The people who designed these things for the Mac and MeXT back in up the ‘80s and ‘90s had a solid background in psychology. They used things like Fitts’ Law to model behaviour and then they conducted user studies looking at failure modes. Microsoft had a room full of cameras that watched people using Windows and Office and they then analysed the footage to see how long things took and what folks struggled with. In contrast, I get the impression that a lot of the newer things are built around someone’s subjective aesthetic judgements, I’ve seen this kind of thread in open source GUI projects (long threads of opinions, no one ever bringing up the idea that this is stuff you can actually measure) but it seems to have permeated the proprietary world as well.
I suspect the beginning of the end was in the early days of OS X when Apple started measuring what animations being shown in stores led to increased sales.
A small part of my job is designing and implementing small GUIs and the worst part is that everyone thinks that they can participate in GUI design sessions and decisions. Every fool who has a pair of eyes in their head thinks that they can participate because photons hitting their retina qualifies them. And yes, the problem is exactly the lack of knowledge about psychology, color spaces, color blindness, disabilities & accessibility, internationalization, Fitts’ law, display technologies, input devices, etc. Meanwhile, you have a super important real-time control loop in the heart of the device and nobody cares about how shoddy the implementation is because it’s not made of colored squares that are one pixel too wide and have the wrong color. And of course the GUI has to be redone every so-and-so many months or years, often with a worse and less practical design. Ultimately, these things are what turns me off GUI programming the most. I’d rather just create beautiful and robust designs that are invisible to the opinionated laypeople.
I’ve found some electron apps like Slack completely inaccessible. They are effectively empty windows on macOS. Same with if you save a site as a chrome app.
Even the most basic things on some sites don’t work. For example, GitHub’s text areas can’t be accessed by dictation commands. The web version of Slack has a really buggy message field.
For example, GitHub’s text areas can’t be accessed by dictation commands.
Could you expand on this? I find Github’s text areas to be the least bad I’ve ever had to deal with (because they’re true <textarea>s, not some JS eldritch horror that pretends to be a textarea). I’m very interested in learning about what dictation commands are, how they’re implemented, how they’re integrated with the browser, how Github breaks them and everything else about them.
I use Voice Control on macOS and iOS. A lot of sites with custom text fields don’t provide a target for the OS to know it’s there, requiring using a grid to make mouse clicks to focus on the field. Text navigation commands fail completely. Dictating new words sometimes works, but you can’t edit anything and it seems when certain JavaScript fires the field stops accepting input completely and nothing gets it working again without reloading the page.
Whatever GitHub is doing for things like pull requests, they aren’t using just a basic textarea, if they were the browser would be automatically providing hooks for voice control.
This site, on the other hand, is fully accessible.
I have been hating what they are doing to scroll bars for many years. Hiding them, making them smaller. How about trying to resize a window - in Windows it seems like the “window border” that used to measure how much space you had to grab onto has become near enough to 1 pixel as to make it criminally hard to use (might depend on application).
I don’t mind these scroll bars on touchscreen platforms. On a phone or tablet, they’re a secondary visual clue about your position in the frame and you have good fine control via direct manipulation.
One of the roots of the problem seems to be the quest for UI metaphors that work everywhere. You wouldn’t think something that works well on a laptop screen that you sit half a metre from and use with a trackpad is a good idea for a TV screen on the other side of room and use via a remote control, but people seem to have no problem taking things from a phone and putting them on a laptop or desktop. This is almost certainly led by marketing folks saying ‘our brand identity needs to be the same everywhere’.
Agreed. I refuse to use a Linux desktop that doesn’t let me resize a window just by grabbing-and-dragging it with alt-right-click. Or moving it with alt-left-click. I have no idea how I would survive using only Windows, which doesn’t have these. On Linux even if a WM doesn’t have these defaults, you can usually configure it to behave this way. Not even having that possibility makes me immediately trash it.
Sorry to be obscure and glad you got it. In some of my circles this acronym is common to the point of cliche, often applied to things like institutional racism.
For me as a valid techie, visible scrollbars fill two very important purposes when I’m viewing something:
They tell me how big the thing is (big document, small scrollbar).
They tell me where I am (scroll bar on the bottom, I’m almost done reading).
Saving a few pixels on the side of my screen is not worth losing such a valuable anchor. If you want to save pixels, save them from the top or bottom. Let me move the task bar to the side. Let me move my tabs to the side for Christ’s sake. We’re not in a 4:3 world any more.
I set my Firefox scroll bar to chonk/4 (as per article) a while ago and I don’t think I’m off the mean in terms of ability to resolve desktop stuff or aim with the mouse etc, but it’s much nicer imho. I will rue the day this feature disappears.
Scrollbars are terrible UI design. You move them in the opposite direction to the way you want the page to move. And how far they need to move depends on the total length. I always looked for how to disable them in the past. It’s good riddance as far as I’m concerned.
You move them in the opposite direction to the way you want the page to move.
I think this depends on perspective. I’ve never thought of it as the direction I want the page to move, I think of it as the direction I want to go in the document. In the same way, when using a mouse scroll wheel, scrolling down moves me down in the document, scrolling up moves me up.
Conversely, the default scroll behavior being opposite in Mac throws me for a loop every time I grab one. But I don’t think this makes for a poor UI any more than having to change the stick layout on an FPS to “inverted” makes the default a poor UI.
To me, the value to a scrollbar is to be able to see where you are in a document and be able to move around it quickly. I fully accept “Reader Mode” will remove things like that, but when I’m heavily editing a document with dozens of pages, I don’t want to have to spend all day with the page down key figuring out where I am and where I need to go.
Anything else notwithstanding, but the notion that people are actually clicking on the tiny arrows like in the 90s feels weird to me.
I don’t exactly know when I got my first mouse with a scroll wheel, but it must have been 20 years by now. First laptop with touch scrolling? 19 years. I’ve not clicked these a lot in those 20 years, so the size simply does not matter one bit.
I know, this doesn’t help the people not using these devices, but I suppose it’s also not that they’ve been designed to click on. So, different reason, same problem.
Trying to help random people use computers is pretty eye opening in a lot of ways. Last night, I helped someone try to place a Wal Mart order after her phone died… thinking the google account password was the same one as the wal mart account, thinking pre-filled forms were requirements instead of suggestions, seeing a bold “Inbox” label as meaning you’re already on the inbox, typing things then clicking on “submit” rather than using tab and enter, not even realizing some things could scroll… the list goes on and on.
Clicking the little button is nice for people like that since at least they already are familiar with the concept of clicking buttons on the screen.
I don’t exactly know when I got my first mouse with a scroll wheel, but it must have been 20 years by now. First laptop with touch scrolling? 19 years. I’ve not clicked these a lot in those 20 years, so the size simply does not matter one bit.
For a long series of boring reasons, I spent a lot of time teaching people who aren’t particularly technically adept to use computers. There’s still a surprisingly large proportion of people who clicks on the arrows instead of using the scrollbar. As in over the pandemic I’ve come to know at least a dozen, out of… maybe fourty or so?
They don’t use it because they never know which way to scroll. They find it confusing that, in order to move the paper in Word upwards, or to go to a higher-numbered page, you have to get the wheel to go down. The arrows make sense because they associate it with “move this way in the document” and there’s no ambiguity about it – the direction they show correlates with the direction of the scroll bar’s motion, which works like a map of sorts.
Touchpads with “natural” scrolling sound like a good idea but they’re not (that* good. I’ve tried to teach it to dozens of people, I don’t think any of them picked it up. First, scrolling gestures done on a tiny surface on a laptop’s case instead of the screen don’t make a lot of sense for some people, and they find it really hard to perform, because they need to look at their fingers and the touchpad, and if they do that, they don’t look at the screen. Second, if you’re just a little clumsy and not very good with technical devices, even simple gestures like two-finger scrolling are just hard to do on the cheap touchpads that make up most of the laptop demographics, rather than the fancy-ass Macbook Pro that designers are using.
Touchpads with “natural” scrolling sound like a good idea but they’re not
I find them frustrating and immediately find the setting to turn them back. Moving from scrollbars to scroll wheels felt natural: I moved the wheel in the direction I would move the scroll bar. Using touchscreens for scrolling also feels natural: I grab the thing and move it. ‘Natural’ scrolling on touchpads is in a weird place because I’m not directly manipulating the thing but I’m expected to move it as if I am. Without a decade of using scrollbars before I got my first wheel mouse, I probably wouldn’t have found those useful either.
Second, if you’re just a little clumsy and not very good with technical devices, even simple gestures like two-finger scrolling are just hard to do on the cheap touchpads that make up most of the laptop demographics, rather than the fancy-ass Macbook Pro that designers are using.
I can definitely relate to this. The Surface Book 2 has one of the best non-Mac trackpads that I’ve used and I still made mistakes. I tried using a cheap Dell laptop and scrolling with the trackpad was totally impossible and I just used keyboard shortcuts or the scrollbars.
Dragging the scrollbar or using a mouse wheel might result in smooth scrolling. While the buttons can scroll exactly one line or other defined distance, which is sometimes more desirable. And the wheel usually scrolls only vertically (trackballs are great for scrolling also horizontally). Besides that, that buttons with small arrows serve as a visual symbol that says what this control element is for.
I understand the efforts to make UI simpler, but these little buttons do not bother me much and I am slightly more inclined to keep them.
It’s also worth noting that NeXT made the buttons a lot more usable by putting them both at the same end of the scrollbar. This meant that, if you scrolled a bit too far, you had a small movement to go the other way, not a large mouse cursor displacement to go to the other end. I can’t remember with OS X whether this was the default or an option that you could set.
Yes! I think, it was default, see screenshots of Rhapsody or 10.5. BTW: Apple/Mac is quite controversial to me – some parts are great, some terrible. IMHO best is to take the good bits as inspiration and port them to GNU/Linux :-)
JFTR, I did in no way suggest or even entertain the idea to do anything with them, especially make them usable in a worse way.
It was just an observation how I don’t usually focus on them, and also the non-technical users I know and I remember watching, are not using them, that’s why I’m wondering.
On the one paw, recompile my GUI theme? - but hey I’m a gentoo user I do that every friday anyway, I could just drop a patch in I guess… RECOMPILE MY GUI THEME??
Oh, I actually did that with Breeze. I did it because, okay, Breeze is well-supported by KDE applications, I’m not sure how much longer QtCurve is going to be a thing, and now that a bunch of KDE applications are using Kirigami, they look weird with themes other than Breeze. It’s not entirely bad but it’s huge, which is a pet peeve of mine – widgets are so big there’s hardly any space left for the thing I’m actually interested in seeing and it just pisses me off.
The fact that you get to recompile it is one of the things that surprised me the least, that used to be a thing with GTK engines back in the day. I’m not sure what surprised me the most. The fact that it has a D-Bus interface is… probably up there at the top.
Anyway, thankfully QtCurve still seems to work so *sigh* at least it feels slightly less cramped than my first monitor. Reliving the cramped UX of Windows 95 in 800x600 resolution is not what I had on my bingo card for 2023.
The people who designed these things for the Mac and MeXT back in up the ‘80s and ‘90s had a solid background in psychology. They used things like Fitts’ Law to model behaviour and then they conducted user studies looking at failure modes. Microsoft had a room full of cameras that watched people using Windows and Office and they then analysed the footage to see how long things took and what folks struggled with. In contrast, I get the impression that a lot of the newer things are built around someone’s subjective aesthetic judgements, I’ve seen this kind of thread in open source GUI projects (long threads of opinions, no one ever bringing up the idea that this is stuff you can actually measure) but it seems to have permeated the proprietary world as well.
I suspect the beginning of the end was in the early days of OS X when Apple started measuring what animations being shown in stores led to increased sales.
A small part of my job is designing and implementing small GUIs and the worst part is that everyone thinks that they can participate in GUI design sessions and decisions. Every fool who has a pair of eyes in their head thinks that they can participate because photons hitting their retina qualifies them. And yes, the problem is exactly the lack of knowledge about psychology, color spaces, color blindness, disabilities & accessibility, internationalization, Fitts’ law, display technologies, input devices, etc. Meanwhile, you have a super important real-time control loop in the heart of the device and nobody cares about how shoddy the implementation is because it’s not made of colored squares that are one pixel too wide and have the wrong color. And of course the GUI has to be redone every so-and-so many months or years, often with a worse and less practical design. Ultimately, these things are what turns me off GUI programming the most. I’d rather just create beautiful and robust designs that are invisible to the opinionated laypeople.
I’ve found some electron apps like Slack completely inaccessible. They are effectively empty windows on macOS. Same with if you save a site as a chrome app.
Even the most basic things on some sites don’t work. For example, GitHub’s text areas can’t be accessed by dictation commands. The web version of Slack has a really buggy message field.
Could you expand on this? I find Github’s text areas to be the least bad I’ve ever had to deal with (because they’re true <textarea>s, not some JS eldritch horror that pretends to be a textarea). I’m very interested in learning about what dictation commands are, how they’re implemented, how they’re integrated with the browser, how Github breaks them and everything else about them.
I use Voice Control on macOS and iOS. A lot of sites with custom text fields don’t provide a target for the OS to know it’s there, requiring using a grid to make mouse clicks to focus on the field. Text navigation commands fail completely. Dictating new words sometimes works, but you can’t edit anything and it seems when certain JavaScript fires the field stops accepting input completely and nothing gets it working again without reloading the page.
Whatever GitHub is doing for things like pull requests, they aren’t using just a basic textarea, if they were the browser would be automatically providing hooks for voice control.
This site, on the other hand, is fully accessible.
I have been hating what they are doing to scroll bars for many years. Hiding them, making them smaller. How about trying to resize a window - in Windows it seems like the “window border” that used to measure how much space you had to grab onto has become near enough to 1 pixel as to make it criminally hard to use (might depend on application).
I don’t mind these scroll bars on touchscreen platforms. On a phone or tablet, they’re a secondary visual clue about your position in the frame and you have good fine control via direct manipulation.
One of the roots of the problem seems to be the quest for UI metaphors that work everywhere. You wouldn’t think something that works well on a laptop screen that you sit half a metre from and use with a trackpad is a good idea for a TV screen on the other side of room and use via a remote control, but people seem to have no problem taking things from a phone and putting them on a laptop or desktop. This is almost certainly led by marketing folks saying ‘our brand identity needs to be the same everywhere’.
Agreed. I refuse to use a Linux desktop that doesn’t let me resize a window just by grabbing-and-dragging it with alt-right-click. Or moving it with alt-left-click. I have no idea how I would survive using only Windows, which doesn’t have these. On Linux even if a WM doesn’t have these defaults, you can usually configure it to behave this way. Not even having that possibility makes me immediately trash it.
I believe I’ve used AltDrag in Windows before to get that behaviour. Or if not that, a similar tool.
and macOS has “Easy Move+Resize” although it’s nowhere near as snappy as an e.g. native KDE behavior. But works.
Contemporary UI trends are obviously set by people who hate desktops/workstations.
POSIWID
They are taking agency from users every way they can. Other examples are infinite scroll and videos without controls, or even runtimes.
For those, like me, who had never seen POSIWID before, I think it in this context it’s supposed to mean:
“The purpose of a system is what it does”
Sorry to be obscure and glad you got it. In some of my circles this acronym is common to the point of cliche, often applied to things like institutional racism.
For me as a valid techie, visible scrollbars fill two very important purposes when I’m viewing something:
Saving a few pixels on the side of my screen is not worth losing such a valuable anchor. If you want to save pixels, save them from the top or bottom. Let me move the task bar to the side. Let me move my tabs to the side for Christ’s sake. We’re not in a 4:3 world any more.
I set my Firefox scroll bar to chonk/4 (as per article) a while ago and I don’t think I’m off the mean in terms of ability to resolve desktop stuff or aim with the mouse etc, but it’s much nicer imho. I will rue the day this feature disappears.
Thanks for calling that out, it is indeed quite pleasant.
Scrollbars are terrible UI design. You move them in the opposite direction to the way you want the page to move. And how far they need to move depends on the total length. I always looked for how to disable them in the past. It’s good riddance as far as I’m concerned.
I think this depends on perspective. I’ve never thought of it as the direction I want the page to move, I think of it as the direction I want to go in the document. In the same way, when using a mouse scroll wheel, scrolling down moves me down in the document, scrolling up moves me up.
Conversely, the default scroll behavior being opposite in Mac throws me for a loop every time I grab one. But I don’t think this makes for a poor UI any more than having to change the stick layout on an FPS to “inverted” makes the default a poor UI.
To me, the value to a scrollbar is to be able to see where you are in a document and be able to move around it quickly. I fully accept “Reader Mode” will remove things like that, but when I’m heavily editing a document with dozens of pages, I don’t want to have to spend all day with the page down key figuring out where I am and where I need to go.
Anything else notwithstanding, but the notion that people are actually clicking on the tiny arrows like in the 90s feels weird to me.
I don’t exactly know when I got my first mouse with a scroll wheel, but it must have been 20 years by now. First laptop with touch scrolling? 19 years. I’ve not clicked these a lot in those 20 years, so the size simply does not matter one bit.
I know, this doesn’t help the people not using these devices, but I suppose it’s also not that they’ve been designed to click on. So, different reason, same problem.
Trying to help random people use computers is pretty eye opening in a lot of ways. Last night, I helped someone try to place a Wal Mart order after her phone died… thinking the google account password was the same one as the wal mart account, thinking pre-filled forms were requirements instead of suggestions, seeing a bold “Inbox” label as meaning you’re already on the inbox, typing things then clicking on “submit” rather than using tab and enter, not even realizing some things could scroll… the list goes on and on.
Clicking the little button is nice for people like that since at least they already are familiar with the concept of clicking buttons on the screen.
For a long series of boring reasons, I spent a lot of time teaching people who aren’t particularly technically adept to use computers. There’s still a surprisingly large proportion of people who clicks on the arrows instead of using the scrollbar. As in over the pandemic I’ve come to know at least a dozen, out of… maybe fourty or so?
They don’t use it because they never know which way to scroll. They find it confusing that, in order to move the paper in Word upwards, or to go to a higher-numbered page, you have to get the wheel to go down. The arrows make sense because they associate it with “move this way in the document” and there’s no ambiguity about it – the direction they show correlates with the direction of the scroll bar’s motion, which works like a map of sorts.
Touchpads with “natural” scrolling sound like a good idea but they’re not (that* good. I’ve tried to teach it to dozens of people, I don’t think any of them picked it up. First, scrolling gestures done on a tiny surface on a laptop’s case instead of the screen don’t make a lot of sense for some people, and they find it really hard to perform, because they need to look at their fingers and the touchpad, and if they do that, they don’t look at the screen. Second, if you’re just a little clumsy and not very good with technical devices, even simple gestures like two-finger scrolling are just hard to do on the cheap touchpads that make up most of the laptop demographics, rather than the fancy-ass Macbook Pro that designers are using.
I find them frustrating and immediately find the setting to turn them back. Moving from scrollbars to scroll wheels felt natural: I moved the wheel in the direction I would move the scroll bar. Using touchscreens for scrolling also feels natural: I grab the thing and move it. ‘Natural’ scrolling on touchpads is in a weird place because I’m not directly manipulating the thing but I’m expected to move it as if I am. Without a decade of using scrollbars before I got my first wheel mouse, I probably wouldn’t have found those useful either.
I can definitely relate to this. The Surface Book 2 has one of the best non-Mac trackpads that I’ve used and I still made mistakes. I tried using a cheap Dell laptop and scrolling with the trackpad was totally impossible and I just used keyboard shortcuts or the scrollbars.
Dragging the scrollbar or using a mouse wheel might result in smooth scrolling. While the buttons can scroll exactly one line or other defined distance, which is sometimes more desirable. And the wheel usually scrolls only vertically (trackballs are great for scrolling also horizontally). Besides that, that buttons with small arrows serve as a visual symbol that says what this control element is for.
I understand the efforts to make UI simpler, but these little buttons do not bother me much and I am slightly more inclined to keep them.
It’s also worth noting that NeXT made the buttons a lot more usable by putting them both at the same end of the scrollbar. This meant that, if you scrolled a bit too far, you had a small movement to go the other way, not a large mouse cursor displacement to go to the other end. I can’t remember with OS X whether this was the default or an option that you could set.
Yes! I think, it was default, see screenshots of Rhapsody or 10.5. BTW: Apple/Mac is quite controversial to me – some parts are great, some terrible. IMHO best is to take the good bits as inspiration and port them to GNU/Linux :-)
I used to like having both: up at the top then both up and down at the bottom. KDE would let you do that.
JFTR, I did in no way suggest or even entertain the idea to do anything with them, especially make them usable in a worse way.
It was just an observation how I don’t usually focus on them, and also the non-technical users I know and I remember watching, are not using them, that’s why I’m wondering.
Oh, I actually did that with Breeze. I did it because, okay, Breeze is well-supported by KDE applications, I’m not sure how much longer QtCurve is going to be a thing, and now that a bunch of KDE applications are using Kirigami, they look weird with themes other than Breeze. It’s not entirely bad but it’s huge, which is a pet peeve of mine – widgets are so big there’s hardly any space left for the thing I’m actually interested in seeing and it just pisses me off.
The fact that you get to recompile it is one of the things that surprised me the least, that used to be a thing with GTK engines back in the day. I’m not sure what surprised me the most. The fact that it has a D-Bus interface is… probably up there at the top.
Anyway, thankfully QtCurve still seems to work so *sigh* at least it feels slightly less cramped than my first monitor. Reliving the cramped UX of Windows 95 in 800x600 resolution is not what I had on my bingo card for 2023.
I’ve found the scrollbar problem is particually bad on mobile. I often can’t even see or use the scrollbars in Chrome and Firefox on Andriod.
it’s not, GTK4 supports it (it’s in the a11y settings now) and webkit uses that setting too