Submitted by Mark Shuttleworth, founder of Canonical, recently did an interview with derStandard in which he discussed issues with Ubuntu’s latest release, innovation, the future of GNOME, and other subjects. Perhaps the most interesting thing he said is that Linux does not yet deliver “a good enough user experience.” Of course, you could say that of any operating system.
Editor’s Note: QT-based Gnome was also an interesting point-of-view.
QT based GNOME would actually be very cool. I prefer the look and behaviour of QT based apps rather than GTK. That’s my personal preference.
I hate using KDE and QT-apps(no disrespect to the projects, just personal preference.)
However I think this is a very interesting tings. This way KDE and GNOME coild have a common API.
And why shouldn’t GNOME be able to write programs according to their philosophy on the QT-plattform? (Their philosophy is as we all know a lot more geared towards simplicity.)
GNOME would consider Qt if Nokia donated Qt to GNOME foundation and re-licensed it LGPL, but it will not replace GTK+. Qt would be just like XUL, SWT or any other GUI toolkit that feels (and works) naturally on GNOME desktop… maybe as accepted dependency on GNOME platform!
I don’t know why people keep forgetting something very important: GNOME is not just a user of GTK+, GNOME owns GTK+ (technically wrong but practically correct). Those guys would take GTK+ wherever they feel like doing so for the benefit of^A the GNOME project. [This is something very important] KDE, the biggest user of Qt, are lucky to get if for free, but they have no control over the direction of Qt, they are under the mercy of Nokia’s business strategy. Why would GNOME get itself in such a trap?
I don’t think that’s correct. Many KDE developers work for Trolltech^WNokia so they can influence the direction Qt goes. A good example of this would be the new QGraphicsView features in Qt 4.4 that were added at the request of Plasma developers.
Quite a bit of the work that’s gone into Qt 4.3, 4.4, and 4.5 has been driven by KDE4.
It’s true that they don’t have the same relationship that GNOME has with GTK, though.
But if things were to go too badly, KDE could always just fork it. Then they would have complete control over it as well, they’d just need to come up with the developers.
Er, no. Tt/Nokia has maybe 2 or 3 developers working on KDE. KDE is their showcase “product”, but their business hardly revolves around it.
Certainly they draw from KDE, such as their adoption of phonon, but that just makes sense, it’s a reciprocal arrangement. But Qt is hardly controlled by KDE.
there are only a few of us paid to work on KDE full time, yes. but there are several Qt developers who *also* work on KDE (some during creative fridays) and even more that work with KDE while working on Qt. the new printing support in Qt is a great example of the latter, in fact.
(there are other KDE shops as well, of course; KDAB with Kontact is probably a prime example of that.)
reciprocal is a good term for it.
(i’d also point to the xquery, webkit and printing support for further examples)
indeed; and imho that’s a good thing.
Qt is influenced to varying degrees by KDE (and vice versa), but control would probably not be a great thing.
while a close cooperation exists, the independence of both Qt and KDE keeps each side “honest”. there’s also a lot more to Qt usage than KDE as well, ranging from the well known examples (Scribus, Skype, Google Earth, ..) to thousands of ISV apps. it’s a pretty diverse ecosystem, and we all tend to get along and collaborate pretty well within it.
The LGPL goes against FSF principles, the only reason for it’s existence is in the case that a particular set of libraries or framework cannot offer enough capability or functionality over existing proprietary options, it’s an acceptable compromise. Qt doesn’t seem to have that issue, and has managed to provide GPL *or* commercial licensing for some years now.
GTK, on the other hand, requires LGPL, because people would likely not pay for it. That’s not to say that it’s not a capable toolkit, just that it doesn’t offer enough over alternatives to justify paying for it. Things like roadmaps and centralized support do mean something for many developers.
So why would Nokia LGPL it and turn it over to Gnome? Makes no sense.
[/q]
Gnome does not own GTK, *technically* or not. GTK was a graphics toolkit developed for GIMP, and it’s now bundled with a bunch of other separately maintained packages and libraries as a development environment. No single organization *owns* GTK, and that is part of the problem.
Several months prior to Nokia acquiring Tt, they lambasted the fact that GTK had no direction. The Gnome developers are arguing over what direction GTK should take, with the upcoming version. The recent discussions about Gnome 3.0, and more relevantly GTK 3.0, has led to many posts from developers amounting to “about time” and “are we going to fix xxx?” It’s one thing to say that the community “owns” GTK and controls it’s destiny, it’s an entirely different thing to see that put into play. Even the KDE devs realized through the challenges of KDE 4.0 that developers tend to focus on features, and nobody wants to work on the un-sexy low-level library work.
At the very least KDE has Tt taking care of much of the grunt work that OSS developers tend to shun. That’s the kind of thing that goes into spit and polish, as well as stability.
I’m not trying to diminish GTK, but please don’t dismiss Qt by saying “but GTK is free and open and owned by the community!”. The community likes to talk about GTK, but few of them are actually investing the time and energy to drive it forward. The emphasis seems to be “let’s just work with what we’ve got”.
As I said above, Qt doesn’t need the LGPL, because it offers enough that developers will pay the commercial license, or forgo that to utilize the GPL (or any of the other free licenses it is available under). The cross-platform capabilities alone, which span from platforms like *nix, Win and OSX down to portable devices, is staggering, and not something to be easily dismissed because it’s not “free”.
IBM, RH and Sun are interested in GTK as a framework for their various flavors of what Java should be. Novell wants GTK as a framework for mono. The community is going to have to step up if they want GTK to move forward from what it is already, or cede control to the the corporates they think they’re protecting themselves against.
Otherwise, Gnome will never really control their own destiny, and that would be a shame.
Just my 2c…
I thought that GTK key developers now all work for Imendio.
That would be incorrect. GTK has multiple vendors involved. Mathias Clasen who is the primary upstream maintainer works for Red Hat. The list of major developers is given at
http://www.gtk.org/development.html
Tim Janik Imendio AB
Matthias Clasen Red Hat
Behdad Esfahbod Red Hat
Federico Mena Quintero Novell
Alexander Larsson Red Hat
Tor Lillqvist Novell
Richard Hult Imendio AB
Kristian Rietveld Imendio AB
Michael Natterer Imendio AB
Emmanuele Bassi Opened Hand
Qt is licensed under GPL v3.
If Nokia decided to take Qt in a direction that was unsuitable for KDE, KDE could just fork the code and call it something else (say, for argument’s sake … cutie) and carry on.
It strikes me as a rather confused interview. For example, he mentions that OS X is growing well but omits that, surely, this growth is driven by Apple’s innovations in hardware (iPod, iPhone and “Airbook”), with which Linux cannot compete, rather than in software. Then there’s the problem with releasing an LTS version of Ubuntu which contains new and still flakey stuff like Pulse Audio as well as, by his own admission, a fairly high bug-count. That’s not going to attract the stability crowd. And there’s the stuff about KDE and Gnome. Hmmn, but much as he might like to think so, neither is driven by Ubuntu and neither would be fundamentally changed in any way if Ubuntu suddenly disappeared.
My impression is that Ubuntu is trying to do far, far too much given their resources. Trying to be all things to everyone – the interview even gets in nods towards Sun and towards acting as some kind of industry-wide release coordinator – just doesn’t work. There are too many bases to cover well so the risk is you cover them all less well than more modestly focused competitors who concentrate only on a few things. It’s hard to see Red Hat losing much sleep. Quite the reverse. By giving Linux a higher profile, Ubuntu has probably given Red Hat new customers.
Perhaps some retrenchment wouldn’t come amiss – less talk, more action, more focus on some core competencies. I recently installed an XFCE-based SuSE 11.0. It’s much better than Xubuntu, imho, but you wouldn’t think so from all the noise from the *buntu camp.
Ubuntu has done marvellous things. It’s been good news, but this interview suggests to me that if they don’t calm down a little they’ll end up running exhausted into the sands.
Edited 2008-07-16 23:22 UTC
Kubuntu and Xubuntu really are sub-par compared to other distros that focus on either of those desktops. Mint’s KDE edition, for example, provides a much better KDE experience on top of what is essentially Ubuntu, and the same can be said of Mint’s XFCE edition (speaking of an old version, current release doesn’t have an XFCE edition yet – current KDE edition will be out soon). Even when I tried OpenSUSE 10.1, as bug-laden as it was, KDE was used so much better than Kubuntu used it.
Ubuntu is primarily a Gnome distro, whatever the community likes to think, and KDE and XFCE just don’t get the same level of attention.
Could you expand on that? Last time I tried it, Mint’s KDE was beta and essentially a Kubuntu rebuild as far as I could tell.
Mint KDE is more or less the same as Mint Gnome, though they try and use existing KDE tools and applications where possible instead of just dropping KDE on top of the Gnome version. Although it’s slow work at the moment, the Mint Tools, like their control panel, updater and so forth, are being ported to Qt. I can’t put my finger on exactly why, but Mint KDE seemed a whole lot more complete than Kubuntu; and not in the same way Mint Gnome feels different to Ubuntu.
That man does spin a yarn or two in there… Others benifiting from Ubuntu’s kernel work? even though the view of the kernel developers is that Ubuntu does not contribute?
It may be a good distribution, but I think he has only decided to cry Upstream! Upstream! Upstream! after coming unstuck with higher than expected maintenance burden of 8.04 LTS (where other distro’s have chosen different core component versions).
Don’t believe everything you read, see or hear on the intertubes.
GKH was wrong:
http://blog.phunnypharm.org/2008/07/canonical-and-linux-kernel.html
Quote from that blog posting:
That’s the problem. Cannonical is way too understaffed for the standards they claim for themselfes.
According to Wikipedia, Novell employs over 500 people to work on SUSE Linux. Every SUSE Enterprise release has at least (!) seven years of support http://support.novell.com/lifecycle/ — same with Red Hat: http://www.press.redhat.com/2008/05/21/the-red-hat-enterprise-linux…
Cannonical only guaranies 5 years on the server and 3 years for desktop components on LTS releases. http://www.ubuntu.com/products/whatisubuntu/serveredition/benefits/…
Those 3 years of desktop support is an especially short time when you consider that every free openSUSE release is supported for two years.
If Cannonical doesn’t have the man power to support their own product, Cannonical can’t make significant contributions to OSS projects — be it the Linux kernel (<10 people), KDE (1 or 2 people), or any other OSS project.
Shuttleworth, just replace GNOME with KDE 4.1+ in Ubuntu.
And/Or give users choice, but make KDE 4.1+ the default DE.
Edited 2008-07-17 00:50 UTC
I’m sorry to disagree with you, but part of the reason Ubuntu has taken off as much as it has is because it uses Gnome. If you had seen it from the beginning, you’d have seen that part of their original mission statement was to take Debian Sid + the newest Gnome and create a stable desktop with the latest software.
Debian Sid itself usually has the very latest KDE (though notably the 4.x version has been in experimental as of now). And sadly, in a lot of ways, Debian Sid has been more stable than Ubuntu releases have been (with the exception of Dapper, which has been pretty solid.)
The reason that Gnome was chosen for Ubuntu is that it has a 6 month release cycle.
As much as I wish it was otherwise, the simple truth is that KDE 4.1 won’t be ready for general consumption. GNOME 2 is stable and largely ‘done’. It gives a good base to build an easy to use general purpose desktop. KDE4 simply isn’t there yet. Maybe it will be with 4.2 or 4.3, but 4.1 won’t be and deciding to use KDE4 before KDE4 is ready to use is not a good idea.
As much as I wish it was otherwise, the simple truth is that KDE 4.1 won’t be ready for general consumption. GNOME 2 is stable and largely ‘done’. It gives a good base to build an easy to use general purpose desktop. KDE4 simply isn’t there yet. Maybe it will be with 4.2 or 4.3, but 4.1 won’t be and deciding to use KDE4 before KDE4 is ready to use is not a good idea.
Ah… you say that KDE 4.1 is not ready yet, it certainly wont be “fully ready” but it’s getting there, and at least KDE4 is in a better position than GNOME in many ways and it’s going to be very stable and usable real soon.
I find it a much better idea to go with KDE 4.1+ rather than rewrite GNOME to use Qt, imagine how much time will it take to rewrite/port GNOME to Qt and make it stable. KDE4 certainly is already there, it only needs to polish a bit more, stabilize, get Amarok 2, and it’s already there.
I certain agree that KDE4 has the potential for greatness and the potential to be a better base than GNOME to build a great Open Source Desktop. However as it stands KDE4 has far more potential than substance going for it, and I say that as a fan of KDE4. Switching to KDE as their main desktop today would in my opinion be a bad move for Ubuntu. Once KDE4 has become more stable and more of the potential has been realized, then it’s a discussion worth having. Even as a fan of KDE4 I think Ubuntu is better of sticking to GNOME until at least KDE 4.2
No one is suggesting one re-writes GNOME using Qt. Everybody agrees it would be a waste of time. All Mr. Shuttleworth said was the he thought it would be technically possible, not that it would be practical or even a good idea.
And here I disagree, how can something lacking, by your own admission, both polish and stability be already there? KDE is heading in the right direction and is well on its way, but it’s not “already there”.
I have got KDE 4.1 RC installed under Kubuntu64 and I am running it right now as I post this.
I am having a good deal of success with it, it seems pretty stable, and it is as fast as KDE 3.5.9 on this same machine. I have got essentially equivalent functionality under KDE 4.1 as I have under KDE 3.5.9 (installed on the same machine at the same time).
If the few bugs that remain in KDE 4.1 RC can be shaken out by the time of release, I’m not at all sure that it is necessary to wait.
The crop of mainstream distribution versions that will come out in the latter half of the year will all be able to offer KDE 4.1 as a stable, useful desktop choice. KOffice 2 and Amarok 2 should also be available with it by then.
http://dot.kde.org/1149518002/
http://amarok.kde.org/
http://dot.kde.org/1173761811/
KDE 4.1 will be in pretty good shape even just a little later this year. Give it two months.
Speaking about security, are you going to stay with AppArmor now that Novell has dropped the ball on it or are you going to change over to SELinux?
The above question from the article was a surprise, I missed the news that Novell is no longer promoting AppArmor. Can anyone provide a bit more context or a link to what’s going on? Are Novell and openSUSE moving away from AppArmor?
Here ya go… http://news.cnet.com/8301-13580_3-9796140-39.html?part=rss&subj=new…
Novell has layed off the entire team that developed Apparmor in one go but hasn’t made any moves towards a supported and actively developed solution like SELinux unlike they did when they abandoned reiserfs and moved to ext3 recently. So it is quite puzzling what their strategy is really.
This came right after Ubuntu said they are going to adopt it instead of SELinux which leaves them in a bad state. The only way out for either of them is to invest on technologies, atleast the major ones by hiring developers to develop and support it and not abandon them instead. Ubuntu later also added partial support for SELinux and they have now two incomplete solutions which apparently leads shuttleworth to claim that Ubuntu has better security track record than Red Hat. The entire thing makes no sense to me.
.. because he’s using KDE these days
Life isn’t always nice to Gnome-users…. 
Just kidding.
It is good that Shuttleworth is humble about Linux and Ubuntu. However, although price and “freedom” (what ever “freedom” may mean in this context?) are important, there are many other reasons to want to be on the Linux platform too.
Especially top class security – no serious virus threats, for example – is a major benefit of running Linux (or (BSD) too.
Someone needing an easy to manage Internet PC to just send email, browse the web, do online shoppings, and pay the bills could love running an easy-to-use and secure Linux distro like Ubuntu (especially if she had a Linux geek in the family to manage the more difficult tasks for her..
However, we should realize that non-experienced Windows users do need the help of their family geeks all the time too in managing the more difficult PC tasks…)
Also, Linux suits the world’s biggest super computers as well as tiny mobile devices. Maybe that kind of flexibility could be counted as a part of “freedom” and open source though? (I like freedom – but hate the vagueness of the word “freedom” as a concept, however, as it can mean almost anything – depending on the point of view – like total opposites in the US politics where the word freedom has been very popular…)
Other reasons (related to freedom again, however): platform independence, lots of very different distributions to suit different purposes, transparency and openness of development, no vendor-lock in, continuity (no big danger of Linux going away), high degree of compatibility with other operating systems (file systems etc.) etc.
Edited 2008-07-17 09:01 UTC
I’m pretty confused by some of the conclusions he draws. Good interview, but to my eyes, he has a strangely skewed world view in some areas.