KDE today releases the first Beta version of the next-generation Plasma workspace. The Plasma team would like to ask the wider Free Software community to test this release and give any feedback . Plasma Next is built using QML and runs on top of a fully hardware-accelerated graphics stack using Qt 5, QtQuick 2 and an OpenGL(-ES) scenegraph. Plasma Next provides a core desktop experience that will be easy and familiar for current users of KDE workspaces or alternative Free Software or proprietary offerings.
The Breeze Plasma theme looks flat and dull, but otherwise it looks like a promising release and hopefully other less flat themes will arrive eventually. Also, FINALLY they seem to have scrapped the cashew that so many people gripe on even to this day on KDE4.
edit: Oh it seems the cashew is still there but disguised with three lines instead of the familiar icon in the screenshots.
Edited 2014-05-14 22:41 UTC
I don’t think the cashew is going to dissapear any time soon, after all, it is a monument to Aaron’s Seigo ego.
It appears they are trying to copy the flatness of Windows8 instead of keeping an independent design.
This, for example, really put me off http://wstaw.org/m/2014/04/14/sample1.png
Edited 2014-05-15 04:46 UTC
Flat is in, gradients are out.
OS X is even moving there.
Personally, I find it infinitely better, and hope 3D and gradients die in a fire.
I like flat, when it is not overdone. Like when toolbars run together, or with the colorless and infinitely ugly icons of Visual Studio, which look like they were ripped from an ancient copy of Word Perfect.
I am really happy that many people do not agree with you. And perhaps one day one of the OS developers dare to deviate and bring back a decent looking UI that is not flat.
In fire?
1) This isn’t a fashion show. Nothing is “in” or “out”. Usability, unification of designs, accessibility and ease of development are what drives the trends. And I am allowing the word trend here even though it negates my point
Just because Microsoft used flat widgets in windows 8 ,eventhough they look like something from 15 years ago, doesn’t mean we need to follow them like zombies.
However, we shouldn’t hate a design just because it looks like window8. So, if you can tell me why flat widgets are better, I am all for them, otherwise no thank you…not in this life time. I prefer the 3d buttons. I can easily tell this is a button without trying to adjust to a new perspective.
A lot of HID principles are being broken. That in itself is not necessarily a bad thing BUT we need to standardize new ones. For example, firefox getting rid of the menu bad by default changes things they followed the principle of saving height on wide displays.
Saying this is “in” and that is “out” is childish in my opinion if not accompanied with an explanation for why this or that are better.
2) “3d” and non-flat widgets are not just aesthetics. They serve to save a few ms of your time when you look at different windows and try to distinguish between them. If we didn’t need them, we could simply get rid of the window manager concept altogether and adopt a one application on your screen at a time like android does.
Now in terms of pure looks, gradients aren’t bad when they are use simple styling. Oxygen Qt style uses them to outline the top of the window. This gives a usability improvement. On new wide/flat screens, people try to decrease clutter because there isn’t much “height” space. Notice how the menu title in Air theme is slim. To help with that, Oxygen theme uses gradients to highlight that part of the window.
Gradients have their uses
You just need to use stylesheets and not hardcoded pixmaps 
Edited 2014-05-15 14:00 UTC
Gradients were put in there because designers wanted things to look more like real-world items, and to show off their graphical capabilities.
Flat design gets rid of all of that showoffiness, reducing things back to where they should have stayed – minimal, with only what is strictly necessary.
When you’re using a terminal, you don’t have silly gradients etc. eating up space, so there’s no need for it in GUIs either.
I admit that the initial post was a little gloaty, but the fact of the matter is that 3D design, bevels, gradients etc. don’t give the improvement to usability that is so often claimed… at least, I’ve not found that to be the case.
That being said, I tend to remove all toolbars and menubars, as well as having no desktop and no desktop widgets – maybe that’s why I celebrate moving towards a cleaner-looking aesthetic.
Don’t mention oxygen to me :p
First thing I do after installing KDE is install QTCurve with a custom flat theme.
That’s the key point: aesthetics. Not everyone has the same. 3D design etc may not be much more usable, but I personally prefer it. I find it helps me locate things far more easier than colour does, which is all you’re limited to if you use flat design.
I almost always have a terminal open (I have been using linux since the 90s). My distribution doesn’t have a gui package managers or administration tools. For the last few years, I have also been compiling all packages from source
Plus you can have non-flat icons without anything shiny.
Remember the tango icon color palette? elegant, easily recognized forms without any shiny edges…and they aren’t flat.
By Oxygen, I mean the Qt style and NOT the icon theme.
Flat or gradient, when executed right, became pleasant for the eyes. There’s none really good/bad design concept, cause almost all is subjective
I think they’re finally realizing the theme and style that got people really excited about KDE 4 in the first place. That design probably has little to do with Windows 8. Even thought I must add that the flatness of new versions of Windows and iOS probably gave them incentive to actually do it.
Just look at this:
http://kde-look.org/content/show.php?content=28476
Edited 2014-05-16 12:32 UTC
Having been programming using Qt5 for the past year and fully utilizing the power (read: simplicity) of C++11 that Qt5 supports, I don’t really see why you would want to implement a massive project using QML. I’ve tried QML, mostly before biting the bullet and learning C++11. Even with the natural extra bloat in C++11 syntax, you actually don’t write that much more code than you would in QML. The good thing about Qt C++ API is there is hardly any boilerplate code you need to write.
I do bash C++ from time to time, mainly due to its C heritage and lack of modules.
However C++11/14 are quite refreshing, specially when making use of modern C++ idioms.
QML part is nice, but I think the driving force behind was the desire to cater to young developer generations raised with node.js.
The problem with C++ is that all the latest niceties are wrapped in a tough layer of extremely ugly, convoluted, and difficult syntax. Templates, iterators, smart pointers, the lot. And while the K&R manual for C is a model for clarity and conciseness, Stroustrop’s C++ manual is exactly the opposite.
Of the base Object Oriented layer, I particularly hate the fact that you must show your private parts in the include files, or use a something like a Cheshire Cat, an abstruse pattern that really should not be necessary.
With automatic type deduction, this criticism is mostly out of date. eg:
auto i = my_vector.begin()
or even better:
for(auto i : my_vector) …
And to initialize a vector:
auto v = vector<int>{1, 2, 3, 4, 5, 6};
Smart pointers:
auto p = make_shared<MyType>(/* constructor arguments */);
BUT smart pointers aren’t really needed because most of the time you don’t even need to explicitly allocate anything on the stack.
And C is full of security holes that modern C++11 completely avoids by default. You trade one for another.
Not to mention C++ templates for operations like sort are inherently faster than C’s qsort, because templates bake in the comparison function, whereas qsort goes through a function pointer, which is not as efficient.
Private members are mostly useless these days. I personally find no benefit with that kind of information hiding. This kind of dogmatic or cargo cult practice of declaring private members in a class is more Java’s thing than C++.
I of course meant heap.
Although I agree that C++ is a monumental POS language (even though I have to use it on a daily basis), to the point where I have day dreamed about hitting Stroustrup with a wet sock. The concept of “private” is actually necessary in the context of C++ classes, it’s perhaps not the best name for what it represents though.
Why is it necessary? C doesn’t need the concept of private members and people who love C over C++ seems to be fine with that.
How else should the compiler know how much memory your class will take if it cannot see the whole implementation?
Use the pimpl idiom if you need to hide stuff. It might not be pretty, but remember that all code generated by the compiler is done in advance, and can easily be run on small microcontrollers, mobile phone, huge computing intensive clusters, etc.
The only drawback with C++ is the fact that it’s too powerful for most people. It’s obviously not for everyone.
By reading the symbol table metadata in a module, like in other languages with module support.
Looking forward to see modules make into C++17.
C is not an object oriented language, so I have no idea what your point is in this case. The concept of “Private” is integral to the object architecture of C++, and it has little to do with what part of the code a programmer “should be able to see” (which seems to be a common misunderstanding).
QML has nothing to do with catering to younger developers. It’s about doing UI’s in a better way. A declarative language is far more suited to UI than doing it in C++ directly. Great way to enforce separation of UI and business logic as well.
QML is nothing new in concept, it has been done many times in different programming languages and RAD environments.
Even Qt already had one in the form of .ui files.
It is catering to young developers, because it is JavaScript, with C++ being pushed down the stack for the rare cases where there isn’t a QML component available or not delivering the desired performance.
“Plasma Next is built using QML and runs on top of a fully hardware-accelerated graphics stack…”
So what does this mean for Plasma Next on systems without good 3D driver support? I’m thinking about booting up a live CD on a system with AMD or Nvidia graphics before configuring the proprietary drivers– what will happen? Will it fail to run? Is there a low graphics “2D” mode? Or will it burn up the CPU with something like LLVMpipe? Or something else?
Edited 2014-05-15 04:49 UTC
At least with intel and AMD, the open source drivers can do 3d well enough to drive kde and the effects.
looks pretty good. Really worth to study.
Any concise KDE primers/tutorials available?
These primers/tutorials should just teach where to click to get the most often used functionalities. An 80% coverage is absolutely sufficient for my needs.
Greetings,
pica
I’ll take a look at it
pica
So far I only have implicitly stated a need:
The GUI should be designed in a way I consider nice.
I like to work with tools I consider nice. Furthermore I have the idea I work better using tools I consider nice. Just an idea, not a proof.
Now my main requirement
I work as an architect. I perform enterprise, solution and application architecture tasks. Doing these I often have to combine information provided by several sources. These sources could be text or presentation documents, a scan of a hand drawing or hand-written notes, a photo of a flip/white board, …
In most cases I have more than a single source I have to have available at once.
The <Alt>-<Tab> mechanism provided by several OS GUIs I consider handy, if you have exactly 2 application (in my case 1 source document, 1 target document) to switch between. Having more than 2 applications (translates to more than a single source document) to switch between, I consider the <Alt>-<Tab> mechanism a pain to use.
I consider the <Windows> mechanism provided by Gnome3 handy for that very task. I simply press the <Windows> key and I get my 4 to 6 windows/applications/documents presented in a standardized way. I then can pick the document I require simply be clicking on the associated window image. I consider that way of switching really handy.
What is the preferred KDE way to do that trick?
Greetings,
pica
If you go to system settings and then configure desktop effects, you should be able to find the effect called “Present Windows”. The default shortcut for that seems to be Ctrl-F10, but you can change it to your own.
KDE does have the annoying thing where you can’t set a shortcut to just the <Windows> key. <Windows>+z would be the closest.
Edited 2014-05-15 11:01 UTC
https://github.com/hanschen/ksuperkey
There’s an addon you can install which allows the super key to be mapped to another combination, and still allow it to be used as a modifier.
I have mine set to super+space.
If you’re running arch, it’s available in the AUR, too.
Do you mean the “present windows” effect?
http://www.kde.org/announcements/4.0/screenshots/kwin-presentwindow…
Yes, it is available in KDE. I’m sure it can be mapped to any key.
As always in kde, many different ways of doing that.
Currently, I have a hot corner configured, where I just move my mouse to and the present windows effect is triggered.
Oh how I hope that one day,
Kicker makes a glorious return,
Since Kicker runs on all your,
CRT monitors without issues,
It doesn’t cause desktop to hang,
and act all weird when you add monitor,
Oh how I hope that one day,
Kicker returns to my KDE,
“You weren’t perfect but that’s all right!”
Oh how I hope that one day,
Kicker makes a glorious return,
“Just look how unreliable Plasma panels are!”
Hopefully one day Kicker replaces,
that Plasma junk in my KDE.
Edited 2014-05-16 12:47 UTC
I tried the beta (via the Neon image) and am cautiously happy with the new release. The performance is excellent, the layout is very much like KDE4 (though perhaps a touch calmer, less distracting). Memory usage was about the same as KDE4. I did not much care for the default “flat” theme, I find it hard to tell what is a button and what is just background with flat themes, but it seems easy enough to customize to a look I like better.
Basically, from the end user point of view, the only difference seems to be that KDE5 is faster and less visually distracting. I think the beta is very close to production ready.
By the way, someone above asked what would happen if a person ran KDE5 on a system without hardware acceleration or 3-D support. It works fine. KDE5 does not need 3-D support or hardware acceleration, but it can make use of the feature if the hardware/drivers support it.
I can’t explain what it is exactly but I think what has always held KDE back is that it is subtly just ugly. And this transcends what is changeable via themes because the KDE-ness always remains. Gnome always seemed (before it died) rougher around the edges but I always preferred it just because it felt better to look at.
I always had the opposite opinion.
I think it’s the icons? Oxygen is less unbearable than Tango IMO, although most people seem to quite like tango.
Which version of Qt should currently be used for compiling? 5.3.0-rc or 5.2?