones/1280-458994239...
Do all web forms have a left aligned button by default? I just realized they do on HN but if you'd asked me 5 minutes ago I'd have told you it was on the right.
No specific advantage has actually been attributed to either choice of order. What matters is (1) keeping the order consistent throughout an application, and (2) following the provided style guide for the platform you're developing on.
(Also, the order may need to be changed if the primary action is destructive, such as a "Reset" button.)
I did an extensive write-up about this topic if you're curious on more details: https://dev.to/sergix/ux-illuminating-intention-198k
Not even just an application. The whole phone needs to be consistent. Every app in it. Otherwise it's just repeatedly discovering and forgetting how to use each and every app's unique interpretation of what their average user wants and what their average user thinks is intuitive.
The biggest reason why consistency is more important than following platform guidelines is for cross-platform apps available on multiple devices that have different platform design guidelines. It's obviously not feasible (and I would argue not user-friendly either) to switch the order of action controls for the same app on different devices, especially when the app is available via a web interface as well as native.
For example I would think this makes much more sense:
To: In the first example the "Submit" is aligned with what you're filling in: Name, email, that checkbox, etc. Your eye will naturally fall to the "submit" because that's the next in the series.This also works much better if you put the labels above the inputs:
If you click (or tab) through every one of the inputs one-by-one then you'll end up on the "Submit" if it's on the left, but you need to jump to the right if it's placed there.And people read things from left-to-right; I find left-alignment almost always more natural; this is why most navigation sidebars are also on the left (which is often also inverted in websites using right-to-left scripts).
It's even worse if you also place a reset button; I would imagine more than a few people in a hurry will accidentally reset forms if it's placed on the left and has equal prominence to submit. Aas the article already mentioned, you probably shouldn't have a reset button at all, and most forms these days don't.
Anyway, I think "it goes forward in the procedure, in time" is overthinking these things too much in too abstract terms. Just put things where people's eyes and mouse cursor will naturally go for these kind of things will get you a long way.
All of that being said, consistency is also important, so if there's a system/platform where things are consistently on the right then sticking to that is usually more important.
Like spoken languages, the language of design changes over time, and what used to be normal can be quickly outdated (like the reset button on forms).
What is more "natural" is a pointless discussion anyway IMO; I regret phrasing it like that and I wish I could edit it. As I said, the key is to put things where people's eyes and mouse cursor will go, and that's rarely a "jump" to an entirely different place on the screen (on mobile this problem exists less because the screens are small). While it doesn't matter too much on narrow forms (less of a jump), on wide forms it's a bigger issue (or if it's placed to the right of the form inputs take).
Windows: "Right-align commit buttons in a single row across the bottom of the dialog box, but above the footnote area. Do this even if there is a single commit button (such as OK)."
Mac: "Any buttons in the bottom right of a dialog should dismiss the dialog. An action button, which initiates the dialog’s primary action, should be farthest to the right." (Also noteworthy: "Separate destructive buttons from nondestructive buttons.")
Having it on the left, imo, people might look there first but continue to scan to see what the other control is. When you finish reading a sentence of text in a paragraph, you certainly have the last word of it stick out, how is this different?
When forms were implemented, it was important to keep the behavior consistent with what people experienced in their OS.
Now that we have touch phones, the situation has changed, but the best practices have not.
1: This field requires a lot skill.
2: I am not skilled in this field.
3: It's obvious to me <observation that might require skill in domain>.
And then there's no appeal to an expert in the field, or observed behavior.
It's fair that sometimes negative effects are obvious, and if the writer was observing the deleterious effects of the button placement, I could see where they're coming from.
But recognizing when it's wrong, not so much.
The only way to get good design is to hire good designers. Money doesn't buy good design. This UI was shipped by Google. They have money.
Determining which those are can be really hard depending on how your users use your app. And won't always be universal. The example of the "commonly used" closed window versus quit app... the destructiveness of the former is equivalent.
When one is a common action and the other causes more pain when performed accidentally, it's been more helpful to add an option to ask if that was the intended action than to retrain now-generations of users on keyboard commands. Cmd+Q/Cmd+W have been used this way since I was in diapers.
Try finding the menu item that brings back the app once you've quit.
And every browser has a way to reload tabs from the previous session. Both are equally destructive if the tab/s had state you were not finished with and don’t preserve that state when reopened. At least quit can be guarded with an “are you sure?” prompt out of the box.
> Try finding the menu item that brings back the app once you've quit.
The um... Dock or Taskbar or Start Menu or Spotlight or every other thing people use to launch apps every day?
Not if that tab was a google meet meeting.
We don't have to go back to skeuomorphism of the old iphones, but whenever I use my phone I have no idea what's tappable and what's not. There's even this trend now of not making it obvious what's even a text box or what's not. It /looks/ nice, but it's infuriating to use. And then you have the weird gestures that are completely undiscoverable. Like pulling down from the top right to get the utility menu thing. I mean, yeah, I know it's there, but I never would have actually discovered that on my own. Also now there's this trend of just hiding everything to make an app look minimalistic and simple even though it's not. So I have no idea what it can even do when I look at it. And I /still/ have no idea how to properly line up apps side by side on my iPad. I mean I kind of do, but I constantly forget, and worse, once I do have them lined up, it's hard to get rid of them. It's all insanely undiscoverable.
I wish UX designers would realize it's not all about being pretty, you have to actually give people an idea of what things actually DO. All the clunky old interfaces with the bevelled buttons and huge scroll bars and stuff might have been ugly as sin, but at least they weren't a constant confusion.
The problem with flat UIs is that they're also in the middle of abandoning any conventions on what's interactable as well.
Ironically Windows circa 3.1 and definitely by 95 had this nailed: is it greyed out? 3D? You can interact with it. Not 3D? You can't. 3D but greyed out? It is contextually disabled.
Simple and clear at a glance. What that interface got wrong was the MDI motif - multiple document interface never really worked as well as Microsoft wanted, although if they'd made the leap of making it tiling by default they would've got there.
But it's not all bad; you can't put heaps of buttons on a mobile interface because of the limited screen size and lack of precision with pressing them, so some amount of 'magic' and hiding is necessary imo.
In contrast to my father, who gets off work and tries to do something with his computer and has an issue. It's simply faster for him to call me from my room and have me, whose invested tens of hours poking through all the pokable things already to solve his issue. Simply put, between working, commuting, being an adult, etc, my father had no real time to invest this time in pure discovery, when what little precious time he had as an adult had to be divied up in the most valuable way.
I see that now as an adult, free time is precious. You don't have the time to learn like you did as a kid, when you could just throw 8 hours at something. I barely find the time to play my guitar for a half hour a day between working and being exhausted after the working day. I can't imagine having to learn how to use a computer, at this age. I simply have no free time to invest in such with all the other things life throws at you to prioritize, and what free time I do have I'm mentally tapped at that point, and I fully expect in a few decades at this rate to become a technology dinosaur just like my parents and grandparents were.
A couple of examples getting in my way at the moment, the android (or nokia) phone app added a full screen "call your favorite contacts with just one tap" image to the favorites screen. I could do this until they added the obnoxious message, now I have to scroll down to even see them. The other would be netflix constantly AB testing me on whether to show the next episode button or jump back to the home screen, just make a decision, the constantly shifting interface is worse than either option.
In my opinion, my interfaces are getting worse rather than better. I feel like a few years ago, a lot of the major companies hit a tipping point with UI optimization and realized a LOT of what they were doing was unnecessary (and perhaps even costing the business money). Anything that didn't clearly serve a purpose got removed - streamlining for a few core usecases.
- Swipe from bottom to wake. - Swipe from bottom whilst inside an app, peaks all open apps, once let go it fully minimizes the app. During a peek you also see the number of unread notifications on the left. - Swipe from left to go back. - Swipe from bottom, then to the right to access the BlackBerry hub, which aggregates ALL emails/IMs/notifications in a single list (can be customized into groups). - Swipe from bottom left corner towards center hides on-screen-keyboard. - Top swipe displays app specific option/help/misc links. - 2-finger swipe from top revealed quick settings (wifi, flashlight, etc). No notifications here (they're in the hub)! - Not swiping, but it had a clean, single unified location for all app notification/permissions/etc which feels much easier than Android.
I feel if BlackBerry released the Q10 2-3 years earlier we would have had a totally different phone ecosystem today. I still miss it.
On the home screen, swiping down gave you the system quick settings, and in an app it gave you the app's settings, so the two-finger gesture was just a power-user shortcut to the system quick settings while in an app.
Swipe left to go back was very discoverable, and worked differently than iOS or Android, because the entire OS was built around the idea of stacks of pages. Swiping left was just pulling a page off the stack (fluidly animated and cancellable, and you could start anywhere on the screen). There was also a back button on the left-side of the toolbar that you could tap to do the same thing.
Swipe up and to the right to go to the notifications Hub was also discoverable. The Hub was the left-most page on the home screen, so "up-right" simply combined the gestures into one fluid gesture; it was also totally optional.
Swipe left to go back worked anywhere (you didn't have to start from the edge of the screen) and dynamically showed the page being pulled off the stack, so it was super discoverable. All of the swipe gestures worked this way—"peek" was core to the interaction model because it made users feel in control, and let them cancel an action by just dragging back to where they started.
The two-finger swipe from top was only to bring the system quick settings while in an app (the two-finger gesture was also only added late in the OS's life, at the behest of power users); the quick settings were available on the home screen with a single swipe. While in an app, a single swipe brings up the app's settings.
The only non-discoverable gesture was swiping the corner of the keyboard to dismiss it, and it was also totally useless. The advertised way to dismiss the keyboard was by long-pressing the space bar (there was an icon showing this).
It's a pity more haven't experienced the useful aspects of BB10's interface (10.3.2 IIRC was when it received a much appreciated UI appearance update, fwiw).
I've tried Android, iOS, Windows Phone and BB10 is still my favorite. Windows Phone in particular was surprising in how unintuitive various of the gestures were in comparison.
As a side note, on their phones with a physical keyboard (which featured touch detection on the surface of the entire key array) text interactions were much improved. Finessing text selections was particularly easy compared to using the touch-screen. Double-tapping the keys (without depressing them) brings up the loupe and from there one can hold down Shift while gliding around the keys to adjust the selection, much like a laptop with a touchpad.
512 x 342 x 1 bit color created some pretty fantastic usability.
I've wanted a modern consistent, monochrome interface for a while as a general computing interface.
I've been looking at the e-ink devices for inspiration.
Forced contrast, forced visibility, it all has to be apparent. I've been hacking lua for this holy grail for about 6 years now, before that in perl, then in C since around 2001 or so.
I've got foot pedals, midi controllers I use for general computing, lots of little hacks. Multiple 4k monitors in portrait mode, lots of hacking with arduino sensor packs. Still not there yet.
Interfacing is the current limitation in computing. I've got 128 cores, hundreds of gb of ram, and no good way to use it other than the current paradigms. There's gotta be something better
These things have huge resolution now and the user can pick their language; would it really kill them to write words over the little icons? If in English, perhaps "ANSWER", "HANG UP", "MUTE" and so on, but the magic of words is that it doesn't have to be those exact words ("HANG UP" and "END CALL" are completely different sets of words yet in the context of a phone call, mean the same thing to almost everyone - magic).
Words. They carry so much information. So much. I know, people want to find some magic picture that carries full meaning to all cultures, but there simply ain't no such picture and there never will be; there's no shame in using words. Please. Use words. I'd even happily take them in a language I don't even speak, so long as it used an alphabet I could read (or even not - I can read some common software related words in Japanese simply through having sounded them out a few times). For me at least, words are easy; the ever-changing mist of icon-style-du-jour is not.
But each time it gets updated while you're editing a field, it gets rid of the keyboard, and maybe of the info you just added.
It's extremely painful to use.
I had assumed you'd just tap the "answer" button, but that failed more often than not. It never would have occurred to me to swipe a minimum of 3cm, starting with the answer button, and I must assume this knowledge has spread to users by osmosis rather than discovery.
Typically it is a tap and drag in some direction, but the direction to drag for answer vs hang up is always different between OEMs.
So that's fun. :/
Having said that, text has an annoying feature of needing to be localized, and localization of button texts is tricky because you want a very short string, in order to not overflow, because you have very small screen estate on mobile.
(It doesn't also help that some native built-in components have APIs that show the icons only, without text)
There is no creativity any more. Everybody is too busy following apple
I want ugly buttons with clear text on them.
What I definitely don't want is four almost identical icons with nothing to tell me what any of them do, or swipe gestures that are impossible to discover, or (the worst) swipe gestures that expose icons.
And then, the absolute worst of all, the useless "upgrade" that offers zero additional functionality, but changes the location or design of all the icons of an app that you've already learned.
This[1] is what it looks like when the alarm goes off. (Swipe left to the "zzz" snooze icon or swipe right to the off icon). I can't always rely on muscle memory because I am not always consistent in which direction I put my phone down on the nightstand. This is not the best UI for a person who has just been woken out of sleep and still groggy.
1. https://www.androidpolice.com/wp-content/uploads/2015/07/nex...
Phone interfaces are a minefield for me for some reason. It's too easy to accidentally push a button you didn't realize was a button when you were trying to scroll down the screen but pressed it too hard when you did so. I run into lots of pitfalls like this. I grew up in the era where the OS interface was a "READY" prompt, so I've seen most of the paradigms out there over the years. Nothing has annoyed me as much in recent memory as having a touch-sensitive interface on a slick, compact phone with rounded edges that I desperately don't want to drop.
- most of those apps with the hidden interfaces have no business being on a pocket computer in the first place
If we stopped playing the game of maximising engagement, and returned to "practical usefulness" as the primary driver for application design, we'd probably make better apps.
The solution is:
1) Publish Human Interface Guidelines that detail a rich set of standard gestures, how various tappable elements MUST be marked, and how they SHOULD be arranged.
2) Publish abridged HIG for end-users as a part of user manual for the device/platform. Aim for closed-world reasoning, i.e. the user must be able to build a mental model of, "if I can't see this functionality here, here or here, it does not exist", and not "it may be hidden somewhere else".
3) Tell app developers to stick to the HIG or GTFO.
I know, wishful thinking.
The decay started long, long ago. The other day, someone commented on HN with a link to a piece of old Microsoft WinAPI documentation, I think Windows 95 era or older, where there was a side note on window styling and "escape hatches" that unfortunately had to be built in, because marketers are marketers and desperately want to fuck up usability to put branding on things. Back then, platforms already gave away too much control over styling to software vendors (where originally this control resided with end-user).
It's interesting that nowadays, with all the work gone into sandboxes and frameworks, nobody seems able to do that.
I recently got an Android 10 phone, from a vendor I had never heard of: Ulefone.
I was coming from a Sony Xperia compact running Android 8 - roughly the same hardware, except not rugged. I presume Sony polished Android a little bit more than Ulefone did, but even taking all that into account, Android 10 was full of shockingly bad UI "decisions".
First: the Do Not Disturb icon in the pull-down thingy is a but when enabled a appears in the top bar.
Second: Pulling down the pull-down thingy once displays a row of 6 icons with no labels. Pulling twice displays 5 columns of icons with labels. It takes extra planning to rearrange the widgets in 5-column mode such that I get the 6 I want in the first row but also a logical grouping in 5 columns.
Maybe these are Ulefone-isms, but whenever I trip over them I imagine how hard Steve Jobs would have fired someone who put this stuff in front of him.
Just yesterday, I wanted to make a normal voice call to someone I usually contact via WhatsApp. When on their contact page, I couldn't tell which icon would make a WhatsApp call and which would make a PSTN call. The WhatsApp phonecall option appeared with the phone's "phone" icon, and the PSTN option had no icon at all. Both had text saying "voice call".
Oh, and don't even get me started about my Android TV. FFS.
I noticed a lot of UI problems like this when my parents changed from Windows Phone to Android and I had to help them. There are a lot of small actions that make sense if you used Android in the past (since they've been slowly introduced) but are completely bonkers to anyone picking it up now:
- To reject a call, swipe the "Accept call" icon down (this one is particularly horrible); - To enable Wi-Fi, bluetooth, etc. swipe from the top; - To dismiss a notification, swipe it to the side.
Spotify on Android is pretty awful at this point. Now I can't tap and hold to get to the context menu anymore, now I have to use the triple dot button. Why remove that? Usability is also pretty terrible. Why can't I load the list view of an album I have saved? What the hell. Same with entire playlists that I have saved and downloaded.
Snapchat (I wish I didn't have to use it, but it's the main for of communication a lot of my peers use) keep changing things constantly when nothing was every broken. Over the past three years, they've changed the order and number of tabs they have at least four times.
Google Photos on Android is also annoying separated IMO. Give me an order by date and an order by album/folder. Please just load the folder structure. I assume they do things the way they do to maximize use of their cloud storage, so it is likely intentionally awful for local images.
The official Reddit app is absolutely awful and slow. Reddit Is Fun (third party app) is, and always will be, my favorite Reddit experience. Clean and fast.
That was kind of ranty, but I just wish UI's were simpler on mobile.
The thing the author doesn't seem to realize / acknowledge is that UI/UX design is about balancing enormous numbers of competing constraints and concerns. It's a human problem, and like most problems of this genre there is no "best" answer, only different sets of weights for different, often competing, concerns. Simplicity and ease of use are good things, but they are almost always at odds with flexibility and power...also good things. Do you make things bigger with more space so they're easier to see, or do you make them smaller so you can fit more things on the page? Do you use color so you can communicate more and catch the eye more quickly, or do you avoid that so that your app is friendly to colorblind users? (Yes, I know there are color schemes that can achieve both to a decent degree.)
These kinds of tradeoffs are lurking almost everywhere you look in UI/UX design, but this kind of nuance seems to be lost on the author. He only seems to see his set of priorities for a UI. Yes, there are plenty of cases where one thing is pretty objectively worse than another, but usually it's more subtle than that. I'm way more impressed with someone who can talk intelligently about the tradeoffs than I am with someone who fixates on something that very well might have been traded off and rant about it.
Everyone and their mother has an opinion on what they think is good design. I am thankful that as a developer I am not having to fight over pixel pushing.
As an example, my favourite pet peeve was using Visual Studio with old-school, upfront locking source control (like TFS), and then accidentally drag and dropping a file or folder due to lag in remote desktop, or a failing mouse which sent two click events in 1 ms or something. VS duly pre-emptively locks the 10k files in the folder you just dragged, and begins a 5 minute operation you'll have to somehow undo later, even though it should be fairly obvious from the click events that it was non-intentional.
Going back to the meeting example, surely solid, accurate taps in the center of the hang-up icon could be taken as intentional, but a kind of glancing, less accurate one needing confirmation?
Another instance of the same principle is that if an unexpected button/element appears and I click it in <30ms or whatever the fastest possible read+react time is, the click isn’t intentional and should be ignored or confirmed. This should scale over time based on user familiarity and speed.
This culminates at:
> as the technology became more sophisticated the controls were made touch-sensitive - you merely had to brush the panels with your fingers; now all you had to do was wave your hand in the general direction of the components and hope. [HHGTG]
> if an unexpected button/element appears and I click it in <30ms
Easily solvable by locking the buttons for a few seconds (or less).
(I confess that the first time I encountered this I opened Task Manager and found and killed a Zoom.exe process all by keyboard, just on principle. That was when I discovered that Zoom has two processes so that the main one can restart the call if the call process crashes!)
But you know one potential factor for my never hitting it by accident? They use text labels (“Leave Call” / “End Call”) rather than an icon. Much easier to get right.
The more contextually 'intelligent' a system is, the harder it is for the user to model it, and ease of user modelling is often more important than reducing the number of interactions need to complete a task.
In this particular case, it wouldn't be possible to know if your keyboard sequence that quits a call would feature an unnecessary enter at the end or not.
In my experience designers are the worst people to hire for UX because they often sacrifice usability for aesthetics. Programmers fare a bit better (usable UI usually doesn't conflict with the code quality), still not perfect though. Casual users are probably best, with some education in usability of course.
https://news.ycombinator.com/item?id=24760194
If you happen to do design work and are looking for a job, I'd love to chat and see if there are any possibilities for collaboration. If you're interested, feel free to drop me a line at my username at google's mail service.
For this specific problem (Google Meet UX/UI) sure anyone agree with: don't put a destructive action that requires no confirmation (leaving the meeting) next to a common action (mute/unmute yourself). If designers don't get that right, sorry but they are not competent designers. It's like a programmer that, in order to "balance enormous numbers of competing constraints and concerns" decides to not escape user-provided HTML in the frontend. Well, that programmer is not a competent one.
I agree with this point, but if the designer had a phone call in mind (with a pre-pandemic mindset), then the design feels more reasonable.
In a one-on-one call, 'mute' is much closer in intent to 'hang up' - i.e. I am not currently participating in this call.
To underline this point, I think in general design suffers from a lot of bikeshedding at tech companies. There are likely many designs tucked away in discarded files that addressed this specific pain point, too, but were discarded at the request of some PM, or manager, or director. Then there's the process of user testing and experiment design that is used to validate UI/UX of products like this. This design may have actually tested well even if it wasn't the team's favorite... I've worked places where that data is used to override a designer's opinion.
Now apart from all that of course, Occam's razor probably applies as well I guess... perhaps this bit of UI was just poorly designed. But I see a lot of chatter on HN regularly about design being superfluous, subversive, unintuitive, "bad" when really I think many (most?) designers are unhappy with the designs that ship out as well.
I can't be the only one getting confused about this on a regular basis.
It's rare to see it solved cleanly. I try to avoid interface elements that attempt to combine these two roles entirely.
I quite like the way the Twitter "Follow" button works: when you click it the text changes to "Following", which I think is just clear enough, but only because it's a button that every Twitter uses frequently enough that they are likely to remember how it works.
This and "Your video is OFF" are states I care about when video conferencing because the reason to enter them is important.
There is perhaps a patent which prevents it for financial reasons.
[x] Mute microphone
(apparently HN doesn't permit the checkbox character)
Yeah, there are countries that use keyboards where Q and W are apart, but that's not common.
Firstly, I've observed a phenomenon when I talk to a programmer about a problem I want to solve. They might get excited, learn a bit of the problem domain and then go off and build something that kind of solves part of my problem. But then I'm stuck in a endless cycle of explaining the rest of the problem, to someone who isn't really interested, then waiting for a new iteration and then evaluating how it doesn't completely the solve the problem until the programmer gets bored and goes off to find something new and exciting to do with computers. Programming the computerwas always the end for them. They don't actually care about the problem. I care about the problem. So for me computers are a tool to solve my problem. I might as well write the program myself because it's less work to become a mediocre programmer than to intimately understand the problem.
My theory is that something like this happens in interface design. There are designers who love to create beautiful designs, and we love that, but beauty is the end of it for them. They don't actually care about the problem that the program (website, app, whatever) is supposed to solve. If they were stuck in a job filling out forms all day they would quickly learn the lesson to put the reset button out of the way. If they were intimately familiar with the problem, the better design would be obvious. And so it is that people with no "design skills" can point out the obvious mistakes of the designers.
It's also true that good design is hard. Starting with a blank sheet and creating something, let alone something good, is daunting. Perhaps the hardest part is having the humility to admit, "I don't understand the problem sufficiently" and the empathy to care about the problem enough to learn it well.
There's this widespread belief now that software is only good if a user who never saw it before can become proficient in it in seconds to minutes. I think this is one of the most devastating, dangerous ideas in computing. The only way you can achieve a learning curve like this is by removing almost all functionality from software - make it so dumb that it really takes only a minute to figure it out entirely. Sadly, this is what we see in mobile and web applications these days.
What worries me here is that we've conditioned everyone to assume software is immediately and fully discoverable. Nobody is expected to read the manual these days, and so manuals are not provided, and since manuals are not provided, any feature that cannot be made apparent without explaining it in the manual goes away.
(Even with kitchen appliances, the situation isn't that bad. When a person sees a particular appliance for the first time, they do read the manual, or get someone to show them how to operate it. Maybe it comes with the fact that buying appliances is expensive and overall a hassle, whereas software is too easy to procure?)
--
[0] - I highlight that condition, because it's my belief that most software vendors don't care about delivering value to users. They care about making money off users, and there are many cheaper ways to do that than creating a truly useful and ergonomic product.
I *try* to make my software initially intuitive and gradually discoverable. I think a gentle learning curve is better than a completely flat one.
I thought quite a bit recently about the problem that you are talking about (diminishing end user value in software). I think the source is that we have a lot of devs and designers whose only experience is designing for maximized conversion rate or maximized engagement. As opposed to professional software.
They have no habit of designing for value.
And *that’s the culture*. Go to any awwwards gallery - all of the websites or webapps mentioned there are essentially ads with minimal content.
Which isn’t too bad, since it leaves a market untapped through the collective arrogance of the incumbents...
Right. We're prioritizing learnability over usability.
> We should be making the easy things easy, hard things possible, and forget about the whole intuitiveness thing.
Whoa there... Another way of thinking about 'intuitiveness' is in terms of affordances. We mustn't throw the baby out with the bathwater. While it is too much to ask that every function should be obvious upon first seeing the UI, it is not too much too ask that every function should at least be obvious in retrospect after trying to use it or having it demonstrated, and of course leveraging affordances and interaction patterns the user is likely familiar with from elsewhere should be given priority.
Yes, of course. Thanks for bringing this up. I apologize, I went a bit too far there - what I meant was just "intuitiveness" in the sense of expecting people to be immediately able to work well with something they see for the very first time, with no explicit learning or training.
I also don't mean to ignore familiarity with UIs in general - yes, unless you have a good reason, it's a good idea to copy design elements users are well familiar with (if they're not completely insane, or dark patterns). This matters particularly on mobile and desktop. On the Web, everyone is used to websites looking different from each other, but there are still higher-level patterns (like footer with company info, "contact" link somewhere on the site, site logo redirecting to home, etc.).
I totally agree about affordances, and mental handles in general. "Obvious in retrospect" is a great way of putting it - once you know a feature exists, or used it briefly, it should be easy to find it again. Once you familiarize yourself with a bunch of features, it should be obvious where to find them, because they should fit a consistent mental model. In a way, it's the job of the UI - to let the user learn the correct mental model of the application, and how it manipulates underlying resources.
For that to happen though, you as a software team need at least to a) have a consistent mental model yourself, and b) design both "backend" and UI around that model. I think this is one part where we fail frequently, but unintentionally - the developers and the designers don't spend enough time ensuring they have a shared mental model. When this happens, you have UI that may be consistent with itself, but feels off when used, and every now and then you see surprising behavior or incomprehensible error messages - that's the "backend" model leaking out.
The fact that the code is available means someone can fork it and make their own changes and improvements.
Therefore, in this project you'll be helping a professor from some non-CS department write educational software. You have to balance accessibility with pedagogical accuracy, as guided by someone who doesn't secretly know that some specific algorithm is the magic key to solving the problem.
If the stakeholder has such a clear vision of their end result, they probably won't need you to help them implement it!
This is giving a false complacency to what outta be a tight intergration between design + developers with an oversight from the product owner. It is simply inexcusable. We're actively creating a culture of disregard/complacency/ignorance by not shedding light on this as a huge problem in UX/UI design.
I want more people to make noise about this.
I do believe there are programmers for every category though.
[1] https://blog.gardeviance.org/2015/03/on-pioneers-settlers-to...
The thing is: both design and programming are more often than not messy when things get real, but every programmer wants to create clean and beautiful code and every designer wants to create aesthetical designs. Many intuitively perceive this "beauty" to be in tension with the complexity of the problem that needs solving. The goal of programming and design is however not to create beauty, but to solve problems and do so beautifully and clearly communicate these solutions, because it doesn't matter how beautiful your code/design is if nobody understands it. Beauty is the cherry on top that you get to achieve once the cake below is done, tastes well and puts a smile on the faces of all the party guests. If you are good you might already plan the cherry into the shape of the cake at the very beginning — but confusing the cherry for the cake IMO means you are either not there yet in terms of your profession or the thing is a toy project to try things out specifically — which is totally acceptable if everybody involved agrees and totally egoistic and shitty if they don't. Don't agree to serious projects you are not willing to commit to once the initial interest fades.
Really good designers are about integrating both usabilty and aesthetics in an iterative process, and do so in such a way the problem is solved, and solved beautifully, all while reducing the cognitive overhead needed by users.
And this cognitive overhead is really what should drive us. The work of programmers and designers is so beautiful/dangerous because it multiplies to a thousand people for a thousand hours. Our decision affect people on a daily basis. Removing a papercut might seem like nothing, but if it avoids irritating even ten other people who use that thing 10 times a day, isn't it the obvious thing to not half-ass on, not to toy around with?
Btw. Design that is only about aesthetics is something that I call "styling" maybe we need a similar word for programmers who toy around?
There's a research field called Semiotics of Human-Computer Interaction studying user interfaces from that angle. Developers usually have a technical background from maths or engineering, and often are not aware of the importance of linguistics in our field.
Both programming languages and GUIs are languages (artificial, sure, but linguistics also study those) which are used to write expressions that can be read by humans. Semiotics, the study of signs and their meaning, provide methods to analyze how users make sense of the software artifacts (products and tools) delivered to them.
One researcher called the user interface a designer's deputy, i.e. a messenger that conveys in its entirety a message that the programmers wants to communicate to the users. This deputy must stand on its own, since it's the only information available to the user.
Users then perform sense-making on the symbols in the interface, to infer the meanings of all elements [1]. Notice that the user can't see what each interface element actually does, since they don't have access to the code; they can only guess what it does from the available symbols. If the symbols lead the user to infer an incorrect meaning, communication breaks.
[1] https://blog.prototypr.io/how-semiotics-can-help-us-in-ux-on...
That’s how I identify what I understand and what I don’t.
I think a lot of the methodological knowledge built by usability people has probably been ignored in favour of a metric-driven approach these days. That approach is very good at identifying bottlenecks in existing workflows or deciding between two ideas, but ultimately lacks the empathy you describe. You can only really build that by engaging directly with users, and watching them suffer at the hands of your creations.
Do normal users know all this? Or do they just only use the 10% easy functionality of the phone and is happy with that? Maybe articles like "17 Things You Didn't Know Your iPhone's Home Button Could Do" is a sign usability has been sacrificed in order to have a clean&neat design.
Of course menus don't translate well to touch devices, and menu systems can quickly become unorganized junk drawers for functionality, but it feels like discoverability was never fully solved on touch devices.
And it also has the same problem you describe - you don't know how to find something unless you know what it's called. And even if you do find something, there's no breadcrumb trail to learn how to find it without using search. You have to look at what's on the current page and like some kind of reverse engineering sherlock, think about where a designer could possibly have put this page. It's insane.
Sure, discoverability is good to have, but it is not a failure that you need to be told about it as long as it only contains features useful to power users.
Requiring experts to undergo some training to make the most out of their devices is acceptable UI if the final interactions are optimized for their use; double so if it doesn't interfere with usage by untrained users.
I wonder if it's kinda a stockholm syndrome or if it caters to one's ego somehow. After discovering a feature, one feels smarter / more connected to the device, compared to if it was actually easier to use from the get go.
For example, he texts frequently but doesn't know how to check his email. Because I am the tech support child, he reaches out to me for iPhone help. However, since I am not an iPhone user, I always struggle and sometimes fail to solve his problem.
Perhaps this is a sign I have not dedicated enough time to learning this new OS in my family ecosystem, but I do feel it is the least-similar to the other OS's I have used. It's like if my dad was learning programming and asked if I'd help him with his Lisp program. No thanks.
Though maybe I am just a luddite and the universe is written in iOS.
Not just unfamiliar, difficult.
When that phone stopped working I got a decent Android phone and it felt so much more comfortable.
I will say that when it first came out the direct manipulation of the iPhone UI was a qualitatively different experience compared to any other touchscreen for a very long time. Back then most Android reviews included caveats like "stuttery", "janky", or "low resolution", and these had the cumulative effect of spoiling the illusion of direct manipulation.
Today the gap is much smaller and easier to ignore, and the iPhone has added lots of hidden affordances like swipes and double, triple, and force taps to cater to expert users at the expense of novices.
One danger is if the affordance is triggered accidentally. For example, my kids like pushing buttons and swiping when I wouldn't think to do those things and accidentally switch apps or enter into guided access modes without meaning to. In my case I occasionally trigger things like sticky keys unintentionally.
Another danger is if application developers start assuming that they can rely on users knowing hidden affordances and use swiping, double-tapping, force pressing, right-clicking, etc in the core application workflow. Not an Android iOS example, but last night one of my kids started playing Stardew Valley as their first non-tablet game. It has a steep learning curve since it relies on multiple different keyboard keys and on differentiating between right and left-click, whereas I wouldn't consider right-clicking to be a novice skill.
For complete novices, the first thing they do is click the mouse, and since there are only two buttons it doesn't take long to figure out that left-click is the primary button, and right-click brings up a menu with handy options (practically everything has a context menu in Windows). But with the Web, where context menus are rare, I wouldn't be surprised if it's not as well-known anymore.
I have theories -- maybe it's because most web apps do not bother with context menus (outside of text operations), maybe it's because more people have developed their mental model from touchscreen devices where context menus are less commonly used, maybe more people are on laptops with trackpads that do not make right-clicking as obvious, maybe because there's no visual affordance indicating which onscreen elements have a useful context menu and which have the standard webpage context menu so it's a guessing game. I don't honestly know. But I am confident that if you place key functionality in a context menu of a web-based app that most novice users will not discover it on their own. As per Jakob Nielsen: "...be warned: less skilled users rarely use these [context] menus." [1]
Once end-users learn how to use your app (and therefore are no longer novices) then they seem to have no trouble remembering and using context menus, so it's a great way to expose expert affordances.
In the case of my kids I can also say that the Stardew Valley user interface (such as keeping right-click and left-click actions straight) has been the most difficult part of the game so far. And it doesn't seem like it was particularly necessary distinction to make -- there seem to be few cases where both right- and left-click actions are equally appropriate.
[1] https://www.nngroup.com/articles/feature-richness-and-user-e...
While that's the only way to build empathy, you should be aware (and beware) that it is also a good way to build contempt (as counterproductive as that often is). Many of the "dark patterns" probably originated that way.
I know, there are open source business models which work for some companies and that's true but generally in the sorts of technical problems for IT people. Compared to the IT industry as a whole those are edge cases. They seem more important to us embedded in the IT industry, but compared to the overall multi-trillion dollar global IT industry they're peanuts.
Did you mean MacOS?
Seriously, Linux has a lot of UX issues, but hiding the full path of files is a problem that plagues literally every single OS file manager. It seems that at some point, all the major OS makers agreed that users are too stupid to understand a file hierarchy, and came up with tricks to hide it.
I find this annoying as well, but it only happens when you use quick access shortcuts. And then once you're in a sub folder it'll display the full path in the title bar provided you turn on "Display the full path in the title bar" in the Folder Options.
Navigating directly to c:\Users\MyUser\Downloads will also show the full path in the title bar.
It seems some of these shortcuts have "special" behaviours in explorer.
And this is why I find myself mostly in the text shell.
Were I to go graphical the minimum I want is a simple address bar that I can override the current path with a specified path at will.
Maybe a tree view like in Windows 95 Explorer just to pretty things up. Some of the modern comforts like thumbnail images would also be welcome.
I wish they would stop insulting their users who have taken the time to understand the underlying technicalities of a file system. All this should be user-configurable in a good file manager. SpaceFM and PcmanFM come closest to what I've specified.
That's not a difficult question. The only time it makes sense to expose the internals is if you build a specialist tool for people who like internals exposed.
This means that no, it shouldn't be "user configurable" either. You don't want a combinatorial explosion of states in your UI for a tiny set of users in the general purpose case, and you don't want to create a honking fat tool for specialists who'll only use the specialist path.
The computer industry is discovering what the mechanical tools industry has known for a while: You can build a general purpose tool that's simple & straightforward, or you can make a specialist tool.
Pretty much any specialist will be slightly unhappy with the specialist tool and modify it to their own requirements, but you can't build a specialist tool that makes even close to all specialists happy. The specialists will also continue to use the generalist tool if it gets the job done, but they'll complain every step of the way.
- Drag and drop the file into a terminal to insert the path
- Copy and paste the file into the terminal, which inserts the path
- Press Command+I to show a popover describing the path to the file
- Enable the "show path bar" option in the Finder's "view" menu, which will show a nice graphical representation of the path to the current location at the bottom of the window (which lets you copy the POSIX path to each item)
- For more advanced users, run the command `defaults write com.apple.finder _FXShowPosixPathInTitle -bool YES` to show the full path to the current folder in each Finder window title
THANK YOU!
- use columns view
Ah, and also if you alt-rightclick on an item, you can copy its path to the clipboard. Or you can open the ‘Edit’ menu with alt pressed and copy the path to the current folder. Apparently there's even a shortcut for this.
It's certainly fine if you work a lot locally with the Mac; but for me, the cost-effective thing has been to fumble around each time (since I only do this a few times a year) rather than memorize handling that I use so seldom.
Also, nothing to memorize about copy & paste or drag & drop. I would say you’re used to Linux GUIs not doing what you expect so the obvious solutions aren’t even considered.
Showing the Finder path bar is the more obvious option. Nautilus, the file browser on some Linux distros, supports this as well.
To a terminal. I thought Linux was the one where you needed to open a terminal to do basic things, yet from my experience it's usually MacOS that needs a terminal for such basic things as turning off mouse acceleration.
- Drag and drop the file into a terminal to insert the path
- Copy and paste the file into the terminal, which inserts the path
- Press Alt + Enter to see details describing the path to the file
- Press Ctrl+L to see full path. Or just use better file manager like Nemo
In macos I don't have to edit random .conf files in system/application directories just to make my wifi work correctly. The graphical UI is configurable enough that I rarely need to modify anything outside of my normal user directories.
Where as in ubuntu the forums/stack overflow usually say "go to this file at /sys/whatever and write in this". Hence the need for full file paths.
Quick example, because I'm trying to make my nas usable on a mac and it's given me a headache.
https://support.apple.com/en-us/HT208209
The first thing I do on any Mac is to drop into Terminal to run this:
And don't get me started on the wiping `/private/var/db/mds/messages/${UID}/se_SecurityMessages` after each update, without which SMB doesn't work.It's quite a basic thing, most games include it in their settings yet it doesn't seem important enough to have a setting in the OS itself.
Terminal -> Finder:
Finder -> Terminal:I made a Bash function:
True of any hotkey.
> which is why requiring it for a common UI function is bad design.
it isn't required.
In open source the developer has the power, they get to decide what they work on and nobody can tell them otherwise. The best a user can do is post a begging letter in a bug tracker. It doesn't matter what the user thinks, even a majority of users, all they can do is ask.
In commercial software it's the user putting bread on the developer's table. The user feeds and clothes the developer's children, and/or pays for their supply of mountain dew. If the user wants something, bye and large they get it, or at least they have a pretty solid chance of it more often than not. A vote for a feature speaks a lot more convincingly when it's backed up by a wallet.
There's nothing wrong with open source or free and libre software. It's great, I love it, but a lot of it's proponents seem to think proprietary software is some sort of crime and genuinely don't understand why proprietary software dominates so completely in so many domains outside of IT infrastructure.
Developers know best, even when they don't, because they're the ones doing the actual work. But in the FOSS ecosystem you have: i) the freedom to do the changes yourself ii) the freedom to offer help with design, documentation, ideas iii) a public bug tracking and issues manager
Let me know when you can do this commercially.
On a tangent: it is indeed a crime to use commercial software when there are libre alternatives; every use moves the needle one tick deeper towards Eternal September
For me, a large part of the fun of being a developer is enabling others to do stuff they otherwise couldn't. As such we absolutely entertain feature requests and similar, and implement a lot of them.
When sales come back from a sales presentation laughing and telling of jaws hitting the floor, it's almost always due to features that started as a suggestion from one of our users.
Very often though the feature requests are trying to solve XY problems. Often there's a better route to achieving what the user wants, which almost always is some way of avoiding redundant work or other workflow simplifications.
Us devs often do know best when it comes to edge cases and limitations, and about other use-cases that this particular user haven't considered.
However most requests are born from something real, so we will usually inquire what the user is after, in an effort to determine the impact and alternate routes. I might even contact other customers who I know use that module or have a similar work flow and ask them what they think.
And based on that implement changes that make the program better not just for that user but for all our customers.
unless it's accepted upstream, you've got an ongoing maintenance problem on your hands. getting an idea in isn't always just about time/resources/money. if your idea doesn't fit their 'vision', it won't be accepted, regardless of how much you have funded your feature. do you now take on maintaining a fork? sometimes the answer might be 'yes', but I suspect in most cases it's going to be 'no'.
(I've given up either reporting bugs or, where at all possible, using their software, as it's abundantly clear my interests and theirs are not in the least aligned.)
If they are critical bugs, they might or might not be addressed.
At a certain scale (much smaller than Google's, probably already at about 100 customers) it's impossible to please _all_ users of your software, so you try to please the majority. And whatever you change, there will always be that 'one guy' whose workflow will break.
The Google model dominates the proprietary world presently, and even long-term shrinkwrap / clickwrap vendors such as Microsoft are shifting in whole or part to advertising-supported software.
What I pay for Google software is indirect, but given a roughly $100 billion global spend on online advertising, allocated largely among the world's richest 1 billion people, that amounts to about $100/year for the privilege of tools which frustrate rather than delight me.
As I've described in "The Tyranny of the Minimum Viable User", odds are strong that mass-market software of any stripe, including proprietary whether paid, subscription, or advertising-supported, will fail to address power-user / elite-user interests:
https://old.reddit.com/r/dredmorbius/comments/69wk8y/the_tyr...
> why can't I easily make a shortcut in ubuntu? > how about the fact that you can't get full file paths shown in the default file manager,
Is it any worse than needing to find the setting to display file extension in windows?
I am on Cinnamon desktop as it's a bit old school and predictable (for people who have been around since windows 98). I right click a file and I have an option to create a link. And my default file manager shows the full file path in the address bar . So I don't think it's a Linux problem rather than a Gnome 3 problem. But I am sure plenty of people don't mind Gnome 3 (everyone else at my work used the standard Ubuntu desktop.
Blame GNOME 3. After a decade it's still lacking polish everywhere. The day Canonical decided to drop the ball on Unity was a sad day.
It was scary how close we got to some published prices.
What do you mean by "proprietary" in this context? I believe that the availability of the software's source code to its users is unrelated to the fact of whether the actual development of such software is paid or not.
Moreover, free and open source software (free as in "freedom", not necessarily free as in "free beer"), makes it generally simpler for the regular users to provide input regarding the features they need changed. So, in a way, it also helps to achieve the goal you describe:
> The best way to ensure people pay close attention to those last areas of fit and finish...
Don't get me wrong, a good interface needs to take user feedback into account. But trying to accommodate the union of all features needed by all users is a recipe for madness. Somewhere you need an engineer, PM, designer, CxO, or someone who can make judgement calls and decide which user needs are more important than others.
This is true, and I've seen that happen too, one of the companies I worked with would sell source licenses to customers, and as I understand it this was very common in the mainframe business going way back. The software vendor retained rights to sell and distribute the software though, that's what I mean by proprietary.
I've seen these arguments for open source and libre software before many times, but there's a huge discontinuity between the theory and what actually happens in reality. In the real world there are tens of thousands of small and large software houses producing niche software for diverse use cases for businesses all over the world. Hundreds of niche engineering design, test and optimisation applications, B2B services, audio and video tools, chemical engineering tools, automation and industrial control systems, booking and billing systems, here are an almost infinite variety. Most of them are only known to people actually in these niche specialisms. In that world customers paying for customisations is stock in trade, it's entirely normal. In fact the company I'm at right now is paying the vendor for customisations to our incident and change management ticketing system.
In comparison open source, outside nerdy IT oriented tech projects, might as well not exist. It's minuscule. Barely even a footnote.
If only that were true. I've watched a completely non-IT person work with both vanilla Gnome 3 and Windows 10; they found the former far more intuitive and unobtrusive (not looking at the individual apps, just the basic desktop UI). Proprietary design-by-committee doesn't necessary make for better solutions: hence the numerous complaints recently from people (often with accessibility issues) trying to book vaccination slots on government websites.
If you go straight from problem to coder, you shouldn't be surprised that you don't have a proper solution, because nobody really made a proper solution. That's like building a house without an architect. Hire an analyst.
This is what agile was created to solve. It was an acceptance of the reality and an attempt to live with that rather than trying to bend it to your will.
Agile does not exclude different roles, on the contrary.
And if they are, it often ends with impossible to implement/maintain designs.
Do you think business analysis is about software architecture? Oh boy.
Given the tendency for software architectures to mirror org-charts, the two aren't as disparate as you think.
Have you ever played that telephone game?
There should actually be testing associated with every step:
Is X a real problem worth solving (severity or cost x frequency)?
Would Y be an appropriate solution (test a mockup, prototype, or stub UI)?
Does the Z implementation work (as in, not just does the problem get solved when Z is used, but does Z actually get used to solve the problem)?
In defence of the developers/programmers, many people are really bad at explaining what they want, not least because it's often not what they actually need.
If a designer is designing something that looks good but doesn't work well (i.e., solve the problem), they are bad at usability/UX. They should be getting user feedback on their design before it's implemented - that's what user-centred design is about.
There's often a gap when it comes to understanding requirements - between the end-user and the developer, between the designer and developer, etc. Requirements gathering is actually a specialised skill and one of the key duties of a business analyst.
Some days ago I wrote here in HN about the change my stock broker made on their default home broker. It immediately becomes obvious that the people responsible for that design have never performed anything nontrivial with stocks. They created a symmetrical, colorful platform. The symmetry is enforced, therefore you don't have the flexibility of setting up your quote-boxes any longer (perhaps because that would break the artist's concept). Just to mention one of many issues the new design brought.
The same applies to the Google Meet example in the article. Functionally, you'd never want those buttons presented that way. But placing them like that makes them look nice and symmetrical and Gestalt-related stuff, so that's the way to go and deal with it, dear users.
> Perhaps the hardest part is having the humility to admit, "I don't understand the problem sufficiently" and the empathy to care about the problem enough to learn it well
I agree, but don't see that happening, not in the short term at least. I had an argument with a designer some time ago about how much longer the project was going to take and how her ideas were actually substracting value from a user perspective. But she ended the discussion with a "this project has my signature, my reputation is in there". And that's what my criticism against current UX trends is all about. Artist's concept trumps user needs, project maintainability and everything. They are investing all the resources on graphic design matters and this is obstructing the whole field.
Some UX folks argue that designers like those "aren't true UX designers". I agree, but those "false" designers seems to be outnumbering the "true" ones. The latter may eventually have to found a new discipline.
Concurrency is excellent as an idea for systems design but horrible for user interfaces. Nothing seems to provide a stable interface and some concurrent process considers it it's privilege to suddenly modify a list of things I'm choosing from .. resulting in the items shifting just a few milliseconds before I click or tap my choice, which results in the wrong choice. This shift happens not only in vertical lists, but also tab-bar style buttons.
To be precise concurrency isn't to blame for it, but it's more like laziness. The interface elements should be locked in place if the system detects that I'm about to select something. Whatever else is waiting to show up can wait, because I obviously didn't need to know about them a few milliseconds earlier.
Or a window from a different app pops up (the app took a while to start), steals the Enter key press, interprets Enter as "Yes do [something]", and then does it (but I didn't want that!).
Why can't OS windows be click & mouse disabled for 2 seconds, after they open
But what I think you really mean is that people are often in a blame culture that strategically puts up responsibility defenses which distracts them from producing a good result and focuses them on red taping to cover their asses. It is quite possible in these organizational structures to produce absolute shit and tick all the boxes and get everything approved and make sure no one in the position of actually doing stuff gets blamed for anything.
Not saying that it's their personal fault that certain companies are that way, but it is possible to have some self respect and march back up to the manager and say hey this is shit and we should do a better design rather than waste time on this (ok, yes in a more articulate and respectful way) - if you think that risks your job then it's probably going to be more fruitful working somewhere else anyway. Disclaimer: yes I know real life has other restrictions that means not everyone can do this.
However, with dynamic display-based interfaces, you can do just that. Much as I tend to loathe Apple's design, the way they handle iPad shutdowns is quite good: first, you press a button; then, the entire display changes to focus on this task. If you want to shutdown, you then have to perform an entirely different gesture to confirm. This layering of visual feedback and input types escapes the control panel paradigm and correctly takes advantage of the freedom that act lends to better communicate with the user.
Unfortunately, doing so often means eschewing standards and best practices to find a solution that should work better for users. That is extremely difficult to get right; it's not very surprising that designers would purposely decide to use a flawed but known model instead.
We need to get it out of our heads that designers only do things because they're visually satisfying. They are making purposeful and critical decisions to meet a design objective.
They're essentially aestheticians, not designers. Yep, I hate them too.
The big takeaway is that humans shouldn’t be required to think like computers when they are interacting with one.
“Warning : Failed to load library” <ok>
Why did the library failed to load? Why are we being informed? Why does it say “Ok” when it is not ok?
As a lovely example. And on the regular you still have these problems. The daily wtf is full of current day examples.
Many graphic design refugees are interested in learning about usability, user research, heuristics, Fitts' Law, GOMS/KLM, etc. But it takes time and energy. I've found that some small companies with no established UX team start out gravitating to shiny portfolio examples and end up putting visual design (VX) folks in charge of interaction design which sometimes goes poorly e.g. https://medium.com/intercom-inside/the-dribbblisation-of-des...
It is kind of a tautology to say it but this hints at it being a human organizational social issue behind the pathology if it keeps cropping up.
I agree that good design is hard and often involves substantial work beyond just making something on the sheet.
I try to make myself interested in your problems. However, most of them seem really ''otherworldy'' to me or are in a domain I don't care about about. I became a programmer because I like the art of the trade, and I keep myself interested by writing interesting code.
It's not my job as a coder to care about your problem past what has been described to me. It is your job, as product lead, to care about the problem and design a sufficient information architecture, subdomain map, and other documentation to model your problem.
This goes both ways. It is not your job to care about my problems past how it impacts the product. I don't expect you to know or care about how we design the software or how it is implemented; whether we use snake or camelCase, if we use openssl or pgp, if we choose mysql or postgresql.
I just think it's important to outline that much of your comment can be applied in the inverse, and saying that "It's less work to become a mediocre programmer than to intimately understand the problem", is the same as saying "It's less work to become a mediocre product analyst than it is to intimately understand how to code".
I agree with your last point, it's important for every department to have humility and empathy for the problems of their peers. But, it's not a problem to be figured out by designers and coders; it's instead an issue that extends further and requires effort from everyone involved, and at every level, to support.
In particular, it is frequently the case where a product lead says something like “we need a way to...” and leaves the implementation open to discussion to all.
This often leads to a programmer, who is more deeply concerned with how that function will operate on the data, coming up with a “how about a button that...” and/or just implementing the button as a suggestion.
And then the UX architect and product lead, knowing that they don’t want to piss off this coder for fear of future pushback on their ideas, just caves and says “fine”.
That’s the most common scenario I’ve seen at larger companies. That, and having woefully inexperienced UX and UI designers in the first place.
You're spot on re: humility and empathy. I see so much hubris in programmers/computer people thinking they/we can fully understand the world's problems, much less solve them.
For starters, what context have you even experienced this in? From your description it sounds like both you and the programmer (or designer, as the case may be) are just doing this as a side project. Of course there's not going to be the incentive to follow-up and do the hard work of finding actual product-market fit if it was supposed to just be a fun learning experience from the beginning. Did you ever communicate with them what your expectations are, what the end goal is, how much of a time commitment you expect, and was the other person on the same page and equally invested in it? Also, why are you expecting the programmer to both figure out the nuances of the product and implement it? What is even your role in this, what do you bring to the table?
It sounds to me like you did not communicate well or did not set expectations properly about this project, and now you're blaming the other person for doing a bad/lazy job.
Oh man, I actually remapped quit on Firefox because this bit me too many times and Firefox kept removing ways to confirm quit. I will die on the hill that "confirm quit" is the correct behavior.
(I have, in fact, died on this hill enough times I may be outing my HN burner with this comment.)
[1] https://bugzilla.mozilla.org/show_bug.cgi?id=52821 (only took 21 years, jeez)
I can't imagine how often you need to delete everything from a file without using a text editor. I'm sure it's not that common that you need to give it a one-letter flag (instead of --remove or --delete) and put it right next to the fucking edit flag.
I really wonder if there is any script on any distro that uses this behavior. Most distros did go into the sane "let's split the crontab into many files" option.
> I will die on the hill that "confirm quit" is the correct behavior.
Google Chrome has a weird 'solution': on cmd-q press it popups a temporary overlay which reads "Hold cmd-q to quit"
By a decade or two, yes.
Right — Control-W should always¹ be ‘erase word’. Windows/IBM really mucked things up by stealing Control for GUI operations (among other things like conflating Tab with Next-Field because they're the same thing on punch cards).
¹ Okay, ‘end of transmission block’ is also acceptable.
Not all UI elements should have the same visual weight
If you have 3 buttons, probably one of those buttons will be pressed 80% of the time, and the other buttons pressed less than 10% of the time. How do you style those buttons?
As a programmer, we like to think of the three buttons symmetrically. We want all the buttons to look the same and behave the same, because then they're easier to style and easier to reason about. Our instinct is to make a button class and then place all the buttons next to each other in a nice neat table.
To a user, the three buttons are different, and it should be obvious which one is the button you're expected to press most of the time. From the user's perspective, "Submit form" isn't really the same type of UI element as "reset form". The submit button should be big, bold, colorful and obvious. My eyes should naturally settle on it. The reset form button (if it exists) should be small and non-obvious. Its an advanced feature. It should be out of the way and most people should never notice it.
My email compose window has this problem. It has 3 buttons - "Send", "Save Draft" and "Discard". When I'm writing an email, I'm not choosing between 3 equivalent options. I hit send about 80% of the time I type an email. Once I've written an email, if I visually hunt for the obvious button on the screen, my eyes should naturally settle on "Send". Styling should make that obvious. But no - there are 3 buttons I have to choose from. They're all next to each other, and they're all styled in an identical manner. The interface makes users actively hunt for the "Send" action. This is bad UX.
What really gets me is the touch bar on macbooks where I touch it by mistake when typing numbers, sending all my windows dancing around. Drives me up the wall. Leave more space there, Apple! Or just get rid of that stupid touch screen gimmick.
https://store.storeimages.cdn-apple.com/4668/as-images.apple...
I’m quite positive this is a byproduct of the product->design->engineering waterfall, because while we say we are agile, nothing really gets iterated on once an engineer starts coding. If it isn’t caught in the wireframe phase, it’ll take a redesign to address.