Today sees the release of Kubuntu 8.10 featuring the latest KDE 4.1 desktop. The Kubuntu developers have been hard at work, integrating this major new version into a completed desktop. The settings and artwork have been kept close to KDE's defaults to ensure the best face of our favourite desktop shines through. Desktop effects have been enabled by default for cards which support it thanks to the wonderful KWin and package management comes via a brand new version of Adept. Printer Applet and System Config Printer KDE were written to ensure a complete user experience, both are now part of KDE itself. Update Manager, Language Selector and plenty other tools have been reworked for KDE 4. Upgrade, download or request a free CD.
Dot Categories:
Comments
Maybe they should have called KDE4 "KDE-X alpha alpha", but they didn't. It was released with a 4.0 stable build number, giving the impression of a stable release.
Still, you've ultimately missed my point. While the bugs, flaws and general broken nature of KDE 4 is somewhat disheartening, the truly scary thing is the direction it's taken. It really wants to be Vista; not even Windows users generally care for Vista. It's also trying to dumb down the KDE experience, which has always had an advantage over Gnome precisisely because it /didn't/ dumb things down, but rather gave the user lots of options should they choose to exercise them. You can't out-vista Vista, and you can't out-gnome Gnome. What kind of sense would that even make anyway? If we wanted Vista, Leopard, or Gnome, we could simply opt to run those.
As Anon nicely says, you seem to assume that things are " the way they are now "by design" ". There is nothing like a "broken nature of KDE 4". It is not its nature. It is its temporary state.
In many places, I have seen: "take care, KDE 4 may not be for you." And this warning is still available on Kubuntu website for the 8.10 release...
Even in the limited state in which KDE 4 currently is, it already has many many features. I think KDE policy remains the same. Just wait a bit more (6 month to one year) to see KDE fully back! and actually better than it has ever been.
I really and sincerely hope you're right. I'm the type that can make due with TWM if I have to, but I'll really mourn the loss of KDE if it doesn't find its way back to its powerful roots.
"I really and sincerely hope you're right."
I'm a recent dev for KDE, and am focusing on restoring some functionality to Konqueror4 that was lost as a result of the port to Qt4, starting off with features that when implemented in Dolphin would automatically benefit Konqueror (my old favourites like previews-in-tooltips, expand-folder-on-hover, etc). Every dev I've interacted with has invested a great deal of their time and energy in helping me to do this, so from my (admittedly so far limited :)) experience, there is absolutely no conspiracy to "dumb-down" KDE. All that is needed is more people stepping up to help :)
That's honestly the best KDE news I've heard in a while. I hope you succeed!
Here here!
It is 'hear hear', but SSJ hits the nail on the head with a very good, personal example. That's exactly what's going on: lack of time. 4.2 will be a lot better, just try the upcoming alpha. And further down the road, think 4.3 and 4.4, we'll really start innovating and doing cool, impressive stuff. Sure, it took us 3 years to build the foundations & get back to 3.5 in terms of functionality - but I think the future will prove it was well worth it.
"While the bugs, flaws and general broken nature of KDE 4 is somewhat disheartening, the truly scary thing is the direction it's taken."
Please provide some tangible examples of the "direction" KDE has taken and why it's the wrong direction. Seriously. You are so strong with your opinion, that answering this simple question should not be hard at all.
What I see is that KDE has added a lot of groundwork and technology that allows KDE to become truly kick-ass system in the future. KDE3 was a dead end. It represented a desktop that was basically Windows95 on steroids.
can't stand the truth? the only thing reminding on win95 is the current stability of kde4.
it's bitter - i know.
Oh I can stand the truth just fine. And I can understand the fact that KDE3 represented ideology of the static-desktop that was common with Windows 95 and it's contemporaries. KDE3 was an excellent desktop of that type, but fact is that world and technology is moving elsewhere.
And you too are complaining about stability of desktop that is 10 months old. KDE3 is currently 6.5 years old. Looking at the longevity of KDE3, KDE4 is literally still taking it's first steps. And we need to remember that move from KDE3 to 4 is a lot bigger more than move from KDE2 to 3 was. If we look at past releases, move from 1 to 2 was about as huge. And KDE2 was buggy, crashy and semi-functional for a long time after release.
Not sure if you'll really get me. My poor English doesn't suffice for seriously arguments. Sorry.
What I am really convinced of, is that the main cause for the KDE4 disaster is to think of a desktop environment as of a matter of ideology.
"ideology of the static-desktop" - i can only hope that you had a grin at your face while writing that ...
But honestly, all intents based on inflated ideologies are doomed to failure.
More in detail now - its easier for me.
For daily work i mostly use bash, lets call shell here. If i would start vi with a terribly broken ~/.vimrc vi would crash. That's what you would expect. But if the underlying shell crashes conjointly, I would cry, curse and rant the hell - Everyone used to use a shell would call that a failure by design! But the very same thing happens with KDE4. A stupid single widget crashes not only itself, but the whole underlying desktop shell too! Really!
So a gravely failure by design - not the first or last one so far. But the big surprise is, how many people doesn't seem to realize that.
Did you really never wonder about that?
Back to your win95 comparison. Microsoft wasn't ever able to stabilize that platform - guess why? Because it's not possible to hide a misconception. They finally had to drop that crap and switch to NT.
And an ideologically slipslop here only obscures the clear view.
Considering that, the critical posters here are much more constructive than the bunch of those, telling the devs will fix it. They didn't in 4.1 and they won't in 4.2 or later, if some fundamental changes won't happen. Even worse, defending the current state and pretending it can be fixed without changing some fundamentals will only delay the badly needed redesign.
And about the devs - you cannot blame those. Most of them working at applications and aren't responsible at all for that kde4 misconception.
If I recall windows 98 was based on the same codebase as 95. It still was horrible mainly cos it's kernel sucked. But if I recall it was far more stable than 95...
I have had plasma crash, but afterwards it started right up again. Sure it is annoying but hardly the end of the world
... yup and windows me was another weak try to stabilize the faulty design by cosmetics - with the very same result as before - it stayed _unstable_. I thought the point was clear what I'd meant without having to mention all those stranded works at the surface.
But you are absolutely right, by introducing that automatically restart it was obvious that the KDE is wanting to go the win95/98/me-way, dealing with the symptoms instead of redesign the real cause.
> Sure it is annoying but hardly the end of the world
But for long time kde users like me, in _need_ for a stable platform and now forced to take a glance at enlightment, wm or even gnome again, it comes rather close to that ;-)
> dealing with the symptoms instead of redesign the real cause
Actually they are dealing with the real cause (non-sandboxed code) and designed it that it would allow avoiding it (sandboxed widgets, e.g. written in JavaScript)
And what about those widgets written in c++? Proof me wrong, but if a little stupid widget can crash the whole desktop shell, then they actually did not dealing with the real cause. And there isn't an excuse for a flaw like that.
Just coming to my mind - even web browsers - in the past not especially known for security, stability or clean design - currently trying to isolate their different tabs, so one process should never be able to crash others or the whole browser itself.
those are the core widgets, and doubtless they will get more stable with time...
Plasma in it's current incarnation is less than a year old
yup, got that already.
But then the question is, how the separation is achieved between core (-widgets) and others? The chosen language as an isolating model? Honestly? And that's not only a matter of faulty design but also of security somehow. I'm sure you can imagine what is possible for those nasty guys with some c++ experience ;-) All they have to do is making an really appealing, seemingly harmless little widget...
They could do that already... Make a nasty little plugin for prog X, then uh oh. Why would you use plugin X? Because you know it comes from a trustworthy source.
You can't download with ghns binary plasma packages, only scripted plasma widgets. Any thing else will either be included in your distro packages (you trust them, right?) or you'll have to compile yourself (in which case you take the risk).
Scripted plasma widgets use the scriptengine which serves as a buffer between the applet api (in language X) and the native c++ api.
> Make a nasty little plugin for prog X ...
I did answer that post in the second thread with you, ~10 posts below ...
http://dot.kde.org/1225379191/1225397878/1225398986/1225399609/122557451...
"What I am really convinced of, is that the main cause for the KDE4 disaster"
There is no "KDE4 disaster". There is just software that simply doesn't yet do all the things it's users would like it to do. That's hardly a "disaster".
"But honestly, all intents based on inflated ideologies are doomed to failure. "
Inflated ideologies like "free software"?
"For daily work i mostly use bash"
Most users don't. If you are a CLI-user, are you really in a valid position to offer critique about a GUI?
"Back to your win95 comparison. Microsoft wasn't ever able to stabilize that platform - guess why? Because it's not possible to hide a misconception. They finally had to drop that crap and switch to NT."
Switch to NT was about switching the kernel of the OS, it had very little to do with the GUI itself. GUI and the concepts behind it of XP and Win95 (or Win98) is fundamentally the same.
"Considering that, the critical posters here are much more constructive than the bunch of those, telling the devs will fix it."
They are not constructive at all. Their critique is vague and unspecific. They keep on telling how "KDE4 is unusable". That's not constructive at all. If they wanted to be constructive, they would provide tangible examples of what exactly is broken.
The most tangible examples we have been told in this discussion is basically that "some functionality that is present in KDE3 is absent in KDE4". And I agree. That claim is true. But the fact is that that missing functionality is being added back as we speak. This is a question of resources and time. In the migration to new codebase, some functionality was lost (and some was gained). But as time goes on, that functionality is added back.
You and many others seem to think that KDE4 is about removal of features. It's not. You are complaining about a release that is 10 months old. You are comparing it to a release that has 6.5 years of developement behind it (even more, if we consider KDE2 to be kind of "pre-KDE3").
Compare KDE4.0 to 4.1. HUGE amount of improvements happened between those releases, wouldn't you agree? And they did that in 6 months. Sure, you might say that 4.1 isn't good enough yet. But they are on a right track. In just 6 months they fixed large part of the complaints users had about 4.0. It doesn't take a rocket-scientist to see that KDE4 is developing very, very fast.
"They didn't in 4.1 and they won't in 4.2 or later"
Many of the requested features were added in 4.1, and more will be added for 4.2. Were ALL of them added back in 4.1? No. There's only so much time and resources to do this stuff.
What exactly are YOU doing to help make KDE4 better? Am I correct in assuming that your contributions so far consist on sitting on your arse and complaining?
"if some fundamental changes won't happen."
And what would those be? Come on, let's hear some "constructive criticism".
"Even worse, defending the current state and pretending it can be fixed without changing some fundamentals will only delay the badly needed redesign."
What changes are needed? What kind of "redesign" would you want? What exactly is broken in KDE4? If you want to be constructive, now is your chance. Vague comments about "redesign" and "fundamental changes" are NOT constructive.
And are you really in a position to demand for "fundamental changes" and/or "redesign"?
> Most users don't. ...
You didn't even read my post! The comparison with bash and vi should demonstrate, what you would expect of a shell. That also applies to a graphical shell like a desktop environment.
> ... are you really in a valid position to offer critique about a GUI?
So stop! It's an impertinence not being able to understand (or deliberately to ignore) what i am talking about, but judging about 'valid positions' - every one using a software is in it. Period!
All your following lines mostly dealing with the missing features in v.4. I didn't ever mentioned those! If you want to make another Obama/McCain drivel here, answering questions i never asked - I'm not the guy for those time wasting nonsense.
> And what would those be? Come on, let's hear some "constructive criticism".
Read the part about what a shell is supposed to be again.
> The comparison with bash and vi should demonstrate, what you would expect of a shell. That also applies to a graphical shell like a desktop environment.
I agree that this is comparable, however one needs to check whether a specific example is comparable.
If an application launched by the desktop shell would take it down when it crashes, that would be bad.
Fortunately that doesn't happen.
The only things that might crash the desktop shell are non-sandboxed plugins, just like non-sandboxed plugins of a shell interpreter can make it crash.
Therefore one of the desktop shell's design goals was to allow sandboxed plugins, comparable to shell functions and scripts in your analogy.
OK, yup I've got it.
So now, is it possible that when I'd write yet another silly analogue clock widget in c++, that widget would be a "non-sandboxed plugin"? Is the choice of language the isolating concept? And if so, that you would not call a design flaw?
The core widgets will be in c++. But in future I expect that they will encourage people to write code that is sandboxed.
It would be helpful if you described what widgets were bringing plasma down, or maybe heaven forbid, file a bug report?
It's a compromise: keeping widgets in process allows for reduced memory consumption and better visual integration, but native widgets will bring down the Plasma process (just as in KDE3, a dodgy applet would kill your panels).
Pick your poison.
No, if I remember right - the opened applications and the desktop shell (that was the subject matter) continued to run quiet well. KDE3.x seems to have a far better modular design to achieve such isolating.
"It's a compromise" - isn't that the most common phrase when it comes to a defense of design flaws? ;-) where i heard that lately? ... our company last Friday :-D
He is talking about kicker applets. So kicker goes down.The desktop remains, so what good is that to you beyond being able to click on an icon?
You are aware the krunner still works if plasma goes down? As do any apps still running and pretty much the rest of kde. What could you possibly be doing on the desktop that means means it is a disaster if it goes down?
> ... so what good is that to you ...
It simply only had crashed kicker. The applications still in front, i can uninterrupted continuing my work or whatever I'm currently up. Alt+F2, MMB etc. all remaining untouched - That in fact is the advantage a modular design can provide you with. No more, not less, but personally i think there are much smaller things related to the 'desktop experience' and usability that had been widely discussed here...
> What could you possibly be doing on the desktop that
> means means it is a disaster if it goes down?
Besides the clear advantage mentioned above, really it shouldn't be as hard for example to imagine a widget crashing continuously to more than only disturb your workflow ...
Ingenuous, would you want your employees work with that, wasting their time and your money ... ;-)
If plasma crashes you can still use alt+F2, and apps remain untouched and kwin still works so you can also Alt+TAB.
Also note that if a widget crashes plasma, either it won't be added when plasma restores or it will be in a previous stable state (plasma saves its config periodically), or if plasma keeps crashing it gives up after a few tries and lets you start it manually.
But seriously what distro are you using that has such crashy behaviour? And what did you try when plasma did crash, for you not to know this?
No, you are simply wrong. First, there is a difference if a 'plugin' you've deliberately installed can crash an application or if a widget (also deliberately installed) can crash the whole desktop shell. Second, a plugin architecture should ever try to isolate its core against those plugins.
Finally, a concept where the chosen language qualifies for privileged 'non-sandboxed' API calls !? Of course you have no doubts about that ;-)
Its always the same when it comes to lousy concepts. First it will be claimed there isn't something like that, and later you can hear that it isn't as bad or others do worse.
But what I learned just now is, that there isn't only the automatic restart, nope! plasma also saves its config periodically. (doesn't sound suspicious?) Thats what I'd call cosmetics - with the result cosmetics ever gain, read the posts here or remind yourself on win95/98/me. Btw. even the ones claiming to be happy with the sadly state kde4 is in now, have to admit that it runs not quite stable. A Desktop shell!? 2008 not 1998!?
> But seriously what distro are you using ...
Mostly I'm running debian, sometimes gentoo, but my latest kde4 experiences i've had with an ubuntu notebook. And also seriously: having plasma crashes there wasn't the same as suffering a kicker crash (must be years ago), in terms of effects and what it felt and in terms of frequency!
When running "stable" i.e. 4.1 I only ever got one or two crashes. Even when running trunk I have surprisingly few crashes.
Saving its config makes sense even in the most stable of stable apps. Besides I've never claimed it was completely stable, Plasma is still very young ( ~6 means that it had lots of time to become stable. Unfortunately it's code was pretty unmaintainable and couldn't be ported to Qt4 easily which was the reason they dropped it.
In my experience if you want a good kde environment you should try either mandriva or suse.
Cause your post only deals with the symptoms, I'll take the absence of reply to the question regarding the subject matter as an answer -
to be said further (though it really wasn't the point):
> Plasma is still very young ( Saving its config makes sense even in the most stable of stable apps
Periodically? (Btw. how often?) What stable apps do you mean?
Anyway, a need for that caused by faulty design isn't just what you would expect after a 'fresh software redesign'. Besides, it would be very illustrative to hear if kicker ever had the same mechanisms ... if not - why? You never wondered? Seemingly others felt the difference between kicker and plasma crashes too. In quality and quantity.
Thanks for the distro advices.
But i'll stay as long as possible with kde3.5. And then, if there hasn't been a fundamental change in kde4, i have to take a look at old friends and foes ;-) unfortunately.
But maybe a new QT desktop has evolved in the meanwhile - a professional one in terms of robustness by concept, one to attract developers to base their software on it -
Just what exactly is faulty by design? (and don't answer KDE or Plasma, I mean which concept in the algorithmic sense, ie what to fix)
If your answer is un-sandboxed widgets, then everything is already explained. Base widgets (ie shipped with your distro) can run un-sandboxed, which allows for grater visual integration. Other widgets (the kind you get through ghns) are all scripted and run in a sandbox.
I'm a new KDE user, I couldn't stand 3.5, which always felt off, and am currently running kdesvn on a production box with very little pain considering.
> Just what exactly is faulty by design?
Thought after the longest thread i ever wrote that's pretty clear now. Blaming my English ...
> If your answer is un-sandboxed widgets, ...
Yup, that's an important part of.
> Base widgets (ie shipped with your distro) can run un-sandboxed...
> Other widgets (the kind you get through ghns)
> are all scripted and run in a sandbox.
That in turn simply isn't just the truth. Every one (not only distros) can write or ship 'base widgets' running un-sandboxed. Naturally not via GHNS, but that - as you surely know - isn't caused by and wouldn't suffice as a security model.
"... choosing which language to write your applet in. C++ and ECMA script will both be available by default"
[ http://plasma.kde.org/cms/1084 ]
And don't get me wrong here - I'm not demanding to close the c++ API. In contrary, who would write in say ruby if he could in c++ :-D
That's why "nearly ever plasmoid on KDE-Apps is a C++ plasmoid"
[ http://mail.kde.org/pipermail/plasma-devel/2008-July/000078.html ]
Probably the point gets somewhat clearer ...
I will try again: Every shell (even most apps with a certain extent) needs to separate its core from additional functionality it probably wants to provide. So first it has to be decided, what functionality belongs to core and what is extension and second how to achieve a clear separation of the scopes. You ever want to shield the inner / privileged against the outer without loosing the benefit that those are coupled. You may have more scopes than only these two.
And in both points kde4 fails deadly!
Again, if a silly analogue clock widget can crash the whole shell, than:
- either it belongs to a far to central scope
- or (and) the privilege separation sucks
> I'm a new KDE user
and I'm a long time one - maybe that's why I'm worried and you apparently don't
gosh! what a mess :-/
to be said here in addition: every single widget should be encapsulated not only against the (core)shell but against each other widget - regardless of the language, of course
o man!
OK, I re-read your comment. I must admit, that before I just glanced through your comments regarding Bash, Vi etc. But, to address your comments:
You are concerned about Plasma crashing? crashes are naturally always bad. But doesn't plasma automatically restart itself in case of a crash? And I haven't witnesses any plasma-crashes in 4.1.
That said, would plasma crashing REALLY be that bad? It would not affect apps at all. Apps are not related to plasma, and neither are app-windows. If plasma crashed, the worst that could happen is that the panel and the widgets disappear. But your app would still be there.
I would be more concerned about Kwin crashing, since that could actually take the app-windows with it (although it would probably just take the windeco with it, leaving the app untouched)
In some ways plasma i somewhat similar to Explorer.exe (which drawn the desktop) in Windows-land. And I have had Explorer crash on me when I had had bunch of apps open. And after I restarted explorer, my apps were still there.
"Read the part about what a shell is supposed to be again."
It's supposed to be the area where the user gets his work done. Be it CLI or a desktop. How does KDE4 and/or plasma divert from that definition? If anything, it provides the user more tools to get his work done. Is that a bad thing?
Or is your point that shell should be about running applications, and it should not concern itself with stuff like plasmoids at all? Plasmoids are just an extension of what we have had since the beginning. Plasmoids are just another step in the road that was started by systray, taskbar, Kmenu, icons and the like.
If you just want a desktop that runs your apps, why not use something like TWM? Fluxbox? What you seem to want is a barebones windowmanager, not a full-blown desktop.
If you want KDE to move towards barebones WM, then it should be said that KDE is not about that, and it has never been about that.
The system I am using at a client company has KDE 3.2 installed. KDE 4.2 is almost upon us, so it's okay to compare the two. KDE4 has much more eye candy, and looks really slick, but when it comes to actual functionality and usability, KDE 3.2 edges it out. I see a lot of scaffolding for what will surely be awesome stuff in the future, but in the meantime KDE3 (even the ancient 3.2) is more productive for me.
Wow, nice to hear something different than "kde4-looks-so-nice" post. I agree with you, and would like to add that with systemsettings it takes a lot longer to navigate compared to kcontrol with its tree view. OTOH, bigger icon everywhere. Progress!!! ;-)
System settings are much more organized than they used to be. Kcontrol was fine for experts, not for all users.
Right. Cause now we have "I wish I was OS X" control panel instead.
Well, if it works better...
Yeah, let's add the expert and basic tabs. Then it will be even more organized. Look, if you create something for stupid users, only the stupid ones will use it. Do you really want that?
yeah, there a 2 possiblities : organize Hierachically or by Subtextsearch.
the old KControl had this. Now i have to guess where my settings are. Great !
I believe someone was working on a KControl look for the systemsettings. Shouldn't be hard to do. Feel free to volunteer...
I want to change some keyboard properties.
It appears in:
General -> Personal -> Country/Language
General -> System admin. -> Keyboard & mouse
Advanced -> Advanced user options -> key combinations
As a "beginner" I don#t know of deadacute keyboard maps!
I know of the difference between "é" and "'e" and I want that I sdwitching desktops is possible using Ctrl+Alt+Left/Right
Now please tell me, how I find these options as a "new user"
Kcontrol was better for both, beginners and "professionals". I agree it is easier to visually find items in a huge icon panel compared with a listview. But the items themselves are badly organized. I bet you agree you were wrong.
> When we say the menu sucks
Then "we" failed to note that the traditional menu is still available.
> When we say kicker is broken
Then "we" failed to realize that it can't be cause there is no kicker any longer.
> When we say Dolphin sucks functionality-wise compared to Konqueror
Then "we" failed to realize that Konqueror is still there.
> Shame on you.
Nope, shame on "we" for wasting space in the internet.
I think tooth's confidence just paid off and all that nonsense was proven wrong :D
Wow, how does it feel to be a cheerleader? Seriously.
But really, to me his arguments sounded like a bunch of excuses. "You can work around the stupid design philosophy by doing X or Y, so you are stoopid", but never actually addressing the stupid design philosophy itself.
Cheers,
-Ben
> Wow, how does it feel to be a cheerleader?
Please tell us.
> like a bunch of excuses
And yet you never did contribute anything. Is this your first contact with FOSS? or probably even with other humans?
## Wow, how does it feel to be a cheerleader?
#Please tell us.
What am I a cheerleader for exactly? I guess I'm just so stupid I miss your point?
## like a bunch of excuses
#And yet you never did contribute anything. Is this your first contact with FOSS? or probably even with other humans?
Negative comments are a contribution. Want something more specific? Try this: "Please, please, for the LOVE OF GOD, don't try to make KDE like Vista, or like Leopard, or like Gnome."
Because I don't support KDE blindly doesn't mean I don't know anything about human contact, or about FOSS. I've submitted my share of patches, though admittedly not to KDE. I've also donated to the FSF and a few others. Not that I think it earns me any cred, karma, whatever, but that love me or hate me, I'm a member of the community too. Just because I agree with Bob about KDE4 doesn't mean I am a hater. But I have a special distaste for those who blindly follow something or make excuses for it, especially when we can potentially make it better. Pretending it already IS better is just a hindrance in the long run, not a help.