It's the elimination of window borders. Aside from not being able to differentiate one window from another similarly colored window in the background, it's nearly impossible to click and hold on anything along the edge to resize the window.
It's the overloading of the title bar with so much shit like search boxes and extraneous buttons that a user has almost no place to grip to move the window.
It's the way that tabbing between text boxes either doesn't behave the way you'd expect, or doesn't work at all.
It's all the tooltips that interrupt and litter the interface and, at times, block out things that you are looking at. And 95% of the time, the information provided in these tooltips are redundant or useless.
It's amazing how much damage these cargo-cult UI/UX morons have done in the past ten years. They threw out several decades of usability pioneered by real HID experts for something that looks pretty but doesn't fucking work for a lot of people.
Applications like Postman, Teams (and pretty much all of MSFT's applications these days), Chrome, and Insomnia should be case studies on how to not design user interfaces. They are about as bad as desktop software gets.
The biggest sin is that this would be a non-issue if these things were configurable at the windowing system level and could not be overriden by app developers. But the trend has gone in the opposite direction; instead of providing more configurability, Windows and Gnome/GTK are actually taking away options that have existed before.
“Morons” is the right word. I don’t like to gratuitously shit on people doing their jobs. But what the hell? I can’t move a window anymore without clicking around like an idiot to figure out what’s part of the title bar and what’s a button. Every time I start “New Teams,” it asks me if I want to go back to Old Teams. If I open a PDF in New Teams from the file browser, it’s not obvious how to close the PDF without losing your place in the file system. And on top of that, everything is grindingly slow.
A lot of people working at Microsoft/Apple/Google are bad their jobs and should feel bad.
That title bar thing, good god damn. I have to concentrate my vision on the title bar to place my fucking cursor between non-outlined active sections just to grab that motherfucker and shove it over. It is entirely a user experience regression. There's no benefit gained, whatsoever, from moving content to the title bar. Easily moving a window by dragging the title bar was something that "just worked" for at least two decades if not three.
Maybe two-odd decades is about the time it takes for enough people to have forgotten the reasons and decide to just remove Chesterton's Fence[0] because, despite the fact that screen area is at an all time high, we still need to squeeze more shit in around the edges.
I've already whinged about both scroll bars[1].
I'll also whinge about "dead space" rarity on UIs like DevOps and Jira. "No, i want out of all contexts! Did I accidentally switch that slider by clicking like 10cm to the right of it?" (actual example from within DevOps right now).
> There's no benefit gained, whatsoever, from moving content to the title bar
The one benefit I can see is to gain some extra vertical space on the skinny 16:9 screens, especially now that the task bar in windows 11 got thicker and can no longer be moved to a side. They, of course, still haven't gotten around to fixing the auto-hide behavior, so that's still not an option.
The rare times I have to use windows, I use edge and I quite like the vertical tabs. This allows for a usable title bar, too.
You can't actually use reason to understand modern UX/UI designers choices. I've just got Windows 11 at work. They've increased the blank space between all the items in the file exporer quite significantly. They clearly don't care about vertical space because I know have to scroll first in places I could click first before.
Feel free to enlighten all of us, instead of using ad hominem on this particular comment. I, a lowly uninformed user, would love to know what UX/UI is versus what it pretends to be, so that I might understand the decisions being made in Windows 11.
Alt-Space, 'M' is your friendly keyboard method to move a window around.
After pressing those three keys, use the cursor keys to move the window around as you desire, and press the Enter key when you're happy with the window's new location or press the Esc key to cancel the move op.
Works really well when the window's title bar has been moved offscreen for some reason.
Once you've moved the window with the arrow keys you can also use the mouse to move the window. The problem with this is that it's _slow_, and it doesn't actually work on some software. Some non-native window frameworks like to disable (or don't implement) all hotkeys by default, including f10, Alt+space and sometimes even the tab key. IMO the only reason for a program to disable the standard windowing mechanics is because it's intended to be used as an overlay (programs like launchy or slickrun).
You wanna freak out anyone under the age of about 45-50, use the [Ctrl]-[Ins] / [Shift]-[Ins] / [Shift]-[Del] shortcuts while they're watching your keyboard. Everyone nowadays only knows those Apple-copied [Ctrl]-[C]/[V]/[X] ones, and thinks they're the God-given Standard.
FWIW, and I know it's not as good as patching the root cause, on Windows environment I have found relief in using AltDrag (and/or its fork AltSnap [1]). You can grab and move a window by pressing Alt and then clicking wherever you want on the window. There are also a few other tricks regarding window maximizing/resizing etc.
[1] https://github.com/RamonUnch/AltSnap
Ironic, to some degree. I agree with the overall sentiment,but:
> There's no benefit gained, whatsoever, from moving content to the title bar.
But also:
> Chesterton's Fence
>>The more modern type of reformer goes gaily up to it and says, 'I don't see the use of this; let us clear it away.' To which the more intelligent type of reformer will do well to answer: 'If you don't see the use of it, I certainly won't let you clear it away. Go away and think. Then, when you can come back and tell me that you do see the use of it, I may allow you to destroy it.'
There are always pros and cons to every design decision. As technology develops, we create new tools to interact with an ever-expanding content base. By moving things to the title bar, you gain back some real estate that you can dedicate to this ever expanding set of other tools and content.
There are other design paradigms that can be used for this, but it's a fairly simple (and naive) implementation / solution to this problem to move some of the stuff to the title bar which (visually) appears to be wasted space. But visual space != interactive space, so
> By moving things to the title bar, you gain back some real estate that you can dedicate to this ever expanding set of other tools and content.
But the loss of the title bar causes actual pain and interferes with practical use. It's hard to imagine any tools/content that would be worth that loss.
The only programs I have that are putting things in the titlebar are putting tabs there, and those programs do not cause me notable problems. For me the positives are much stronger than the negatives.
Visual Studio actually puts the menubar in the titlebar; I never have any place to click left unless the window is maximized (or, at least, full width).
I'm ok with adding more functions to the titlebar, it's something I experimented with myself for a desktop app some 15 years ago. I found there was a heck of a lot of special behaviours tied to the titlebar and that overriding them meant a lot of work to fix edge cases like - being able to move windows. So I never followed through with it.
Maybe the addition of dedicated Move button next to the minimise, maxmimise and close button would be a reasonable compromise?
That would provide a consistent target to click on with your mouse. Obviously a button is a lot smaller than the entire titlebar for clicking on, so there would be some efficiency loss for people that regularly move windows (I am one of them). Getting move use out of the titlebar space would be worth the minor inconvenience of a more accurate click-drag to move operation.
If you need to use the move button, the move button needs to stay on screen. I may be the only one, but I regularly position windows partially off screen when I only need one side of the window and I don't have enough desktop space for the whole thing.
Oh yeah, that's a little bit annoying. It looks like I can drag on the second toolbar on the very left but it would be much better to force some dragging space both left and right on the title bar. Otherwise I think menus there is okay, I don't need the entire title bar as a dragging area.
Browser tabs I actually find OK, I didn't even really consider them as "using" title bar space, since each tab is a title bar and contains no "active" areas that prevent the simple 'click and drag' action.
This does become more problematic when there are so many tabs they become tiny and the whole title bar of the browser is taken up by tabs so that there's no dead-space to drag the entire browser window to another screen. I try to keep minimal contexts (due to personal brain capacity issues), so I don't run up against this very often - I manage links that I want to keep and go back to using other means, or if I forget them then they weren't very important in the first place.
That's actually a case against modern overfilled title bars.
"Real" title bars can be removed if not needed (e.g. by switching to another windowing manager like dwm on Linux with X11; not sure how it's done in Wayland or in Windows but there must be a way).
"Crowded" title-bar from Gnome can't be easily removed, because they are not managed by the window manager.
> But title bars are mandatory, so that's why not.
They're mandatory because they're the way to move the window with your mouse. So fucking them up with a bunch of stuff that hinders that is just as bad as not having them in the first place.
Congratulations, you've circumvented the mandatory requirement!
Which is just as bad as disregarding it in the first place.
That's a screenshot across the entire width (of the top) of my monitor, as you can see from NP++ being maximised behind MS Word. Pretty darn busy, isn't it?
Sure, I can clean it up quite a bit, I can find out how to customise that. But naive users? They'll probably freak out. (And, TBF, I just noticed you can grab on to at least some of those busy-looking fields and drag it. But I only noticed that just now, while futzing around to screenshot it. And again, would my Mom? Or yours?) No, on the whole: What large software providers do to the Windows ittle bar nowadays is shit. The larger, the shittier. So of course Microsoft itself is worst of them all.
> But title bars are mandatory, so that's why not.
What does that even mean? That's the whole point of the discussion - standard UI elements have been bastardized so much, might as well just keep the trend going.
> But to ask "why even have it" is to fall into a discussion more about backwards compatibility than design.
No, it is a question about design. Window environments like Windows 11, macOS, Gnome or lots of other Linux window managers are designed for desktop PCs, not tablets. I would include Laptops under desktop PCs for this discussion, since most people -- that are working the whole day on their laptop -- add one or more screens to it. And on those large modern displays, it is more effective than ever to use overlapping non-fullscreened windows. Then it is absolutely necessary to have good standardized UI components to manage those windows. But those customized title bars break that standardization and customization. On modern Gnome I often have to search for the window that is in focus. It used to be clearly visible with a significant color difference in the titlebar. Today, there is slight difference in gray shading. That is ridiculous. And half of modern programs redefine the colors of their title bars, so you have to know for earch application which shade of gray stands for focus and which stands for not in focus. This is fucking stupid.
Just use a tiling window manager. In mine you move windows around by holding middle mouse button anywhere on the window, or left button with some key combo.
Oh wow I thought the New Teams menu was an internal dog food thing for Microsoft employees. I’m so sorry, it’s absolutely maddening. Nobody likes Teams at Microsoft either, and many hours a week are wasted complaining about it.
I used to use Teams, and it had this feature where if you clicked on the app icon (macOs) enough times a debug menu appeared. It was clearly aimed for internal testing. But there was no timer attached to it, so if, over the course of your day, you clicked on it in the dock, eventually the menu appeared. It seemed pretty half baked to have a debug menu available to your users in the wild.
So what do employees say about it? How could it be fixed? And since you're not in that dept I'll ask something else too, why are Microsoft employees not using MS made tech? Like it seems almost nobody is using WinUI3, or MAUI internally but it's all React and Electron?
> A lot of people working at Microsoft/Apple/Google are bad their jobs and should feel bad
That's the thing, I don't think they feel bad at all. Their UX departments exist solely to justify their own existence and are increasingly distant from the opinions and issues facing actual users using their products. They're still well paid, so they have no reason to feel bad.
I agree. The distro I am using lets me move windows by holding down Meta/Super key and left mouse button - anywhere in the window. Firefox is an example of how terrible it can get, see this screenshot: https://i.imgur.com/zs2wekv.png. To move this window I can either click to the left of the Firefox icon (to the left in image) or between the arrow down and minimize to the right.
Yes? We managed on 1024x768 screens and all that “wasted” real estate. Meanwhile I have two 2880x2560 monitors hanging off my dock and UX designers are tripping over themselves to hide more shit in the overflow menu to save me 5 pixels of vertical space.
Going back a long way, this is what always bothered me about UI efforts in the Linux world like the Enlightenment window manager. It seemed like the very definition of “nice house, nobody home.”
The problem boils down to allowing aesthetics to dominate over functionality - form supersedes function. There are plenty of people happy to inflict that on others. I sometimes wonder what they use themselves.
> I can’t move a window anymore without clicking around like an idiot to figure out what’s part of the title bar and what’s a button.
I don't know about MacOS and Windows nowadays but most linux/BSD window managers + Haiku allows one to move or resize a window using a keyboard shortcut so that you can do it regardless of where you put the cursor.
It is a must have to know because there is inevitably moments where you have a window showing at the wrong place. I remember back in the days I was using windows sometimes when switching from multiple screens to single screen a window would start/appear out of the frame so you'd have to move it using one of those shortcuts.
I would be surprised Microsoft and Apple have removed these. Someone to confirm?
> I don't know about MacOS and Windows nowadays but most linux/BSD window managers + Haiku allows one to move or resize a window using a keyboard shortcut so that you can do it regardless of where you put the cursor.
Except this is an X11 feature, and they're all frantically switching to Wayland... Which probably doesn't have that feature, right?
>Every time I start “New Teams,” it asks me if I want to go back to Old Teams
Are you sure you’re not accidentally opening the old Teams application? When you update to new Teams it leaves the old version in place so you have both, and if you open Teams with the new shortcut/application you won’t get prompted to switch back.
I think this was a bug that lasted a couple weeks or so because I had the same exact issue: Open New Teams, get prompted to switch back each time, including the first run (!?)
I am coming to the opinion that Windows XP was the maximum usable interface, and we should have stopped there [0]. But then, I might just be old and telling the kids to get off my lawn.
[0] this is definitely true for Windows. All subsequent versions of Windows have had worse UX.
Everything old is new again? I remember when people were deriding Windows XP for being "Fisher-Price" in appearance and preferring to run Windows 2000 because it was more stable (until roughly SP2 came out, which basically was an entirely new OS released as a service pack).
Windows 7, which was Vista's equivalent to SP2 also has some fans. Heck, I'd even personally argue that Windows ME has some niceties not seen in subsequent releases, particularly when combined with Active Desktop (which I'm pretty sure did not survive the transition to XP).
I wonder if/when some subset of users will prefer a more controlled window layout engine, one which always opens windows into their own "space" on the screen. To that end, I don't actually mind how iPad does its window layout, I just wish iPad had all the features and openness of macOS to install your own drivers, background apps and virtual devices/mixing. A simpler UI can actually be a good thing, but you shouldn't lose functionality when making things simpler. A simpler UI doesn't necessarily mean less complex operating system...
> For me personally, Windows XP means XP SP2. Anything before that wasn’t worth using.
That is because Microsoft had its own versioning system:
- RC was alpha quality.
- Releases were betas.
and then starting with SP2 one can talk about a released product.
NT had 6 SPs, 2000 had 4, XP had 2. Even in Windows 11 they fixed things (taskbar) after release.
This really tells a not so nice story about engineering and quality at Microsoft.
The only thing I remember is that I would never install SP2 because of EULA. I don't remember the specifics though. Around that time I started mostly using Linux anyway.
> Everything old is new again? I remember when people were deriding Windows XP for being "Fisher-Price" in appearance and preferring to run Windows 2000 because it was more stable
That’s more of a truth universally acknowledged. XP was goofy then, it’s goofy now, and it was goofy at every instant between.
Yes but in hindsight endearingly so. And the UX itself was still really good. The start menu in particular was an actual improvement over the old one, as were the clickable login screen icons.
I wouldn't say orthogonal. Making something inviting, approachable and feel like a safe place to experiment/explore for a novice is a critical and overlooked part of usability. The childlike choices of "bright primary colours" and "no sharp edges" is one way to attempt that.
It's what the general public needs that us nerds can least relate to.
It might have been good for brand-new users and children, but not all computer users are like that. It would have been much better if they had had selectable themes.
> It might have been good for brand-new users and children, but not all computer users are like that.
Well, are you sure which of those groups you belong to? At least as far as Windows is concerned? Because:
> It would have been much better if they had had selectable themes.
It did. One click in the right place in the Control Panel and it looked like W95/NT4/W2K. (So did Windows 7, and therefore I must assume Vista, too. [Skipped that one myself.])
But, hey, brand-new users and children couldn't be expected to find that, right?
>I wonder if/when some subset of users will prefer a more controlled window layout engine, one which always opens windows into their own "space" on the screen.
Ya know one of the first things I do on an XP install is to go in and disable as much of UI faffing about as I can. It takes a couple of clicks, but once you disable themes in XP you get something that looks quite a lot like Windows 2000.
No offense, but I cannot possibly trust your judgement when you use a DE that, by default, is the classic example of no window borders for the mouse to grab onto.
I do not remember how I have configured it, perhaps the window border width is determined by the window theme (I use Chicago95), but I have also been using XFCE for many years and I have decent-sized window borders on my 4k monitors and my mouse theme (Hackneyed 48x48) has extremely obvious mouse cursor shape changes whenever you touch the window borders, so they are easy to grab.
The default theme that XFCE ships with has a 1px grabbing area set. So it's not that the cursor doesn't change, it's just you have to be very precise.
There are "fixes" where people hack the theme to increase the size of some transparent images to force a larger grabble area or telling people that they are holding it wrong.
> A good interface would allow enough configuration that we both could have what we want.
Like Windows, from at least 3.1 (ca 1987?) to at least mid-life of Windows 7 (ca 2012-14?). The item under discussion, window border width, used to be settable from 0 up to... Pretty much all of your screen. (But from some update somewhere halfway through W7, it defaulted to at least one pixel however hard you tried to set it to zero.)
I don't "agree to disagree", they offer a function that's unusable.
It would be the equivalent of MacOS breaking the ability to double click on an app.
But it's okay because you could use Spotlight.
That's great and all, but if everything else allows double clicking and MacOS allowed double clicking but only on the absolute centre pixel, we would all consider it broken.
Yep I got into the habit of doing that because I run XP in a virtual machine for legacy junk. I actually had opportunity to run classic MacOS recently and realize that I largely prefer the 10.x UI especially once they started toning down the decorations. Classic MacOS got a lot of things right that are missing in current operating systems, but man is it ugly.
> I remember when people were deriding Windows XP for being "Fisher-Price" in appearance and preferring to run Windows 2000 because it was more stable
Yeah, most people apparently never noticed the Control Panel setting in XP where you could with one click activate the W95/NT4/W2K interface in stead of the "Fisher-Price" one. Worked in Vista and Windows 7 too, but there they started to remove ever more of the detailed UI customisation settings from the Control Panel. (For a while, you could change them by editing the Registry, but towards the end of W7's lifetime, at least some of those didn't take effect any more.)
Windows 2000 was the last truly great Windows for me. It was the last Windows version that had a working "Find in File Contents" feature without having to install grepwin.
This is something that's always puzzled me - why is searching in Windows always so gosh darn slow? You'd think with an entire background service dedicated to file indexing, I'd be able to quickly find files in a directory tree with the word "foo" in their file name or contents, but even with a modern SSD it still takes at least a minute on any remotely populated directory.
File search in graphical Linux is terrible as well.
Just the other day i was trying to do sone simple file search in Mint and oh god the GUI search function is completely useless. I had to jump around installing 3 "file search" programs, none of them really worked and at the end I settled with some console bash magic. ... its 2023 people.
Microsoft should just buy Everything off Voidtools because they have perfected Windows search and it boggles my mind that Microsoft can't recreate that.
I was a fan of Google Desktop Search - that was actually able to find relevant content in my files, and it was fast. But Google killed it as soon as they decided to drive everyone into the cloud. Supposedly all major OSes have such an (indexed) search functionality integrated nowadays, but for some reason it doesn't really work, and I'm never motivated enough to find out how to fix it...
May be they are so busy figuring out how to upload our data to their servers and sneak in clauses in their terms and conditions to cover it ... that they forgot to build the actual functionality of showing search results to user.
Same with Google Talk (as in, was good, then killed by Google). This is their text chat and voice chat app before crapware like Hangouts. Not tried Duo or Meet.
Recursive agree to all the above, purely in terms of UI/UX.
Firefox currently has about 20x20 pixels next to the tabs to grab and move the window, and "always show scrollbars" is off by default. This would be reasonable on a phone, but not on any kind of machine with a mouse.
Can you not right click and enable menu bar to get the traditional menu (plus a title bar of sorts) back for this?
I'm not really defending firefox UI to strongly here, I use it every day as my browser of choice and I find the defaults wonky but at least customizable.
> Can you not right click and enable menu bar to get the traditional menu
FF has a great compromise setting, keeps the menu bar hidden by default but it appears when you press (the CUA standard keys) [Alt] or [F10]. (At least on Windows, that is.)
Especially with the Dashboard shell [1]. Was very convenient: compact, always at your fingertips, with extended virtual desktop, fully configurable and stable: you lay out things some way, they stay this way.
It is interesting that modern UIs keep removing agency from the user. There is very little space in the system that feels mine because I am the sole owner of it. On Windows 3.1 I managed everything in the program groups. It was not too convenient (compared to that Dashboard shell or Mac), but still, if the system added a group now and then, I could always take it apart and rearrange to my liking.
Now the Start menu is completely taken over. The ability to manually arrange things in a folder is taken away. The folders that should be rightfully mine (such as Documents) are constantly invaded by every program that feels entitled to put things there. As a result “we do not go there anymore.” The only places that are left for me to manage are: the area to the right of the start menu and the desktop. But I do not like the start menu and do not like cluttering the desktop, so the only place I still govern is a tiny custom toolbar I added to the taskbar.
The Windows 10 Start Menu scrollbar is unclickably skinny.
It is supposed to expand when you stop your workflow to undertake the Hover-Your-Pointer-Tip-Over-The-Miniscule-Scrollbar Ordeal. If you win, the scrollbar will typically expand to something less difficult. But expansion doesn't happen so reliably that it has your full trust.
> Fair, but after win3.1 there wasn't a menu up in that corner anymore.
WTF are you on about?!? There still is!
Source: Just clicked on the top-left icons on a few programs in Windows 10. Not guideline-defying shit like Microsoft Office / Outlook / Edge / Teams, of course, but programs that follow the Windows developer guidelines display the standard menu just like they've always done: Restore, Move, Size, Minimize, Maximize, Close (usually also displaying the shortcut key, [Alt-F4]). Notepad++ does, WinSCP does, 7-Zip does, PuTTY does, SAS client software does, Teradata client software does... Seems most built-in Windows utilities do, too: File Explorer does, Task Manager does, Command Prompt does, Notepad does, Paint does, Snipping Tool does...
[Edit:] And even in Word / Excel / Edge / Outlook / Teams, at least that menu appears up in the top-left corner when you press the age-old shortcut [Alt]-[Space], even though there is no icon for it. (That was what the W3.x icon was supposed to depict, BTW: The space bar, not a hyphen or minus sign.) [/Edit]
Seems you have to go out of your way to build a Windows app that doesn't have a system menu up in the top-left corner. (Start building a Windows application in most IDEs I've used, and unless you change the default settings it sure gets one automatically.) So why do you spout such drivel, when you obviously have no idea what you're talking about?
Still quite a bit of time between "after win3.1" and "over 5 years". Like, almost thirty years now where that exact icon has, OK, "been absent" -- but replaced with other icons in the exact same spot. Did you never, even by mistake, click on one of those in all those over twenty years? Wow, you must be a lot better than I at mouse control. And / or a lot less curious.
I won't argue against Windows 2000's usability. But I will say that to me, as someone who first started using computers in the XP era, any Windows earlier than Windows 7 looks so god damn ugly to me, and Windows 2000 even more so than XP.
Interestingly though, I don't have the same impression when looking at the old Apple OSes. To me, Mac OS 9 looks old, but not ugly.
To me, XP and Vista looked like a terrible attempt at replicating Mac OS X when that all wow'ed us - it's hard to imagine these days just how big of a deal the iMac and later on OS X were in terms of moving away from grey boxes everywhere (though sometimes too far, since Windows 7 really toned down and refined XP/Vista's UI).
For me, I grew up on Amiga Workbench 2.0 and MS-DOS, so Windows NT/2000 felt like a good evolution from it's User Interface. Very functional, for sure.
100% agree on Classic MacOS. As outdated as the tech was, it was beautiful because it was elegant without being distracting.
To the contrary, XP and Vista' soft, round buttons and edges always made me think of a Douglas-Adamsian dystopia. Win2000 was bare bones, square and functional.
I swear they actually are actively trying to make the "Control Panel" / system settings experience worse with each iteration for sure. 90% of 10 and 11's system settings feel like they just took the original Control Panel screen and hid it, exposed 2-3 of the least valuable options, and displayed them in a weird solid-colored screen. To me the only way to adjust most windows settings is to find the "Advanced settings" link in the solid-colored window that will bring up the original control panels.
I absolutely agree. For at least 10 years now, my dream has been to essentially recreate the Windows 2000/XP UX as a Linux desktop environment.
That said, doing so would basically require creating everything from the ground up, because merely theming some existing toolkit and window manager to look like XP is not at all the point. And frankly, there HAVE been SOME useful UI/UX improvements in the past couple decades, like, FTA, minimaps. But can you imagine a much more conservative desktop, with highly integrated apps, based on the "good old days" of desktop design with just a touch of modern elements when they improve on things?
... Truth told, the closest thing in existence is probably SerenityOS... but it really would be more practical if there was a Linux desktop for this. Every so often I fantasize about how I might start such a project. It's been like that forever now.
I believe the biggest obstacle is dealing with different toolkits, since gtk3 and later are constantly breaking theme compatibility with new versions, and nobody have the time to maintain good alternative themes in these conditions.
XFCE with Chicago95 already exists (and so does Q4OS with XPQ4) as proof of concept that the window manager and toolbar side can be replicated.
The problem is that important pieces of software are not relying on any toolkit directly, but implementing their own version, so even if you support qt, gtk2 and gtk3+ applications with 3 different themes, there will be software running under electron or implementing their own client side decorations and breaking all the rules.
I think the easiest way to start would be to simply collect and maintain existing themes with a unified look and make a distro based on Ubuntu LTS or even RHEL to avoid breaking changes at maximum. This distro would have pre-selected software that behaves well with the selected themes, and need to have at absolute minimum a working browser with extensions following that theme.
If implementing from the bottom up, then something similar to helloSystem could also be a way, but I do not see it working in practice due to the time it would take.
You can find well maintained applications for almost everything in either qt or gtk3. So you can select one toolkit to support. KDE might even accept your theme if you did enough work on it to make it work - and this is something that is feasible for one person. The hard part is those few exceptions, if you need one GTK3 app because KDE doesn't have a good alternative (or one qt app because GTK3 doesn't) then you have twice as much work and it is probably beyond what one person can do. Making that missing app for qt/gtk3 probably needs a large team and a few years and so is beyond you (unless you are rich and can afford to hire someone to fulfill your vision) - and there are many missing applications, not to mention existing ones which are just barely acceptable
It's does not seem well known but it looks and works very closely to old Windows.
I have a side project plan to take this, combine it with a decent laptop, pre install libreoffice and a browser and sell it as "Stable OS" with a long term "no-change" policy to anyone who just needs a basic computer where not everything changes at the whim of someone looking for a item on their resume.
I did like XFCE at one point, but it never really did quite nail the UX for me. A lot of the problem is more on GTK+2 than it was anything else. (Obligatory mention of no thumbnails in file dialog; I used to patch GTK by hand for this!) Now with newer versions of GTK, this problem has unfortunately gotten worse for me, and I don't really prefer XFCE very much. Worseyet, I've gotten used to wlroots giving me good support for things that never really worked well in X11, and XFCE's plan regarding X11 and Wayland going into the future is still somewhat unclear.
I also tried LXDE and LXQt as well. I actually thought LXQt might be the one for me, but it was a bit buggy and lacked a lot of much-needed maintenance.
Also, although I'm glad Linux developers don't always make a habit of just simply throwing out ideas and code that works fine, it does bother me how antiquated some things have gotten. DBus is the best desktop IPC option on Linux, and that's not really a great thing in my opinion. Likewise, it sucks that most file archivers still work by popening some command and parsing its output, occasionally breaking on edge cases, usually having limited support for the kinds of advanced functionality you'd get in Windows archive tools. I do understand that making actually-good applications takes time, but over the course of the past couple decades it doesn't feel like much progress has been made here. I mean hell, speaking of archive tools... FreeBSD has been working on libarchive for ages and it'll likely be integrated with fucking Windows before it's integrated with common Linux desktops, which will probably mostly farm out to CLI tools like unar and zip directly. This is the same approach, with the same exact problems, that I remember being used as far back as when I started using Linux, in programs like Xfe.
Of all the versions of the Windows UI, Win 7 was -- for me -- the best by far. Nothing else even comes close, and everything from Win 8 on has been substantially worse.
I stopped using windows after Using Win10 for a short time, but ClassicShell was always maintained and worked flawlessly... oh, dev stopped in 2017. oh well.
Win95 default was peak Windows UI for me. It had proper shading so you can see the window edges. Min/max/close buttons were high contrast and easy to see, and were big enough to easily click. Window titles were high contrast and clear. The UI was light and snappy unlike XP’s.
Nothing. Those are just touch ups of the Win95 UI. Smoother color gradients from using more colors and smoother animation.
XP was the next big change. Never liked it. Felt like I was dragging around bitmaps. It’s “heavily” for no reason - no usability improvement nor does it even look that good.
Edit: Oh right I forgot about Active Desktop and MS’s attempt to turn Windows Explorer into Internet Explorer. Not a fan of it but at least it didn’t impair usability that much. Visually it still use the same “language”.
Active Desktop and IE Windows Explorer, for one thing.
But also, generally, the difference is that Windows 95 still feels like a "workstation" computer, designed for maximum readability and clarity, while staying in the background and not drawing any attention to itself.
With Windows 98, the OS begins trying to look pretty, which means having more elements and style choices which take attention away from the work.
If I remember correctly turning off active desktop on 98 was almost a requirement if you wanted a ... I was going to say stable system but lets be real this was 98 the best you could hope for was a sort of stable system.
It can be turned off, but it can't be turned off. It's still there, lurking just around the corner, and still showing up in some dialogs.
And I think this is exactly what many people's complaints are in this thread: an interface that is faulty out of the box that needs to be "tuned" in order to change it from an entertainer back into a workstation.
I haven't played with Windows 98 and Windows 2000 in a while, but I remember all sorts of "improvements" over Windows 95, such as sliding menus, "shiny" icons that draw attention to themselves, "shiny" gradient title bars, and so on.
The article complained about things which were not configurable or not easily configurable. So did many comments.
Judicious animation has well studied benefits. I do not see how Windows 98's icons were more attention grabbing. Gradient title bars help locate the buttons in my experience.
I do not use Windows 98 as my daily driver, and I can't give you precise examples.
But I know that many areas of the OS, such as parts of the Control Panel, like the Display control panel require, load into memory, and include IE modules.
> I do not see how Windows 98's icons were more attention grabbing. Gradient title bars help locate the buttons in my experience.
All this crap to the left in Explorer windows that takes up space for no good reason. Active Desktop. Uglier icons that are inconsistent with the many older Win 95-style icons that remain everywhere in the OS. It's not that bad on the whole, and you can reverse most of it, but it's a bloating and bastardization of the (unironically) finely honed aesthetic of Win95.
The left area was used for information like file size, disk usage, and image preview. There have been better implementations. But Windows 98 could suit either preference.
Windows 98's icons look very similar to Windows 95's. Windows 2000's look better and more consistent to me.
I might be getting old, too. (Born 1987.) I have also come to regard Windows XP as the pinnacle of personal computing UI. Almost everything that followed, regardless of vendor, feels like a step in the wrong direction.
Same, and same. It was pretty optimal too, where i remember getting the number of BG tasks down to 15 and I knew what each every one did. I haven't been able to make that claim for an OS for many many years now.
From Windows 95 to Windows 7, the UI fundamentally stayed the same, with a few details here and there, including, scroll bars, window title bars and borders.
Essentially, what changed is just cosmetic and mostly as a result of better hardware. Earlier versions of Windows were designed for 256 color displays, Windows XP fully embraced 24-bit color, and Windows 7 was designed with GPUs in mind. That's how we got from pixel art to shading to semi-transparency.
And I must admit I liked Windows 7 because it was both functional and pretty. A nice, modern for the time skin on top of tried and tested UI core concepts.
Windows 8 broke everything as an attempt to unify the desktop and mobile experience, and we haven't recovered since then. In fact, I think mobile killed desktop usability. We are in a conundrum that we didn't find a way to resolve: desktop and mobile are fundamentally different platforms, so they would need different UI paradigms, but there is an overlap functionality as many apps are present on both platforms, so it would be nice to provide the same experience on both.
To that, add the fact that the web is often third or a fourth option (mobile and desktop web), and the ability to use native controls in web browsers is rather poor, so people make their own, but then, you want the website to look like the desktop app. Same kind of problem.
I'd love the windows 2000 style classic theme with the overall customizability improvements in XP with the UI feature sets of newer versions.
E.g. booting up 2000 by itself it absolutely sucks. I mean, it looks great... but you can't tile windows, you can't resize the command prompt dynamically, you don't have DPI support, a lot of the little victories like pinning and grouping aren't there, the way it draws often causes ghost trails, virtual desktops aren't a thing, a surprising number of things were modal when they didn't need to be. It has great parts but the problem isn't newer versions threw them all away with no improvement it's that not every change was an improvement.
Same time frame: GNUStep/Windowmaker on the Linux side of things, also something of a high water mark. It's really weird to me how much interfaces have degraded in the past two decades.
Mac is all secret handshakes, nothing discoverable, terrible af keyboard-primary work. Windows was very good at letting you stay on the keyboard.
In the spirit of scroll bar discussion, the standard of Ctrl-Home, Crtl-End, Shift-Ctrl-Home and Shift-Ctrl-End are so good. No such "begin of doc" and "end of doc" on a Mac except if an app wills it.
> It's amazing how much damage these cargo-cult UI/UX morons have done in the past ten years. They threw out several decades of usability pioneered by real HID experts for something that looks pretty but doesn't fucking work for a lot of people.
No need to throw insults. I love HN because it's one of the few places where civil debate trumps the hateful tone of all other platforms.
I fundamentally agree with usability being more important than aesthetics. But I don't know what you mean by UX/UI cargo-cult morons.
I'm a UX/UI leader with 20+ years of experience. To me, the main culprits of crimes against usability are business leaders and marketers, not designers (although these do bear some of the blame).
Yes, there are designers who think form is more important than function and push for small scrollbars. However, when you explain the issues they often back down and create usable designs.
I wish the same was true for C-level leaders, marketing leaders and managers. In UI terms they're both ignorant and opinionated. A dangerous combination. Their demands are typically "I like the scrollbar of this website", " the design doesn't look modern" and similar. The amount of fighting that takes to push for usability and accessibility is excruciating.
We need some roles and ranks to act more professionally, and trust the experts. And yes, we also need some designers to think usability first.
> In UI terms they're both ignorant and opinionated.
These are the UX/UI cargo-cult morons you're looking for.
It seems in the software world we've yet to establish our own version of industrial design[1] so "regular" designers gets used. And leaders who decide aren't technical so they don't get usability.
I think the lack of competition also make the terrible designs seem successful.
I'm not using Teams because it looks great. I'm using it because my org uses it for chat and meetings. It's not like they let people try a few different looks for Teams and then see which people like the most.
> To me, the main culprits of crimes against usability are business leaders and marketers, not designers (although these do bear some of the blame). [...] However, when you explain the issues they often back down and create usable designs.
Honestly, my experience is the opposite. It is almost always designers who push for what they consider aesthetically pleasing with very little regard to usability.
Yes, when you push back a lot, you sometimes win. But it takes too much fighting, you need to be at powerful position to have the chance to win and they simply do not seem to care about usability.
It's more, there's no attempt to structure information anymore.
All the settings menus are a nightmare of lists upon lists and each app has a completely random layout for it's menu-ing, often with links to external sites, for what could be set locally.
It is geniunely confusing to try and understand the designers intent. I can never tell if I missed a setting, or if it was never meant to exist in this menu at all.
The amount of times I've had to Google how to change basic settings is way too high, even for Games where the experience is half the ticket value.
CS2 has a settings menu with text buttons at the top (like a tab) that drag you down to an arbitrary point in a long list of settings. There is no tabs, it's impossible to mentally separate what the text button represents mentally, from every other option, because they all point to the same page anyway. Why bother with the buttons?
The semantic meaning of concepts don't relate to each other and are not structured accordingly. It is waaay too unnecessarily difficult to navigate around computing.
I miss Macromedia Flash's UI. I didn't use the program much, but it was so simple and easy to use.
One my biggest drives for privacy and self made solutions, is to get away from the experience and knowledge pollution we've been seeing lately.
This was one of the biggest powers of Mac OS X, before Electron became big.
Not many companies and people developed for Mac OS X, but the ones that did pretty strongly adhered to Apple’s design guidelines, partially because it just made dev live easier and partially because your application would stick out like a sore thumb if you didn’t.
At least the global menu still lives on macOS. I’m sad Gnome 3 didn’t pilfer it from Unity.
>your application would stick out like a sore thumb if you didn’t.
Anybody else remember the UI from the Photoshop plugins from Kai[0]? You clicked the plugin, and then you were looking at some alien organic texture full screen with a few adjustment sliders. That was someone that said, "I see your UI suggestions, but I have other ideas"
This is the most intensely 90s software aesthetic I have ever seen.
Edit: Holy shit, Bryce! I think I had that on my family PC. We were really into Myst and Riven at the time and I guess someone thought it would be fun to try to make our own 3D worlds. It turns out that making digital art actually requires art skills and I don't think I ever created anything with it.
>It turns out that making digital art actually requires art skills
it used to. now, you're only skills in art need to be to properly describe the scene and let some generative model do the work for you. i hope we continue to have distinctions made from digital art created by a generative model and not actually give the artist title to the people driving the prompts. to me, that's no different than me telling Bob Ross that I want a painting with a happy little bush in the left corner, a waterfall in the right, with a nice cabin in the woods in between them and then saying I was the artist because I was able to describe the scene.
Alas, Mac is one of the biggest offenders of crappy design and anti-usability.
When I first had to use one for work I thought it was buggy and defective and losing data. Turns out it was just hiding it because scrollbars were entirely hidden and disabled by default. They also hid a lot of other stuff. What kind of idiot created that design guideline?
Then there's the fact that the editing keys don't work at all consistently or correctly (and sometimes just don't work at all).
And nowadays they like to hide functionality behind obscure multi-finger touchpad gestures.
Their 'magic' mouse was a magic louse, that didn't work very well, ate batteries like they thought they were in a hotdog eating contest, and just randomly did stuff you didn't want based on where your finger happened to touch them.
I hope someday they hire a designer that has used a computer before and can upgrade their UI/UX standards to at least the Windows 95 level.
Old MacOS was good, until they started gradually making it phonelike. The original Macintosh Human Interface Guidelines¹ was solid work. (The big flaw, as usual, forced on the professionals from above: the form-over-function one-button mouse, requiring the undiscoverable double-click — it's almost forgotten now, but originally people had to be trained to double click ‘correctly’.)
> Turns out it was just hiding it because scrollbars were entirely hidden and disabled by default
They are if you use the touchpad, and appear whenever you scroll. The logic is that as you are scrolling with the touchpad, you will not click on the scrollbar, so in that case it is a useless appendage.
> Then there's the fact that the editing keys don't work at all consistently or correctly (and sometimes just don't work at all)
I am not sure what you mean by “editing keys”, but Apple takes painstaking effort to make sure most shortcuts are global, and in text entry fields, Emacs shortcuts always work.
> And nowadays they like to hide functionality behind obscure multi-finger touchpad gestures.
Sorry but this is just a laughable complaint. Apple’s touchpad and accompanying gestures are considered best-in-class and have been so for more than a decade. Microsoft and various Linux distributions have tried to catch up during that decade+ and still aren’t close.
There are many things one can complain about with macOS ever since Lion started the iOS-ification, and I’d be right with you, but saying the old Mac OS X UI was lower standards than Windows 95 (and I assume XP) is a supremely contrarian statement that few people of that era would have made.
> Mac apps that are unapologetically _Mac_ apps. They’re platform-specific and they’re not trying to wow us with all their custom not-Mac-like UI (which often isn’t very accessible).
I wonder if there's a directory of such applications somewhere, actually.
Seriously. When I think of my daily tools I’m not sure I have even one application with a standard UI. Not one. JetBrains IDEs may be the closest, but I’ve already seen their UI reboot coming (soon mandatory I assume) and it’s disgusting.
> it's nearly impossible to click and hold on anything along the edge to resize the window.
This is what the resize box is/was for. (It's actually draggable edges with no "meat" to them, which introduce the conflict by overloading an element, which is also just a pixel wide.)
*) The resize box used to sit below the very bottom of the vertical scrollbar, just below the down arrow button, as a distinct and dedicated handle for resizing windows. Mind how this is in the direct vicinity of any scroll buttons which may be used to see more than what is exposed by the initial view. It's the quasi "natural" origin and anchor of any basic interactions with the viewport. (Somewhat ironically, while the resize box has mostly vanished from application UIs, it is still rendered by web browsers in certain circumstances. E.g., with textarea elements with scrollbars and resize enabled. With most UIs, there will be no related scroll buttons anymore, though.)
IIRC in that era, the leftmost part of the title bar could be clicked to get a menu of options that contained Minimize, Maximize, Move, Resize, Close, and maybe one or two others. If you chose Move or Resize, the mouse would change cursors and just moving the mouse would change the window, then you'd click to get out of it.
I'm pretty sure this had a button there in like Windows 3.11 or 98, that I think was later hidden so you'd just click the title bar to get it?
You'd click the application icon in the title bar, then in WinXP you could reach it by right-clicking anywhere in the title bar.
And in the Windows 3 era, it was an actual button (IIRC the minus on the left wasn't minimize, it was this menu, and the down arrow on the right was minimize): https://www.computerhope.com/jargon/w/win3x.htm
In modern Windows 10, you still get that in well-designed applications. In poorly-designed applications (hate to throw shade at Firefox, but it is one), you may be able to still get that menu by hitting ALT+SPACE. It just pops out of nowhere because there's nowhere to click to get to it.
ALT+SPACE, M, arrow keys is forever in my brain due to so many poorly behaved programs which would open their window somewhere off-screen, and that was the way to move them back onto the screen.
> ALT+SPACE, M, arrow keys is forever in my brain due to so many poorly behaved programs which would open their window somewhere off-screen, and that was the way to move them back onto the screen.
Happens quite often with dialog boxes that pop up to inform you of some error. Particulalrly if you're working on multiple screens and have moved the application off your main one. (Which is what at least I tend to do when working on a laptop attached to one or more external screen/s.)
You brought back a muscle memory I forgot I had: ALT+Space would activate the Control Menu [1] and then pressed down twice and enter to resize the window using the keyboard arrow keys.
> IIRC in that era, the leftmost part of the title bar could be clicked to get a menu of options that contained Minimize, Maximize, Move, Resize ... I'm pretty sure this had a button there in like Windows 3.11 or 98, that I think was later hidden so you'd just click the title bar to get it?
Still there, even in Windows10. (And AFAICR even W11; only used that a couple months last year.) It's called the System Menu, IIRC. If you see nothing to click, try pressing [Alt]-[Space] -- even MS Office apps, which have gotten rid of the icon, still pop up the menu.
> And in the Windows 3 era, it was an actual button (IIRC the minus on the left wasn't minimize, it was this menu,
I also called it "the big fat hyphen" when I first started using Windows 3, but AIUI now, that was actually supposed to be a picture of the space bar. As in, you know, [Alt]-[Space].
> I also called it "the big fat hyphen" when I first started using Windows 3, but AIUI now, that was actually supposed to be a picture of the space bar. As in, you know, [Alt]-[Space].
A pic of the old Windows 3.x Program Manager with its Program Groups within the Multiple Document Interface, linked somewhere in this thread, reminded me: Sub-windows within an MDI had a shorter bar for an icon, and IIRC you got the system menu (document menu?) for the focused one by pressing [Alt]-[hyphen] (or [Alt]-[minus]).
I have that quite often as I RDP in from various desktop sizes. You can bring up the Move command with Alt+SPACE then "M" and then use cur cursor keys to move the whole pane. That allows you to bring one of the corners into view so you can then drag it.
That does make it easier to resize a window from the bottom left corner, which is indeed where you want to resize the window from in the majority of cases.
However, if you want to resize the window from the top or from the left, that does not help you much.
While this is true, it also turns a task that takes only one action (resize the window up or to the left from an edge or corner) into one that requires three or four (move window, resize window from bottom right corner, readjust window position).
I guess, it's a question of perspective. For me, maybe due to habit, these are two distinct actions: moving the origin and resizing the viewport. And I'm actually not comfortable with having to combine them in a single interaction. More often, I do not want to combine both, especially on a larger screen. Moving the viewport or its origin is a matter of the desktop metaphor, as well as altering stacking in the Z-order. Resizing it is on an entirely different level and relates to another kind of mental model or mapping, relating to aspect ratio and field of view, which, if changed, requires a remapping of the situational awareness regarding everything that makes the desktop metaphor. (BTW, I also hate snapping windows.)
Personally, I think, it's this kind of constant overloading, which makes modern interfaces harder to use, or at least, causing them to put more strain on the user. (I guess, as interaction is becoming more and more the message and thus the media, like in constantly pulling the content area for more to load or swiping one thing away to see another thing, this is probably just as it ought to be. But this is essentially slot machine territory and not a working environment.)
Possible meant: "Window's resize widget is outside the viewport, so have to first move the window by dragging the titlebar until the resize widget is accessible, then resize, then drag the titlebar again to approximately restore the window position".
Still, this seems to be more logical to me than moving the window borders, since moving something in a suitable place first, in order to maybe manipulate it there, is much more like we would proceed with real-life objects.
(However, the window being rendered for an initial view in a way where not all the crucial controls are accessible is the real problem and arguable shouldn't happen at all. At this point, every action to fix this will probably be flawed. E.g., if we resize the window by dragging any of its visible borders, any action and/or cancel buttons will be still tied to the bottom part of the window, which is still off-screen, meaning, we're still stuck. On the other hand, dragging the window by its title bar, while maybe a more promising approach, may be impossible, as there is probably not sufficient screen estate left for this, to begin with.)
> That does make it easier to resize a window from the bottom left corner, which is indeed where you want to resize the window from in the majority of cases.
> *) The resize box used to sit below the very bottom of the vertical scrollbar, just below the down arrow button, as a distinct and dedicated handle for resizing windows.
At the intersection between the vertical and the horizontal scrollbar, if both were present. Or below them both, at the right end of the status bar, if the app has one (which many did, back in the day). Still there in f'rinstance Notepad++ 8.5.4, from June this year.
You have to scroll faster than the combination of bandwidth size and JS computation. Turn it into a game: visit the site progressively working backwards through your old phones until you can 'win'.
Then see if you've got time to take a screenshot.
If you actually need to click a link in the footer the only way to win is not to play.
Even better: automatic infinite scrolling. Wanna click on something in the footer, where useful contact details can typically be found? Ooops too late, it's gone!
Glad to see Postman mentioned. It has probably the most counterintuitive UI I've ever used. Sometimes I procrastinate by counting how many hamburger menus I can see simultaneously in Postman.
For years I thought it was odd that GitHub didn't let you do basic tasks in their mobile site. You couldn't browse releases. You couldn't look at the files in a PR. Then one day I accidentally hit my thumb over the middle of the screen and it scrolled. I had no idea all this info was there but required horizontal scrolling because there was no scrollbar. Moreover, the tabs all perfectly fit on the screen so there was no indication that anything existed beyond them.
It's still somewhat of a problem, but now I see GitHub doesn't perfectly align its tabs. It's obvious that some text is cut off and you need to scroll to read it all. Other UI elements are now placed behind a menu button and that's nice, too. But, it's amazing just how poor the mobile UX is everywhere overall.
This is common on sites like YouTube as well. Try to scroll through your front page without your mouse accidentally hitting random videos and causing them to auto play. It drives me insane. The best spot for the mouse that I've found is a small maybe 20px area between the menu bar and the leftmost video. Other than that, you're autoplaying something.
I agree with everything you said as well. Huge complaints from me over time. What if a Chrome window has fallen to the background behind another window and I want to grab it? Oops switched tabs because I can't get my mouse to stop in that little 10px area up top. Oof.
Oh, it can be, but it loves to just turn itself on. And I doubt anyone in charge of allocating FTE-time at YouTube gives a crap, because autoplay drives "engagement" and "views". Just actively user-hostile and safe to assume purposeful. My subscriptions don't randomly disappear, so why does this setting on my account?
The "autoplay on hover" thing can also be turned off, but get this, it's seemingly stored client-side so you have to turn it off on every computer you use, and if you reset your browser or clear the cache, it turns itself back on. I'm sure YT couldn't find the space to store a boolean.
It's a basic evolutionary principle. When there's no selective pressure, species breed for superficial beauty at the expense of other traits. So to get rid of that problem, we need to allow users to apply selective pressure. But the question is how. Individually our voices are lost, and when you need to log into app X to use the service within, you get no choice of app Y with a different UI.
Discord’s interface is extremely busy but if you ever used IRC I think it’s fairly intuitive. It’s a much more complicated app than Telegram with a lot more functionality.
As a long-time IRC user who never had problems with various IRC clients, I find Discord's interface quite counter-intuitive.
First example that comes to my head: I click on a channel that I haven't visited for some time, the view focuses on old messages (from the last time I visited it) rather than on the newest ones (which is typically what I want). And going to the newest ones isn't even trivial, I don't know a way other than scrolling down all the way, which may be quite annoying if the channel is very active and there is a lot of material to scroll past.
Unfortunately Telegram is getting more and more useless gimmick anti-features. It was better 3-5 years ago, when it didn't have so much useless junk. Still the most streamlined experience.
While voice and video calls, screen sharing is nice, they are not very polished, good quality, and UI is overloaded, but the last useful feature additions.
(useless junk: animated stickers, custom wallpapers, emoji responses (though I use them), premium emojis, collectible usernames, (i get that it needs money to run infra), channels and social-media-platform related stuff (i guess those are revenue drivers and signup-funnels as well)
Discord has one killer UX feature: voice chatrooms. It is great for gaming, and great for remote work too for small flat-hierarchy companies. (no setting up of a meeting, just dropping in to a "meeting room". I have worked on a product with similarly streamlined ad-hoc remote meeting user experience, but it is defunct now unfortunately.)
Funny that I would find the other basic comms channels integrated a good thing, if they would be properly polished, and the stuff you like is what I find junk.
I'm always baffled when people say this. I understood Discord immediately and fail to understand why people are so confused by it.
You have a list of servers you're in on the far left, you click one, and you get a list of channels. Click the channel, and if it's a voice channel, you can now talk. If it's a text channel, it shows the channel and you can chat. Voice channels have a speaker icon that make it obvious it's a voice channel.
Some servers will initially hide all the channels except one where you need to click a reaction on some bot's message to acknowledge their rules.
The only thing that Discord screws up on, IMO, is the configuration of notifications. You've got your account-wide notifications, server notifications, channel notifications, and the handling of @everyone and @here. Getting the exact behavior you want is tricky sometimes.
I don't know if I think it has more shady stuff than whatsapp or other mainstream platforms (so sad we have to write this as there is no open federated standard *in widespread use* as with email), which are backed by megacorps famous for spying on users and selling it to whoever, and being digital goons of governments when it comes to privacy and freedom of speech. All are the same from my perspective, with UX being a main differentiator.
man every time I want to respond to a message in telegram I end up in the forward mode, it's so intuitive in whatsapp but so backwards in telegram. Likewise when trying to react to a message with any emote other than the one set by default. And then I couldn't even find the setting to change the default emote reaction. So I guess I disagree with that assessment. I do agree about discord tho.
I blame designers/devs cargo-culting on fancy new stuff (webdevs on new css features, 2000s custom bitmapped UIs), and especially widescreen monitors on desktop, and pushing too much complexity to too small screen real estate on mobile as drivers of these processes.
Actually I found Windows Phone Metro UI as very well designed and clean UX for the small/touch-only screens, always showing a single view, but they backed out of that and made the current monstrous hybrids, with some worst of both words. (I was never really fond of the skeuomorph stuff on computers, and mobile UIs with lots of hidden interactions.)
I responded to something UI-related recently and closed with "if the cloud made shoes there'd be different ones for grass and concrete, but they'd all be the same size so you'd have to stuff them with prosthetics if your feet were too small or cut off toes if they were to large" and some kind stranger gifted me with a word:
seems like what all these issues have in common is when apps just go "fuck your configs, we want our app to look like/do this instead"
that last paragraph would really solve things, but also not being able to have that level of control does sorta suck for when you might really need it (and not just want to piss of the user lol)
I'm sure it's negligible these days, but there's also the concern of an increased graphics load since you'll have to do an extra pass for the transparency. But I don't want to dive too far into the mindset of "everything must be optimized to the max".
Computers are very fast. GPUs are way faster than you think. Of course, it doesn't feel like that way since all the abstractions and layers kill performance by a thousand cuts.
Remember, "Aero" was a thing almost 20 years ago and that was a graphically intensive window renderer at that time. Nowadays we could have 50x better effects without a problem but UI designers are not willing to.
Oh for sure, I'm with you. Old graphics work has this mindset of avoiding transparency passes stuck in my head. Certainly a small part of the bigger problem.
My 0.3GHz iMac in 1996 could draw and resize window drop shadows at 60fps while simultaneously encoding an MP3 without a hitch. It’s seriously inexpensive computationally.
Sorry, it was 1998. I’m pretty sure Mac OS 8 had drop shadows but even if it didn’t, I took that machine all the way to 10.4 and every version of OS X had drop shadows.
Classic Mac OS drop shadows were opaque, hence a lot cheaper.
I don't think OS X is a good example of your point because resizing windows was incredibly sluggish for a long time. I always figured it was somehow due to the combination of live resizing plus ubiquitous compositing, but I never understood why it was so slow.
Resizing is slow because it is a back-and-force between the compositor (insanely fast, no bottleneck here) and a given user application, which receives a resize event multiple times a second, has to relayout its UI (CPU-bound - OS X might have had a phase when it had to run a constraint solver), and then rerender every widget in the new size. Shadows don’t matter here at all.
There were live resizing hacks on classic MacOS that often worked pretty well, depending on the app. So it seemed pretty shocking to me how huge the regression was in the early days of OS X.
I don’t think the graphics hardware was always “insanely fast” back in those days, but even so, there must have been some terrible bottlenecks in the software.
As a user there was no way around it, and if I recall right, even as a developer it was hard to get decent resizing performance out of the system widgets.
One of the very few times I can recall where Apple has shipped something with such poor performance. Maybe most people didn’t notice or didn’t care because they just don’t resize windows very often?
I used to have that problem when using xterm: white text on black background.
I worked around it by writing a script that gave each new xterm instance a different dark background colour. Each window by its own felt like it was still black but when you put two next to each other you'd notice the difference in hue.
Then I switched to mate-terminal which put a nice big scrollbar in the right-side border of each window, so I didn't need that script any more.
I love not having window borders waste screen estate, and solve the grabbing issue by having the windows key + left mouse button bound to moving windows and windows key + right mouse button to resize. I can "grab" the window anywhere, with no need to hit borders or a title bar (which I also don't have...)
Things being "a waste of screen real estate" was a bigger issue when we only had 1024x768.
Now we have extremely high resolution monitors, and we use that to get really smooth angled lines and anti-aliasing, scaling up the visuals and decreasing the logical DPI, but the lack of a hard edge to grab doesn't get wider when the visuals are scaled up: it's still a single pixel wide.
The 4x3 aspect monitors were bigger. The 16x9 added only horizontal space.
We still need the same "real" size of items in the real world. And there is a trend to occupy the space with useless staff: ribbon, big G logo, status messages not about what you do but about bugs and internal program stucture.
> The 4x3 aspect monitors were bigger. The 16x9 added only horizontal space.
Unless you had a large monitor back then and a small monitor now, this is incredibly incorrect.
Back in the 1024x768 days, most people probably had a 15" monitor. You might have had 17" if you were lucky. Even at 17", that monitor is only 13.6 x 10.2"
Now, People on a 16:9 are likely running a minimum 22". At 16:9, that's 19.2 x 10.8".
...shit you might be right. But really only if you had a large 4:3 monitor and now have a small 16:9 monitor. 4:3 15" is 12 x 9", and 16:9 24" is 20.9 x 11.8. 27" is 23.5 x 13.2.
Laptop screens are of course a different beast. Still though, with resolutions being higher, certainly they could have used a pixel or two to create an actual window edge.
It's pretty much as big an issue now, because my field of view has not gotten any bigger.
And I think you missed the essential part of my comment: The reason it's a waste of screen real estate for me is that there is no need for any edge to grab when I can grab the window anywhere by combining with a keypress.
I have more real estate to "grab" than you can ever get with borders that way, without dedicating any space to it at all.
Modern design looooves wasting screen real estate on white space. Partially to look "clean" and partially as a result of avoiding any hint of skeuomorphism. Have to separate different UI elements somehow.
Trying to claw some of it back from window borders seems like it's attacking the wrong problem.
I hardly have any UI elements to begin with, it's almost all text. But that's also largely besides the point, which was more that window borders only matters beyond visual separation in the first place because of a UI paradigm that wants you to hunt for a specific place on the edge of the window in order to move or resize it, but it doesn't need to be like that at all.
The point is not to deny your preferences, but rather that they shouldn't be forced upon everyone else, especially when they break decades of proven usability design and practice.
My point is that the "proven" usability design and practice isn't all that proven.
There is a whole lot of unexplored or underexplored design space, and when people are complaining about having space to grab a window border away from them due to a tension between competing needs it's worth considering that there are alternatives where that tension just goes away:
You only need wider window borders because you've already decided that you need to grab the border, rather than decided you just want an efficient way of moving or resizing the window. Maybe that genuinely is the only option that works for you, but most people haven't tried alternatives.
I'll note that lots of user interfaces has this kind of modal mechanism to indicate which class of action you want to take on an object so you don't need to target some tiny visual area, which is always going to be too small for some people nearly no matter how big you make it. We're used to doing that. Most people just don't do it with windows.
Even if we were going to dumb computers down to that level - and really there are tons of things in any modern UI that is entirely undiscoverable - that is something trivially solvable by popping up a tooltip explaining the options the first few times a user grabs the border.
No, based on their keybindings it sounds like they use dwm, a window manager which you have to compile yourself, configure yourself in a .c source file, and comes without window decorations by default, I believe. It's neat but made for 0.1% of computer users at most.
Although I want to add the "Windows Key + Left/Right Mouse Button" bindings are incredibly convenient and should be supported by every OS by default imho. The area you have to hit with the mouse cursor to resize a window turns from "a few pixels in the window corner" to "the entire damn window no matter if there's a button or not".
I use bspwm, which is still not easy for beginners though not as extreme as dwm, but it's also kinda besides the point which is that a lot of the complaints about borders really are complaints about the difficulty of moving and resizing windows or otherwise manipulating rhem, and once you look at it from that point of view there are alternative solutions, one of which is this one.
Pretty sure you can configure any WM to use those key combinations. I used them in i3 and currently use them in bspwm. Not sure why you jumped to dwm immediately.
I'm sure you can also use them in a fully fledged DE, and someone mentioned using alt + click on windows to drag.
> Applications like Postman, Teams (and pretty much all of MSFT's applications these days), Chrome, and Insomnia should be case studies on how to not design user interfaces. They are about as bad as desktop software gets.
Sadly Tunderbird devs drank their coolaid, as evident in the latest release.
The border between UI churn and “I am old and dislike any changes” is very blurry. I have only tried it for a short time and I think they did a good job.
The same for Jetbrains’ new UI. If you put every UI change in the same bucket, how can the situation improve?
The UI is not terrible, but it feels very unpolished. The font line spacing seems off and too narrow, the toolbar and other UI components drift and shift when you switch tabs, the main window menu is in the wrong location, etc.
Edit: Somehow this URL leads to some spot on the Web page of the URL but does not always lead to the actual post. To find the post on that Web page, the post starts with
You have to open their disquss profile, ctrl+f search for "to me, a" then click on the "7 comments" part of "Discussion on Littlefield Advisors 7 comments" then ctrl+f search for "to me, a" again and there you click on the ".. days ago" link under the username.
Now everyone can bookmark it, or print it and glue it to the wall.
edit:
You should really be using firefox, click "view" > "page style" > "no style" to read that page. The anchor doesn't work anymore if you disable the style. They could have just <a name="comment-6294665157"></a> ?
The only solution I've found for addressing this in any meaningful way is AltSnap[0], which lets me avoid needing to find window borders for any resizing or moving operations.
I don’t understand why a power user wouldn’t use a tool like this. IIRC I used AutoHotKey for this when I was on windows. Most Linux window managers have this built-in.
I mostly stopped grabbing window borders and title bars about 15 years ago.
On Linux (e.g. KDE), you can simply Ctrl + mouse drag the window, and then click anywhere you want. This makes it even much simpler than the need of clicking on the border, even with larger borders. You actually don't need any window border at all then, or it can be full of other utilities.
There is also some software for Windows to do the same.
> You actually don't need any window border at all then
Even if that was a great solution for everybody, you still need a window border. Or at least I do. I need that visual separation from all of the other windows on my desktop, and I need it to be thicker than one or two damned pixels.
"Pretty" is subjective. I think a lot of it just looks like an experiment in minimalism, by people who forgot that there are colors besides jewel tones.
It almost seems like the people who make this stuff don't actually care about usability. They probably have top notch motor skills and can click tiny stuff instantly, or they're not actually using the GUI much, they do everything in CLI and have no idea what a GUI is even for.
We need better input tools to deal with this. AI smart mouse assist that makes the stuff you want "grab" the mouse, combined with algorithms that can cancel out stiction-related jumps. Maybe even eye tracking that knows what you want to click and pulls the mouse towards it. A zoomed view for when you're trying to highlight some text that auto pops up on click and drag.
> AI smart mouse assist that makes the stuff you want "grab" the mouse
I recommend trying out an iPad with a mouse — there the cursor sort of locks into the closest clickable element, and has to have a sort of “escape velocity” before it will leave the given button, only living as a free-form cursor in-between.
Apple has here a unique way ahead - due to those UIs being designed for touch-first, they have surplus information compared to desktop-only software (e.g. certain apps understand touchpad gestures as well, simply from the appropriate swipe gesture on mobile).
Some of those things you mention are configurable at the windowing system level depending on your system. Using SwayWM with xdg-desktop-portal gives control of some of the things you mention, and with keystrokes and tiling so that there is never a need to grab windows by the edge, or almost ever use the mouse for that matter.
> it's nearly impossible to click and hold on anything along the edge to resize the window.
That might be the proper nudge to get rid of yet another old&persistent UI mistake - why do you need to hunt for a tiny border when you could, e.g., have some tap/hold key(combo) and have the resize border become as wide as ~1/3 of the window width?
(same thing with the title bar, a wide and tall box in the center should be more than enough)
> They threw out several decades of usability pioneered by real HID experts
Yeah, no, those experts were just as smart as the current "UI/UX morons", with tiny borders and tiny buttons highlighted in the post being the testament of their expertise
that really bugs me any time i have to use a Windows computer. on linux (on pop os or zorin os at least) you just hold down the super/windows key and then left click and drag anywhere on the window to move it.
pop os also puts a thick border around that active window which makes it much easier to see where you are
There is a Powertoy to introduce this behaviour in MS Windows, I think it is Fancy Zones. OTOH, there is always Open Source Software to the rescue, in this case AltDrag and/or AltSnap.
Not that you're wrong, but IMO the moving and resizing complaints are resolved by the Ubuntu behavior of (IIRC) alt-leftclick-drag and alt-middleclick-drag. It completely eliminates the need to find any special target within the window: just click anywhere inside it.
That's just the top of another iceberg: features are increasingly harder to discover.
I use i3wm, which is a tiling WM. There's no window borders, so this resizing feature (W-leftclick and W-rightclick) has been there for many years, alongside a keyboard-only way to move or resize. That's hard to find unless you read the documentation, but that's alright, because the target user is supposed to be a power user.
But how are casual users of Windows or Gnome supposed to resize a no-tile window? Or to find a feature with no visual clue? I've had this problem myself on web sites where I can't guess where the links/buttons are, or with phone apps where I have to try several gestures until one of them does what I wanted.
This X11 (probably also in Wayland) feature Meta+mousebutton for resize, move, etc. has been there for the last 30 years or something. Yet it is not widely known, so let's be honest, it IS hard to discover. It can't be a justificiation for destroying the UI, esp. window borders.
Contrary opinion apparently, but I almost NEVER nned to resize windows OR use the title bar. I'm glad I have the extra real estate to do things that I do most frequently.
If I need to switch window, I am on a mac so I can easily just use the content of the window to choose my program. Much faster than reading for me.
The problem is also waste of real estate, empty space for aestetics. Reading a conversation in teams etc. is outright painful. My suspicion is that it's some company internal engagement metric gambling ala every click on the scrollbar is a "use", much "use",less time spend in other apps equals promotion.
I don't know what windows has, but in gnome, you can grab anything in the title bar to drag the windrow. You don't have to search for some empty pixel there. The edges of the window also easily work for resizing as well
I honestly don’t relate to anything you say. It’s like for every UI evolution you find everything that could theoretically go wrong, and say that it’s definitely a problem.
Like:
> Aside from not being able to differentiate one window from another similarly colored window
What? When has that ever happened to anyone?
I use Postman and Chrome daily and I have no idea what you think is so unusable about them?
Ok, but how is that an actual problem? You can use visual context to differentiate both windows, you’re not a visual AI. Also when does that happen that you have 2 terminals that overlap? I’d rather have my UIs be optimized for the 99% of usage, I don’t care if they lack affordances for cases that never happen in practice
Resizable windows are an outdated concept, full-screen applications with tiling capability (think emacs) are much better. Add multiple workspaces just as they are today and it is perfect.
Unfortunately Windows and MacOS don't really support this kind of environment very well. Sure you can tile windows, but they are still windows. When you open an application it is not immediately tiled or opened in another workspace etc.
Everybody says "floating windows are dead", but I fail to get sold on the tiling concept. For people working on 3 windows, that may be acceptable, but if one's juggling 10 windows on a couple of desktops, things get hairy fast.
Window resizing also helps a lot of things in terminal. Sometimes programs write log lines so long, I have to resize the terminal wider than the screen itself to make sure that every log is written in single line to be able debug things with some efficiency.
Also, not everyone uses GNOME. KDE has very nice features for floating windows like "dim inactive" which makes working with many windows a breeze, while effectively cutting eye strain, because your average screen brightness is lower, without turning down your lighting. Also, enabling window shadowing convincingly raises active windows over others, so brain's depth perception can isolate said window pretty easily.
Both ways (floating and tiling) have advantages over each other in some use cases. I use an hybrid approach (manual tiling of floating windows with snapping), but the scenarios I prefer tiling is really rare, I may say.
I am thinking that most of the time you want your applications full screen with tiling been an advanced user kind of thing
I myself mostly use MacOS without tilling, I just keep all my applications full screen and use multiple workspaces. I prefer to have a single big (32'' at the moment) monitor and just switch workspaces instead
But if MacOS had better tilling support I would probably use it more.
> I am thinking that most of the time you want your applications full screen with tiling been an advanced user kind of thing.
If using a 13" MacBook, yes. If using a 16" MacBook, maybe, but if using a 27" Linux desktop, the answer is definitely no for me, unless I'm running a multi-pane IDE or other specialist software.
MacOS is very optimized for that kind of workflow (I'm writing this comment on a 13" MacBook), but as the screen goes larger, the wasted space becomes too much. I even sometimes divide a workspace to two applications on this machine, to see more on a single screen.
For manual tiling, I sometimes enable Magnet and snap windows to corners, esp. if I'm away from my Linux desktop and need to do some system administration across a couple of machines.
> I am thinking that most of the time you want your applications full screen
I have a large screen. It's very rare that I want any application to be full screen. I want to use some of that screen to see other things at the same time.
> I just keep all my applications full screen and use multiple workspaces
I also like this setup, but do you have a workaround for the godawful animation that MacOS uses when switching apps? I can disable the panning effect, but not the fading one.
wait how do you disable the panning effect? I hate that. I just want the workspace to switch I don't need the stupid visuals that just slow everything down.
I have come to a funny realisation recently. It's not my eyesight which is becoming worse, it's the UIs which are becoming worse. The tiny scrollbars with laughable contrast are in no way accessible to anyone. I've recently switched to using KDE with the Oxygen theme and it's a joy to use without any eye strain.
These scrollbars are absolutely pathetic, with no room for customisation. (good luck theming a locked-down application) It's blatant how UI designers don't give a damn about the users' needs, not even in FOSSland. I am not entirely sure why this is the case but it's a sad regression from the days when we had good-looking, functional, accessible and snappy software. Not locked-down, unthemable electron bullshit.
Also, we wouldn't need "dark mode" or the such if the UIs had any contrast and were legible.
... Ya know, I haven't thought about that before, but screens are a lot brighter now aren't they. I wonder how much of the dark mode thing is just that the old good defaults are now eye-burningly bright, so of course people don't like them as much.
Screens can get brighter now, but if your screen is too bright, that's entirely on you. Adjusting the brightness is trivial and you can even do it without using the buttons on the display (DDC/CI).
I have a (marginally) HDR monitor, and I run Windows 10's HDR mode. Immediately it was noticeably blander than non-HDR, but I've quite grown to like it. I find it very pleasant to use for long stretches for work and such.
When I play games I get the vivid colors and contrast so all good.
Interesting, I have an HDR monitor but never tried actually using the HDR function. It does seem to make a difference and feels nicer on the eyes. I'll have to see how it affects games too.
About 3 years ago, I actually did research about monitors before buying one, for the first time in my tech-using life (which started around 1990, as I recall). Up until that point, it was never a question of "what is the best monitor for my application?", but rather "what's the biggest, highest-resolution monitor in my (pitiful) price range, that Best Buy (or whoever) has on hand right now?"
It began for me 15 years or so ago when backlights started to make high-pitched noises and 16:10 monitors started to become less common. It’s been rough since for various reasons, but brightness not going low enough isn't something I've experienced as a widespread issue. SRGB mode is pretty dim at 120 cd/m², so monitors usually are able to go at least that low, and typically much lower.
"Too bright" depends entirely on the context. I control-tab and alt-tab like a maniac these days (might reduce that by getting a fourth display, used to think that would be excessive but the Overton window has shifted and I'm coming around on the idea).
The default UI here on HN features your username in #828282 on a background of #F6F6EF [1], a contrast ratio of 3.54:1. The up and downvote arrows are #999999 on #F6F6EF [2], a ratio of 2.62:1. On a high-brightness screen, these less-intense contrasts look great. And this is far from the only place with 'nice calm greys' that are intentionally used by designers to reduce eye strain.
I run Visual Studio in dark mode, and Windows Explorer in dark mode, and Notepad++ in dark mode, and Omron Sysmac Studio in dark mode, and Autocad Electrical in dark mode, and Alibre CAD at the default (light gray) theme, and they look great. I'm in a nice, bright office, with (4) daylight 4' T8 LED bulbs directly overhead, and a whiteboard as the backdrop to my monitors, so it's not like I'm a recluse in a dark cave. The monitor brightness is not wrong.
At least, it's not wrong until I win-right or alt-tab over an Excel document or an old version of Studio 5000, where the color profile is stuck at black text on a white background. Then I'm instantly blinded. I can't set half the monitor (just the part over the Excel spreadsheet) to the right contrast and brightness to make Excel acceptable, because then I can't see better-designed apps.
And don't get me started on opening up a movie or game after work's done. "Set your brightness so the logo is barely visible." Yeah, no. "Game of Thrones is a cinematic show and therefore you have to watch it like you’re at a cinema: in a darkened room." Wagner and Snyder are watching their productions on studio-grade OLEDs. I won't put a show on the integrated screen of my old Thinkpad or Precision laptops, I know those displays are trash (and they're small, I only put stuff that's easy to see on them), but no brightness setting on my relatively nice IPS LCDs can comfortably handle the diversity of content they're used to display.
Yeah - I'm glad Android finally has the "extra dim" mode to reduce that with oleds (only barely effective with LCDs), but it comes at the cost of awful contrast. Lower output is much better in comparison.
I like being able to see my phone in sunlight nowadays, but there have definitely been some tradeoffs.
This is something that has always boggled my mind.
>Gee, this stark white brightness is hurting my eyes, maybe I should turn down the brightness on my display...
>Nah, I'll just demand every single company completely overhaul the CSS and contrast of everything while begrudgingly suffering the eye-burning whiteness of companies that haven't yet overhauled their CSS
If everyone would design against properly calibrated monitors it would be fine, you could just set it to the standard calibration, but they don't so there is no universally good brightness you'd never need to change for. It's like webpage sizes, if everyone built and tested their UIs at the standard physical screen scaling then you would run into a lot less variance in website sizes.
If everyone had calibrated monitors and every app and website was the same brightness, yes.
Instead, it varies wildly, often for good reasons... and then switching windows exposes you to extreme shifts. E.g. switch between photograph editing and a giant white text document, nothing's gonna save you then - stuff that looks correct and good for photography is absurd when most of your screen is the same as the sun.
Calibration is one thing, perceived brightness of the whole screen with specific content is another. And it's heavily influenced by how many nits are available.
Calibration in this case isn't about providing "and so everything then looks like it's the same brightness, because it's calibrated" it's about "and so the same modifications can be applied to everything consistently, because it's all calibrated". I.e. just lowering the brightness a bit below the intended value now is impossible. One thing built on a monitor with it's brightness curve undercalibrated and one with its brightness curve overcalibrated will look two different kinds of wrong when shifted by such a transform. That would not be the case with calibrated sources, everything would shift in the same way and you're able to have it look "wrong" (i.e. darker, capped, brighter, whatever) exactly the way you want, consistently.
Taking it to the DPI example, having things built at a standardised DPI isn't about making everything appear the same physical size it's about making everything tuned against a consistent physical size for the exact same reason, default is always intended and your global adjustments are always consistently resulting in the source material being larger than intended or smaller than intended instead of "well, depends how uncalibrated the source was if it's still smaller or larger than intended".
> Calibration in this case isn't about providing "and so everything then looks like it's the same brightness, because it's calibrated" it's about "and so the same modifications can be applied to everything consistently, because it's all calibrated".
You seem to be assuming the commenter you replied to didn't know that. As I understood it, all they were saying is that advice about calibration is useless for people who don't do photo editing, but whose problem is exactly that different windows have such wildly varying brightness that switching from one to another often makes everything look either pitch dark or third-degree-interrogation light in your eyes.
Which category of people do you think there are more of? My bet is on the latter. They need... Well, if you don't want to call it "another kind of calibration", you're free to come up with another term.
Does being outside generally hurt your eyes? Most complaints about brightness of screens has to do with the contrast between it and ambient lighting. When I'm in a dark room a bright screen is going to be hard to use. In a well lit room it's not an issue. It's kind of why ambient light sensing is a nice feature.
screens output way more nits these days, I remember not being able to see a screen properly because a lamp was on due to the monitors only having 200ish nits, these days you can get +1000 nit monitors, this is especially true for HDR monitors
To play PS5 I bought a monitor instead of a TV, because the place where I had to install the screen in the rented apartment is tiny, and tiny TVs are just crap with tons of input lag. So I got a gaming LG monitor.
Many PS5 games, specially those with HDR support, offer help in adjusting the brightness, often in the form of showing a very dark and a very bright image side by side, and telling you to adjust your settings until both are visible.
I found out that no matter what I do, this never happens. In the end the best setting is when NEITHER are visible. If the dark image is visible, the screen is so bright it feels like staring into a flashlight. If the bright image is visible, the screen is so dark that I can't see the contents of the screen with my curtains open or the lights turned on. I can't wrap my head around how someone can make a screen be so crap.
True. I use my screen with 0% brightness during the day and I have to reduce contrast during the night (which leads to reduced colour depth). Using dark mode is unavoidable for this reason.
Display manufacturers should really not just look at the maximum brightness but also the minimum.
I've been using dark themes whenever possible for a few years now, but recently realized much of unpleasantness of light mode comes for me from it being too blue w.r.t. the ambient lighting conditions. Playing with the slider in "Night Color" under KDE or "Night light" under Windows does the trick.
But the SW sliders are much more convenient; say, at work the desirable blue light level varies quite a lot (e.g., a sunny afternoon with blinds half-closed vs. a rainy morning), and a smooth adjustment is very nice to have.
> Also, we wouldn't need "dark mode" or the such if the UIs had any contrast and were legible.
Contrast is a factor yes, but the other thing that's happened with the flat UI epidemic is banishment of mid grays and light grays in favor of stark white and off-whites, making "light mode" much more bright looking than it had been previously. It's no wonder people were clamoring for dark mode after blinding flat UI had taken over.
Not only that but it seems like color from icons has gone - now we get black and white ones. I'd get it if these companies making the UX were small, because yes its super hard to deal with accessibility and different color blindness but I agree - we've lost anything but the extremest of the extreme colors.
The ironic thing about icons losing their color and detail is that it happened right as high DPI displays began to become commonplace in consumer hardware. We have these amazing screens with barely visible pixels and excellent color reproduction being wasted on rendering monochrome lines.
There's also an increasing prevalence and popularity of e-ink displays. Though some of these are colour-capable, most are monochrome, and I've definitely had challenges negotiating apps and content which use colour as a key differentiator.
Is there actually an increasing prevalence and popularity of e-ink displays? I know there technically exists some "high refresh rate" e-ink screens these days (meaning you can get 10 FPS if you're comfortable with a whole lot of ghosting), but aren't e-ink screens still pretty much only used in dedicated e-book readers?
Personally, I think it's just that monochrome icons are fashionable. I can't imagine that the GIMP developers for example chose to make GIMP's icons monochrome because that works better on e-ink screens.
There's a wide range of e-ink tablets and displays available. Many of the tablets run Android and are largely the same as any other Android device. The displays can be used with any OS, obviously. There are also e-ink phones (smart and otherwise), watches, and other devices.
For text and graphics, you'd generally want to use a higher-quality, lower-refresh-rate mode. Which is to say that there's quite a bit of established UI/UX knowledge that wants a refresh. I've distilled a set of basic principles: persistence is free, pixels are cheap, paints are slow, colour is (mostly) nonexistent, and the more ambient light the better. Paginated-navigation (whole screen changes in one go), line-art, and dithered or halftoned images work relatively well, raster images not quite so much, though can be acceptable.
My argument here (<https://news.ycombinator.com/item?id=37869476>) is that across all OSes, and for the Web generally, being mindful of monochrome, low-refresh-rate displays is something that should be front-of-mind for UI/UX teams and design.
Many of these issues already exist, e.g., for documents which are likely to be printed on monochrome devices. Colour is useful where available, but it is not universally available.
(And that's before addressing issues such as colourblindness or other disability / accessibility considerations.)
My original comment was not at all specific to Apple (you'd introduced them to the discussion), other vendors, or any specific OS or application. Simply that monochrome and e-ink are increasingly concerns.
But since you raised the topic of Apple, and since, contrary to your initial statement that, broadly speaking, monochrome was "a display technology they never used", I'm both correcting the record, and noting that should there be a monochrome or e-ink prospect on Apple's roadmap, and further noting that Apple rather famously does not announce such roadmaps in advance, that flat/white would in fact fit with e-ink remarkably well.
What that has to say about whether or not Apple is contemplating any such move I am, of course, entirely ignorant.
Since decades being mindful of monochrome displays and color blindness meant selecting appropriate colors and using different shapes along with different colors. Removing color where available is not useful.
Any explanation of the anti color trend must explain Apple's participation. Especially because Microsoft and GNOME appeared to imitate Apple.
I did not say monochrome was a display technology Apple never used. You named e ink. And monochrome is not a display technology.
Other trends included more transparency in content areas, links almost unidentifiable without color, and other reliance on small shade differences. And Apple made all app icons the same shape. All worse for e ink.
Monochrome lines are cool tho, ink or charcoal can be gorgeous, and “barely visible pixels” are genuinely needed to get them.
That aside, high dpi displays are still a rare sight on desktops, there’s not many of them and the aspect ratios and ancillary features are extremely limited.
>That aside, high dpi displays are still a rare sight on desktops
No, they aren't. ALL displays these days are high-dpi, when you compare to the 640x480 and 800x600 screens that were normal in the late 1990s. But the window managers back then were FAR better (in functionality and appearance) than what most of us use today.
> ALL displays these days are high-dpi, when you compare to the 640x480 and 800x600 screens that were normal in the late 1990s.
Leaving side your wilful misunderstanding of fairly standard concepts, no they are not: the monitors you are talking about were generally tiny, and the minor increase in median pixel density was largely counteracted by the blocky precision of LCD pixels.
In modern parlance, "high-DPI" or "HiDPI" does not refer to your 24" 1080p screen. That's solidly a "low-DPI" screen; it's perfectly usable with UI elements rendered at "1x".
I often wish we had more analog control in the digital world. I wish there were turn knobs instead up/down buttons with 10 steps (where it is easy to accidentally leave a menu... When trying to get back to where you were, you are suddenly adjusting something else entirely). Manual control with digital devices has become too digital and too cumbersome.
It used to be easy to adjust contrast and brightness on a display, volume on an amplifiee, an analog TV, a termostat, a car radio, etc.
I realize that we are not really on course (yet?) for reintroducing a lot of analog controls, but in the end, our world is analog. Input is analog via speech, muscle motion, etc. Output is analog, via light and other vibrations that reach our senses. Why isn't control more analog? It's probably a cost thing.
I would totally buy a display or a laptop with analog controls. I don't even care if the turn dial actually has 16M steps, so long as the response is pretty much immediate and feels like a real potentiometer. It should feel like direct manipulation and like you're in control, instead of these digital roundabout abominations.
As to the subject, I imagine having some knobs that I can adjust under different circumstances to quickly vary intensity or cycle through alternatives in order to make things more readable or audible.
90% Of what we do is in the browser today. Browsers could have an "accessibility" API such that turn knobs (bluetooth? whatever) could be used for control. Like scroll wheels but on steroids?
And when you finally have a along controls everything is overloaded.
Dish forced a new remote on us last time a device broke and they replaced it. It has far fewer buttons. I’m sure it helps getting familiar with it on a super basic level. But the old one wasn’t that complicated anyhow.
But here’s the kicker: there’s no fast forward or rewind buttons. There no stop button. No record button. All of these (and more) have been turned into menu items and/or secret chords on the remote.
Oh. And it has a mic on it too. Hard pass.
Tv remote is just the easy example. I see it all over the place. Sleek no longer is pretty to my eyes. If I see something that I have to interact with these days and it looks sleek, I see frustration.
I just got an Apple TV and got crazy with the unintentional swipes and touches. Then I discovered you can turn that off, to click-only. It‘s much better now, maybe that helps you too.
Yes direkt a 1:1 mapping of an analog (button/knob) would work better than what you describe.
Multifunction buttons… that never existed in the analog world? Or is that what is called “mode”? Send vs receive etc. Does anyone have a concrete example?
I’m not a remote control designer but I would think it would be fun to give it a try! Maybe I will :)
Most plugins used in music software do take that approach. Where the plugin interfaces are modelled to look 'like hardware' in most cases, with nobs, sliders, all the things you'd expect on a hardware compressor / synth etc.
Even better, I connect a midi keyboard so I can use its knobs and faders to manipulate the plugins. Much better feel, and for something like eq, it's nice to be able to manipulate multiple things at once
“…it's nice to be able to manipulate multiple things at once.”
Particularly if one thing affects another in a ripple-effect or chain reaction. It is great to be able to adjust at least two values at once. Keeps you from going back-and-forth adjusting in ever smaller increments to get to the settings you want.
Not this in this use case, in fact quite the opposite.
generally most want to have a knob they can adjust and not worry about the underlying value. It's in line with the philosophy of tweaking settings and not using your eyes to judge if its in the right position.
I hope you see this, just noticed your reply! Thanks for the encouragement, music production/DJing is something I really want to get into, so encouragement is extremely welcome. :) My problem when I tried using it was that the knobs were really difficult to adjust with my mouse, like it didn’t feel responsive. Am I doing something wrong?
It seems we are so amazingly primitive that we still think touch screens are a pretty neat idea for anything.
But I’m sure analog controls will come back at some time in the future: First in really expensive cars, as expression of luxury. Only cheap cars will still have the cheap touch screens. Slowly, the old new way of interaction will trickle down to anything else.
Like it went for the digital watch, ubiquitous for a short time, then a return to analog. Although those Casios seem to have a retro-retro comeback lately. I think even real keyboards might be a cool feature of future high-end smartphones.
I mean, for most things a touchscreen is pretty much the ultimate human interface — we are fundamentally hand-based creatures and the ability to re-render the same glasspane into a controller for anything is still magic.
There are of course things where more specialized inputs are required, but for the rest, touchscreens are here to stay*
* One improvement I would like to see it about their surface — we can no longer blind type on phones, because we can’t feel the borders of the buttons - but I think we have the tech to dynamically make the screen’s surface rougher/smoother. Another idea is to bring back 3D touch (and potentially improve on that - maybe it could even take some 3D vector as input?)
This is actually one of the things I love about my Mini Cooper; they have cool physical toggle switches that feel like you're in a cockpit instead of the all-digital interfaces that feel like a coffee bar.
> It's blatant how UI designers don't give a damn about the users' needs, not even in FOSSland.
Especially with this new "don't theme my app" movement. It's really unfortunate. The issues with CSS stylesheets in GTK should be solved with replacing CSS with a better way to describe styles, not with just throwing it all away and not letting the user set their own themes.
Honestly I want the app to just set its look exactly. I really don't care if different programs look slightly different as long as its internally consistent in app.
What I hate is installing a program that was designed for a KDE distro and because you run gnome all the icons are the same color as the background and alignments on things are whack.
Package it all up in a flatpak so it looks and works exactly the same on all distros. One complete and verified experience.
"One complete and verified experience." These words sound sickening to me. None of these have anything to do with accessibility, or respecting the user.
The user should be in control of his computer. That includes the programs on it. Locking it down so you can have "verified authentic design" is just straight-up user-hostile.
And who will make sure that the program actually works properly in that one-off hacked up, “press spacebar to heat up CPU” configuration you have?
Linux userspace is buggy, and I think this very-minimalistic approach sort of grown out from frustration about that. And they may be right, maintainers really don’t have the resources to do much better, especially when most frameworks they depend on has their own share of fatal, almost impossible to fix bugs.
This is the main reason I switched to MacOS. Was sick of everything on Linux being contentiously broken out of the box with the excuse "Oh you are meant to just configure the icon colours and toolbar padding yourself"
The programs can still be open source so you can recompile with whatever look you want, but they should bundle a working experiance by default.
But linux users seem more obsessed with curating custom themes over having things actually work.
I shouldn't have to recompile something to customise it. Make the defaults whatever you want sure, but even if you don't provide GUI config, provide an environment variable or a config file. "just recompile it" is incredibly user-hostile.
I disagree. The user has accessibility needs and preferences that the app developers necessarily can't know about. App developers shouldn't be trying to enforce styles/themes/icons on the user when they might be completely unusable depending on their needs. The inconsistencies between KDE and GNOME apps/themes are annoying, but that's a solvable problem.
There's also the consideration that a decent style sheet can be ~100 - 200 lines long. Yet I see ~5000 line css files that have so much custom formatting for JS frameworks that build everything with divs, rather than using semantic HTML5 which has accessibility baked in.
It seems like there's two wildly different ways that CSS gets used. In one style, a CSS class is a semantic category of things on the page, and the style sheet defines how it looks. In the other, a CSS class has no semantic meaning, and the style sheet defines a list of options to choose from. It's the difference between `<div class="user-comment>` and `<div class="bg-lt-gray-fg-dark-gray-bold-text">`.
The first type of CSS can be written succinctly, but the second cannot.
"don't theme my app" is not about users theming their apps, but instead against distros theming by default and more importantly the expectation that you could theme every app with a generic theme.
Same difference. If you don’t support theming done by distros, you don’t support theming full stop.
I suspect even their attitude towards theming done by end-users is more akin to ‘well sure, if you insist, you have the full source code, I can’t stop you, but you’re on your own’ than considering it a fully supported configuration.
Hard disagree with your first paragraph. An application can have endless customization abilities, but still doesn't need to let the system choose a theme for it to support "full theming" as you put it.
There can also be legitimate reasons not to support it. On the top of my head, for example, the default theme being the most tested, etc., etc. In a perfect world, they'd all be tested and well integrated, but there's limited man power going around, and the world isn't perfect.
Distros SHOULD be theming all applications. Once thing a distro can do well is create a theme for their distro so everything looks the same.
Maybe distros are not doing a good job of this, but that is one thing they have the power to do. QT, GTK2, SDL, wxwidgets, and more toolkits I can't even think of should all fit together on my desktop and look the same. xfce, gnome, and kde all have some nice apps, they should mix and match. This is a HARD problem, but that isn't an excuse to not face it.
I would also like to live in a world where it is easy for end-users and distros to theme apps. But with current technologies and cultural expectations this results in an untenable testing and support burden on app developers.
We could shift cultural expectations so that distro maintainers become the first tier of support for all app issues in their distro, and escalate issues to app developers if and only if they've confirmed that the issue is distro-agnostic. That would be sustainable for app developers but could be a heavy burden for small distro teams.
> This is a HARD problem, but that isn't an excuse to not face it.
Agreed, I just don't think it's the responsibility of app developers to solve.
> Distros SHOULD be theming all applications. Once thing a distro can do well is create a theme for their distro so everything looks the same.
I would reject a distro that did this (unless I could disable it). But users should be able to make everything look the way that works the best for them.
> I have come to a funny realisation recently. It's not my eyesight which is becoming worse, it's the UIs which are becoming worse.
We've hit peak usability about twenty years ago, in the Windows 2000 era. The screenshot in the article is actually from about that era, it's one of the early 10.x (10.3 I believe) OS X releases.
Arguably, Snow Leopard was peak interaction design. Just for the highly functional visual grouping of elements. (MacOS has drifted off from this quite a distance, since.)
Today in HN: A billion armchair rants about Windows 2000 being the best UI forgetting just how much it actually sucked at usability or accessibility features.
I'll have to admit that I don't know about its accessability features, but I was at least as productive in Windows 2000 and XP as I am in todays systems.
In Windows 10 i lost a lot of time because of UI bugs: double click registered on single mouse clicks, UI elements dissapearing when connecting monitors, no scrollbars, no window border, no titlebar.
> we wouldn't need "dark mode" or the such if the UIs had any contrast and were legible
Have you tried dark mode in Windows 11? It turns the entire window bar black, if multiple window bars happen to overlap it is impossible to tell them apart visually. On top of that in some programs it also turns the close/min/max window buttons completely black unless hovered by the cursor!
Microsoft should fire everyone at the company who ever made a UX design decision - except for the people who made the Windows Phone 8 UI, their only consistent and intuitive UI in the last 15 years.
I use "dark" mode in Windows 11, but it turns the title bar a nice blue shade since I actually went and configured it (it's right next to the mode setting). I actually can't configure the color in "light" mode where it's always white if I recall, though.
The other criticisms are valid though, the last good Explorer UI was Windows 7.
Windows Phone was visually stunning. Thanks for reminding me of it!
I don't know what you mean about turning the window bar black. Did you set the theme to "high contrast"? I've used dark mode on windows since it became an OS feature and never had that problem.
Microsoft's UX has been poor since Win8, but not for graphical reasons. 10 and 11 look great out of the box - far better than any other OS I've seen, although tbf I haven't done much distro-hopping.
The first thin I do on MacOS is turn on the scroll bars. Yes, MacOS Has them set as "Only Show When Scrolling" by default, which is probably the most ridiculous setting.
It seems like most people who scroll use touch, wheel, or other non-scroll-bar interfaces, and the scroll bars are just a visual indicator of document size and position while navigating it. I know I do— and have for 20 years unless I was using a shitty laptop without a scrolling input method. Maybe unless I'm scrolling through literally millions of lines of text, which is infrequent enough that I'd prefer to have them hidden and revealable by scrolling a bit.
Not having them take up screen real estate and working with most people's usage styles while being configurable for others is the right decision. The problem is systems and applications that don't let you configure them.
People talk about accessibility and their preferences like they are absolute truths— it's a lot more complex. Not having scroll bars sucks for people that never got comfortable with other scrolling input methods and people who use sight interfaces, for example. They couldn't possibly be less relevant to people who use screen readers, and as an input method rather than an Univision visual indicator, only slightly more relevant to people on phones and tablets. Phone and tablet usage has outpaced regular computer usage for quite some time, and that informs users interaction style. Adding visible UI elements adds cognitive load, and for people who never worked with scroll bars out of necessity, they're just another bulky animated distraction on the screen. Most younger users would probably think them about as useful as an always-visible on-screen keyboard.
I guarantee you— the designers at Apple who chose to hide those scroll bars probably thought more about scrolling more than every one of these commenters combined, and that was before they did formal usability studies.
The days of usability studies are long gone.* It's now change for the sake of change.
Spot the indication that not all settings sections are shown in this screenshot, in all display settings and lighting conditions that your device might be used in.
* Were they ever there? I can't believe this dialog background - https://arstechnica.com/gadgets/2001/10/macosx-10-1/4/ - was the "winner" in any usability study. This was removed in 2003-11 with the release of 10.3.
a) That scroll bar decision was made a long time ago
b) Problems with specific dialog boxes or dynamically created features is evidence of problems with those dialog boxes or dynamically created features. As someone who's worked both as a developer and a designer, often the engineers win.
c) Assuming that these topics are simple enough for off-the-cuff reckonin' to validate or invalidate them are exactly why the engineers shouldn't win when it comes to interface decisions.
Because designers and devs don't want to have to do multiple designs. Everyone's aiming for one design that works for everything. The problem is that such a thing is impossible.
Vollying back your flipness-- This sort of reductive assumption about software design is why companies that need people who aren't developers to actually use their products hire interface designers.
Most employers I've had are all about getting engineers to be everything. Designers, infrastructure, QA, UI/UX, and even product/project managers.. everything.
Sounds like you need new employers. The only people who want that are shortsighted managers who don't care that they know nothing about users' needs, and that's certainly not a universal disposition in the software business.
This has been true of every employer I have ever had. It's true outside of "tech" as well but nearly all of my experience has been as a software engineer.
People are generationally creatures of habit. Newer generations are entering the workforce for whom their first personal computing device was a touchscreen smartphone. With that context in mind, it's not surprising that a "native" responsive experience to them will be mobile centric rather than desktop centric. The market then evolves to cater to the needs of its largest cohorts.
Yes. To engineers, the interface is something you use to interact with software. To users, the interface is the software. When they buy (with money or creepy analytics for marketers) software, they're buying an interface. As anybody in business knows, you're not in the business you think you're in, you're in the business your customers think you're in, and if they expect software to behave in a way it can't, their competitor is a half-inch down the app store results/google listing/etc.
Controls? Regardless of whether or not you like the changes, device usage changes the way people use other devices. What you deem "correct" is not the objective universal truth you think it is. Not knowing that is why almost all FOSS interfaces are an absolute dumpster fire.
Does anyone apply formal usability studies anymore? (not just do them - do they apply the results). I was trained in formal UX long ago, and many UIs today outright violate common rules we learned back then. High contrast was considered important for low vision users even if it looked ugly then. Now most UIs are lower contrast and while they look nicer, they are less usable.
Of course I learned long ago and so I'm willing to accept that what was best practices back then might be wrong given new research that I'm not aware of. However this seems unlikely.
Formal usability testing is still practiced in enterprise, military, and professional products.
From what I've seen from colleagues in the consumer space they've predominantly switched to A/B testing and statistical clickstream/log analyses, though large companies sometimes still do run lab studies for important decisions and to rationalize the giant one-way mirror they installed.
It depends on how big/important/popular of a product you're talking about, but most software that your average joe uses (operating systems, native communication apps, national websites) likely either conduct significant usability testing or largely inform their designs using great usability testing from people like the mozilla foundation and the nielsen norman group. And no-- neither of those organizations have to be perfect for that to be a perfectly valid approach, and if you follow that guidance, you'll never end up with a 'very low contrast' interface.
I go to the store to buy a hammer. I’m a hammer user and I use the hammer.
Or I go to the store and I buy heroin. I’m a heroin user and I use the heroin.
These UIs we speak of, are they hammers, or are they heroin?
Because I’m currently looking at a high contrast screen using a mixture of Unix tools created from the late 60s to yesterday.
I’m looking at an Excel spreadsheet made by a team of lawyers that is in a UI. The lawyers themselves highlighted and handled the contrast of the document. Beyond the table itself, Excel’s menu bar is easy to read for the basic operating needs of a spreadsheet. I mean come on, the pivot table UI is clearly masterful. (If Excel was a native plaintext tabular format and you could pipe stdin/out to the GUI app? Gimme!)
However, a lot of software is basically useless these days. It’s not designed from the beginning to be productive in any shape or form.
Some stoned dance by a junkie under a highway overpass is definitely some kind of expressive art form. How much more productive is the expressive art forms promoted by TikTok, Twitter or Twitch?
There’s clearly a scale at play. I’m as big fan of Twitch as much as I’m a big fan of the Buffalo Bills. I’ve always appreciated the sociological aspects of sports. To add some context to my definition of productivity, I would say that group sporting events are overall productive from a sociological perspective, be it IPL cricket or Fortnite tournaments.
TikTok is clearly and quite obviously towards the junkie dance scale of things.
From an experiential position there was not much of a difference between my observations of a woman filming herself doing a weird little dance on the sidewalk while completely oblivious to the rest of the world and the junky on the next block doing a weird little dance on the sidewalk while completely oblivious to the rest of the world. This basic object-level assessment of the actions that we and others experience with our physical bodies is core to the very notion of “meaning”. The “meaning” is just the stoned dance.
I know of no junkies famous for only being junkies but there are definitely TikTokers famous for only being TikTokers so don’t confuse this for some strict equivalence.
I’m just trying to establish a kind of framework for discussing the relationship between productive UI and productive software tools.
A)Yes. I've been in the software business in some capacity for the better part of 25 years and software usability right now is miles beyond what it used to be. Thinking otherwise is a function of nostalgia.
B) if you're doing dev work, you're probably not entirely cognizant of how different your software selection and usage patterns are than most people's, even for non-dev things.
C) there's a whole lot of interfaces out there and the great ones work so naturally and intuitively that you don't even notice them. Those are the ones for which many assume the design decisions were so obvious that they didn't even need designers. That's so so not true.Look at the most popular user-facing software: web browsers are all obsessed with usability (mozilla actually does some of the best open usability research out there.) Instagram, Tik Tok, Facebook, Twitter, et al... Incredibly clear and usable for their target market. Can't get much higher contrast than you do in email applications and other communication apps. There are tools to guage if your text color/background color/ and text size make for readable text according to the WCAG and I have never had a job, either as a designer or as a developer, where AA compliance wasn't a minimum for all functional elements. Most common user-facing interfaces that regularly use light or white text on colored backgrounds, like the iMessage interface or signal, benefit from OS-level accessibility enhancements like increased text size, which are very commonly used among the people who need them. My non-tech-savvy elderly relatives all knew about them through the setup process on their devices when they first bought them, and never looked back. Looking through every communication or other popular user-facing software package on my Galaxy U22 using a default theme shows nothing even remotely low-contrast.
> the designers at Apple who chose to hide those scroll bars probably thought more about scrolling more than every one of these commenters combined, and that was before they did formal usability studies.
This blind appeal to authority shouldn't convince anyone.
That point is how you justify all your conjecture. Rethink what you wrote because all you're really doing is relying on anecdotes. I can literally say the opposite (I always use the scrollbar for scrolling and everybody I know does too) and we'd have made zero progress.
You haven't actually made an argument worth arguing about. It's devoid of value and it's not a productive addition to the conversation.
And don't bother wasting any more of my time. I don't care. You clearly have nothing to say.
I think you're missing the point. Usability is not prescriptive. You implement what your users have the easiest time using. Touch interfaces change how people interact with interfaces. If you care about usability, you work with people on that and not against them.
I saw those fat candy looking aqua scrollbars in the linked article and I immediately thought how great OS X used to look and work. Truly sad the mess macOS is now.
Not getting it. I mean, if Mac OS lets you configure scroll bars (and I haven't noticed a problem with the defaults so far either) that's much much more than can be said about gnome isn't it? Where gnome theming has gone from black magic in v3 to ... anathema in v4.
> Yes, MacOS Has them set as "Only Show When Scrolling" by default, which is probably the most ridiculous setting.
I have used MacOS every day for 10 years and i've never, except in exceptional circumstances, used the scrollbar to scroll. I don't see anybody in my circle using it. They all have the "Only Show When Scrolling" setting on.
I don't think this is ridiculous and I would enable the setting if turned off by default.
I also spend 99% percent of my time in full-screen mode in safari, vscode or terminal. So the top bar is also hidden by default. I think this is much more focused.
The scroll bar, for me at. least, is also used to gauge how far along in a document I am. So its use is more than just for moving. If I have to find out here I am in a document by scrolling that just means extra effort on my part.
Yes, we should absolutely make everything less accessible in the name of aesthetics. That's made computer interfaces so much better in the past 20 years.
"Dark mode" is needed in dark rooms, since screens don't autoadjust their backlight based on the room's white point.
The concept of "white point" is often spoken of in terms of color, but brightness very much does matter too. It is painful for there to be something (especially something large) brighter than "white".
**
That said, we should have automatic dark mode based in inverting HSL's "lightness" (this can trivially be done directly in RGB, just check the min and max color channel of the pixel, then adjust all channels to flip them), rather than requiring ad-hoc color schemes all over the place.
This is one thing Apple got right a veeeery long time ago. Every screen they have sold since like the mid 2000s, both integrated and standalone, has had auto brightness. I don’t understand why other manufacturers still can’t that right, to this day, no matter the price category. Even most non-Apple smartphones these days still suck at this.
I’m surprised Linux doesn’t have some ubiquitous “use the camera to do auto-brightness” option (pinging the camera on a proprietary OS would be creepy, but if it is open source…).
That wouldn’t work very well because pretty much every webcam has auto exposure adjustment. Also a camera is good at determining how much light is in the room behind you, not how much light is hitting the screen. There really is no substitute for a photodiode and a halfway decent control algorithm. It only costs a penny so there really is no excuse not to include one.
I know a lot of people who have covered their camera with tape. Some laptops even come with a physical shutter. This is something company policy often recommends - if the camera is physically covered an attacker can't see the next product we are working on.
When I buy a laptop, one of the things I'm looking for is that it doesn't have an integrated camera at all. It's happened sometimes that I can't find one, then I cover the camera with tape.
The Lunar dev has a variety of recipes for a hardware widget you can build which allows their macOS-only brightness-control app to implement adaptive brightness on any monitor that permits backlight control over DDC/CI. I'm sure a more-motivated and clever person than me could tap into that hardware and use it to drive ddcutil/ddccontrol outputs on Linux.
CRT monitors had easy dials to quickly adjust brightness and contrast. You almost did it automatically without thinking when you had developed the motor memory. Nowadays monitor controls are often fiddly and awkward. But I still use them to adjust brightness to the environment.
There are some tools out there that will allow you to bind hotkeys (or at least provide a widget) to send brightness-adjustment commands via DDC/CI over the HDMI/DP cable to your desktop monitor. This may improve your UX. My 2015 and 2021 Dell monitors both support the functionality, with differing degrees of finesse.
> Also, we wouldn't need "dark mode" or the such …
People with cataracts may benefit from dark mode. People with glaucoma or higher order aberrations (e.g. me) may benefit from light mode. Web sites should use `prefers-color-scheme`; native GUI programs should use the native equivalents (or default colours); command-line programs should not assume a particular background.
The main problem is that there is no dark or light mode. It is light gray on white or dark gray on black (and the white and black are also gray). It is just sick.
My current favorite is the vanishing scrollbars that appear for two seconds after you start scrolling. Chrome is awful with this. When I'm trying to review a couple-thousand-line PR, Chrome stutters and freezes, and the scrollbar won't show up for several seconds and then it immediately vanishes as the UI catches up with the input. Or sometimes I click on the ephemeral scrollbar but my click doesn't register until after it's vanished again. It's infuriating.
My experience is 'developers' have no issue with scrollbars; it's the higherups and designers that find scrollbars antiquated and obsolete in a world with touchscreen devices.
One thing I just love about KDE is that when you right-click on an "empty" area of a scrollbar, it offers you options to go to the very top or the very bottom of whatever view you're seeing.
I seem to recall that something like QtCurve offered buttons to do that within the scrollbar.
If only everyone did that, many time wasted in scrolling would be saved and "go to the top" buttons in websites wouldn't exist
That sounds nice, but I'm using KDE right now and just tried it on several apps and it doesn't work. In Chrome, it just brings up the regular right-click menu the browser normally provides. In actual KDE apps, it does nothing.
It's also sad that 10-15 years ago near-everything I used was just themed based on system/windows manager and it was possible to make those changes system-wide (with some mess if you used QT apps with gnome and vice versa, but still), and now we're locked in hellhole of HTML and CSS "native" apps
The tiny scrollbars with laughable contrast are in no way accessible to anyone...It's blatant how UI designers don't give a damn about the users’ needs
I’ve seen this trend going on for over a decade now, like a slow parabolic arc. Somehow the aim of developers and designers turned around 180 degrees and became less and less about pleasing the users and more and more about impressing other people in your field. (Both programming and design.)
You're right about the scrollbar contrast, even though I've customized the hell out of the scrollbar in my collAnon pwa I made the mistake to make it less contrasting for design appeal reasons alone. I guess that's gonna go in the next release
True, but not relevant and not a contradiction of the comment. Pannoniae says that the empirical decline in usability of software is attributable to changes in the UI and not attributable to changes in his eyes. The fact that changes in his eyes exist is not a refutation.
One thing that I've found interesting (in my 40's and slight astigmatism) is how the effects of eyesight show up in interfaces or where there's contrast with brightness/darkness.
In the physical world that'd be issues like bright points like car headlamps or street lights having a large corona/starburst, or if you're looking at a bright object against a dark background such as reading a book where it's lit in a darker room the 'bloom' will obscure what's behind it. On a screen with much brightness I'll get similar effects if there's contrast. I've been window shopping for a HDR display for a while now and wondering how much benefit I'd get out of it seeing as the main selling points are the brightness/contrast, especially when you're getting into the various forms of local dimming to present the media at its best.
For me, with (mostly-corrected) astigmatism and some (uncorrectable) higher-order aberration, dark on light is the only thing I can read fluently. The worst case is ‘black holes’ on a light page, as are fashionable for illustrating command line examples (and f—ing github's CI log view), especially combined with Las Vegas syntax colouring of varying brightness.
I run into so many websites that I think are completely broken because they have a scollable popover or some other weird element, but it's impossible to tell that there is content because the scroll bar is hidden on this small interior frame.
Absolutely infuriating. I blame iOS and macOS for encouraging this insanity and starting the fashion trend, and tangentially whoever pushed along this whole "flat" UI trend that makes it so hard to guess at what is what.
UI is communication, and UI designers have decided that mumbling is cool.
In the context of the original iPhone at least, hiding the scrollbar when not scrolling makes some amount of sense… when your screen is 3.5" in size and your goal is to render content with "desktop" fidelity, there's not a whole lot of room for a scrollbar, plus most people aren't going to be interacting with the scrollbar.
For desktop OSes where the smallest screen being used is much larger on the other hand there's not much of a good reason to hide them.
It looks nicer + you can just use the scroll wheel. A lot of the UI affordances lost over the last decade or two are unfortunate, but complaints about scroll bars seem to be purely baby duck syndrome.
As the article says, not everyone can use a scroll wheel, and not everyone even has a scroll wheel or touchpad. I personally find using mouse scroll wheels quite painful due to RSI caused from decades of mouse abuse. Yet, designers find scroll bars and up-and-down scroll arrows icky so fuck me, I guess. :(
Obviously not, but you also shouldn't expect the whole world to be designed around your needs. If you have difficulty using the expected input method, you should use alternatives with the same functionality, not expect others to redesign their apps in your preferred way.
Often, you don't even need to lift a finger, design-wise or implementation-wise, to have good accessibility. The system-wide or toolkit defaults are usually quite accessible right out of the box. macOS's horrible disappearing scrollbars being a huge exception.
Most of the terrible designs and accessibility problems stem from software going out its way to write custom controls or force controls to look and behave in a non-default way. It's not an afterthought--people are deliberately adding code to make their software worse.
"Accessibility" here refers to one person's very particular disability. You can't possibly predict the entire gamut of disabilities people might face. Trying to do so "first and foremost" is utter madness.
Nobody's asking you to predict anything, they're asking you to leave the functional design elements of a scrollbar alone and not fuck them up. Probably if you rent an angle grinder and remove the handrails from the steps at your local library it'll look nicer, but you don't do that, because there's a good reason for those steps to have handrails.
Designing with accessibility in mind makes software better for everyone, period.
You design a clear, visible, distinctive scroll bar and buttons? Boom, you've catered to everyone with bad vision (for any reason: from blindness to eye surgery), reduced motor skills (for any reason: from arthritis to old age to hand injuries), non-technical users (so they don't have to hunt around the interface hoping to discover hidden features).
And this goes for everything in software.
Disability is a spectrum, and you yourself will be disabled in one way or another, multiple times, during the course of your life. Be it from old age, surgeries, injuries, strain from sport or mundane tasks like holding a baby.
There are a number of common disabilities. Everyone starts losing near vision just after 40, which works out to almost half of the people using a computer expected lifespan is around 78, and babies don't use computers). 10% of the population is color blind of some sort. Most people will have a broken arm sometime in their life (the only statistic I can find is 6 million people in the US break their arm every year: this could also be a few clumsy people breaking their arm several times per year and most don't ever. I think most people breaking their arm at some point seems more likely)
It does not look nicer. It looks more mysterious. Where in the list am I? Is there anything below to scroll to? Above? To the right, everyone's favorite?
And no, trying to scroll every control just to find this out is not my ideal of usability and comfort.
In an age of information overload, perhaps the mystery is exactly what we need.
We know the exact temperature and humidity outside, we have logs of the exact times that our smart lights turned on and off. Our watches tell us our heart rate throughout the day and even its variability. Our phones monitor how much we use them, and in which apps, and report back weekly. When I had outlook in the office, it would summarize how much time I spent in meetings, seemingly to increase my stress level one more notch.
The lack of a scroll bar returns mystery, but even more, it brings us back to a simpler time when we weren't overloaded by information such as "how much is left of this text?" or "is there a button hidden somewhere on this page that allows me to complete my task, or is the app just broken?"
These are the sorts of small joys that we miss in the Information Age.
It's so you can doomscroll for hours without realizing how much content you have consumed. It's pointless to know how much is left because the feed is infinite. You can't ever get to the bottom that a scroll bar would imply as new content is loaded.
It is an enabler of information overload, not a solution or respite.
I don't mind having them hidden as an option, but there should be an OS user level override that all apps and sites *obey*. macOS has the first half covered, but electron apps and some browsers have yet to cover the latter half.
(For humour purposes only:
Fuck me, I've been holding it wrong the whole time!)
On my desktops I've got more screen width than any of my open windows know what to do with, and they still want to skimp on scroll bar width. I can understand with mobile, but mobile design decisions are crowding out desktop usability. It's lazy and cheap but the market is monopolised enough that the approaching horizon can only look "no worse" at best.
> It looks nicer + you can just use the scroll wheel.
"Beauty is in the eye of the beholder".
My work laptop does not have a "scroll wheel". There is a posibility to scroll using the touch pad but it is undocumented and hit or miss.
> UI designers have decided that mumbling is cool.
At risk of hijacking this complaint thread onto an even more inflammatory set of rails:
They may just be copying modern cinema. I can't watch many movies or TV shows these days without subtitles - even when the kids are quiet / sleeping / at school.
Once again I have to give firefox credit for giving users the tools in about:config to disable this. It's cute fluff at best and annoying/abusive at worst.
The browser's UI should more or less be fully off the table when it comes to what a website can change, and that includes scrollbars.
And on the flip side, Chrome - just had the biggest laugh when I got to this point:
"Imagine being able to configure anything useful in chrome ever."
Overall, a very well written article - kudos to the author.
(I've made the switch to FF again myself after Edge just got too annoying with their constant nagging about Bing search and other features. And the prompt to restore tabs. Just STFU and stop it already.)
Could you explain this point to me? I just checked and out of those two, it was _Firefox_ that has tiny, barely visible, vanishing scroll bars and _Chrome_ that has thick, well-visible, persistent ones. (And no, I didn't configure that on either)
The only thing that IMHO needs improvement in Chrome scroll bars is that the knob is light grey on lighter grey.
One useful thing you can enable in Chrome is tab bar scrolling. Without this setting the tab bar is broken by default, if you open too many tabs on it the bar will simply not show some of them because there's no room left, while all remaining ones are just a few pixels wide.
> The browser's UI should more or less be fully off the table when it comes to what a website can change, and that includes scrollbars.
I have a very strong recollection of adamantly defending IE6's ability to style scroll bars on forums in the early 2000s, whereas the Mozilla crowd at the time called it an abomination.
As an older man with both the power of hindsight and the weakness of failing sight, I can admit I was wrong. It's far too apt to abuse
If it were restricted just to style abominations, that would be one thing. It's the sites — and the libraries — that hijack the native behaviour, changing the speed / acceleration / easing of the scroll that are the really bad offenders.
It's not just your eyes, we were using very different screen resolutions back then and UI was lot clunkier looking too. Having the option is nice, but being able to disable that is critical
Let's include smooth scrolling in that list, too (whether scrolling with the mouse wheel or moving between words I Ctrl-F for on the page). Thankfully you can override this with uBlock Origin.
Also, kudos for the extended family of extension writers. The article author ends on praise for minimap sidebars as an upgrade over conventional scrollbars, and what do you know, there's something for that:
The problem with about:config options is that Mozilla sometimes decides to remove them (based on telemetry numbers perhaps). The usual path for an unpopular option seems to be GUI -> about:config -> The Void.
about:config options are workarounds, not solutions. This is for two reasons: first, they can change or vanish at any time, as you point out. Second, they're not very discoverable. You generally have to be told about them to know they're there.
Does the suggestion in the article not work? There are few options in about:config for scrollbars so I'd look at the others too. If that doesn't work I'd bet https://old.reddit.com/r/FirefoxCSS/ has a solution.
That is confusing... I'm not sure what the unit of measurement there even is. Something like pixels would be nice, but that can't be it with a default of 1!
about:config is invaluable, but better documentation on the options would really be nice.
Not to mention it is the only program I ever used which lets you directly rearrange the UI. Three clicks and you can reorder or remove buttons, margins etc, we need something like that in every program including a color chooser for every element.
The author sees half the problem and offers the worse half of a solution.
A scrollbar is not only a control, it's also an indicator. It shows where your viewport is relative to the bigger view, e.g. a long list. Hiding scrollbars, like some [redacted] GUIs like macOS do, is just acutely impolite, much like making buttons indistinguishable from text, or putting light grey text on a light grey background.
A solution exists, and is nearly universally implemented: scroll wheels on a mouse, and scroll gestures on a trackpad. Very, very rarely do I find a view which can be scrolled by scrollbars, but cannot be scrolled by the standard trackpad gesture, or by a mouse wheel. (BTW my mouse also allows horizontal scrolling using the wheel; so do trackpoint controls, too.)
Such scrolling does not require the pointer to even be at a scrollbar; the pointer just need to hover over the desired view / control / widget. This is really easy to achieve with imprecise input devices, unsteady hands, and poor vision.
But with this natural, easy way of scrolling one starts to really miss scrollbars in their indicator role.
(Regarding minimaps: sometimes they are helpful, sometimes not so much. I in particular find them unhelpful and bulky for text editing, but of course I'm all in favor of having them as an option for those who enjoy them.)
Scroll wheels and gestures are paging mechanisms, a scrollbar does that as well but it’s also a “take me to the middle of the document” control.
Assuming a properly responsive document (not something you can take for granted today, especially on the web), you can readily use the scrollbar to navigate large ranges within a document, especially large ones.
Countless times, on large, old PDFs with no links and that use that kind of “B-29” section page numbering have I essentially used the scrollbar to binary search the document for some buried page.
Of course the modern web has practically destroyed the applicability of the scrollbar as indicator because of the rampant use of lazy loading and endless scrolling. Many times on the Mac I’ve tried to use Cmd-down arrow to jump to the end of the document, hoping that there IS an end, and that it will load all of embeds that wreak havoc with the formatting.
But, alas, I find I’m on some endless train, with no hope of knowing how far I’ve gone or how far I have to go.
All that said, I happen to have a weighted mouse wheel on a bearing that is specifically designed for high velocity doom scrolling, partly because the scrollbar is effectively useless.
> Scroll wheels and gestures are paging mechanisms, a scrollbar does that as well but it’s also a “take me to the middle of the document” control.
IME, only on Linux am I able to click on an arbitrary location beneath or atop the scrollbar and be taken there immediately. On Windows, it seems that clicking below the scrollbar is equivalent to scrolling down once. If you want to go all the way to the bottom of the document or webpage, you have to drag the scrollbar all the way down manually. I actually find it kind of annoying. I don't know about MacOS, as I haven't used it since I was a kid.
>On Windows, it seems that clicking below the scrollbar is equivalent to scrolling down once. If you want to go all the way to the bottom of the document or webpage, you have to drag the scrollbar all the way down manually. I actually find it kind of annoying.
What's stopping those input methods from having a special scroll gesture that is optimal for their input medium? But sure, have an accessibility override for giant scrollbars.
I thinks that of all things, eye trackers and voice controllers have the easiest time sending whatever scrolling commands, without a need to fiddle with scrollbars.
> A scrollbar is not only a control, it's also an indicator.
Thank you! finally someone understands what a scrollbar is for. It primary job is to show the user that the document is bigger than the window, and secondly, to show which part of the document is visible. Letting the user scroll around is not the primary function!
I was genuinely surprised when Apple started hiding scrollbars by default in macOS. Their UI designers clearly don't have a clue what the basic UI controls really do.
Not just that "the document is bigger than the window" but the best scrollbars tell you, approximately, how tall the document is in relation to the window height! Proper scrollbars are, possibly, one of the best UI elements ever invented.
Input devices for scrolling are not universally available.
Wacom/pen users don't have a scroll wheel.
Many trackballs don't have a scroll wheel. Some have a scroll ring around the ball though.
I got RSI in my scrolling finger, so I removed the scroll wheel from my mice. And there have been times that I have overstretched my mouse-arm and had to use a trackball with my other hand for a few weeks.
There are extensions for chrome and firefox[0] that can allow for continued scrolling by scroll wheel to exponentionally (as in, actually exponential) speed up. With this, I pretty much never use the scrollbar UI element as an input.
On windows at least the middle button could be used to anchor a virtual scroll wheel or whatever that feature was a called. I’ve used it quite often on large content with small windows.
??? macOS (and iOS) is designed this way because of the gesture you want! You don't need the scrollbar because you have the gesture. If you want to see where you are in the page you just scroll a teeny bit. It works great, I haven't seen a scrollbar in years and its never a problem. Sure, some accessibility options are nice for some people.
A gesture is an action. It's good, and it works as intended.
A scrollbar is (also) an indication, glanceable, not requiring any action.
If someone among your friends and family has an iOS device, remember how many times you had to show them that certain screens can be scrolled down? Oh, not just down, but also to the right? There used to be a few perfectly aligned views that had no indication whatsoever that there's more to them is you scroll.
Narrow but visible scrollbars would solve this easily.
On the contrary, I've professionally seen even technical users get confused by their OS hiding the scroll bar, and them not realizing what they're looking for requires scrolling.
People literally did not know that you could scroll the share menu on some versions of android until they specifically added a "fold" to hide half an icon just so people realised that there was more options available. Hiding a scrollbar is so unnecessary. It's one of the best UI/UX elements ever created and people got it right so long ago
Don’t underestimate how much pulling the rug out from under people via constant UI churn negatively affects their lives and interactions with computers. Leads to lower confidence, helplessness, etc.
On macOS at least, it's not one change, it's at least four:
- scrollbars overlaying the content, rather than having dedicated space
- scrollbars that auto-hide when not in use
- scrollbars that are initially narrow, then widen when you mouse-over
- removal of the scroll buttons
That's not including the visual design changes, which are arguably important too (e.g. the scroll "thumb" now has a very flat design rather than textured so it looks draggable).
Edit to add: oh yeah, one more I forgot: in classic Mac OS the scroll thumb had a fixed size, now it reflects the size of the viewport. That one is a useful and positive change, though. Not sure where that innovation came from but it wasn't the Mac.
Many changes are still many changes even if they happened at the same time. You can't just lump together all the changes to the design to claim its only one.
I don’t have citations handy but I don’t think they changed it all at once. There might have a been a single major switch on macOS but I think all this stuff was developed incrementally on iOS.
My theory is that designers target users viewing 1 full-screen window at a time. Their model user never moves or resizes a window, or opens more than 3 tabs/documents a day, as they sit at their 13" laptop in an ideally-lit room with no reflections.
Then you get programs having splits or panels or whatever they're called for multiple documents, which is just an internal window manager that's not as good as my regular window manager.
Whenever I read a piece like this or another UX blog, it just becomes so obvious how little we care about a11y.
“Good” (as in accessible) design is pretty boring and decluttered compared to most modern expectations of web apps. I read Adam Silver’s book on forms and came away realizing that we’re doing it entirely wrong from an a11y standpoint but that’s just not a priority.
Numerical ones don’t seem to exist outside of IT, but other fields do the same strategy of abbreviating common long words by keeping only a letter or two. The non-numerical way is to just replace the rest with an “x” like “txn” for “transaction” in finance or “pax” for “passengers” in transportation or “sx” for “symptoms” in medicine.
> Is it still okay to make "nothing sucks like a VAX" jokes?
Please!
I was a stage II CS student learning assembly programming on a VAX simulator. Wrote a VAX simulator, in VAX assembly then ran VAX assembly programmes on it.
I claim your time would be better spent figuring out why it makes you unhappy and working out how to not mind it - this is a somewhat common means of abbreviation, and you'll probably encounter it again.
It's a longstanding abbreviation for "accessibility", along the same lines of i18n being an abbreviation for internationalisation. There's the first letter, and the last letter, and between them a count of the number of letters elidded.
It's really uncommon to see, I wouldn't call it "long-standing" at all. It's really obscure jargon that even most technical people don't know. At least "i18n" is widespread enough that most people will see it, though that's stupid too because it's incredibly unclear. I had no idea what it meant until this year despite having seen it for a decade or more.
The word you're looking for is "jargon". It refers to niche-specific words that are hard for outsiders to understand, and isn't just tech-specific. "Lingo" is a much broader term.
True. I was grasping at straws for the right word. Thank you.
The Greek/Latin/<jargon> structure should be standard or taught in schools in some capacity.
Basic programming concepts are popular, useful and should probably be taught as a sub category of English... given that programming is supposed to be a language, and pull its roots (somewhat) from English or "natural language".
Society would produce higher quality code if the basic concepts were considered as a literacy requirement for children.
It's a word that starts with a, then 11 letters, then ends in y. The format is not hard to grok once you understand it then context and sheer elimination leaves only a few choices
edit: I will say this is a bad example because looking it up there is 40+ words that fit this description so maybe I am biased by experience. i18n works better but I think my point is no longer correct.
I would hope (read: pray) that on HN a11y isn't jargon. On a thread about scrollbars (i.e., UI and UX) a11y should be as accessible as saying UI or UX.
Again, obscure and jargon are a function of context.
Spitballing...Keep in mind, without context, "accessibility" is also jargon in the sense not everyone is going to know what and how that means. My mum certainly wouldn't understand a11y, I agree. But she also would understand "accessibility" or even "website accessibility."
Full disclosure: I'm 5x more picky than the next person when it comes to comms. Absolutely, words matter. But one of the foundations to comms is context. The context in this case is HN.
It’s not exactly wrong from a business standpoint. The math just comes out that a business does better when they do all the things they can’t do if they want an accessible website. Dark patterns, especially, are very anti-accessibility.
Laws to mandate accessibility, laws to mandate provable correctness, laws to mandate fairness, and laws to break up the big, for only the big could satisfy all the other laws.
No reference I know of. I'm just struggling to balance on the one hand each field's demand to be taken seriously (and since the public doesn't take them seriously, experts appeal to legislators for validation), and on the other hand the Brandeisian push against Bigness that arises after decades of regulatory capture.
I have no stake in a11y, but I hear the same appeal re software correctness and statistical significance and many other individually reasonable asks, that nevertheless add up to a gridlock of regulations.
You know what's an accessibility problem for a lot of people over 40? White text on a dark background. I don't think I'll ever get tired of hearing people advocate for making web pages readable for people my age.
Yeah, opening up the inspector and updating the author's CSS made it readable for me, but made the shell script unreadable.
I'm under 40 and this puts immense strain on my eyes. After reading a few paragraphs if I look over to the wall I will see persistent shadow-text in my vision for 30 seconds or more. Very uncomfortable.
It's to the point that if highlighting all the text doesn't mitigate the dark/white contrast, I'll just avoid reading that page entirely.
I’m using Dark Reader extension on sites like that. It may sound absurd to do that, but it actually re-colors bs colorschemes and has quick settings for relative brightness, contrast and saturation, so you can tune it to acceptable levels. Without it my eyes bleed on dark toxic sites as much as yours. It also works good with syntax highlighting, because it uses “dynamic coloring” which guesses best color maps for every page.
Unfortunately a lot of accessibility concerns are of competing interest against each other and people making webpages don't have infinite time to allocate to ensuring they have options for everyone. You might consider just disabling css altogether.
Or get in the habit of using the browsers reader mode. You can style it to your liking and it generally does a great job, at least for me in Firefox where I use it frequently
Use for browsers, you might try to play around with the "Dark Reader" addon. While normally used for setting a universal dark mode, you can also a white mode, either globally or per site.
Biology. I’m in my 50s and get an eye exam every year. According to my optometrist my prescription has been stable for about a decade, but I can definitely tell the difference. My eyes do not focus very fast anymore, my eyes do not adjust going from a light space to a dark space or vice versa very fast anymore. I now have to hold books and my phone away from me to bring them into focus. Same with white text on dark backgrounds. I used to use dark themes for everything, but it just doesn’t work for me anymore. I’ve mentioned these to my optometrist and they say that it happens to everyone with age.
What about yellow text on brown background like [0] ?
I use this combo in my editor to avoid changing to light theme on bright days and dark theme at night.
Not that old, but I have astigmatism. And every spot of light has a glow around it. Black on white is better because the characters only got thinner. But light characters on dark background is like having tiny projectors and the shapes are not recognizable at a glance.
Man, everyone hates bright lights on tall SUVs except their owners, who are blissfully unconcerned with the trail of migraines they leave in their wake.
The web says that astigmatism leads to glowing, blurry text when it is shown white on dark. The iris narrowing in bright conditions seems to reduce the effect, while the widened iris in low light seems to worsen it.
I agree the scrollbar is dying and becoming unusable. A scroll bar needs to be:
- Wide, clearly visible and easy to hit.
- Proportional sized, indicating how much of the document is currently in view.
- Movable parts should have features indicating friction, as opposed to slippy-ness.
- With arrow buttons, placed together instead of at opposing ends.
- It should change colors to indicate hover and mouse press.
Of the classic scrollbars shown here https://scrollbars.matoseb.com/ , the Nextstep one comes closest to having it all and the Mac OS 8 is the overall prettiest. IMHO.
All of this, with the addition that you should have the option of panning in both dimensions from a single toolbar widget if the document supports it. I've seen that implemented but never with the arrow buttons together rather than at opposite ends.
> - Wide, clearly visible and easy to hit.
Way back in the day, there used to be a field of study called Human Computer Interaction, where people would learn about things like Fitt's Law, and the fact that the edges of the screen were particularly valuable because they effectively had infinite size.
I am on OS X. For my sins, I have MS Teams open. It is hard aligned to the right of the screen. On the right edge of the window is a scroll bar. I shove my mouse to the right, I click to grab the scrollbar (which has conveniently expanded on mouse-over, and now extends to the edge pixels), and... the entire window gets dragged.
I don't know how we have ended up here. There was a period when people were actively researching what made good user interfaces, and it was feeding in to end user experiences, and it showed.
The Windows Start button is/was placed in the bottom left so users could just flick the cursor to the bottom left corner, no need for precise motor control. Could be moved to the top left or right as desired, too.
All that went away with Windows 11, though thankfully someone clearly had enough power to say "You are taking my bottom left Start button over my dead body.".
> Way back in the day, there used to be a field of study called Human Computer Interaction, where people would learn about things like Fitt's Law...
Which, despite the name, is not a law and should not be called such. I've had so many people claim that I'm wrong for disliking the global menu in MacOS because "Fitt's Law" says it's better. And yet, it's still less usable for me despite the claim that this "law" makes or what Apple's defenders say.
If people's subjective preference can be contrary to what Fitt observed, then it's not a law. A law is something that is true for everyone, it is not a matter of opinion. I have no doubt that this is a useful maxim, but it's not a law and I really wish it wasn't named so poorly.
The law is that the thing at the corner of the screen has the fastest seek time, because it has the largest available area. That's not subjective.
You might not agree with what is occupying that spot but that says nothing about the physiology of selecting it.
I certainly agree that the global menu, being almost never used day to day, is a stunningly pointless use of that space. It ought to have been handed over to foreground applications, exactly because of Fitt's Law.
With arrow buttons, placed together instead of at opposing ends.
Huh, I've never really thought about that but it does make sense to have them next to each other instead of at the opposite ends. Wonder why it's so rare to see that, apart from "it's always been this way".
I have to assume the natural question is then "which end?" If you put them at the top and the left of vertical and horizontal bars, they're inconveniently far apart when you want to pan different dimensions. If you put them bottom and right, it's very easy to move the window so they're off the screen entirely. But then that's a problem you have with split controls anyway. It feels like there is no good choice here, and split is on average least bad.
> designers do not sit with non technical users to conduct usability testing
It wouldn’t matter if they did. I’ve been a web developer for nearly 25 years and have worked with many many designers. You know they care about? Pixel perfect layouts that match their “vision”. UX isn’t even a fleeting thought.
While not all designers are like this, many are. I've seen designers come out of actual UX studies involving real users, saying "Well, while this is one data point, we think those users are wrong." There's no getting through.
Regarding minimaps: I thought, these were great, too, but then I discovered that I find them rather irritating and that I actually don't use them. I guess, this is, because you have to maintain a mental mapping of that document and how it is organized, already, for the minimap to make sense. But, then, referring to that visual mapping of what you're already mapping in a much more abstract sense, seems – at least for me – to be rather adding friction than being helpful.
The scrollbar, being both highly abstract and visceral/tangible at the same time, is hard to beat.
Agreed. Once a file/page reaches a certain size, I find minimaps to be a mostly unintelligible blob. Maybe it'd be solved by just making it bigger, but then we push into the problems of screen real estate and what not.
I think, it's also, instead of just approximating the position in the document, we now feel compelled to translate our mental image to a visual one and to match this with the rendering in the minimap. Which introduces an additional step of dis-abstraction, and what is essentially abstract (text) now has to be treated like a material object, as well. Moreover, if the overall structure isn't as varied as it may be the case with code, it doesn't provide much significance, anyways.
(It may depend on where you're located on the aphantasia to photographic memory continuum, though. I'm definitely closer to the former than to the latter extreme.)
The best part is that every website now feels it necessary to completely override the unusable skinny built-in browser scroll bar, then add a useless, non-interactive horizontal progress bar at the top of the screen instead.
As I'd commented on the Fediverse thread discussing this post, the distraction of having a horizontal animation responding to a vertical movement is absolutely maddening.
On desktops, I'll remove that horizontal bar via uBlock Origin's element remover or a Stylus CSS rule.
I know it feels clunky (perhaps) aesthetically to have thick scroll bars and thick sliding dividers and thick window borders, but I would choose ergnomics over prettiness.
Even as a normal human with good motor control, good equipment, and decent eyesight, I routinely encounter situations (macOS) where I have to carefully move the pointer back and forth across a region where I know the "line" (divider, border) is so I get the opportunity to move something.
Another related problem is the overloading of the title bar. In the past, the title bar was always there for each window. Grabbing and moving a window was very easy. Now, many apps try to move menus and other controls into that space, leaving the user to find the few pixels here or there which can be grabbed to move the window (rather than taking some action within the app). Some apps allow you to revert to normal title bar (thank you Firefox), but some don't. And true, you gain some valuable screen space from what would normally be a wasted big area of title bar, but the tradeoff sucks when you need to move something.
> Ok so a lot of folks have fine motor control problems.
The scrollbar is not there so you can use it to change your scroll position. It's there solely for the purpose of visualizing what your current scroll position is. Every device and every application should have a way to scroll that does not rely on the assumption that people will hunt out some narrow area of their screen and hold it while dragging to scroll. That's just a horrible interaction that never should have been considered in any setting.
At least that's my opinion.
I am going to disagree strongly there. The alternative way to scroll is, in most cases, the mouse scroll wheel. And that gets tedious (and strenuous for some people) really fast when the document gets large. I absolutely need the ability to use the scroll bar to change my position in a document.
I can't remember the last time I had to resort to manually scrolling by clicking the scrollbar
In order of use, I scroll one click, middle-click and drag (Best for constant movement), use the dedicated keys, use arrows, or use vimium bindings (Which I try to avoid, because I'd prefer to not get used to non-CUA keybindings on programs I might have to use in other computers)
One Alternative would be autoscroll. Ie clicking with the middle mouse button and then moving the mouse up or down. The distance and direction between the click position and the cursor controls the scroll speed.
But it feels like this is primarily a windows thing for some reason. Firefox supports it cross platform but that's about it
You're right that clicking and dragging the indicator part of the scroll bar to scroll around is silly and nobody really ever did it.
But the scroll bar isn't just for visualization and it was commonly used to jump around the page - you'd click in the "negative space" on the scroll bar, which would then scroll the page to that position.
I drag the indicator when scanning a document. It's only a single click, and then I move the mouse to part of the document I'm interested in (and then can keep moving it if I want to look at another part). If I was to do this with negative space it would be multiple clicks (and even then negative space clicks tend to jump to the position, whereas a drag will show everything in between).
Though I find it's more common for a negative space click to just move a page and not jump to the location, so maybe that's influenced me.
Add the newest Steam UI to the list of shame, not only for having skinny scrollbars, but also for having the 5px wide hitbox end several pixels short so that slamming the mouse against the right edge of the maximized window misses the scrollbar.
This is why, in the good old days, UI folks got all worked up when programmers reimplemented basic controls, because they always did it half-assed and could not connect into all the other parts of the OS that users can configure, or have connections to accessibility software, etc...
I remember some horrors in my lifetime: WinAMP skins, browser skins, people recklessly ignoring all of the wisdom collected on asktog.com...
But the horrors became so frequent that we all became numb, and now even Windows ships with more UI disarray than could have been imagined in the early 2000s...
The hardest part of becoming a greybeard is realizing what could have been, that we have abandoned.
Actually, I don't really mind the scrollbar behavior in Steam, as the UNCHANGEABLE font size is so friggin tiny I can't see anything anyway (no, the "scale text" option does not fix this for me).
They're getting less usable as well. One of the mice I use daily has had a broken scroll wheel for a few months now and I've been too lazy and/or busy to replace it, and it's shocking how broken scroll bars are if you try to actually use them. Infinitely scrolling content and content that resizes itself when it loads in very frequently combine in ways that make the scroll bar jump around in ways that make navigation very difficult.
I've bookmarked this post and have had this frustration for a long time. I implemented the Firefox fixes and have a nice, rectangular scrollbar that my old eyes can actually see.
Only thing it took a while to figure out was setting the "Always On" part. I struggled to find that within Firefox and then realized it was an OS setting. I found it in Preferences on my Macbook and now, my FF life is markedly better.
Chalk me up as preferring the older 3D style interfaces with the skeuomorphic features. I have to use Windows 10 at work and the new flat interfaces with thin or invisible elements are hell.
I think too many people with OCD are getting into tech, I’d have nothing against them if they realised they have an issue and would not destroy everything around them, like remove bad looking but essential for usability elements or having to align vertically everything making it a challenge to understand what line is what, that said, I curse you guys, unless you get help or get out of usability/ coding
My fellow readers, do not succumb to the heresies expressed in that blog post! Clearly, the future is in eye tracking data streamed to the cloud, where an AI will predict the scroll target, to be revealed by the application. Most cheerfully, this will not only be a subscription based service, but will also solve the problems with ad attention. ;-)
Scrollbars are becoming indeed a problem, but what about struggling to find the right spot on a window corner/border to grab in order to resize it? Did they test how much of a hassle can be to grab a two pixel wide area on a 13" screen laptop? I can understand they need real estate to be used for actual content, but seriously how hard would be to have a small "handle" that protrudes out of the closest window corner/border when the pointer is like less than 10 pixel away from that, and disappears after say 1 second that it moved away?
It’s very funny and strange that one of the defining characteristics of modern GUI design is that everything is very large. Lots of whitespace, buttons twice as tall as the cursor, windows that have to be maximized in order to see more than seven lines of content, and yet the scrollbars are the size of a toothpick.
And the entire content area for some no-longer-relevant reason is maxed at 1000px. It's like the designers got stuck with bootstrap from 2014 and can't seem to get unstuck.
I'm generally not a person to die on hills like this, but the disappearing scroll bars on macOS have really been a point of significant contention for me. Sometimes I just want to click and drag a scrollbar. And it seemingly takes an act of god to get it to appear (you can't just hover, you have to two-finger scroll first). And then in order to get it to stay, you need to keep your mouse horizontally positioned on the scroll bar. If you click+drag and your mouse moves away from the scroll bar area and release the mouse button (even if your mouse does not move), the scroll bar disappears.
In my perfect world, the scroll bar would appear if you hover in the right spot for long enough. And if you drag it and release, it would stick around until 1.) you move the cursor >0 pixels and 2.) a reasonable timeout passes, or it's obvious you've moving your cursor away from the scroll bar.
I'll note that Slack's desktop app _just works_ the way I wish scroll bars would work (it just appears and stays if your cursor is in the content area), though I think it's because they implement them manually.
I've found most macOS software is actually pretty good at respecting the scrollbar visibility preference, not that we should expect anything less given how often we are told that the Mac is where you go for good UX, and that software developers for it know what they're doing.
(Acrobat recently got this wrong, and I was hearted to discover there's a fairly active thread an the Adobe forums about how much of a fuckup this is. I am not alone! I added a comment or two, of course)
>And it seemingly takes an act of god to get it to appear (you can't just hover, you have to two-finger scroll first)
It's supposed to show up by just doing a two finger tap. It works in native apps. For some reason chrome doesn't respect this though, I think because maybe they reimplement some part of it on their own.
Looking from web browser perspective, scrollbar should be part of browser UI ("chrome" if you will), but not webpage. Sure, some designers love that they can customize colors of scrollbar and change its width, but it's a huge accessibility problem, even for people without disabilities. And when clicking on scrollbar counts as interaction with website and starts autoplaying video somewhere in the background, you know that someone messed up. With screens getting wider, scrollbars width should increase proportionally.
I agree mostly; one of the tricky areas is pages with split layout containers that are individually scrollable, where it doesn't make sense to use the page/body level scrollbar to control that content. Another common case are smaller widgets or modules that have their own scrolling context.
The other half of the issue is that native scrollbars, especially on Windows, are a huge eyesore in those scenarios because they disrupt the layout or conflict heavily with the app or site's branding in some way. Take any eyesore in real life and it produces the same psychological effect — scrollbars too often feel like they don't belong.
Of course, many here (and myself) would argue that usability is far more important than branding.
I think a happy medium would be to enforce a certain size and stay with fundamental usability principles, while allowing customization of colors / arrow images at most, probably.
A good, usability-focused designer alongside their engineers would keep these things in mind, ideally, but alas — a great many experiences are driven by amateur-level professionals who are still learning about these blind spots, and favor aesthetics over functionality.
Went to the HTTP RFC 9110 to check on the proper meaning of a HTTP response [0]. The list of sections on the right is scrollable, but there's no indication that it is, and no scroll bar unless you click on it and try to scroll it. The page as it loads on my screen looks like it stops at section 4.2.1. and there's no indication that there are more sections. Took me a few moments to work out how to get to what I needed. Which is bad design, right? If you're confusing the user to make some aesthetic "improvement" then that's bad. I don't even know that removing scroll bars is an aesthetic improvement. Why do this?
I defence of that web site, I do see a vertical scroll bar for this section. But I'm also one of those funny individuals who check that "Show scrollbars" is set to "always" in the system preferences, after each update. (On macOS, it's not the default, but it's still an option.)
At least on macOS, this is a gloabal system setting, which also applies to any web browsers. (I can't speak on the subject of other OSes/GUIs, though.)
I don’t think this is UX/UI designers not knowing/caring about accessibility. It’s their bosses saying the tiny scrollbar is preferred because it sells 20% more due to the aesthetics being better. Who cares about the 0.5% of disabled people that weren’t likely to buy your app anyway.
Same for browsers, FF may care about accessibility, but they won’t lose the browser wars (bit of a moot point by now, but eh) because most people like Chrome’s tiny scrollbars.
> It’s their bosses saying the tiny scrollbar is preferred because it sells 20% more due to the aesthetics being better.
I think we need to stop shifting blame to "the bosses." The bosses are (generally) not getting involved with the minutia of how scroll bars work or what color they are, or how accessible they are. This problem falls squarely in the lap of 1. the UX designers who think scrollbars are icky and get in the way of their pure vision of the app and 2. the developers who don't push back on the madness.
Yes, I've had "bosses" who annoyingly micromanaged the UX of our product, and yes, it's irritating when that happens. But I highly doubt it's the general case. Designers have run amok in the never-ending quest to fill out their portfolios, and accessible things like thick scrollbars, large fonts, contrasting bold colors, and an information architecture that lends itself to screen readers get in the way of their sterile, minimalist aesthetic.
Perhaps, but I think a large part of this is designers assuming the scrollbar isn't actually for scrolling anymore, and treating is as simply a position indicator. Which for users who exlusively use a trackpad/touchscreen/scrollwheel for scrolling, it is. Ignorance tends to be the largest cause of accessibility issues.
It would be extremely useful for touchscreen users as well, a scrollbar lets you scroll a whole document (and go near specific points) in an instant instead of swiping like an idiot for a minute.
And for trackpad and scrollwheel users it's the same, I absolutely adore chiral scrolling but you still need to use the scrollbars sometimes (or often).
All these things only replace the arrows of scrollbars (enhancing them with some speed control), but the arrows are just a tiny and stupid aspect of scrollbars (I don't remember ever seeing someone using them).
Exactly, a current UI would be far from optimal on 30 years old hardware. Furthermore it would look alien, not in a good way, even if Windows 10 would probably feel old fashioned to us in 1995, because of the flat look that was forced when hardware was more limited in the 80s. It's almost a Windows 2 with a few more colors and less borders. Check this site for screenshots https://guidebookgallery.org/screenshots/win203
It would be more sensible to show them a current Windows. Or Windows 7. Or a Mac with visible scroll bars. Or ask them if they would like Windows 11 more with the scroll bar size setting removed.
But it's not just disabled people. I'm not disabled, but curse the scrollbar and window frame situation every day. Their minimization makes actually using the UI a greater pain in the ass.
My pet peeve is Azure Log Analytics. The query results window has a completely non-standard scrollbar setup, even compared to the rest of Azure.
The default result set shows 100% useless columns, all of which have the same constant values repeated for each row. E.g.: "subcription id", "location", etc...
The useful values are invariably off-to-the-right somewhere, so you have to scroll horizontally. Which you can. With a 2-pixel thick horizontal scrollbar that fades out and disappears permanently after about a second.
You have to be quick, otherwise you miss your chance to see your data.
It would be hilarious if it didn't cost me 50 cents each time I had to re-run a query so I can have another try at clicking the vanishing scrollbar before it's too late.
“X” that sometimes means back and sometimes means several steps back.
… and the browser back button sometimes does what you expect, and sometimes not.
“X” stacked directly under one of the aforementioned “X”s, which means close, but the other “X” is back, all on a flat ui with the same background and no border.
And yeah, the log analytics that feel like some alien intruder or parasite on the host UI.
I still prefer it over AWS, which has separate portals for each product in each region. All of which show random internal identifiers instead of human-readable text. None of which support a folder structure, a technology invented before the advent of computers.
Every cloud “dashboard” is total shit (there are whole businesses that do nothing but provide a better UI for them, that’s how bad they are) but AWS does indeed take the prize for worst.
The general degradation of GUI usability over the past 10-15 years is honestly the most depressing development in computing. It’s a constant fight to keep things halfway usable, and it seems to be getting worse with each major (and sometimes minor) software update.
The really depressing thing is that GUI usability regression isn't the most depressing development in modern computing. Allow me to gesture broadly at literally everything.
Yeah, I have to disagree, I wrote this deliberately. I could live much better with the other developments if at least we had intuitive, consistent, efficient, native GUIs.
For chrome and firefox there is an extension I use that helps dramatically. "Scrollbar Anywhere".
The way I have it set up, I can right click and drag anywhere on "whitespace" on the page and scroll in any direction I drag the mouse. You can set it up to go as fast or slow as you like relative to the amount you move your mouse, and you can even have it "glide" when you let up on the mouse if you like.
While we're at it, if there is a scrollbar and i click the middle, I want to go to the middle of thr content, not move up or down one page closer to the middle.
Large, visible, and provides room for metadata like selected lines, errors, etc. I want them practically everywhere, even when they don't seem to make sense - widescreen monitors mean there's more than enough wasted horizonal space, it wasn't doing anything useful anyway.
I hate to be That Guy, but I wonder if there is any USA case history precedent where the Americans With Disabilities Act (ADA) was used to successfully 'motivate' companies to fix inaccessible software? It seems like an obvious fit. Clearly, The Market is failing to adequately solve the problem.
On one hand, no scrollbars, so when you're scrolling, you have no idea how far 'down' you are... on the other, 80% of the screen is filled with useless junk, ads, sidebars, app menu bars, search bars, address bars, status bars, side panels... but not enough space for a scrollbar.
I have a lot of respect for designers. What they do is very important. I think we should all work together to help them do it better.
For instance, let's rethink the API's they use. Imagine a sleek new minimal API. Now imagine it even more minimal. Keep repeating "minimalism" the word "handcuffs" comes to mind. Think how happy they'll be once we've removed all of that clutter.
Let's build that API for them so they can really just design their hearts out.
> The simple fact that these skinny scroll bars exist are evidence that designers do not sit with non technical users to conduct usability testing.
This kind of hits one of the superficial symptoms of the problem, but I think it’s a much more fundamental problem than this.
I’m sure we’re all familiar with the type of programmer who’s incredibly productive and cranks out features like nobody’s business, but has little knowledge of, and even less respect for, computer science fundamentals. Short term these are great for getting an MVP out the door, but they become a liability when you need to scale. When stakeholders see their short-term productivity and miss the long term implications of their work, you breed a very particular technical culture.
The UX design world has the exact same problem. Designers who come from a pure graphics design background and no experience with, or respect for, decades of HCI research. They produce beautiful, slick designs that dazzle their stakeholders, while not accounting for the actual interactive component of their work.
Or, in short — the same attitude that leads to developers saying “premature optimisation is the root of all evil” in response to any performance concerns is what causes designers to produce this sort of work.
So what? Next thing you know, people are gonna be crying for "accessibility features", like "being able to use computers if you have a disability". Or even worse, normal people asking for changes, just because they want a different (they say "better", but whatever) experience as a user! Who cares if a few million people have a hard time using my app? Works fine for me! And that's what matters.
Scrollbars are terrible, but it's crazy that instead of improving them they just got rid of them, like saying "well let's pretend this problem doesn't exist".
When you have a huge page, the drag island becomes miniscule. They at least set a minimum height for it, but you still need to carefully hunt for it.
When you have a huge page, dragging the island jumps a huge amount, sometimes multiple pages.
Scrolling gradually requires keeping your finger firmly on the mouse button which screws up your hand muscles.
When infinite scrolling the cursor keeps smacking into the bottom of the screen, or loading new content causes the absolute scroll to suddenly jump pages.
Removing the island, having scroll non-linear based on movement distance, double clicking for absolute movement, an interaction for continuous scrolling (right click?), etc. Or maybe doing something like youtube does with the video scrollbar, pulling up for zoom in etc. There's a million things to try here... and what they do try is sweeping it under the rug.
Mind that continuous scrolling was a function of the scroll buttons (arrow buttons), which were axed in favor of visual simplicity. You just kept them pressed for continuous scrolling. On macOS, absolute scrolling is still Option-clicking on the scroll track, while simple clicking is paging. (I don't know, if this is still a thing on other GUIs, though, but I hope it is.) The size of the scroll thumb gives an indication for what a screen size is in relation to the document and what amount of movement is required when dragging it. (This was a noteworthy contribution by Windows, while it was also arguably bad about minimum sizes.)
Most of this had been addressed by the original design. Loading additional content will be always a problem, though, since it necessarily breaks any spacial relations already established. (But this is really a problem of infinite-scroll designs, which simply do not match the established UI controls, without caring to provide any suitable replacement or work around or even suggesting a metaphor for this. Which, in fairness, is to be expected for what is essentially a marketing instrument.) Scroll wheels and trackpad gestures will be arguably superior for continuous scrolling for most users, but cover continuous scrolling only and none of the other options provided by the scrollbar (especially, a fully implemented one.) – There's much to do in order to catch up.
My god... I've been using iOS for ten years and your comment just made me realize you can actually press onto and grab the razor-thin scrollbar that disappears about a second of idle time. What a farce.
That actually gives me a reverse accessibility problem on iOS, because I will try to drag-scroll at the edge of the screen but accidentally grab the scrollbar, which goes in the opposite direction.
The unprecedented success of the iPhone is mostly to blame for all of this madness. Too much success can lead to missteps. When you’re selling gazillions of hammers, everything starts to look like nails.
Hiding most of the UI was genius for such a tiny screen with a simple and very limited OS. People loved it. It’s was tremendously empowering, they could finally own a computing device they can understand or at least not screw up.
But a computer has a much larger display and does a lot more. It’s OK for it to show more UI because there are simply more features.
macOS these days is full of tiny things that only appear on hover that would have brought Mr Fitts or even the original Apple Human Interface Guidelines team to tears.
An airplane dashboard doesn’t need to look clean, it should be useful to trained professionals.
It's definitely somewhat annoying when they don't exist. For example I had to build a relatively simple feature that required a user scroll down, and its not completely obvious sometimes that you should do so due to the lack of scroll bars. So I had to force them on.
Worse is when UI designers intentionally put a delay to scrolling to give it an elastic band feeling. Bro I just want to scroll down, not wait 400 milliseconds for your stupid thing to respond. Why can't they just minimize latency and make things intuitive. No rows of indecipherable icons that we are forced to memorize. No having to hover our mouse over each icon and waiting 1.5 seconds for the text to explain what it is that each icon is doing. Put plain text there, or make it so the tooltip immediately displays. We are people, we use 10s of pieces of software, we can't memorize 10*10 different icons.
It's ironic: 20+ years ago, UI software had FAR more configurability and interesting usability settings like the one you mention here. Yet computers were FAR slower in CPU speed (and memory size too).
Now, computers are infinitely faster, with far more memory to play with, yet software usability has gotten much worse, as functionality has been removed, which logically should result in simpler, faster software, but instead it's the opposite: software has all become dog-slow, even though it doesn't DO nearly as much as the software a quarter-century ago.
Ello has been down for months now with issues at TalentHouse AG (apparently in or near bankruptcy), and archives of that site at both the Wayback Machine and the (usually successful) Archive Today ... show no actual content best I can tell.
Getting rid of scroll bars was something that the guys with infinity scarves stuck us with, right around the time they started doing 200-weight fonts, and removing all the standard web components we had standardized on since the early 90s. I understand designers need freedom, but a lot of the changes in the last 10-15 years have been focused on "pretty" because the tech finally allowed things to be pretty. I hope the pendulum will start swing back towards usability.
Apple's been on a mission to kill the scroll bar since 2015 or so. Others followed suit. (Here's an article from 2016 explaining how to re-enable scroll bars for all the reasons people stated in this post https://www.howtogeek.com/207749/how-to-enable-scroll-bars-i... )
Preach! I complained about this a while back, noting that even when we had a 512x342 screen on the original Mac, there were chunky usable scrollbars... but now with 4k screens, we can't spare the pixels! http://jfloren.net/b/2022/7/15/0
Maybe the solution is to improve the accessibility tools, not hack scrollbars. Assistive technologies should not just try to emulate the computer, but try to also use their specific strengths. For example, shouldn't eyetrackers be able to use blink or eyeroll gestures, or have a virtual "scroll down" "button" (focus target) underneath the screen?
> Maybe the solution is to improve the accessibility tools, not hack scrollbars
I doubt you meant it, but your post reads like you're advocating less-abled people purchase (often expensive) eye-trackers instead of making scrollbars generally more usable for everyone.
The article specifically mentions eye tracker users as an affected group.
Everybody who tries to actually use the scroll bar as a handle (rather than a non-interactive position indicator for use with the other scrolling methods) would have a problem, because they're not really meant to be used that way and most people don't use them that way (which is why they keep shrinking).
Unfortunately the article doesn't mention how people who don't use an eye tracker, can't use a scroll wheel/touchscreen-dragging (or adapted replacement for either!), but could use a "normal-sized" scrollbar, are actually interacting with their devices. I imagine that overlapping group to be rather small.
> ...how do you think people scrolled documents in Word before mice had scroll-wheels?
With the page up and page down keys, mostly, though you could drag the scrollbar.
Not sure about Word, but I believe many apps had a keyboard + mouse combination that enabled scrolling with any mouse movement without interacting with the scollbar, as well.
Very painfully. Have you ever noticed how fountains in cities are increasingly impractical to fill water jugs from? (Just because eons ago it was meant to be used that way, doesn't mean it still is.)
> Naturally the script has to override flatpak separately because heaven forbid my flatpak applications look the way I themed my system.
Triggered.
There's a bug currently where you have to set QT_AUTO_SCREEN_SCALE_FACTOR on QT apps for Hidpi. Trying to get it set on flatpak apps though is an absolute nightmare, so all my flatpak apps are really small on my screen.
I'm imagining similar post about the removal of dedicated copy/paste buttons. "But I don't want to use the keyboard to copy"
> Ok so a lot of folks have fine motor control problems
That's what accessibility settings are for. These aren't an afterthought. Good interfaces are designed with them in mind.
> somewhat inaccurate pointing devices like eye-trackers [...]
voice/sound controlling their computer
Desktop interfaces aren't a good fit for these modalities anyways, and I don't want a 1% use-case detracting from my ui. If you designed an interface from the ground up for eye-tracking, it probably also wouldn't have any scrollbars (see Apple Vision)
A bigger problem are signifiers that something can be scrolled. Mostly it's self-evident that there's more content outside the viewport, but sometimes you run into edge-cases: iOS used to have some bad spots in the share sheet and Apple Music.
whenever you try to view an image in the browser it no longer allows you to go to the image link and instead redirects you to their wonky image viewer (why???)
i can't scroll horizontally, i can't use the arrow keys to scroll at all, nor page up/down or home/end
i can only scroll using my mouse, and if you need to scroll horizontally you have to manage to find their microscopic scroll bar that sits above an aesthetically displeasing footer to scroll sideways
which is infuriating for large images
and its zooming dysfunctional. on normal websites where you can view large images you can click to the location you want to zoom in, and it ballpark brings you to the right place.
not reddit though. wherever you click zoom it always zooms to the top of the image and its up to you to scroll the entire image on your own
Imho scroolbars are unneeded on computers with properly made keyboards. People are desperate to use scrollbars only when the PgUp, PgDown, Home and End keys are not quick to reach.
This is the real battle to fight for. Vote with your money and stop buying computers with stupid keyboards.
Scroll bars are a visual indication of how much content there is off screen/window and where I am in it. There are more uses to it than just dragging or clicking it with the mouse.
I can count the amount of non technical people I have met who use the navigation keys, on a single hand, and every single one of them has a finance related job.
Even most devs, particularly younger devs, never use navigation keys.
Bad scrollbars (and other UI elements) are all symptoms of the origin of the web as a document/form platform and subsequent attempts to use it for interactive apps. This has has popularized some really useful paradigms in applications (back buttons, breadcrumbs), but the lack of capable standardized controls has been painful. I'm hopeful we'll eventually claw our way out of these dark ages, but there's a lot of maturing that needs to happen in web UI designers for that to happen. It's not a lack of skill on their part, but rather a lack of taste. They simply don't know how responsive and useful an interface could be because they've grown up on shitty interfaces.
Even though I hardly use scroll bars by grabbing them with the mouse cursor, I still want them around as visual hint what part of the page is currently shown (first thing I do on a new Mac is to enable scrollbars to be always visible).
This article made me realize I can never even see the scroll position in Safari, especially on a dark page like that one. And I've gotten used to never seeing it. I don't even look anymore.
Safari's scrollbar on that page is white on white. ???
I filed a Radar for this years ago, rdar://problem/46773896. Doubt anyone looked at it. I am confused how nobody seems to have also ran into this considering the WebKit homepage demonstrates this quite well.
Can someone tell me how to make the Mac os native scroll bar wider? It's infuriating trying to grab the little skinny bar. I'm asking here because it seems related and hopefully someone reading this works at Apple or has a solution.
The whole "modern" design has decided to to tell big fat Fuck You to any semblance of ergonomics and sanity. Having to scroll big empty screen to see anything even remotely useful is the apotheosis. Total disrespect of the users.
Regarding quickly scrolling around in a large document, there is also scroll wheel acceleration, i.e. the users finger scroll speed is not just a linear function mapped onto the software scroll speed but rather it can accelerate.
MacOS, iOS and Android have this anyway, and a few custom software as well.
I implemented a cross platform user-space variant of this, to get mouse scroll wheel acceleration. You can even use this in addition to the native scroll wheel acceleration on MacOS.
In addition to all the crap already mentioned, I would add the scrollbars of vscode which are horrendous. I never know where I am, where the "holder" of the scrollbar is and where I can scroll.
I hate most auto-hide scrollbars. Something I didn't see mentioned in TFA was the fact that I mostly use scrollbars as a read-only tool; sure they are good for the occasional scrubbing through content, but mostly I scroll with the keyboard or scroll-wheel.
Glancing to the side and seeing both how much scrolling there is (by the size of the scrollbar grab target[1]) and how much space I have to go (by the position of the scrollbar grab target is something I use all the freaking time.
1: This is a relatively modern feature. The first decade or so of me using scrollbars they were a fixed size)
In Firefox, I also add the following to userChrome.css to make the scroll thumb high-contrast versus the scroll bar itself. The default Firefox colors are, in my opinion, too low-contrast.
I was going to add to the complaints but realized I always scroll with touchpad or drag up/down on my phone. Maybe scroll bars are kind of obsolete except for visual clues.
I also rarely click the scrollbar (except when "jumping to approximate absolute position" in a long document), however, I always look at it to get the current position/document length.
Therefore, my usual problem is terrible contrast of scrollbars. I have modified GTK and Qt themes to show dark blue slider on a light gray background, but only about half of the apps respect it.
I'm not a heavy mobile user and every time I do need to scroll a long page on mobile, I absolutely hate the experience. I cannot easily jump to a specific point the document. I have to scroll, scroll, scroll, scroll until I reach the paragraph I wanted to read.
The wheel doesn't tell you where you are in the page, which is why scrollbars are good (I don't know that I've ever used them to change position, personally; that's not what they're "for" in my brain).
It's like how Safari has this weird insistence on hiding URLs. Every other browser tooltips or status bars a link you hover over, but no: we're apparently just supposed to click blind. For that matter the address bar doesn't even show you the full URL you're at right now unless you enter edit mode.
Windows 11 has these awful scroll bars - at least on dark mode, that are a hair thin and the bar color is the same almost as the background scroll area and I for the life of me, at under 40 years old, cannot ever figure out /where/ the damn button portion is. Windows 11 also likes to auto-hide scroll bars, and minimize them otherwise which is super weird to me since its supposedly a touch friendly OS (Designed for Surface and Phones as Windows 10X supposedly)
Mobile devices have no visible scrollbar unless you are actively scrolling. This is the look they were going for when they say “touch first”. Users are not expected to use the scroll wheel to jump, or click, or do any maneuver more precise than selecting a block of text.
1. Out of context bare-ui style scroll bars in web textareas and mis-sized iframes in most web browsers.
2. Lack of screen real estate on early, lower resolution mobile devices left designers looking for something better.
On the web design front, scrollbars have been customizable via JS and CSS for some time now. On the mobile front, swiping is fun, but it's really nice when you can just tap about where you want to be in whatever you are viewing.
I wonder how that might work in a virtual environment. Scrollbars are about overflow of content for a given screen+window size; with AR/VR in a way we're working without any screen size, so this UX could take some different form.
For example, Apple's TouchBar - a separate screen - in my opinion offered good scrolling UX and often with minimap support too. I'd like to see an extension of that, i.e. a large unified VR scrollbar.
Seems historically similar to the 70s/80s when automakers started positioning an alternator in such a way that it required removing an axle to replace it.
I also like the small and usually invisible scroll bars that the author is against, like the ones in MacOS. Though I only just realized that I can’t remember the last time I have used one to scroll the page - for me they always act as indicators.
The only scroll bar I ever use as a scroll bar is the one in VS Code, which I absolutely love. Displayed next to a miniature, much taller version of your document, with little bands for highlighted or searched for variables, that scroll bars is, in my opinion, peak design.
One place where hidden scroll bars are a pain - side-scrolling large tables on webpages. It’s not obvious you there’s more data and it’s not intuitive for even technical users to randomly decide to side scroll in that scenario. Excel users are used to this since the scroll bar is always shown, but achieving a nice UX for data/table heavy web applications is not straightforward.
I recently got really frustrated with Firefoxes almost invisibly thin scrollbar. Messed around in about:config and managed to make it fatter, but it made all the scrollbars inside web pages also fatter - not what I wanted at all. I'll try if only changing what was mentioned in the article helps with that.
Bring back reasonably sized scrollbars, please.
I have the opposite problem - I don't want to see scrollbars. I have the touchpad and scrollwheel for that. I don't need anything taking away space from actual content, especially in narrow sidebar menus. A semi-transparent strip visible only while scrolling sounds perfectly fine.
Unfortunately, as the author notes, its not always easy to get this customized.
For those looking to accessibly scroll webpages: if not visually impaired, could I introduce you to the powerful combination of hitting spacebar to scroll down webpages and (in Firefox, at least) hitting the / (forward slash) key to bring up the "quick search" window? From there, you can type in whatever you want to select or bring into focus that you see on the screen. It's a great way to quickly move the focus to another part of the webpage. (If you just hit tab without quick searching first, you'll scroll back up to the last link that was in focus, but by using the / key quick search, you can basically navigate with just your keyboard and the tab key.) While this doesn't help if scrollbars are embedded in a site, if it's a traditional webpage such as the parent article, it works great :)
As a result, while I do have and use a trackpad, I rarely feel like I'm missing scroll bars because I can just hit spacebar (or page down) to scroll down most times. If you need to page up or down on a Mac, it's fn key + up to go up a page, and fn + down to go down a page. Meanwhile cmd + up is like the Home key and cmd + down is the End key, so you can quickly scroll to the top or bottom of a page (or text entry box).
Not to say this isn't a valid article, but there /are/ workarounds to missing visual scrollbars if you absolutely need this functionality.
Admittedly, these workarounds do not work that well on mobile, where I sometimes wish the iOS Contacts app A-Z scroll helper had somehow "caught on" within webpages, to scroll them. We're missing an IDE-style control that helps us "quick scroll" and jump between headings and sections of longer webpages, that's for sure. As the article mentions, I'd be in support of minimaps, but my version of them would look a bit more like the contacts UI, small icons to represent each section that I could click on to navigate to it, and maybe if I swiped left on the right side of the screen it would show me the table of contents in full, with a search box? (Just spit-balling here, but we need more UI innovation and standardization, not less. I really don't think we've found the "perfect" UI, I think we just forget what made old UIs great when we build new UIs... And sometimes we invent greatness: see the Alfred-like Spotlight search box that you can even get on Windows via power tools that lets you quick switch to any application just by typing its name... literally genius, but not something easily available before OS X Lion.)
Could I introduce you to space bar typing space or pressing a button instead? Could I introduce you to someone who uses eye tracking and voice control not having the keyboard dexterity you assumed?
Sorry if it came across that way, everyone’s needs are different. As far as I know both Mac and Windows have preferences to make scrollbars permanent and/or giant, though not every app is as accessible, and webpages can, of course, do whatever they want to some extent. You could probably find a browser and OS combination that is completely customizable to suit, though.
macOS cannot restore scroll bar width or buttons. The article documented the poor accessibility of options for GTK, Qt, and Firefox. It implied Chrome had none. Windows 11 is similar to Firefox. The GTK, Firefox, and Windows options are following the paths of previously removed options. And apps doing what they want was part of the article also.
I find `overflow: scroll` is the worst (fair, it has it's uses), but I'm perplexed by seeing sites that has scrollbars on Windows when content absolutely perfectly fits on the screen.
Regularly do I visit sites that's infested with scrollbars, because every second element have this set - but it looks good on macOS, right?
> And while the bars were shrinking, another feature silently disappeared: buttons to click and hold down to scroll left/right in increments.
This feature still exists, with a larger clickable surface, but with lower affordance – the entire area above and below the scroll bar is a "button" for exactly that purpose.
Not really the same feature. Typically, scroll-bar clicks will scroll by pages, while the little buttons will scroll by lines.
Also: using the little buttons, you can position the cursor once and then click repeatedly for repeated movements, taking you ultimately to the very top or bottom of the page. But when clicking the scroll bar, page movement generally stops as soon as the slider moves underneath your cursor.
However, this area moves the screen a page per click, instead of a line per click. It actually replaces another, now long forgotten scrollbar button found on the Xerox Star and in some Lisa applications.
I can imagine minimaps being useful on graphical content but the author gave the example of it being used dogmatically and unhelpful: code editor. There is nothing to see in "the shape of the code"?
The minimap could (and I usually set this the first thing in vscode) be replaced with actual useful markers like svc changes
Right on point. I don't want to play a sniper game when scrolling things. My hand was getting tired of just trying to grab a scrollbar with a mouse. I did the same thing some time ago, increased min width of GTK and Firefox scrollbars to like 32 or something, my life only got better after that.
This isn't limited to being a problem for those with fine motor control issues.
The wheel has a linear/constant speed, which is a hindrance for very long pages. Also, wheel scroll can behave differently depending on where the mouse cursor is or has focus. The scroll bar has none of these problems.
* Thankfully, Firefox is at least fixable (for now…),
Here, I set the size to 50. I don’t want it this big myself but by goly am I glad I can make it this big anyway.*
Only 50? I can finally have something useful to fill that massive blank space between my content and the edge of the page.
Yes, I agree. Something that has been increasingly bothering me: how hard it is often getting to see the scrollbar, or grab it if you can find it. It's being hidden and minimised. Firefox is quite bad here, and some GTK apps. Frustrating.
Apple started the trend and the web picked it up. But as usual, web had no underlying mechanisms to retain the UX integrity.
Thin scrollbars work by being transparently overlaid over text and grow in size on hover, without bothering the main layout. Browsers can’t do that universally. So what you get is a windows/linux-style static scrollbar of an annoyingly small width. Or a js abomination that partially simulates it and stops working randomly on some mobile devices.
Sure, you may blame designers. But these original Apple scrollbars are actually fine and they look good. The problem is where it was for the last 20+ years: we are using slow and clumsy 40MLoCs as an applications and media platform. They need TENS of years to add support for primitive things like flexible boxing or grid layouts or linear layout constraints (oh wait, no). So there’s a chance we’ll get these overlay scrollbars in 2040+.
> 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??
Gentoo is a great learning exercise, but Arch has me focusing on doing things.
Maybe scrolling itself is the real problem, and bad scrollbar design is just a symptom. Think of how long it took you to get to this comment, and all the other things you could have been doing with your life. Was it worth it?
not just that, but for example my mousewheels sensor is basically unusable (when I scroll either direction its just roulette if it wants to scroll up or down or a mix of both) so I often use the middle mouse click to scroll... ends up not a lot of applications other than browser based ones support this.
Scrollbars were never meant to jump instantly anywhere. They are intended for progressive movement in either direction.
It's a much more recent feature implemented as "minimap" in several IDEs. The difference is that minimap actually renders preview of the content to jump to, so you don't jump blind.
Add: autohide is there because oftentimes it is obvious content can be scrolled by observing it does not fit entirely in the viewbox.
There was nothing wrong with the big wide scroll bar with the nice obvious buttons on either side. Every person involved in their removal needs to be ejected from the software world immediately without prejudice.
Scrollbars getting smaller was something I was noticing subconsciously, but was happening slowly enough that I was getting frog-boiled. Now they're annoying me even more than usual :^).
This occurred to me the other day when I was using notepad. The scroll bar isn’t awful but it’s not great. If someone had motor control problems it would be a massive source of frustration.
With Windows, I can’t tell which app I am looking at especially with browsers which look similar (chrome, edge). It’s hard to tell which application is highlighted in the task bar.
Looking at you, Spotify, with your scroll area so close to the track actions hamburger. God you make managing a playlist such a chore. HMU if you want ideas bruh. I gotz em.
The way I've heard it told GNOME 2 was the cat's meow. I myself only switched to GNU/Linux in 2012, so I missed it. Unity (Ubuntu's erstwhile competitor to GNOME 3) with the HUD was quite the joy to use, though.
Please discuss scrollbars issues while using slakc, Its getting tough to scroll upwards for more chats in history on slak, due to clunky behaviour of scrolls.
>gsettings set org.gnome.desktop.interface overlay-scrolling false
"Don't rely on gsettings. They are private and users are not supposed to be setting them theirselves. Don't complain when they suddenly stop working." -- every gtk dev I've ever talked to.
That said, I've applied those firefox about:config fixes and it's much better. Whew.
gsettings don't just apply to gnome desktop environment. They impact the entire gtk ecosystem of dozens of desktop environments and innumerable gtk programs.
Not sure why anyone wants scrollbars. On Mac they're pretty much universally hidden until you start scrolling. I never use them and hardly notice them which is great. I only scroll with 2 fingers on the track pad. I'm always shocked when I see people interact with a scroll bar directly using the mouse.
is there a way to go to the end of the Reddit or Twitter stream other than spending some long seconds manually swiping down and finding the last remembered post headline?
It's the elimination of window borders. Aside from not being able to differentiate one window from another similarly colored window in the background, it's nearly impossible to click and hold on anything along the edge to resize the window.
It's the overloading of the title bar with so much shit like search boxes and extraneous buttons that a user has almost no place to grip to move the window.
It's the way that tabbing between text boxes either doesn't behave the way you'd expect, or doesn't work at all.
It's all the tooltips that interrupt and litter the interface and, at times, block out things that you are looking at. And 95% of the time, the information provided in these tooltips are redundant or useless.
It's amazing how much damage these cargo-cult UI/UX morons have done in the past ten years. They threw out several decades of usability pioneered by real HID experts for something that looks pretty but doesn't fucking work for a lot of people.
Applications like Postman, Teams (and pretty much all of MSFT's applications these days), Chrome, and Insomnia should be case studies on how to not design user interfaces. They are about as bad as desktop software gets.
The biggest sin is that this would be a non-issue if these things were configurable at the windowing system level and could not be overriden by app developers. But the trend has gone in the opposite direction; instead of providing more configurability, Windows and Gnome/GTK are actually taking away options that have existed before.