Ever year, Linus Torvalds goes on vacation to Australia, during which he usually also visits linux.conf.au. During his stay this year he gave an interview to ComputerWorld, in which he talked about the success of point releases and the important topic of file systems in Linux, which is quite an active field today with ext4 and Btrfs. He also gave some insights into why he switched away from KDE, moving to GNOME instead, and he shares his thoughts on Windows 7.
Because of the inability to add support for ZFS to the Linux kernel due to the GPL being incompatible with Sun’s CDDL license, Oracle started work on Btrfs, a GPL-licensed file system that should bring ZFS-like functionality to Linux. Currently, Btrfs is in the beta stage, but support for it has been added to the mainline Linux kernel – despite its beta status. According to Torvalds, there are two camps when it comes to file systems: the one that wants stability, and the one that wants to release often. “You want filesystems to be stable, but you can’t be in beta forever,” Torvalds notes, “Btrfs is developmental, but it was merged in the main kernel to help people test it.” This obviously indicates that Torvalds and his crew of kernel maintainers acknowledges that Linux needs something to compete with Sun’s ZFS, and therefore, they decided it made sense to make it as easy as possible for Linux users to test the new file system and report possible bugs.
While Torvalds is critical of early netbooks, which according to him showed signs of “teething problems”, he is hopeful for the future of this new market niche. “I’m hoping the next generation will be more powerful and offer a better user experience,” Torvals said, “I was doing kernel development on a netbook and it was not at all horrible. The screen was too small, but we are getting to a stage where you can get a cheap good laptop.”
Moving on to Windows 7, Torvalds says that the big problem Microsoft has is the fact that they can’t do a yearly upgrade like Linux distributions can. Microsoft has to charge for their releases, and people won’t shell out an amount of money each year just to be up-to-date. The consequence is that Microsoft has to ave longer release cycles, which means more breakage between releases, which inevitably leads to customer dissatisfaction. He notes that “the cost of the pain is likely to be higher than the cost of the operating system which is why people are slow to upgrade.”
He does have some advice for Microsoft, though. He thinks they should decouple the operating system from the applications, and this is actually what Microsoft is doing right now by moving vital functionality of the operating system, such as photo organising/editing and the email application, to Windows Live, which has its own release cycle separate from Windows itself.
Torvalds also explains why he moved away from KDE. “I thought KDE 4.0 was such a disaster I switched to GNOME,” he explains, “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.” This is rather curious, seeing Torvalds’ advice to “use KDE” – which, coincidentally, was one of the most-commented stories on OSNews.
There’s more interesting stuff in the interview, so be sure to give it a read.
It would be great if Osnews actually bothered to contextualize the information that they are purporting to convery. Here’s what Linus actually said:
“I realise 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 re-try KDE, but I suspect I’m not the only person they lost.”
And then he adds:
“I got the update through Fedora and there was a mismatch from KDE 3 to KDE 4.0. The desktop was not as functional and it was just a bad experience for me. I’ll revisit it when I reinstall the next machine which tends to be every six to eight months.
The GNOME people are talking about doing major surgery so it could also go the other way.”
So, yeah, Linus didn’t heed or read the warning signs that KDE 4.0 was a developer’s only release and as an end-user of his distribution he had a bad experience.
In my opinion, KDE communicated this release poorly, which they have since corrected, and no self-respecting distribution should have shipped kde 4.0 as its default desktop. People should have had to go out of their way to select it.
But the pain was largely worth it. KDE 4.2 is such an incredible desktop environment that eventually kde 4.0 will be remembered as a necessary stepping stone towards a brilliant collection of software.
Edited 2009-01-24 11:47 UTC
No KDE 4.0 will be remembered as a mistake and an extremely unpleasant user experience, just as early releases of Gnome 2 are still remembered with bitterness.
KDE 4.2 is a vast improvement over 4.0 and the first one I consider even marginally usable. That does not make KDE 4.2 “incredible” though and there is much to be done before it catches up with KDE 3.X. KDE 4.X is a good direction but the team definitely did not communicate well and over promised prior to release.
The snippets OS News included in the summary were accurate. Linus moved away from KDE due to the awful experience of 4.0. Linus, like many long term Linux users, switches between desktop environments based on which works for him at the time. This does not change the accuracy of the quote.
The early versions of Gnome 2.x were really quite flaky but they simply had to spend some time pushing forwards some of their infrastructure. They just didn’t have choice. They’re desperately trying to find ways to not do it again, but at some point in the not-too-distant future they are going to have to. Are those early versions of Gnome 2.x going to stop people from using it now? No, and it will be the same with KDE 4 as it was to a slightly lesser extent for KDE 3.
With more manpower, more resources and if there was a real pressing need to maintain backwards compatibility (lots of binary applications written for KDE 3 and dependencies on it) then a different approach would need to be taken. However, there isn’t the manpower, the resources to do that nor the dependencies on KDE 3 to make it worth while. Gnome has pretty much decided that they want to maintain compatibility with 2.x, but they don’t have the resources to go for a completely new Gnome 3 and do both at the same time. The result has been stability, but stagnation. That equation will almost certainly need to change for any prospective KDE 5.
Linus even talks about this in relation to Btrfs:
Edited 2009-01-24 16:32 UTC
manpower?
If manpower is an issue why did they make Plasma?
If understand the reason to make Solid, Phonon, etc but Plasma seems to me really be a waste of time rewritting something for no particular reason instead of focussing on polishing the end-user experiment..
Just from the top of my head (and I’m (yet) no plasma or KDE dev, just using Qt4 to develop my apps):
– break up the old cruft that several parts of the ui became over the course of the KDE2 -> KDE3 years (panel’s source code being in bad shape and extremely difficult to maintain is prominently mentioned by aseigo as a main reason for starting plasma, for example)
– lower the barrier for new contributors (artists, usability specialists, etc.) to work on the ui without having to muck with the C++ base of the desktop
– break up the close relationship between the presentation layer and the buisness logic behind menus, taskbars, panels, etc. allowing the creation of different, tailor-made user interfaces while keeping the base code in a reusable library (libplasma fits nicely into this). Recently, there was a brainstorming session underway which dealt with new interface approaches for educational settings.
– enable KDE4 to reach new target platforms (media centers, PDA/Smartphone, netbooks, etc.) without the need to reinvent the wheel for all these different use cases every time from scratch (remember, KDE4 always had this additional emphasis on the “framework” part).
– keeping up with the expectations of users coming from OSes like Mac OSX or Windows Vista/7 (Besides, plasma is reported to run on Windows, albeit stability and performance may have to improve to be really useful there).
– Making use of the new or improved opportunies in Qt4 (graphicsview framework, Widgets-on-canvas, OpenGL rendered widgets, etc.)
– Plasma allowed the KDE4 devs to introduce the concept of an activity aware desktop, which is arguably (still) in an early stage, but should allow different approaches to the old “one person, different hats” problem. (And judging from my experience so far, it should be possible to keep it out of the way of the user if he/she does not want to use activities).
This is just from the top of my head, as always please feel free to correct me and/or augment this list in case of error on my side.
Regards
Martin
PS.: Besides, the GNOME folks work on quite similar technology (GNOME shell) for GNOME 3.0, so it is not soley a case of “what will those crazy KDE whackos do next?” 🙂
EDIT: Some small corrections and rearrangements
Edited 2009-01-24 19:42 UTC
The problem is not so much with the ideals of Plasma, but the fact when it locks up the entire desktop is completely useless for KDE 4 apps. Of course I can kill X11 and come back with a broken kde4, or I can hunt down certain commands to run in commandline and restart X11 [startx] manually. Truly not what one calls a User Experience to be praised.
Plasma didn’t want to be akin to Apple’s Dashboard. It wanted to make Dashboard then entire top layer of the UI experience.
It’s an impediment when it craps out that kills that user experience.
or… you can just fire up plasma again , which it should do automaticly.
of course the idea is for it not to crap up.
what if X craps out ? what if the kernel craps out ? what if the electricity craps out ? what if planet earth craps out ? its all useless for kde4 apps.
and even if plasma craps out , that in no way closes any application, only the desktop containers are affected. which you can just fire a terminal and write : plasma &.
either way , using kde 4.2 rc1 and plasma as yet to crap out even once.
The GP asked about the motivation behind plasma (or at least I had the impression that this was what interested him, anyway), so I naturally concentrated on the “ideals” of this layer. Yes, the “experience” for the user left a lot to be desired in 4.0 (calling it unusable was absolutly appropiate for a lot of different setups) and was still flaky for a lot of hardware/software combinations during the 4.1.x series, if user feedback is something to go by.
During the course of the last year, a lot of bugs in the infrastructure were exposed (and slowly fixed) so that the FOSS graphics stack is today in a lot better shape compared to one year earlier. Personaly, I would have prefered if better communications between up- and downstream would have taken place prior to the release of KDE 4.0 (e.g. KDE4 devs inquiring the state of the X11/Xrender drivers prior to making them a prerequiste for a good desktop experience, distrobutions like Fedora keeping more than one ear on the ground to determine the state of software like KDE4 or pulseaudio more closely before they start to throw it at their users, distros trying to work closer with upstream to coordinate backporting of features better in order to avoid duplication of work or shipping applications missing crucial functionality – the ark/Mandrive problem comes to mind, important infrastructure projects like Xorg being more outreaching to their recipients, trying to forsee the needs of their “customers” before a lack of QA bits them in their rears, etc. ). And if something positive can be said of this whole last year, than (at least) the awareness of this lack of working communication channes has increased extremly over the last year.
KDE4 is currently installed on about 8 computers that I have to administer, which covers a wide range of hardware specifications, ranging from nearly seven year old desktop computers with pre Pentium4 era processors to a PhenumX4 9550 with 8GByte of RAM. It does not work flawlessy on each of them but it works very acceptable on most of them by now, and judging from my experiences so far KDE 4.2 will improve the situation a lot, so that I can start to migrate the non-technical users that rely on me being their Linux-guy from KDE3 to KDE4, if they want to.
It would be great if FOSS users would not have to fill out bug reports and answer tedious, technical questiosn about their computers in order to get a working desktop within a year. One year ago, I would have said that it is the job of the distrobution to kown their target audience well enought to shield those parts of the communicty from this tasks that are neither interested nor willing to take part in this process. It is, after all, one big motivation behind the whole “distribution” layer we have in the ecosystem, but this seems to have been a misconception, or users may have been using the *wrong* distrobution for their needs, I don’t know.
If I may be so bold to hijack this reply for a little infomercial:
If you experience crashes of KDE4 (and especially plasma), please consider filing bug reports with enough information for the devs to bisect the bug quickly. I stumbled upon a lot of issues that only occured with certain combinations of Xorg, video driver (esp. nvidia), xorg.conf setttings and KDE4. Sometimes, it takes a certain usage pattern to trigger a “blatantly obvious bug”, but if no one reports those (and if they are limited to a certain segment of the feature test matrix) then it may be hard to detect them on behalf of the devs. Instructions for helpful and complete crash reports can for example be obtained via
http://techbase.kde.org/Development/Tutorials/Debugging/How_to_crea…
a discussion of other (from the pov of the KDE4 devs) helpful practices can be found at
http://aseigo.blogspot.com/2009/01/bugskdeorg.html
Thanks for the attention 🙂
To move the open source desktop on beyond what we have today, continue to compete with what Windows and Mac will be introducing from a functionality and presentation point of view and ensure that we’re not left behind with the equivalent of CDE in the Unix world for the sake of politics. Again. I can clearly see some seat squirming going on with this over the next couple of years.
How can you rewrite something when the functionality of Plasma didn’t exist before?
We didn’t have a resolution independent desktop before. We didn’t have a scripting framework to enable more applications and applets to be written and deployed easily. We didn’t have any real graphical capabilities or SVG. If we didn’t want to be left with CDE in the open source desktop world then Plasma had to happen at some point.
1) It is unfortunate people still think plasma is the desktop.
It isn’t.
I would describe plasma as a tool with which one can create a desktop ui.
2) the old kdesktop+panel was a mess and unmaintainable. so here goes the old saying “You can polish a turd, but it still is a turd”.
I think there is some confusion here between the Gnome the desktop, GTK+, and other foundation libraries being developed. Lumping all of them in together is sort of like saying that Qt, Plasma, Solid, etc. are “KDE” and are all doing the same thing at the same time.
As to the desktop, there is a very concerted effort going on right now to construct a *new* desktop shell for Gnome 3. (See http://live.gnome.org/GnomeShell, for example.) This shell is intended to act in a very different manner and will be using some new technologies that Gnome developers have been developing for two or more years (Clutter, gnome-instrospection, gjs, for example).
As for GTK+, there is a somewhat contraversial plan among the Gnome developers to break API compatibility of GTK+ to fix some creaky foundations in a way that will make it much easier to extend GTK+ in the future but without adding many new features immediately. On the one hand, this should make porting GTK+ 2 applications to GTK+ 3 applications much easier. On the other hand, it might be seen as unnecessarily antagonizing developers to break API for not much short-term gain. Some Gnome developers are against this.
Developers affiliated with Gnome have been developing some new, exciting libraries in parallel to the official Gnome development for the past few years: Clutter, Tidy, gnome-introspection, gjs, seed, Vala, Telepathy, Empathy, GStreamer, webkit-gtk, dbus, pulseaudio, libsydney, Tracker, devicekit, etc. (Note: I am not saying these are *Gnome* technologies per se, but technologies on which many Gnome developers are working and which they want to incorporate into Gnome.) The developers have been slowly incorporating these libraries into Gnome. The new shell should be using them even more heavily. So, you might say that the Gnome developers have been pushing their next-generation infrastructure in parallel to the stable core library releases for the past few years. (Perhaps this is not unlike the KDE 3.5.x releases while KDE4 foundation libraries were being developed.)
I don’t think this comment is true at all. (Please see above.) There was some stagnation on the desktop shell because a) there was a big push to maintain stability, b) no one agreed to what the next big shift would be, and c) no one stepped forward to “just do it” and code the next generation shell. That has all changed in the last year.
I agree, however, with the main point of your post: anytime people make a big change in the shell or in the libraries, it will inevitably take some time to iron out the wrinkles and provide many of the desired features. This was true with KDE2, Gnome 2, and now KDE4. I’m sure the same will be true with Gnome 3.
As for Linus’ switch, I don’t think many Gnome developers put much weight into it. After all, Linus flamed the Gnome developers on their own mailing lists two years ago or so, saying everyone should just use KDE. The Gnome developers took his comments with a grain of salt then, and they are doing so now.
Cheers!
What do you mean “Linus didn’t heed the warnings that KDE 4 was for developers only”? Linus *is* a developer! (admittedly, not of KDE 4 🙂 )
As I recall, this was the entire foundation of the KDE4 folks’ defense of the brain-dead way they released. They prattled and prattled on saying that they could call it 4.0 Gold and the distros would protect the users from the KDE project’s deceptive advertising.
Well, now that all is said and done we can see how that worked out. In this case, it’s probably for the best. I don’t mind counting Linus Torvalds, “Father of Linux”, as a Gnome convert and KDE4 refugee. 😉
Edited 2009-01-25 23:32 UTC
KDE4 also desperately needs a full theme (for plasma, Qt widgets and window borders [+ icons]) that doesn’t suck ass. I hate the default theme and to make it worse, there aren’t any to replace it with. There are only two alternatives, Bespin and Sculpture, and the only theme I recon as usable is the latter one. And even it has its (dis)usablity problems. Plasma’s getting better when the 4.2 is released.
The only thing that keeps me from switching back to GNOME is that KWin runs more smoothly than Compiz Fusion and its effects are more polished and usable than the Compiz implementations (Expo, Cover Switch, etc).
Edited 2009-01-24 12:07 UTC
i dont really have problems with the default theme.
bespin is a nice theme, it even has a menubar like kde3 or MacOSX. to me the problem with bespin is that it has so many options to configure ( which in fact is good, because you can always choose not to configure it or just import a theme (like the ones over at kde-look) ).
to me the problem with bespin is that i feel it has alot of hacks on its code.
if you are still using kde4 , then just file bug reports for the bad usuability.
also , i guess maybe they were expecting more community support , with more plasmoids and themes.
I agree that some major work needs to be done on the default theme. The current one looks old-fashioned in that everything is gray, and widgets, bars etc tend to take up way more space than they did in KDE 3.x. What ever happened to the promise of a pretty, shiny new OS? It seems like they took the complaints of 3.x being “too bright” too much to heart, to the point where the 4.x theme is actually stodgy, despite all the high-powered tech underlying it.
My actual theory for this and all of KDE 4.x’s shortcomings up to this point is simply that all the attention has been focused on developing new technologies and comparatively little on end-user experience. The whole “no icons on the desktop” fiasco is a key example of this situation of programmers’ priorities taking precedent over user priorities. This has been rather disappointing for me as I was hoping the end-user experience would be the area that would most benefit from the 4.x release over the 3.x.
But, you say, the technologies are supposed to enable a better end-user experience. True, but in reality they are just making the programmers’ lives easier while the status quo is being maintained in the GUI department. To create a good user experience, dedicated interaction designers are needed in addition to programmers, people who don’t just see things from the perspective of how it can be implemented in code.
Perhaps things will change with time. But the fact is that much of the KDE desktop is already “finished” and unlikely to change at this point. Developing an entire piece of software and then hoping to remodel it as an afterthought is an approach that nearly always leads to a sub-optimal user interface. Modularity alone will not save the day.
Edited 2009-01-24 12:35 UTC
Can you actually name one of these alleged “shortcomings”? One can have icons on the desktop with KDE 4 (in that as early as KDE 4.0 one could have a transparent folderview plasmoid). Now, in the imminent KDE 4.2, one can have the desktop with icons exactly the same as they used to be in KDE 3. Combine this type of flexibility with multiple desktops … and KDE is rapidly passing all other contemporary desktops.
One right off the top — at least through Mandriva’s 4.1.3, mouse button actions in the root menu (what you configure in KDE3 with KControl > Desktop > Behavior > Mouse Button Actions) are not user-configurable, at least as far as I’ve been able to find out. And furthermore (after queries in a number of fora) I can’t find out whether this will ever be fixed.
For what it’s worth, that has been an essential component of the window managers I’ve used, ever since OpenWindows on a SPARC2 ages ago, and it is a drop-dead requirement for the sort of work I do.
Fair enough. Most critics of KDE4 don’t actually have any beef, they are all bluster. You are one to have a genuine issue, so I understand that it isn’t for you without this feature. Likewise, GNOME’s shortcomings in respect of the clipboard mean that GNOME isn’t for me. This is all fair enough.
I guess I just have a problem with posters who seemingly don’t use KDE4 and haven’t tried it recently dismissing it loudly at every opportunity. There is a great deal of faked moaning and groaning going on about KDE4, almost as if there is a concerted astroturf smear campaign going on.
That is close to delusional. There have been many valid criticisms to KDE 4.X. The fact that you choose to not listen to them and mark it up as a “grassroots smear campaign” is ridiculous.
– Lack of mouse button options. Right click on the desktop in specific for me.
– poor/broken support for multiple monitors (this is with an intel based laptop so not an issue with NVidia or ATI drivers)
– hiding of panels (partially fixed in 4.2)
– ability to bind a hidden panel to a corner rather than an edge.
– setting a delay for hiding of a panel.
– keyboard shortcuts (completely broken in 4.0 and 4.1 and spotty in 4.2 RC1)
– multikey keyboard shortcuts (completely absent)
And these are just off the top of my head. I deliberately left the cashew off the list even though I despise it. It was a design choice.
KDE 4.1 may work for you but to dismiss criticism of KDE 4.X as some sort of smear campaign is silly. When people do not like something they comment. Especially when it is a replacement for something they liked a great deal more (KDE 3.X).
KDE 4 may be a great direction but it still needs more work before it will satisfy many of the fans of KDE 3.
I think it will never satisfy some of the fans of kde3. Some people are set in their ways and will not accept a new/different way of doing things. That is too bad.
The kde devs have stated many times that kde4 is not kde3 and that they are not remaking kde3 with Qt4. They are making a more easy and generally more flexible (the balance of which is difficult) desktop. As a result some features will be different or replaced with others.
Granted, many of your issues are still valid in this case, a lot of issues others have are simply asking to turn backk time.
Continue to tell yourself that. Anyone who has used Linux in a Windows dominated world for a prolonged period of time tends to be fairly flexible.
Most of the complaints KDE 3 users have about KDE 4 is the lack of ability to customize the desktop to their comfort. For instance all of the examples I gave involve customization of the desktop. If I were a user who took a desktop how it is given to me I would be much more likely to acept and use KDE 4.
I use ultraportable laptops and customize my desktop to optimize for the limited space. Then when I can connect to an external monitor I need to be able to do so without a fuss. KDE4 fails at even these two goals. When you get into more specialized elements (I use a tablet PC) KDE 4 fails even worse. Rotating using xrandr crashes KDE4 but works in fluxbox, gnome, KDE 3, and xfce.
So you can say some people are too stuck in their ways to appreciate KDE 4 but I say generally anyone who uses Linux on a day to day basis is not the type of person afraid to try new things. KDE 4 has failings, why not simply accept that not everyone can work with these limitations. The fact that KDE 4.1 and 4.2 were major pushes to add the missing features implies that they are acknowledged as important.
What I said was that your gripes (minus one) were quite justified, but that many other people will be disapointed because the feature they wanted is not coming back in the same form as it existed in kde 3.
An example of this is Kcontrol. Personally, I prefer the current icon system but I have no problems with the old way (it was a bit too cluttered though). However, some people grouched that the new way sucked and that they wanted kcontrol back. This is the kind of thing, that I was referring.
I already the mentioned the theme’s unfinished feel and the icons-on-the-desktop issue (which is apparently achievable in 4.2 “by using a Folder View as the desktop background”–is this the default or not?) Other than those examples, and just a general sense of glitchiness, I guess my main disappointment was how little has changed user-interface-wise since v3. The icon-based preferences, for instance, are a nice step forward, but the preference dialogs themselves are much the same as before, and just as confusingly arranged. I was kind of hoping for more of a rethink of the UI that would match the extent of the technology overhaul. Admittedly Dolphin, the new dock and Plasmoids are nice and new. However, it seems most of the rest of the UI has just been ported….
I really don’t mean to sound negative. The KDE project has my full support, and I know that many developers’ blood, sweat and tears are in this project, and I am deeply thankful to them. I guess I just held very (overly) high expectations for 4.x that were in certain regards not met….
No, it is not the default, and I am happy with that, I prefer having few folderview than having ~/Desktop crap. I can have an access to ~/, one to my production server /var/www using KIO-slaves and one shoing my current/in progress file using nepomuk indexing query. It is far better than icon based desktop space. With folderview theme like naked, it is also nice. And with the theme editor/mixer, I can use default KDE4.2 theme with folderview skin from naked, even better.
I’m confused. Can’t you have folderview on top of the “Desktop crap”?
If no, that strikes me as rather poor usability. If yes, then why isn’t the Desktop turned on by default, if it wouldn’t make a difference to you anyway?
Edited 2009-01-24 17:42 UTC
yes you can have a folderview plasmoid (or any other plasmoid) inside a folderview containment.
Desktop might not be turned on (will depend on your distro), because ~/Desktop sucks royally.
Kde default settings are quite sensible, but then it is up to the distros to customize it based on their userbase.
If the userbase is conservative on average it will use conservative settings and vice versa.
Edited 2009-01-24 18:07 UTC
I’m sorry, but I don’t understand what you are talking about with your repetition of the phrase “/Desktop crap”. What makes it crap? Please clarify. You are aware you don’t have to put anything in there, right? Why should the user not be allowed to place icons on the desktop by default if they want to? It seems like you are just trying to justify the rather dubious decision of the KDE devs…
Because a traditional desktop in any form has made no attempt whatsoever to try and help you organise your desktop – the very reason why people want to use computers in the first place. Get it to help you do the dirty work. I mean, on a real desktop you might actually have a desk tidy. There’s no such thing on a computer desktop.
They’re not the only desktop developers to be questioning current, traditional desktop layout and organisation and trying to find something ‘else’:
http://www.osnews.com/story/20328/Rethinking_the_Taskbar
http://arstechnica.com/news.ars/post/20081028-first-look-at-windows…
I agree, there’s something wrong with needing a ~/Desktop directory. You only get to realise that when you think about it. 🙂
That’s a good change in KDE 4.
If you’re really going to use the metaphor of a real desk, then a /Desktop directory is a necessity. After all, there has to be some “lowest level” where things are not “put away” into folders.
Even if you don’t agree, I don’t see how letting you put things on the lowest level by default is hurting anyone. You don’t like it, don’t use it, it’s that simple. Meanwhile traditionalists and average users are not turned off because they don’t grock this whole “folderview plasmoid” thingy.
Creating a new folderview plasmoid may seem easy, but it’s still not as easy as a right-click directly on the desktop and then clicking “new shortcut”. The question is, is it really easy enough for the average person to use, or will we end up with the majority of people inexplicably having no icons at all on their desktops, period, because they can’t figure out how to use the new system?
All of the KDE 4.0 and KDE 4.1 distributions that I have seen have included a folderview of the ~/Desktop folder by default. Within the area defined by the folderview, one can indeed right-click directly on the desktop and then click “new shortcut”.
In KDE 4.2, one can now have a folderview area occupy the entire screen area. One can also now have icons in the folderview placed anywhere one wants them.
I daresay most KDE distributions will from now on ship with a full-screen folderview of the ~/Desktop folder set up as desktop #1. This will be entirely indistinguishable from KDE 3 for most users.
Wouldn’t (shouldn’t) that be ~/ then? Why an arbitrary folder that is separate from your “lowest level” home directory?
For the simple reason that the user should be allowed to have it empty if they so choose.
what should they be allowed to have empty? /Desktop?
then how is folderview any worse? You can choose any folder to be your /Desktop and not have it arbitrarily chosen for you.
You can even choose not to have any icons at all thereby preventing yourself of bad habits.
Um…. Dude, if you care that much about a folder named “Desktop” being in your home directory then by all means choose a different folder….
I really don’t get what it is that you don’t get… It’s a very simple concept that has always existed…. If you want to put icons on your desktop, you can… If you don’t you don’t have to…
Well, it is just that now I can put them where I want and am not forced to use /desktop. I think this is better, you can use it if you want but you have the flexibilty of even having multiple folders in a tidy fashion by sticking it directly into documents and whatnot, intead of having that 1-2 jump desktop->documents. It’s nothing revolutionary, just better
I agree totally that it’s a better system!
My point was just that I think having a /Desktop folder is a logical default for the majority of (non-tweak-minded) users. Those who know better can and will change it anyway .
with a /desktop it became a dumping ground full of crap. With folderview I have several different folders open and I put things exactly where they belong…instead of desktop-> ~/folder
Sure I think /Desktop is crap; you might not, but it still is a bad habit.
I believe I’m running KDE 4.1.3 on my test environment at the moment, although don’t hold me to that. It actually takes work to find the “About KDE” these days, and I don’t normally run KDE 4.x.
Here’s the thing. Plasma is a competing UI paradigm with QT4 / KWin. I have a folder view, which is a plasmoid, but when I click on a folder inside, I get dolphin. If I want to resize the window, I can’t grab a corner, I have to grab the resize tool, and it doesn’t work at all like a regular KWin window, even though it kinda looks like one. Multiple UI paradigms on the same desktop are a Really Bad Idea(tm).
The plasma taskbar, which replaces the flaky 3.x panel UI, is also flaky, and has less functionality. Moving objects is unpredictable. At the moment, the Windows 7 taskbar (which is now very similar aside from virtual desktop support) is much smoother and user-friendly by comparison.
Someone commented that Plasma isn’t the desktop– the problem is, it’s what people see. It’s what they interact with. Novices don’t know the difference between Plasma and KWin, they just think “KDE 4 is a pain in the @#$”.
Multi-monitor support still doesn’t (as far as I can tell, I need to test it again) support the idea of a single desktop spanning two monitors, even though that’s what Xrandr is reporting.
Finally, tell me this: Can KDE 4.2 allow me to have file objects on the desktop, without a window, AND have functional plasmoids on the same desktop? Can I do this without editing XML? It appears currently, I can have a folder view on my desktop (see previous complaints), or I can have a folder view *as* my desktop… but if I do that, then I can’t have other plasmoids on the desktop.
Don’t get me wrong– Aaron has done fantastic work, but you don’t develop a new UI paradigm on top of what people have been led to believe is a production environment. You develop it in isolation, do usability testing, THEN you release it. Releasing it piecemeal and incomplete results in annoyed users, and gives KDE4 a really poor, and undeserved, reputation.
The KDE4/QT4 library rewrites, the porting of KDE3 apps to the new libraries, this has all been fantastic, and I would prefer to run in a pure KDE4 environment. But plasma severely limits my options, and forces me to do things not the way I want, but the way IT wants me to, and that’s not the KDE paradigm I’m used to.
Yes, it is. This is because the Plasma desktop is intended for a different role.
You would use this typically for setting up different “contexts” for different things you are working on. One set of files or locations and applications are important for one context (say, doing your taxes), and another set for another context (say e-mailing your friends), and yet another set would apply to another context (say working at home on the such-and-such project).
OK, Plasma now lets you set up folderviews, application shortcuts, plasmoids etc for a given context, to make it optimal for you to work in that context. You can have as many different contexts as you like, and you can save the current desktop settings as a named context.
The idea is … say you want to work on your taxes … you can load a previously-saved “My Taxes” context, and suddenly your whole desktop is set up for you to work conveniently in that context. Load the “My Taxes” context, and shortcuts to tax-related applications and all your current tax files suddenly appear on your desktop.
Then you get a notification of incoming email … so you load a new “My email” context (perhaps on desktop #2), you are set up neatly for emailing. Done that … either switch back to desktop #1, or re-load your “My Taxes” context.
I believe you can have a context apply to just the one desktop, so you could have one configuration on one desktop for one context, and another desktop for another context. You can organize it however you like, and switch easily between different setups, as you like.
It is not meant for you to laboriously change sizes of folderviews, and places where a folderview points to. Plasma is not meant to be a window manager type of paradigm, it is meant to be a “quickly arrange my desktop for doing this or that type of task” type paradigm. A new paradigm that no desktop system yet has been powerful enough to achieve. Set it up however you like it once (optimal for a particular task) then save that setup.
No other current desktop paradigm has anywhere near this power, flexibility and ease-of-use. KDE4 is revolutionary, rather than evolutionary.
… or you can use it in the same old way as you use KDE3, or GNOME, or Windows.
Your choice.
Edited 2009-01-25 10:34 UTC
OS/2 had this, in the concept of “Workplace folders”.
But you’re telling me that in KDE 4.2, all this context jazz is fully implemented? It wasn’t two weeks ago when I looked at the nightly snapshot.
Plasmoids are not meant to replace full applications. They are only there to increase uses and flexibility of the desktop.
Previously the desktop was only a dumping ground to peoples crap and shortcuts. The folderview plasmoid alows one to skip a step and put things right where they belong. With 4.2 you can have a folderview containment (which is the desktop in this case, another containment is the panel). And yes you can have other plasmoids in the folderview containment (even a folderview plasmoids, if you so wish).
Multiple screens have been heavily worked on in 4.2 so things that didn’t work before will work now.
Yes, and if you do the same thing in KDE3, you get Konqueror. In Gnome you get Nautilus. In XFCE you get Thunar I believe and In Windows you get Windows Explorer. Your point is what?
Edited 2009-01-25 15:58 UTC
Yes, but the icons on the desktop behave the same as the filemanager UI in every one of those environments… except KDE 4.x.
Folder view is inconsistent with Dolphin/Konqueror. Inconsistency is a bad idea in a desktop UI. That’s my point.
Thats because it’s not a filemanager and there for shouldn’t mimic it in every way. Folder View is meant to look like that and so are widgets, they are not meant as a replacement(as someone already said) so there for you dont make them consistant.
Nonsense. Graphic artists, professional usability experts and developers have been working in unison with kde4 and there are many visible improvements already in place.
Is it perfect? Probably not, but whose definition of perfect are we working with?
Is it suitable for a huge percentage, probably the majority of computer users? Definitely so, and ever more so with each iteration. You see, KDE and any piece of free software do not stop improving just because they have just released a new version. There are newer bugs to fix, features to implement and ideas to test.
KDE4 as it exists already rocks and once all the dust settles with all the major applications fully ported, we are almost there by the way, it will be the most impressive cross-platform set of applications in existence. Remember that soon mac and win users will be able to get a taste of the goodness as many applications have been ported or are being ported to both platforms.
KDE4.2 will allow you to have folders and icons on the desktop again.
I love the default theme (Oxygen). Just goes to show eh?
Yeah, I don’t get the complaints with Oxygen. I personally find it elegant, professional and very modern looking. It’s a work in progress, and little oversights like the aliased window corners and the contrast of the titlebar buttons has been addressed as of 4.2 and 4.3 alpha, respectively.
I’m glad they went with grey windows rather than following the theme on plasma. I think transparency on the desktop is like chrome on a car; less is more. The transparent titlebars and plasma widgets provide just enough “chrome” for the system. Giving the windows a Vista-like dose of transparency would have made the system look gaudy, and flat out terrible when compositing is unavailable. Much like Vista. IMO of course.
I like the default theme (Oxygen) better than any other theme KDE has ever had. Just goes to show, eh?
There’s qtcurve.
http://www.kde-look.org/content/show.php/QtCurve+(KDE4%2C+KDE3~…)?content=40492
I don’t get it… He was comfortable with KDE3. KDE4 he didn’t like.. Why not just stick to KDE3 – what he was using?
Why all the drama?
i dont understand that also.expecially coming from linus.
i mean , should we all have switched to windows when kernel 2.6.0 through 2.6.10 was out ?
I guess it is because KDE4 was used by LInux evangelists as a major boon to switch to linux. And it isn`t, so it might actually scare people away. To each is own, I will stick to my Windows 7, Mac OSX and the occassional Linux Mint with Gnome on my laptop
It is now.
I can’t say I recall that at all. I can only remember a very small handfull of people saying it was anything except broken crap at the time of release, actually.
What *drama*? He’s being very upfront about it, and it seems he hasn’t made any fuss about it.
Furthermore, he’s showing us he’s willing to try something else instead of bitching about his problem with KDE4 (for now).
The explanation for it are further up in the article, he uses Fedora. And Fedora seriously messed this up, as they jumped to KDE4 and removed KDE3 as an option. Removing that choice for their users.
In addition to not get the message from upstream, the Fedora team seriously miss-calculated their user-base. They put all their users in the category as early adopters and calculated that it also was true for their desktop usage. The Fedora team got lots of, well deserved, flack for it.
OK, this explains it. Debian has probably got it more correct, as KDE4 is AFAIK only available in Debian experimental (unstable) and backports.
http://blog.creonfx.com/linux/install-kde-4-on-debian-unstable-sid
http://kde4.debian.net/
After KDE 4.2 is released, KDE4 might get promoted to Debian testing.
At least with Debian, the names pre-warn you of what you are getting.
I agree… same thing with Slackware – in testing. So if he’s using Fedora and he didn’t like it, then who’s got it right?
OpenSUSE…?
Arch…?
FreeBSD…?
Mandriva…?
I’ve never really cared much for gnome – I’ve tried it and didn’t set well. Although I’m using KDE4 with Archlinux, It took a couple of days to get used to, but I have no complaints. I agree with Pat putting KDE4 in testing for Slackware as well as the Debian team until 4.2 is ready.
From what I understand there will be some library changes and for me that’s quite logical and I will bare the virtue of patience til then. I think the KDE team has done quite well with the efforts of this release.
Give it time….
Well this is what Fedora do anyway, they dont lay around with old tech, they have new bleeding edge tech pretty much before anyone else.
Even though Fedora 10 has a KDE4.1 version(they backported some KDE4.2 features as well), the main version is defaulted to GNOME. I think people knew what KDE4.0 was going to be, they just did take noticed of the plan and missinformed themselves.
Probably because fedora mixed kde3 and 4 and broke it?
Stop blaming Fedora, blame the true responsable of the KDE 4.0-.1 mess, Aaron “I want all your attention” Seigo.
The main problem is that after you try GNOME and get used to its stability,smooth working and the “just works”, you cannot back to KDE.
wow …
unless aseigo now is the manager of kde on fedora, i dont see how he is to blame and not fedora, for that specific mess.
for me it is the other way around ! after using kde , macOSX , even windows, i really cant go back to gnome !
HECK even after using gnome 1.x , i cant go “back” to gnome 2.x.
but hey , everybody has different tastes
I talked to a Fedora package maintainer. He said, that they got wrong information from the KDE project previously to the 4.0 release. They got told that 4.0 will be a real final release. However, when they got told that 4.0 wasn’t meant for endusers it was to late. It was short before the 4.0 release and Fedora had KDE 3 already replaced (KDE 3 and KDE 4 cannot be both installed because of conflicts).
To switch back to KDE 3 with Fedora 10 wasn’t an option. That would triple the breaks for the users (f8 -> KDE3, f9 -> KDE4, f10 -> KDE3, f11 -> KDE4).
However, I think it would have been nice if Fedora would have expanded the support period of Fedora 8 by one month, so one can directly switch from a KDE 3.5 distribution to a KDE 4.2 distribution. :/
Currently I’m evaluation Fedora 10 + KDE 4.2 in VirtualBox and I’m frightened. It’s very instable, plasma keeps crashing (desktop, panel, etc.) and it seems not to be possible to remove a plasmoid from the desktop. Every time I relogin all the old plasmoids are back. Sometimes a big grey window covers almost the whole desktop and I can’t click anything on it any more. Only relogin helps for that. Also “expose” doesn’t work right. Sometimes I have to click twice or even more times on a window to switch to it.
I hope these bugs are because I use it in VirtualBox, but most of them I also experienced when I used the LiveCD. (The ones I didn’t experience where features that didn’t work at all on the LiveCD (e.g. expose) because of missing support of my graphics card etc.)
that is very very weird !!!! aseigo always blogged that kde 4 was not 4.0. always blogged that 4.0 was a release for developers to start building apps for kde4. dont know how fedora kde manager would think that kde4 4.0 was the real final release. it goes against everything every single developer or user blogged about.
about the crashing and etc, i have been using opensuse kde factory and unstable. even on unstable kde4.2 svn releases ( last year ) it was never like that.
i can say that i have 4 activities, each with own plasmoids ( and alot of them really ) , with vertical panels, with sometimes bespin with a menu bar in another panel, and it doesnt crash or forget any settings.
i dont want to be suspicious , but i am starting to think fedora has a hidden agenda for kde4.
just for testing sakes ( and i dont know about virtual box is to blame ), install opensuse and try factory kde repositories ( currently kde 4.1.96 ). you should find it extremely stable.
Sounds like a problem with your distro. I’ve used archlinux with kde4 without a single crash in months of leaving kde 4.1 running.
-Tom
Hm, archlinux. Maybe I should try that. However, the OpenSuSE LiveCD had mostly the same bugs as I reported about Fedora (grey window covering the desktop etc.). Anyway, I would never use SuSE, because frankly YaST is crap. Its freaking slow and does not expose options I usually have in the kde control center provided by guidance (not a Fedora extension to the control center!). So SuSE actually removes functionality from the control center to replace it with much much slower, crappier and uglyer tools. (WTF?) At least it was that way the last time I tried it (I think that was a KDE 4.1 LiveCD).
Fedora screwed up, big time, then. KDE3 and KDE4 can co-exist on the same disk quite nicely.
FreeBSD, (K)Ubuntu, and OpenSuSE all figured out how to do this from the beginning. KDE3 goes in /usr (Linux) or /usr/local (FreeBSD) and KDE4 goes in /usr/kde4 or /usr/local/kde4.
User configs go in .kde or .kde4.
Everything runs nicely. You can stick with a KDE3 desktop using KDE4 apps, or use a KDE4 desktop with KDE3 apps.
Sounds like more bugs in Fedora’s implementation of KDE4, than of bugs in KDE4 iteself. At least to me. KDE 4.2 is running nicely on FreeBSD. Haven’t updated my Kubuntu install yet.
That violates the FSH standard.
As I said in another post, I experienced almost all those bugs with an (earlier) OpenSuSE LiveCD, too.
However, I still like KDE much better than any other DE (GNOME, Windows, XFCE, OS X). Well, overall. And I still like Fedora very much, because they have better KDE support and longer maintenance periods then KUbuntu. KUbuntu actually *deletes* the old rpms after 6 months, so you end up with a broken installation, in case you haven’t downloaded all the latest rpms when you had the chance. In such a case you can’t e.g. download the foo-devel pkg to the already installed foo pkg. You can’t install it from the CD either, because the foo-devel on the CD is older than the foo pkg installed on your system.
No it doesn’t.
Would you mind stopping with your FUD?
First Kubuntu does use deb, not rpm.
And second the support goes longer than 6 months it is 1 and a half years and then you still can download the packages. The only thing that changes is that you do not get any updates.
That’s absolute crap as per usual when Fedora/Red Hat does anything KDE related. Any distributor with half a brain would have kept KDE 3.x in as the default (Suse managed it) and then monitored the desktop as it moved along as to when the best time would be to switch the default version. Regardless of what they were told, you don’t switch a major desktop version like than and I’ll bet that they will never do it where Gnome 2 -> Gnome 3 is concerned. If we ever get to see Gnome 3.
Just like with Ubuntu and the LTS version, KDE 4.0 was used as a rod for KDE’s back even when KDE developers had reiterated time and again that 3.x would continued to be maintained and point releases would still be made.
Rubbish. Suse managed it.
There’s a difference between, can KDE 3 and KDE 4 be installed on the same system at the same time, or, can a distribution offer both KDE 3 and 4 in one release? Obviously, the latter is true, but is the former too? (Idunno)
The reason that Fedora dropped KDE3 is that they often just want to move on to the latest and greatest, even if that’s not always the most glitch-free option. They’re probably reasoning, if you want to stick to KDE3 for the moment, why upgrade your existing Fedora system.
That’s the choice they made; as a Gnome user, I think KDE 4 on Fedora 10 is very promising, feels snappy and solid, and it’s already quite usable unless you can’t stand a few glitches here and there. It would be completely un-Fedora to keep supporting KDE 3 even in Fedora 10.
Yes, it is possible. Suse managed it and had KDE 3 and KDE 4 apps running fine side-by-side. It defaulted to KDE 3.x and offered KDE 4 as an option. That will most likely change with 4.2.
indeed it will, the next version of suse will only ship kde 4.2 maybe 4.3. That is still a long way off (~6/7 months)
You are a little behind the times:-) it has already changed. OpenSuse 11.1 defaults to 4.1.something, and KDE3 is pushed down to the ‘other’ category when selecting desktop.
But as you said, it’s defiantly possible to have them side by side, Mandriva also handled it very well. I think the only distributions making an issue out of it and screwing their user-base was Fedora and Kubuntu. All others, if they provide KDE4, had it installable side by side with KDE3.
They don’t conflict. One can have both KDE 4 and KDE 3 installed on the same system at the same time.
What you will get, however, is duplicates. There might be a KDE4 version of konsole installed, and also a KDE3 version. In this case, KDE4 must be running the base desktop … because KDE4 can run KDE3 applications (as well as GTK applications), but KDE3 is not able to run KDE4 applications. If you run KDE3 as the base desktop, none of the KDE4 applications will run.
That’s incorrect, actually. openSuSE allows this with no problem. I currently run KDE 3.5.x, but use KDE4 versions of Kopete and Kmail. I do occasionally wind up with two wallets open, but that’s no big deal.
What about KDE4 applications that expect the Phonon/Solid etc underlying framework?
I can see how Phonon can have a pseudo aRts emulation, for example, in order to be able to support running KDE3 audio applications written to utilise the aRts API rather than ALSA, but I can’t see how KDE3 and aRts can possibly have a pseudo Phonon API to host KDE 4 applications.
Maybe SuSe lucked out with Kopete and Kmail?
Those frameworks allow applications to have access to their functionality even when the run on something as “foreign” as Windows, so I think it is quite natural to assume they can still provide the functionality when the application has just been started on a different workspace.
One thing to remember is that neither Phonon nor Solid are separate from the application.
If one of their backends is implemented as a client to some service, e.g. the HAL backend for Solid being a HAL client, it just needs to have this runtime requirement fullfilled.
In the case of HAL this would be true independent of whether a user logs into a KDE4, KDE3, GNOME or XFCE session.
As for Phonon and aRts this is no problem either.
Independent of the backend used by Phonon, e.g. GStreamer, it can either be configured to output to aRts or use the same output path as aRts, e.g. stream mixing at ALSA level.
The probably best way to estimate how a KDE4 application would run on a different workspace is to pick a backend for those frameworks, e.g. HAL for Solid and GStreamer for Phonon and check how a non-KDE4 application would be impacted when it was directly using the chosen components.
For example a KDE4 application with the GStreamer Phonon backend would look to the underlying sound setup just like any other GStreamer application.
A KDE4 application with the HAL Solid backend loaded would have access to the same information as any application directly using HAL’s D-Bus API.
Then obviously you have to install that framework and then the application will work just fine on a KDE3.5 desktop.
> (KDE 3 and KDE 4 cannot be both installed because of conflicts)
Seems other distributors like Suse didn’t agreed and offered both versions.
Hmmmmmm. Well Suse managed to do this, as well as any distributor with half a brain. They installed KDE 4.x side-by-side with KDE 3.x, with 3 being the default, and it’s only now that 4.2 is planned to be the default KDE desktop.
One can only guess as to why the incompetents at Fedora and Red Hat (as well as Ubuntu) chose to turn 4.0, and 4.1, into a far bigger horror show for the benefit of their own users ;-).
Well no. There wasn’t any breakage because of that since Fedora shipped almost a completely KDE 4 desktop and just a few apps like k3b which has no KDE 4 equivalents like every other distribution does. Other than a few minor backports, Fedora made no changes to the upstream KDE 4 and features or regressions are all from upstream
Refer
http://magazine.redhat.com/2008/02/18/kde-sig-talks-about-kde-4/
http://www.redhatmagazine.com/2008/05/14/fedora-9-and-the-road-to-k…
http://fedoraproject.org/wiki/Releases/FeatureKDE4
http://rdieter.livejournal.com/7424.html
Edited 2009-01-24 14:42 UTC
How to be able to stick to KDE3 when your distributor (Fedora) no longer support a release of their distribution that includes KDE3? Just ignoring security issues is not an option. Support for Fedora 8 ended at the beginning of January.
I was wondering that myself. I would guess he probably just doesn’t want to get stuck with an old release of a desktop environment for the next 6-8 months, which is understandable.
I’ve heard mixed opinions (mostly bad) about KDE4.x so far, tried 4.0.x and 4.1.x and hated them both. It would be nice if KDE would redeem themselves, but so far, it seems like it’s going to be a long way. I don’t know of a way to try 4.2 yet either (live cd would be nice), so I guess I’ll just wait. I’m not going to get my hopes up, because people said that 4.1.x was “good” too.
I’m kind of nervous that Patrick (Slackware) and Tomas (Slax) are both considering (and have been) “upgrading” to KDE4 already in a not-too-distant release. Oh well… there possibly goes my ability to run them nice and fast on 256 megs of memory (without using Xfce or another desktop). Unless some serious performance optimizations have been done.
Depending on your level of adventuresomeness, you could
give the Chakra project (a rather new live-cd distro based on Arch Linux and their KDEMod4 sub-project) a spin:
http://chakra-project.org/
They tend to track the KDE development very closely, and due to the similarities between Arch and Slackware (as little as possible distro specific tweaks, etc.), it is probably a good indicator of what to expect from a Slackware KDE4 desktop. But be aware that Chakra is quite new (although based on proven technology) and my experience with it is rather limited (although my short flyby in the last week was a very positive experience).
Additionally, the nice folks from KDEMod provide KDE3 packages for those who did not want (or were not able) to follow Arch in its migration to KDE4 around 4.1 last summer :
http://kdemod.ath.cx/download-kdemod3.html
Personally, I have to say that I’m quite sure that my usage patterns differ from your ways of doing things since
a.) I don’t hate KDE4 so far (since 4.1, I’m quite happy, to tell the truth) and
b.) that 256MByte RAM were a bit too low for me to work with KDE 3.5.x in the past (On machines with as little memory as that, I use to run XFCE or Fluxbox).
I maintain a KDE4 installation at work on an old AthlonXP 2200+ with 512 MByte RAM and an old Rage128 32MByte video card (no desktop effects) and the performance is very acceptable ( it suffices for my development and testing work and I get no complaints from my colleagues 😉 )
Measuring memory on linux is iffy, and I have no hard, solid data available, but from my experience so far, KDE4 “feels” about as memory intensive as KDE3, YMMV, of course. I’m typing this from a 2GHz Celeron with a 5200 FX NVidia card (173.x binary driver) and 1GByte RAM and KDE4 (including desktop effects) runs very well on this rig (getting rid of the approximately 0.5Kg of dust that the CPU fan accumulated in the course of the last six years recently helped a lot, performance wise 🙂 ).
At the risk of stating the obvious, I found these little tricks to improve the performance of KDE4 on
low-spec computers quite a bit:
– If in doubt, disable strigi indexing and/or Nepomuk. Especially Strigi’s performance seems to have improved a lot over the last months (apx. comparable with tracker indexer now, from my POV), but it may be a cheap way to reduce disk accesses
(System Settings -> Advanced -> Nepomuk)
– Less aggressive Anti – Aliasing of fonts
(System Settings -> Appearance -> Fonts -> Configure Anti – Aliasing). Reducing to “Slight” helped a lot on the beforementioned Athlon system.
– There is a new environment variable available for KDE 4.2 which disables ARGB rendering (which can cause a lot of trouble in the performance department esp. on older graphic cards) at the expense of eye-candy
http://aseigo.blogspot.com/2009/01/todays-tip-kdeskipargbvisuals.ht…
– Disabling unwanted runners in krunner ( Alt-F2 -> Preferences menu ) can improve the responsiveness of this little app quite a bit on slow systems
I hope this helps (at least a little)
Regards
Martin
EDIT s/RGBA/ARGB
Edited 2009-01-24 19:07 UTC
Crap, forget to mention that Chakra alpha 1 (currently available) defaults to 4.1.3 while KDE 4.2 should be available on a (hopefully-soon-to-be-available) alpha2 release of chakra.
To my defense: This piece of information was there but was able to sneak out of my comment due to budget cuts, err, my intents to shorten the comment a bit.
Sorry for that 🙂
Edited 2009-01-24 20:13 UTC
http://distrowatch.com/?newsid=05290
(Mandriva 2009.1)
http://distrowatch.com/?newsid=05286
(Kubuntu 9.04)
http://pkg-kde.alioth.debian.org/kde4livecd.html
(Debian KDE 4.2 live CD)
All of these are alpha pre-releases, but they will allow you to run KDE 4.2 (RC1).
KDE 4.2 itself is not due for release until Jan 27th, AFAIK.
KDE 4.1.x is good … in that it can run anything that KDE3 can. It is not however feature-complete KDE4 as yet.
KDE 4.2 is considerably closer to being feature complete, as well as being considerably faster.
The latest version of Nvidia drivers, which fix the Nvidia Xrender bug which severely affected KDE4 performance, should also now be available in these latest distributions, and this also will bring a chalk-and-cheese improvement in KDE4 performance for those systems formerly affected by that bug.
As long as your graphics hardware is working properly (and not, say, affected by the Nvidia Xrender bug), then KDE4 should actually be the fastest full-featured Linux desktop that can run on your system.
Edited 2009-01-25 06:55 UTC
Because it’s never good, especially for a developer, to stick with “outdated” software? Okay, as a kernel developer, Linus won’t run into major problems because he does not develop KDE software directly, but… it’s so easy today to get problems installing some software if you’re not exactly up to date with underlying libraries or desktop systems.
Allthough KDE 3 will be available for quite some time, development seems to target KDE 4, and that’s what is to be used in the future. Or do still people exist who are comfortable with KDE 1 and still use this?
Don’t get me wrong: I do belong to those users (NB: not admins, as an admin I have a different point of view) who install a system *once*, and then keep it using. If I wouldn’t have had a file system meltdown in July 2008, I’d still happily be using FreeBSD 5 and XFree86 (yes, instead of X.org)… 🙂
I remember when FreeBSD 5.x was released – there was much wailing and gnashing of teeth. Long time FreeBSD users maintained that 4.x was the best thing next to sliced bread and 5.x was epic fail.
Well fast forward a few months and spawned from that big change has been two of the best releases – 6.x and 7.x because of this major leap forward.
That same thing will happen once KDE 4.x reaches maturity – there will be much wailing and gnashing of teeth until things settle down. People will then turn around and talk about how great it is and how there are features they couldn’t live without.
IMHO when they released 4.0 they should have called it, “KDE 4.0 Developer Release 1.0” and kept upping the 1.0 onto the end until it was at user release quality to which one then can change it to ‘KDE 4.0’ and put in a blurb, “user release”.
They did say it was developer quality but at the same time it didn’t help when vendors were shipping KDE 4.0 with distributions and end users assuming that if it is shipping with a distribution it must therefore be user ready.
As for GNOME, there is a heck of alot that needs to be done when it comes to integration between the various applications and tapping into global services such as spellchecker and so forth.
There also needs to be a move to make these desktops more abstracted so that it is easier to port them from one operating system to another; even if that requires KDE and GNOME working together on a abstraction layer that sits below both desktops as to make supporting alternative platforms easier.
Edited 2009-01-24 12:21 UTC
What’s the time frame between FreeBSD5.x to FreeBSD 7.x?
Compare that with KDE4 to the present.
A few months, you say?
Depends on how you define a few months. The fact you’re getting anal about months/weeks/years etc. shows you missed the point of the post entirely. I was pointing to the fact that changes might not give immediate results, and in some cases feel as though there has been a move backwards BUT when one looks at it from the long term the changes were necessary.
That is the point I was trying to make, hence the comparison to FreeBSD. You missed the point entirely and speaks volumes to the fact you can see outside the box.
I also think that was an error to release version 4.0 as stable and that the biggest error was the use of KDE 4.0 as the default KDE version of some of the last versions of several linux distributions.
Even some of the KDE project developers say that the true “KDE 4” is the next 4.2 version. But KDE 4.x is a necessary and big improvement for KDE and for me it has much more chance to compete with MacOS X and Windows 7 desktops.
But wait! KDE 4.1 was the “true” KDE!
Two months from now, it’ll be “Wait for KDE 4.3, it’ll be awesome!”, and everyone will be running Gnome.
Actually KDE devs have been saying since KDE 4.0 that the first proper release that is designed to be used for average users is KDE 4.3.
Its uers like you who don’t know the meaning of the word “this is not a stable release, the next stable release will be x” has created half of the problems that KDE devs have to put up with
Wow, may I remind you that the article mentions a lot more than just Linus dropping KDE?
Considering what a shitstorm Linus started when he publicly blasted Gnome and hailed KDE, I’m not surprised it’s all that was taken away by many. It probably pisses both camps off, really, and those outside either camp are probably wondering why the *freaking creator of linux* couldn’t handle the upgrade or handle blacklisting the KDE updates he obviosly didn’t want. This is what most of us who feel the same way he does did. Even more probably wonder why the hell he uses fedora at all if he really wants total control over his own system.
Linus Torvalds said:
“It’s a huge job to do a distribution. The reason there are hundreds is it is easy to start your own, but if you want to be a leader and introduce new code, the testing and Q&A involved is enormous. It depends on having enough users that you get coverage and it is unreasonable to expect too many large distributions. Ubuntu grew surprisingly quickly and maybe that can happen again.”
Here I agree with Torvalds, although I don’t think another phenomenon like Ubuntu will happen soon. Mark Shuttleworth is an exceptionally wealthy man who has the resources to make things happen quickly. Still, he didn’t create Ubuntu from scratch. He duplicated Debian’s infrastructure and hired a bunch of top Debian developers to work for Ubuntu. Of course, Shuttleworth made also some other smart moves, like dropping support for less popular processor architectures, and he made a large number of Ubuntu’s packages depend on community support (and ultimately on Debian’s unstable branch).
I expect Linus Torvalds might still change his mind about KDE4. I’ve used the latest release candidate of KDE 4.2, and I don’t miss much of the functionality that was in KDE 3.x. Currently my biggest gripe with KDE4 is that editing bookmarks in Konqueror seems way too difficult. But after changing the plasma theme from the default, switching applications menu to the traditional layout, and changing window decorations to the old plastik theme, I find that using KDE 4.2 is a pretty similar experience as using KDE 3.x — only KDE4 looks much, much nicer.
Out of curiosity, which Plasma theme did you switch to?
Silicon.
Thank you.
My opinion exactly.
I used the classic configuration of KDE 4.x so that it looked like KDE 3.5 and now prefer it to KDE 3.5.
I will stick with KDE 4, now.
What is interesting is that KDE 4 doesn’t certify a lot of classic applications for KDE4 compliance, but they work anyway. Kubuntu Linux is a great system that way. I can still use my programs even after an update. Requires a tad of software bloat to do so (but I have lots of hard drive capacity and RAM).
Torvalds is using Gnome because it is slimmer for netbook usage (which he says he is trying out).
KDE4 is still in the Occam’s razor stage — slimming and becoming more efficient.
But it will eventually be the dominant desktop interface, for all OS’s.
Good article but why link to page 5 of 6? TFS also gave me the impression he would be discussion ext4 at all, instead of just ZFS and btrfs specifically.
Using GNOME is not one of the best experiences around, but after KDE devels chose to break the whole thing apart, it gaves us no other mainstream choices on the desktop.
I do find that GNOME issues are far more easy to fix to regain users’ happiness about their requests, than KDE4.0 becoming any sort of desktop killer any time soon. KDE4.0 a plain SHOT on right on the foot. It’s starting to hurt now, and you will halt your walk slowly.
I just don’t understand how the distro vendors could put a piece of software in there that was so obviously unfinished and unusable.
Does nobody at Fedora sit down and say ‘hey wait a sec, the software we’re shipping does not work very well, lets not do this.’?
I totally understand that KDE4 is a work in progress, but it’s clear linux distro vendors are just basically shovelling crapware onto a CD, instead of crafting a quality end-user product.
This ‘we take the crap from upstream,and deliver the crap to you’ attitude does not inspire confidence.
The problem here is fedora cares very little about kde. Hence crap decisions get made.
They are improving but at the kde3 time it wasn’t so nice. With kde4 they are slowly coming around and are getting interested again, and hopefully it will only get better in the future.
“To some degree Btrfs does what ZFS does. Some uni ran ZFS as a module in Linux so using it with Linux can be done. The biggest thing Sun did with ZFS is they were good with PR and marketing.”
From http://btrfs.wiki.kernel.org/index.php/Main_Page:
“Btrfs is under heavy development, and is not suitable for any uses other than benchmarking and review. The Btrfs disk format is not yet finalized, but it will only be changed if a critical bug is found and no workarounds are possible.”
Above is pure evidence that Linus does not know so much as he should about even btrfs. Trying to compare fully functional storage subsystem like ZFS (which integrates classic FS and VM and part of BIO layers on one subsystem) with something like only plot of the new FS (which keeps all in classic way of thinking about storage) is very naive.
btrfs it is nothing that only next example NIH syndrome (Not Invented Here). Head one of biggest FOSS project shows next time only pure incompetence on trying compare btrfs and ZFS.
“. The Btrfs disk format is not yet finalized”
but it kind of is now which probobly is whay it get integrated for beta testers.
wikipedia is not always fast to update better to got to the project website
Just because an engineer is proficient in their background with kernels, what makes you think they are remotely well-versed in filesystems?
Specialties are quite common, especially in Computer Science.
If you are proficient engineer proficient with kernel background and if you really saw ZFS in action more than few minutes you should be able distinguish between classic FS and something which is not.
I’m not talking about Linus professional background but only about what he know and what he doesn’t know in context his public speech.
Simple he talks about things he doesn’t know and this is completely unprofessional behaviour.
It is quite stupid of Linus to imply that ZFS has got much attention because of SUN has good marketing division.
ZFS stands on it’s own, without marketing. The proof of the pudding is in eating it. When people try ZFS for a while, they instantly change their minds. ZFS is not about marketing, it IS truly revolutionzing.
Huh, not about marketing?
In Linux you can use LVM + any other filesystem and have all things ZFS does. Maybe less convenient but still works. But btrfs will be simple better – just read specs and datasheets, btrfs is a really powerful thing – by design. They simply taken the best from a bunch of other good filesystems. And that’s why final result will be better than anything before. Sun is going to lose this battle in long terms. Because now bunch of filesystem gurus will improve btrfs and not zfs. Who loses from this scenario? Sun. The one and the only. Well, Sun has been kicked into teeth by openvz\virtuozzo – now Linux beats Solaris in this feature to the hell. Next will be filesystems and Sun boosted this process itself, ironically. They provided a great challenge. And Linux gurus did responded. With a new filesystem design which looking very interesting. In terms of possible speed, features and reliability.
Funny, as btrfs will be a Linux only thing, whereas ZFS is already running under Mac OS X, FreeBSD, NetBSD…
Where is the granular resource management like in Solaris Containers? With Crossbow Network virtualization, Sun will set the bar even higher. Linux is still a toy, when it comes to resource management.
That’s just nonsense…Btrfs will be enterprise ready earliest in 2012 including only basic features. Every feature they plan to implement in btrfs after that later is already available on ZFS.
Btrfs is doomed, as it is a Linux only thing, it is also too little too late. I would not be surprised if other non-GPL operating systems adopt ZFS as well…
Except in-filesystem snapshots, pooled storage, revertable snapshots, transferable and saveable snapshots, thin-provisioning, automatic rebuild using hot-spares, multiple-copy storage even on non-redundant hardware, checksums of all data and metadata, in-filesystem compression, the ability to swap out drives for larger ones and automatically have the space available right away, and a bunch of other features I can’t think of right now.
LVM is useful to a point, and is much better than using raw devices with partition tables. But it doesn’t compare to pooled storage systems in any way.
Maybe, will be, sometime in the not-so-near future, possibly. Really strong endorsements there. ZFS is here, has undergone nearly a decade of development, testing, and deployments, while btrfs hasn’t even released 1.0 yet. It hasn’t even been released as part of a vanilla Linux kernel yet.
And I guess you think ZFS is a poor, weak, broken thing — by design?
It is quite obvious that you dont know much about ZFS if you believe that LVM can do everything ZFS does. You sound like a Windows user saying “Windows can do everything Mac/Linux can do”. Yes, Windows can, but Windows can be quite cumbersome. You CAN cut down a tree with a knife, but it is quite a difference between a saw and a knife?
But, not only is ZFS more convenient, but it can also do things no other filesystem can do. For instance, you can snapshot the root partition and patch, and upon boot in GRUB, you can choose which snapshot to boot from. And if it doesnt work, you can rollback immediately. One guy had 10.000 snapshots, using 4GB of space. You can thus have many patch levels and boot from any one of these. ZFS is forbidden to write on a snapshot, so if you get a virus and destroy that snapshot, the virus will totally disappear from the disc.
Not like windows, where you patch and then try to rollback – and windows have written all over the place, the register, etc – making it impossible to totally undo the patch.
There is no fsck (fsck a big linux filesystem took over a week – it would be faster to use backup) in ZFS. You “scrub” while the discs are online instead.
But the greatest feature is probably the safety. ZFS detects and repairs silent corruption. That is the main point of using ZFS:
http://opensolaris.org/os/community/zfs/docs/zfs_last.pdf
“Measurements at CERN (using Linux)
â— Wrote a simple application to write/verify 1GB file
â— Write 1MB, sleep 1 second, etc. until 1GB has been written
â— Read 1MB, verify, sleep 1 second, etc.
â— Ran on 3000 rack servers with HW RAID card
â— After 3 weeks, found 152 instances of silent data corruption
â— Previously thought “everything was fineâ€
◠HW RAID only detected “noisy†data errors
â— Need end-to-end verification to catch silent data corruption”
http://www.miracleas.com/BAARF/RAID5_versus_RAID10.txt
“When a drive returns garbage, since RAID5 does not EVER check parity on read (RAID3 & RAID4 do BTW and both perform better for databases than RAID5 to boot) if you write a garbage sector back garbage parity will
be calculated and your RAID5 integrity is lost! Similarly if a drive fails and one of the remaining drives is flaky the replacement will be rebuilt with garbage also propagating the problem to two blocks instead of just one.”
[url]http://74.125.77.132/search?q=cacheA0-6Ufp8a8J:gridmon.dl.ac.uk/nfn…]
One failure for every 10**15 bits read.
RAID arrays may not handle block re-maps satisfactorily or take so long
that the system has given up.
RAID5 not perfect protection – finite chance of 2 disks failing!
SCSI interconnects corrupts data or give CRC errors
Filesystems (ext2) corrupt for no apparent cause (EXT2 errors)
Silent data corruption happens (checksums are vital).
Fixing data corruptions can take a huge amount of time (> 1 week). Data loss possible
And I dont agree that you say that a planned design as btrfs is better than ZFS. You know, the big difficulty is realising and implementing the design. Anyone can write fancy design papers. And when btrfs has gotten into production, ZFS will be improved much more, it gets more and more functionality all the time. btrfs seems surprisingly similar to ZFS, to me at least.
I promise that if you tried out ZFS, then you would see what all the fuzz is about. ZFS is not like LVM – it has unique features no other filesystem has. Seriously, I mean it. If you believe that ZFS is like LVM – then this is proof that SUNs marketing division has failed about ZFS. As we all know, SUNs marketing division is not that good.
This is a good read about future filesystems and the problems they have to solve. Read it, and then you will see at brtfs with new eyes:
http://queue.acm.org/detail.cfm?id=1317400
Edited 2009-01-27 13:21 UTC
You havent tried ZFS, thats for sure. ZFS does have unique features that no other file system have. If there are people believing ZFS is only marketing, then SUN marketing division has surely failed.
The main point of ZFS is that your data is safe. See here about Linux:
http://opensolaris.org/os/community/zfs/docs/zfs_last.pdf
“Measurements at CERN
â— Wrote a simple application to write/verify 1GB file
â— Write 1MB, sleep 1 second, etc. until 1GB has been written
â— Read 1MB, verify, sleep 1 second, etc.
â— Ran on 3000 rack servers with HW RAID card
â— After 3 weeks, found 152 instances of silent data corruption
â— Previously thought “everything was fineâ€
◠HW RAID only detected “noisy†data errors
â— Need end-to-end verification to catch silent data corruption”
Here is some good reading about future file systems:
http://queue.acm.org/detail.cfm?id=1317400
Then you will see what kind of problems brtfs will have to solve.
This is nice article series from a linux guy building a new home server:
http://breden.org.uk/2008/03/02/a-home-fileserver-using-zfs/
BTW, I dont agree with you comparing btrfs with ZFS. The difficult part is to implement a design. There exists lots of good ideas, but not many are implemented. And the implementation is the crucial part. I can write down a design: “The Best OS in the world, it will be able to run 256 bit program blindingly fast. etc etc” Now, someone has to implement it.
Besides ZFS evolves all the time. It doesnt stand still.
Seems to me that the level of debate in OSNews is reaching an all time low.
Now to the facts:
Fedora 9 was released with KDE 4.0.3 with no KDE 3.5.x options due to the following reasons:
A. To the best of my knowledge, the Fedora’s KDE SIG (special interest group) felt that it didn’t have the necessary resources to maintain two concurrent KDE releases. (3.5.x and 4.0)
B. Furthermore, given the support time-frame of F9, the KDE SIG assumed that it may not have the necessary resources to maintain KDE 3.5 (and especially QT 3.x) post its EOL. [1]
C. It was assumed that people who rather wait for a stable (and/or user-friendly) KDE release will keep F8 until KDE 4.1 is released and then switch. (Worked for me just fine, thank you)
Grated, it would have been nice if the Fedora KDE SIG could have extended the F8 EOL till March/09 (wasn’t their decision) and/or keep KDE 3.5 in F9 and F10 (See above), but given their resources and Fedora’s built in limitation (6/13 month release/support cycles), I fail to see how they could have made any other decision.
Linus’ decision aside (he has every right to use what-ever DE he sees fit; Plus GNOME/F10 is a very solid choice), I’m currently using KDE 4.2 on F10 (KDE Packaging Project – home of the KDE SIG) and I’ve to hit any obvious bugs. (Even the annoying corrupted tray icon background is more-or-less gone)
– Gilboa
[1] http://fedoraproject.org/wiki/SIGs/KDE/KDE4FAQ
Edited 2009-01-25 06:32 UTC
I liked the way it was framed there ( http://fedoraproject.org/wiki/SIGs/KDE/KDE4FAQ ):
To me, it makes sense, given the Fedora philosophy, to do what they did. If one is running a working Fedora + KDE3 installation, and mindlessly upgrades without wondering whether there’ll still be any KDE 3 on it, I wonder if Fedora is really the distro for you.
All this has given Linus T. the wonderful opportunity to get used to the not-that-horrific-after-all Gnome desktop.. 😉 What else could one desire?
KDE 4.0 certainly was not good enough for most users. I think it was a mistake calling it KDE 4.0. Maybe it should have been called KDE Technical Preview or something like that. But I certainly think it was a good idea to release it. I think release early and release often is definitely a strength of the open source model.
I am using KDE 4.2 as my desktop for quite some while now and for me KDE 4.2 is the best Linux desktop I ever used. KWin always was my preferred Window Manager and I think KWin in KDE 4 is the best window manager out there. In my opinion Oxygen and Plasma look beautiful and professional. Many of the KDE 4 applications improved over the KDE 3 versions and KDE 4 adds some great new applications like Dolphin (well, there was a KDE 3 version, I know), Okular and Marble.
But the most important thing is, that, as a developer, I think KDE 4 is exciting. Qt4/KDE4 provide a rich framework to write applications, the APIs are nice, there is plenty of documentation. With QtCreator, there is also a quite nice IDE to learn to code with Qt. I think KDE 4 has a very bright future and in a couple of years noone will complain about the rough start anymore, but people will enjoy the best desktop environment out there.
4.0 was basically a pre alpha release – horribly unstable and missing key features.
4.1 is basically an early beta – slightly more stable, but still very flaky.
4.2 is getting buzz that it’s great and solves a lot of the problems of 4.0 and 4.1.
I’ll keep trying KDE 4.x, but only in VMs or Live CDs, just to play around with it.
But I won’t do a native install on an existing native Linux install, and try to rely on KDE 4.x as a main, everyday desktop, for a long time probably.
I reckon that when KDE 4.x has made it into Debian stable, it will be ready for every day use. That’s probably at least a few years off. 😉
Meanwhile, I’m quite happy with Gnome, and KDE 3.
And I actually generally prefer Gnome over KDE, probably by a 55/45 margin. I’m a “less is more” kinda guy when it comes to technology. I like “basic” features, with advanced features kept out of immediate view. I don’t like my cell phones, DVD players, TVs, Office software, development tools, etc etc to suffer from feature bloat.
“I thought KDE 4.0 was such a disaster I switched to GNOME,” he explains, “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 could have said exactly the same thing, except that I am still using KDE 3.5.
But, just like Linus, I’d rather move to GNOME than use KDE 4.
And no, KDE4.2 is still far from being usable. Ladislav of DistroWatch reckons that the first usable release will be KDE 4.4
Ladislav of DistroWatch has a huge bias against KDE 4.
This is a bit like saying that Steve Jobs doesn’t think Vista is any good.
Does he? Honestly I have never realized it.
Oh, that explains why I have been using it every day for the last couple of months. I must just have imagined that it was usable for me.
It is also a matter of personal taste. Some friends of mine do everything in a Terminal and they use a very light GUI, if any.
BTW, how comes that you have been using KDE 4.2 “every day for the last couple of months” if it isn’t even released yet?
Obviously I was talking about KDE4 in general, including 4.2.
because he’s been using the svn snapshots, beta’s and rc. What that says about the quality I leave to others to determine…
Obviusly both you and him are wrong, and it’s proven every day by Mandriva and OpenSuse users running their KDE 4.1 based desktops.
Obviously having different opinions doesn’t mean being right or wrong. Who is right, Mandriva or openSUSE which push KDE 4 or Debian and Slackware which haven’t adopted it yet? Personally I prefer Debian and Slackware more conservative approach.
I tried move to GNOME, but i allways use KDE, its dificult to move. And im happy do 4.1 and tomorrow (27.01) is comming the 4.2 i hope that rigth click is back
I am curious, what is this right click that you and linus are referring to?
Linus’ primary complaint with Gnome, and the reason he switched to KDE, was that he could not configure right click to do the action he wanted. I think it was on the title bar of a window but it may have been the desktop (both configurable in KDE 3). Regardless Linus was frustrated with the tight UI restrictions of Gnome as well as something involving printing and in his unique manner went over with flames trailing him to KDE.
Now I believe he is saying that he will have to live with the inability to configure this until KDE 4 sorts itself out. Or until something else comes along….
EDIT: I do not think right click configurability is back in KDE as of 4.2. I work from KDE 4 sometimes to get a feel for progress but have not delved too deep yet. Configuration wise the options seem a bit anemic.
Edited 2009-01-26 21:35 UTC
I always liked him, but now I am really getting to love this man! He always says things I wished I said, but he features a certain way of saying that it will disalodge people from their comfortable chairs.
Exactly. Well said. I wish I could mod you up.
I bet BTRFS will rock the world =). Sun made a mistake – they released their good ZFS filesystem. They raised the bar and now you can not agree with anything less than this. But well, they did not allowed to include it into Linux by using their stupid CDDL. So, Linux gurus and Oracle did what they should – they created btrfs, which inspired by best filesystems, including ZFS, too and accumulates their advantages and lacks disadvantages. So, it will be even better than ZFS. And guys from Sun are losers, once more. Now they have to compete with Linux and all filesystem gurus around rather than have them to collaborate on ZFS. On other hand, it was worth of it. The thing is that btrfs is a really promising filesystem. It is raw and unstable today but *already* kicks asses in many patterns during benchmarking, beating everyone to the hell, including zfs, too. It is hard to imagine what will happen when btrfs will be finished and polished. Competitors will simply have nothing to put on opposite side of the ring, he-he.
Strange logic from a Linux fanboy.
Show me some better code. Oh, it’s not written yet?
In fact it’s the GPL that forbids the usage of CDDL code, not the other way around.
This is where it gets funny. Hey, btrfs is not yet ready, and lacks most of ZFS’ features.
Btrfs has to play a little catch-up first…
My guess is, as btrfs will be enterprise ready earliest in 2012 (according to the developers), many users will switch to OpenSolaris as they are bored to wait.
Promising, ok, but not enterprise ready for many years to come…
Like everything in Linux: “We don’t care if it’s correct, as long as it is fast”.
Please do a feature comparison. Don’t forget, ZFS will not stay still, and wait for btrfs…