It depends on if you like features and configuration options or not. If you don't like features and don't want to be able to set any configuration options, go for Gnome. Otherwise, KDE.
KDE these days seems like configurability-overload. I appreciate the flexibility but it just feels excessive to me. I think a lot of the options should be tucked away into an equivalent of Firefox's about:config. For me, something like XFCE has the ideal config panel complexity. (Of course, it's all subjective.)
There are ways that the configuration can be done for you, however. There's a "Use Desktop Layout From Theme" that would pull in any layout changes from a theme you install.
When I found out that you could apply a theme and, with a few clicks, have KDE fashioned into something that looks indistinguishable from Unity, I almost switched to it just for the novelty.
KDE could use some work in the design department, with a lot of the menus and UI stuff getting an overhaul, but without actually removing functionality (just moving it around to be more sensible, and not overloading new users with too many options at once).
If you want customization(do not like the default) is safer to use KDE, with Gnome you need extensions and if you check the Gnome reddits and forums you will notice that latest updates broke the popular extensions again.
Also for accessibility reasons, AFAIK Gnome does not show the font size in the default settings app, this is the first thing that i need when I want to install Linux, I put the LiveCD in and then go and enable the zoom and large fonts.
Last I tried, you couldn't type direct paths into the fileselector or the file browser, basically requiring that you mouse your way around. I found that extremely aggravating and went back to KDE almost immediately.
Well, it wasn't advertised, and I argue it's a basic feature of a file picker / browser so it should be very obvious. To be honest, that was just the final straw. I was already unhappy with Gnome for a variety of reasons, that was just one of them.
Having a key sequence is fine. Remembering the key sequence is annoying. Every OS and UI wants to have it different. Modern OSes even get rid of hints and text so people don't even know that a key sequence is available.
This is true, but having just a keyboard shortcut is still a disaster for discoverability. I've used Windows a fair bit and have spent a decade using various Linux DEs; it was only two weeks ago that I discovered Ctrl+L.
It's a wonderful shortcut, useful in browsers and a huge number of other applications, but having buttons and menus and other UI features - which may take up a tiny bit of space and piss off designers who want ultra simple UIs - are what allow people to become good at using software naturally, without having to actively seek out knowledge from the few enlightened ones.
> I've used Windows a fair bit and have spent a decade using various Linux DEs; it was only two weeks ago that I discovered Ctrl+L.
It wasn't available in Explorer until recently.
More specifically: sometimes I have to work with Windows 2008 servers and it is not available there.
I think in older Windows Explorers you could use an F-key or something but I don't work that much with Windows anymore.
> but having buttons and menus and other UI features - which may take up a tiny bit of space and piss off designers who want ultra simple UIs - are what allow people to become good at using software naturally, without having to actively seek out knowledge from the few enlightened ones.
I tvink we agree. I have had a feeling for a while that a lot of the simplification of UX we have done in the last 10 years is more dumbification, and that by removing the learning curve we've also removed the learning.
Do I mean software should be hard to use? No. Use as clear ux as possible but if you are designing power user apps for existing power users (or people that should become power users) don't be afraid to add some configurability.
If I should try to come up with one rule of thumb: if a ux designer is removing options the users want they might be dumbing down the UX instead of simplifying it.
Uhm, which is to say, horrid? Because polish is the last thing that comes to mind with the abomination that KDE is.
The defaults are literally to spam you with notifications because a single fricking file copy finished. Hello! We have SSDs now! No one is fucking waiting for their file copy to finish! If you copy 100 files, no one wants 100 notifications!
No hierarchy, bad sizes, bad padding, no different colours, etc etc. It's just terrible design, as I've said in another comment which has been promptly flagged. Seems like the only proper criticism is to write the code myself to fix it? Because I'm passing, I don't use KDE, and I don't need to write code to see when a design is amateurish.
All the UI and UX in general looks like it's from 2007. I know that it's volunteer work and I don't want to downplay the generosity of those who contributed to it, but I do wonder what's causing the delay in catching up to modern aesthetics. Is it a technical issue? Is there too much bureaucracy to overcome? Is there a shortage of UI and UX developers?
Basically creative people by and large never subscribed to the idea that they should give away their creativity for free. So all the talented designers are employed and happily working on paid projects, especially less established ones get approached with ridiculous low offers all the time.
In short outside of the bizarre software/tech world very few subscribe to the idea that intellectual property should be shared and used freely.
I'd bet that isn't true and that plenty of artists can and do peruse places like deviantart and quickly improve upon someone else's images.
But design for a GUI is more like industrial design. Getting a designer to help improve KDE's look-and-feel would be like squatters in a dilapidated hotel convincing an interior decorator to spruce the place up for fun.
Actually, that sounds like it would be a blast. Perhaps open source people are simply terrible at marketing.
Actually, there are many things I like a lot better about the 2007ish KDE/Oxygen/Breeze style than about more modern approaches like for example Windows 10.
I am not saying it is perfect, it is not, but especially in design terms, 'newer' isn't always better. And in general, I like KDE look very much.
I agree with you that newer interfaces (Windows 10 in particular) are mostly not very good. But I wouldn't say personally that KDE is a good example of 2000s era design. I think Gnome 2 was the absolute pinnacle of PC GUI design, actually. Having instantaneous access to Applications / Places / System (with text labels!) makes so much sense; it's one of the few desktop environments that pretty much any competent computer user could sit down at and immediately use.
Actually, I have my KDE set up to be an approximation of Gnome 2 with a few tweaks that work better for me, but I'm not under any impression that it approaches the immediate usability of Gnome 2.
> Having instantaneous access to Applications / Places / System (with text labels!) makes so much sense
I don't really get that. I suppose I've just never gotten used to it, but I just find that typing the name of the application is so much easier and faster. If I know I want to e.g launch Nautilus, why would I first spend time thinking about whether it's a "place" or an "application" or a part of the "system"? Is my text editor an "educational" program or an "office" program or an "accessory"? Or maybe _that_ is a "system" tool?
I'm actually genuinely curious about this; how is it better to mouse through hierarchical categories and look for an application, instead of just hitting super and typing the first couple of characters? Especially something like KDE's start menu, where you can hit super and start typing, or hit super (or click the icon) and mouse through similar categories to not lose the discoverability benefits of categories.
>I don't really get that. I suppose I've just never gotten used to it, but I just find that typing the name of the application is so much easier and faster. If I know I want to e.g launch Nautilus, why would I first spend time thinking about whether it's a "place" or an "application" or a part of the "system"? Is my text editor an "educational" program or an "office" program or an "accessory"? Or maybe _that_ is a "system" tool?
Well, there are several aspects to this. For an ordinary user, who's just started with the system, they don't even know the file manager is named "nautilus", so they wouldn't know to type it in. The Gnome 2 system aids discovery. Second, a user might not even know they could search for things, or be less comfortable with the keyboard. (Personally I do set things up this way: I have caps lock mapped to KDE's search widget, in the style of Chromebooks. I don't have a "Application" menu at all.)
But maybe most importantly, one usually thinks "I want to look at my pictures", not "I want to open the file manager." In Gnome 2, you would open the Places menu and click Pictures. Which, of course, would open Nautilus already pointed at the correct directory. I have a Places widget with a folder icon on my KDE taskbar for this purpose.
KDE let's me setup myy font sizes, I always use huge fonts and I hate with a passion when devs pixel allign things and break the GUI when I use larger fonts.
I also get exasperated when I get asked to pixel perfect align checkboxes and the labels in web pages, each browser has different checkbox dimensions and styles, and aligning them how the designer wants means a lot of extra styling that will break on users that don't use the defaults(but the designerr wants his pixel aligned stuff on his Mac screen... also rewording messages to align in message boxes or labels
I agree. That said I switched to KDE a few weeks ago and turned into a fan boy during the configuration process. I'm very happy with my current configuration (probably a mix of it actually being good and the Ikea effect). I love that you can just configure almost anything almost exactly to your liking with not all that much effort.
I'm not a huge fan of the default configuration, but the ability to completely change it (with the many panel options) is just amazing (kind of contrary to GNOME).
I think I have figured out at least part of the reason for the spacings always feeling just a bit (or very) off (although I think it's bearable with my theme):
Their Kirigami design support library has a set of pre defined spacings and sizes developers are advised to rigidly stick to. These probably break down in many contexts and might not even be chosen well in the first place.
Another reason might be that KDE doesn't have a design group that is as strong as Gnome's and KDE might not be organized top-down as much (which is probably related to financing).
I think you can fix or mitigate well a lot of the mistakes (Qt is great) by using a good Qt curve theme.
You're correct that KDE's design group isn't organized from the top down. While that might be a problem for a company, I think it's actually a good thing for a community of volunteers because communities need to grow. Everyone has their own lives outside of the work they do for KDE, so it's important that the work load is spread across many people. Anyone can start start designing for KDE, you just need to talk to other KDE designers and submit patches.
I am a fairly new KDE designer who started contributing icons in the spring of 2018 because it was easy and fun.
I apologize, but I don't understand how anything you wrote in any way responds to what I asked, who I asked it to, or could be considered "helpful" in any way.
Please don't inject meaningless flag waving platform propaganda into genuine conversations.
Most GTK based DEs are pretty decent in this regard (GNOME, XFCE, Cinnamon, Budgie). Qt based DEs tend to be more utilitarian and weaker when it comes to design.
On the other hand, one thing they do really well is theme support and configurability. I agree with you that out of the box it's pretty bad and they need an real designer to redo Breeze, but themes are allowed to adjust margins and it's possible to make KDE look really nice.
Personally I find that using Source Sans Pro as the font and the Arc theme for most of the rest works well.
Glad I’m not alone, KDE’s design system (or lack of such) has always been like nails on a chalkboard for me. GNOME takes it a bit too far in the other direction but is more tolerable (especially with a compact theme). XFCE is probably my favorite and because it strikes a great balance between design, readability, and info density.
Xfce is great, and by far my favourite. But it's harder to sell to the average user because it looks so much older than windows and Mac. At least by default
You are not alone. Typography is questionable but the worst parts are padding and alignment.
I have to apologize in advance because it's going to sound like a harsh comment and I know this is volunteer work. I really think that KDE is technically very good. The amount of basic design mistakes however is just unbearable as far as I am concerned.
I mean, let's take a look at this improved notification panel. The closing and setting icons seem slightly lower than the icon on the left of the title. The large hamburger menu icon is actually two pixels on the left of the closing icon. There are 17 pixels of top padding but only 12 pixels of left padding. The large image is one pixel of center (232 pixels on the left, 230 on the right). Every single line of text has a different interline.
And it's same for every application of KDE. They really need a design team to polish it up.
Yeah, this is the real tragedy of KDE, and it's been like this forever: technically, it's fantastic, but the design and polish are really lacking. Instead, all the people good at visual design are working on Gnome, which has technical underpinnings that are crap by comparison.
GNOME does terrible waste of screen real estate so if those "people good at visual design" are responsible for that, i'd rather them stay in GNOME instead of infecting other desktop environments.
Design quality for a UI, ordered easiest to hardest:
1) Information dense
2) Looks good in screenshots
3) Information dense and easy to use
4) Information dense and easy to use and looks good in screenshots
An awful lot of designers stop at 2 because a 4 is way harder to achieve, yet a 2's pretty easy to sell (to "stakeholders"). The result is at least sometimes (sometimes) better than a 1. They may also drop a 3 to a 2, though, and it's really hard to later turn a 2 into a 3 (since it's "uglier", so harder to sell) and a 2 is so different in character from a 4 that it can't even really be considered progress toward it (a 3 might be).
I kind of feel bad about my first comment because I know it takes a lot of time to do these modifications, the new interface is a clear improvement from what was there before and UX wise it's not bad at all.
It's just that every time I use KDE, I have this feeling that great work doesn't shine as it should because of padding, alignment and box in a box in a box issues. I know these adjustments are a pain to make and that it's a very ungrateful work but it shows.
So, thank you to the design team I guess and keep up the good work.
I never could use KDE, because it systematically mistreats words. It's not only typography (few WMs get it right, tbh), but as a language user, I feel uneasy everytime I have to use "Konsole" or "Konqueror".
Does it solve the infinite appending issue though?
Basically, every Slack message would append to the same notification, so after 4-5 notifications, I'd just see old ones and a ... line, making the notifications useless. This also went for the update centre, etc.
Also when notifications decide to just stay on screen indefinitely because of... Reasons.
I've never experienced this. Particularly, I don't experience this with Discord, and Discord is to Slack what a bee is to an ant, which is why I juxtapose it. (And I would ask if it might be something on the Slack side, but you say it happens with everything.)
>Also when notifications decide to just stay on screen indefinitely
Never had this either.
Might have been bugs. If you're on Ubuntu or another point release distro, those kinds of bugs also stick around, IME, because your DE version stays stagnant and the bugs are only fixed in subsequent versions.
On the other hand, I've had it with Slack, Discord, Skype, and most anything that uses the OS notification system; using Ubuntu 14.04 and 16.04. So while it's not KDE, it's certainly something relatable.
I am not sure if there really is a lack. I mean, I think I see what you mean, but given the fact that those are notifications, they are probably meant to feel 'thin' to not attract too much attention.
Another factor might be that those small, focussed screenshots are just not the appropriate format to show how those notifications would look like in the context of a complete desktop.
But in the end, it is also possible that there is an unintended lack of horizontal padding.
Lack of horizontal padding? What padding do you want? There's clear whitespace between everything.
I for one definitely like and want more compact information. I definitely do not like the trend for making information more sparse. Leave that for your blog; business runs on information.
Information density is not hampered by the presence of proper negative space, on the contrary, using the right amount of negative space actually allows for very information dense layouts to still be readable. A good example is a traditional paper phone book
What you define to be proper is likely not what I define to be proper. It's certainly a context-sensitive opinion.
The amount and location depends on what you want to do. Yes, just like a phone book, it needs whitespace. A notification window is not a paper phone book though. A paper phone book typically has a lot less whitespace than we're discussing here, so I'm not even sure if you're agreeing or disagreeing or trying to point something else out.
The most egregious example that stood out to me here is the x-padding between the edge of the notification panel and any text inside the panel is less than the padding between the notification title and the text below it.
The problem isn't that they need more padding, it's that their padding makes no sense. If they made a design decision to have tight padding that would be one thing, but they've got a horizontal layout with tight padding in the horizontal direction and wide padding in the vertical.
here's a really quick-and-dirty edit to even out the padding around each text block to be equal to the line height, which actually tightens up the total density by correcting the too-high line height in the notification text: https://imgur.com/uVcY6Sw
Programmatically, that assumes that that information is available (query a font rasterizer: where is the first pixel? consider also kerning and foreign languages; also query icon: where is the first non-background pixel? consider also theme information where window background might not show the icon well!).
Programmatically, you've un-aligned everything: now the amount of whitespace at the top of the notification is different than the amount of whitespace on the left of the notification.
Sure you might say it looks better. But how much time do you think it would require to make it look "pretty" (for you)? As-is, it's functional and isn't ugly. If you want to improve it, then go contribute to the libraries and applications being used.
As you’re keen to point out regarding everyone else’s comments, that’s your opinion.
Personally I think it’s functional and ugly. They could skip the drop shadow, the subtle transparency, the big icon on the right, the rounded corners, and just draw a white box. As long as they fixed the left padding it would look better than this.
Lots of attention to detail on the wrong things before they have the basics, IMO.
Then the vertical padding should be reduced. As it is, they're inconsistent and looks bad. I really want to like KDE and I used KDE 3, KDE 4 (from 4.8, since every version before that was a shitshow) and Plasma 5 for a long time as a main driver and is still my go-to in my backup laptop; but the visual aspect has always been an utter failure despite the massive improvements made in version 5.
I don't know how to explain. Those look exactly like the KDE notifications I see all the time, but they never looked compressed inside the window on my desktops. Instead, it always bothered me that they were too spacey.
I also noticed that, came here to comment precisely the same thing.
I think the problem might also be exacerbated by the fact that the padding doesn't seem to be scaled based on DPI. (Unlike line height/icon size) So it looks acceptable on a low resolution display, but super close to the edges on a "Retina" display.
When looking at all the screenshots, it’s clear that padding and alignment is neither professional (looks amateur), nor consistent.
If you’re going to rip off Apple, at least do it to Apple quality! Otherwise it’s a cheap knock off. Before people grip that this is free software, would you be happy if KDE’s core components were poorly written? Free or not, we should take pride in our work and put the bar at high quality.
Just har a look at this as well as the fine article. Colors are similar but that's about it in my non designer opinion; even the colors are different as the ones from Apple seems to be somewhat translucent (kind of like Vista IIRC ;-).
Thank god I'm not the only one who is annoyed by that. I remember 1 or 2 years ago when I first installed KDE I thought I messed up the installation because the padding/margins looked so jarring, I was convinced there was something wrong with the UI theme.
And to be honest, I think the notifications area required a bit of optimization. To me, it feels like it is also one of the components that received more optimizations than any other component already, but it also seems to be a little more complex than you would expect on first sight.
> The old notification plasmoid was originally written in 2011
Where "few years" means 8 years? :D By all means rewrite software when it is 8 years old. A lot has changed in that time and the author notes that user expectations have changed, which leads me to the...
...false dichotomy you offered. User expectations and the code base not supporting those is the stated reason.
They don't. What you are probably referring to is how they reimplemented lots of components from kde3 to kde4. This was due to a switch to a new major qt version, at which point porting is a big enough amount of work that you might as well improve things while you are at it.
If KDE was a monolitic project with insufficient tests, those details wouldn't improve. Instead, I'd bet developers would be fast at removing current functionality when unavoidable changes add hard to track bugs.
every few weeks KDE forgets all my panels because of connecting/disconnecting external displays and KDE having horrible multi display support.
KDE is so full of bugs like that its really frustrating, like the task manager forgetting which screen they are on or shift getting stuck when connecting an external keyword, etc. etc.
But I've made my peace with all the bugs being a fact of life, using it for 8+ years.
You are not alone. I really like KDE and enjoy using it mostly but it's full of bugs but every other day I'll attempt to use GNOME and I'm back in no time.
I'll share the display issues - something deep inside calls random() on monitor resolution and position - I'm sure!
I've switched to Arc Dark and Numix Circle Icons and it's not ugly anymore. I don't know why they still ship Breeze with that broken padding by default.
Please stop redesigning things that work (and are debugged) just for the fun of it. I've only just upgraded past the iteration of KDE which decided to steal focus with its notifications, no matter how much I customized kwin's focus stealing prevention. My current iteration's plasma system tray will stop expanding every couple of days. Reproducible on two systems. It's fun.
I mean, how many times can you redesign a logout screen, really?
Not sure why there's so much negativity in this thread. Thanks for spending the time to improve this. I love that KDE can show notifications from my phone via KDE Connect. It's not without its problems (the design could improve, and I'd love support for android quick respond actions if they're possible) but overall seems like great work.
I've been using KDE Connect on my phone with GSConnect in Gnome and I can safely say that KDE connect is what made me boot Linux more than Windows. I've been dual booting for ages but the ability to copy text and send files across the network with a standardised interface made my life a load easier.
107 comments
[ 2.4 ms ] story [ 172 ms ] threadWhen I found out that you could apply a theme and, with a few clicks, have KDE fashioned into something that looks indistinguishable from Unity, I almost switched to it just for the novelty.
https://www.addictivetips.com/ubuntu-linux-tips/make-kde-pla...
Also for accessibility reasons, AFAIK Gnome does not show the font size in the default settings app, this is the first thing that i need when I want to install Linux, I put the LiveCD in and then go and enable the zoom and large fonts.
It's a wonderful shortcut, useful in browsers and a huge number of other applications, but having buttons and menus and other UI features - which may take up a tiny bit of space and piss off designers who want ultra simple UIs - are what allow people to become good at using software naturally, without having to actively seek out knowledge from the few enlightened ones.
It wasn't available in Explorer until recently.
More specifically: sometimes I have to work with Windows 2008 servers and it is not available there.
I think in older Windows Explorers you could use an F-key or something but I don't work that much with Windows anymore.
> but having buttons and menus and other UI features - which may take up a tiny bit of space and piss off designers who want ultra simple UIs - are what allow people to become good at using software naturally, without having to actively seek out knowledge from the few enlightened ones.
I tvink we agree. I have had a feeling for a while that a lot of the simplification of UX we have done in the last 10 years is more dumbification, and that by removing the learning curve we've also removed the learning.
Do I mean software should be hard to use? No. Use as clear ux as possible but if you are designing power user apps for existing power users (or people that should become power users) don't be afraid to add some configurability.
If I should try to come up with one rule of thumb: if a ux designer is removing options the users want they might be dumbing down the UX instead of simplifying it.
We've just officially passed the 15 year mark. It's even become a meme now:
https://wiki.installgentoo.com/index.php/File_Picker_meme
The defaults are literally to spam you with notifications because a single fricking file copy finished. Hello! We have SSDs now! No one is fucking waiting for their file copy to finish! If you copy 100 files, no one wants 100 notifications!
I see that notification pane and aside from color, nothing gives it typographical hierarchy. It just looks odd to me.
In short outside of the bizarre software/tech world very few subscribe to the idea that intellectual property should be shared and used freely.
But design for a GUI is more like industrial design. Getting a designer to help improve KDE's look-and-feel would be like squatters in a dilapidated hotel convincing an interior decorator to spruce the place up for fun.
Actually, that sounds like it would be a blast. Perhaps open source people are simply terrible at marketing.
I am not saying it is perfect, it is not, but especially in design terms, 'newer' isn't always better. And in general, I like KDE look very much.
Actually, I have my KDE set up to be an approximation of Gnome 2 with a few tweaks that work better for me, but I'm not under any impression that it approaches the immediate usability of Gnome 2.
It even looked nice! https://upload.wikimedia.org/wikipedia/commons/f/f7/Ubuntu_8...
I don't really get that. I suppose I've just never gotten used to it, but I just find that typing the name of the application is so much easier and faster. If I know I want to e.g launch Nautilus, why would I first spend time thinking about whether it's a "place" or an "application" or a part of the "system"? Is my text editor an "educational" program or an "office" program or an "accessory"? Or maybe _that_ is a "system" tool?
I'm actually genuinely curious about this; how is it better to mouse through hierarchical categories and look for an application, instead of just hitting super and typing the first couple of characters? Especially something like KDE's start menu, where you can hit super and start typing, or hit super (or click the icon) and mouse through similar categories to not lose the discoverability benefits of categories.
Well, there are several aspects to this. For an ordinary user, who's just started with the system, they don't even know the file manager is named "nautilus", so they wouldn't know to type it in. The Gnome 2 system aids discovery. Second, a user might not even know they could search for things, or be less comfortable with the keyboard. (Personally I do set things up this way: I have caps lock mapped to KDE's search widget, in the style of Chromebooks. I don't have a "Application" menu at all.)
But maybe most importantly, one usually thinks "I want to look at my pictures", not "I want to open the file manager." In Gnome 2, you would open the Places menu and click Pictures. Which, of course, would open Nautilus already pointed at the correct directory. I have a Places widget with a folder icon on my KDE taskbar for this purpose.
Yeah. That probably KDE's most compelling feature.
KDE UI never gets Typography, spacing/padding, or alignment correct.
It's a shame because they are so close on most other criteria.
I also get exasperated when I get asked to pixel perfect align checkboxes and the labels in web pages, each browser has different checkbox dimensions and styles, and aligning them how the designer wants means a lot of extra styling that will break on users that don't use the defaults(but the designerr wants his pixel aligned stuff on his Mac screen... also rewording messages to align in message boxes or labels
I'm not a huge fan of the default configuration, but the ability to completely change it (with the many panel options) is just amazing (kind of contrary to GNOME).
I think I have figured out at least part of the reason for the spacings always feeling just a bit (or very) off (although I think it's bearable with my theme):
Their Kirigami design support library has a set of pre defined spacings and sizes developers are advised to rigidly stick to. These probably break down in many contexts and might not even be chosen well in the first place.
Another reason might be that KDE doesn't have a design group that is as strong as Gnome's and KDE might not be organized top-down as much (which is probably related to financing).
I think you can fix or mitigate well a lot of the mistakes (Qt is great) by using a good Qt curve theme.
I am a fairly new KDE designer who started contributing icons in the spring of 2018 because it was easy and fun.
That being said, on Linux I think Elementary OS is at least trying to create a coherent visual design language. Maybe there are others?
Please don't inject meaningless flag waving platform propaganda into genuine conversations.
Personally I find that using Source Sans Pro as the font and the Arc theme for most of the rest works well.
I have to apologize in advance because it's going to sound like a harsh comment and I know this is volunteer work. I really think that KDE is technically very good. The amount of basic design mistakes however is just unbearable as far as I am concerned.
I mean, let's take a look at this improved notification panel. The closing and setting icons seem slightly lower than the icon on the left of the title. The large hamburger menu icon is actually two pixels on the left of the closing icon. There are 17 pixels of top padding but only 12 pixels of left padding. The large image is one pixel of center (232 pixels on the left, 230 on the right). Every single line of text has a different interline.
And it's same for every application of KDE. They really need a design team to polish it up.
1) Information dense
2) Looks good in screenshots
3) Information dense and easy to use
4) Information dense and easy to use and looks good in screenshots
An awful lot of designers stop at 2 because a 4 is way harder to achieve, yet a 2's pretty easy to sell (to "stakeholders"). The result is at least sometimes (sometimes) better than a 1. They may also drop a 3 to a 2, though, and it's really hard to later turn a 2 into a 3 (since it's "uglier", so harder to sell) and a 2 is so different in character from a 4 that it can't even really be considered progress toward it (a 3 might be).
For progress in polishing stuff (not only visual), see the blog https://pointieststick.com/ for weekly updates.
I kind of feel bad about my first comment because I know it takes a lot of time to do these modifications, the new interface is a clear improvement from what was there before and UX wise it's not bad at all.
It's just that every time I use KDE, I have this feeling that great work doesn't shine as it should because of padding, alignment and box in a box in a box issues. I know these adjustments are a pain to make and that it's a very ungrateful work but it shows.
So, thank you to the design team I guess and keep up the good work.
Basically, every Slack message would append to the same notification, so after 4-5 notifications, I'd just see old ones and a ... line, making the notifications useless. This also went for the update centre, etc.
Also when notifications decide to just stay on screen indefinitely because of... Reasons.
I've never experienced this. Particularly, I don't experience this with Discord, and Discord is to Slack what a bee is to an ant, which is why I juxtapose it. (And I would ask if it might be something on the Slack side, but you say it happens with everything.)
>Also when notifications decide to just stay on screen indefinitely
Never had this either.
Might have been bugs. If you're on Ubuntu or another point release distro, those kinds of bugs also stick around, IME, because your DE version stays stagnant and the bugs are only fixed in subsequent versions.
That's exactly why we've seen scientific slide presentations in Comic Sans. It looks great to someone != it's good.
Another factor might be that those small, focussed screenshots are just not the appropriate format to show how those notifications would look like in the context of a complete desktop.
But in the end, it is also possible that there is an unintended lack of horizontal padding.
I for one definitely like and want more compact information. I definitely do not like the trend for making information more sparse. Leave that for your blog; business runs on information.
The amount and location depends on what you want to do. Yes, just like a phone book, it needs whitespace. A notification window is not a paper phone book though. A paper phone book typically has a lot less whitespace than we're discussing here, so I'm not even sure if you're agreeing or disagreeing or trying to point something else out.
The problem isn't that they need more padding, it's that their padding makes no sense. If they made a design decision to have tight padding that would be one thing, but they've got a horizontal layout with tight padding in the horizontal direction and wide padding in the vertical.
here's a really quick-and-dirty edit to even out the padding around each text block to be equal to the line height, which actually tightens up the total density by correcting the too-high line height in the notification text: https://imgur.com/uVcY6Sw
Programmatically, that assumes that that information is available (query a font rasterizer: where is the first pixel? consider also kerning and foreign languages; also query icon: where is the first non-background pixel? consider also theme information where window background might not show the icon well!).
Programmatically, you've un-aligned everything: now the amount of whitespace at the top of the notification is different than the amount of whitespace on the left of the notification.
Sure you might say it looks better. But how much time do you think it would require to make it look "pretty" (for you)? As-is, it's functional and isn't ugly. If you want to improve it, then go contribute to the libraries and applications being used.
As you’re keen to point out regarding everyone else’s comments, that’s your opinion.
Personally I think it’s functional and ugly. They could skip the drop shadow, the subtle transparency, the big icon on the right, the rounded corners, and just draw a white box. As long as they fixed the left padding it would look better than this.
Lots of attention to detail on the wrong things before they have the basics, IMO.
EDIT: Here's one for all you folks that aren't bothered by iffy typography https://i.imgur.com/4tBNA0w.jpg
thanks you pretty much perfectly summarized KDE
I wont say it can't get better but I'd be hesitant to anyone who comes from the outside if they try to to make to many changes at once.
Again good UX exists but recently it seems it is all about
1. Copying Mac (see Unity) and Chrome
2. Removing options and configurability (see any modern app)
proposing design changes to open source projects when you haven't already been a contributor for a hundred years never goes well.
I think the problem might also be exacerbated by the fact that the padding doesn't seem to be scaled based on DPI. (Unlike line height/icon size) So it looks acceptable on a low resolution display, but super close to the edges on a "Retina" display.
If you’re going to rip off Apple, at least do it to Apple quality! Otherwise it’s a cheap knock off. Before people grip that this is free software, would you be happy if KDE’s core components were poorly written? Free or not, we should take pride in our work and put the bar at high quality.
edit: i mean this as the conscious aesthetic, no disrespect is implied
You know, the thing about an open-source project is that you can contribute.
Since you think you are able to do it better yourself, why don't you fix it?
KDE isn't similar to Apple at all.
Source: have used both for years.
And to be honest, I think the notifications area required a bit of optimization. To me, it feels like it is also one of the components that received more optimizations than any other component already, but it also seems to be a little more complex than you would expect on first sight.
Where "few years" means 8 years? :D By all means rewrite software when it is 8 years old. A lot has changed in that time and the author notes that user expectations have changed, which leads me to the...
...false dichotomy you offered. User expectations and the code base not supporting those is the stated reason.
If KDE was a monolitic project with insufficient tests, those details wouldn't improve. Instead, I'd bet developers would be fast at removing current functionality when unavoidable changes add hard to track bugs.
I'll share the display issues - something deep inside calls random() on monitor resolution and position - I'm sure!
I've switched to Arc Dark and Numix Circle Icons and it's not ugly anymore. I don't know why they still ship Breeze with that broken padding by default.
But it's really getting better every release.
I mean, how many times can you redesign a logout screen, really?