JUL
12
2008

11 Myths about KDE

As a response to recent negativity on the Internet, we've been working with Groklaw to get a story running detailing facts about questions such as "Releasing KDE 4.0 was a mistake", "I am forced to use the kickoff menu", "The whole KDE4 desktop interface is radically new". among others. Thanks go out to Pamela Jones for giving the KDE community a chance to rectify certain points that have recently been said in public. This way, we hope to make it easier for journalists to put KDE's direction, recent decisions and put simple myths into the right light.

Comments

Hello,

sad to hear about Jacob getting lost. By "hostage" I mean that the release was only possible once Plasma had reached a barely useable state, and that took longer than anticipated.

About Aaron, I hardly would have wanted Aaron to do something he didn't want to do, not that I could have, I also didn't want it. My complaint is fueled by the "No plasma on Windows" part, I got the impression that Aaron had indeed was not being reasonable about Plasma.

And about Sonnet. It's a real pity, because with it, people would take more about enterprise readiness of KDE4. Such a unique feature would tremendously benefit the workflow of many people. I still hope it will be done one day. Sadly one few can do it, it appears.

And no doubt, working on Plasma is not bad. I have no concern that in the end Plasma will not be a killer application for KDE4. It will be.

I just have it on my mind, that with not as aggressive hype/vaporware Plasma, we probably would have both the old stuff and the new stuff and be happy. I can't prove it, but it seems unusual to me how that Plasma thing went.

Did KDE ever before abdandom such critical components before there was code? It could well be a sign of maturity.

Yours,
Kay


By Debian User at Sat, 2008/07/12 - 5:00am

> About Aaron, I hardly would have wanted Aaron to do something he didn't want to do, not that I could have, I also didn't want it. My complaint is fueled by the "No plasma on Windows" part, I got the impression that Aaron had indeed was not being reasonable about Plasma.

What do you mean by "No plasma on Windows"? I might have missed talk on that, so I just want to be sure what exactly you're meaning. What I do know is that libplasma is platform agnostic (i.e. no direct dependencies to Xlib etc.) and for example is used in Amarok, even on Windows.

Further, Plasma uses/has to use X.org libraries to integrate with the X server. There is simply no way to avoid X.org libraries when you're implementing a desktop. The advantage with Plasma (again AFAIK) is that most of it is platform agnostic, offering a Windows developer the opportunity to possibly port it to Windows. (Even on Windows, it is possible to swap the desktop shell and set Konqueror as default file manager although explorer.exe will go defiant at that point.)


By Stefan Majewsky at Sun, 2008/07/13 - 5:00am

Hello,

I am referring to statements on a blog no longer open to the public. But once upon a time, there was discussion why Free Software should not be ported to Windows. That was of course before the decision was made that KDE will be ported. And later there was a definite "No" to a Plasma port in that blog. I think the argument was that Plasma should be "Linux" (or alike) specific.

There is and must be a part of Plasma that abstracts the underlying display technology, but it certainly need not be Xorg that is abstracted. It could well be MacOS or Windows stuff there.

Yours,
Kay


By Debian User at Sun, 2008/07/13 - 5:00am

The arguments for not porting Plasma(The context was always, the Plasma desktop) are rather simple, straight forward and obvious.

The non X11 platforms, Windows and OSX, already have desktop shells.

Since running two desktop shells simultaneously are a real mess, it would require Plasma to replace the native ones. But since the native desktop shells are more or less required for applications to run on those platforms, it would requre that Plasma needs to implement the necessary functionality to stay compatible. Not only would it be lot of work and rather quickly run out of scope of Plasma, it would most likely also requre invasive changes to Plasma to make it work.

So from a technical angle it is basicly a really bad idea. And when you look at it from a Free Software perspective, it would be rather braindead. Why waste lots of work on trying to replace the standard solution on proritare platforms, when you can spend the resources on improving the Free solution and continue to make it better than the alternatives.


By Morty at Sun, 2008/07/13 - 5:00am

Hello,

I accept the point that technically it will be difficult.

I not accept your point on Free Software perspective. You know that Free Software _did_ replace everything on everything so far. From replacing the UNIX tar, ls, shell, etc. to the Windows IE, people have and will replace everything they can.

And why would KDE bring its applications to Windows with your argumentation. Why then an Amarok where there are already "standard" media players? Why a Konqueror where there is already an IE and Explorer?

Once Plasma is better than other desktop shells, I believe it should be ported to the other systems. For a simple reason: More users, more developers, more exposure, more press, more ideas, ....

And who to do it? Of course not us who care about Linux, but obviously KDE is going to attract enough Windows/Mac developers, starting with its applications. And probably some smart guys will work out how to deal with the system tray.

Yours,
Kay


By Debian User at Sun, 2008/07/13 - 5:00am

One if his reasons, I guess, might be to give linux SOME advantage over windows... Anyway, his point is HE won't be working on it - anyone else is free to do so, this is FOSS.


By jos poortvliet at Mon, 2008/07/14 - 5:00am

"I am referring to statements on a blog no longer open to the public."

incorrect.

"I think the argument was that Plasma should be "Linux" (or alike) specific."

the plasma workspace will remain on Free operating sytems, yes. libplasma has been portable from day 1.

there may well be efforts to port plasma as a desktop shell replacement at some point by others, but my own focus remains free software systems.

this in no way hindered any effort within kde, and i resent you making such an insinuation.


By Aaron Seigo at Sun, 2008/07/13 - 5:00am

What about libplasma being used in Amarok? I assume that it will be compiled for Windows, so plasma effects to an extent will be available on Windows. I'm guessing the ZUI, taskbar, containments and the like won't show up on Windows however. Is this correct?


By T. J. Brumfield at Mon, 2008/07/14 - 5:00am

As he said - he won't work on it, but this is free software so somebody else can.


By jos poortvliet at Mon, 2008/07/14 - 5:00am

Hello Aaron,

I welcome that the blog is again open, that's great by itself.

I also stand corrected and ashamed, as my googling showed that you never said what I thought you said. I might say I picked it up in another forum, but I don't think so.

Well, so guess what, I am happy now and hope you are not too angry with me.

Yours,
Kay


By Debian User at Mon, 2008/07/14 - 5:00am

>"saying "KDE4 is finished" was a Myth is sort of funny. I don't recall anybody >ever say that"

> Plenty of people seem to be assuming it, if not outright stating it.

The problem is that is was officially released. Yes there were notes that 4.0 was for development and testing. For me that really means that it is Alpha or Beta code. An official release says to the work that is is ready to use in a production environment. KDE 4.0 was clearly not ready for that.


By Craig at Mon, 2008/07/14 - 5:00am

> Replacing the core desktop components kicker/plasma before the replacement was ready, was in my eyes a mistake.

> So, that's why I would say and suggest KDE that "Replacements shall replace when they are ready."

Good points. However, I am not so sure using KDE3 Kicker and KDesktop or no panel/kmenu and desktop at all would have been understood either.

A release without workspace components would probably be possible technically, but I wonder what people's reactions would have been.
This is quite easier for non-workspace applications, e.g. the KDE PIM apps, but even there the strategy of keeping the existing implementation as long as necessary has created quite some confusion.

Now imaging doing this with workspace apps like Kicker. Some people would probably complain that it is still the same, some would complain that they would need two sets of libraries, some people would probably complain about not having widgets despite some developers working on that.

It is always easy to compare options after the fact but I personally I think that shipping a basic workspace built on top of the new stack still looks better than not shipping workspace at all and relying on distributors to ship a combination with a different workspace, e.g. KDE3's one.

Of course some people, including yourself, would have preferred this option, but since this is more a matter of runtime, e.g. which programs are auto started with a KDE session, it is still available to users, admins and distributors, while keeping the first option available to those who would like to try it instead.


By Kevin Krammer at Sat, 2008/07/12 - 5:00am

> Replacing the core desktop components kicker/plasma before the replacement was > ready, was in my eyes a mistake.

and who was going to get the kde4 ports of kicker and kdesktop working well? at what point would plasma get testing and work?

let's see if it was a mistake in a month's time, in 6 months time and in a year's time.

> Loosing the Sonnet developer was probably related to Plasma, probably not,

that's an accusation not based in fact at all. it's a really low thing to accuse the plasma team over. poor form.

> taking the project hostage to focus on Plasma

the only people who focused on plasma was the plasma team. everyone else worked on their own parts, just as they always have.

i spent a huge amount of my time working on kdelibs as well. so even i wasn't held hostage.

may i ask what your supporting evidence for your claims above are?

> I fail to understand why a non-ready Plasma had to be part of 4.0.0 at all.

because .. it's a part of kde4. not shipping it with 4.0 would've been even more rediculous. 4.0 was a release that it fit into just fine, given what the goal for 4.0 was

> I tend to dispute that the port of kicker/kdesktop was not actively
> discouraged, because I remember it differently

let's not rely on your memory. provide some documentary evidence and we can talk from there.


By Aaron Seigo at Mon, 2008/07/14 - 5:00am

Hello Aaron,

Regarding memory: I can't google your Blog anymore to cite you. That's my excuse for relying on memory. You bet, I would try to back it up. Is it easily possible to make the old Blogs accessible ?

I don't think it's wrong to say that you encouraged work on Plasma instead of work on kicker/kdesktop, calling it dead early on. That was not a necessary step to do and that was the mistake. I am not calling it a mistake that you yourself didn't do it, you sure beared a lot maintaining it in KDE 3. And I am not considering Plasma anything but (yet another) genious deed from you, even if its currently visible state is anything but pityful (I was using 4.1 SVN for some time).

I would like developers to not push their own projects at the cost of other projects, that's my point.

And that accusation, I was just trying to give an example. I didn't say I knew for real. But that's my argument, with Plasma breaking the desktop experience not only for 4.0.0 users badly, but also all the time during development, the incentive to use that desktop was reduced. That may have cost KDE valuable developers. I somehow always assumed that the lack of Sonnet was related to KDE4 taking forever to get usable. I take it could be something unrelated to KDE though. So don't see it as an accusation, just me trying to back up with an example.

In reverse, if you - only for a minute - accepted my theory that KDE4 would be stronger in other domains, if Plasma had been developed slower, but with less breakage of the existing desktop. Wouldn't you imagine that a more functional desktop at late Alpha time, would have encouraged even more participation and more pillars of KDE4 to succeed earlier or at all?

Your fundamental assumption is that a broken desktop shell didn't harm the development of KDE4 at all. That can't possibly be right, can it? For sure, people could work around it, using KDE3 as desktop shell and only a handful of KDE4 applications. But does that generate the excitement and overall inclusion into KDE4 that many people wanted? For me it did not and I can imagine for others it did not as well.

And just to note: And your time in kdelibs was very well spent. I have learned a lot just from observing your refactorings there. That was really fun, I guess to you too. And thanks to Trolltech for letting you have the time, I am 100% certain KDE 4 libraries are a big time better due to your involvment. You really make a difference.

So please, the trouble and all that aside. There is a point to be had. I think you should not have discouraged kicker/kdesktop porting from the start. You could have done Plasma even when a KDE 4 alternative to it existed.

Obviously now, all is sane, and KDE 4 and Plasma do generate the excitement it takes. After 4.1 and 4.2 releases much relief will be had, and it will all be history. But can we still learn from it?

Yours,
Kay


By Debian User at Mon, 2008/07/14 - 5:00am

If Aaron and the team had ported kicker etc instead of developing plasma, when do you think it would be a suitable point to release Plasma? Within the KDE4 cycle or hold it over until KDE5? Can you imagine the reaction if plasma replaced kicker etc later in KDE4 development cycle? There are enough people here that couldn't read or comprehend the very simple rationale of the KDE4 4.0 release as it is.
It would be a complete waste of resources to port kicker etc all then throw it away.

KDE3 to KDE4 is a paradigm shift so it makes perfect sense to make the decision to develop and introduce Plasma in KDE4.

Stick to 3.5 until KDE4 suites you.


By Ian at Mon, 2008/07/14 - 5:00am

Hello,

there are two important things about release times: Release early. And release when it's ready.

The suitable time to replace would have been around 4.2 when Plasma will have become fully able to mimick kicker/kdesktop by means of configuration. That would not have caused uproad but rejoicing instead. And until then, people would have had the freedom to use it already.

About that waiste of resources, that would be in the hands of those who do it, to decide. But do you really consider that the community and developers have not lost more time, because of it?

I think Plasma would have received only positive feedback ever, had somebody stepped up to port kicker/kdesktop combination, even if only for a limited time. That large amounts of negative feedback itself must have worked to the detriment of Plasma. Who wants to be a part of a project that is seen negatively? May have turned people away from it.

And Plasma in KDE4. Sure. But not necessarily exclusively. And while I give up the notion that the port was being actively discouraged (there is no proof for it), I still uphold that the port would have been a huge benefit to the project and not resource loss.

I will try to list the losses we incured:

a) Momentum. There was a time, when people started to think alternatives would be practically dead in the water once KDE4 will start releases.
b) Nobody even considers honestly to make 4.1 the default desktop, because 3.x remains a must-have rival. The desktop shell of 4.1 may still not be stable enough to build on it, although it certainly is getting there. With a kicker/kdesktop port as stable as the old one, it would be a no-brainer.
c) I feel there was a chance to loose Aaron as a developer. We definitely lost some of open attitude that was a huge community bond. He is making such a huge difference on both fronts, we won't want to loose him.
d) Delayed 4.0 release. Plasma was one of the reasons that caused the delay. With an alternative ready, it would have been simple to delay it past 4.0.0 easily.
e) Less feedback on 4.0 before release, because the lack of usable desktop shell prevented early adopters to run/bug report the code for an extended time.

All of that makes me regret that I have not have undertaken the effort to port kicker/kdesktop. I mean, how hard could it have been. But in all honesty, I didn't think that Plasma would be so much work, so the issue came only up once all the damage was already done or happening.

Yours,
Kay


By Debian User at Tue, 2008/07/15 - 5:00am

4.1 is absolutely a completely fine desktop. I don't doubt most distributions will make it the default in their next release (or 4.2, which should be released in six months or so, if their next release coincides with that). The latest up