Linus Torvalds piped up in the comments of a Google+ posting by Linux kernel hacker Dave Jones to air his true feelings about Gnome 3: “it’s not that I have rendering problems with gnome3 (although I do have those too), it’s that the user experience of Gnome3 even without rendering problems is unacceptable.” People care what Linus thinks, and when he ditched KDE for Gnome a couple of years ago, people took note. Now he’s using Xfce.Torvalds covered his beefs with KDE in depth with in a 2009 interview with Computerworld:
“I used to be a KDE user. I thought KDE 4.0 was such a disaster, I switched to GNOME. I hate the fact that my right button doesn’t do what I want it to do. But the whole “break everything” model is painful for users, and they can choose to use something else. I realize the reason for the 4.0 release, but I think they did it badly. They did so may changes, it was a half-baked release. It may turn out to be the right decision in the end, and I will retry KDE, but I suspect I’m not the only person they lost.”
This is of course after, in 2005, saying:
“I personally just encourage people to switch to KDE. This ‘users are idiots, and are confused by functionality’ mentality of Gnome is a disease. If you think your users are idiots, only idiots will use it. I don’t use Gnome, because in striving to be simple, it has long since reached the point where it simply doesn’t do what I need it to do. Please, just tell people to use KDE.”
Is this evidence that Linus Torvalds is hot-headed and hard-to-please, or is it just indicative of the overall dysfunction of Linux desktop environments in general? Or is it an indictment of the open source collaborative development system not being well suited for something as personal and subjective as GUI design?
Title says it all.
I think Linus know that he can be a difficult person.
That is why he used git in “github”. They parody themselves sometime.
You should understand soft software dev. culture though.
I find the mailing lists and discussion posting comical at tongue in the cheek.
I don’t think he is more in difficult than say Bill Gates or Steve Jobs though from what I’ve read.
I’ve read some of his interviews and you realise Linus does not follow some people’s attitude that Linux is beyond reproach.
It is a joke .. because Linus created GIT. A Revision Control System.
Edited 2011-08-03 20:22 UTC
Github itself has nothing to do with Linus – however, he did create ‘git’, which the site is based on. And yes, there’s a great quote from him about his tendency to name projects after himself… first Linux, then Git.
[flame?]
Linus should take care about the Git user experience, before going after Gnome (particularly the mess with git commands).
Afterall, this guy shouldn’t be talking about user experience.
[/flame?]
Eh, git is no worse than any other source control system, and better than some. They’re all hard to learn, and all do things slightly differently from all the others – at least git is one of the ones where you get some reward for the time spent learning it…
I’m curious what would Linus say about Ubuntu’s Unity interface. (And also what does he think about KDE 4 now that it has matured two years.)
You should be careful of following person. A lot of people who would love Gnome 3 then realize that Linus doesn’t like it then they dislike it.
Linus is a Kernel Developer and leads the Kernel Development effort, he seems like a nice guy and really smart too… However his views on everything doesn’t make him right, or should change your opinion.
well, i for one don’t need to follow anyone. i was pretty much hooked on shell… until i saw this
http://jimmac.musichall.cz/log/?p=1181
after reading this plan i simply stopped following gs and my adoption of it. making phone from my desktop is way too unacceptable for me.
now, i’ll simply wait 6 more months and if this change comes into default… i’m switching elsewhere
Linux is a good programmer, but he is neither a usability expert nor an average computer user. He might not even be an average Linux user. If Linus doesn’t like the GUI he can change it, the average use can not. This is why we should design for them and not for people like Linus. What he thinks about Gnome, KDE or Unity is not more relevant than anybody else’s opinion.
The desktop train have left the station, and Linux isn’t on it. The next train is all about mobility, and on that train, Linux already have reserved first class tickets. Both Apple and Microsoft realizes where things are going and are trying to unify mobile and desktop user experience. The Linux must do the same. Gnome3 with its simplified interface is a step in the right direction.
Personally (being a developer like Linus) I felt that Gnome3 was very awkward at first, but after learning some keyboard short cuts, it have improved my productivity tremendously and I wouldn’t switch back to Gnome2.
Sure, there are bugs and minor glitches that need to be fixed, but that always happen on x.0 releases of all software. It happened to Gnome 2.x, that wasn’t much good until say 2.12, it happened to KDE 4.x that now at 4.7 is starting to look good. Over time Gnome3 will improve, and people will get used to it, and like it. Give it some time.
Finally, you have to have done something right if Apple, with lots of usability experts on their payroll blatantly copies large parts of your user interface, as they have done with Gnome Shell in their new MacOS-X Lion release.
After Ubuntu switched to Unity, I have been using the fallback of the Gnome 2 classic desktop. As of Ubuntu 11.10, Gnome 3 will be the fallback. Fortunately, Linux Mint is sticking with Gnome 2 for now, so I have switched to Mint. It is such a clean desktop. I have tired of Ubuntu trying to live so far out on the edge. I stopped using Fedora for this very reason. I must not be in the target audience the “New Gnome”. Like Linus, I left KDE after the 4.0 debacle, and have yet to find a KDE 4.0 implementation that is reasonable for me. Now Gnome is doing the same thing!
Then again, I’m old.
GNOME 2 all the way here, on my Linux desktop. Unity treats me like a child (those big icons are so incredibly condescending) and is very inflexible, whereas GNOME3 is basically a design study into how the annoy the crap out of me. I’m not kidding – it’s like the design document said “Let’s annoy the crap out of Thom”. I stopped giving a shit about KDE4 right around KDE 4.5 was supposed to fix the horrible Kwin performance but didn’t. Again. I tried KDE 4.6 afterwards just to have a go, and there, too, Kwin was a mess. I’m sure Kwin is now fixed in 4.7. Again.
So, GNOME2 on my Linux desktop all the way. If GNOME2 ever starts getting really long in the tooth due to the lack of maintenance, Xfce here I come.
I like Unity. The Windows key basically turns into an intelligent command-line opener. The big icons are nice for quickly switching between workspaces.
Right.
It is perfect for RIGHT HANDED PEOPLE.
I use the mouse in my LEFT HAND
Whoever designed that load of codswallop needs to be taken out and thrashed with a buggy whip. They must be right handed (or a German)
I’m pretty ambidextrous but using a mouse in my right hand? forget it. It just don’t work.
Gnome 2 for me all the way.
I’m left-handed as well, and both my tablet and my mouse are on the left side of my keyboard. And you know what? I haven’t got a single problem using KDE. I haven’t even used systemsettings to set my mouse to the left-handed button order. I’m completely in the dark what you can be complaining about.
He was replying to a post about unity, what does it have to do with KDE?
Gnome Shell behaves more or less the same way – hit the Windows key, type a few characters of the application (or recent file, etc), and press enter.
And yeah, I’m pretty happy with that part of it. Both Shell and Unity have things I don’t like much, but application launching isn’t one of them.
If you don’t like the big icons why not make them smaller, how difficult is it to slide a slider?
Edited 2011-08-03 17:39 UTC
You’ve nailed basically my entire Linux desktop computing dilemma. Unity can DIAF, and it’s the reason I no longer play around with Ubuntu. I tried Gnome 3 but it was a no-go on my hardware; both my P4 desktop with Intel graphics and my Core2Duo laptop with Nvidia Quadro graphics are somehow not powerful enough to run it. Never mind that they both handle Compiz-Fusion like a champ, and modern games are quite playable on the laptop under Windows 7. I did try the gnome-fallback option on the laptop and it ran, but it’s horribly simplified from the already simple Gnome2. Just from the screenshots of the regular Gnome3 interface, I can see why it would annoy you, and I think it would do the same to me.
KDE causes lockups and kernel panics on the desktop, and is horribly slow and buggy on the laptop. I do like the look and feel of KDE now, after nearly 10 years of hating it, but it’s not stable enough for daily use.
Since Arch is my distro of choice these days, and it doesn’t allow me to stay with Gnome2 (though oddly enough there is an unsupported method to stay with KDE 3.x) I have elected to stick with my old standby, Xfce. I started using it when Slackware was my distro and Pat dropped Gnome. It’s very much a “Gnome lite” but it can be brought up to nearly the same level of DE functionality with some tweaking and extra packages. On both my computers it’s fast, but especially on the laptop.
It may not be that your hardware doesn’t support it, just that you don’t have the right drivers. I own an Acer laptop that has a Nvidia 330m chip in it and if i want to install ANY 3d-compositing environment, (gnome3, unity or otherwise) i have to download the correct drivers from Nvidia’s website. I use Arch as my main system and this works fairly well.
On the laptop with the Nvidia chip, I’ve used both Nouveau and the binary-only Nvidia supplied drivers. With Nouveau I get the “not supported” message with Gnome3, and with the binary driver it tries to load the new interface but locks up. Either way I have to do the “gnome-fallback” method. On KDE it loads the desktop fine with either driver, but it’s just a slow and choppy interface.
On the desktop it’s Intel graphics, for which there are only the open source drivers and Gnome3 is still not supported beyond fallback. On KDE it locks up and sometimes kernel panics before the interface loads. I think that’s a KMS thing though; I’m going to try a suggestion I found online when I get back home. Though, I suspect if my Nvidia-based laptop is slow with KDE the desktop may be even worse.
I do appreciate the help, but I’ve been using GNU/Linux as well as BSD in various incarnations for about 12 years now, on just about any hardware imaginable (including ARM and PPC) so I have a thorough understanding of driver issues.
I’m having the same issue with intel and KMS. My “solution” has been to stay on kernel 2.6.31, which is rock solid. If you have found another solution, please share it so I can try it as well.
Unfortunately not. The KMS setting (i915.modeset=1) did nothing to stop the kernel panic under KWin. However, it did allow me to run Gnome3 without resorting to the fallback mode; unfortunately Gnome3 KPed on me when I launched more than one application.
I’m on the lookout for an Nvidia or ATI PCI Express x1 card (the only type of express slot in the desktop) so I can try to get it working. Until then it’s back to Xfce.
Cards for “larger” PCIe slots should work fine in any of the smaller ones – as long as no components on the motherboard obstruct the way for cards’ edge connector, and as long as the PCIe x1 slot is “open” on the side where the rest of a larger slot would be …and if it isn’t, carefully cutting out that opening should work.
(but yeah, the last one gets tricky
)
Edited 2011-08-08 13:17 UTC
Just curious, but what apps are you running when you get kernel panics and lock-ups in KDE4?
I use KDE4.4.4 each and everyday and only shut it down when I need to move my laptop on the weekends.
I haven’t had a single lock-up and i don’t think I’ve ever even seen a kernel panic in the 11 years I’ve been using linux. Oh wait, yes, I saw one once when I tried to compile my own kernel.
I’m not saying KDE is the best, but it is what I prefer. This is a major reason I use linux, because I can choose what I like to use.
I never get to run any apps, the KP happens during loading the interface.
I think I’ve tracked it down to a KMS issue with the Intel driver, I’m going to try a fix when I’m at home again.
I feel exactly the same way; Linux allows you choice, and while I like what I’ve seen of KDE I’m choosing to stick with Xfce for stability and speed.
KDE doesn’t cause any such thing as “lockups and kernel panics” under Arch. Running Arch and KDE4 is perfectly stable.
I run Arch and KDE4 with the open source ATI/AMD Gallium3D graphics drivers from Xorg. Stable as a rock, and excellent performance.
http://www.x.org/wiki/RadeonFeature
http://www.phoronix.com/scan.php?page=article&item=mesa_711&num=1
Unlike the adapted-from-Windows “binary blob” drivers from graphics card OEMs, the open source Xorg drivers are actually designed for Linux.
If you run Arch and KDE4 with different graphics drivers, and you really do experience “lockups and kernel panics”, then surely the graphics drivers are far and away the most likely cause of your issues.
Edited 2011-08-04 02:11 UTC
He doesn’t have an ATI card … so it isn’t a solution for him is it?
Edited 2011-08-04 05:59 UTC
That is not the point. The point was that he was blaming KDE4 for kernel panics when KDE4 runs just fine … why not blame something actually within the kernel, such as using the graphics card “binary blob” driver designed for Windows?
There happens to be an open source driver designed for Linux for nVidia cards also.
http://nouveau.freedesktop.org/wiki/FeatureMatrix
Even though, due to the fact that this driver has to be reverse-engineered because nVidia won’t release programming specs, this nouveau driver is a fair way behind the Intel and AMD/ATI open source drivers, it is still a driver designed for Linux, and it does support a composited KDE4 desktop.
That would be a solution.
Edited 2011-08-04 06:26 UTC
You are laughable … He has a Quadro card which means it is likely he is using Graphics professionally, Nouveau driver is not an alternative.
Binary Blob designed for Windows … what planet do you live on?? What in the world gives you that idea? Does it come in an .exe file with WHQL in the filename?
Stop making shit up.
Edited 2011-08-04 09:06 UTC
As usual, I can easily back up what I say, so I have some links for you:
The proprietary driver from nVidia for Linux is a binary blob plus an open source “wrapper” to interface with the Linux kernel. The binay blob part shares most of its code with the nVidia driver for Windows.
http://askville.amazon.com/Linux-drivers-work-distros/AnswerViewer….
“Binary-only drivers are more complicated. For example, NVIDIA currently only provides a binary graphics driver for Linux. NVIDIA^aEURTMs binary driver can be used on virtually any Linux system, with the requirement being the version of the Linux kernel. The driver communicates with the kernel via a small open-source kernel module which is compiled against the source code for the running kernel. With this method, NVIDIA is able to ship a unified driver that works across different distros. According to NVIDIA, the drivers for Linux and Windows share a common code base that makes their driver development much easier.”
http://en.wikipedia.org/wiki/Binary_blob
Despite being reverse engineered, the nouveau driver does quite well:
http://us.generation-nt.com/answer/free-software-nouveau-driver-now…
“While the Linux 2.6.38 kernel has been out for less than one week, if you use NVIDIA graphics, particularly with a low-end GPU, start counting down the days to the release of the Linux 2.6.39 kernel. Particularly on lower-end NVIDIA GeForce graphics cards, the reverse-engineered open-source Nouveau driver now meets or exceeds the speed of NVIDIA’s official proprietary driver in a number of OpenGL test cases.
Up to this point the OpenGL performance of the Nouveau Linux driver has been quite admirable considering that it’s largely a community project and it’s developed without NVIDIA’s help but by clean-room reverse-engineering their binary-only Linux driver.”
Edited 2011-08-04 09:43 UTC
Err No, you are still talking crap.
it is kinda normal for code to be reused … all that suggests to me that they have modularised their code base for the driver correctly.
That however does not mean it is “designed for Windows”. That is called good Software Engineering.
FFS … apparently you are in Software Development … and you come out with this nonsense seriously?
The rest was tl;dr … but it the normal excuses as yo why they can’t design a driver ABI properly, which ever OS on the Radar seems to manage just fine with.
As for portability … Desktop is pretty much x86, if another competing instruction set comes out and actually gains traction there will be binary drivers produced for it as well … so who really cares except for the GNU warriors?
As for the Nouveau driver … Seriously the guy is rocking a Quadro card?
Edited 2011-08-04 10:27 UTC
It is perhaps good Software Engineering, but there is absolutely no doubt that the driver code is designed for Windows, and then adapted and pressed into service for Linux.
So are you suggesting they start over entirely? Instead of taking porting the common components over (which makes perfect sense).
Having a set of common components that can be moved between OS does not mean it is “designed for Windows”, It means it is just well designed in the first place and the developer actually put some thought into it before actually coding it up.
Edited 2011-08-04 11:54 UTC
No, I can’t see how that would be cost effective, since nVidia aren’t expert at Linux, and Linux represents perhaps only about 8% of desktop usage anyway.
http://chenlevy.devio.us/writings/2011/08/03/count_linux_users.html
To me the best approach for nVidia would be to simply publish programming specifications for their GPUs.
Like this:
http://www.x.org/docs/AMD/
No-one can actually build a competing GPU from programming specs, so no IP is given away.
Then, open source driver developers (perhaps Xorg) who are expert at Linux can write a suitable open source driver for Linux.
Everybody wins.
Edited 2011-08-04 11:59 UTC
That would be nice. But from what I understand, nvidia licenses third party ip and agreements may prevent specs from being released. This is a similar situation to UVD. AMD will not release specs to UVD because of two reasons:
– specs might allow people to bypass decoding DRM (it would be an improvement, but others clearly don’t think so)
– third party IP.
Edited 2011-08-04 12:31 UTC
Programming specifications are not real IP, they do not require a skilled author and a creative process. They are simple facts.
They may however be kept as “trade secrets” in order to provide a small level of “security through obscurity” to protect DRM.
You must be new to the game? Programming specifications are written by someone – implicating copyright – which is IP. Also, documenting how a technology you licensed from a 3rd party works, giving insight to it’s internals, is not only IP infingmenet (they literally copied something from someone else), it certainly is a blatant licence violation. Or do you really think no NDA’s were involved when licensing such technology? Therefor, releasing ANY information without explicit permission of the 3rd parties will be an IP violation. Like it or not.
I don’t think so either. But, third party IP providers might specify obscurity as part of their licensing agreement.
Edited 2011-08-04 15:04 UTC
Indeed, they might … but to what end one has to wonder. !!!
How is anyone served by obscuring programming specs? I will point out once again that no-one can build a competing GPU chip from the information provided in programming specs.
Also an interesting question would be … who exactly would benefit if programming specs for hardware devices were obscured and kept as trade secrets?
That, my friend, is the right question.
Edited 2011-08-04 23:22 UTC
Except Nvidia because usually the Open Source Drivers are substandard (Remember with Firefox 4, GPU acceleration was not recommended to anyone who had anything except for a Nvidia graphics card) and it would take years (like the AMD driver) to get anywhere near the performance of the Proprietary Drivers.
It is probably cost effective for Nvidia, because they have a lot of customers in the 3d Graphics market that use Linux. This is where they sell the most expensive accelerators.
To both of you: Re-read my comment please. I don’t have an ATI card at all. The Nvidia Quadro chip in my laptop isn’t giving me any real issues with KDE apart from being a bit slow and some flickering (the “bugginess” I referred to). I simply prefer Xfce, and formerly Gnome, on that machine for the speed and lack of graphic anomalies.
On my desktop with Intel graphics is where the KPs and lockups are occurring. On that machine, ONLY Xfce and regular WMs like Fluxbox, Openbox and Enlightenment work well.
Well,
There you have it. Some versions of the intel drivers misidentify support for ogl features that are either buggy or don’t work. As a result compositing assumes that they work and hence crashes and other nastiness happens.
That was pretty much what I had found out, and there’s a KMS value I can set that might fix it. I’ll attempt that and see.
Although the Linux drivers for Intel graphics GPUs are open source, Intel write them. They are Intel’s code.
In many respects, Intel’s Linux code sometimes leaves a lot to be desired. Intel have been struggling for some years now to come up with a decent Linux GPU driver.
For some of Intel’s chipsets the Linux drivers are truly abysmal.
http://en.wikipedia.org/wiki/Intel_GMA#GMA_500_on_Linux
http://www.phoronix.com/scan.php?page=news_item&px=NzAyOQ
I generally like to support Intel because they do support open source, but in this one area I have to note that Intel’s performance in the provision of open source code has not been commendable.
PS: The point is, this problem is still far, far more likely to be caused by the GPU driver than it is by the KDE desktop. Get the graphics drivers fixed, and KDE4 runs like a charm.
Edited 2011-08-04 22:55 UTC
So, which is it? In a previous post you said Intel and AMD’s open source drivers are far better than the Nvidia drivers, open source or closed binary. Now you’re saying Intel drivers are crap.
I think you just like to argue for argument’s sake.
And for the record, I really like Intel’s open source drivers in Linux, and I also enjoy the fact that they as a company are willing to help the F/OSS community. I don’t like having to use the binary Nvidia drivers, but Nouveau just isn’t quite there yet for 3D and direct rendering. I haven’t had a machine with ATI/AMD graphics in years so I won’t comment on them right now.
For the record, here is what I actually said when I mentioned Intel and AMD open source drivers.
This statement does not compare Intel and AMD’s open source drivers to the Nvidia closed drivers, it compares the open source nouveau driver for nVidia GPUs to the open source drivers for Intel and AMD. It is essentially a comment on the difficulty of making adequate progress via reverse engineering.
Fair enough. Quite a few of Intel’s GPU drivers for Linux are quite reasonable … just bear in mind that there are some GPU models for which Intel’s drivers for Linux, unfortunately, are utterly abysmal.
Not there yet for gaming and other areas of 3D and direct rendering, I agree, but nouveau is fine for the purpose of running the KDE4 composited desktop.
I have several machine with different ATI/AMD graphics, and they all run just fine. Perfectly stable. Performance is still being tuned so there is still a little way to go in that respect, but nevertheless IMO they are rapidly becoming the best solution for 3D and direct rendering on Linux.
Edited 2011-08-05 01:56 UTC
Sorry I missed the word desktop in the section where you were commenting on the kernel panics.
Read my comment again. I’m only experiencing the KPs and lockups on the desktop with the Intel drivers. As the only drivers for Intel chips are the open source ones, I can’t really try any others besides VESA, which I’m sure would be terrible performance.
If KDE4 is “perfectly stable”, why is it the only thing on that computer I’m experiencing KPs with? Gnome3 under fallback works fine, Xfce is rock solid, Fluxbox and Openbox are amazingly fast and stable. I’ve researched the issue and it is a KMS thing with the open source Intel drivers and does indeed only affect KWin. Try to know what you’re talking about before you open your mouth and prove yourself a jackass.
I think your problem with KWin is when it’s running with composite effects turned on. Have you tried to disable KWin effects?
You can do it two ways: from System Settings, and with a keyboard shortcut ALT + Shift + F12.
Have you checked your mesa installation, it’s ok the output of glxinfo?
Mesa is fine, and I think you’re exactly right about the compositing. The problem is, I can’t even get the K desktop to load fully to be able to turn it off. It dumps on me right after logging into KDM.
As I said in other posts, I’m going to try a KMS setting in Grub when I get home later this evening. It should get me in, and if compositing is still an issue I can turn it off at that point.
Either way, I think I’ll be most happy with Xfce for now.
It is shame but Windows 7 would work perfectly fine on either machine … Accepted the p4 would drop down to Aero Basic Graphics … but it will still work fine.
EDIT: Reread you comment and missed the Windows 7 bit … Sorry.
Edited 2011-08-04 05:55 UTC
No problem, and you’re exactly right: Windows 7 works amazingly well on the laptop. I only have one Windows 7 license so it’s Windows XP on the P4, which is absurdly fast and stable on it. Granted I won’t have support for that OS much longer, but I only use Windows for gaming on the laptop and to serve media files to my Xbox 360 on the desktop. If I find a way to get Linux to serve media to the Xbox (Samba maybe?) I’ll ditch Windows XP for good.
I have a netbook which has an AMD Ontario Fusion chip, a combined CPU and GPU, which is called an APU by AMD.
http://techreport.com/articles.x/20415
“The Aspire uses the C-50 APU, an Ontario part with two cores and a 9W thermal envelope. Each core offers 512KB of L2 cache and a 1GHz clock speed, while the integrated Radeon HD 6250 ticks along at 280MHz.”
The CPU part is 64-bit dual core, but it is clocked at only 1Ghz.
Despite this modest CPU, because it has a reasonable GPU (significantly more powerful than Intel GPUs) the netbook runs the KDE4 desktop superbly well. “absurdly fast and stable” would be a fair enough description.
As for Windows 7 purportedly working amazingly well, that too is a very subjective opinion:
http://i.imgur.com/FJVn7.png
Edited 2011-08-04 11:49 UTC
To be fair, this netbook of mine dual boots to KDE4 and to Windows 7.
Windows 7 is stable and runs moderately well on this hardware, but it is quite pedestrian (and no more stable) compared to KDE4.
Edited 2011-08-04 12:40 UTC
Sounds like a great machine! I do love the “hidden” power modern netbooks have.
No, no it isn’t. It’s a direct observation. I’ve been running Windows 7 and Linux (first Slackware, now Arch) on my laptop for nearly a year now, and it’s been a great experience on both fronts.
Look, I know you are very anti-Windows, based on your posting history, but please don’t try to use me and my good experience as a lever to bash the OS. I don’t care for Microsoft any more than I care for Apple when it comes to business practices and politics, but both of their current OSes (and really all versions of Mac OS) are, in my opinion, great at what they do.
(emphasis mine)
Surely you must simply still count iOS as Mac OS there; surely you can’t mean there, couldn’t have forgotten so quickly about some of the later Mac OS Classic releases? ;P (and basically also the earliest OS X ones)
Its anecdotal I know, but with desktop effects turned off kde 4.7 runs just fine on my whimpy little atom netbook. So its certainly possible to run on slower hardware as long as you don’t go overboard on the desktop effects. I think the problem is more, certain hardware is buggy as hell with kwin, rather than kwin just being a hog in general.
Gnome 3 of course won’t run at all.
If kde ever gets too much for me I shall probably drop down to xfce like everyone else seems to be doing. Though I might take to detour through Awesome first to see if I could get used to a tiling WM.
Why not? Most atom netbooks have Intel GMA graphics, which is perfectly capable of running gnome 3. Do you have something else?
Yeppers, I have Gnome 3 running well on an Atom netbook as well. Some animations may not run at 60fps, but window management is generally quite smooth.
Its an eeepc 1000HE, and when testing on a fedora live usb, its either 1) does work and I get a black screen 2) Boots into fallback mode 3) boots into gnome shell but its buggy and pretty much unusable. The uncertainty alone is enough to keep me from using it.
Edited 2011-08-03 20:27 UTC
damn it, I have been making so many typos today I should be banned from the internet. I meant to say 1) Does notwork and I get a black screen.
Might want to look at using Fuduntu … they are staying with Gnome 2 til fedora 14 EOL … Not sure after that .. you probably want to hit the forums.
Yeah I left Gnome just recently, I simply loath Unity and Gnome 3 … let’s just say with both of these I’d rather not use a tablet/phone/kiosk interface on my Desktop/Lappy.
I’ve been using KDE 4.6.x, not bad. I find it a bit buggier than Classic Gnome but the bugs are being worked out.
Unity and Gnome 3 is useless for my needs.
Funny. I just started seriously trying Gnome3 yesterday. I actually looked at parts of it and said to myself: Thom has to hate this with a passion.
It works great, but there are several usability tradeoffs I can see at play. You can sort of see the arrogance of the designers dripping across the screen. ( this is so great, you don’t NEED a minimize! Giant space at the top of the window? It HAS to be that way in order to make it easier to select the menu without triggering the application switcher thingy). If you just focus on Getting stuff done, its not terrible. If you actually pay attention to the use of screen real estate or find use cases for minimize, its kinda cringe worthy.
You do realize you do not need to use kwin in KDE, right?
Besides, it runs perfectly fine in my Intel GMA 950…
I doubt your system is less powerfull than mine…
Regards
kwin in KDE SC 4.7 has improved performance and reduced demands of GPUs via its ability to run OpenGL ES, to the extent that now it can run even on mobiles and handhelds.
http://kde.org/announcements/4.7/plasma.php
“The Plasma Workspaces window manager, KWin, has received extensive cleanup of its code and can now run on OpenGL ES supporting hardware, making it better suited for mobile devices, and also improvements for desktop users. KWin has a new shadow system, improving support for users of older hardware or hardware for which driver support for OpenGL is limited (Xrender backend). KWin’s performance improves noticably thanks to numerous optimizations in certain painting operations.”
If you can get KDE SC 4.7 running on your distribution, it might be worth another look at kwin even for low end graphics hardware machines.
We have a fall back atleast on the KDE 3 front http://www.trinitydesktop.org/ is a fork of KDE to keep KDE 3 alive.
All we can do is hope someone does the same with GNOME 2
My story is much the same but I left Ubuntu when Linux Mint 9 came out. Still on Mint 10 but I too am old.
I’d just like to point a few things. Gnome-Shell is not the same as Gnome3. The next version of Unity will be based on Gnome3 libraries. Second, Ubuntu 11.10 will have Unity2d as the fallback. Gnome-Shell will never be on ubuntu unless you install it yourself.
Are you telling me that on Ubuntu 11.10, if I choose Ubuntu classic (neither Unity 2D or 3D), that it will be Gnome 3, but it will look like the classic Gnome 2 desktop (i.e., not what most people think of when they see Gnome3). For instance, I will still have my gnome applets such as GWeather and the Frequency Scaling applets avaialable? That’s good to know!
No actually, sorry to get your hopes up. After this next release ONLY unity will be present, gnome2, ie gnome-panel, is gone for good and gnome-shell has already been laughed of stage. For Ubuntu 11.10 Unity3D will be default with Unity2D the fall back, no-others there except on the ubuntu-based distro’s.
“laughed of stage” ? Have the Ubuntu people ever looked or used Unity (in any form or shape) themselves ?
It never ran stable for me… ever… even gnome3/shell didn’t crash on me 2 seconds after loading the desktop on a clean ubuntu install.
Not that I like gnome3/shell btw… I find all the reinventing the desktop paradigm’s horrible failures. All are terrible for productivity. (and on KDE4 they have poor performance, crash happy subsystems and poor pixel perfectness there always seem to be weird pixels and redraw errors on everything in kde nowadays.)
It’s the same problem as with virtual-keyboards and touch interfaces. All nice and eye-candy, all fancy and shiny… but you cannot get any freaking work done… virtual keyboards are 66% unreliable and never fast enough, same goes for touch screens… ever seen somebody selecting a new stencil in photoshop with a touch screen ? no of-course not… for the same reason that nobody types reports or documents on virtual keyboards… it’s slow, it’s irritating and your thumbs will die long before you get to anything useful.
Sorry to rain on your comment… this is a more general rant…
I guess I really must be missing something because to me none of the desktops today are as productive as CDE (solaris) was or as functional as 3DWM (sgi) specially for their time.
(Yes this means XFCE is the only desktop environments that fits the bill for me atm, although I also happily live with gnome2 for my daily workstation)
Ubuntu 11.10 will fallback to unity-2d, not gnome 3.
You are never going to find a KDE 4.0 that is reasonable for you, since KDE SC is now at release 4.7.
If you used KDE 3.5.xx before some distributions released the not-ready-for-users KDE 4.0, then you will be pleased to learn that KDE SC 4.6 or later has every feature that you were used to in KDE 3.5.x, and a great deal more power and flexibility besides. KDE SC 4.6 or later is better in every way than KDE 3.5.x ever was.
A number of Linux distributions which are based on KDE, such as PCLinuxOS, MEPIS, Slackware et al, did not release KDE 4.x desktops until about KDE 4.4. Why are you, like Linus apparently, judging KDE solely by those distributions which released KDE 4.0 to users too early, and ignoring those distributions which didn’t release it to users until it was ready for users?
I’d really like to know … what is the thinking here?
Edited 2011-08-03 23:18 UTC
Obviously I meant the 4.x SERIES. I have tried recent Fedora, Mandriva, PCLinuxOS KDE versions with recent KDE versions. You had to know I (and probably Linux) were talking about the v4+ versions of KDE – i.e, after the 3.x series.
You said: “I left KDE after the 4.0 debacle, and have yet to find a KDE 4.0 implementation that is reasonable for me.”
If you found KDE 3.x reasonable, then [citation needed] as to what it had that is lacking in KDE 4.x now.
http://kde.org/announcements/4.7/
Edited 2011-08-04 00:03 UTC
You know, maybe he just doesn’t like it? Jesus but you’re a prick.
And no, I’m not trying to “start anything” with you, I just find your belligerence and god complex a bit disturbing. The dude has an opinion based on his own experience, and just as you did with me and others here, you are trying to tell him he’s wrong about what he experienced and wrong to feel the way he does, as if you were there watching and have supreme reign over his decisions.
I happen to like chocolate ice cream more than strawberry. Are you going to tell me that I can’t possibly be right, that strawberry is the ultimate in sweet frozen dairy goodness and I’m a loser/moron if I don’t eat strawberry just like you? No, because that would be a dick move and make you appear to be a lunatic and/or an asshole.
Citation fucking needed? No! He doesn’t have to give you shit. He simply doesn’t like it. Now quit crying over it and man up, Nancy.
Everyone is, of course, perfectly entitled to their opinion.
What is not acceptable is to state that their opinion is based on a particular reason (in this case that KDE4 had lost functionality compared to KDE3) when that is, in fact, not true.
I only seek to correct the misinformation being propogated, it makes no difference to me what people do and do not like, or what they choose to use.
The reason why I want to correct misinformation is that other people might make decisions based on what they read here. If someone’s only reason for not trying KDE4 is because they had been lead to believe by what they read here that KDE4 had lost functionality compared to KDE3, how is that a good thing?
The idea, surely, is to help people make the best decisions for them, based on correct information.
That comment is so dripping with arrogance and self-importance it’s disgusting. Do you honestly think that you, and only you, can guide us poor feeble-minded lost souls into the One True Path? I was right about the god complex, sheesh.
Look dude, we’re just as smart as you, and I dare say some may be smarter. Have you ever considered that your average alternative OS user is apt to get a second or third opinion before committing? Assuming that you can “save” us from what you consider to be a bad decision is both unnecessary and, at least to me, sorely unwanted.
Pffft. Facts are facts, information is information. Correct information has nothing whatsoever with “guiding”.
A lot of people don’t have the time to check. Correcting misinformation merely saves them that time.
What, are you saying that you would rather everyone be fed misinformation? WTF?
Oh, and I almost forgot:
O RLY?
Once again you completely contradict yourself. Which is it, you don’t give a shit what people decide, or your mission is to help them decide?
My but you are confused.
Providing people with correct information is wrong … how exactly?
Providing people with correct information is helping people to decide for themselves, OK, but it is “giving a shit” about what they decide how exactly?
This will be interesting to hear.
With the big four.zero looming close overhead I’ve decided to resist the urge to settle in my old ways (at least when it comes to using WM’s, that is…) and use different WM’s on different machines just to keep me on my toes. *
I use IceWM and GNOME 2.0 on VMs.
I use GNOME 3 on a test machine.
I use XFCE and on my 1201n netbook.
I use KDE 4.6 on my development workstations.
Now, to be honest, GNOME 3 is at least 3-4 minor releases before being anywhere close to being stable and its very unlikely that I’ll use it on my main workstation(s) – though you never know – lets see what the KDE devs have in mind for KDE 5.0
– Gilboa
* Heck, I even write C++ code *gasp* from time to time, (instead of plain C) just to feel hip again
Edited 2011-08-04 08:19 UTC
I can’t stand what GNOME has become. I’ll keep using the 2.x series as much as I can, then switch to Fluxbox. Don’t even get me started on Unity…
I think Desktops and windowing environments are silly eye candy that should have been thrown out long ago. I see them like those cool interfaces I’ve seen in movies that are never implemented because of how impracticable they’d be to use.
I think a modal interface (eg. fullscreen applications) makes more sense.
I have used xfce since 2007. I can’t say I have ever been disappointed. It works slick and is lighter on the resources than gnome or kde.
well, i’ve compiled my first XFCE desktop when Olivier was just trying to release a “CDE clone” ( it was how many years ago ? ).
Never looked back… now running a xfce 4.8 desktop on slack13.37, and it even looks pretty.
Being able to setup Xfce like CDE is one of my favorite things about it. That, and it works like I think it should. I’ve never worked with a piece of software that is so easy to use.
I’ve been using Xubuntu since about that time myself. Xubuntu has everything you like about Ubuntu but with XFCE. It’s on ALL my systems, including my PS3.
Hopefully more people will realise the simple elegance that is XFCE
Yes. I also run XFce, it has a nice blend of speed and functionality. KDE4 is nice, but, it uses too much RAM.
Slackware 13.37 + XFce 4.6.2 = win
Anyone remember this phrase? A very common saying about 10 years ago. Now a decade later, KDE and GNOME are still going their separate ways, unable to agree on anything. People at Ubuntu are obviously smoking crack, thinking that “Unity” has a future.
It seems like Linux actually managed to take a step back in regards to the desktop.
The KDE devs don’t know how to design good user interfaces. To them, number of features and buttons equals quality. Their UI design is ugly as well, looks like something made for kids. What’s with those huge ugly icons in the panel? They look just like they did in 1999.
GNOME used to have potential back when they tried hard to mimic OSX. With 3.0 they have completely changed direction and while it does have some potential, using it feels very unnatural, even after a while.
In my opinion these two projects need to try harder to mimic the desktops of the two companies that dominate the market, Apple and Microsoft. The Linux desktop will never be a game changer, and they will always lag behind. Trying to reinvent the desktop has failed, time to copy someone who knows what they are doing.
Linus jumping from desktop to desktop is a good example of what I’m saying. None of the desktops are good overall. They do some things right but they all seem to be developed with developers in mind and not real users.
That used to be true. How long has it been since you used KDE?
It might have started out with that intent, but GNOME2 in the end is quite a bit different to osx.
Screenshot of a standard desktop KDE 4.7:
http://kde.org/announcements/4.7/screenshots/plasma.png
is it “post your desktop’s screenshot” week already ? whooohooo !!!
http://oi54.tinypic.com/2z4ck6q.jpg
/*
i refrained from screenshooting my wallpaper, since it might be considered NSFW, and the I/O is caused by a DVD-burning on background
*/
Oh yes … the KDE devs still don’t understand the idea of contrast in design …
The task bar is barely readable. Slightly Dark grey on medium grey … seriously?
Edited 2011-08-03 20:26 UTC
“Now a decade later, KDE and GNOME are still going their separate ways, unable to agree on anything.”
Actually, and stripped of your hyperbole, KDE and GNOME have agreed on a whole lot of things.
O Rly!!!
http://www.phoronix.com/scan.php?page=news_item&px=OTY5OQ
Yes Rly!!!
“KDE and GNOME have agreed on a whole lot of things” does not mean “KDE and GNOME have agreed everything”.
KDE called their Systems Settings “System Settings” first. How do you imagine GNOME would react if KDE had called their new Dolphin File Manager Nautilis instead?
Doesn’t change the fact they are bickering about rather trivial things, rather than just fixing it.
It isn’t a trivial thing at all to have two totally different packages with the same name within a distribution’s repository.
GNOME need to fix the problem they created, but they are bickering about it for some reason.
Changing the name is the trivial thing. Does it really matter if it is called “System Settings” or “KDE System Settings” … it pretty easy to fix but they are being arses for the sake of it.
Anyway it was pretty obvious that was going to happen tbh, since “System Settings” is a pretty generic name.
Also it doesn’t really matter who had it first … it is like two kids arguing over who did what first … it is pathetic for a group of adults …
Edited 2011-08-04 18:25 UTC
Does it really matter if it is called “System Settings” or “GNOME System Settings” … it pretty easy to fix but they are being arses for the sake of it.
There, I fixed that for you.
There are all kinds of laws which give rights to whomever was first. Copyrights, trademarks and trade dress laws are all based on who was first.
You wouldn’t want GNOME to act illegally now, would you?
KDE System Settings
Gnome System Settings
Oh I fixed the dispute … Seriously was it that hard?
FFS … argue the toss why don’t ya. You always argue the toss …
Seriously?
Whoever was first on a name so easily generic doesn’t matter … if it was apple or microsoft I am sure you would be the one complaining they are trademarking/patenting/copyrighing/<insert complaint here> …
Double standard somewhat …
I think with your “quote” you just summed up lemur2 quite nicely.
I have finally realized he is the ultimate troll: He starts out agreeing with you to gain your confidence, then finds one little thing to disagree with you about and drills it into absurdity. There are occasional cracks in his logic, but he glosses over them and continues the decent into madness, all the while claiming a superior intellect and experience level. He even resorts to deliberately misunderstanding the intent of your post to try to infuriate you to the point of fighting back.
The worst part is, he hides behind a lame excuse of a self-diagnosed personality disorder (see his profile), and somehow thinks that it validates his actions. In other words if you don’t “follow the rules” of how to talk to an INTJ, well it’s your fault he trolled you.
So I’m ignoring him from now on, even if he is genuinely nice and non-combative with me. One of my pet peeves is a self-important preening prick like that whose sole pleasure in life is making others angry or wasting their time. He’s had more than enough of mine, and this sentence ends my association with him.
… and your link doesn’t change the fact that they also have agreed on a lot of things.
Edited 2011-08-04 11:21 UTC
But they are still arguing about stupid shit … do we have to go roung in circles and fighting like kids.
The point is that professionals don’t do this … I work with plenty of third parties and regardless on how I feel about their business I remain professional.
Edited 2011-08-04 18:16 UTC
Sorry but that is not true. If you actually read the thread you linked you’d realise that a lot of the people posting there were constructive and trying to solve a certain issue at hand.
To assume that no discussion will get heated is naive. There are heated discussions inside any community so it is no wonder that some discussions between comunites can be a little heated too.
Edit: Just look at the recent Google news
Edited 2011-08-04 22:31 UTC
They were arguing about a generic name FFS … seriously this is like proper trivial.
TBH … the fact that it created so much grief … that it got to a Tech News website proves in itself that it was a serious spat.
Ok I have to say three things:
1. I really like Gnome3/Gnome-Shell. It is great for me, and not to mention my wife.
2. I really like the diversity in Linux Desktop/Window Managers right now and look forward to more diversity in the future. Gnome does not look or act or feel like KDE and neither of those like XFCE nor like E17. It is great! No one is stuck with just one like the other major OSes. We should celebrate that it is no longer the Gnome looks like KDE which looks like Windows 95.
3. Gnome3 has some great ways to open up multiple windows of the same application. Drag and drop to the desired workspace is great as well as right click on the icon and select open new window.
I like GNOME 3 on Fedora 15. Already got used to the way you are supposed to switch applications and keep doing that on Mac and Ubuntu (Unity) as well and then doesn’t work. Annoying
Sure, it needs lots of work. NetworkManager is not quite there yet and I miss the weather applet. But other than that! Awesome!
It’s strange to me how separated the group responsible for kernel development and desktop environments in Linux is.
um. Yea. You can run (and many large systems do) linux completely without a desktop environment. There is absolutely no reason why the kernel devs should ever do desktop development (unless they want to).
Edited 2011-08-03 17:34 UTC
I seem to just grok it. I hope people realize there are other options. Some of us do really, really like using gnome-shell versus all the other options. I’m stuck on Unity at home for the time being, and it kills me to not be able to use gnome-shell on it.
Why not? I’m using the GNOME 3 PPA and haven’t had any problems:
https://launchpad.net/~gnome3-team/+archive/gnome3
I did that during the beta but had problems with fullscreen gl apps (see: games). I will be trying again with ocelot.
I don’t blame Torvalds at all. The problem these days is that developers keep experimenting with UI changes without giving you a way to go back to things that worked out for you. The new trend is squeezing half-baked tablet UI’s into desktop PC environments. Eventually you’ll get tired of having to re-learn everything over and over and either stop using the computer altogether, or stay with older versions of software to get your work done in the way that works best for you.
I see interesting how GNOME 3 causes all these mixes feelings.
Personally, I love it.
“Linus Torvalds Not a Fan of Gnome 3”
Obviously…
XFCE is nice but KDE is nicer, albeit a little bit heavier on resources (not by much, maybe a few hundred megabytes). But when I use a computer with AMD Phenom II x4 965 BE and 8GB RAM, why not take advantage of it?
KDE 4.0 was a disaster in a sense. But it improved a lot since then.
It was primarily a disaster b/c the distributions moved all the normal users over to it – which was NOT what KDE desired or intended. (In fact, they pretty much told distributions NOT to do that.)
And, of course, if you use Fedora you get the latest untested bleeding edge. That’s what Fedora is for. Linus was using Fedora, and if he didn’t want to be a guinea pig, he should have known better and used something different.
However, ubuntu 8.04 LTS is NOT for that; but yet that is what was done.
Gentoo staved off KDE4 a little longer – not making the switch until KDE 4.3 was available. But it was one of the few (other than say RHEL, Debian, and possibly SLES) that did not do so so quickly.
And now it is easily the best desktop software available, bar none, for desktop, laptop and even netbook machines built in the last six years or so.
I started out using KDE when I first started using Linux a lot. When 4 came out I didn’t really like it so I switched to Gnome. Now Gnome 3 is out I have switched to Xfce. Its not perfect, but it has potential.
My biggest complaint with Gnome devs is that they think there is no need to ever change the way they think the desktop should be. The bar across the top of Gnome 3 is a complete waste of space. You get a hot spot, the current task, time, and system notifications. Thats it. In a world where screen real estate is getting wider but not taller, how is this an acceptable design goal. I can’t put any apps on the bar, and the clock is missing the built in calendar. On top of that the top bar of applications seems to be fatter too.
I don’t mind the new look or the new technologies, but I demand the ability to change the way things are set up. Why CANT I have another panel? Who’s it going to hurt? Why CANT I have NO panels. Would it be such a problem to make things a little bit flexible. I have a hard time respecting a project that seemingly doesn’t respect the end user.
Demand? Respect?
In GNOME 2 there were 2 panels, one in the top and one in the bottom. In GNOME 3 there is only one left. That’s at least a win
Maybe you can create some mockups for how *you* think GNOME 3+x should look? Who knows maybe it makes a lot of sense.
Don’t forget that you are not “the users”, you are just one
In Gnome 2 there were 2 panels by default, but I could add or remove as many as I wanted. I could have zero panels. I could have also have many panels. The whole point is that it is brain dead to think that they are going to make one desktop that EVERYONE likes out of the box.
+-+
So KDE SC 4.7 wins hands down then.
Everything on a KDE4 desktop is a widget. Panels, taskbars, menus, icons, shortcuts, folderviews etc, etc … they are all widgets under Plasma. You can “define” desktop configurations by adding, subtracting, re-sizing, moving and arranging widget sets however you like. You can save your desktop arrangement(s), and give it (them) a name. You can have as many or as few of such arrangements as you like. You can quickly load a given saved desktop arrangement onto any virtual desktop screen, at your whim.
You can make a KDE SC 4.7 desktop look and behave almost exactly like any other desktop at all.
It isn’t going to be a desktop that EVERYONE likes out of the box, but it is arguably the only desktop that users can arrange for themselves no matter how they like it. KDE4 definitely is the only desktop that allows users to have multiple different desktop arrangements (and shortcuts and folder views and files on the desktop) that they can rapidly change depending on the context in which they are working.
Edited 2011-08-03 23:47 UTC
if you need so much flexibility that you want no panel (which you can’t do with windows or osx), you can probably figure out how to install a different de. If not, there’s always xubuntu.
This to me is the funniest argument (and the most prevalent) against gnome 3 and unity: “not customizable enough”, because really, if you like customization that much, you’re free to customize the DE you use, the OS you use, the distro you use, and even the source code if you like. Why do customizers complain about defaults? It really makes no sense.
It’s about time you host a poll about how many like or dislike gnome 3, that should prove very interesting..
It’s not black and white though. I’ve been running it with F15 for a few weeks now, and while I’m *mostly* pleased with it, there are just a few things I can’t stand.
Some of those can be corrected by extensions (e.g having the clock show the date instead of just time), but I’m trying to give the stock version a fair chance before I start looking for workarounds…
Hate to say it but the last few years have really been a disaster for the Linux desktop.
We had KDE3. A complete, reasonably fast desktop environment that could run – not fast, but well enough – on a Pentium II box, given enough RAM. Now that’s considered “obsolete,” and what we have to replace it is both slower and less functional. (And also uglier by default IMO, but that’s another matter.) KDE4 is usable… But it’s definitely not what KDE3 was.
And we had Gnome 2. Okay, it was kind of slow, but it was very complete. Then it became “obsolete.” And now we have Gnome 3, which doesn’t do anything better, and in fact does a lot less in the vaunted 3D desktop mode. It’s usable, but it’s not pleasant to use.
Now we’ve got… Xfce, which works quite well, but doesn’t integrate all the stuff that KDE3 or Gnome2 did. And LXDE, which at least works, but is really a bit of a hackjob, and even less featureful than Xfce. For truly complete desktop environments, that can compete with the Windows or OSX desktop… There’s nothing at all right now.
I’m currently using KDE4 on my laptop and Xfce on my netbook. Xfce I have no problems with; it just is what it is, which happens to be something other than a complete desktop. But KDE I’m getting less and less impressed with. The new version of KWin is sluggish, compatibility with other WMs like Openbox is really bad, performance is highly dependent on graphics hardware, and it takes an annoyingly long time to start up… Longer, in fact, than any desktop I have ever used. Including Windows Vista.
Seems to me that the Year of the Linux Desktop has already passed. I hope the trend turns around, but I’m not feeling very optimistic right now.
P.S. Bonus for KDE: KOffice 1.x had pretty good interoperability with MS Office formats, and could save documents in several of them. KOffice 2.x can only save in one format, ODT. What the heck?
why do you say “desktops that can compete with windows and osx” and in the same statement call gnome and kde4 slow? Can you run windows 7 or osx on a pentium 2? I doubt it.
All these years I’ve been using GNOME as the primary DE for my relatively current home desktop. GNUstep/ROX-filer combo or XFCE for less powerful machines and mobiles. I’ve settled for LXDE for my home machines.
It’s lighter (ergo faster) than XFCE (GNOME-light IMHO). And most of the time my applications are in full screen so all I need is a taskbar and menu/launcher. Maybe I’ll find time to experiment once E17 goes final
one word guys: Awesome
yeah, awesome as in awesomewm
http://en.wikipedia.org/wiki/Awesome_%28window_manager%29
Im pretty sure that Linus is using Awesome or xmonad or something like that off the record :p
I’m sorry but Tiling window managers aren’t the end all for a lot of people.
Yes, they might work well for terminal windows and/or people with a lot of screen real estate, but I couldn’t possibly use one of these on a 13″ screen. At most I can fit two windows side by side (and definitely not a browser).
Sometimes I wonder what people with tiling window managers do with their computers…
Real Work^A(R). Which in internetland seems to be coding.
I was just wondering what Linus was using the other day, when I (reluctantly) switched to Gnome 3, which is what I’m typing on right now. I *knew* he wasn’t using G3, even though he’d gone Gnome when KDE 4.0 came out.
For some reason, I feel compelled to say why I’m using Gnome 3, and what I did to make it useable.
First things first: I have never really found *any* of the desktops to be really usable or attractive OOTB. I think the dock metaphor for both application launching and window/application management is better than any other by a LOT, and have thought this since NEXTStep. So every time I use any environment at all, I also run Docky, and this fixes 99% of the complaints I have with most of the linux desktop shells, including (and especially!) Gnome 3.
So now that we’ve established that my workflow will more or less remain the same no matter what shell I’m using: why am I using Gnome 3? *Process of elimination*, plain and simple. Once I don’t have to care about how my workflow might change, the criterea change to things like resource usage, speed, beauty, stability, and completeness (sound framework, session-management, integrated file i/o stuff, etc).
– LXDE/Openbox/Enlightenment/WindowMaker/EverythingElse: I like stability, minimalism, and resource-friendliness; but I also like features and completeness. The year is 2011, and I *expect* a 3D-accelerated desktop, font-smoothing, animations, and session-management (as opposed to simple window-management). So that knocks out everything except E17, and that didn’t last too long – simply too sparse.
– XFCE has some goodness to it – its like a fork of early Gnome 2 with a mo’ better panel. But its compositor is completely broken and S-L-O-W. Using Compiz instead of Xfwm4 resulted in similar performance and tearing issues, which I *do not* understand. And on my machine, once you really get rolling, it uses more RAM than Gnome 2, because many of the Good Apps are Gnome Apps, which require loading a lot of the Gnome libraries. And while it doesn’t look *bad*, it does look antiquated.
– I’d be using Unity right now, because panel-on-top, dock-on-left is how I’ve been doing it for a long time… but the Unity released with 11.04 was unstable, had a memory leak you could drive a mack truck through, (+2mb every *second* if you ran the system load indicator), and simply *isn’t* configurable. I can’t even take the Files/Folders lens off the dock without uninstalling the ability to search through files and folders from the entire shell! Its still in the running, but I won’t see it again until 11.10, and only then if I can actualy control what tiles go on my do… er, “launcher”.
– I just finished running KDE 4.6 for some weeks. I’ll be honest, I liked it. A lot. KWin, on *my* hardware, was at least as snappy as compiz, and I didn’t have vsync issues in my fullscreen movies like I do in Compiz. They went a LONG way with organizing the ocean of options in the Control Panel, oxygen-gtk makes GTK apps 100% transparent, and the apps seem far more capable and integrated with each other than in Gnome. But you know what? I hate Plasma. I hate Plasma crashes. I hate random app crashes (though no more full-on system freezes, yay!). I hate the 10-second startup load time. I hate the 800+ megs of RAM it allocates immediately. On top of it all, I think it is ugly; the thin buttons, the cluttered interface, yesteryear’s color scheme, the round and tiny window control buttons, the glassy, shiney, gradient-ridden widgets and windows – KDE *really* needs some artists and designers.
– Gnome 2 has no future.
– Gnome 3 compatability mode has no real future, either, and does not have feature-parity with Gnome 2’s panel.
– Gnome 3 was a last resort, because I object to it *on principle*. I HATE virtual desktops, which are now a critical part of their workflow. The “Activities” “area” is a *horrible* idea, causing users to hit that upper-left corner about a jillion times a minute to get any work done at all. Using the whole screen to show me what used to be presented to me in a drop-down menu is also a horrible idea. Removing the minimize button and feature without giving users a *functional* replacement is stupidity defined.
Its like they decided that rather than do work with your computer, your time would be better spent playing gesture-based mini-games with your windows instead.
BUT.
But, it doesn’t have a memory leak. Surprisingly, it takes up less memory than Gnome 2 on the same box, even with Docky, even after days of operation. I haven’t seen it crash this entire week, across three different machines with different video cards by different manufacturers. The 3D effects are pleasing, fast/smooth and useful (as opposed to gratuitious). Gnome-tweak-tool and extensions exist (and, in fact, make Gnome 3 bearably useful for me – without them its a no-go). There are some things I haven’t seen and don’t have in any other desktop, like built-in, keystroke-activated desktop recording. And IMO, by damn if they don’t have the most pleasing asthetic/artwork design of all of the desktops combined. I say this about Adwaita, and I hate the color blue, so that’s something.
I can deal with the included and disliked metaphors, because using a dock app allows me to sidestep most of them, and the tweak tools and extensions allow me to sidestep the rest. So I’m here, for now, but only time will tell if I jump ship to Unity or not in the future as they seem hard at work addressing my misgivings.
Edited 2011-08-03 20:05 UTC
I’ve been using Gnome 3 on F-15 myself and I have to say that you don’t ever need the mouse or virtual desktops.
Use the Windows key to open the Activities page and just type part of the name of what you want. I just wish it had Tab complete. Instead you have to use the arrow keys.
If you have Terminal already running, it switches to it. If it isn’t, it launches. Pretty handy.
I tried to use virtual desktops at first but it is a real pain without some way to pin an application to a particular desktop. I like hitting Win and typing Evo to get Evolution. I don’t like having to move down two spaces first to get it on the right virtual. Gnome 3 needs a way to make apps launch on particular virtuals before it’ll be useful.
People keep bringing this idea up WRT Gnome 3, and I’d like to address why it is one of the worst responses to the Gnome 3 design problem.
Mouse-free (keyboard-only) computing is something we already locked down thirty years ago with the invention of the terminal.
The whole *point* of a GUI is to *be* mouse-driven – _completely_ so – first and foremost. It is supposed to change the way we interact with our computers and the software on them, *from* a memory-and-memorization-based, keybaord-driven method, *to* an expose-and-select method. The point of a GUI is to *show* the user what s/he can do with the machine just by looking at it *while* making it easy to do these things.
If you have to remember that the name of your email client is “evolution”, then the shell has failed to live up to its most basic purpose.
While Gnome 3 doesn’t exactly “fail,” it really does score poor marks. It really *is* easier to navigate its shenanigans with a keyboard, which is bad, but what makes it worse is that there doesn’t appear to be an easy way to discover all the keyboard shortcuts it relies so heavily upon.
I am not opposed to *adding* keyboard shortcut wizardry to a GUI, but it is *not* a replacement – much less a suitable one – for a properly designed graphical shell.
There is an extension (windowNavigator, I think its called) that alleviates a lot of this pain .
Well of course I use the mouse or trackpad inside GUI applications.
For browsing for an unknown application in the shell I use the mouse also. But using the mouse is about 10x slower than typing three or four letters of the application I want. Even Alt-Tab is faster than using the mouse and dock on Windows, OS X or Linux.
I make it a point to learn the keyboard shortcuts for applications that I use a lot, including the shells.
In Windows 7 I almost always launch applications by hitting the Windows key and then typing in search. In OS X I use Command-Space and search.
For web browsers I open a new site with Control-L, type a piece of the URL, autocomplete it and I’m there. I don’t use the mouse to browse my bookmarks.
There is zero benefit in grabbing the mouse, locating the mouse pointer, locating the target and clicking it. This takes 5 seconds while hurrying. Using the keyboard takes 1 second.
Now all that said, I don’t see the problem with mouse navigation in Gnome 3. You seem to dislike not having the application list visible at all times. I see that as a feature. When I am using a program I’m not interested in what else is running. And I have a hard time trying to imagine any workflow that needs hitting the application list several times a minute.
For minimizing applications, it does take a bit of getting used to, but the virtual desktops do work for that. Stuff that needs to run but not be seen can be tossed onto another workspace. Personally I just leave that stuff running behind other windows.
To a new/inexperienced user, they’re *all* unknown applications.
This is true *if you know the name of the application you want*. Good UI design doesn’t assume such, however.
Don’t get me wrong, if you’re experienced with your OS and have a favorite set of applications you are already familiar with, then yes, the new GS/Unity/KDE search functions are pretty speedy. I am glad they are there, I use them too. The problem lies with this being *the* most efficient way to do things in G-S.
(More efficient than all of this: single-clicking one of a row of buttons on the left hand side of the screen. Just as fast as key-combination+first two or three letters of an app’s name, and I can do this one-handed… but we already know I cheat with a dock, and that’s just one reason why.)
Respectfully, I disagree. Alt+tab MAY be faster in the *one* case that the window/app I want focused next happens to be the very next one in line; otherwise, clicking a button on the left hand side of the screen is far quicker for me, and I don’t have to place a hand on the keyboard.
I usually do most of my light computing (email, browsing, system configuration, file operations) using only the one hand (on the mouse). To *me*, it is both a detriment and a complication to have to remove my hand from the mouse, and engage my other one as well.
Well, then allow me to explain it to you simply: it is about 2-3x more complicated to do *any given thing* in G-S than it was in G2.
And some things I simply can not do at all, most notably: minimization. There is *no way,* in G-S, to simply get a window out of your sight immediately. The next closest thing – the “recommended” “replacement” – is to hit Activites, grab your thumb’d window, drag it to another/new workspace, and exit activities. They’ve taken an act as simple and long-standing as simply clicking on something and turned it into combination of four actions, two of which being gestures.
And that’s just *one* example of how G-S takes simple operations that have been around since the dawn of WIMP computing, and inflates them into a complicated, near-dance-like series of contortions. I have many more.
What’s really funny about this is that I generally agree with the idea that minimization should be gotten rid of, as it *can* be confusing to some. They can take away the metaphor all they want, but why can’t the replacement metaphor be something as simple and quick, like, say, a “sink to below workspace” button instead?
MY workflow requires exactly that: Eclipse, a file-manager, one or more terminals, and a browser. I’m a developer, and I flip between these four – three of which are usually useless unless they take up a majority of the screen – several times a minute, possibly more. And that’s not counting communication apps, like IM’s and Email that I am also using nearly constantly, and can *not* ignore in the name of “single tasking is better for you, trust us!”.
More to the point, however, is that a good UI should be configurable enough to adapt to YOUR habits and workflows, no matter how “unimaginable” anyone else might think they are. A poor UI makes YOU adapt to IT, and that seems to be exactly what G-S is shooting for.
It’s not that there’s anything inherently *wrong* with G-S – after all, I’m using it (and with the right supplemental programs its the least obnoxious and most functional of the Linux desktops I’ve tried so far). It’s just that it is less convenient, efficient, and adaptable in almost every conceivable way when compared to what it replaced. Perhaps this will get better with time, but if it does it will likely be because the user base makes enough extensions and tweaks to save it.
Excellent post once more! Can’t mod this post up, as I “have already moderated this user recently”. (I don’t subscribe to that metaphor, I think I am moderating posts, and I can only do a limited number of posts from 1 user…. This is an example where OSAlert’ website metaphor is broken I think.)
If a mouse click costs you 5 seconds, then you obviously haven’t learned to use it well enough, and thus underestimate its usefulness.
I have a scenario here (for people doing software development). Let’s say you are at the middle of programming and have modified 10 files. For the next commit, you are only going to commit 2 of those 10 files. You do not know the paths of the 2 files off the top of your head, but you can tell straight away by glancing at the hg/git/svn status output (or whatever your IDE shows to you). What would be your workflow, without using a mouse?
Edited 2011-08-04 13:06 UTC
Please read more carefully, or quote the relevant portion of my message.
I wrote: “Well of course I use the mouse or trackpad inside GUI applications.”
And when I wrote “There is zero benefit in grabbing the mouse, locating the mouse pointer, locating the target and clicking it. This takes 5 seconds while hurrying. Using the keyboard takes 1 second.” it was just after two paragraphs about launching applications or using a web browser.
As for your specific example about picking files to commit to version management, I use the command line. Tab complete is a wonderful thing. And if I felt ordinary tab complete was too slow, I could add a bash completion function for SVN or GIT which would only tab complete modified files for “svn commit” or “git add”
I use tab completion all the times (I hate it when “/etc/in” doesn’t tab complete to “/etc/init.d/” in Ubuntu), but in this particular scenario, I can just type “hg commit” while glancing at the status output, then double-click-select-middle-click-paste the 1st file with my mouse (huge target to select, even larger target to paste, so no precise mouse movement needed), press spacebar with my left thumb, then double-click-select-middle-click-paste the 2nd file. In any day of the week, I can do this faster than tab completion with keyboard.
The thing is, if a programmer doesn’t embrace mouse as an inputdev that can improve productivity, he or she may not even know that such option exists. Then, it becomes a self-fulfilling prophecy where everything to do with a keyboard = fast, while everything to do with a mouse = slow.
Nonsense – the GUI was invented before the mouse was, and the first such environments were 100% keyboard controlled. No, the purpose of a GUI is simply to let the user get their work done, whatever that work might happen to be.
Fair enough, but it’s not too hard to work out that the icon labeled “Check Email” runs your email program, without having to know what it’s called.
Well, sure, if you count what Xerox did, then I guess you have a point. I guess I should have specified and said “the GUIs that actual customers could purchase and use at home back in the day”. Yes, even you, GEOS on the C64, Apple II, and Windows I.
No, it USED to be that easy. My exact complaint is that now, in G-S, its an utter mess.
In G-S, it’s gesture on over to “activities”, click on the “applications” word (because all you’ll see initially is a mapping of any windows you might have open), and then look for something that looks like an envelope from within a scrollable, screen-wide schmorgasboard of icons which are in alphabetical order by application name (lucky, in this case, that “e-Mail” and “evolution” both start with “E”. This is not usually the case).
And THAT is the root of my complaint. From “Applications -> Internet -> Email” to what I described above. I have no idea who thought that would be a good idea, but at risk of sounding like a broken record: I’m glad I have a dock to bail me out into something sensible.
Plus imagine the shivers, the repulsion, some (too many) people might get at the sight of e-word …no way in hell ( ;p ) they’ll ever click it / that was a very poor choice of a name.
In what way are they critical? Sure, they’re more prominant than in previous versions, but if you didn’t use them before, why can’t you just not use them now?
One reason is that “use workspaces” is now the G-S answer to “hey, what happened to the minimize button” and about a dozen other usability questions that arise. G-S was simply designed with workspaces at the heart of How Things Go.
Thankfully – and, again – there’s always Docky.
At least Steve Jobs probably like it, as Apple seam to have borrowed most of the new features in MacOS-X Lion from Gnome 3.
Give the Gnome developers some slack. The first versions of KDE 4 was far from perfect, and now a few versions later it looks better and better. The same thing will happen to Gnome 3. Give it time.
Sure you can give them time. But it is quite apparent that gnome shell wasn’t designed to handle any extra complexity. Who knows, I could be wrong, but the way some GNOME developers carp on about their design philosophy it almost seems like they think it is quite close to ideal already.
Edited 2011-08-03 21:11 UTC
Having enjoyed the Gnome2 setup for some time, gnome3/unity have completely taken the wind out of my love for the DE, I’ve been checking out KDE 4.7 but I’ve fallen in love with LXDE, while it’s a bit rough around the edges, it’s much more sutable for my day to day use then Gnome3 or Unity.
KDE 4.0 made me go back to Windows.
It may be Linus, but still just a personal preference. If he puts forth the opinion that strawberry ice cream is better than chocolate, will that be news too? Important thing is that we have different options, and have the luxury of arguing various DEs against each other. With Windows and OS X, you’re expected to do it their way whether it suits you or not.
Pretty much. If its in any way related to tech Linus’s opinon always make headlines (in our very small corner of the web). It doesn’t hurt that it seems people are always itching for a good DE war. If he has mentioned his browser preference we would be having a Chrome/Firefox debate right now.
I think it’s his hobby… whenever he’s bored, he throws a firebomb onto the net, to liven things up…
Indeed he does. “Digg users – you’re
all a bunch of Wanking Walruses” was my favorite one lol
I think mr.Linus is just normal person, like me. I switched to gnome after kde4, and now i`m with lxde after gnome3. Both are now useless and heavy crap!
Linus feels like me. KDE 4 and GNOME 3 are a fiasco. Xfce is the only decent DE left.
That is, if you want to use Linux at all. I have switched to OS X as my main OS. There are still 3 or 4 Windows programs I need. Problem solved with Parallels or with Boot Camp.
http://www.zimagez.com/zimage/screenshot-08042011-124817am.php
Good to see there’s some growing interest for Xfce. KDE4 is a fiasco alright. GNOME is however getting better except for the “fisher price” UI of Gnome Shell. So let’s have the best of both worlds : GNOME platform + functional desktop = Xfce.
I never used GNOME2 because it uses far too much RAM for what it does. KDE3 was quite ok, but felt like an unfinished business which then turned into a the major UI fiasco and memory hog known as KDE4.
Can’t say I’ve ever really cared what major figures in the open source world use for their desktops. Nor ayone else for that matter outside of my own machines. At the end of the day, Linus is just another user like the rest of us. One who’s exercising the amazing choice the FOSS world affords us all to choose whichever tool we find fits our needs the best.
Openbox + xcompmgr + GLX-Dock + gkrellm and Thunar provides a pretty good lightweight DE when configured properly.
Mine takes up about 160MB on loading, enough eyecandy with a Mac like dock, but is practical for real work and very fast. Set it up to find out how to avoid going to Gnome 3 or Unity when my Ubuntu 10.04 comes to the end of support.
Personally I found that stock Gnome 3 is completely unusable for production! Task switching is a shore!
(Although with Avant Window Navigator (mac-like dock with all the mac-dock features), this gets much alleviated)
The real problem is configuration, you can^A't configure anything, its even less configurable than Mac OS X!? And to think Gnome was such a simple and configurable desktop… Somehow someone thought that a “tablet” like UI was a good idea for the desktop (looks a lot like WebOS in my opinion).
It’s a point zero release. Configuration options will be added in as it evolves. Gnome 2 was even worse in this respect when it first came out.
Yes Going from Gnome 1.4 to Gnome 2.0 drove me to xfce 3.
It’s one thing to have stability issues or poorly implemented ideas in a point-zero release.
But when did no longer being able to do or configure something at all become an acceptable trade-off for running a point-zero release?
It’s a clean slate, in the most literal sense it’s not possible to no longer be able to do something you were never able to do with a given product in the first place.
I’d rather they have a limited set of cleanly implemented features as a starting point than a more “full” set of half done or buggy features. In either case, my advice would be the same. If the features you need aren’t in the current release, no one’s holding a gun to your head forcing you to migrate.
Edited 2011-08-04 21:45 UTC
If that were true, it wouldn’t be called “GNOME” 3, it’d be called “New Product Clean Slate Thing v. 0.1”.
There is a reasonable expectation that comes with calling the software “Gnome”, and that expectation is that this is the next version of Gnome… not a “clean slate”.
I’d rather they release when they reach feature parity with the previous version of the product. This current mindset of “a limited set of cleanly implemented features or bust” is a disease that doesn’t last five minutes under scrutiny.
Whether or not I, personally, migrate or use G-S is separate from the discussion about whether or not releasing too early, before maintaining feature parity, is a mistake. Making it about me instead of about the topic at hand is defensive and counter-productive.
The point zero should clue people into the “new beginning/paradigm” part of things, just like it did with Gnome 2, if it were a continuation it’d be 2.XX
NO other software product in the history of sotfware development releases *to any success at all* if it is less capable than a lower-versioned product. It does not matter if the version bump is major or minor. It is simple common knowledge, backed with forty years of experience, that “Version 3.0” of a given thing is generally understood to be “Like Version 2.0, only better” – where “like” means “can do all the stuff version 2 could do”.
This crux of my disagreement is, in fact, that this practice of treating a dot-zero release as a completely different product *is an error*, because it flies in the face of common knowledge and about forty or fifty years of software development.
Just look at some software development history if you doubt this:
– KDE 4.0 release, which was by *every* account a disaster from every angle. Biggest complaint: “It can’t do a LOT the stuff 3.0 can do! How is that possible? Who thought it was a good idea to release a 4 with an inferior feature set when compared to 3?”
– The Gnome 2 release (which was absolutely not a success, for the same reasons as above)
– Windows has maintained feature partiy with every release, AND they change the name of the product when the changes are too extreme
– The Mac OS changed the name of the product (OS X – from regular numbers to Roman numerals) coupled with a HUGE marketing campaign when they decided to draw a line in the sand and change the entirety of the OS and break backward compatability/feature parity. They did so because every time they broke backward compatibility/feature parity in the past WITHOUT changing the name of the product or doing heavy marketing, it resulted in financial disaster and notable public outcry (see: version 6 to version 7, and I think version 3 to version 4 before it).
So, NO, the “dot-zero number” should not, and DOES not, “clue people into the clean-slate-ness of things”. That strategy has *always been* a failure, *even in Gnome’s own history* – which makes their willingness to repeat such a blunder, and your willingness to defend it, even more confounding than usual.
It’s the way things are done in the open source world. Look at where Gnome 2 and KDE 4 are now, and tell me with a straight face that they failed. Yeah, the launch of a new development line can be rough, especially for those that don’t expect it, but it generally sorts itself out.
Just because the *project* still exists, doens’t mean that the *launch* wasn’t a failure.
But you know, since you asked for it, I can certainly tell you with a straight face that KDE 4.0 was a failure, and has yet to recover from that blunder. Its userbase has shrunk considerably as a result of that maneuver, and that launch is often mentioned in conversations about what not to do during a launch.
Only if you repeat the mistakes of others who have launched poorly.
You know what’s better than “hey, this hurts, but it’ll go away eventually”? How about “hey, this doesn’t hurt at all”?
For the life of me I do not understand why you’re defending an obviously inferior way of doing things.
GNOME3 has been better than any other desktop I’ve used since it’s 2.28 release.
I use windows 7 every day at work and find myself fling the mouse to the activities hot spot, trying to snap windows to the left and right sides of the screen.
I play WINE and native games that don’t suffer because of the shell. I feel really comfortable in an environment I’ve been using for about two years now.
The shell has changed significantly since it first started but what I see on my screen every day is the best WIMP interface going.
What do you mean trying to snap windows to the left and right? You mean dragging them to the sides so they snap to the edge of the screen and take up half of the screen? aka, easy tiling?
That was in windows 7 since it was released. Gnome3 copied the functionality /exactly/.
I suspect Linux really just wants a solid, fast, simple desktop much like what we used to have. Windows XP, KDE 3, Gnome 2. That’s a personal preference, he doesn’t see the benefits of what the new desktops do, and to a degree he’s right.
XFCE is probably the best solution for him. It would be interesting to hear his thoughts on KDE now that it’s a little more mature, though, and brought back a lot of the missing functionality 4.0 lost.
I wouldn’t call KDE4 “mature”, sure it has improved a lot since 4.0, but I’m still able to crash plasma by just resizing down the panel in KDE 4.7, this is far from “mature”, it’s lame.
And I suspect Linus answer to that question would be “Sure it has matured a lot since 4.0 but it’s still not as stable as 3.5 or GNOME 2.32 was.”
Edited 2011-08-04 01:58 UTC
KDE SC 4.7 has all of the functionality of KDE 3.5.x. No-one who makes the claim that there is still “lost” functionality from the 3.5.x series to the 4.x.x series is ever able to actually say what it is.
In other words, [citation needed] if you want to make such a claim.
KDE SC 4.7 has so much functionality that you could very probably configure (and save as an activity which you could make the default or re-load at your whim) a desktop setup that you named “XFCE mockup”. I’ll leave it to the readers to guess what “XFCE mockup” would actually look and behave like …
Edited 2011-08-04 05:21 UTC
The point is during that time frame where they were putting the missing features back in … they lost the trust of the user base.
http://www.osnews.com/permalink?483459
A number of KDE-based Linux distributions lost no KDE functionality at all at any point. Why are you blaming KDE for the actions of some KDE-based distributions which jumped the gun?
Right now, a number of GNOME distributions are jumping to GNOME 3.0 or switching to Unity. Some other GNOME distributions, such as Mint, are sticking with GNOME 2.xx. Ring a bell? Deja Vu, anyone?
PS: The word from the KDE developers when they released KDE 4.0 was … “it is missing a lot of functionality, but we wanted some user feedback about what was implemented so far.”
The word from GNOME developers seems to be … “you don’t want overlapping windows. You don’t want a minimise button”. Hmmmmm. What was that about “the trust of the user base”?
Edited 2011-08-04 06:34 UTC
Still doesn’t change the fact … by many of the comments on this thread being an example, that quite a lot of the user base were alienated.
Not really: stability and low resource usages are *major* functionnalities, and some users are apparently not getting the same experience from KDE 4 as they had with KDE 3.5.x, as I read quite a few complaints about Plasma, Strigi/Nepomuk..
Now that it works with OpenGL ES, by sharing the load between the CPU and GPU KDE4 works now with lower resources than KDE 3 did.
Strigi does take some time to index the local drives (less than Windows desktop search, but more than no search at all). If you don’t want it, simply disable it and it won’t run.
KDE4 is perfectly stable and has been for some time now.
Edited 2011-08-04 09:53 UTC
He might be referring to RAM more than anything. KDE *is* the heaviest desktop around in this respect.
Neopmuk kept summoning a crash dialog for me. I never had a problem with Strigi, but never found myself using it either, so I eventually turned it off.
I wouldn’t say that.
I *will* say that the base desktop is pretty rock solid, if you ignore a few driver-related issues (re-sizing a terminal window in KDE 4.6 would hard-lock an nVidia based system. True, nVidia fixed the problem in their drivers, but one must wonder why *only* KDE exposed this issue.)
However, some of the software certainly is not stable. Plasma, which I wish I could somehow disable entirely, often crashed and re-started itself… sometimes even when I wasn’t even sitting at the computer. If I’ve done anything that required Sudo access and then try to shut down the machine, I got a wallet/polkit-1 crash window. Neopmuk issues. Etc. etc.
It’s a far cry from the machine-freezing KDE 4.0 of old, and I’d agree that the desktop environment itself is pretty solid… but to call the whole software collection “perfectly stable”? Nah. I can’t agree.
It doesn’t matter how a GPU driver bug is exposed, it remains a GPU driver bug.
Across several machines I haven’t had a single issue with KDE4 crashing, stopping, freezing or even hesitating since KDE 4.2 (Jan 2009).
I can’t say the same for Windows 7 in the same timeframe. This is a bigger statement than it seems at first glance, because the machines in question were all designed to run Windows.
Admittedly I had to go without a composited desktop for KDE 4.2 and KDE 4.5 because it was slow if I enabled compositing, but once again these were GPU driver issues, and I didn’t want to use fglrx.
However, it is now KDE SC 4.7, and the whole desktop software collection is indeed perfectly stable. Moreso than Windows 7 on the same machine.
Oh, and KDE SC 4.7 is fast, also. My Acer Aspire One 522, for example, has only 1GB of RAM, and KDE SC 4.7 runs in that limited resource far better than Windows 7 on the same machine.
Edited 2011-08-04 23:28 UTC
From a high-level, in-theory-only perspective, I agree.
In *practice*? It certainly matters if only one out of many desktop environments are affected. Why is it that everyone *else* holds stability even in the face of driver problems as more of a priority than KDE?
My only experience is with 4.6. KDE itself was very stable; some of its apps and helper programs, not so much.
1) No one can blame you for not wanting to use FGLRX.
2) Going without a composited desktop makes the competition invalid – apples to oranges. If you can run Compiz, xcompmgr, and Win7 on the box, but *not* KWin-with-composite because of problems, then KDE is clearly trailing behind in an apples-to-apples comparison.
No KDE hate here, I quite like it. I just don’t agree that it can be said to be on “perfectly stable” or on par with all the other environments regarding stability and features if you have to disable a feature to keep it stable and responsive.
I kind of want to see the new OpenGL KWin stuff.
True, KDE 4.2 and KDE 4.5 both suffered performance-wise in apples-to-apples comparison with other composited desktops. Whether or not this was due to KDE, kwin or to GPU drivers is highly debatable.
This issue really only affected performance, it didn’t affect stability or functionality. I was talking about stability when I talked about this.
One could say that KDE 4.2 and KDE 4.5 did lose performance compared to other versions of KDE because of these issues … but KDE is at version 4.7 now and the issues with kwin have been resolved.
To keep performance after KDE 4.2 a viable strategy for me was to skip KDE 4.5. I installed it as a trial, and I was able to make the decision to stick with KDE 4.4 in a very short time. When KDE 4.6 was released, I trialled that too, and I was then happy to upgrade.
Other people could have done different things, it is up to them, but if they are not interested enough to keep a close track on it, then correct information is what they need, not mere uninformed bagging of KDE.
Edited 2011-08-05 03:40 UTC
People don’t like the truth but there it is. “Now he’s using Xfce.”
Gnome and KDE both suck. Sorry folks.
Xfce wins.
We have something in common.
Like many people here, I liked KDE very much, but I left it after they introduced version 4. I switched to gnome, and I liked it very much – to the point that I found gnome plays nicer with the user than KDE did, and looks better. Anyway, I’ve always was interested in KDE project, because I believe they had good intentions but lack of manpower and funding to do proper testing and bugfixing, and when KDE 4.7 came along, I decided to install it on my desktop to check it out. What strucked me is that it has a lot better dual-monitor support than gnome, and – after getting rid of this horrible default settings – it can be just as friendly as gnome is. Not sure about what’s faster and lighter for the CPU, because I don’t see the difference. Now I’m a KDE user and I’m proud of it
Debian Squeeze, a 2.6.38 Kernel from backports. Problem solved.
The only thing that’s a small pain is getting an up-to-date browser.
But even that is quite easy to solve. Chrome sets up its own repository, Firefox can be installed by hand quite easily.
I will switch to Gnome 3 at some point, I tried it and I feel it’s not that bad. 3.2 will solve many problems and 3.4 could be fine enough to switch over.
But for now, I am completely happy with my setup – and I am planning to stay on it for at least another year or so…
.. that use it. I used to be mostly Red Hat (before the Enterprise Linux / Fedora split), but after having using Ubuntu on/off for 6.5 years, I’ve become accustomed to the ease of which Ubuntu sorts things out for me and I like the massive support community.
Fedora 15, on the other hand, feels like a beta release with very limited support, and the NVIDIA driver support is very poorly done. The installation still leaves the MESA OpenGL libraries as the default (causing crashes) and SELinux seemingly interferes with attempting to fix this.
I love GNOME 3 and consider it a much surer future bet than Unity, and so I hate that Ubuntu decided to go their own way here. Maybe we need a Gnobuntu.
The Gnome3 PPA is there, and its what I’m using in 11.04. The water is *just fine,* come on in.
…is fickle.
Someone who knows when he doesn’t like something, and has a strong opinion about it. Same as Torvalds, if he doesn’t like it, it will be known. Linus however can’t just enforce these changes, the other guy can (and does), and it’s making the company billions…
Anyway, I have much respect for Linus. Throughout the years, I’ve yet to encounter a statement of him I didn’t agree with. I however have turned my back on “linux on the desktop” completely, I currently only run an Ubuntu in vmware for development purposes, running a massive amount of consoles on my work laptop (running Windows 7). I haven’t tried KDE since the 3.x series, and the only thing I did on my Ubuntu running Gnome (no idea what version) is starting a terminal and launching QT creator, so I can’t really say if Linus is right or not, but I do tend to believe him.
I do love Linux though for everything else. For servers, it’s just a joy to work with, but the desktop experience? It has massively improved, but still feels limited and weird. It’s strange when you have powershell in windows, know it pretty well and then almost never use it, while on Linux, I do pretty much _everything_ in bash. On a mac, same situation as in windows, and there I have a complete bash shell. That must say something about the GUI right?
I find it surprising that linus doesn’t use a tiling window manager. Maybe if the xfce developers manage to mess up xfce enough he will give awesome or even dwm a try.
Is this David Adams trying to justify a vacuous post about one guy’s preferences in DEs?
Linus has a good eye for design, IMHO. Is he any less qualified than Thom, Eugina, or David?
No and I never said otherwise – because that wasn’t my point. The relevance of a vox populi piece would be tenuous enough (especially without a formal methodology), let alone the anecdotal account of one guy on Facebook, regardless of however much you happen to vouch for him (and lets face it, his field of expertise is not UI design). Adams not only presented a false choice but his two latter options were outright non sequiturs, while the first was, frankly, insulting (like that famous Gilbert Gottfried joke, minus the joke). This is clearly an attempt to think up an excuse to report on what a famous person x thinks about product/project y.
I was really excited about Gnome3/Gnome-Shell. I thought it looked great and was reading all about it pre-release. When it came out though and I actually started using it I came to hate it. Things like not having Shutdown and Restart appear in the drop down session/user menu, not having a minimize button on the titlebars, space being wasted with useless chrome for no reason. There is so much about it that I couldn’t stand. I used it for a while and grew increasingly frustrated. One day when I booted up it just hit me like an epiphany that “I hate this environment.”
I reluctantly gave Unity a shot. And ended up loving it. When I started using Unity I was expecting to not like it. But the more I used it, the more I liked it. It kept growing on me.
As of today, I can say without hesitation that I like Unity better than any other desktop environment I’ve ever used (linux, mac, or windows).
I think Linus is right. Gnome/Shell is a terrible design. I’m just so thankful that Ubuntu developed Unity.
I don’t really care what Linus says about a Desktop Environment… it is a matter of taste. I am a KDE user but I also use Gnome on occasions an even XFCE. All of them have their pros and cons. But I like any of them.
It is just a matter of taste. I thought the long flame wars between KDE vs GNOME were behind us, but they keep coming back
In the end it is just a matter of taste. Be happy and use whatever you feel comfortable with.
Edited 2011-08-05 05:01 UTC
With Ubuntu 11.04 I switched back to the gnome 2.3 classic mode which works fine for me.
I tried gnome 3 but I hate the way the screw around with the desktop. Over-large title bars and more bars and some extra bars which results in a very small usable window.
Most screens are 16:9 nowadays and don’t have that much precious vertical pixels left.
I dove off the good ship KDE when 4.0 debuted and switched to Gnome. I was a happy Gnome 2.x user for years until KDE 4.3 came out and I found it to finally have settled down in terms of features and stability. In my opinion, KDE since 4.3 has simply gotten better and better. I just upgraded to 4.7 and I’m finding it further refines and improves the overall KDE experience to a high degree. And, Kwin performance is finally back up to snuff! Of course, I use Compiz-fusion which I have always found to be faster than Kwin while offering more bells and whistles (I’m a sucker for eye candy.) I tried Gnome 3 for a day and found it to be completely unusable and insulting. Gnome devs apparently think its users are idiots and keep reducing (or hiding) functionality as if we were playing with Fischer-Price blocks! As for Canonical’s Unity desktop, I simply don’t like it because I feel as though it’s incomplete and not as configurable. I’ll withhold judgment for a few releases and see how it matures. Let’s hope Canonical’s developers are more receptive to user feedback than the Gnome developers are, apparently. While I always knew that KDE 4 would get there eventually (and I think it has come a very long way since its inaugural release), I have serious doubts about the direction of Gnome 3. I hope the Gnome devs read all these user comments all over the web and these forums and have a summit in which some major changes are made for Gnome 3.1 or 3.2. Otherwise, I fear users will leap to Xfce, LXDE or other lesser known (but more traditional) desktop environments in droves.
This year has been really difficult for me and for the linux convertees I always push around.
GNOME 3 as in Gnome-Shell turns to be a beautiful clever smartphone-esque interface, yet a frustrating experience for the power user. Want to see it? Open 10 applications at the same time and see how you go about with it. My suggestion is that GNOME guys go back to the 2 panel layout mode, just like GNOME2 was. The magic here is to take advantage of new technologies and continue the tradition of GNOME2. Perhaps make the “fallback” mode not a second-class mode but actually a definitive option for who don’t want a smartphone? It would be so easy to set up a ON-OFF button: Classic Mode or Smartphone mode? Simple as that! And please get rid of that horrible Cantarell font, which is ugly as hell. Give me back sane, logical, and consistent menus. Linus want it. I want it.
GNOME2 and its lack of maintenance is very discouraging to use at all at this point, not to mention some bugs still around for this platform.
XFCE, if only one person could find the solution to make it look like GNOME2, and I mean, an automated solution, to make it look more polished and so on, it would be a huge step towards the legacy of GNOME2, without being GNOME2. But I suspect no one wants to be like the other else.
KDE 4.6 has come a long way, and I tried Kubuntu for a audio recording workstation. It looks great, if it was not for the default settings – tiny titlebar buttons, dozens of “useless” applications to me, messy classic menu, and too tiny fonts. But spending one hour tweaking it, it came very very close to what a desktop should be like, very usable, elegant and simple. I believe if KDE fixes some mistakes and try to be a bit simpler, it could eat everybody’s lunch this time around. Fortunately it’s Kubuntu that I will be deploying.
UNITY, on the other hand, I hated since the beginning. The global menu never really made my mind, and I think UNITY will warm the heart more of Mac users than Windows users. I may be wrong. But since next Ubuntu will be either Unity 3D or Unity 2D, main stock Ubuntu is definitely over for me, as a distro. I think Mark Shuttleworth blew it, and blew it big time. Every convertee I had moved to Ubuntu in the past 3 years is now switching to Windows 7, all because of Unity, and not just Unity: but this stupid eager will to deploy unfinished and disfuncional software to meet deadline standards, which is a disease to me.
Edited 2011-08-08 15:19 UTC