03 August, 2010

KDE strategy for openSUSE

As I mentioned in two earlier blogs now, within openSUSE a strategic discussion is going on - what direction should we, as a distribution community, take?

I would like to address a few things in this post. First of all, why a strategy, and what will it and won't it do? Second, there is one strategy I'd like to mention specifically, as I think it's disrupting but as a community proposal it deserves to be discussed as any other strategy. That's about the KDE strategy for openSUSE.


But first about the idea of a strategy in the first place. The strategy portal page talks about it plenty and I won't repeat that. I only want to stress what a strategy does and doesn't do.

It does:

  • help make project wide decisions; for example say we choose the home for developers proposal and the liveCD is full. Do we remove the second media player or the second debug tool? The first it is...

  • help focus marketing; say the marketing/promo team wants to set up a campaign targeting a group of ppl, who should they choose? A clear focus helps a lot.

  • help attract contributors; having a clear story and purpose helps attract contributors. It will also attract a specific kind of contributors, to be precise the ones to whom the strategy appeals and who are thus likely to implement it.

  • help in making a choice; if you're a contributor or user you have certain needs. Looking at the strategy and marketing can help you make a decision for one or another distribution!

Note that all these serve to make the impact of the strategy bigger over time - people who like it will start using it, voice their opinion, get involved, steer...

It does not:

  • mean we become less open; so if you want to focus on creating a pro-audio spin while the community has chosen for the focus on developers - go ahead. Nothing will change in that department: who codes, decides, and we're an open community.

  • mean we will actively remove things which don't fit the strategy; so if we focus on being being a mobile and cloud distribution, we won't remove OpenOffice for Google docs! We might put a Google Docs button in the menu, next to OO.o, or we might put resources in google docs import/export instead of MS Office 12 support.

So a strategy gives focus and direction, but does not limit much - except when it comes to either-or questions where it gives direction. A strategy is also broad - it has influence on pretty much everything you do. Example: server technology. The cloud proposal has influence on openSUSE as a server - we would integrate and ship things like OwnCloud, Etherpad and similar server technologies in an early stage. As a base for deriviatives we would make sure setting up a server can be done easily from the SUSE Studio GUI. And when we aim for developers, the build service should be integrated so developers can write their application and build it for over 10 distributions with one click.
And third, besides giving focus and being broadly applicable, a strategy also unites. It gives everyone in the community a common goal, lets us focus on our strengths, and binds us.

the KDE proposal

Looking at those three goals of a strategy you might understand a bit better why the number 1 KDE strategy ain't the best of the proposals. While you all know me as a KDE guy, I can't really support this idea. Talking to the strategy team and the dude who submitted it initially (Marcus), it has been improved a bit. It used to just talk about KDE - as in, let's be Ubuntu for KDE. That is focusing on a solution, not stating the goal or the problem you try to solve. Now it proposes to have a strong end-user focus, making it a bit more inclusive. You can then choose the right technology for the right job.


Still, while the strategy focuses on a traditional strength of openSUSE (a great integration of KDE apps and a good Plasma Desktop setup), it does not bind but it segregates.that is another traditional strength of openSUSE: being a broad, all encompassing community. This strategy is not broad at all (it is still only about one desktop technology) and does thus not give direction for a large part of the openSUSE community. Moreover, it's too specific and technical to attract most 'common' users. They aren't interested in technology but in the result.


I think it might be good for KDE and in the long run might work for openSUSE. Even though focusing on technology instead of the goal (end users) Not so sure about Free Software in general however. We, as in the Free Software desktop community, were just starting to build real bridges between each other - next year will have a Desktop Summit again!

Perhaps more important, this proposal would chase away an important part of our community - the many non-KDE users and contributors. And the costs could be serious and in many area's. KDE and Gnome technologies can help each other. A good example of that, something I've been lately involved in (yet I needed Bryan to remind me of it) would be a11y or accessibility. This is something which has been moving within the KDE community lately, in part due to some inquiries a government organisation did at last Linuxtag. However, there currently are very few good tools like screen readers written on KDE technology, to eg the Orca screenreader has to be used. Which is fine - and something openSUSE has an edge in as we ship both good KDE and Gnome libs and apps!

What it should be

I think it would be good to have a proposal focusing on end user products, on something aunt Tilly can work with. openSUSE could be a distribution aiming for polish, the final touch. Working on creating a great end user product. And both the Gnome and KDE people would be able to work with it, as would the Apache team, the Kernel team and all others in the community!

my offer

I'd be willing to write such a proposal (yes, short notice, I know) if ppl think we should have it. I'm NOT saying here that that's the direction we, as in openSUSE, should choose - personally I like the poweruser proposal as well as the developer proposal. Oh and the cloudy one as well... Besides, I've been involved only so short, my vote doesn't count as I'm not even an openSUSE Member right now. So the openSUSE community should vote - not me. I'm just here to help!


  1. I like the power user proposal. It plays to the strengths of OpenSUSE and what attracted me to OpenSUSE to begin with. Power User doesn't mean hard to use, it actually means the goal should be to make it easy to do 'advanced' things with your computer. (SpiderOak is a good example of this in practice.)

    Actually its not much different from the "status quo" proposal really.

  2. Agreed, I like that one too. At some point we need to consolidate the proposals, and the 'poweruser' and 'home for developer' strategies can probably be merged. They would constitute a strategy which strongly plays on traditional openSUSE strengths while at the same time bringing the needed focus.

  3. I think the "be the Ubuntu for KDE" strategy would actually work really well. The way I see it every time work is done on openSUSE it has to either be done twice or written in GTK or Qt (for seamless integration). This doubles the workload. Sticking to one tool kit and one desktop environment would mean all work contributed would be used in one place. You talk about making openSUSE aunt Tilly proof but Ubuntu has already done this with Gnome (or is slowly getting there). Why do something that's already being done? I think with a bit of similar polish KDE is far more attractive and is based on much better technologies than Gnome, like Qt for instance (just my opinion). Obviously a lot of Gnome developers would be lost but just do what Ubuntu does and steal their work from upstream :P. Still having Gnome as a supported (almost second class) desktop environment would may be stop them from leaving as they would still be able to make their software for Gnome? Making a distro Aunt Tilly proof doesn't mean you can't install all the IDEs you want either, so it doesn't conflict with the home for developers proposal either.

    I really like the home for developers proposal as more people writing software on openSUSE can obviously only be good.

  4. Evey time the KDE fanboys bring the same argument; focus on KDE, and openSUSE will be more successful - well, I'll let you in on a secret: IT DIDN'T WORK FOR MANDRIVA!

    KDE as a focus does NOTHING to make a distribution more popular; no matter how good it is. Now, for 11.3, the KDE-guys brute-forced their approach and made KDE default desktop for SUSE and promised us all kind of great things would happen, like more contributers for openSUSE, more popularity - guess what? IT DIDN'T WORK!

    openSUSE with KDE as default desktop environment is still less sucessful now (for example on distrowatch) than the early GNOME.-centric releases immediately after SLED came out were.

    KDE is NOT better, is NOT more polished, is NOT more functional than GNOME.

    Forget it KDE-fanboys, the larger audience prefers GNOME, and THAT is why Ubuntu is so popular; because they focus on the desktop that users want.

  5. Very much like the description of what a strategy is/is not.

    As for the existing proposals, the PowerUser proposal sounds really good. It does describe what brought me into the community all those years ago. I'd love to see your fleshed-out proposal, but it is rather late.

  6. "IT DIDN'T WORK FOR MANDRIVA!" The last I heard Mandriva still makes KDE4 look and feel just like KDE3 and comes with loads of GTK apps be deafult so its hardly a good example of what a KDE4 distro could do.

    "KDE is NOT better, is NOT more polished, is NOT more functional than GNOME." Which is what the stratergy is. Make it more polished. And what can Gnome do that KDE couldn't without a little bit of work?

    "KDE as a focus does NOTHING to make a distribution more popular"

    Prove it?

    I thought that all openSUSE did was make KDE the default selection on the DVD. Nothing more. Although I am probably wrong.

  7. To be brutally honest, I don't entirely agree with what connel says. I love using OpenSuSE. It's been great to me, and YAST is really impressive (imho). However, I prefer Gnome over KDE. I don't really find much appeal with KDE- and it would be a grand shame if the community focus was shifted to KDE exclusively. It would mean I would have to use another distro and not appreciate the great portions of SuSE I thoroughly enjoy.

    I say nay. Let there be both.

    Per Qt: I think GTK is still far more mature than Qt. I've fiddled around with KDE from several versions, including 3.5, 4.1 and latest. I still feel there's a lot of bloat, and there's been a few times where applications have crashed on me suddenly (as rare as that may be to that of other OS', it was a more prominent occurence in KDE than it ever was/has been in Gnome).


    I think the 'Power user' label should also encompass gamers somewhat. I've been using OpenSuSE since 11.1 and for a plethora of reasons. Natural every day use is fine, but I'm a gamer at heart. I've been really impressed with the traction that Wine has been making, and currently play all my games through it (although I'm waiting for 2.6.35 kernel to be released for 11.3 since it apparently has a patch fixing WoW crash at login).

    Sorry for rambling. Ultimately my point is that there are certain little caveats with OpenSuSE that only showed up in 11.3. Some games do not run although they do on nearly every other distribution (Ubuntu, Fedora, Mint, etc). Fallout 3 is an example. OpenSuSE 11.2 worked like a charm for it. The howto on the website was clean and clear and it installed perfectly (I will also point out that this is definitely not a wine bug, it seems to be distro-centric).

    Would be great if a little polish was thrown offhandedly in that department. People like me would giggle like higschool girls at the thought alone.

  8. As a user of the KDE Desktop and Live KDE cd's. This is what I do to make the OpenSUSE releases work as a Desktop for me.

    Change the default kernel to the desktop kernel.
    Add the Pacman and VCL repositories.
    Add bogofilter and clamav
    Remove all related virtualization software
    Use VLC as the default media player and Phonon back-end
    Replace OpenOffice with KOffice (as of ver 2.0+)
    Add proprietary Opera Browser and remove Firefox (My Preference)
    Delete all games (My Preference)
    Oh, and put the Trash Icon on the Desktop.

    It takes up 2.7 Gigs of my 11 Gig hard drive (Shared with XP on a 9 Gig partition.

  9. As a lizard user since SuSE 7.0 (2000) I heard these remarks about (Open)SuSE the most over the years:
    - best KDE-desktop
    - most and best polished desktop
    - best SOHO distro (server & desktop)
    - very good development platform
    - good for power users & novices

    Focus on these points. OpenSUSE doesn't have to be ranked at n° 1 at polls or Distrowatch.
    Keep focus and it will always be a top distrobution.
    - the longetivity of support for the releases should go up to at least 3 years will OpenSUSE remain the best SOHO distrobution or others will take it's spot!
    - the repositories should be drastically simplified

  10. @Henry:
    Please, do everyone a favour and read Mark Shuttleworth's post about tribalism. Also read Aaron Seigo's post about the same subject.
    Hopefully you'll stop ranting, stop behaving like an animal and start being more civil. I also hope you'll come with a more useful discussion, with better arguments to prove your point, and if possible with charts, numbers and/or studies to back them up.
    Greetings, Ricardo

  11. This comment has been removed by the author.

  12. What will be done about it?:

    I see that there is a strategic discussion of openSUSE (http://news.opensuse.org/2010/07/21/timeline-for-opensuses-strategy-discussion/), I'll be very happy if they (or you, or even the community, because it will start talks between developers and community) in this discussion, decide that openSUSE, now has more autonomy from Novell, and to take major decision as the creation of the Foundation openSUSE (http://www.osnews.com/story/23474/openSUSE_Linux_Seeks_More_Autonomy_From_Novell), and that the foundation now has more autonomy from Novell, and decides to return or replace the SaX2 (https://features.opensuse.org/308357) because the removal of SaX2 (http://en.opensuse.org/Archive:SaX2) of openSUSE, was a decision made by Novell, who preferred not to invest in the development of SaX2, for business issues (http://lists.opensuse.org/opensuse-factory/2009-12/msg00017.html), for this and other reasons, I am in favor of creation, of Foundation openSUSE, and also decide that the openSUSE, switching to versions LTS (Long Term Support), and that support will last for more of three years, among other points that must be discussed. I just want a better openSUSE.

  13. I think the KDE proposal actually helps building bridges, because Novell itself focuses on GTK-related technologies for its business: GNOME is default in SLED and KDE Desktop is quite hidden compared to oS.
    As for netbooks Novell seems also to be interested in MeeGo whose Netbook UX is based on Clutter and GTK.
    On top of that Xfce and LXDE are supported on oS as well.

    The KDE proposal levels the field a bit.
    And with Novell's commercial interest in GNOME and MeeGo Netbook it's pretty much ensured that GNOME users have absolutely no reason to fear oS.

    Maybe the KDE proposal could lead to some positive tweaks to SUSE's GNOME as well, eg. shipping some Qt-only applications like Marble-Qt together with GNOME.

  14. I think the discussion on strategy is very important (I pushed for something like it long ago myself). However I would like to hear plans, not general descriptions even at this stage.

    Someone might think it is premature, but a strategy without a detailed analysis of the available resources does not make any sense, and discussing of the resources after the strategy has been decided makes even less sense to me.

    In other words:

    - Novell should clarify explicitly what will be its level of involvement in the strategy, what component of the distribution it will support, and for how long (I know there is a support period of 18 months, I do not refer to that, but to how long Novell right now wants to commit to maintaining say the components "X" and "Y" in the next releases).

    - Then, once this is clarified who in the community makes the commitment of maintaining the remaining parts to allow the strategy to be successful?

    You wrote that what is not part of the strategy won't be dropped. This is another unclear point. Who will take care of it?

    I feel that without a clear answer and clear ideas on these points, which essentially require a quite accurate estimate of the available resources we are just trying to decide what we hope openSUSE will be, and not a real strategy.

  15. "As a lizard user since SuSE 7.0 (2000) I heard these remarks about (Open)SuSE the most over the years:
    - best KDE-desktop
    - most and best polished desktop
    - best SOHO distro (server & desktop)
    - very good development platform
    - good for power users & novices

    Focus on these points. OpenSUSE doesn't have to be ranked at n° 1 at polls or Distrowatch.
    Keep focus and it will always be a top distrobution.
    - the longetivity of support for the releases should go up to at least 3 years will OpenSUSE remain the best SOHO distrobution or others will take it's spot!
    - the repositories should be drastically simplified"
    snip ----------------

    Makes sense to me.

  16. AlbertoP wrote:
    "...a strategy without a detailed analysis of the available resources does not make any sense..."

    Very good point, a lot of the comments on the strategy discussion seem to assume that Novell employees will automatically shift their focus according to whatever strategy is chosen.

    But the truth is that there's zero commitment from Novell to this strategy - except probably the openSUSE Boosters team of roughly a dozen people. The rest of Novell employees will just go about their business as usual, and change nothing.

    Volunteer contributors are even harder to manage - they each have their own agenda too.

    Therefore the strategy must fit well with what Novell developers are doing for SLE, and it must have the highest possible level of support from volunteer contributors. Or the strategy discussion will at best have been a waste of time.

    I hope the election process will take this into account. So we don't end up with a strategy that was elected by say 30% of total votes. If no proposal gets more than 50% of the votes in the first round, we need a second round or third even. Until one proposal has >50-60% of the votes.

  17. IMHO openSUSE should strife to deliver the best desktop experience, full stop. You have to compete with Ubuntu, there is no getting 'round it.

    My advice is to deliver GNOME 3 builds early.

    GNOME 3 will very probably be the only really new and exciting thing happening in the next two years; KDE 4 has come and failed to excite. GNOME 3 sets new standards for user-friendliness, and those distributions who adapt late will be left behind.

    So, if we focus this discussion too much on KDE, we will miss entirely miss out on that one. And I have little doubt that GNOME will steal followers from KDE massively, once GNOME 3 is out in March, as well as spotlight.

    Furthermore, we should discuss whether to support the aim announced at GUADEC, to make 'GNOME OS' a more unified experience across distributions, like Android OS.

    So, I think a focus on KDE is short-sighted and besides the point.

  18. I would like openSUSE to be the "Ubuntu for KDE". Ubuntu, as we all know is focussed on GNOME as its customised desktop. With each openSUSE release, I hope to see similar customisation on the KDE desktop.

    KDE SC has provided tons of ideas and technologies but I feel these need to be streamlined (such as activities) and present them to a user in a way which is most intuitive. openSUSE, IMO, is best placed to do this.

  19. KDE is designed as a very 'poweruser-style' desktop environment. Thus you will never be able to turn it into an Ubuntu-GNOME-mock-up.

    Or you might as well simply use the real GNOME instead.

  20. While the 'power user' proposal does appeal to me, there is a real need for 'Aunt Tilly' spins too. Mandriva's current problems make it less appealing than it used to be, and *buntu doesn't appeal to everyone. The solid base of OpenSUSE would be a really good place for the 'I only want to work' people, while the rest of us have the full range to go for.

  21. I agree that KDE should not be the defining strategy for OpenSUSE. I liked KDE a lot in the past and it was my main desktop for 8 years or so, but it is not what makes me stay. I use Gnome nowadays since either I have outgrown KDE, or KDE has found another audience.

    The strengths I see OpenSuse have are increasingly in the package management area. Zypper is the best package management tool around, and the build service and hosted repositories are great. There is much potential for growth there, work to bring even more software easily accessible to the user. Perhaps something like an "app store" can be developed on this infrastructure. It may sound as going against free software ideals, but the convenience of purchasing software as an integrated experience could work also for free software. The Click'N'Run effort from Linspire seems to have failed, but OpenSuse is much better positioned for something like this.

    The Suse Gallery and the Suse Studio also builds on these strengths. The next step as I see it is integration with the cloud. For example, make it possible to build Amazon EC2 images, or provide a custom hosts for virtualization, could work well as a revenue stream too.

  22. > Working on creating a great end user product.
    > And both the Gnome and KDE people would be able to work with it,
    > as would the Apache team, the Kernel team and all others in the community!

    +1 for involving more people, unifying things, fostering cooperation and becoming central role in there. :)

    Even getting the kernel people to cooperate with the desktop people, or Apache people is really awesome. This is something that could move us all forward.

  23. >My advice is to deliver GNOME 3 builds early.


  24. @Henry
    > Forget it KDE-fanboys, the larger audience prefers GNOME

    Is that so?
    What about the many Ubuntu users who have installed the kubuntu-desktop package, therefore turning it into KUbuntu - did you count those?

    > and THAT is why Ubuntu is so popular; because they focus
    > on the desktop that users want.

    I think you got your causality a little mixed up.
    If I had to speculate, I'd guess the number of people who only use Gnome because of Ubuntu, is MUCH higher than the number of people who only use Ubuntu because of Gnome.

  25. Ubuntu is a Debian GNOME desktop, simplified to the extreme.That's certainly as far away from KDE as you can get.

  26. guys, please don't feed the troll. The GNOME-vs-KDE thing is old. I don't like to sensor so I won't be removing the comment just now but if everyone starts feeding him/her with responses I have to...

    Besides that, thanks for the comments.

    With regard to the position of Novell: Novell is a community member. Like any community member, Novell mostly works on what Novell cares about AND invests a bit in the betterment of openSUSE in general. The first might or might not align with the strategy, the second I will push for to be aligned with it. But I don't call all the shots.

    So imho it is NOT needed for this strategy to be completely aligned with Novell. The community has to push it forward - and as I wrote in what the strategy will and won't do, the choice will have its effect there.

  27. I am very impressed with your involvement in openSUSE Jos. I think that you have very good ideas about the direction the SUSE distro should take. However, I believe it is of great importance that, not only you, but all the directives involved in openSUSE get together and perfect your plans. Although they are good, I think that they lack detail. I agree that there is not a lot to say about something that you can't divine, such as SUSE's success in the market. But details are important. Through the emails that you have sent, and the other ones sent by the community leaders, they seem to talk mostly about what to do with the details.
    Set up a roadmap. Make clear goals, names, places, numbers. They all help. Right now the "#1 KDE Distro" is a good ideal, but unclear.
    How is KDE the best in openSUSE? What is lacking right now? What will be the steps to make KDE the best in openSUSE?

    if you want, you can check my blog where I try to get to design details in KDE to customize SUSE and then make openSUSE unique and outstanding.

    Thank you for your work.


  28. Andy,

    Thank you for your kind words. I think you are very much right - detail is still lacking in my ideas. But please note I want to be an enabler - not set directions but help the community to do so.

    Moreover, I'm still very much new and simply lack the knowledge to go into details.

    Your blog is impressive, you have very interesting design ideas. I would like to urge you to either get involved with the KDE design team (Nuno & friends). Another option (in case the KDE artists have different ideas or directions) you could consider working with some openSUSE peeps to implement your vision for openSUSE. There is no reason why oS cannot have it's own distinct visual identity for KDE - we can have our own (oxygen based or not) theme as default!


  30. Dear Jos,

    the main indent of the proposal was to give oS a clear focus, a direction and something to differ from other mainstream distributions. As oS is historically known as a great KDE distribution, why not focus on these strength.

    GNOME or other DE's won't loose that much (if even anything). Novell (as the main sponsor) will still support GNOME and a spin could be made available. (Please note that I am personally a GNOME user, but I still think this direction would be the best for the distribution in general).

    So why don't we just let the community decide which way to go?


  31. I´m using Kubuntu since 7.10 and I´m always waiting for the "next" release to be more stable and mature, but such a thing never happens, Ubuntu is a Gnome oriented distro, period, Kubuntu is an alternative project without real Canonical support.
    I heard some comments about OpenSuse´s KDE desktop and I´m gonna try it as soon as I can, I have to say goodbye to apt but I hope zypper will fix that void.

  32. Hi dhunter,

    You can count on zypper being very nice to you. At least that has been my experience - it's a pretty good package tool. Fast, too, and light on resources as it uses delta packages on upgrading instead of downloading the whole bajeera again.

    Yast ain't the most sexy or easy to use tool to manage software graphically, but it is fairly powerful. And smart - offering solutions if you want something difficult instead of doing something stupid or telling you it can't be done like other tools do. Zypper does the same on the commandline, btw.

    And yes, both KDE *and* GNOME are good in openSUSE, quite a feature ;-)

    It also means you can more easily mix and match things between those - eg I currently use Banshee as music player as I really like it.


Say something smart and be polite please!