 It's on silent, and not to kill your laptops from the desks. Who cares about what a room coordinator has to say about the speaker? It's quite obvious a lot of people care about Debian, so I'll hand it over straight away to Stefan. Will you join me in welcoming him? So thanks everybody. It's a real pleasure to be here. It's my first time in Australia and my first time at LCA, and I must say I'm really much enjoying the conference. So the kind of teasing title of this talk is because there are quite some rumors about Debian and the role of which Debian is playing in today's ecosystem, and my answer to that question is that you should care about Debian because quite some parts of today's free software ecosystem depends on Debian, even if you don't necessarily know about that. So I'm trying to convince you of this thesis, clearly I'm a bit biased on this topic being the Debian project leader, but I'll try very hard to hide that fact from you during the talk. So to speak about Debian is often useful to start from the beginning. So there are a couple of historic facts which are relevant to this discussion. So let's start from how Debian started. So the small paragraph of text you see at the top of this slide is actually the message that Jan Marduk, the founder of Debian, sent to a news group, back then we used heavily news group to communicate, called CompUS Linux Development. And in that message, Jan was announcing the imminent compression of a new Debian Linux release called Debian. So I think it was a bit optimistic because we nowadays know that free software distribution are essentially never complete. So 17 years later Debian is not yet complete, it will never be. So Jan was defined a bit optimistic back then. But in that message, you could find some of the main reasons for creating Debian and for having Debian around. So one of the first reasons, one of the first motives to have Debian back then was to create a non-commercial distro but which was still able to compete in the commercial OS market. So that was the prime motive to run Debian, to create Debian, sorry. And Jan wanted to have something easy to install, of course the criteria for being easy to install back then are not quite the same we have today. So I don't think Jan was thinking back then to have a CD put in your reader and ten minutes later have a system installed. I don't think he was thinking of that back then. And another point for him was to actually have a system which was built collaboratively and collaboratively by a group of experts which were experts in every single package that we're putting in the distribution. So the idea was that if you're maintaining a package, well, then you know a lot about it and you are an expert of the software you are delivering to your users. And finally, another criteria for creating Debian back then was to actually have the first major distribution which was actually run in an open manner in the spirit of the GNU project. And actually Debian has been GNU supported for a while in its early existence. So this is how all it started and then there are two more important events in the history of Debian which are relevant for this discussion. So the first one is the creation of the Debian social contract in 1997. So this social contract is an agreement between the people who make Debian, so Debian developers in general, and not only Debian users but the free software community at large. And in that contract you find four main points. So the first one is a commitment of having Debian as a purely 100% free operating system. The second one is a commitment to give back to the free software community, meaning that every change you do in Debian, well, you commit yourself to giving it back to the upstream author so that everyone can benefit from the change you have made in Debian. Then there is the spirit of openness which Debian has been joining since the beginning which is defined as don't hide problems. And finally, the document defines the priority of Debian which are essentially true. So free software on one side and users on the other side. So this kind of document, the social contract is still in effect and it's what binds together people who make Debian and user of Debian. And the other important point in Debian history which is relevant for us today is the Debian constitution which has been created for the first time in 1998. And what the document is essentially a set of structures and roles to govern a free software compatible democracy. So why I say free software compatible, well, because we don't really have to go through the burden of voting on every single change, we want to enable people to work as they like in open source and free software and do whatever they want unless at some points the community feel the need of voting on something. So that is what the Debian constitution enabled people to do. So there you find decision procedures and I'll get back to that shortly. You will find how democracy is used in Debian and you will find some scaffolding like the Debian project leader or the secretary to actually run votes. So this is in starting from 1993 up to now and 17 years later the landscape is a bit different. So we now have in Debian something like 30,000 binary packages is probably the largest free software distribution existing today. And in 17 years we have released 11 times and the 12 times Debian squeeze is going to be released most likely next in the weekend of 5 and 6 of February. The project has been growing a lot. So we are now today 900 Debian developers, 100 Debian maintainers which are a different technical role in the project, plus thousands of contributors like porters, translators, battery alger and so on and so forth. And we also have in Debian the largest number of ports among mainstream Linux distribution. So we had 12 for Debian landing and starting with Debian squeeze you are going to have two, for the first time, two non-Linux ports. So we're going to have two ports based on the free BSD kernel blend with a new user land. And finally we have that in all these years something like 120 other distribution which has based their work on Debian. And well if you ask me this is a pretty good success. So it's quite impressive considering where Debian started from. Nonetheless the landscape has completely changed. So back then there were like two, three Linux distribution and today there are kind of hundreds of them. So if you look at source forage you find hundreds and hundreds of Linux distributions and Debian is just one of them. So there are many. So why should one choose Debian? Why should one care about Debian in the first place? Especially considering that some of those other distribution have very nice features. For instance some of them release more frequently like every six months. Some of them have more users. Some of them innovate more, whatever that means. Some of them get more credit, press, you find news about them in the newspaper and this kind of stuff. Or whatever other food you might, FUD you might have about Debian add it to that list. So the question is who the bloody hell cares about Debian anymore? Okay, thanks. And more importantly, oh come on, go to another talk then, you're useless here. So is Debian still relevant? And yes it is. I think it is, but I don't want you to take my word for that. I want to explain why it is the case. So there are two main arguments for that. The first one is that Debian does it better. There are some features which you find in Debian and which you won't find altogether in many other distributions, possibly not at all. So the first one, there are four of them. The first one, in my opinion, is package quality. So Debian has a focus on the quality of every single package in the archive. So I like to call it a culture of technical excellence which appears in various forms. So for instance, we have a clear definition in our EGOS document called the Debian policy of how a package should look like. So in that document, you find the definition of where any single file you might want to ship in a package should be installed, for instance, where you should put documentation, where you should put images, where you should put other kind of artworks. And this gives to users a sense of consistency through all the operating system, which is quality. A second form of this quality you can find in Debian is an extensive package testing. So every package is run through a set of an automatic tester called Lincoln, which actually checks whether your package adheres or not to the policy. And then we do a lot of other testing, like stress testing of package installation. So periodically, we test all the package in the archive to see if you can install or not install in all possible situations, trying to ensure that you will always be able to install a package provided its dependency are satisfied. And then we have periodic master build of all the packages in the archive. So in every single stable release of Debian, you have the guarantee that every single package in the archive can be rebuilt from scratch using the other package in the archive. And this is more important than why I think. Because what is the point of FreeSofter if the ability to change it and rebuild your own package is just theoretical? So unless you can really change a package and have the ability to rebuild it, well, then you have FreeSofter in some way, but you cannot really enjoy the benefit of it. For us, this is a very important quality point. And then we still have the criteria that a package maintainer or a team of people maintaining a package are generally expert in the package they maintain. That means that when you ask questions about the package, you can interact with people which know the details of the software and not only details of packaging. Imagine scientific software, for instance. It makes quite a hell of a difference if you can interact with someone who knows what you're talking about and not only just how to package software. Finally, all packages in the Debian archive are equal. So there is no distinction about first class and second class packages. So the kind of support you get for a Debian package is the same, no matter what the package is. And all this is summarized quite well in the Debian release mantra, which is re-release when it's ready, which actually means that for us, what is more important is the quality of the packages, rather than some artificial debt that you might want to put on your release cycle. And so when I told you before that on the weekend of February 5th and 6th, we are going to release Debian squeeze, well, we hope we're going to do that. And I'm quite confident we will. But if the day before that we will find some release blocker, well, the rest assure that we will not release and fix that before being able to release. So this is the first point, package quality. The second point is, well, maybe quite obvious, but it is freedom. And here I mean software freedom because the social contact is still in effect and guarantees that there is a specific tie based on specific software freedom principles that tie together users and developers. And we have been doing that. We have been promoting the culture of C3 software since 1993. And Debian is free the bottom up. What does that mean? Well, all the software that you get via Debian, it's free. Firmware included started with squeeze. We moved away all the non-free firmware bits from the kernel to non-free. So by default in Debian, you will get only 100% free software. But not only that. Also the infrastructure we use to make Debian is completely free software. So in Debian, no one will accept to give users tools to do bug reports, for instance, which are not free. Nobody in Debian will accept to use a piece of our infrastructure like BuildDemon, which is not free. This is really not acceptable. And users in general are aware of this choice of ours and are happy about them and trust Debian not to betray those principles of free software that are written in our social contact. And all this in the ecosystem as a role of setting a very high bar for free software advocates. So Debian is a kind of a benchmark. If you see if you can do better or not than Debian in defending software freedom, well, this is an important role that Debian still play. Third reason is independence. And I care a lot about this. So what does it mean to be an independent project like Debian? Well, there is no single company which is actually behind Debian. So Debian is a community distribution but it's not associated to any single company. That means that all our means are donations of hardware or of money. Sometimes they come from company but they came from different companies. And most of the time come from users which actually want to support Debian and donate money to us. And everything else is a gift economy based on exchange of work among volunteers. This is very much remarkable. So think of all the, I use the term mainstream distribution, which is not really meaningful. But imagine the most popular distribution you can find out there. Well, most of them are one way or another run by companies or supported by companies, by specific companies. Debian is not the same. In Debian we are very much more independent from the need of gaining money of some company. And that means that people trust Debian choices not to be driven by profit. Fourth and last reason why Debian does it better is decision making. So a lot of people make fun of the way they can make decisions but it's not true that we vote on every single thing, actually. The basic criteria for deciding in Debian is what I like to call a do-work-racy. So do-work-racy means that if you are doing something, if you are responsible for something, as long as you work on it, you are free to decide to do whatever you want on that specific project area. Usually it is a package. So if you are responsible for a package, you can do whatever you want as long as you are working on it, okay? And the second point, which is a second class way of making decisions, is democracy. So when that fails, when there are conflicts, well, we might want to resort to democracy to take decision by a vote in the project. So what does that mean? Well, the reputation in the project follow works. So if you are someone well known in the project, that means that you are doing a lot of work for the well-being of the project. It means that there is no benevolent dictator in the project, not even me, I'm just a kind of coordinator of activities which happen in the project. And there are no imposed decision by who has money, by who has the infrastructure, or by who is employing people to work on Debian. So this is the first argument for me to say that you should care about Debian, because it plays a very important role in the free software ecosystem. And so I think that Debian should live long and prosper. Freedom and independence are very good, and I'm kind of scared at thinking at a distribution market in which every single distribution is run by companies. Because one day, their interest will clash with our interest. And we have seen recently the example of companies which used to be very good friends of free software, changing their mind on being bought by other companies and not being very good friends of free software anymore. Every reference to recent event is completely... Yes, absolutely, a science fiction, I would dare to say. And also having Debian around is a kind of... It's a good example even for distribution run by company to actually invite those companies to enable every passing day more and more their distribution to run their own boat. Okay, so this was my first argument for the relevance of Debian. The second one, if you are a more pragmatic type of guy, type of people, it's Debian is the root of a high tree of derivative distributions. So what is a derivative? Well, a derivative essentially saw a very big software project in which you apply software freedom to a whole distribution. So you take an existing distribution, you add some packages, you change something, you patch whatever you want to patch, and you rebuild your own package and periodically you sync with your mother distribution. So this is the idea of a derivative distribution. And derivatives have changed the game of software distribution in the past, I would say, 10 years. Because with derivatives, the focus of a software project moves to customization. So all the people, power you will need in your derivative distribution will be on customizing it for your specific goal. Everything else you can reuse from the distribution you descend from. And if this is done properly, that is very good. Everybody wins. Why? Well, derivative gain a lot in terms of work they can reuse. This is a clear advantage. And the so-called mother distro gains a lot in terms of the public it can reach out to via the derivative distribution. And this is true both for users you can reach out to and also for contributors you might be able to get. So the idea of derivative distribution is very good. And Debian has been chosen as a mother distribution, let's say for a loss of derivative distribution. So if you look at DistroWatch, you will find like 120 distribution which have been based on Debian. Some of them are not alive anymore. But those that teleport here are essentially whole alive and kicking. And so that shows that Debian is a nice choice of a base if you want to create your own derivative. So why it is so? Well, first of all we have quite some guarantee on the packages, like you have guarantee of quality and you have guarantee of licensing assurance. So all the review work which has been done on those packages by Debian is something you can reuse for free. This is the beautiful of the software after all. And then you have a very solid by system. You have a lot of packages. So we have seen like 30,000 packages in Debian. And finally Debian is not necessarily customized for any specific use. So we call it the universal operating system, meaning that we try to address various kinds of needs, trying not to be too specific on any of them. So it means that it's a very good base for customization. So starting from Debian you have quite some freedom in customizing it for your specific need. Let's see as a specific example. So there is this minor distribution you might have heard of called Obuntu. And its distribution started in 2004 by Canonical with initially was a target user, was a desktop user. So the motto was Linux for human beings. And they've been very successful in that. And it's a Debian derivative. It's very popular. So if you look at popularity contest data it seems that they have something like 10 times the user base of Debian. So it's very difficult to count the number of user of a software project because on that basis it's something like that. Historically there have been a division in main and universe in Obuntu with the main part being worked on mainly by the, let's say like the corporation part but it's more blurry than that. And the universe part being both worked on by the community. And it's heavily customized in the main part of the distribution. Actually they have quite some packages which have been forked for Debian. So the package is a little bit different. And it's very much more close to Debian in the universe part. And on the right you see the number of packages which populate on Ubuntu release. So this is for I think links but they haven't changed substantially after that. So you see that three quarter of the archive so 74% are unmodified Debian packages. You have like 18% of packages which are Debian packages patched and then which then become part of the Ubuntu distribution. And then you have a 7% of packages which are not packaged in Debian and which ends up in Ubuntu. So this is a representative example in terms of our derivative works. So there are quite a lot of packages which come directly from the distribution you choose to base your work on. But it's this part, so the part which came unmodified from Debian is way higher in other derivative distribution. And that's normal if they have less people power to customize distribution they tend to have more unmodified Debian packages. So what does that mean? Well, it means that if you are running a Debian derivative or a derivative of a derivative or a derivative of a derivative well chances are that you very much depend on Debian. And this is so even if your distribution doesn't tell you so. So this is the second reason why you should care about Debian because like it or not you depend a lot on the work which is being done on Debian. So the question is how we can make all this sustainable, fun and useful for free software in general and for all involved parties. Well, the situation is getting worse. So people are starting to create a derivative distribution from Ubuntu and it is just normal that it is the case. And what does this mean is that we are seeing today a new way of distributing software to people and it's what I call a new software distribution pipeline. So since a few years ago it was something like this. So we used to have upstream software auto on the left then we have some in-between distribution like Debian or Red Hat or any other distribution which was there like 10 to 15 years ago and then we had users on the right. And there was a kind of double flow of information from one direction and the other. So from left to right you see software which flows from who creates the software to users on the right. And in the other direction we have bug report and patches. And now the thing has changed dramatically and we have way more in-between links. So you have different distribution which use the work from each other and at every single step in that chain every single link in that chain you can have users. Okay? That is wonderful because we get way, way more users than we were able to get before. So at every single link you might get new users because you have customized the distribution for a specific need which wasn't addressed before. And we have more high balls which are able to swallow more bugs and we have also a lot more potential contributors. So at each single link we might get potential contributor. But we must be very careful because all this must be sustainable for every single link in the chain and should benefit free software as a whole. Because I don't know what is the case for you but in Debian we do this not for the benefit of Debian but in general for the benefits of free software. So that is what we care about. So what are my let's say golden rules for making all this sustainable? So the first of it of them is keeping in mind that free software is actually bigger and more important than every single link in the chain. It's more important than Debian, it's more important than the Dubuntu, it's more important than every derivative distribution or every single project out there. And then the two rules are that at each single link in the chain we should care about giving back because if we don't give back a single patch which is contributed in a specific link well then the contribution is lost. It stays there and it doesn't flow up to the left and it does not benefit other people. The second point is giving credit. So paying attention that users are very much aware of where the work they benefit from comes from. And this is true at every single level. It's true also for Debian that should be very careful in explaining to their users where the free software they use comes from. So summarizing, there was three main messages I had to tell you in this talk. So my main argument is that you should care about Debian for two reasons. The first one is that it plays a very important role in the free software ecosystem and it's kind of remaining one of the very few distribution doing that. And Debian is also the root of a huge tree of derivative distribution and all of them benefit from work being done in Debian. So if you are a user of one of these distribution you benefit from the work and you should care about the wellbeing of Debian. And finally, we all need to realize that free software is better served by collaboration and code exchange along those three of derivative distribution. So I try to keep this rather short. So I will welcome every question on these topics or any other Debian-related topic you might have. Thank you. Now, I fully agree with you on the importance of Debian but I'm thinking what we're seeing now is sort of a decline of an individual Debian release simply because the release schedule is a fairly long one and there's a number of projects that are on such quick release cycles adding features that users really want that by the time a Debian package actually moves to stable that software is essentially almost obsolete which has a lot of people ending up if they're actually running Debian, running sort of a hybrid between stable and back ports or something of that nature. But what you said here as in Debian being this root of a boatload of other distributions isn't that something that you should reinforce even more strongly in the public image of Debian because I heard you made this point here in this talk but everything that I've seen in that regard on the Debian website proper was one remark saying Ubuntu is based on Debian. And maybe you want to extend that message a little bit further so more people actually get that point. So I agree. Let's start from this last point which enables me showing the first of two spam slides I have available. So we are actually running right now two initiatives related to that. So starting I think six months ago we created what we call the derivative found desk which is a point where a discussion list and a forum and some resources for all Debian derivative distribution to actually meet and discuss how to exchange code among them. But what is more in line with what you just said that we are running right now a census of existing Debian derivative and this I think it's the first step in actually doing what you're proposing. So in communicate more clearly the role of Debian with respect to all these derivatives and being clear about this kind of all Debian is playing. So yes, I totally agree with that and we are trying to fix this. Regarding the first part of your question. So I agree. In fact, what I think is that it's not a single release of Debian which is kind of diminishing in importance but there are new target of users which not care much about stable releases. So that explains the interest in rolling releases which is becoming I have the impression more and more relevant. So a lot of people are using this kind of distribution but don't forget that actually Debian has one such rolling release. So actually I think we have been kind of the first one in having it, it's called testing. It's not meant to be a final user release but it's actually pretty good. So if you use it on your laptop you will find that you have a very good mixture of fresh software and software which has been tested a little bit anyhow. So it has been stayed around for at least 10 days without people noticing serious bug in it. So we do have one such release and I think we should look into how to use it more and how to make it more suitable for final users. So but I don't think it's actually the importance of stable release which have diminished but more than there is a whole new class of Linux users which cares about this kind of very fresh software and something we didn't have I say 10 years ago. So I mean I disagree that there's this whole thing about the stable release cycle. It is a long cycle and looking at the history here it was a one year cycle up until about midway through and then it's been three long. The average is two years for the past four years. And in fact having worked in an ISP that's actually pretty good. A two year release cycle is something that is reasonably workable. So long as the finished product is pretty stable and you can do your own ports to change just the things you need to do. So Ubuntu has come along and added their LTS stuff for servers. I'm not entirely sure whether or not a stable release every six months is a goal that we want because other people are doing that. So I agree with the, so the fact is that there are different kind of users. So the user you just took as an example an ISP is a user which is perfectly served by stable release which came out every two years. And that's fine. Exactly, so I think we in the last 10 years we saw a rise of desktop user, in Linux desktop user. We should be all happy for that, about that. And we just need to figure out how to best address their needs which are not necessarily the same of stable users. So I agree with that. And to some degree that people want to get stuck or prepared to run testing or unstable are doing it. And if they're not prepared to do that, they don't need to go into it. And that will be done. Yeah, sure. This is actually more of a comment than a question. I feel that Debbie and one thing you've done badly is that you've undersold the testing distribution. Most users can use it. It is perfectly safe even non-technical people. There's a big bladder made about releases. Everybody beats it up and says the people say it's taking two years, three years, four years. It really doesn't matter. People can use testing straight away. They don't have to wait for the stable distribution. Yes and no. So I might agree that we undersold them but because we tend to be conservative in that. We prefer not to tell people that something is pretty good and risking that we eat the corner which is not that good than the other way around. So we chosen to stay on the conservative side and not saying that. But in fact, you are right. Testing is a pretty good distribution but it has some glitches which make it not yet suitable for final users. For instance, you have no guarantee that you'll have a Debian installer which is working every single day to install Debian testing. So it might happen that in a specific day you try to install Debian testing and you are not able to do that because the release of Debian installers has not in sync with the testing distribution. So there is some interesting fix in that. So there is an internal project in which some people have interest which is called CUT, constantly usable testing. It's not something we have committed to release or anything yet but there is interesting doing that and the idea of that project is actually fixing those glitches which make testing a not a suitable distribution every single day. So yes, it's fairly good but there are some issues that makes it not yet entirely suitable for every possible use. So for more corporate market, we still have servers that are surge. They're not supported. I think we've got about two months before they're pretty much all gonna be gone at least the public ones. But the Debian release cycle, I mean, one of the things I'm hearing from other people is that the Debian release cycle is too frequent. Of course using, I know, of course using REL as the major example with REL 5 was, for all the back ports might be needed at the end of a Debian release cycle, REL 5 has been just about unusable without a similar system as similarly. So yeah, that was more a comment than a question, I guess, but we have been working on trying to have some sort of LTS support for Debian as well. So it's not there yet for squeeze but the security team is working on that. And right now the security lifetime you could expect for a Debian distribution is something like three here and a half. So this is the time span you can expect right now. So to reach levels of other LTS distribution we just need to add one year and a half of support. And we are actually working on that so it's not there yet. I mean, this is all sort of funny to me because when REL first hit the market one of my comments to one of my counterparts at Red Hat was, oh, you finally figured out how to do a stable release. And it really sort of interesting there's this dynamic tension all the time. And what you find is anytime somebody is building or deploying a new system they always want the latest, freshest, greatest bits because they know that once it goes into production and it's locked down they won't be able to touch it again forever. And so the idea of going out the door based on something that isn't reasonably fresh I mean, you want some confidence in it but if it isn't reasonably fresh it feels like you're sort of behind the curve before you even start. On the other hand, once it's out you really do want this really long extended support cycle if you can get it. And so this is the thing that I think every distribution struggles with and it's why I've been so intrigued in watching this set of Debian derivative distributions to see which ones sort of bias in different directions. One direction on the other. Yeah, because I think at the end of the day it's probably always going to be impossible to make everybody completely happy but it's intriguing that there's so many of us with very different objectives who all sort of find some sweet spot to live in in and around the Debian ecosystem. Thanks Peter. Just like to clarify something here in my humble opinion it's not just the users it's also the relationship that Debian has with their respective upstream projects. Debian relies very, very heavily on fixes from upstream even though there's a security team working on it, et cetera, et cetera. Red Hat or Novell for their long-term distributions they can pay people to fix bugs in the software. Debian doesn't have that luxury so it strongly relies on upstream to fix issues. And if you have an upstream project that says well folks you're using a branch that we've considered obsolete for a year and a half then Debian has a bit of a problem and that's just something that needs to be addressed. Of course if you have like a derivative distribution that caters to this and that perhaps even does have a commercial organization behind it that can pay people to fix these things then great. But I think Debian by itself has a hard time kind of delivering on that sort of expectation. So I agree but this for me is the real challenge. So as I said in the beginning of the talk we are kind of one of the few remaining community store among the most popular distribution. And we are completely volunteer driven and this for me is a real challenge. Whether we can face or not the market with this difference that we don't have the luxury of paying people. This is quite exciting but you know what I think that overall we are doing pretty well. Because I mean the Debian maintainers know the software they are packaging and so they can maintain some code back port fixes or this kind of stuff. Of course it's not the same firepower of big companies but considering the difference I think we're doing pretty well. I have two things to say. First is a compliment. I actually like that I always know what state that Debian is in. Regardless of what I install or what I uninstall it usually always goes back to a relatively stable state even in testing which is fantastic not many other distributions are able to stick to that. Thanks I forward the comment to the Debian community. However on that topic of the testing I was wondering if it was possible to perhaps have a couple supported so you could perhaps do a yearly release based on the testing software and once it's been tested for a while in that somewhat stable testing release it can then go from the somewhat stable to the very stable if people really want to be stable. So I start observing that stable testing is kind of oxymoron. It sort of is but you get what I'm saying. But yeah I get what you're saying and in fact within the discussion of this CUT project which is being going on there are two models which are being proposed and which are being evaluating. One model is purely rolling distribution in which you just need to fix some glitches of the testing distribution and another one is the idea of taking periodic snapshots. So finding some moment in which you have some guarantees out of the current testing you take a snapshot of that and six months later you take another snapshot. So that's what? Because 10 days is not long. Yeah absolutely. So that's kind of one of the idea which is being discussed but it's not there yet. So if you are testing into that you might want to join the initiative because we pretty much need people that are willing to look into that. And you know having an impact in Davion is quite easy after all. You just need to show that you're able to do things and you will be able to have an impact on the project. Hi. Where is it? I realize it's a bit of old history at this point but the OpenSSL issued that Davion had a couple of years back now. And as Davion becomes more of a, there are more downstream distributions pulling from Davion and bugs like that can affect downstream users. I was wondering what sort of infrastructure there is in place for downstream distros and users to address that. So the problem on that, so there is not much infrastructure in place. The infrastructure we have is that all our services are open and usually have some sort of standardized interface so it's relatively easy to hack and use them. The problem in that respect is that among all the derivative distribution we have had based on Davion the only very big one is Ubuntu. So the first moment in which this problem of which interface we use among the two has arrived only with Ubuntu. So we are facing now the problem that we need to make things work well at that boundary and maybe figure out together with other derivative distribution how to make things work better for them. So this is exactly what we're trying to do with the derivative from this but in some sense it is a novel problem because thus far the other distribution were kind of 99% relying on Davion and modifying one person of package and they just kept the changes and didn't care about that because there is no real need to actually in their change much code they were just pulling and not really pushing back. So this is kind of a novel problem we are trying to attack right now. Hi, my question is along the similar lines. What's being done to ensure that bugs and that make their way upstream having all these derivatives of derivatives and that sort of thing. So I think it's been kind of interesting situation once more in the relationship with Ubuntu because we have been bequals complaining that we are not giving back to us and that made us realize that maybe we are not doing everything properly ourselves with respect to upstream. So that's a pretty exciting time in which we need to realize that every single link in that same should do the same and should are there to pushing changes back to upstream. Regarding what we have done and what we have been doing practically well for us it's a standard practice to forward bugs upstream. So it's nothing novel it's something we have always been doing and then of course you have no guarantee that every single maintain is going to do that but let's say the social pressure for doing that is there, is absolutely there. Yeah, that's true. So at your right. Can I just offer my congratulations to yourself and any other Debrion contributors in the room on the work that's being done to remove non-free software from the kernel. It's a fantastic achievement. And also to those which are not in their own. And by the way, this is a very nice example of collaboration with upstream. So the Debian kernel team has been working for like the past two or three release cycle in doing that. So we were being kind of hammered by others saying look you have not putting away bits from the kernel non-free bits but okay we made the exception for two release cycles but the kernel team especially has kept on working on that together with upstream kernel and now it's done. So it's been a very nice example. Yeah, let me also say a word of thanks for the update mechanism which you put in place. Which makes it possible for- The update? The updating mechanism, apt which you put in place makes it possible for somebody non-technical like me to update for 12 years and have nothing broken at all. Questions? It's a bit of a noob question. Is there any way you could just quickly describe what the upstream projects look like? What do you mean the upstream project? Well is it- Ah, the upstream. So that's something which I'm really excited about as well. So that's an initiative run started by some open-source people in particular by Vincent Onth. I'm not sure I can pronounce it properly. Which actually I met at some free software conference a few months ago and said you know what? We have been trying to do some cross distro work on application installer. So meaning the package manager but some kind of package manager which is more suited for final users. So a kind of package manager which is the granularity of applications rather than a granularity of packages. Like I want InScape. I don't care if it is split in three packages. I don't care if I need to install a library or something like that. Something like Software Center which we have in Debian which is in Ubuntu N. So the idea was to work on that and to actually try to do that in a collaborative manner cause different distribution so that we can share package metadata. Like screenshots, like tags are all stuff that can be easily shared among different distributions. So what happened is that they had a meeting and the meeting was attended by people from Debian, from Ubuntu, from Magia, from OpenSUSE and so on and so forth. And they worked on that. They adopted quite some Debian technology like APT, XabianIndex, like DevTags, like ScreenshotDebianNet and they are doing to create a software project to put together those technologies and technology from other distribution to actually have a, let's say, a user interface for package selection which is suitable for end users and which tries to share data from all distributions. So that's kind of the idea. And it's pretty exciting, not only the idea but actually the way in which it has been worked on in a cross-distribution manner. To my, the best of my memory is the first large-scale example of cross-distribution collaboration. So it's pretty exciting. I'm sorry, just one spam slide remaining. So if everything goes well, we are nine days from the release of Debian Squeeze and on the Debian Wiki, you will find a list of release parties and one is missing in Brisbane. So... I'm sure you'll agree that it's great to have such an authoritative speaker on such an influential topic. So join with me to thank Stefan.