Slashdot Mirror
KDE SC 4.7 May Use OpenGL 3 For Compositing
An anonymous reader writes "KDE SC 4.5 is about to be released and KDE SC 4.6 is being discussed. However, Martin Graesslin has revealed some details about what they are planning for KDE 4.7. According to Martin's blog post, they are looking at OpenGL 3.0 to provide the compositing effects in KDE SC 4.7. OpenGL 3.0 provides support for frame buffer objects, hardware instancing, vertex array objects, and sRGB framebuffers."
I love eye candy, as long as there is an easy way to turn it off. I don't need my linux box booting as slow as my windows.
For the ignorant, please explain what KDE currently uses for composting? I know on my machine it's hardware accelerated and DirectX isn't available on Linux. Doesn't that mean, by default, that they used OpenGL?
or you just don't turn on the optional compositing in the first place? this is a new feature created and intended for newer machines.
No, you can turn off compositing. Unlike akonadi, which already makes KDE unusable.
Lacking <sarcasm> tags,
More things to brag to my friends about. My e-penis will be massive by the time it comes out.
KDE can use XRender and/or OpenGL 2.1.
Just the need to upgrade how Kwin uses OpenGL currently to do rendering. Right now its still using the old OpenGL 1.1 - style rendering (fixed-function rendering pipeline) to a programmable one using vertex and fragment shaders. This way, it'll be easier to port it on embedded devices that uses OpenGL 2.0 by default
More like the winds of stupidity. GNOME is still designed for idiots, and the KDE developers decided that being a rock solid DE with a good OLE model was less important than having cool looking visual effects and trendy desktop applets.
Palm trees and 8
Good luck with that.
People will be blaming KDE for the following issues until they abandon the idea:
1. Half the Intel users will blame KDE for the kernel panics you get when using a Hello World shader with some of the Intel drivers..
2. The other half of the Intel users will blame KDE that they can't use any of the items listed (frame buffer objects, hardware instancing, vertex array objects, and sRGB framebuffers) because they still only support OpenGL 1.5 from 1865..
3. Then there will be the Linux people complaining about it running very slowly because some software driver is used by X11 due to distribution issues with distros and proprietary drivers.
4. The AMD users will probably be using some old buggy version of their driver that has buggy implementation of frame buffer objects or whatever.
5. See #4 but replace AMD with Nvidia.
6. Then there's the army of Linux users that do have a Nvidia or AMD card, but their card is from 1765 and therefore doesn't support OpenGL 3.0.
But besides all that OpenGL 3+ is pretty neat and you can do some fun shaders for your compositing. I wish them the best of luck!
This seems like it can only be a good thing. The major place where we're lacking (AFAIK) is in driver support, and having a major software suite such as KDE use OpenGL 3 will help the driver writers manage some of these bugs (the same way Compiz appearing on the scene majorly improved graphics drivers in Linux a few years ago). Perhaps this will also help to push Intel to OpenGL 3 (or 4 - I mean, COME ON!). At the same time, I have some Linux machines that don't have OpenGL 3 support (one has a GeForce 6600), so I really hope they keep functionality with OpenGL 2 for a while (that machine isn't getting upgraded - the next thing I do to it will be to replace it).
I really wanted to change my sig to something witty, but all I could come up with is this.
I don't know that we need any more eye candy in KDE 4. It already has a ridiculous amount of aesthetically pleasing features. How about we squash some existing bugs and add more usability features.
This is not the penguin you're looking for.
Oh for god sakes people. Kwin provides pluggable back ends for rendering engines for compositing. Currently we support xrender and OpenGL 1.1, soon we will support the next version of OpenGL. Big deal. You can turn compositing on or off, or choose which engine is best for your platform. We will not remove the old engines or force everyone to use compositing. So stop your trolling.
No, then you just carry on using OpenGL 1.1 or xrender
Funny, I don't recall you ever having paid me anything
If such a change happened, I'd imagine there would be alternate rendering paths, just as you can fall back to xrender right now; old hardware likely won't be left out.
I can tell you that 4.5rc2 automagically loads up akonadi and all of its fluff/garbage/helpers if you have a clock plasmoid. Without option to turn it off.
Akonadi and Nepomuk are simply jokes. Enforcing them on the user, specially considering how useless both services are, is a really bad idea©.
Nepomuk can be disabled easily, not so much for Akonadi. You literally need to cheat it by giving empty path strings, or no clock.
I'm a major KDE advocate, but those two services get on my nerves way too much, specially because they are rather hefty for what they do (for me, nothing at all, for others, very limited usage).
Rant mode ON:
KDE seems exceedingly dependent on itself right now. And integration efforts (with popular apps out of KDE) are pretty much non-existent or unknown even among devs (I discovered after a friendly rant about the current "closed" state of things, that Krunner now does index Firefox bookmarks. The person who corrected me learned it by pure chance it seems, as no "user friendly media" (getting deeeeeeeeeeeep into mailing lists and all the bulk of svn commits is not user friendly, it costs more than mere minutes to check all that) reported on it at all).
I don't know who is to blame but whoever is responsible for this, is not helping the already damaged (by 4.0) reputation of KDE. Half-baked and/or mandatory apps are not helping. Neither does the silly "KDE SC" gimmick.
I can only think something in the management chain is broken, leading to absurd/rushed/experimental decisions pulled off. Either that or the exceeding majority of the 6-month release cycles is translation/bugfixing. As new features talked about during the release of "KDE 4.X" are implemented in "KDE 4.X+1" in the same state shown during the 4.X release (Look at tiled windows in the 4.5 branch. It's there, but...)
Rant mode OFF
Sorry, I really needed to put that up for discussion. Whenever Akonadi is mentioned I go berserk as I am reminded of stuff like it being a requisite for the standard clock.
The worst is that I am an enthusiastic KDE user and I follow development closely, trying betas and reporting bugs. I don't feel "betrayed" or anything like that, but some things are too annoying/habit breaking/RAM eating. Krunner, a Quicksilver/Kupfer-like launcher, can't be disabled and I was told by KDE people that it governs over logout functions (WHY THE LAUNCHER? why can't I just have my alternative of choice without option to take it out or disable it?).
Well, at least the project is dynamic and a good fix/decision changes for better can happen eventually.
Why not? Serious question. There are some companies and government agencies out there that still use 1960s mainframes. This idea that we should code only for new hardware is only appropriate for a certain washington based software company whose profits depend on making people buy entirely new software suites every time they upgrade their hardware.
There's not as much profit in selling a minor software upgrade that works on an older computer compared with selling a full new software version that works only on a newer computer. By deliberately making newer software unusable on old hardware, customers are forced to upgrade their hardware, and incidentally have to buy a new bundled Windows OS, which then forces them to upgrade all the otheir software that they use as well. It's a mug's game.
There's no reason why open source should follow that model. It's free, and it's intended to *help* users make the most of what they have, not just grab the most of their money. Moreover, making software run on older and different hardware is a great way for developers to find bugs, and thereby improves the quality of the code. And that means that other developers will have more confidence to reuse it for their own projects.
Open source should have a 50+ year outlook. That's how the real world works. Look around you, how many buildings, roads, bridges, companies, laws etc, are 50 years old? How would you live if you had an arbitrary rule that you couldn't enter a building or cross a bridge or drive in a car built before the year 2000, and could only do business exclusively with companies founded after 2000, etc?
I would have to agree about Akonadi. I recently did an install of Arch Linux, and used KDE for about a week before it started annoying me more than Windows 7 does. I removed KDE and just went back to Gnome. At least there I can strip out all of the stuff I don't want to use.
As computers get more powerful, things become feasible that were not in the past. People want those features. Problem is, making software that uses them doesn't work with older systems.
That is just life. Now as for your example with mainframes, in that case someone chooses to pay for support for a system. They cost a ton to maintain. Also, you do not, in fact, get new software, just support on what you have. If you own an IBM/390, as we do, you don't get to run the new version of zOS on it. You are stuck with old software. Supported software, but old software.
Nowhere did I see anything that said support for old KDE would stop, just htat new KDE may need hardware to do composition. I fail to see the problem here.
No, you're not a customer. If you were, you would know that KDE falls back to no compositing when there is no HW acceleration available. You're just a whiner on an internet message board.
Oh, shut it. OS X 10.0 was barely beta quality as well, and somehow people stopped complaining when it started becoming usable, even though the upgrade to 10.2 cost money. Same with Windows Vista (6.0) --> Win7 (6.1). With KDE4, you were even warned not to use 4.0. But you still had to run off and use it, didn't you?
You're a user, not a customer. There is a difference. Unless you know some place that sold you KDE, and in that case you got ripped off, because you can download it right from your distro's repository like the rest of us users.
Because i dont subscribe to upgrading just because there is new and shiny available. The old functions perfectly well, and its appalling that people code like they do today, rendering perfectably good hardware simi-functional.
---- Booth was a patriot ----
about OpenGL decorating my windows.
DO care about things like "desktop works" and "can find a fast, professional theme that makes taskbar look like window title bars," neither of which is available with KDE since KDE 4 was released.
Yes, I have recently tried KDE, up to and including KDE 4.4.5 on Fedora. It continues to suck eggs. KDE 3 was professional and powerful. KDE 4 seems to have all the options I don't want, none of the options I actually used, no way to get a unified KDE/GNOME/Plasma theme (hell, you can't even get a unified kwin/plasma theme), ugly artifacting with 3D compositing off, craptacular stability and a distinct inability to remember many settings, dog-slow previews compared to Nautilus, no "compact" mode in Dolphin, either, poor dual-display support that fails to automatically handle them elegantly, and a distinct lack of KDE4-specific, complete alternate icon themes at kde-look.org to do away with the bright colors (I don't want red icons and blue icons both on my desktop at the same time; my desktop PC is not an Icee machine, it's totally unprofessional).
In short, I find KDE 4 totally considerably less usable than GNOME or KDE 3.5 and I'm fairly sure that pouring more development hours into 3D compositing is not going to make it moreso. How about just fixing the artifacting with 2D rendering? That I could actually give a damn about, though it would be one problem solved amongst many, many problems that didn't exist until KDE 4.
STOP . AMERICA . NOW
To truly hate akonadi, you need to be logging in with $HOME on an nfs mount. And shutting down the box from time to time.
What happens is that KDE issues telinit 6 without waiting for akonadi and mysqld to terminate, which means that your nfs mount is still active at shutdown, so when the system forces the unmount the database is not coherent. Thus you get the dreaded "akonadi could not start" error on next login. Well, that's easy enough to solve by just whiffing $HOME/.local/share/akonadi -- as long as you don't have anything useful stored in there.
Which the KDE team is making harder to do all the time. Good thing the system backs up that akonadi database on a regular basis.
Oh, wait ...
Lacking <sarcasm> tags,
Ah, yes, the good old days of KDE, back when it had exactly five options that could be configured, and the only way to modify the menu was by hacking an XML file.
Funnily enough I recently made the reverse migration. Xfce served me well for a while, but every single recent version has replaced something that worked fine with a rewritten version that has fewer features and/or simply doesn't work properly at all. KDE meanwhile is very pleasant to use, runs perfectly fast even on my underpowered netbook, and is the only mainstream Linux desktop environment that actually bothers to support widescreen monitors properly by implementing usable vertical panels.
Well said. I agree completely. If you haven't already, you should check out one of Aaron Seigo's posts from earlier this year on his blog:
http://aseigo.blogspot.com/2010/05/i-dont-need-no-stinking-nepomuk-right.html
He attempts to justify and defend the thorough integration of neopomuk and akonadi with KDE4 in his post and the subsequent comments. He mostly fails.
In my opinion Aaron Seigo needs to go. He seems like a really nice guy and all, but he still defends the release strategy of KDE4.0 (and this despite being one of the lead devs of the -at the time- completely bug-ridden and barely functional plasma), and seems to always be at the forefront of KDE4's questionable future plans. They've reached feature parity(?) with 3.5.X. Now they need to work on stability and speed. Stability and speed. Stability and speed. The obsession with social networking integration is stupid and shortsighted. The SC naming scheme is lame. And almost as many users are now annoyed by neopomuk and akonadi as they are by that damn cashew.
I've been "trying" KDE 4 for maybe a year or so. I like some things, but I hate most of them. At 4.5 it still feels like someone's abandoned alpha. Every new release brings new UI candy, yet breaks long-standing functionality or fails to address real usability problems (like that stupid desktop peanut - whose idea was that?).
What particularly irritates me is that they seem to be reinventing non-desktop features. Not only is this very much against the "Unix way", but they're doing a terrible job of it and the whole mess is wholly unnecessary. I don't know if we as users are doing a poor job of informing the devs about desired functionality, but I would love to meet (and murder) the person who thought Akonadi would be a good idea.
Perhaps I'm a minimalist, but I like KDE for mostly one thing: KIO slaves. I love the fact that I can open up a file browser and treat remote files almost as though they were local. That makes my life as a developer and sysadmin so much easier. Everything else is fluff to me, as long as I can fire up Kate and edit my remote server's configs I'm happy. On the flip side, everything that gets in the way of that location-shifting goodness is EVIL! Akonadi is evil. Half-assed transitions to libssh2 are evil. Godawful "toaster" notifications and ambiguous error messages are evil. The plasma interface engine randomly crashing every few hours is evil. All those unfinished K apps that nobody uses are evil. I could go on...
It seems the KDE people have forgotten that, above all, we just want a GUI to make our lives easier. Streamline it, trim off the fat, we're Linux users for fuck's sake. People are flocking to minimalist interfaces like Fluxbox, just to get out of KDE hell.
-Billco, Fnarg.com
It doesn't take an old machine to make kde 4.x unusable.
Now that I think about it, kde 4.x is like the hot but totally completely brainless girlfriend... Nice eyecandy, initially very pleasing and exciting, then rapidly becomes tiresome as the reality that most of your interaction occurs outside of the bedroom sets in.
Likewise, kde 4.x is very pretty and the first half hour is spent checking out all the cute/neato things it does. Horray, kmail finally keeps responding while checking mail in! Konqueror supports more of web 2.0!
Then the dream begins to crack: Why does my desktop's framerate crash to a slideshow when I first move a wobbly window? Why does my desktop require me to install a complete SQL server for something that doesn't work worth a damn anyway?
Then you wake up and the house is on fire: Wait, I swear Konqueror used to fit twice as many file icons on this page... Did its html engine just completely stop updating the display on some random webpage? Why does it seem to randomly forget settings?
Then you call the fire department and realize that everything 4.x does, 3.5.10 does faster, better, more stably and while using half the ram.
Every new version that's released I try out, and since 4.2 (when the showstopping breakage was mostly fixed) every time I end up going back to 3.5 muttering about how they don't seem to have fixed a damned thing.
Let me assure you that Aaron is *not* generally thought of as being selfish and elitist. He is a very smart guy who sees the big picture in things and who also listens to other people a lot before he makes up his mind. He also has a good way with words, which may not go well down with people who have other agendas. Those of us who often interact and work with Aaron sees what an immense load of bullshit he has to put up with from anonymous cowards. We know he is a pleasant guy, and we not only like him, but also pity him sometimes for the flack he has to endure. Like the parent.
He might not be wrong, but I think his attempt is doomed. Nepomuk and Akonadi are not applications, so to the user they are meaningless. However KDE4 is generating messages about them, so that's confusing at least - usually annoying, too. On top of that there seem to be no applications which actually use them in a way which would get the user interested - he states himself that he turned off Nepomuk on his own system, so apparently he hasn't found a use for it either.
The KDE developers want these services available to applications, and that makes a certain amount of sense. However they cause problems and eat a lot of resources which leads to user complaints. Instead of starting the services by default (and using a setup which consumes lots of memory and CPU) they should default to off. Then when an application is first started which uses these services, that should start a wizard which lets you configure the services. Then you could decide which service to run and control the resource requirements. The user would understand what the services are for in this context. Right now he needs to find out what they are for by noticing the high system load and identifying the process which eats up all the resources - that's not a good experience.