Next Generation KDE Plasma Notifications
blog.broulik.de
blog.broulik.de
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
KDE isn't similar to Apple at all.
Source: have used both for years.
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.
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.
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:
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.
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.