The Art of Release

Monday, May 12th, 2008

An update on the long term plans for Ubuntu release management. 8.04 LTS represented a very significant step forward in our release management thinking. To the best of my knowledge there has never been an “enterprise platform” release delivered exactly on schedule, to the day, in any proprietary or Linux OS. Not only did it prove that we could execute an LTS release in the standard 6-month timeframe, but it showed that we could commit to such an LTS the cycle beforehand. Kudos to the technical decision-makers, the release managers, and the whole community who aligned our efforts with that goal.

As a result, we can commit that the next LTS release of Ubuntu will be 10.04 LTS, in April 2010.

This represents one of the most extraordinary, and to me somewhat unexpected, benefits of free software to those who deploy it. Most people would assume that precise release management would depend on having total control of all the moving parts – and hence only be possible in a proprietary setting. Microsoft writes (almost) every line of code in Windows, so you would think they would be able to set, and hit, a precise target date for delivery. But in fact the reverse is true -  free software distributions or OSV’s can provide much better assurances with regard to delivery dates than proprietary OSV’s, because we can focus on the critical role of component selection, integration, testing, patch management and distribution rather than the pieces which upstream projects are better able to handle – core component feature development. This is in my mind a very compelling reason for distributions to focus on distribution – that’s the one thing they do which the upstreams don’t, so they need to invest heavily in that in order to serve as the most efficient conduit of upstream’s work.

We also committed, for the first time, to a regular set of point releases for 8.04 LTS. These will start three months after the LTS, and be repeated every six months until the next LTS is out. These point releases will include support for new hardware as well as rolling up all the updates published in that series to date. So a fresh install of a point release will work on newer hardware and will also not require a big download of additional updates.

Gerry Carr at Canonical put together this diagram which describes the release management plan very nicely:

Ubuntu Release Cycle

The Ubuntu team does an amazing job of ensuring that one can update from release to release, and from LTS release to LTS release directly, too. I’m very proud to be part of this community! With the addition of some capability to support newer hardware in LTS releases, I think we are doing our part in the free software community – helping to deliver the excellent work of thousands of other teams, from kernel.org to GNOME and KDE, safely to a huge audience.

There’s one thing that could convince me to change the date of the next Ubuntu LTS: the opportunity to collaborate with the other, large distributions on a coordinated major / minor release cycle. If two out of three of Red Hat (RHEL), Novell (SLES) and Debian are willing to agree in advance on a date to the nearest month, and thereby on a combination of kernel, compiler toolchain, GNOME/KDE, X and OpenOffice versions, and agree to a six-month and 2-3 year long term cycle, then I would happily realign Ubuntu’s short and long-term cycles around that. I think the benefits of this sort of alignment to users, upstreams and the distributions themselves would be enormous. I’ll write more about this idea in due course, for now let’s just call it my dream of true free software syncronicity.

166 comments:

  1. Arek says: (permalink)
    May 12th, 2008 at 11:57 am

    Sorry to write this.
    I use Ubuntu for a long time now, and I like it very much. For sure I will not change to Win or MacOS and probably not to other distribution of Linux.

    Unfortunately 8.04 was not a success. The number of little bugs is a bigger then for all previous releases all together (maybe I have bad luck but I see same problems in forum and in bug database for other people). This is specially irritating if it is called “LTS”.
    Take your time, don’t be fixed with dates, but with stability (yes, not even functionality).

    I know you guys are doing hard job. And it is also really great you support this project financially.
    Keep going. Don’t worry about time. We will be waiting for next version.

  2. Simeon Walker says: (permalink)
    May 12th, 2008 at 12:43 pm

    Nice. I’m glad there will be continuous point releases of 8.04. Doing this may make it easier for me to keep all my servers on the LTS release.

  3. PhilK says: (permalink)
    May 12th, 2008 at 12:51 pm

    That’s a beautiful chart, perfectly describes the idea of LTS. What program was it created in?

  4. BUGabundo says: (permalink)
    May 12th, 2008 at 12:58 pm

    i too believe that having major distros launching at around the same time would be great.
    But then again there is the other side: the technical and marketing: launching so many distros at the same time would put a really heavy load on mirror server; several distros lunched at the same time, would make news dimmed on some points of a distro, or to not announce one or more of them.

  5. Vadim P. says: (permalink)
    May 12th, 2008 at 1:19 pm

    Congratulations on a very excellent first, and the kept promises.

    And good luck on the syncronicity idea – it’s great, but people are being stubborn and whatever anti-ubuntu sentiments they have cloud their decision on this -very- great idea, that nobody but FOSS can pull off.

  6. Rodrigo S. Amaya says: (permalink)
    May 12th, 2008 at 2:10 pm

    That is exactly why ubuntu is loved by so many, you don’t have to wait 3 or for years for a new mayor release (like slackware), or wait to buy the latest desktop eye candy (like Vista), with ubuntu you can have it every release.
    Fantastic!, Thanks Mike and all the team to make such a fantastic contribution to the Open Source community.

  7. Tom says: (permalink)
    May 12th, 2008 at 2:19 pm

    Hello Mark,

    congrats on Hardy. It is really nice. I also really like the idea of synchronizing. I think that would have given Hardy that extra few bug fixes that are so important for a excellent first impression.
    But even if Red Hat and Debian say No ( they can be a little stuborn at times ) you should maybe think about releasing LTS releases a little later .. the early april release date really worked well in the past cause it adds a lot of exposure to Ubuntu .. but I think you have to admit now that a lot of people think “Ubuntu = Linux” meeting exspectations is the real issue cause people are starting to really want a “Just works in any case”-system, which shows how far you have come in a really short time.
    I think LTS releases should land X.06 like dapper .. might be tough for LTS+1, but it would give the LTS.06.0 that extra polish ( all the fixes from Fedora and OpenSuse + those from extra RCs which will probably end up in 8.04.1 now )
    Dont you think so?

    Cheers,

    - Tom -

  8. Zone of Mr. Frosti » Blog Archive says: (permalink)
    May 12th, 2008 at 2:27 pm

    [...] interesting quote from Mark Shuttleworth: Most people would assume that precise release management would depend on having total control of [...]

  9. Me says: (permalink)
    May 12th, 2008 at 2:28 pm

    @ Vadim P.

    > but people are being stubborn and whatever anti-ubuntu sentiments

    Please, come on.
    Possible release date of the next RHEL : march 2009
    Release date of the next Ubuntu LTS : april 2010

    Come on Ubuntu, set the next LTS to march/april 2009 !

  10. Eiriks forfatterblogg :: Dagens ironiske: Linux redder Windows XP says: (permalink)
    May 12th, 2008 at 2:38 pm

    [...] Siden Vista er for tungt til å brukes i dette markedet (noe Microsofts egne krav til maskinvaren bare bidrar til å sementere, latterlig nok) kommer XP til å være kommersielt tilgjengelig i en eller annen form til den lovede, modulære arvtakeren ankommer. For de av som etterhvert er kommet til å se Microsoft som det største hinderet for innovasjon i databransjen, er den gode nyheten at det som måtte eksistere av strategi på operativsystemområdet blir enda mer pulverisert enn de ørten variantene av Vista skulle tilsi. Kaotisk og reaktiv til det paniske, er vel ordene som faller meg i hu, når jeg setter det opp mot lanseringsstrategiene til f.eks. Apple og Ubuntu. [...]

  11. Kaminix says: (permalink)
    May 12th, 2008 at 3:04 pm

    Hope my wireless network card, which has worked two releases before, will stop being broken in the next point release then.
    Also, it would be nice if my computer stopped looking up whenever I use tab on a base install (though it works post-desktop packages), another easter egg which came with 8.04 LTS.
    Anyway, just wanted to comment on the ‘perfect release cycle’ of Hardy Heron. The new release schedule looks nice.

  12. Jari Saarelainen says: (permalink)
    May 12th, 2008 at 3:35 pm

    IMHO, I don’t think actual release should be carved on stone. If some of the goals remain unsolved to due date, I could compromise the release date. Be honest, if we fail, so be it. I would rather present a mature release that “just works” the presumtive ubuntists

  13. Saad Raza Abbasi says: (permalink)
    May 12th, 2008 at 3:38 pm

    Mr. Shuttleworth, your ambitions need to be applauded. I look forward to seeing the next releases of Ubuntu. I really like the chart above =) and was wondering if you could tell me which software was used to make this chart? Please.

    Mark Shuttleworth says: Looking at the image properties, I think it was made in Adobe ImageReady.

  14. Mark Shuttleworth: El Arte de las Versiones | Blog de Marcelo Ramos | Lo que hago, lo que me interesa, lo que me llama la atención says: (permalink)
    May 12th, 2008 at 3:49 pm

    [...] del post: Una actualización sobre los planes a largo plazo para la gestión de versiones de Ubuntu. 8.04 [...]

  15. Geert says: (permalink)
    May 12th, 2008 at 3:50 pm

    It’s a great release but this time the side-effect of releasing on time is having quite some bugs on the desktop front. (I personally encountered compiz bugs, a buggy FF3b5, bash bugs, installer bugs). Most of them fixed in hardy-proposed, except for FF3b5 ofcourse.

  16. Jadd says: (permalink)
    May 12th, 2008 at 3:52 pm

    Personnally, the “Hardy” promise was much more important to me than the date promise.

  17. Max Randor says: (permalink)
    May 12th, 2008 at 3:52 pm

    Ubuntu is my favourite operating system. but hardy did not really hit its release date – it was not ready for release at the release date but was released anyway.
    There are still important outstanding bugs that have not yet been fixed.
    When I upgraded to the beta – it felt like alpha. The final release felt like the beta release if the -17 kernal and other updates fix the outstanding bugs then maybe with them hardy will finaly be ready for release.
    Yes Ubuntu was released on time, but it was not ready to be released so that does not count.
    Still the best operating system in existence though. Reflects badly on Ubuntu for a LTS release to have critical bugs (IMHO) outstanding weeks after release.
    The point releases will help though – I can install 8.04.1 on windows users computers – until then I will have to wait.

    Sorry but it had to be said.
    (Ubuntu is great)

  18. bixbi » Blog Archive » Ciclo Ubuntu fino aprile 2011 says: (permalink)
    May 12th, 2008 at 4:09 pm

    [...] semestrali con aggiornamenti in termini si di sicurezza che software. Il suo fondatore e sponzor Mark Shuttleworth ha divulgato la scaletta , chi è abituato ad usare e scaricare ubuntu non è una grossa novità, [...]

  19. cosmix says: (permalink)
    May 12th, 2008 at 4:52 pm

    Ubuntu has been nothing short of revolutionary in the linux world, not in terms of the quality of the software or the packaging, but because of its appearance at a time when the linux world was rapidly expanding and its impeccable community infrastructure and wide-ranging support. Its focus on free, community-focused yet well-supported and functional software certainly created a reference point, spurred competition and opened the door for linux to the Small-Medium Enterprise world.

    Yet calling Ubuntu 8.04 an ‘enterprise platform’, with its half-baked, semi-functioning gvfs, its largely unintegrated and unsupported pulseaudio and a beta version of Firefox 3.0 as its default browser is in no way an accurate characterisation.

    Moreover, it’s somewhat ironic that the punctuality of the release is employed in this manner to showcase the merits of the project. Yes, the Ubuntu developers, the volunteers, the users, the community as a whole is commendable. But the fact that Ubuntu came out ‘on time’ means nothing when ‘on time’ translates to beta quality software at times inferior to the version it succeeded.

    It’s a shame 8.04 was tagged as an LTS release. While a great ‘normal’ release, it’s certainly quite far from the robust, polished and functionally conservative featureset that any ‘enterprise’ release is typically associated with. In fact it’s exactly the opposite; flamboyant, bleeding-edge and somewhat buggy. I’m certain that the decision was conscious and that Canonical is expecting 8.04.1 to remedy all that’s wrong with 8.04. Hopefully it will, yet I cannot but feel that 8.04 LTS should’ve been 8.06 LTS, or at least 8.04 sans-LTS. The fact that it isn’t, coupled with this post introduces a political layer to its development and one that could potentially affect its popularity where the ‘LTS’ label has any meaning.

  20. David says: (permalink)
    May 12th, 2008 at 5:10 pm

    I am curious…what tool was used to create that roadmap diagram?

    Mark Shuttleworth says: looking at the image properties, Adobe ImageReady.

  21. DaveB says: (permalink)
    May 12th, 2008 at 5:21 pm

    cosmix says:
    > flamboyant, bleeding-edge and somewhat buggy.

    I respectfully disagree. The stability of 8.04 during the beta cycle, and now that it has released, is very good and certainly on a par with both other Linux distributions, and with operating system releases from other large operating system vendors. Hardy exploits the hardware available in current systems really well, and has very up-to-date versions of software such as Firefox and Open Office. The alternative would have been for the LTS release to have drivers, software, etc which is 9-12 months old, and be stuck on that base for the next five years. I think Hardy has hit the mark with incorporating the right level of new technology in a way that is supportable and consumable.

    My expectation is that by the first point release will add additional stability, and if I was rolling Ubuntu out across 1000 systems in an enterprise, I would be
    doing my development/test work on Hardy as it exists today, and rolling out the first point release into production.

    Mark raises a good point that while the technology is cool, the release process is more impressive to those of us who have worked on large software projects. Its a subtle point that the general public will never recognize, and my hat is off to Canonical and the whole Ubuntu community. I hope they don’t get cocky, because it won’t get repeated without the same level of hard work that made it all come together this time.

    Dave

  22. WildSubnet says: (permalink)
    May 12th, 2008 at 5:45 pm

    I enabled proposed updates in Hardy and grabbed the new kernel rev. Already seeing stability improvements. They made change to the desktop scheduler in that and it shows in fewer mystery hangs.

    I’m thinking I’ll stick with Hardy for awhile. The point release schedule makes it more compelling to do so.

  23. Niklas Andersson, TechWorld Open Source says: (permalink)
    May 12th, 2008 at 5:47 pm

    Rhytm = Music that you can dance to

    I agree. A predictable release schedule adds so much more value than the “we deliver when it’s ready”… Ubuntu givs the open source field a needed rhythm and pace.

    When you know the bus leaves as scheduled others can plan around that. It seems to me other projects are already planning to get their stuff out in order to get it into Ubuntu.

    I’m hoping to see interesting stuff such as Multi Pointer X, OOo 3, Wine 1, Gnome 2.24, KDE 4.1 in Intrepid Ibex.

    A six month release schedule is ultimate.

    “Don’t stop the music”

    Niklas Andersson, TechWorld Open Source

  24. Stephane Marchesin says: (permalink)
    May 12th, 2008 at 5:56 pm

    As one of the X.Org guys, I have a question about the upstream release schedule you want to enforce. Talking about the case of X.Org specifically, we don’t have the resources for any kind of regular releases. So would Canonical be ready to pay a release manager if needed ?

    Mark Shuttleworth says:
    Yes, if X.org is willing to embrace some of the practices and tools that make it easier to achieve an always-releasable tree!

  25. dominiko says: (permalink)
    May 12th, 2008 at 6:06 pm

    First of all, congrats for Ubuntu. Knowing when to expect the next release of Ubuntu is nice. Having said that, we have to acknowledge that there is a trade-off between releasing it exactly on the due date and ironing out some showstopper bugs. In my opinion, there is no hard deadline and there is no shame in release it a couple of months later than the fixed 6 months release schedule if it helps to fix important bugs, especially for a LTS. My experience so far with Hardy indicates that it needed a bit more time. Still, I’m using Ubuntu to post this, so you must have been doing something right. Thanks!

  26. Meeting Upgrade Release Dates: Windows v. Linux « Legal Technology Blog says: (permalink)
    May 12th, 2008 at 6:16 pm

    [...] May 12, 2008 Meeting Upgrade Release Dates: Windows v. Linux Posted by Steven Richardson under Operating Systems | Tags: linux, technology |   I have posted many times on this blog about Linux, and the Ubuntu distribution in particular, and how it compares to Microsoft Windows in the small and solo law office.  In this post, I want to talk briefly about an aspect of OS support that we are all familiar with: software upgrades and their (sometimes far off) release dates.  We have all seen how long it took Microsoft to come out with Vista after the initial release of XP, with many blown deadlines along the way. By contrast, Canonical, the publisher of Ubuntu Linux, has come out with an upgrade every six months (in April and October) for awhile now, and in fact came out with a major long term service (LTS) release last month for both the desktop and server versions (version 8.04 or Hardy Heron). This consistency in meeting release dates is at the very least refreshing, as well as unexpected. In his blog, Canonical’s Mark Shuttleworth commented that, [...]

  27. Demian says: (permalink)
    May 12th, 2008 at 6:18 pm

    Cool!
    Nice to hear this.
    I don’t know what the people at other distros may answer, but if any deal can be reached, it will be great news for the whole GNU/Linux community.
    A big “thank you!” from me to all the people involved in the release of the distro.
    I only have one personal request, given that I’m a designer: the people at Canonical and the Ubuntu community goes to great efforts to release the distro in time. I’d love to see a more marketing-wise release, a-la-Mac OS X. I already suggested something when Gutsy was in beta: http://ubuntuforums.org/showthread.php?t=536957
    Of course, as an ubuntu-artwork collaborator I can no less than offer my help on this task if you want to put up a site for this event.

  28. Imric says: (permalink)
    May 12th, 2008 at 6:28 pm

    I must agree with the previous posters irritated with the problems that the 8.04 release had/has. Until now, Ubuntu/Kubuntu has been almost flawless for me; on upgrade to 8.04 I’ve been showered with odd bugs. They are different on each machine, too (tv capture-cards without sound, AC97 mobo sound stopped working, and nvidia problems were the main ones.) I’m not used to this. Now, these weren’t show-stoppers as I was able to find workarounds for everything (well – I didn’t have the patience by the time I got to the AC97 sound – I just added a sound card and wrote it off) on the forums – but for the first time in years, I feel like I’m using a Linux distro that matches some of the the stereotypes that anyone-but-linux types use. Really, I can’t consider this to be a successful release; it felt like more of a step backwards than forwards. It seems to me as if it was released prematurely for the sake releasing, as opposed to being released because it was ready. This CAN’T be good for an LTS distibution of Ubuntu.

    I do understand the magnitude of the task at hand, mid you – and I’m not deriding Ubuntu. It IS really good; it DOES fill my needs, and I AM satisfied now that I have things working – but if this release is a sample of what we can expect of on-time releases then I for one will NOT upgrade on the same schedule…

    Sorry to be so negative (REALLY sorry) – but you need to hear this feedback as surely as devs need bugs reported.

  29. cosmix says: (permalink)
    May 12th, 2008 at 6:50 pm

    The stability of 8.04 during the beta cycle, and now that it has released, is very good and certainly on a par with both other Linux distributions, and with operating system releases from other large operating system vendors.

    I suppose that’s partly true, but in my opinion only for a very limited number of well defined use-cases — again from the point of view of an enterprise settings. In many respects, 8.04 is a step backwards (viz. gvfs).

    The alternative would have been for the LTS release to have drivers, software, etc which is 9-12 months old, and be stuck on that base for the next five years

    Which alternative would that be? I believe the best decision would’ve been to postpone the LTS release until October. This would have two major advantages: the userbase would help iron out the numerous bugs with new frameworks and libraries (pulseaudio and gvfs being the two most prominent in daily enterprise use) and 8.10 would coincide with the GNOME 2.24 release, arguably a much more refined release than 2.22 and one that’s to my eyes much more suited for enterprise use than 2.22. Sadly many of those improvements slated for 2.24 may never be incorporated in 8.04 unless explicitly backported, something not always possible or straightforward.

    It’s surprising to me that you claim that experience with large software projects would support the 8.04 release being an LTS. In my view, familiarity with enterprise computing, large scale software project development, but also knowledge of the release schedules, current state and roadmaps of the open source projects that Ubuntu builds upon would certainly suggest a mid-summer or even October LTS release. In any case, 8.04 is out. Hopefully the community will manage to sufficiently polish it by 8.04.1.

    Mark Shuttleworth says:

    2.24 will have refinements, but it will also introduce new issues. We were left a little at loose ends by the gvfs changes introduced in 2.22, perhaps we failed to make a strong enough case in that upstream community of the value of stability during this cycle. I remain convinced that the best way to handle this is for the distributions to articulate a common vision, to which I have no doubt upstream projects would respond in whatever fashion is appropriate for them to.

  30. Jonathan Carter says: (permalink)
    May 12th, 2008 at 6:51 pm

    Your blog entry has excellent timing. I was just explaining to a bunch of (government) customers in a meeting today when they asked why they should use Ubuntu instead of Fedora or OpenSUSE. I explained to them about the predictable release and support cycles, which neither Fedora or OpenSUSE can match, and that in a large-scale environment, you can’t really even consider using Fedore or OpenSUSE. Now, I don’t want to come across as a Mark Shuttleworth fanboy, but you have really done a remarkable job putting together the team that you did, and did an awesome job on getting Ubuntu where it is today, even though you have lots of money. I don’t think there are many people who could have pulled it off even with your resources. Really, a honest thanks for Ubuntu, it’s really just brilliant.

  31. Mark Shuttleworth nos desvela la fecha de la proxima LTS « Ubuntu Life says: (permalink)
    May 12th, 2008 at 7:02 pm

    [...] Mayo 12, 2008 Posted by superpiwi in Linux, Noticias, Ubuntu. trackback En su blog, Mark shuttleworth nos habra de las bondades de ubuntu, del equipo de desarrollo y de la fecha de la proxima LTS (Long [...]

  32. Albucasis says: (permalink)
    May 12th, 2008 at 9:32 pm

    Good point, Mark. Syncronicity of the major GNULinux distros is a goal that deserves a big push. The unification idea has made me dream sometimes too, but I never got to see under this release scope. Please do keep it going forward. I hope that people is touched by the idea of melting together all the good brains and efforts that the FLOSS community, from programmers to translators, activists and users have been creating for so long now into a big, nice little system that overcomes all the expectations. I dream it would create a tidal wave of people migrating to GNULinux as we have never seen before, a big push in overcoming frontiers and threats of propietary software companies. A big push towards digital freedom. What is more, a big awakening chance too. It sounds so good to me… El pueblo FLOSS unido… :)

  33. Shuttleworth on Ubuntu Release Cycle « The Never Ending Internet says: (permalink)
    May 12th, 2008 at 9:37 pm

    [...] 12, 2008 Mark Shuttleworth the man behind Ubuntu, writes a nice blog posts on his blog about the release cycle of Ubuntu. If you haven’t read it and are a Ubuntu fan, I would go [...]

  34. Me says: (permalink)
    May 12th, 2008 at 9:53 pm

    Jonathan Carter Says:
    > Your blog entry has excellent timing.

    So the best feature of the next Ubuntu LTS is its release date…

  35. Flounder says: (permalink)
    May 12th, 2008 at 9:53 pm

    The release cycle is (generally) the right thing to do (implement what’s possible and drop what isn’t possible). Software is never perfect, and preventing releases until every last corner case is perfect would take (literally) forever. I have never seen an enterprise or commercial product released without requiring important fixes across the first 9-18 months after the release cycle, and it is rare to see a new OS, database, or other product adopted before an evaluation of at lease 3 months (and generally longer).

    However

    1. 8.04 LTS will require significantly more paid key developer effort to resolve bugs and regressions then Intrepid for the next 3-9 months. (paid because bug fixing sucks, and most developers will work around rather than fixing annoying bugs if given the choice)

    2. The initial release of an LTS release should probably be LTS-rc with the .1 being the actually gold LTS label. LTS releases really need that extra 3 months of rigor to meet the quality requirements of an LTS release. Too late now, but it’s something to keep in mind.

    3. There should be special LTS release notes which cover things like recommended hardware (either stuff specifically tested on or stuff explicitly not meeting LTS supportability goals), the need to do a fresh install, recommended use cases (or precluded use cases). LTS implies a quality level not just a term of support level.

    Hmmm that gives me an idea…. A “safe” upgrade which automatically preserves /home contents, backs up /etc and /home/user/./*, and does fresh install….

  36. cosmix says: (permalink)
    May 12th, 2008 at 10:47 pm

    Mark,

    Thank you for your comment. It’s really appreciated.

    Regarding the common vision part of your response, I couldn’t agree more. Sadly, many open source projects (and the respective upstream communities) are largely impervious to third-party vendor/distro/user needs and requests. I don’t think the problems with gvfs were the result of a failure of the Ubuntu community to impress upon the upstream developers how important 8.04 was for our users. I’m sure they knew it, but didn’t really care as much as you’d probably expect them to. Moreover, it would’ve been presumptuous to think that they’d do anything different to what they did, just because Ubuntu happens to be of the most popular distributions around. :)

    The rule of the dev jungle is still at play here: Novell, Redhat, Canonical (and the Ubuntu community), independent developers, everyone does what they feel they should be doing and steer the direction of the software towards where they (and their customers/users) feel it should go. And that’s good of course. As such, I guess some things tend to follow the release schedules of these respective organisations. Without any dominant vendor or contained development ‘house’, the burden of selecting both the software and the time to release it falls upon the distribution provider. I’m convinced that regular 6month releases and regular biannual LTS releases are an excellent strategy, but I firmly believe that those releases have to reflect a certain level of quality, especially when they are aiming at the enterprise. I fear that in some ways 8.04LTS missed this target, even if your and the other Ubuntu community leader intentions were good in the first place. While an excellent, and polished ‘normal’ release, it lacks the maturity I (at least) would expect from an LTS.

    Regards gvfs: the problem is that the Ubuntu dev. community will have to backport a significant portion of the relevant 2.24-2.26 code to 2.22-ubuntu in order to provide a usable network volume browsing and mounting experience within GNOME. Given Ubuntu’s limited development resources, I’m still not sure whether waiting until 8.10 for an LTS and using 8.04 as a prep release would’ve been such a bad idea, given GNOME’s (upstream) roadmap.

  37. Truefire says: (permalink)
    May 12th, 2008 at 11:27 pm

    Very nice. It’s good to see we will be able to keep Ubuntu releases more bug free this way.
    All the stuff that won’t make it to the ‘official’ date will be presented in the ‘point’ dates.
    So many of the things Ubuntu does with OS’s are revolutionary, this is no less.

    I actually tell everyone with Vista woes about Ubuntu, and they almost instantly switch.
    I found so much need for a better, more reliable OS than Windows, and cheaper than Mac (though Ubuntu is better than both,
    depending on hardware for Apple’s case) that I created a website with another Ubuntu fan dedicated to ease the switch
    from Windows/Macintosh to Ubuntu.

    Check it out (Development) here: http://www.easygeek.com

  38. Truefire says: (permalink)
    May 12th, 2008 at 11:28 pm

    EDIT: website ends with .ORG not .COM.

    Sincere apologies,
    Truefire

  39. meneame.net says: (permalink)
    May 12th, 2008 at 11:54 pm

    Mark Shuttleworth nos desvela la fecha de la proxima LTS [en]…

    En su blog, Mark shuttleworth nos habra de las bondades de Ubuntu, del equipo de desarrollo y de la fecha de la proxima LTS (Long Time Support) de Ubuntu: Sera Ubuntu 10.04 (en Abril del 2010). Tambien podemos ver un grafico de Canonical sobre el ciclo…

  40. Feng Wai says: (permalink)
    May 13th, 2008 at 1:27 am

    Excuse me Mark, you did ship “an enterprise platform release delivered exactly on schedule” but you forgot the more important thing, which is Quality Assurance. Quite simply the 8.04 release is the buggiest ubuntu experience I had since I knew ubuntu (I switched back to gutsy).
    Why not hire 10 more full time QA guys?

  41. Daniel Hardy » Blog Archive » One Upgrade Down says: (permalink)
    May 13th, 2008 at 2:21 am

    [...] drivers (come on Nvidia, open it up! everyone is doing it) the install went very well.  With today’s announcement I think Hardy may be staying on the quad-core for a long time to come. Tags:Ubuntu Hardy, upgrade [...]

  42. buddy says: (permalink)
    May 13th, 2008 at 3:29 am

    First of all, thank you Mark, and all ubuntu developers all over the world. God bless you all for putting such a hard effort to deliver the linux world a system that will stand and be counted among other OSes. But Mark, may be its true that punctuality means good, but stability is a must. If i will have to choose between a buggy but on time released software and a mature one, i think will go for the mature one. In general, ubuntu still rocks! tha’s all i have to say.

  43. Demian says: (permalink)
    May 13th, 2008 at 6:34 am

    @buddy: switch to Debian then… :-D

    (Sorry, felt urged to say that, but I’m kidding :-P )

  44. quidam says: (permalink)
    May 13th, 2008 at 8:05 am

    I think Ubuntu will stay for a long time an end-user distro and it’s for that reason Mark & Canonical are very anxious for future. Many enterprises choose RedHat, Novell because these distro offer services. Ubuntu is very young and had never created nothing in this way. Mark, remember that RedHat went to Cap after only 4 years…
    Have you a real plan development for Enterprises ???

  45. andy says: (permalink)
    May 13th, 2008 at 9:19 am

    Hi all,

    About syncing the releases with other distros: I have read here a few times that people can be stubborn and that this may be (or become) a problem.

    Yes, people can be stubborn, but there may be other reasons: Debian, for example, is maintained/developed almost entirely by volunteers. And there is no way to _force_ a volunteer to get his/her work done in a given time. I don’t know how Red Hat or SuSe handle their releases but for Debian I don’t see a way to make this happen.

  46. chakl says: (permalink)
    May 13th, 2008 at 10:12 am

    Mark,

    the OpenBSD project has been releasing exactly on schedule for several years now.
    Nice to see Ubuntu going the same way.

  47. Ubuntu-Diktator : Karl-Tux-Stadt says: (permalink)
    May 13th, 2008 at 11:08 am

    [...] Shuttleworth, hat in seinem Blog die nächsten Releasezyklen für Ubuntu bekannt gegeben. An den eigentlichen Zyklen [...]

  48. Ubuntu LTS says: (permalink)
    May 13th, 2008 at 11:12 am

    [...] Yesterday Mark Shuttleworth wrote about next Ubuntu LTS release. The next LTS release of Ubuntu will be 10.04 LTS, in April 2010. Please read more information in the Mark’s blog. [...]

  49. The Art of the Open Source Release | distros | Fair or Unfair says: (permalink)
    May 13th, 2008 at 11:34 am

    [...] Shuttleworth writes about another advantage that free software sets enjoys over monolithic, proprietary code [...]

  50. Rudd-O says: (permalink)
    May 13th, 2008 at 12:08 pm

    And with this olive branch, single-handedly, Mark achieved the consolidation that for one decade Linux distributions have lacked.

    Let’s hope the dream of synchronicity will take us to new heights and easier cross-distribution efforts.

  51. Tim Richardson says: (permalink)
    May 13th, 2008 at 12:54 pm

    Two releases ago I swapped from Ubuntu to Debian because Ubuntu’s OpenOffice Base was badly broken. I was amazed that a version could be released with such a bad bug, and I concluded that the Debian approach is better: you can use the stable version for rock-solid stability, testing for a pretty bullet proof version, and unstable, the developers’ sandbox, for cutting-edge (I personally use testing on my main personal computer and on my home-server). The debian developers remain very concerned about the 2.22 Gnome Nautilus; it is still held in experimental, it has not been released even into Debian unstable; yet Ubuntu came to a very different conclusion and released it in 8.04. I think this shows a strength of Debian. Two competing operating systems, OS X and Microsoft, are very, very conservative about releases; Ubuntu is actually almost radically aggressive in bringing new technologies into a release version. If you prefer stability, you might need to look at another distribution. The strict release schedule of Ubuntu is, I think, too dogmatic. I am not convinced it is good thing; some key components such as Gnome need to slow down and focus on quality releases, not rush along at Ubuntu’s frenetic pace.

  52. Vincent says: (permalink)
    May 13th, 2008 at 1:34 pm

    A great idea, but why should you colllaborate with two out of three? Imagine if you’d synchronize with just Red Hat, Novell or Debian – wouldn’t it be logical for either of the other two to follow, or at least keep an eye on how it’s working out?

  53. Christophe Olinger says: (permalink)
    May 13th, 2008 at 2:15 pm

    Hey evrybody,

    What I also realized, that the new feature list is also quite similar when comparing it between the major distros. What about having a common blueprint manager for all distros. If I see an interesting feature for one distro and work on it, I would be sure that it would be also included in other distros. Maybe we can avoid a bit of the “ten people working on the same thing but do not know about each other.” Hmm, I realize that this comes close to: “lets all have only one distro and concentrate our mental power”. But thats a completely different question (Is it really?)

    Cheers

  54. Benjamin Schweizer. Operating Systems Lifecycle Chart. says: (permalink)
    May 13th, 2008 at 2:15 pm

    [...] 10.04LTS Apr 24, 2010 – Apr [...]

  55. ubuntupedia.de » Nächstes LTS-Ubuntu heißt 10.04 says: (permalink)
    May 13th, 2008 at 2:53 pm

    [...] ist die aktuelle Ubuntu-Version keinen Monat alt, da veröffentlicht Mark Shuttleworth in seinem Blog bereits Pläne für die kommenden Versionen. Eines steht dabei fest: die nächste LTS-Version [...]

  56. sourcode » The Art of Release says: (permalink)
    May 13th, 2008 at 5:01 pm

    [...] via markshuttleworth.com [...]

  57. Weeber says: (permalink)
    May 13th, 2008 at 5:48 pm

    I upgraded to Hardy and is a good release but I have to say that it isn’t as stable as Dapper was when it came out. Dapper was rock solid and I’m sure that many users started to respect Ubuntu for its stability from there. Hardy desktop version on the other hand need some workarounds that I don’t think it should have been shipped with, like the PulseAudio workaround to get youtube to play sound (I know it is suppose to be an Adobe flash bug but who will the end user blame?). Certainly the PulseAudio community is the one to workaround that issue.

    To follow I will quote the Arstechnica conclusion of the Hardy Heron review: http://arstechnica.com/reviews/os/hardy-heron-review.ars

    “Ubuntu 7.10 set the standard for power and ease of use for Linux desktop distributions. Ubuntu 8.04 pushes forward, but trips over its own feet because of the PulseAudio integration problems. The clear failure to improve some of the weak spots that we identified in 7.10—like Tracker’s mediocre search tool—also detracts from the value of this release.

    Despite the problems, Ubuntu 8.04 still offers a better and more polished desktop than its predecessor. Users who are willing to work around the PulseAudio problems will have no trouble adopting it now, but those who aren’t comfortable with that might want to wait until the release of 8.04.1, which will resolve some of the most significant bugs that users have discovered in the release.

    With an incremental six-month release cycle, bugs are easily justified as transitional problems introduced by new technologies that haven’t had a chance to mature yet. The problems are harder to excuse for this version because, as a long-term support release, it was supposed to be more robust. The previous release was the “gutsy” one with experimental new technologies and this was supposed to be the “hardy” one with an emphasis on stability. I’m quite happy to suffer with the bugs for six months, but I can’t imagine anybody using it on the desktop for three years in its current state. This release is disappointing because it falls short of what was promised, but it still delivers a lot of value for experienced users who will be able to work around the weaknesses.

    Ubuntu has achieved Linux desktop dominance by offering a significantly better experience and more usability than competing distributions, but with the others are starting to catch up in those areas, this is a bad time for Ubuntu to be stumbling. Fedora 9 and openSUSE 11 are both right around the corner and they might finally give Ubuntu a real challenge on the desktop.

    Ubuntu 8.10, which is codenamed Intrepid Ibex and is tentatively scheduled for release in October, will include improvements to mobile computing and desktop scalability. A strong 8.10 release with fewer problems out of the box would help put Ubuntu back on track.”

    I couldn’t have said it better myself.

  58. Ubuntu 10.04 será “LTS” « Tux Azteca says: (permalink)
    May 13th, 2008 at 5:54 pm

    [...] * El anuncio. [...]

  59. Fecha de la pŕoxima LTS de Ubuntu. « La Casa de Tux says: (permalink)
    May 13th, 2008 at 6:07 pm

    [...] de la pŕoxima LTS de Ubuntu. 13 05 2008 En su blog, Mark shuttleworth nos habla de las bondades de ubuntu, del equipo de desarrollo y de la fecha de la proxima LTS (Long [...]

  60. Andre Noel » Blog Archive » Mark Shuttleworth fala a respeito das liberações de versões do Ubuntu says: (permalink)
    May 13th, 2008 at 7:01 pm

    [...] grande Mark Shuttleworth (SABDFL) escreveu um post a respeito das liberações de versões do Ubuntu. Leitura [...]

  61. frank_0 says: (permalink)
    May 13th, 2008 at 9:37 pm

    How are OEMs supposed to cope with a major
    release every six months?

    The 6 month cycle for major releases is
    just pandering to the fanboys – who then
    discover the bugs and start whining.

    Here is one who would appreciate a more
    prudent approach: a major release once
    a year followed by a maintenance release
    six months later.

    You would then be able to walk into a
    bookshop and find ubuntu books which are
    not stale and obsolete.

  62. pheeror says: (permalink)
    May 13th, 2008 at 9:50 pm

    I’m sorry but I must agree that Hardy is a little more error-prone than the previous releases. The worse is that I’m still looking forward to next release of ubuntu ;-)

  63. A. Peon says: (permalink)
    May 14th, 2008 at 12:39 am

    I can appreciate the concern for “synchronicity,” but from a stability standpoint — most of Hardy’s showstopping warts are in the kernel (certain ATAPI devices magically rendered incompatible, various chipset support quirks that 6.06′s kernel happened not to step on), and in that respect 1. distributors will forever be at the mercy of Linus and the gang, and 2. users served by the diversity of the mess as it stands may suddenly lose out on *every* distro’s stable release if identical kernels and so forth are adopted across the board.

    ‘Synchronization’ as it’s being promoted mostly benefits users (and non-free software companies) with binaries that need to be deployed widely. Perhaps what’s really needed is a truce such that all major distros will maintain library version N while offering N+1, so that software intended for binary distribution can link to a common stable platform while the individual distributors move on with the introduction of new stuff, and N+1 would receive more testing before becoming the new ‘N’. This would solve for, say, Adobe distributing their non-free blobs linked to the latest and greatest libs and ‘forcing’ users of a stable system to perform an entire distribution upgrade to get them. [It wouldn't solve for nVidia releasing blobs tied to specific kernel releases, but I have more sympathy for a software company trying to distribute working software than a hardware company that refuses to document the hardware they sell.]

    Note that, as it stands, I’m left praying a point release will support my *old* hardware as well as 6.06 did. And yes, I’ve filed bugs.

    For QA purposes, it might be best to release some extremely cut-down ‘instrumentation’ images (for CD or USB stick) that check kernel sanity and little else before or during the alpha process; I ended up refraining from testing with full live CDs because of a bug in an init script that could’ve eaten my filesystems, and lo, when release day came, it turns out 8.04 wouldn’t even boot live due to new IDE/ATAPI bugs with my hardware.

  64. Anouk L. says: (permalink)
    May 14th, 2008 at 12:48 am

    Like many others, Ubuntu 8.04 LTS is the worst “thing” that I tried! So, do what you want with your distro, but please, don’t try to impose your views to others! If Debian decides to follow the “Ubuntu” way, I will stop contributing to it and look elsewhere!

  65. Saint Peter says: (permalink)
    May 14th, 2008 at 2:23 am

    I would have expected a LTS release to be a grown up, polished, and clean updated version of the previous release, in this case 7.10. That is, I would have expected the Ubuntu community to have spent six solid months polishing what needed polish from the previous release. That I would be happy to take as a LTS release.

    Instead, the driver for the LTS release was the “shipment date,” with concerns about stability and polish relegated to a secondary role. The result is the poor desktop edition in 8.04 with a swarm of little annoying bugs here and there. In my case for example, a bug (already confirmed in Launchpad) whereby my system freezes whenever I logout from my Gnome session is driving me crazy.

    Fortunately the server edition is not too bad. However, as to provide another example of the questionable quality, the python Trac system widely used with Subversion and other source code management software doesn’t work out of the box, manual twiking is required.

    The best news to me is the upcoming 8.04.1, 10.04.1, … releases. These I believe will be the real LTS releases, unless future LTS releases build on the previous non-LTS release instead of just shooting for the “shipment date.”

    One final comment about synchronization of multiple releases. I wouldn’t expect the distros (RH, Suse, Mandriva, etc.) to agree on a shipment date. There’s too many business, political, and other type of reasons why that has very low possibilities of becoming reality. However, I do believe that release synchronization of major projects such us Gnome, KDE, Xorg, etc., is entirely doable and would be a welcome event for the FOSS community in general. If these major projects get in sync, and only if, I will think the chances of the distros getting in sync as well will be meaningful.

  66. Chris Lees says: (permalink)
    May 14th, 2008 at 3:27 am

    I agree with a previous poster that you should have an extra 4 weeks on LTS releases – support for my mass storage MP3 player broke in Hardy! I repeat: *mass storage device*! Dapper was a great release. It’s also good as a differentiating factor from all the other releases, to say to the enterprise “We had a longer bugfix phase than the normal releases”. Freezing at the same time but having a longer bugfix cycle also makes it possible to have the community sending out beta discs for anyone to test and report bugs on.

  67. La próxima LTS de Ubuntu en Abril del 2010 y… « Entre tuxes y pepinos says: (permalink)
    May 14th, 2008 at 5:23 am

    [...] Fuente: Mark Shuttleworth [...]

  68. Chen says: (permalink)
    May 14th, 2008 at 6:44 am

    Am I the only one that see something wrong with the chart above?

    What a Kubuntu 6.06 LTS corporate user should do between October 2009 and April 2010 (or July 2010 – the release date of 10.04.1 – taking into account the Hardy stability issues)?

    Is it official now that KDE will not have Ubuntu LTS support?

  69. FreeSoftwareUser says: (permalink)
    May 14th, 2008 at 7:16 am

    Hi Mark,

    Why would Red Hat cooperate with Ubuntu, especially now that Ubuntu also has its sights set on the server market. Don’t they consider Ubuntu a threat?

    Mark Shuttleworth

    I don’t think Red Hat would see Ubuntu as a threat, we appeal to different audiences.

    Have you ever noticed that competing car dealerships, fast food restaurants and other very similar businesses all setup shop next to one another? You get food courts in shopping malls, for example, where you have all the take-away food places in one area. You would think that the competition would be bad for them. But counter-intuitively, all the restaurants do better when they are all in the same place, and the same is true of car dealerships. The phenomenon is called “clustering”, and it works because people first decide to go “looking for a car” and then later decide which car, or which dealership. I think Linux is the same – if we coordinate our releases, we send a very strong message to the outside world that will bring more people to Linux in the first place – making the pot bigger for everyone.

  70. ensi says: (permalink)
    May 14th, 2008 at 7:57 am

    With the release cycle being this fast Ubuntu as well as X.org/Gnome/KDE and other projects are in a constant beta cycle. Quality and quality assurance is always secondary to the developers desire to create new stuff. Of course one can always stick with an old version of any given Linux distro and not upgrade but then one might not get a) new applications or b) new versions of old applications with important fixes. This is because things are always changing underneath applications and application writers naturally adapt their code to that. Thus APIs and ABIs are constantly changing and and in order to get new stuff the user ha to commit to a constant upgrade/reinstall cycle. Personally I absolute hate upgrading. I want to spend my time doing my work, not wasting my time with resolving the problems (there always are some) after an upgrade or reinstall and retweak my settings to my liking. Now contrast this with Windows where the API has stayed the same for years and developers just make do with that. Yet new applications emerge every day. Imho, this instability and constant change is the greatest problem with Linux as a whole.

    The bottom line is, you wont get a rock solid system release, untill you freeze features and go through several “bug-fix” cycles only. And this is the greatest challenge the Linux/Ubuntu/Open-Source community ever has to face if they ever seriously plan to become anything more than a side note.

    Mark Shuttleworth says:

    You make a very good point about the pace of releases and one’s ability to make fixes. We can of course choose how many new components we bring into a release that is going to be designated LTS, and this is the primary tool we have for allocating time between new-version-integration and stability/fix work. My sense, though, is that the upstreams would respond very positively to a coordinated cue from distributions w.r.t. their long-term support plans. If upstream knows that multiple distributions will be using a particular version, they can respond appropriately.

  71. Tom says: (permalink)
    May 14th, 2008 at 8:47 am

    Hello again,

    I thought about it some more and I think what you want is not the same release date but the same feature freeze date or something .. that way Debian for example can still release when they think it is ready. So they might release a few weeks/months after Ubuntu but still ship the same ( or very similar ) packages and versions.

    Wouldn’t that make more sense?

  72. Ubuntu 8.04 LTS Released ! - Page 2 - TechEnclave says: (permalink)
    May 14th, 2008 at 8:53 am

    [...] permalink So, Ubuntu 8.04 LTS will have not just 1 but multiple point releases like the Fedora respins which include the latest updates and hardware support. See this blog post by Mark Shuttleworth himself: Mark Shuttleworth Blog Archive The Art of Release [...]

  73. Simon Scullion » Twitter Updates for 2008-05-13 says: (permalink)
    May 14th, 2008 at 9:05 am

    [...] Shuttleworth on plans for Ubuntu releases into 2013, some commitment http://www.markshuttleworth.com/archives/146 [...]

  74. Oliver says: (permalink)
    May 14th, 2008 at 10:01 am

    (Off topic)

    Hi there Mark !

    I just wonder how or if the size of Ubuntu and the other releases is a concern for you ? I mean, they are always around 700 MB but contains more and more useful free software in every release =)

    When do you think Ubuntu will be delivered on two CD’s or one DVD like Windows ?

    Sincerly
    Oliver

  75. Me says: (permalink)
    May 14th, 2008 at 11:11 am

    > I think what you want is not the same release date but the same feature freeze date

    What is the features of Ubuntu 04-10 ?
    Nobody knows. So nobody can agree with Ubuntu.

  76. alex says: (permalink)
    May 14th, 2008 at 1:10 pm

    An important factor for the “don’t release until it’s ready don’t worry about dates” is always the question of just when do you decide enough is enough. It is one of the benefits of a fixed release date provided you trim things from your ambitions to meet it rather than push to do more than can be done.

    Besides the other great Free Software factor is that if you don’t like that approach (dates for releases) you can look for an alternative that only releases when good enough. And largely find that all your applications you are moving with you still work.

    Reality says one release now and again will be worse than others in terms of little bugs. Just so long as it doesn’t happen very often.

  77. Meet Barnaby » Blog Archive » Cross Project Synchronisity says: (permalink)
    May 14th, 2008 at 1:58 pm

    [...] Mark Shuttleworth has an idea in his latest blog post. The founder of Canonical Lmt. is planning an ambitious synchronization project.  If several Linux distributions agreed to synchronize their major releases it would put pressure on upstream to create a major releases on a similar schedule. This sounds to me a little bit how Eclipse fixed a growing problem it had, all of the plugins would release whenever and nothing worked with whatever other version, it was hell. So they synchronized the release of all their related projects.  Mind you, and Eclipse distribution is about a fraction of a percent of the work of a linux distribution, but that means that if you multiply the benefits by the larger scale you get a result that will shake the technology world as we know it. [...]

  78. Barnaby says: (permalink)
    May 14th, 2008 at 2:04 pm

    Mark,

    The problem you are tackling of synchronizing upstream projects for a consistent release is not unique to Ubuntu, or even to Linux distributions for that matter. A lot of companies face this problem; my last company did. Some parts we took from upstream were stable but old, some were new but unstable and the end result is inconsistent.

    I wrote a little blog entry on my blog, http://www.meetbarnaby.com, about how your proposed cross project synchronisity would help a lot of others in the FOSS world.

    Keep up the good work.

  79. ConFigures » Blog Archive » Synchronicity and Traceability says: (permalink)
    May 14th, 2008 at 3:26 pm

    [...] Shuttleworth wrote a post about the latest Ubuntu LTS (long term service) release, 8.04, and associated and upcoming release management (and his hopes for coordination of release schedules with other open source efforts).  His entry [...]

  80. CWH says: (permalink)
    May 14th, 2008 at 5:29 pm

    Now the LTS roadmap is available, I would like to see a combined effort to get Ubuntu into the education field. I know there is a specialist version, but there just isn’t enough recognition in the education community that there is an alternative to Microsoft , and no-one is stepping up to the plate and advertising the advantages.

    Colin

  81. Unità, coordinazione e dinamismo! :: mondolinux linux ezine says: (permalink)
    May 14th, 2008 at 6:50 pm

    [...] suo blog il Leader di  Ubuntu e della Canonical spiega che sarebbe un grosso vantaggio per gli utilizzatori [...]

  82. dthomasdigital says: (permalink)
    May 14th, 2008 at 7:45 pm

    The chart to me is amazing for many reasons.
    1. That you would commit so such goals
    2. That you have proven this is just not eye candy and you can hit these goals
    3. The will to work with other vendors, it really does sit you apart

  83. 8chapas.com » Blog Archive » Mark Shuttleworth pide sincronizar la liberación de distros GNU/Linux says: (permalink)
    May 14th, 2008 at 10:58 pm

    [...] apunta en su blog que un calendario de lanzamientos conjuntos beneficiaría tanto a usuarios como a distribuidores. [...]

  84. Steven says: (permalink)
    May 14th, 2008 at 11:11 pm

    I’m just your average home user who got fed up with Windows a couple of years ago and switch to Ubuntu. I did a clean install of Hardy 2 weeks ago and couldn’t be happier. I’ve had no negative issues with it other than some of my favorite Firefox addons are not yet compatible with Firefox 3. I’m known as the “computer guy” amongst my friends, but I decided a few weeks ago that I’ll only offer help and support to those willing to migrate to Ubuntu.

  85. Zygo says: (permalink)
    May 14th, 2008 at 11:38 pm

    I don’t think synchronized release schedules (or for that matter synchronized anything beyond individual vendors satisfying critical technical requirements for their own customers) between Linux distributions is a good idea.

    The recent snafu with the Debian openssl libraries is an excellent example of why moving toward a monoculture is a bad idea. There’s no way to predict when a serious bug like this will occur; however, it’s easy to predict that it *will* occur eventually. This is a simple matter of statistics: if the probability of a vendor producing a bug-free release is anything less than one (and nobody’s perfect, not even Ubuntu ;-), then the probability of a vendor producing a bug-free release every time approaches zero rapidly. Over the long term, failures of similar scope to the openssl bug are therefore certain (just look at the Windows security track record if you ever doubt this). If my distro’s release schedule is significantly different from your distro’s release schedule, then at least one of us has a chance of having a secure system in their data centre, on their desk, or on their lap the day after the security advisories come out.

    As a user, if my needs are not met by one distro’s release schedule, I have a choice of switching to another distro with a more compatible schedule. This is desirable, since not all users have identical requirements: some are better served by fixed quality on a variable schedule, while others have a fixed schedule and their own internal QA processes to deal with quality issues. This freedom of choice, with its inherent opportunities for evolution, diversity, and survivability, is lost if all the Linux distros agree on a single schedule, a single development methodology, a single base kernel revision, a single group of shared library major versions, or a single anything else. In practice, of course, most distributions are fairly similar; however, they’re not so similar that they all share single-line security-destroying code bugs, and the varying release schedules mean that not all users will get the bugs at the same time.

    The more that the content of Linux distro releases are centrally coordinated, the more serious the failure modes become. Imagine if all Linux systems worldwide were affected by the SSL bug, instead of just Debian and Ubuntu. Actually we don’t have to imagine this, since Microsoft keeps providing case studies for us to learn from.

    Red Hat and {Free,Open}BSD, as crummy, annoying and alien as they are, now seem a whole lot more attractive to me since yesterday. Over the past two days I’ve watched as a number of SSL certificate infrastructures have narrowly being composed entirely of weak keys by mere historical accidents of their deployment–accidents that were only possible because people happened to be using a variety of implementations of interoperable software in the past.

    Large-scale synchronization of Linux distros, either of content or of delivery schedule, is such a bad idea, that if the distro I use started participating in such a scheme, I’d immediately start migrating to another–or if one didn’t exist, fork my own. Or at least I would, if I actually used unmodified LTS releases in the first place.

  86. TuxJournal.net » Shuttleworth: sincronizziamo i rilasci delle distribuzioni says: (permalink)
    May 15th, 2008 at 12:05 am

    [...] pagine del suo blog, Shuttleworth sembra avere le idee ben precise su come migliorare la situazione dei rilasci [...]

  87. aaron lea says: (permalink)
    May 15th, 2008 at 6:16 am

    The core of 《易经》 (“The Book Of Changes”, not a very good translation) is “Changing, Unchanging, Simple”(I have only read the Chinese version, not English or others, so don’t know if this is a normal translation).

    Changing: everything is changing and always changing.
    Unchanging: the essence (or called inbeing) is unchanging.
    Simple: the essence of things is always simple.

    Now, the linux desktop is a young man, and the art of ubuntu release, the rhythm of the release rocks the free software world at the right moment. It’s great (but I’m not sure whether the release cycle is suitable for server edition). I didn’t use linux for my desktop until ubuntu appeared. Ubuntu is my primary system now. Changing bring us risks, also bring us opportunities, bring us advancement.
    Free software is growing, Linux is growing, Ubuntu is growing. Maybe someday they matured, the rhythm is not suitable, it’s time to change a new rhythm to dance.

    The Unchanging is the spirit of “humanly to others, following and influencing the wave, keeping creativity, keeping things simple”, I think.

    Thank you all. Mr. Shuttleworth, I’m your fans.

  88. Next Ubuntu LTS in 2010, unless Linuxes synchronize says: (permalink)
    May 15th, 2008 at 6:27 am

    [...] in his blog said a common release schedule would bring benefits both to users and Linux [...]

  89. Mark Shuttleworth por la sincronizacin de lanzamientos Linux at Tod-OS.com :: Te ponemos al dia says: (permalink)
    May 15th, 2008 at 8:44 am

    [...] por la sincronizacin de lanzamientos Linux Published by xender May 15th, 2008 in Ubuntu Mark Shuttleworth, director ejecutivo de Ubuntu, ha expresado su opinin de que el usuario se beneficiara si las [...]

  90. jacques says: (permalink)
    May 15th, 2008 at 9:02 am

    I must say it’s quite amazing that this release on schedule. But I was much more amazed by the sheer number of bugs able to make it to a LTS release as a consequence of holding the schedule, and the overall poor quality of the 8.04 release (sluggish, voracious with memory, random sound problems). Whatever marketing spin is added onto it, and despite the fantastic work and dedication of all contributors, this is it for me. I’m moving on to Arch linux. I must say: thanks a lot for having converted me to linux!

  91. Ped7g says: (permalink)
    May 15th, 2008 at 10:48 am

    An LTS release that hit the date with broken GCC. *clap* *clap*
    I’m on Kubuntu only since 6.10 (still sticking to it BTW, but finally I will have to move due to end of support :( ), but with every new version I feel the older HW stop working at some point… it’s either “improved” driver or just unmaintained one not compiled properly for fresh kernel, etc… I don’t blame Ubuntu team for this, the root of the problem is upstream, but it’s sort of pity the QA of Ubuntu can’t afford even more different HW and even more extensive (and expensive) testing if everything works.
    Also in the time many new features were added to linux core, like upstart, new way of naming IDE (SATA) HDDs, audio server, now the KDE4 is emerging … and with all that new functionality I feel entropy getting closer and closer, I’m afraid one day in not so distant future (years or tens of years) GNU/Linux will be total mess and chaos (according both to Mythical Man Month book where the simplified premise says the most stable version of OS/platform is the first version and personal observation).
    I would personally often prefer old functionality with even more stable and cleaner and more efficient code, than new functions or look. It often takes me months to make some new functionality work for me instead of against me, and 60% of new stuff usually is of no help at all for my personal needs. I of course can accept other people may have different needs and maybe they gain much more from current development of things. But for *ME* the Kubuntu 6.10 was of tremendous value, and no later Kubuntu release got close to that level.

    Of course thank you for everything anyway, I still owe you a lot for all that hard work.

  92. Ped7g says: (permalink)
    May 15th, 2008 at 11:03 am

    About the synchronicity idea:
    I think planning the big dates ahead enough can help also volunteers. You can’t force them to hit the date with release, but it’s more likely their projects will be ready at that date (more likely 2-6 months late after that date, but at least they will be in some more stable “major” version).
    So while I think the idea is great, I don’t believe it will work up to the point that at April 2010 there will be rock solid distribution. But together with point releases it may do wonders for 10.04.1 or 10.04.2 for example. So really like these ideas.

    Now I have read rest of previous comments and I see the 8.04 LTS has much more of problems than just broken GCC, and I wonder, should we really call this “LTS version release hit the date”? It looks more to me like the “something hit the fan on date”.
    I hope my thoughts about point releases will prove to be true, I’m really looking forward for 8.04.1.

  93. Aceler says: (permalink)
    May 15th, 2008 at 1:18 pm

    Schedules makes stability.

  94. Harrison Metzger says: (permalink)
    May 15th, 2008 at 4:16 pm

    I think Mark Shuttleworth’s idea is great. While I don’t think the release date should be carved in stone, it would sure be great if all the major distros had the same version of major packages such as the kernel, gnome, kde, etc.. However, then Distros will be less independent, and I feel the purpose of the distros is to have independence. While I like the idea, I’m sure others may not, and it may turn out to be counter productive to GNU/Linux as a whole. Lastly, I wish Ubuntu would drop xubuntu, and kubuntu, and just have ubuntu and have the setup give you a choice of graphical desktop environment. I know you can install GNOME on kubuntu, and KDE on ubuntu, IMHO its sort silly.

  95. Jose Hevia says: (permalink)
    May 15th, 2008 at 4:17 pm

    I updated to hardi, and had problems with the laptop’s nvidia 7300 go graphic card and wacom tablet. I use the screen rotated(tablet mode)so I need hardware accel.

    It worked fine before. I know how to modify de xorg.conf file but a normal user should not.

    I had a strong desire for noveau project to work.

  96. Martin Ankerl says: (permalink)
    May 15th, 2008 at 5:30 pm

    This is truly an excellent idea. Besides that, regular releases are one of the corner stones of agile software development. Extreme programming, SCRUM, etc. all have regular releases as an important part. Fixed cycles will increase the quality of each distribution, and development will be more productive.

  97. Tolan Blundell says: (permalink)
    May 15th, 2008 at 6:00 pm

    I’m very sorry to do this but I really have to agree with the other posters who have concerns about the current release schedule.

    I think that Ubuntu is great, I really do, the amount of innovation Canonical and the community have put into the desktop experience for non-technical users is fantastic.

    But the problem that *keeps* coming back and biting me is the QA, Ubuntu is just too buggy. I can understand the attraction of a 6 month cycle, and I can see some of the value to it, although I would personally make it a year not 6 months, but that’s beside the point.

    The problem as I see it is two-fold. The first is that too much of the 6 month cycle is dedicated to integrating new features. I understand that the rate that new features are added is a large part of the attraction of Ubuntu, but I really think that more of the release cycle needs to be dedicated to bugfixing.

    The second issue is that a strict release schedule is fine in terms of limiting your ambitions in terms of adding new features, but it’s no good if it means you ship with known major issues. This has been an ongoing problem too as far as I’m concerned, it’s just particularly obvious with Hardy.

    I’ve only installed Hardy on my laptop so far but I already have a list of issues as long as my arm, it’s enormously frustrating.

    Anyway, I really hope these issues get sorted, I think if they do then Ubuntu has a great future ahead of it.

    Thanks for all the hard work to everyone involved in Ubuntu, I hope my comments are taken in the spirit they’re meant.

  98. Tolan Blundell says: (permalink)
    May 15th, 2008 at 6:02 pm

    Oops… I forgot to say. Apart from all that, I think the idea of synchronised release schedules is actually a good one, I just think more time needs to be spent on testing, and there’s needs to be a willingness to miss release dates if there are serious issues. (Pulseaudio issues are *serious*! It doesn’t have to be data corruption…)

  99. The Proliferation of Linux » Blog Archive » Shuttleworth Calls For Coordinated Release Cycles says: (permalink)
    May 15th, 2008 at 6:09 pm

    [...] [Mark Shuttleworth via Slashdot] [...]

  100. oshelpdesk » Nächste LTS-Version in 2 Jahren says: (permalink)
    May 15th, 2008 at 6:19 pm

    [...] Hier geht es zum Blog-Eintrag “The Art of Release” von Mark Shuttleworth. [...]

  101. Srinath Madhavan says: (permalink)
    May 15th, 2008 at 6:49 pm

    Mark, I am particularly disappointed at what Ubuntu has become over the last few years. At the very outset, I’d like to say for the record that I’m no GNU/Linux noobie trying to express my frustration here. I’ve been using GNU/Linux since 1998 (the Red Hat era). Experience has taught me never to bother with upgrades and always go for fresh installs. So none of the following problems relate to upgrade issues.

    In the beginning there was all this hype and lots of publicity that Ubuntu gained in being the most user-friendly distribution out there. But as time progressed, I began to notice that problems were more commonplace with newer (and quicker) releases. For instance, while Feisty worked well on my Laptop and my desktop, Gutsy started showing signs of beta software in that there were many unexplained terminal crashes, X lockups etc. I was quite confident that Hardy would not only rise above these small problems, but prove to be the best distro out there. Unfortunately this is not the case. Neither did Hardy allow me to install drivers for my USB TV Tuner (that worked well in Gutsy by the way), but it also magically dropped support for suspend/hibernate on my Laptop. The point is that the hardware in question is not *new*. Add to it, it was well supported by the earlier Ubuntu releases.

    Instead of merely complaining about things that don’t work for me, I decided to poke into the causes of these problems. To this end, I decided to try out different distributions which can be regarded as being on-par with Ubuntu with respect to software versions. These were Debain (Lenny) and Fedora 9. Even though Debian had a Ubuntu-skeleton feel to it, it turned out to be more stable and all my hardware worked well with it. Granted, I still had to build a few drivers myself. On the other hand, Fedora 9 proved to be even better as *everything* in my Laptop (including my TV Tuner) was detected out of the box (something I expected out of Ubuntu Hardy) and even Suspend/Hibernate works flawlessly (so far). Granted F9 is bleeding edge but it’s not like Ubuntu is running a ageing 2.4.x kernel compared to F9.

    I am never shy to experiment with building drivers/kernel modules or even recompiling the Linux kernel. I’ve done these dozens of times on several different distributions. But somehow, every time I tried to support my hardware in this fashion, Hardy kept pissing me off with all kinds of compile errors. It would simply not work with the standard ways of rebuilding kernels and modules. After poking around in the Ubuntu forums, I finally realized that part of the problem stems from Ubuntu’s own unique way of managing things even if it requires violation of Linux Standard Base (LSB) guidelines. This IMO is not a good sign. At the very least, I expect that being based off Debian, Ubuntu should know when and where to stop its modifications/improvements. If you diverge too far from Debain (testing) you risk stability and that is exactly what Hardy is now showing signs of. Part of the problem can be attributed to very short release cycles but that is no excuse, is it? Look at Fedora.

    To sum it up, although Ubuntu is doing a great job weaning people from MS Windows, I fear that its success will be short-lived. I’d hate to see that happen to a distro like Ubuntu which showed great promise in the beginning.

  102. Sean says: (permalink)
    May 15th, 2008 at 6:58 pm

    Firstly, congrats on birthing what has become a revival of a social movement–and furthermore for giving it more steam than it’s ever had before. Ubuntu is fantastic. I’m not quite as annoyed with the 8.04 release as others but I do see their frustration. This project has taken on a life of it’s own and will continue to grow and mature regardless of whether or not it has a regular release cycle.

    From a previous poster…

    “It’s a shame 8.04 was tagged as an LTS release. While a great ‘normal’ release, it’s certainly quite far from the robust, polished and functionally conservative featureset that any ‘enterprise’ release is typically associated with. In fact it’s exactly the opposite; flamboyant, bleeding-edge and somewhat buggy. I’m certain that the decision was conscious and that Canonical is expecting 8.04.1 to remedy all that’s wrong with 8.04.”

    This is indeed very frustrating. This behavior is exactly like Microsoft. Release it and call it gold, hear users complain, then release a series of service packs to make it almost what it was supposed to be. The point release schedule is fantastic and a great idea (buy whomever came up with that a cookie). The problem comes when the intent of such releases appears to be to fix problems rather than add [stable] features.

    But then I thought for a second. If I was in your (Mark Shuttleworth) position, what would I do. There is significance to a regular, predictable and reliable release schedule–especially when the operating system in question is in live business environments. Such a schedule affords the business the luxury of planning upgrades before the release date. However, buggy new software with new useful features can cause more problems than old, maintained stable software.

    Perhaps a better schedule would be to release the standard Ubuntu release every six months with an LTS release coming three months after the previous release every X years, and then basing the upcoming standard release on the previous LTS. That would allow for three months of significant bug fixing for the LTS, better upstream packages, and a stable base for the next regular release.

    Vivé la révolution!

    Mark Shuttleworth says:

    We had long debates at the end of the 7.10 cycle about what we should do with 8.04 – whether it should be an LTS, whether we should do a longer cycle like we did for Dapper, or whether we should declare the first point release the LTS. The community is more conscious than ever about the high standard to which our releases are held, and the responsibility to get things right. At the end, the view that carried the day was that 8.04 should be designated LTS, rather than waiting for 8.04.1. Large deployments will wait for the first point release or two in any event. There was more work done on 8.04′s QA and bug fixing than on any previous Ubuntu release, by a larger team both full-time and volunteer. There were some unfortunate issues with a large GNOME transition that caught us off guard, but we’re fixing those issues one by one now.

  103. Jim Dennis says: (permalink)
    May 15th, 2008 at 8:38 pm

    Mark,

    I think you might have missed one vital point in your comparison of Microsoft and similar closed door product development to our OSV model (development of distributions).

    There is no secret to having predictable release schedules. You cannot reliably combine the that with an MRD (marketing requirements document) listing as yet undeveloped features. So long as all of your components are in a usable state (that is to say that each of them has some usable version available, even if it has to be some old “back rev.” then you can, with a bit of internal engineering discipline, maintain a development tree that can ship on any arbitrary day.

    In large part the Debian developers do this by virtue of the fact that most of them run their systems on “unstable” or “testing” and keep their systems upgraded to within a few days of the head of the project. Thus the Ubuntu team has many opportunities to select and pin down stable points with their proximal upstream sources.

    To put it cynically, Ubuntu can ship any time it wants because it doesn’t commit itself to shipping with any particular set of new enhancements.

    The other advantage that OSVs have over Microsoft is the loose coupling among the components.

    Microsoft’s problems include their own commitment to an MRD. They can’t ship until all these *requirements* are (allegedly) satisfied. Also Microsoft embraces and strives for a much greater degree of couple among components. They call this being “integrated” and savvy market analysts refer to it as “vendor lock in.”

    Right now Microsoft’s biggest competition is not Linux, nor MacOS, or any other outside product offering. Their biggest competitor is their own legacy. They have implemented far more features than any of their customers can use; so they can no longer market upgrades using “new features.” They have *sold* many millions of copies of their software for years past the time that they became aware of the need to change their revenue model into a subscription basis. (They have tried and continue trying, to shift models but there’s quite a bit of inertia to a market of that size).

    This, more than anything, explains Microsoft’s desire to lure hardware manufacturers into abandoning “legacy” support in a way that is “integrated” with (their) operating system. But, here again, they are hampered by their own legacy. They have many millions of lines of code written for the legacy systems architecture and they depend on their compatibility with tens of thousands of other products — essentially all of which are non-portable code.

    (Meanwhile Microsoft’s attempts to supplant the PC with some sort of “legacy free” platform have mostly failed — though most new PCs ship without floppy drives, and USB has largely supplanted old 9-pin serial and 25-pin parallel. The emergence of “ultra mobile” PCs is, in large part, a retrograde trend from Microsoft’s perspective, leaving them scrambling to trim down an XP product offering to address the threat of completely missing out on the opportunities that these ultra cheap systems are opening up along with vast new (low-end) markets).

    Perhaps, you understood all that, Mark. I realize you are a brilliant guy and perhaps it’s just that you didn’t have time to digress into these details when you had other points to make.

    For me the key conclusions to draw from these observations are:

    * Our greatest strength lies in the “agile discipline” of our upstream maintainers — whatever we can do to help them maintain their code in a “shippable state” at the end of each day contributes to our ability to ship the “latest and greatest” version of their code with our distribution.

    * Loose coupling is one of our key strengths — and this implies the need for dynamic integration (essentially meaning that any components which wish to offer integration with others must be capable of detecting and dynamically enabling those features while supporting graceful default behavior).

    * Automated test suites and streamlined feedback are the best contributions we can make back upstream. Launchpad and Apport are fine steps towards the latter. I look forward to learning about what Ubuntu and other community members are doing to encourage the former.

    Mark Shuttleworth says:

    I don’t think it’s cynical to say we can ship any time because we aren’t deeply attached to a particular feature – I think it’s exactly right, and it’s a real competitive advantage that we have over proprietary software. Dan Frye at IBM talks very eloquently about how the open source process is the first time we have managed to separate “R&D” from “Productisation”. In big software shops, people are supposed to be hardcore “upstreams” for a year, then “product guys” while things are getting put into shape for shipment. In the open source world, we’ve separated out the upstreams from the distributions, which means that we should always be able to ship in good shape. And with good branching support in our VCS’s, we can go further, and let the pure-upstream-feature guys do nothing but that, while other people care about releasing. So, I agree wholeheartedly with your point and don’t consider it cynical at all.

    Automated test suites really are the future, but that’s not a meme that has yet attracted a lot of attention in free software circles. Personally, I’d like to see a large test suite ship with every upstream component, and be run every time a package is built in Ubuntu or Debian.

  104. Richard says: (permalink)
    May 15th, 2008 at 8:55 pm

    I’m in the throes of trying to get my very reliable Feisty install upgraded to Gutsy and then Hardy. The upgrade to Gutsy was sufficiently hairy (web browsers crashing etc) that I’ve given up on the idea of upgrading and am now looking at a clean install. For some reason Hardy is also failing in a different way with Firefox 3 (still beta).

    Like others, I think the Ubuntu team really needs to focus above all on STABILITY and bug-fixing – there are far too many serious bugs at the moment, which is disappointing considering I built a brand new PC just to run Ubuntu Hardy. Really Canonical should have delayed Hardy by 2 months or so, as it did with Dapper 6.06 (originally planned as 6.04).

    This self-congratulatory posting about hitting the shipment date is really embarrassing, because it says nothing about quality – as with all projects it’s very easy to measure shipment on time, and much harder to measure quality, and ensure that the quality is right before you release the software – true of commercial software certainly. Most open source projects stick to the “only release final software when it’s ready” and are very, very stable. Ubuntu is more innovative and feature-rich than many open source products, but it mustn’t lose the insanely high quality that generally is a very attractive part of the open source world.

  105. free says: (permalink)
    May 15th, 2008 at 9:32 pm

    Use soft dates, not hard dates. Nobody likes the stress and the shortcuts that are needed to get the date, it creates bad spirits and many bugs. People will feel less free (paid and volunteers)… Besides everybody can wait a bit longer for less bugs. We all want to be proud not date pinned.

    Make your LTS bug free… Take as long as you can. Don’t do anything new with LTS -> be very conservative, use only what you learned and used before. LTS should be a dull release and very stable.
    8.04 is for me not as stable as a LTS should be. I can’t promote it in this state.
    You never should have given firefox beta in it and the other on the edge things. It makes people uncomfortable… People can’t install their needed plugins like “Firebug”
    Maybe when firefox 3 came out an opt in via the update process.

    And use the short term versions to do all the new funky stuf. I love it. I don’t mind bugs and understand that things need to move on.

    PS:
    Thanxs for Ubuntu… Its great and loving the freedom and possibilities from Free Software every day! That is for everybody that is working on the project.

  106. Ian says: (permalink)
    May 15th, 2008 at 10:06 pm

    Sorry but you’re *proud* of 8.04? On my computers 8.04 has been a serious regression. Perhaps if you learned, as did Microsoft about 10 years ago, that soft dates are better than hard dates you wouldn’t have created such a problem.

  107. Alan says: (permalink)
    May 15th, 2008 at 10:34 pm

    There are several very good reasons its a bad idea

    - Mirrors and update sites work rather better with the distribution release dates distributed

    - US publically traded companies have very strict rules on when they can co-operate and what they can communicate about unreleased products.

    - Nobody wants Linux distributions to end up like proprietary products where the “.0″ release is strictly for loons (or before service pack 2 for those not having .0 .1 …) because people rushed for deadlines.

    - You assume the distributions would do the same number of releases over the same time period – why ? A lot of big enteprise actually seems to want their computers to be like their other plant – ripped out every 10 to 20 years. Right now the computing world is way to fast for them

  108. Forrest S. says: (permalink)
    May 15th, 2008 at 10:38 pm

    Mark,

    I think this is a good goal that will benefit the entire community. I think one of the big issues with distribution release cycles now is they are fairly set on a 6 month cycle but not the same cycle. This does not line up with upstream software releases which are all over the place. I see the inclusion of Firefox 3 beta 5 in the 8.04 release as a prime example of this problem. I understand why a “beta” was included in the release but it is not optimal. It still crashes and many plugins are not compatible yet. Not being able to make Kbuntu a LTS because of KDE 4 is a similar situation.

    If distributions were more aligned with release cycles though then upstream projects could potentially and would likely also plan around those dates too. This would result in the latest and greatest released versions of software getting pushed out in distributions sooner without having to resort to including “beta” releases. Certainly not all projects would try to hit certain dates but I think more would. I think the best example of this working well now is GNOME. It is also on a 6 month release cycle so we know that the latest GNOME will always be released about a month before the next Ubuntu release.

    I doubt Firefox wants to switch from a feature development cycle to a time development cycle but I still think if they knew 3-4 of the top Linux distributions were all coming out in April/October (or whatever) then they could plan months and years in advance to hit that target.

  109. Ajay says: (permalink)
    May 15th, 2008 at 10:49 pm

    Ubuntu 8.04 seeems to work well for some, and not so for everyone.
    This is nothing new, but one would agree that some versions of distros are Milestone: quality and/or features,
    I would probably call for more RC (release candidates) versions to weed out major troubles.

    In my view any final-release version of Linux, should be based on quality, not rush, not schedule.
    But of course cannot wait for ever!

    I am ambivalent about the User’s interface (GUI): standardization is good… but kills creativity.
    So we do need both, possibly with different naming system.

    Synchronization seems both very laudable… and very impractical!

    == Common Linux Kernel+Device Drivers Installer ==

    Device Drivers, seems to be as much an issue for Linux as it is with Microsoft!
    I would love to see a simple an common Linux kernel+driver downloader/installer for all Linux Distros.
    Else H/W vendors will give up! How could they support twenty flavors of Linux.

    this was my 2c.
    Ajey

  110. Fran Taylor says: (permalink)
    May 15th, 2008 at 11:34 pm

    I think people do not understand the concept that you don’t release latest code. Releases have fully-formed feature sets and documentation, and they have been tested.

    Hitting a scheduled release date with two thumbs up from QA is a sure indication of quality. It means that the project managers have control over the features that they are introducing. It means that the developers can fix bugs on demand. It means that the customers know what to expect; they have a schedule too. Customers do not want new features, they want new working features, fully tested and fully documented. If you are going to earn a living in Free Software, you have to get someone to pay you, so you have to deliver what they want, on their schedule.

    There are a whole host of reasons why this is a great idea.

    I also think that it is a great testament to the quality of the software and the project management of the core Linux software, that someone can even make a suggestion like this and have it taken seriously.

    This is a real indication that Linux has become a mature, stable, and excellent product.

    I have no connection with this guy and I don’t even use Ubuntu, but I think this is a really excellent idea.

  111. Dan says: (permalink)
    May 15th, 2008 at 11:53 pm

    RE: “Free Software Syncronicity”

    As somone who believes entirely in the philosophies of free software, and someone who also happens to make a living architecting, implementing and rolling out large Linux and mixed-OS networks, the idea of synchronising various distros is appealing.

    I know many will oppose it for a wide variety of reasons. But for me it represents a fantastic way to convince the proprietary market to start looking at Linux as a platform of choice. While I dislike proprietary software on the whole, working with it is a necessary evil for myself and my customers. One of the big problems with the proprietary world is their lack of understanding in the variety of Linux distros, and that frequently when they “support Linux”, they mean that they “support Red Hat” (and quite often legacy releases of Red Hat), which mean all sorts of acrobatics are required to make the products work on a more progressive platform like Ubuntu.

    Again, the Stallman-esque philosopher in me would love a world built wholly and solely on free software, but the “cold turkey” approach won’t happen in the real world no matter how much we desire it. Taking baby steps like bringing a more unified platform to the top three Linux distros is a great way to make Linux in general a plausible target platform for more and more software.

    Again, I’m sure this will see negative feedback (the Slashdot crowd is already foaming at the mouth). Free software is about freedom, for certain, and distros (and users of course) have the freedom to use whatever toolchain and kernel they want. But speaking as someone stuck in the corporate arena, a predictable release cycle between the top three distros is a fantastic idea to make Linux an attractive platform for far more “corporate” software, and is something I’d love to see, even if it is for slightly selfish reasons of preferring Ubuntu to anything else, and desiring to see it listed as a supported platform by more software manufacturers along side Red Hat. :)

  112. Ben says: (permalink)
    May 15th, 2008 at 11:55 pm

    I think what your describing is very similar to what is being accomplish with UnitedLinux. While Caldera Systems/SCO changed to a flawed direction of trying to use threats as a form of making money, the UnitedLinux concept itself could have worked. One of the major differences is that all of the players in UnitedLinux (Caldera, SuSE, TurboLinux and Conectiva) is they also all standardized on the RPM packages format. This had several advantages for software developers/publishers. I like the idea that someone would restart an UnitedLinux style initiative but disappointed on a couple items with how it is being proposed now:

    1) There is no indication that switching Ubuntu to RPM is on the table as a possibility. The current proposal still seems to leave de-standardized package management among the members.

    2) Of the previous UnitedLinux members, which seems to me should be the most interested in re-forming an UnitedLinux like agreement, you have only targeted SuSE. The opinion of TurboLinux and Mandriva (new name of Conectiva) does not seem to matter to you at all. I doubt RedHat will even blink an eye so your two out of three only leaves Novell/SuSE and Debian.

    Either way, I wish you the best of luck in accomplishing your proposal.

    Mark Shuttleworth says:

    Thanks Ben.

    One thing I haven’t articulated very clearly is the difference between binary-compatible-same-package coordination, and “let’s freeze on the same major versions” coordination, which is different. I think it would be extremely constricting for the distributions to try to be binary-compatible (i.e. use exactly the same binary packages of core pieces). UnitedLinux took this approach, and a later effort with Debian (the “DCCA”) did the same, and I was convinced both would fail. They both did – some of course would say the DCCA failed because Ubuntu declined to participate, but I think it was just doomed.

    The key thing for me is SOURCE collaboration, not BINARY collaboration. I.e. distributions and upstreams do their real work with SOURCE code, and the thing we need to achieve is easy collaboration at that level. If we are both using a very similar version of X or OpenOffice, it’s a lot easier for us to collaborate on finding a fix to a particular bug (hell, we are much more likely to have the same bug!). But I would definitely leave plenty of room for distributions to build things differently, because that’s part of how they differentiate themselves. I doubt they would even ship all the same patches, because distributions have genuinely different goals, which is GOOD – they satisfy different users.

  113. Jean-Luc Livi says: (permalink)
    May 16th, 2008 at 2:00 am

    In principle, the idea of synchronising release dates is not bad. However, I still fail to see any benefit it might bring.
    On the practical side, the distributions that fell behind might be “compelled” to release under-tested code, or otherwise degrade the quality of the release.
    So, again on the practical side, I think synchronising should be the least of priorities. Instead, releases should focus on quality and standards compatibility (LSB, file system, etc).

    But, all in all, a BIG “thank you” to Mark and the Ubuntu team. You are one of the few “Good Things” in the world! :-)

  114. Laika says: (permalink)
    May 16th, 2008 at 2:22 am

    Zygo wrote:
    “The recent snafu with the Debian openssl libraries is an excellent example of why moving toward a monoculture is a bad idea. There’s no way to predict when a serious bug like this will occur; however, it’s easy to predict that it *will* occur eventually.”

    Suggesting that an extremely serious (and extremely rare) security problem like that would be “situation normal” in Debian is an outright insult. The Debian policy is deliberately designed to prevent problems like this from happening. People often say that Debian is “bureaucratic” and unflexible because Debian pays too much attention to the training of their developers and to the high quality of their releases. Nevertheless, the openssl bug did happen despite the strict and “bureaucratic” Debian policy. But you can rest assured that the Debian policy will be adjusted soon to make this kind of mistakes in Debian-specific patches even more rare than they were before.

    You should also consider that both Debian and Ubuntu reacted very fast after the problem was discovered. They didn’t try to hide the problem and they provided security updates that fixed the vulnerability. Erich Schubert’s blog ( http://blog.drinsama.de/erich/en/linux/2008051401-consequences-of-sslssh-weakness.html ) suggests that the consequences of this openssl bug are actually more severe in other distros than in Debian and Ubuntu, because the security update in Debian and Ubuntu blacklists the weak keys that were generated by vulnerable Debian and Ubuntu systems. But servers that use some other distro than Debian or Ubuntu may still grant access to keys that were generated on affected Debian or Ubuntu systems and are thus vulnerable to abuse even after the fixed Debian or Ubuntu systems are safe and sound.

    So you chose a poor example to demonstrate that moving toward a monoculture would be a bad idea. In fact, if all distros could immediately apply the same security updates that were provided in Debian and Ubuntu, then this openssl security problem would have been quickly fixed in all distros. But now the servers that use other distros remain vulnerable while the problem has already been fixed in Debian and Ubuntu.

  115. Mark Shuttleworth pide sincronizar la publicación de las distros GNU/Linux « Conocimiento Libre (o lo que está detrás del Software Libre) says: (permalink)
    May 16th, 2008 at 3:08 am

    [...] apunta en su blog que un calendario de lanzamientos conjuntos beneficiaría tanto a usuarios como a distribuidores. “Si dos o tres, Red Hat, Novell y Debian están dispuestos a acordar por adelantado una [...]

  116. postfw says: (permalink)
    May 16th, 2008 at 3:40 am

    In french or portuguese

    Mark, je vous remercie sincèrement pour tout ce que vous faites pour corriger le bug nº1.

    Quand la LTS 6.06 .0 est sortie en juin 2006 e non en avril, les 2 mois supplémentaires on été utilisés pour fournir une distribution stable e de qualité. Malgré cela nous avons assisté à des commentaires sur la fiabilité et aussi à une déception de certains utilisateurs. Une version 6.06.1 corrigée permit à Dapper Drake de faire un bon parcours par la suite. Pour la version LTS 8.04 la situation parait semblable. L´expectative e l’exigence des utilisateurs de Ubuntu était élevée e c’est comme cela a chaque fois que sort une version LTS.

    Deux idées:

    1. A l’image des deux autres grandes distributions Redhat e Suse, nous pourrions avoir deux entités différentes qui produisent deux Ubuntu différents pour des utilisateurs différents, les entreprises e les utilisateurs en général: “Canonical Ubuntu” d’une part et “Ubuntu community ” d’autre part. Personne ne se plaint, ou très peu lorsqu’il y a des bugs dans Fedora, personne ne se plaint de l’absence des technologies immatures dans “Redhat Linux Enterprise”. Idem pour “openSUSE” et “Suse Linux Enterprise”. Chacun obtient ce qu’il cherche dans son produit. Mark, j’ai peur que vous tentiez de faire la quadrature du cercle en tentant de satisfaire tout le monde avec un seul produit. Cette idée comporte certainement des défauts. Le premier, une augmentation des coûts a cause de la création d’une équipe de développement a l’intérieur de Canonical pour la version professionnelle. Vous utilisé déjà beaucoup de vos ressources pour soutenir Ubuntu, laissez moi vous en remercier. Le deuxième défaut c’est l’impact négatif sur la communauté Ubuntu: une version professionnelle certifiée pour les entreprises, est une version payante, pour certains cela va contre la philosophie Ubuntu.

    2. Plus simplement, je pense qu’une version LTS avec moins de révolution technologiques à l’intérieur serait bénéfique, finalement la stabilité est l’objectif poursuivi. C’est selon ce critère que les utilisateurs vont juger. Je sais que cela peut conduire a une distribution moins “sex appeal” par rapport a la concurrence et a la demande de certains utilisateurs avide de nouveautés. Pour eux il y a la version Ubuntu habituelle. C´est dans la version normale de Ubuntu qu’il faut faire le maximum d’avancées technologiques et les consolider plus tard dans la Ubuntu LTS.

    portuguese

    Mark, agradeço-lhe vos sinceramente por tudo aquilo que faz para corrigir o bug nº1.

    Quando a LTS 6.06 .0 saiu em Junho 2006 e não em Abril, os 2 meses suplementares foram utilizados para fornecer una distribuição estável e de qualidade. Mesmo assim assistimos a comentários sobre a fiabilidade e também a uma decepção de alguns utilizadores. Uma versão 6.06.1 corrigida permitiu depois à Dapper Drake de realizar um bom percurso. Para a versão LTS 8.04 a situação parece semelhante. A expectativa e a exigência dos utilizadores do Ubuntu eram elevadas e é assim cada vez que sai uma versão LTS.

    Duas ideias:

    1. Tomando o exemplo das duas outras grandes distribuições Redhat e Suse, nós podíamos ter duas entidades diferentes que produzem dois Ubuntu diferentes para utilizadores diferentes, as empresas e os utilizadores em geral: “Canonical Ubuntu” dum lado e “Ubuntu community ” doutro lado. Ninguém se queixa, ou muito pouco, quando há bugs no Fedora, ninguém se queixa da ausência de tecnologias imaturas no “Redhat Linux Enterprise”. Idem para o “openSUSE” e o “Suse Linux Enterprise”. Cada um obtém aquilo que procura no seu produto. Mark, receio que tente fazer a quadratura do círculo tentando satisfazer toda a gente com um só produto. Esta ideia comporta certamente defeitos. O primeiro, é um aumento dos custos por causa da criação duma equipe de desenvolvimento no interior da Canonical para a versão profissional. Você já utiliza muitos dos seus recursos para manter o Ubuntu, Deixe-me agradecer-lhe. O segundo defeito é o impacto negativo sobre a comunidade Ubuntu: uma versão profissional certificada para as empresas, é uma versão que se paga, para alguns isto vai contra a filosofia Ubuntu.

    2. Mais simplesmente, eu penso que uma versão LTS com menos revoluções tecnológicas no seu interior seria benéfica, afinal a estabilidade é o objectivo a alcançar. É com esse critério que os utilizadores vão julgar. Eu sei que isto pode levar a uma distribuição menos “apelativa” em relação à concorrência e à procura de alguns utilizadores ávidos de novidades. Para eles existe a versão Ubuntu habitual. É na versão normal do Ubuntu que é preciso fazer o máximo de avanços tecnológicos e os consolidar mais tarde no Ubuntu LTS.

  117. Markus McLaughlin says: (permalink)
    May 16th, 2008 at 4:41 am

    Three matters I like to bring up, One : Why isn’t there a method for attending a Linux Convention virtually on a web site? Example of what I mean is on my Linux blog : http://linuxglobe.wordpress.com; Two : Has anyone done a comparison of the recently released Fedora 9, Ubuntu 8.04, and openSUSE 11.0? I like to find other Linux Distro testers to evaluate all 3 major distros and have reviews available that I can add to my Linux Blog…. And Three : Why isn’t there iPhoto/iMovie/iTunes like software available for Linux??? Thanks, have a good summer, and keep on writing good code!!!

  118. Jim Dennis says: (permalink)
    May 16th, 2008 at 8:01 am

    Mark,

    I think you might have missed one vital point in your comparison of Microsoft and similar closed door product development to our OSV model (development of distributions).

    There is no secret to having predictable release schedules. You cannot reliably combine the that with an MRD (marketing requirements document) listing as yet undeveloped features. So long as all of your components are in a usable state (that is to say that each of them has some usable version available, even if it has to be some old “back rev.” then you can, with a bit of internal engineering discipline, maintain a development tree that can ship on any arbitrary day.

    In large part the Debian developers do this by virtue of the fact that most of them run their systems on “unstable” or “testing” and keep their systems upgraded to within a few days of the head of the project. Thus the Ubuntu team has many opportunities to select and pin down stable points with their proximal upstream sources.

    To put it cynically, Ubuntu can ship any time it wants because it doesn’t commit itself to shipping with any particular set of new enhancements.

    The other advantage that OSVs have over Microsoft is the loose coupling among the components.

    Microsoft’s problems include their own commitment to an MRD. They can’t ship until all these *requirements* are (allegedly) satisfied. Also Microsoft embraces and strives for a much greater degree of couple among components. They call this being “integrated” and savvy market analysts refer to it as “vendor lock in.”

    Right now Microsoft’s biggest competition is not Linux, nor MacOS, or any other outside product offering. Their biggest competitor is their own legacy. They have implemented far more features than any of their customers can use; so they can no longer market upgrades using “new features.” They have *sold* many millions of copies of their software for years past the time that they became aware of the need to change their revenue model into a subscription basis. (They have tried and continue trying, to shift models but there’s quite a bit of inertia to a market of that size).

    This, more than anything, explains Microsoft’s desire to lure hardware manufacturers into abandoning “legacy” support in a way that is “integrated” with (their) operating system. But, here again, they are hampered by their own legacy. They have many millions of lines of code written for the legacy systems architecture and they depend on their compatibility with tens of thousands of other products — essentially all of which are non-portable code.

    (Meanwhile Microsoft’s attempts to supplant the PC with some sort of “legacy free” platform have mostly failed — though most new PCs ship without floppy drives, and USB has largely supplanted old 9-pin serial and 25-pin parallel. The emergence of “ultra mobile” PCs is, in large part, a retrograde trend from Microsoft’s perspective, leaving them scrambling to trim down an XP product offering to address the threat of completely missing out on the opportunities that these ultra cheap systems are opening up along with vast new (low-end) markets).

    Perhaps, you understood all that, Mark. I realize you are a brilliant guy and perhaps it’s just that you didn’t have time to digress into these details when you had other points to make.

    For me the key conclusions to draw from these observations are:

    * Our greatest strength lies in the “agile discipline” of our upstream maintainers — whatever we can do to help them maintain their code in a “shippable state” at the end of each day contributes to our ability to ship the “latest and greatest” version of their code with our distribution.

    * Loose coupling is one of our key strengths — and this implies the need for dynamic integration (essentially meaning that any components which wish to offer integration with others must be capable of detecting and dynamically enabling those features while supporting graceful default behavior).

    * Automated test suites and streamlined feedback are the best contributions we can make back upstream. Launchpad and Apport are fine steps towards the latter. I look forward to learning about what Ubuntu and other community members are doing to encourage the former.

  119. Jim Dennis says: (permalink)
    May 16th, 2008 at 8:08 am

    Apology for duplicate posting.

    Mark,

    I got home after work, look at my browser and it, for all the world, looked like I had typed my whole message and forgotten to hit “submit.”

    (Then I immediately went looking back upwards to see what other posts had been made and saw that I had posted it, and that you have already responded to).

    Please feel free to delete the dupe or accept my apologies for cluttering things up for your readership.


    JimD

  120. Ed Daniel says: (permalink)
    May 16th, 2008 at 9:49 am

    “the opportunity to collaborate with the other, large distributions on a coordinated major / minor release cycle”

    Bang on! Great vision Mark, unite the leading distros through co-opetition allows complementary factors to influence the progress in adoption of Linux, which all distros benefit from. Thank you.

  121. Bithead says: (permalink)
    May 16th, 2008 at 12:30 pm

    Mark,
    Focus on what the OS does and how well it does it, not when its released. That’s the strength of linux – lever it. Microsoft software is of poor quality precisely because of that mentality – that superficials are more important than functional substance. Predictability is nice, but don’t rely on it – those who do invariably encounter endemic problems. Distros need to be independent.

  122. LTS de Ubuntu | mayeco::getHome( ); says: (permalink)
    May 16th, 2008 at 6:00 pm

    [...] En un post en su blog Mark Shuttleworth anuncia que la nueva version LTS sera Ubuntu 10.04 LTS ademas de esto anuncia que a partir de la [...]

  123. El próximo Ubuntu LTS llegará en el 2010, ¿o antes? « [ Balrof ] says: (permalink)
    May 16th, 2008 at 7:44 pm

    [...] 16, 2008 Ubuntu Mark Shuttleworth, cabeza de Canonical Ltd. y fundador de proyecto Ubuntu, hizo un llamamiento a los desarrolladores de otras distribuciones GNU/Linux para que sincronicen los lanzamientos de [...]

  124. Harald Hoyer says: (permalink)
    May 16th, 2008 at 8:14 pm

    Here is my comment: http://www.harald-hoyer.de/personal/blog/cross-distro-sync

  125. A. Peon says: (permalink)
    May 16th, 2008 at 9:26 pm

    I just have to come back to this one with a simple, painful observation — the difficulty of actually squeezing a stable, proven, working set out of the open source bazaar is why it takes Debian about a decade to cut a ‘stable’ release, and why Ubuntu found a niche in the first place; there was enough stable software that Debian hadn’t incorporated, or put into the distro “out of the box,” that it was a pretty big win to bundle some obvious, desirable, individually proven projects (X.org, OO.o, Firefox, and to a “lesser extent” Gnome and Linux) and make an instant desktop.

    What has me “worried” (scare quotes because I live with it just like we all do) is that it just might not be possible to get enough of those projects in a sweet spot to actually cut a release that works on a majority of hardware for a majority of basic workloads on any schedule, enforced by political efforts for synchrony or not. In particular, the sheer scope of the desktop environment and kernel projects means they tend to start refactoring themselves well before the previous refactorings even have a chance to become widely understood or stable. [This has been called the 'herd of attention-deficit teenagers' development model, but it happens with the big three-letter companies as well, especially when those companies apply, retract, and reapply resources to FOSS work.]

    I don’t think Ubuntu is free to make massive changes, but is it possible to produce any sort of stable desktop OS with that level of complexity? [Observe that even MS is making an effort to simplify the core for the next round.]

  126. ArtemusGordon says: (permalink)
    May 17th, 2008 at 12:11 am

    Sorry Mark, 8.04 just wasn’t ready when you rolled it out the door, you need to call Terminix for this one. 8.04 reminds me of another Very Large OS that is NEVER ready when it’s rolled out the door, it’s not even ready when the next version comes out and I think you know who in Redmond I mean. Speaking of rolling, instead of trying to get the “Big Three” mentioned above in your blog to agree to a cycle timetable, maybe it’s time for Ubuntu to consider a rolling release system like Gentoo, Arch, or Sidux. That way, the LTS would be permanent, because we would always have an up-to-date system without having to wait for the next release cycle. And also, the bugs would be squashed before they were released into the wild ;) .

  127. Synchronizing releases · DragonFly BSD Digest says: (permalink)
    May 17th, 2008 at 1:27 am

    [...] Shuttleworth, the wallet behind Ubuntu, described on his blog a desire to see major Linux distributions on a common release schedule, so that major releases of associated software can match up.  (via)  This would be useful for the [...]

  128. Tectonic » Synchronise Linux releases, says Ubuntu head says: (permalink)
    May 17th, 2008 at 4:13 am

    [...] talked about his desire to see distributions synchronise their release dates and now in a post on his blog he writes: “If two out of three of Red Hat (RHEL), Novell (SLES) and Debian are willing to [...]

  129. Ubuntu Release, Maintenance and Support Schedule : mypapit gnu/linux blog says: (permalink)
    May 17th, 2008 at 1:12 pm

    [...] more information read Mark Shuttleworth post, The Art of Release Bookmark this article These icons link to social bookmarking sites where readers can share and [...]

  130. Janne Karhunen says: (permalink)
    May 17th, 2008 at 6:46 pm

    Release alignment with major distributions is a GREAT idea and should be implemented ASAP. This would guarantee high quality releases from all vendors as combined bug hunting efforts would benefit everyone.

  131. Dag Wieers says: (permalink)
    May 17th, 2008 at 11:41 pm

    I wrote an article where I play the Devil’s advocate to understand why Canonical would want this. You can find it over here:

    Ubuntu’s need to catch a wave
    http://dag.wieers.com/blog/ubuntus-need-to-catch-a-wave

    Disclaimer: ability to take off Ubuntu hat (and maybe wear Red hat) required when reading :-)

    Mark Shuttleworth says:

    Thanks for your contribution to the debate. I liked the perspective of one of the commenters, who points out that Ubuntu could very easily align with Red Hat, just as CentOS does. Instead, we choose to collaborate most closely with Debian and GNOME + KDE. That’s simply because we love the Debian approach and feel that contributing to Debian is one of the most efficient ways to get work into the upstream bloodstream, so to speak. And GNOME and KDE are the most visible platforms – GNOME’s commitment to a regular release cycle made it possible for us to do the same with Ubuntu.

    I believe Ubuntu will succeed on both desktop and server, regardless of whether the other distributions take up the offer of coordination and collaboration. I think that collaborating and coordinating makes ALL of the distributions more powerful with regard to Windows and the Mac. Us Linux lovers sometimes forget how fractured and fractious the Linux landscape looks to someone who is used to Windows and the Mac, so anything which reduces the pointless friction and increases the potency of Linux as a whole is good for all of us.

  132. Folle Rec says: (permalink)
    May 18th, 2008 at 1:15 am

    Hi Mark,

    So you’re talking about synchronizing release schedules and version numbers? And that’s it?

    I find your comparison to car dealerships and fast food courts to be wrong.

    People shop for cars. They go to many car dealerships. These days the schedules of manufacturing releasing new models isn’t in sync. Some release on a 4-year basis, some 6 years, some 8-10 years. So, while a new Honda Accord was just released a few months ago, a new Toyota Camry might not be released until a few more years. Same stuff, different release schedules.

    Fast food clustering is also a flawed comparison. Yes, they are clustered together. But they are all/most different from each other; different in pricing and different in food category and in cuisine. Same stuff, different versions.

    So again… what are you talking about? That everyone agrees to release a distribution with Gnome 2.24 on October 2008? Not everyone finds Gnome 2.24 as stable on initial release. You might be willing to support that. Others might not.

    I suggest that Ubuntu focus on developing easy to use but complete server configuration tools for the desktop. Then just use RHEL’s srpms and recompile those for your enterprise needs. That way you’re synced with RHEL’s release schedule and version, are stable as well, but give people a reason to switch to Ubuntu.

    Mark Shuttleworth says:

    Yes, in other industries companies “cluster and differentiate”. And that’s exaclty what I think will happen with Linux. We will not aim to achieve the same things for our users, we will aim to differentiate. But the things we choose to differentiate on will be MUCH more meaningful than “kernel 2.6.24 vs kernel 2.6.25″. When you agree to collaborate on the infrastructure you have an increased opportunity to differentiate on what you DO with it. So Fedora should work on just as many computers as Ubuntu does (I would be delighted!) so that Fedora and Ubuntu can differentiate on the OTHER qualities that they have.

    Food halls and car dealerships are clustering by location, but you also have examples of clustering by time. Think of CES – the Consumer Electronics Show. Huge numbers of products across many different categories in electronics are all announced at the same time. They all benefit from the increased attention, coverage, interest and so on. Yet they have plenty of opportunity to differentiate.

  133. i_c-Y says: (permalink)
    May 18th, 2008 at 4:43 am

    Oh come on… you’re just trying to take work from RHEL and Novell who put time and money into this and drag it into ubuntu. Bleh. It’s not a vision to unite distros its just to make ubuntu get free stuff. I hope this doesn’t happen. A sediment also shared on Dag’s blog (http://dag.wieers.com/blog/ubuntus-need-to-catch-a-wave)

  134. Captain Larry Dart says: (permalink)
    May 18th, 2008 at 9:13 am

    It’s been said here already but I’m going to say it again. Giving yourself a pat on the back for meeting a promised delivery schedule is one thing. But if what was delivered wasn’t ready for delivery then I don’t see you’ve got anything to be proud of. Meeting schedules and delivery dates at the cost of product quality doesn’t help anyone. “”Enterprise platform” release”? I don’t know about Ubuntu 8.04 server but if you’re referring to Ubuntu 8.04 desktop then you must be, as they say, “having a laugh”. If 8.04 is enterprise quality then I’m Bill Gates! 8.04 looks and performs like what it is, a release candidate hurriedly cobbled together to meet a deadline. It’s state on release is a strong argument against what you are proposing.
    Enterprises want stability from day one, not, as an example, beta release web browsers. They’re not interested in the fan boy bells and whistles. They need stuff that has been tested and that works. They don’t get excited because “it’s the latest release”. Bleeding edge to them equals risk and risk equals potentially lost revenue. Red Hat don’t call Fedora an enterprise platform for that very reason.
    Synchronise release schedules by all means. I agree that it would be a beneficial thing to open source and Linux, your point about clustering in a marketplace is excellent. But make sure that what is released as a result of this “synchronisity” is stable and ready. Leave the bleeding edge stuff out.

  135. Jupiter Broadcasting » Blog Archive » The Linux Action Show! Season 8 Episode 8 - MP3 says: (permalink)
    May 18th, 2008 at 7:48 pm

    [...] appear on every Asus motherboard Fedora 9, released! VIA Releases 16K-Line FOSS Framebuffer Driver Mark Shuttleworth: The Art of Release Compiz-Check      The Linux Action Show! Season 8 Episode 8 – MP3: Play Now | Play in [...]

  136. ProjectX Blog » Blog Archive » Xlinks digest - 19 / 05 / 2008 says: (permalink)
    May 18th, 2008 at 10:51 pm

    [...] Shuttleworth calls for co-ordinated release cycles Added on 05/16/2008 at 09:28AM [...]

  137. Sean says: (permalink)
    May 19th, 2008 at 2:43 am

    I feel like my experiences with Ubuntu have been very different from posters here. I’ve been running Debian for years. I tried 4.10 and didn’t like it. It wouldn’t install on my old pentiums and had some issues that I can’t remember anymore on my main desktop. I’ve been running Ubuntu at work since 7.04, thanks to Dell. 7.10 had a few minor annoyances, but ran nicely. 8.04 has been great so far. I tried watching YouTube videos and just now realized that the issue wasn’t my fault. It’s a work computer though, so flash not working right isn’t really a priority. The important things to me have been fixed compiz bugs. compiz in 7.04 was blacklisted for me, and when I got around that, it crashed a few times a week. Basically 8.04 DOES feel more polished to me, despite perhaps having some more beta quality software included. The OS as a whole seems a step forward. Even Firefox 3b5 seems more stable than the version that was in Ubuntu 7.10. I still run Debian at home, but for my desktop at work, I prefer slightly new software than Debian releases, and won’t run Debian unstable. Thanks.

  138. The Linux Mint Blog » Blog Archive » The Linux Mint newsletter issue 46 says: (permalink)
    May 19th, 2008 at 10:48 am

    [...] Mark Shuttleworth talks about the Ubuntu release cycle and a possible collaboration with the other, large distributions on a coordinated major / minor release cycle http://www.markshuttleworth.com/archives/146 [...]

  139. Christoph says: (permalink)
    May 19th, 2008 at 12:02 pm

    To the best of my knowledge there has never been an “enterprise platform” release delivered exactly on schedule, to the day, in any proprietary or Linux OS.

    This is wrong. There were some RHEL releases in time as well as CentOS versions.
    On the other Hardy was a little late, wasn’t it?
    So what is the benefit for Red Had or Novell to agree on a schedule with canonical?

    Mark Shuttleworth says:

    Hardy was on time. Dapper, 6.06 LTS, was two months late. We think the lessons from Dapper came through in Hardy, which is why Hardy went out on schedule, with point releases to address issues that arise.

    The benefit to Red Hat and Novell (and Debian and Gentoo and other distributions) of a coordinated *freeze* would be commonality of patches and drivers and documentation. When an end-user Googles for “how do I get X done” they will find solutions which are more likely to be relevant to them, whichever platform they are using. Ubuntu has more users than any of the other distributions, so I think that huge volume of documentation, comments, hints, suggestions and so on would be very valuable to them! We want to share the benefits of everything we do. It’s clear some people think my suggestion of coordinated freezing would benefit Ubuntu at the expense of other distributions – I don’t feel that way at all. I think that it would be a strong mutual benefit, and I think Ubuntu and Canonical have a great deal to add to this sort of collaboration.

    Other distributions have criticised Ubuntu for “not contributing”. That’s amazing to me, anybody who is close to the production of Ubuntu knows that a huge amount of real work gets done in every release which results in a large volume of patches, fixes, documentation, translations and so on, all of which we make every effort to contribute to other distributions like Debian, and upstream. I think anybody who actually looks at that work in aggregate would be delighted with the contribution Ubuntu makes. And all of that would be more accessible to other distributions if we coordinated our freezes.

  140. Shuttleworth: sincronizziamo i rilasci delle distribuzioni | FDS says: (permalink)
    May 19th, 2008 at 12:28 pm

    [...] pagine del suo blog, Shuttleworth sembra avere le idee ben precise su come migliorare la situazione dei rilasci [...]

  141. Aaron Seigo says: (permalink)
    May 19th, 2008 at 11:06 pm

    @Mark: ‘Have you ever noticed that competing car dealerships, fast food restaurants and other very similar businesses all setup shop next to one another? You get food courts in shopping malls, for example, where you have all the take-away food places in one area. You would think that the competition would be bad for them. But counter-intuitively, all the restaurants do better when they are all in the same place, and the same is true of car dealerships. The phenomenon is called “clustering”, and it works because people first decide to go “looking for a car” and then later decide which car, or which dealership. I think Linux is the same’

    The success of retail clustering, which is what you are referring to here, has to do with the customer’s (perceived or self-alloted) time budget and how it is impacted by geographical distance. It really has very little to do with people looking to buy a type of product first; people go to where they are relatively sure they can find the product, and the success of that increases with retail density. Consumers tend to economize their time, travel and budgets within the context of their lifestyle and ingrained habits, resulting in clustered areas being generally more attractive to them.

    In this sense, there is pretty much zero correlation between why the concept of clustering in retail works and anything in the Linux market.

    “I don’t think Red Hat would see Ubuntu as a threat, we appeal to different audiences.”

    This is interesting, however, as it speaks to intra- rather than inter-seller competition: Canonical and Red Hat are different types of companies selling similar products. As long as this remains true, then there might be perceived value for Red Hat here, probably predicated on two things however:

    * Canonical remains a non-direct competitor; given [K]Ubuntu’s deployments at places like Google or the French parliament, this point is perhaps already debatable.

    * Canonical brings something to the table of value for Red Hat. This could come in the form of engineering effort, growing the user population of a uniform base (making it more attractive to ISVs and IHVs as a conglomerate) or as a possible marketing or sales channel. Knowing a bit about Red Hat’s corporate culture, I think convincing them there is value on the table here will be tricky .. but could be doable with enough of a show of commitment.

    You can replace “Red Hat” for any other entity in the above as well; the dynamics will be similar (though the conclusions on perceived competition and culture will differ).

    So really what you’re talking about here is not clustering at all, but creating a strategic partnership, a cooperative or a “virtual conglomerate” (depending on how you wish to look at it). Supply chain management (SCM) probably has a lot more relevance to this conversation than retail location geography. You may find more compelling reasons for your concept within SCM.

    Also note that this is not totally unlike the concept that initially drove initiatives like United Linux; it may be of interest to examine where they fell short as well as how they got off the ground in the first place.

  142. Six Annoyances in Hardy Heron Ubuntu says: (permalink)
    May 20th, 2008 at 8:13 am

    [...] to shake out bugs in the latest incarnation of GNOME. Maybe it will take a while to find the right pulse rate of software releases. If there’s a plan by Ubuntu/Canonical developers to improve this situation, I encourage them [...]

  143. Six Annoyances in Hardy Heron Ubuntu | Search Engine Optimization says: (permalink)
    May 20th, 2008 at 8:27 am

    [...] to shake out bugs in the latest incarnation of GNOME. Maybe it will take a while to find the right pulse rate of software releases. If there’s a plan by Ubuntu/Canonical developers to improve this situation, I encourage them [...]

  144. Simon says: (permalink)
    May 20th, 2008 at 8:58 am

    Hi, I recognise that what mark has done is excellent work, and ubuntu is great. However, i would love to see Kubuntu and ubuntu collaborate
    more in depth. Kubuntu is a joke compared to ubuntu, its suppose to be by the same company and supported/maintained equaly, but this
    is definately not the case. Kubuntu hasn’t improved much since i’ve been using it, (3 years now) which made me go for another distro, (linux mint)
    Although ubuntu gnome is great, most windows users make fun of it because it looks ugly and unproffesional, (disagree if you like, but its a fact)
    A major theme change would be what ubuntu needs to bring over more windows users.
    myself, i use Kde Linux mint. But im appreciate what ubuntu has done, as it is the base for the distro i use. but in my opinion, much better.

  145. deuts says: (permalink)
    May 21st, 2008 at 5:42 pm

    I too have to agree with the many commenters here about Hardy being released too early. It wasn’t just ready considering the bugs it contains out of the box. For me also, I would rather wait for a late release than deal with a prompt release with too many bugs to the extent that I can hardly use my computer anymore!

  146. Lettre Hebdomadaire Ubuntu numéro 91 du 11 au 17 mai 2008 « Lettre Hebdomadaire Ubuntu says: (permalink)
    May 21st, 2008 at 10:10 pm

    [...] http://www.markshuttleworth.com/archives/146 [...]

  147. Luis Jacquez says: (permalink)
    May 28th, 2008 at 8:24 pm

    I said people should not be greedy and enjoy their Linux Operation Sys. because I Know the 98% of the user they did not pay a single penny
    for the OS on their PC,If you want to help submit statistical information on their linux OS flavor and have fun getting your hands dirty on your desktop or laptop if you don`t have time to play with it then buy a Windows M$ or machintos apple pc and shut up!

  148. Ubuntu 8.04 and Onwards - Release, Maintenance and Support Schedule « Linux and Open Source Blog says: (permalink)
    May 28th, 2008 at 9:12 pm

    [...] Standard releases will continue to receive patches and updates for one and a half-year and Server LTS release will continue to receive updates for 5 years. For more information check out Mark Shuttleworth’s post. [...]

  149. Shuttleworth Proposes Synchronized Linux Releases | Networking for Networkers says: (permalink)
    May 29th, 2008 at 3:54 pm

    [...] by design, some Ubuntu releases are supported longer than others. He also makes an interesting proposal: to synchronize the releases of the various major Linux distributions. That would give all the [...]

  150. ArsGeek » Six Annoyances in Hardy Heron Ubuntu says: (permalink)
    May 31st, 2008 at 7:36 am

    [...] to shake out bugs in the latest incarnation of GNOME. Maybe it will take a while to find the right pulse rate of software releases. If there’s a plan by Ubuntu/Canonical developers to improve this situation, I encourage them [...]

  151. Tectonic » Is Firefox 3 ready for prime time? says: (permalink)
    June 2nd, 2008 at 10:03 am

    [...] we see big-name software projects announcing how pleased they are to have met their planned release schedules. It’s something that Ubuntu did with its latest release, Hardy Heron, which has been met by a [...]

  152. Ubuntu Release Management, RedMonk story, MySQL Community: links 07-05-2008 | Commercial Open Source Software says: (permalink)
    June 7th, 2008 at 8:33 am

    [...] The Art of Release – Mark Shuttleworth on Ubuntu release management. [...]

  153. Free Nulled PHP Scripts says: (permalink)
    June 8th, 2008 at 1:46 am

    [...] to shake out bugs in the latest incarnation of GNOME. Maybe it will take a while to find the right pulse rate of software releases. If there’s a plan by Ubuntu/Canonical developers to improve this situation, I encourage them [...]

  154. Ubuntu is THE ‘Linux Home Desktop’ distro; now onto the desktop wars « Limulus says: (permalink)
    June 11th, 2008 at 9:03 am

    [...] more!  Looks like they’ve won the high-end of the market.  Also of note, Ubuntu has set a time schedule for LTS point releases (the first comes 3 months after a LTS release, followed every six months [...]

  155. Ubuntu 8.04.1 due July 3! « Limulus says: (permalink)
    July 2nd, 2008 at 10:55 am

    [...] is Hardy with all the updates since 8.04 was released three months ago; the .2 release will be in six months (three months after 8.10 is [...]

  156. L’art de gérer les versions says: (permalink)
    July 9th, 2008 at 7:26 am

    [...] française de l’article “The Art of Release“. Auteur : Mark Shuttleworth – Traducteur : Bernard [...]

  157. L’étrange déclaration de Monsieur Shuttleworth « Zitrouille s’installe ici says: (permalink)
    July 27th, 2008 at 7:11 pm

    [...] et à son marketing habituellement mieux huilés. Mais ses récentes déclarations, dont celle sur les cycles de développement des projets libres, tendent maintenant à provoquer des réactions assez vives, quoi que certainement démesurées et [...]

  158. Ubuntu Linux OS Version Numbers - Tracking The Progress Of Ubuntu Linux says: (permalink)
    August 19th, 2008 at 1:56 pm

    [...] You can go to the blog (web log) of Mark Shuttleworth (the founder of Ubuntu Linux) and learn a more about the future release dates of Ubuntu Linux: [...]

  159. Mark Shuttleworth pide sincronizar la liberación de las distros GNU/Linux | Proyecto web says: (permalink)
    October 5th, 2008 at 8:04 am

    [...] Hat, Novell y Debian decidan cooperar en un lanzamiento sincronizado en una fecha diferente. Mark apunta en su blog que un calendario de lanzamientos conjuntos beneficiaría tanto a usuarios como a distribuidores. [...]

  160. Et l’on reparle des cycles de release d’Ubuntu … — Hippopota.me/blog : le blog de deK says: (permalink)
    December 9th, 2008 at 7:14 am

    [...] Mark Shuttleworth a publié aujourd’hui un billet sur son blog, intitulé “The Art of Release”. [...]

  161. La próxima versión LTS de Ubuntu será en Abril de 2010 « ZdeS - Zona de Sistemas says: (permalink)
    January 13th, 2009 at 12:23 am

    [...] por | Mark Shuttleworth – en inglés [...]

  162. Is desktop Linux too fragmented to succeed? « BREAKING IT NEWS FOR BUSINESS says: (permalink)
    April 28th, 2009 at 1:36 am

    [...] Shuttleworth would like to go a step further. He has called for the major Linux distributions to synchronize their release schedules so that different flavors of Linux release new versions on a single, predictable timetable. That [...]

  163. Instalation on Ubuntu 9.04 - Zimbra - Forums says: (permalink)
    May 15th, 2009 at 7:25 am

    [...] Posted by d2globalinc So who knows if 10.04 will actually be an LTS. Somebody does: Mark Shuttleworth Blog Archive The Art of Release [...]

  164. L’art de gérer les versions says: (permalink)
    May 16th, 2009 at 6:58 am

    [...] : The Art of Release – Traduction : Bernard [...]

  165. juhuslikud mõtted » Blog Archive » Ubuntu magic: releases says: (permalink)
    September 1st, 2009 at 10:49 am

    [...] The Art of Release (2008-05-12) [...]

  166. Ubuntu Seek | Próximo ubuntu 10.04 será lts says: (permalink)
    September 22nd, 2009 at 5:27 pm

    [...] markshuttleworth – the next LTS release of Ubuntu will be 10.04 LTS [...]