Did Linus Jump Too Soon?

One of the many great things about Linus is that he doesn't bottle it up: he speaks his mind on things that matter to him, without worrying overly about what others might say as a result. And when he mentioned in the course of an interview that he had switched from KDE to GNOME, others soon had plenty to say on the subject. But I don't want to revisit those arguments about which is better today: instead, I want to explore the possibility that Linus decided to jump to GNOME at precisely the time when KDE could soon leapfrog it in important ways.

What has struck me while all the firestorm over Linus' comments and the relative merits of GNOME and KDE has been raging, is that meanwhile, in the background, a couple of very significant developments that might change the state of play in the area of free desktops have been taking place.

The first is the decision by Nokia to release Qt under the LGPL. As I've written elsewhere, this represents the latest stage in a pretty extraordinary journey by the creators of Qt, from proprietary, to non-GPL open source, then to GPL and now LGPL. The importance of that move is not so much on the PC as for other form factors, such as mobile phones. As Mark Shuttleworth was quoted as saying in the Nokia press release announcing the move:

"Qt is used extensively in Kubuntu and KDE applications, and Canonical is delighted to see this breakthrough in its licensing model," said Mark Shuttleworth, founder of the Ubuntu project. "Qt's new licensing terms will help us deliver ever more 'lustful' applications to users. Nokia's continued investment in cross-platform Qt libraries, and the Linux platform, is a major driver of innovation in the free software desktop and mobile device stack."

Already, there are hints of Ubuntu Mobile Internet Device (MID) Edition moving from GNOME to KDE.

Given that the Ubuntu MID Edition is being repositioned to serve the emerging netbook sector, that move assumes an added importance since it may well be through netbooks, rather than desktops, that general users are introduced to GNU/Linux – not so much as a replacement for Windows, but alongside on this new class of machine.

This is one of the interesting facts to emerge from a recent piece of market research (pdf) carried out by PriceGrabber.com:

One in 10 online users own a netbook, 75 percent own a laptop and 83 percent own a desktop (see Table 4). Of those consumers who indicate owning a netbook, 91 percent also own a laptop and 87 percent also own a desktop. Most netbook owners have all three form factors most likely because each serves a distinct purpose.

The netbook claims to be different, not better, than other mobile Internet devices on the market. It may not compromise price and portability, but it does compromise processing speed, comfort and battery life. With a slower processor and two and a half hours of battery life, it cannot run complex local computing applications and it generally will not serve as a practical device for everyday productivity.

The other development is not so much recent, as a continuing project: the native porting of the KDE applications to Windows. At first sight it might appear to work against the interests of free software to allow Windows users to use KDE programs. But I have long been of the view that the best way to convert people to GNU/Linux from Windows is to convert them first to cross-platform applications like Firefox, OpenOffice.org and Thunderbird. Once they are familiar with these, it should be much easier to switch them across to the same apps running on GNU/Linux. In other words, this breaks up the move into smaller, less painful steps.

Viewed in this context, the appearance of KDE apps on Windows could be a huge boon to GNU/Linux. For this takes the idea a stage further by offering a much broader set of integrated applications on both platforms. This would make switching between the Windows and GNU/Linux versions even simpler than it would using separately-developed programs. It would be an even bigger win for KDE, since Windows users, once on GNU/Linux, would be unlikely to move across to GNOME without any clear need to do so.

What interests me about both these developments is that they are essentially orthogonal to all the arguments raging around the comment from Linus. They are about the future, not the present. And in any case, as Linus has shown in the past, he's not wedded to any one desktop platform: as soon as something better comes along, he'll probably have no compunction in moving again. It's best to regard his recent action is provisional, and likely to be flipped as soon as KDE provides him with a reason to change his mind. Given the innovative work occuring around the main KDE code, maybe it won't be too long before he jumps back the other way – and sparks off another flamewar that sadly generates more heat than light.

Glyn Moody writes about open source at opendotdotdot.


Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

I think Micro$oft is doing a

Anonymous's picture

I think Micro$oft is doing a wonderful job subverting GNOME with it's mono. The thing is, you don't have to kill your enemies, it is enough to confuse them and make them waste their resources. That way they can hold onto their monopoly a little bit longer before evolution flushes that ugly dinosaur down the toilet: Once GNOME has developed even more mono apps, their patents will start to talk. M$ was bluffing when it was talking about the patents in Linux (kernel) and we and they know it. This time it's different.

But there's light at the end of the tunnel, as there's xfce and lxde. And at the end of the day, who needs a desktop environment. It's just very sad to see how incredibly gullible and stupid many developers are. Micro$oft is no charity.

#1 programming mistake

Roland's picture

Just found this: http://www.joelonsoftware.com/articles/fog0000000069.html

"... making the single worst strategic mistake that any software company can make:
They decided to rewrite the code from scratch."

"There's a subtle reason that programmers always want to throw away the code and start over. The reason is that they think the old code is a mess. And here is the interesting observation: they are probably wrong. The reason that they think the old code is a mess is because of a cardinal, fundamental law of programming:
It’s harder to read code than to write it.
That's the downside to major innovation.

Good-bye KDE?

ppyo's picture

I have been a KDE user since it first saw the light. But now, like Linus, I am seriously considering moving onto another desktop environment: KDE 4 is so flawed. Eye candy is nice, but it's not worthy if it breaks an environment as bad as KDE 4 did. I am willing to wait to see if KDE 4.2 fixes all (or the majority of) those many problems and annoyances. If not, I'll have to choose among Gnome, Xfce, or Fluxbox. It's a shame, as KDE 3.5 was very good!
Now, that is one of the beauties of Linux: I CAN choose! 8)

Ppyo - A proud Linux user since '96.
Distros in use: Ubuntu, Jolicloud, Android, Zubuntu/Cacko (Zaurus).

The more the merrier

quixote's picture

Ordinary users don't give a fig about which OS they run. To use the overused slogan: it's the apps, stupid. Converting people to FOSS browsers, office apps, games, whatever, will mean only that they'll jump the Windows ship at the next problem. such as, say, when their next computer costs $200 with Ubuntu or $350 with M$.

Microsoft will, of coursee, drop its price, like they do in developing countries. (All the way to $3....) But that has them in a forked stick too. Once they're no longer rich beyond the dreams of avarice, they're going to get a whole lot less respect.

Linus, btw, is right about regressions. Geeks may think that stuff is a challenge, but ordinary users will get "OMG, that's the one that didn't work" fixed in their minds. Bad idea. Because KDE 4 is currently more broken than fixed, I'm gritting my teeth and using Gnome. The gritting happens every time I can't get a window to stay resized, to open where I want it, to name removable media what I want, etc, etc, etc. Hurry up, KDE!

Apps first, OS second

David Brown's picture

Glyn's point that OS apps in windowsland will lure users to Linux is correct. Firebird (before it was renamed Firefox) and Thunderbird came first, because I noticed that all the Windows critical updates were for IE and OE. Then came OpenOffice, because MS Office 2000 couldn't open files from newer versions.

Last year an XP reinstall wouldn't activate after a motherboard replacement, so rather than phoning MS and begging, I tried Linux for the first time. No regrets. Win2K is boxed in a virtual machine for CorelDraw and to let me print pictures on my Epson inkjet.

Make Linux stellar; don't donate to Monopolysoft

Jose_X's picture

Developers doing Winfoss are handing over applications to Monopolysoft who (perhaps with proxy help) will take the foss, polish it/improve it, integrate/extend it, and leave Linux in the cold as the foss versions will be inferior.

Take the time to improve L-I-N-U-X instead of W-I-N-D-O-W-S.

[Sometimes I think: too bad the GPL doesn't apply to the OS that runs the apps.]

PS: Major proprietary companies with stockholders don't have the same win/loss as the smaller players and end users. They have short-term reasons to want to port to Windows. Remember too they can license more restrictively. Restrictions on Windows may actually work to their advantage short term since it's easier for them to deal with Monopolysoft's established ecosystem for near term profits and pretend everything will be OK in the future as well. They might be able to cut out competition from small players by getting high volume discounts. Of course, everyone can win more on Linux long term because there are always other ways to leverage large size. Advice to those that don't need to meet stringent near-term profit goals but want to profit long term: dump Win... and support L-I-N-U-X.

Build community for Linux, not for Windows.. don't port.

Jose_X's picture

I lean to the negative wrt kde on windows. On the one hand, it does make it more likely windows devs will target kde instead of something like mono ( http://boycottnovell.com/2009/02/04/the-api-trap-part-1/ and http://boycottnovell.com/2008/11/25/jose-on-mono/ ). On the other, the value to Monopolysoft of having foss run there is significantly superior (imo) to the issue of windows people getting used to linux. ( http://boycottnovell.com/2009/01/15/gnu-linux-tops/#comment-58683 ). Many people jump on Linux if it is there installed and don't even realize it.

Linux difficulty is the ecosystem. Most people would find it no more difficult than Windows if perhaps easier. The key is when you call your friend up, does that person help you with Windows or with Linux? FOSS migration to Windows hurts Linux. It strengthens the Windows proposition (applications at fingertip from Windows and community supporting Windows). What helps is to focus on improving Linux+FOSS integration and quality and to build local community: http://www.linuxtoday.com/news_story.php3?ltsn=2009-02-03-015-35-IN-KN-0000 .

It just makes too much sense.

Most success on Linux are simply because there is someone nearby to guide you and resolve problems. That is hardly different from what makes Windows succeed for the average person... well, actually, there are more issues (see below), but they have little or nothing to do with apps being too difficult. A quality app will intrigue and attract no matter how unfamiliar it is at first. There is youtube if you want to create a series on using app X on Linux. Remember, most things people do on the computer look almost identical from Linux: surf the web.

The seat of power of Monopolysoft is the OS. The OS has value mostly because of what is able to run on top. Microsoft needs developers developers developers.. and it's to build apps for their platform. Their platform without the applications by others is very very boring.

And there will always be issues in moving to Linux. The app familiarity (a red herring for the most part) does nothing for the real challenges: the ecosystem; service provider support; website issues; and hw issues including dual booting with new Windows versions and acpi issues.

Make it interesting (beyond what you can find normally on Windows) if you want people to migrate. See the link above about adopting/building a community. In short, support Linux. Make Linux interesting and stand out. The more and faster, the better.

OK, one more time:

The key is to provide services useful to people (perhaps little or no relation to OS, or perhaps a computer application), but use Linux as the vehicle. Integrate and build upon openness, upon Linux, not upon Windows.

PS: More Linux tech/biz types should be courting small businesses (or large) to help them leverage Linux' customizability to help their business, both in terms of meeting technical needs as well as marketing needs. Think of all of that bundle of magic that is licensed (all of it) in ways so different from Windows+friends. People exploit the freedom in the licensing when you court businesses and users. There is an awful lot of free stuff. Customize to give the business customer advantages.

Linus' choice was mis-interpreted

Robert Fendt's picture

If I recall that interview correctly, Linus was using Fedora at the time and they pushed KDE 4.0... and let's face it, KDE 4.0 was not for general use, I would not even say KDE 4.1 was ready (and the KDE team never said it was). Torvalds also did say in the same interview that he still hates the fact that GNOME is not as configurable and that he would like to take a look at a future KDE incarnation.

So, what does this mean for KDE4? Absolutely nothing. A bad decision by a distributor (and Fedora is not the only one: bad decisions concerning half-baked KDE4 versions were also made by e.g. Kubuntu) made a user switch to GNOME for the time being, nothing more. So what's the big deal? I do not see it. Let's see how KDE 4.2 fares, since it's the first version meant for the general public.

Did you read too fast ?

Alejandro's picture

In my opinion, many readers focused on Linus' choice and missed the main points of this very interesting article :
- New licensing for QT, means wide adoption, on all sort of devices.
- Kde apps porting to win$, means easier OS switch.
Good night.

KDE 4.1.x

Sergei Mutovkin's picture

Up until a few days ago I was running Kubuntu 8.10 with latest point release of 4.1. It was unusable in many places up until mid-December 2008 for me with my nVidia video card and proprietary driver. However with release of 180.xx nVidia Driver everything has changed: KDE started to run fine and all issues got resolved on their own all of sudden.

All in all, my point is that KDE 4.1.x were really solid releases and were very usable with proper hardware and drivers.

Those who think that KDE 4.2 is not ready for average user, most likely are the same people who claim that Linux desktop is not ready in general. Linux desktop is ready for you only if you are ready to accept that things in other operating systems are done differently from your preferred previous OS/environment.

KDE 4.2 is really beautiful, fast and feature-full environment. I run it here non-stop pretty much since it was released for Kubuntu and have no crashes or other problems. All distributions that will be released this year with new drivers from ATI/nVidia that will accelerate Qt fully (and hopefully also accelerate video decoding via VDPAU, etc) will be amazingly user-friendly and nice looking. It is finally time to remove very very very old, windows-95 looking and featureless Gnome from default desktop environments and let users experience real next-gen user interfaces written in Qt and KDE libs. Some of the KDE applications are really best in class: Okular is outstanding as is Amarok 2.0, Korganizer is great as well.

There are still weak places in KDE though, you have to use slow, painful, badly integrated GTK+ based Firefox, since Konqueror's team decision not to drop KHTML and use latest-greatest Webkit like everyone else was a really stupid idea. Koffice 2.0 is not yet available either. Digikam 0.10 for KDE4 is not yet available (at least on Kubuntu 8.10) and thus not integrated (I believe image management will become #1 feature on the desktop and MUST be integrated really really well). Though last point does not give any point to a slow and frequently crashing F-Spot. KDE does not have a good video player that is as good as Amarok for audio. One day when the last few points will be addressed Gnome environment and many GTK+ applications will become much less relevant.

faeris@faeris:~$ cat

Anonymous's picture

faeris@faeris:~$ cat /etc/apt/sources.list | grep digikam
deb http://ppa.launchpad.net/digikam-experimental/ubuntu intrepid main

(Digikam 0.10 in Ubuntu ;))


JK Wood's picture

The story became popular at just the time when KDE is ready for him. He actually switched a year ago, when Fedora pushed 4.0. This would be an opportune time for him to switch back, true, but this is not something he did in the past two weeks.

Good point

Glyn Moody's picture

It's not clear to me from the interview when he did change - although obviously not, as you say, in the last fortnight: do have a reference for him switching fully a year ago?

My apologies...

JK Wood's picture

I opened my mouth too soon. KDE 4 was made the default desktop in Fedora 9, released in April. I can't really assume to speak for Linus, but I imagine he upgraded sometime fairly soon after that.

Gnome is the default desktop

Anonymous's picture

Gnome is the default desktop manager in Fedora. If you chose to use KDE instead you have to use the 4.X versions.

4.0 should have been called

Anonymous's picture

4.0 should have been called 4.0 ALPHA.
4.0.1 -- 4.0 ALPHA2
4.1 -- BETA
4.1.1 -- BETA 2
4.2 -- 4.0 "USER RELEASE"
4.2.1 -- 4.0.1 (no more disclaimer tags required)

Blogs and tech sites were way more ebullient then they should have been. Any distro that entirely replaced KDE3 with KDE4 before last week fucked up big time. None of them should have made 4.0.x or 4.1.x the only KDE option for their users, period. Other than that (having been using 4.2 since its release), it's difficult to fault the KDE community for their course of action. Hopefully(!) they have learned how /not to/ name a release in the future.

That said, KDE 4.2 is pretty damn great (and previous releases were so unsuited for "users" that the 2 or 3 I tried never lasted more than a few hours on my system). It is incredible how far the KDE community has taken KDE4 in just one year. It's a testament to the power of Free/Libre software and the communities that make it what it is.

Great work KDE Community! Linus will surely come around soon.

KDE 4.2...

Anonymous's picture

KDE 4.2 should have been the first version of KDE4 released at all to the general public, with warnings that most people shoud avoide KDE 4 until 4.3 or 4.4 or they would just have to cope with the problems. In other words, KDE 4.2 is still not really ready for general release and should be considered a beta test version at best.

KDE 4.2 indeed isn't ready

Anonymous's picture

KDE 4.2 indeed isn't ready for prime time - it still hasn't reinstated crypto certificate management, which I consider to be one of the most important features. Else you're not going to do anything than sell users a false sense of security if no KDE application can effectively check against MITM attacks - laptops/netbooks abroad, in unencrypted or weakly encrypted WLANs, is havoc waiting to happen.

The mistake

tracyanne's picture

was to release Distrinbutions with KDE4 as the default desktop, that was just plain stupid. The arrogance was that everyone was expected just have to wear the problems of a half arsed desktop, that may well have wonderful innovations in the pipeline, but which was not in the least ready for general use.

My noobs, my little old ladies, would never have coped with KDE4 (4.1) as it was on Mandriva, and I expect on other Distributions, the one I tried it on begged me to put her back on to the "one that worked" (MDV 2008.1 with KDE 3.5.9), which I readily did, and promptly told all the otehrs that we staying on that release until things improved.

Off topic?

John Stowers's picture

I thought it was quite clear why Linus jumped; a fundamental disagreement with how releases and regressions are handled. Linus comes from the world where regressions are not tolerated, and the KDE 4 release motto seemed to be "shit is going to break, people will have to deal with it, but we will let the KDE apologists sort it out".

Reading anything more into the switch is a complete waste of time.

Window Managers

J S's picture

Linus would seem to be interested in speed as well, having to write tight code for the base kernel, and understanding the amount of resources taken by a window manager. So I always have wondered about his proclaiming one WM vs another is preferred (half expecting a CLI to trounce preferences).

I used to run KDE, but have switched to a mix of XFCE and Fluxbox.

The effects are easiest to see on marginal computers - Fluxbox and XFCE work ok on 500Mhz+, KDE 3.x on 750Mhz+, Gnome on 1Ghz+ systems. Of course Vista seems to need 3Ghz+ multi-core machines to launch properly.


Glyn Moody's picture

I think his view has changed. Initially, he said he wasn't really interested in "user space", but I think he's recognised it's now an important issue for the future of Linux (not least because the server side is going so well).

KDE jumped too soon

Roland's picture

I agree with Linus. I am staying with Kubuntu 8.04 & KDE 3.5.10 for now.
I think KDE 4.3 and/or Kubuntu 9.10 will be great, and that's what I intend to use when they are released. The KDE team is doing wonderful work, but they should not have tied such massive innovation to a schedule. Great innovation doesn't follow schedules.

The important thing...

Glyn Moody's picture

...is to learn from the experience.