 Good morning everybody. Nice to see you here. It's a little bit early in the morning I think not so many people so we can all could all sit around one round table We actually have no round table. It's an octagon table, but it's used by the project also. We are sitting here, okay? Yes, so far for the introduction. I think I will hand over the mic to the representatives of a derivative and for a short introduction and then we keep on walking on the Keep on talking on the following topics if you want to see how to join the gobby talk It's written here There you can see the instruction how to how you can join this the gobby editor and edit your own ideas if Somebody would like to look on I'll see channel debut on auditorium. It would help for be helpful if there are questions from somewhere else So I'm handing over the microphone to Moritz Milnoff. Hi, my name is Moritz Milnoff. I think most of you know me Being a debbie developer for a few years During my regular work time. I work for Univention, which I'm also representing as of today and Univention builds the Univention corporate server a Debian-derived distribution targeted at mostly enterprise customers and Predominantly in the German speaking region. We're quite a large customer base inside Germany and also a fairly significant one in Switzerland and Also a few customers in Austria. There are also a few Customers outside of the German speaking region, but it's only a few ones since a lot of the whole Business being driven like supporting Univention corporate server is usually done by partner companies and the whole network of partner companies is pretty sparse outside of Germany and Switzerland and Univention corporate server is a bit and distinct in regard to many other debbie industries Debian-derived distributions as it's built on stable and we're not Taking snapshots like other distributions, but instead we built on on the latest Debian stable release our current version is still based on Lenny And we're working towards a release and built on squeeze, which will be released at the end of this year and If you're wondering about the target audience Several I think the majority of our customers use Univention corporate server to Deploy workstations based with Windows so that Of course Zumba is one of the very important key factors of Univention corporate server and one of the unique features Which basically is Provided by Univention corporate server is tight all up integration of all the data We have some fairly significant and really large installations And I guess the overall Figure in the base and the installation basis somewhere in the six digits feature Hello, I'm Emma Hickory. I'm going to be speaking on behalf of Ubuntu and I'm not going to go into ground long-length about what that is I'm my field Bokop known as Mika. I'm here for the grammar project live system for System administrators and people who want to rescue data Yeah, we are based on deviant unstable and shipping a few extra packages and that's about it Hello, I'm into Jerry speaking on behalf of tails the amnesic Incognito life system which purpose is to preserve your privacy and anonymity mainly by Forwarding every outgoing internet traffic you may generate through the Tor network and Second purpose is to leave no trace and the computer your the user is using Life system that's built using the Debian live toolkit and it's based on Debian stable and we try to Stay the closest we can to Debian Which means improving Debian some time some times now two years old and well it's now quite well-known and used in the where among the people who use Tor it's listed in the official Tor projects page and We somehow know that one tail system is booted Around the world almost every minute That's all for now So my name is under a stiller. I'm not speaking for for the relative I'm working in Debian for Debian pure blends specifically my plan is Debian made and In principle, there should be somebody speaking for Debian edu we try to Make sure that specific target groups inside Debian are served hopefully well and This could be a preparation for a derivative or it could be a reason not to derive from Debian because you could In principle do everything inside Debian and so I'm always interested What can Debian make? To be more flexible to be more better usable for our users with specific needs and so For I want would like to learn how we can make also the work of the derivatives easier So if if you have some some ideas What's Debian can do better, please? I don't know if it's make sense to have a sequence just say who wants to say something Well, I think from a from a distribution point of There's really not much to ask for the the release cycle is pretty pretty nice and two years fit suits us very well Most of the problems we're having Rather direct upstream problems. There's no specific Problems with regard to Debian that I'm aware of and most of the problems we are running into is Is a lot in the X world because if you want to provide right now most of the customer base is targeted towards service So graphic adapter support doesn't matter very much, but since since we also have a few Few areas where we are actually supposed to graphic drivers especially when it comes to serving thin clients and The the non-back portability of current X or drivers is really one of the major problems But this is of course nothing which which is a problem which can be solved by Debian So from a Debian perspective, I don't see any problem. It's all good The the main thing that Debian could do for Ubuntu is continue maintaining all the software and being friendly to new people joining Debian from Ubuntu Thank you for doing that now and please keep it up The other is that I Hope more people can understand that Ubuntu is not monolithic. I often encounter some individual Debian maintainers saying Ubuntu put this patch on my package. I don't like it and In only one case was that actually me but every time I hear that I feel like I'm at fault And I encourage to contact the person who made something. It's usually in a change log and it can often be resolved relatively easily We are building a greml on Debian on-stair so we don't depend on any release cycle. We need Current software to rescue data so we are pretty happy with over all what we get and I Didn't know a single item I should write on the other question what Debian could do better for us. I'm happy Tales and what tales could benefit from Debian more is Mostly stuff that is specific to live systems for example it happened a few times that we prepared and shipped a new tales release and two days later there was a Debian security announced about Important serious security issues that were fixed in say ice whistle Which is one of the most dangerous internet-facing application we ship in tales and so At this time where we have been poking users to download new easels and where we should just start again and there are worries and tests and publishing process and Have users download yet another easel. So one thing that could be really useful for us would be to be made aware to have some way to be aware of security uploads that are being prepared So that in such a situation we can wait Two days and ship the new ice wither inside our new tales release Security team is answering now. I think I can answer it not in my position as working for Univention, but as someone who prepares security updates the The Mozilla updates are more or less scheduled monthly and and even though they don't announce it publicly If you Google for the next If you Google for the subsequent Firefox version, for example, so if there's Firefox 3.6.17 If you Google for 3.6.18, you will find some link in the Mozilla wiki which sets the release date of the next security update and since also these monthly updates are essentially always bundling security fixes you can more or less estimate that One to three days after that the Debbie and Xuron update will take place It's okay. Thanks. It's great to hear. There is some kind of productivity for us as far as Ice wither is concerned, but it also happened for say many library that ice wither depends on such as elliptic and the usual guilty ones and Also the the kernel So well, probably we should discuss this outside of this room table. Just wanted to mention the issue of pricing. Thanks anyway so I try to summarize we have the to do list is for Debian is Empty because of series 3 and you have pops not it's just not it's to do this But having a better communication channel. It's it is rice. It's okay. So rice, so The to do list for deviant is basically done this place. I'm very happy and perhaps you can Say something what you can do for deviant or what want you to do for deviant to make it's even better well from the Aspect what what is currently being done? Of course one of the One of the aspects is being in the case of financial support like sponsoring that conf which is currently done or also like sponsoring the the seabed booth each year which which gives Debian the The possibility to present it at the seabed conference usually and from a from a cold perspective It's it's a bit the case that since we're only basing on the Stable release of Debian and we try to to be as close as possible. There's not that much really Which needs to be flown back if if there's for example a Significant change in the Debian infrastructure like for example the the move from use flash as the splash screen solution towards Plymouth Then we're rather adapting our own internal solutions rather than working around it and Since we're essentially building on a stable release and since there's also an an unsolvable Lack until the next Unimension Corporate server release can be based on the next ever ease It's usually that up to that point We're essentially basing ourselves on the first or second point release of each Debian release So most of the more awkward box are already also fixed in Debian at that point We made a complete rebuild of squeeze from source which resulted in only I think about 10 failed to build from source failures Which is remarkably good given that it's I think almost 15,000 packages and I think we've sent Patches for all of these or almost all of these some of them are not quite clear where the result is and I think some of these have also been Released now as part of the of the second point release for squeeze and other patches are waiting and In many cases where we are working very close with upstream We're rather working directly with the upstream entertainers are not not channeling all the work for the deputy maintainers For cause for some loosens especially in the whole virtualization areas We need to be quite close to the current upstream release and in and if we are doing any patches in that regard We rather send them directly to the upstream entertainers and bypass the Debian maintainers as long as the Debian packaging is it's not concerned The answer we usually give to this question is we have lots of users I like to believe that we filter the input of these users and provide a source of informed bugs new contributors patches What not in practice the value of this varies hugely depending on the package for packages that are mostly the same As long as we pay attention to it It's useful if it's a package that we neglect then maybe it receives more complaints than help But if it's a package we change a lot, maybe it receives misinformed patches or requests We have eight developers and we have Three of them as DD's two of them are Debian maintainers and in the new maintainer queue to become DD's We have nowadays one developer which is not officially a Debian developer But who does great work in the set shell and one of our most important packages because that's something we Is very special for grammar because it's our default interactive shell and We are contributing to several Important packages for us like in a drama fest tours live boot stuff from the Debian live team full automated installation package and What we can do for Debian is Qa in unstable because we do run daily builds so we catch errors pretty early earlier than most users and we try to report any bugs we find so It's not also Solved for us, but for the user And maybe that's work. That's not that obvious for many people but that's something which is a very nice workflow for us and We are working on contributing back a few even back edges. We built for ourselves All the ones that are not criminal specific already flew back And we have committee bootstrap and criminal such LRC which we try to incorporate in main Debian Well, our main participation Debian as tails is to bring quite a lot of Nante XAV users who would probably not try Debian if it was not thanks to tails to use Debian somehow and Our user communication our communication channels with users show that quite a lot of them Install Debian as their main operating system afterwards Also as a live system we our easel images are Allowed to reproduce bugs quite easily We just have to point Debian package maintainer to a given iso image and it's really easy to to To reproduce a bug this way We participate in the Debian life toolkit development a bit we intend to do more on this side and We are curious enough to start building tails images based on the upcoming being stable release quite early During the Debian release process for example, we start already start in building images based based on testing so at least our build process is somehow some kind of automated test belt test bed for the Installation of a given set of packages with specific requirements and Human testing afterwards Yes, there's some questions from the audience Oh, it's not working. Sorry It is working. Oh, yes. Now. Can you hear me? Yes, but here's who bought off linux magazine I like to ask the Ubuntu representative. I'm sorry. I didn't get your name Hickory Okay, Emmett Hickory is that Emmett I'd like to know a bit more about what Ubuntu contributes back to Debian because This has been a matter of dispute how much you give back and in what ways and you're doing a lot with Debian and I think you have a lot of QA in place also and and Is there a benefit for Debian? so we have mostly the similar packages as Debian and some millions of users and our support teams tend to report bugs into us and our developers tend to review those and Generate patches meant much of the time those patches will be sent to Debian if they aren't open to specific And we encourage that because it makes our lives easier and improves the ecosystem in general When we have new packages, we encourage all new packages to go through the Debian ITP process and for People who want to work on new packages to be Debian maintainers or Debian developers to do that we have trained many developers who are coming to FreeSoft for an encounter to Ubuntu and as people become more Proficient with deb format packages. They tend to say well, I can do this in Debian and we help people go across that and have Tens of developers at least including many that have very significant software stacks that they maintain in Debian or Significant roles within the Debian project Thank you Yeah, I tried to to summarize I think for for those Distribution which are Which are company-driven we have to thank you for sponsoring that conference. Oh, this is one. This is one come Can't take that compliment because from my perspective we are although we have a number of sponsors 15 or 20 We are not So much driven by a company of our yeah, no, I mean Ubuntu is company-driven I won't agree with that Okay, I tried a different word and I thank Univention and Ubuntu for sponsoring Debian It's okay except that when you didn't you should thank Canonical Canonical does great work in Canonical I'm sorry I think I'm not gonna go Thank you intervention but Ubuntu isn't funded well enough to have done that. Yeah, whatever. I think you got the message and The other thing what all these distribution is are doing is bringing more users Indirectly and partly also did directly to Debian. This is also quite helpful, I think and I think there was some sense in the to the next question what makes it Easier for your distribution It might be that we could do some perhaps Relaxed NMU policy because in all those there were two is our Debian developers and if it makes more easy for you to to to build your distribution if some patches gets Faster so it might help in some respect if do you agree or a few other opinions What it's working is it? Yeah, it's work and well, I think with regard to Univention corporate server We don't really need a relaxed NMU policy Because most of the stuff we want to be want to address an unstable It's usually getting getting accepted pretty quickly. I don't think there are any patches Which would need to be NMU since the maintainer usually uploaded them any way it fixed them and For all the patches which affect the stable release I don't think that and relax NMU policy would bypass the stable release managers And I don't think that would be really nice In Ubuntu we don't have a concept of maintainership of individual packages And so everything's always an NMU as we send patches to Debian We tend to get them adopted fairly quickly currently that seems to be working I'm not sure that having us change packages deliberately because a patch worked in Ubuntu without having the review of the maintainer is better for either distribution For the good packages in Grimwell, we try to run peer reviews if anyone changes anything Another developer should double-check it and everything that Derived from Debian should go back to Debian and we try to not NMU it unless the maintainer doesn't want to do the work But the policy works fine the way it is right now and it's not a big deal for us to solve the problems in unstable As far as tiles is concerned, we have no upload right to the RK So we mainly rely on friendly Debian developers who see what we do with a friendly eye and so on probably The case did not happen that often that relax NMU policy would have helped Yeah, you you have no upload rights to the Debian archive, but do you know the past how it could work? Sorry, do you know the past how it could work how you can could get upload rights? Sure. Sure. So Are you working on this? Sure? Okay. We are that is I think it's also quite good because by these derivatives people Have some some stronger interest to to to get absolute right become developer and Some people just become Debian developers because I want to do this and I actually also learn that For my Debian made project. I think if if this Debian made project wouldn't exist. We would have Eight or ten Debian developers less because these people came to Debian to work on this Biological and medical stuff and this I think this is a more important makes Debian stronger so Any any question from the auditorium free free tool to us. Are you running a derivative? I just like to ask Moritz at what point during the Debian release process to you start preparing a new intervention release well for the for the upcoming squeeze release we started the first initial rebuild as soon as squeeze was released and since then we've initially we've first made an effort to to at an initial step to rebuild the package from scratch and this this took about a month and So essentially one a few months after after the next So the I think that the first work started a few weeks after after the final squeeze release Have you considered starting that process when the freeze happens? Actually some some very initial drafting work happened, but I'm usually as as with any company You're always busy with other things also so in order to to avoid additional overhead because at that point where At the very first initial stage of development where you start by by rebuilding all the Debian packages from scratch there's that's not a lot of testing going on if it boots then it's okay and All the all the later refinement like Does everything really work as expected and are there no? Significant errors and and running all the test suites only happens at a far later stage and so for that matter it wouldn't really Change much whether we start with a freeze version or with frozen version or with with the release version Any other question? Perhaps there's there's a representative of the Debian live team or the live team itself Could you could you rise? Do you want I just want to ask you if you could do something with because he has also a live team building They would have with which has live images to do have some questions or no no question. Okay, so I Don't have very much to add, but we're working with toes since a long time ago, and we always appreciate their great patches So please keep doing it Okay so what what I also learned there was recently on the Debian derivatives exchange mailing lists. It's called Debian decks at least or they were not all there was an announcement about a new derivative which is basically Preparing the desktop for young children or for parents to make it as simple as possible and I Had the pleasure to say that in principle there is in Debian something like this like Debian junior And it was a little bit off end and I just gave the link to do the this person that There is something he could We will vitalize inside Debian and to prepare everything what he can do Perhaps this is also a good idea if you have some intent to to support a specific user group just do it inside Debian and Just do the last final bit which which makes your product relabeling is is impermeable. This is the smallest change what I would suggest but It is always a good idea to talk to us first What's what we can can do and if there is something preparation which can be helpful for you So do you have any other comments? Hi, sorry, I just like to respond to the I'm Colin Watson. I work for Ubuntu as well as on Demian I just like to respond to the query from Linux magazine there as well as to your question about About Sort of NMU policy and that sort of thing I apologize for taking so long to reply but it took the Debian BTS that long to answer my query the So Ubuntu a while back started Tiger user tagging everything we submit to Debian partly in order that we could answer this kind of question with actual data rather than with with guesswork the We have about Almost 500 patches outstanding in the in bug tracking system right now the BTS is recording Further 1,600 resolved and presumably there are rather more that have been archived at some point so You know, I think that's a fairly decent rate of both patch flow and actually being applied in Debian the There are certainly cases where it would help if things move more quickly, but I suspect that NMUing wouldn't necessarily be received on all of those cases I would say that it's it's very useful for derivatives to do that kind of user tagging so that we can actually have raw data that we can look at rather than Having to just work purely qualitatively So sorry not really a question No, there's no need for only question any other comments or questions Do you do something about accessibility like testing having an accessibility Group to check that things are accessible to blind death, etc users We used to have a bunch of people using grammar for accessibility Still do actually and I know some people who are using it. Yeah, and we have a problem from this perspective because we lack some hardware to test specific stuff and Whenever we call for testing and Contributing back and we get close to zero return. So I think you're the one maintaining the east. Yeah, thanks for that And If there's anything we could do to improve Working together or testing or maybe we just need I do And maybe we just need some documentation so you can test yourself most of the things because that's what I did for the installer for instance Okay, and maybe it's just about making sure that all derivatives Know this kind of documentation and each and everyone can do their own testing Yeah, I'd appreciate it that if we could talk to each other. Yeah we're starting working for to for accessibility features in tails recently and There are already some options in live boots to enable such feature that we are mostly inherited from Casper I guess I think and mmm Those don't seem to be working that That much nowadays, but for sure we'll fix that in in Debian live Shortly so every other Debian life system will Have clear guidelines to what packages shall be installed and what boot option shall be post Is there a Debian accessibility working group actually yes, there is yes, perfect. So we'll have to talk to each other I think the Samuel and Mario long are the Most active people there, but there are those as well. Yes, and there is Mario long as well. Yeah, my your luck So any other question so my question is Not only about the patches. How about the something that? in your derivations, but Still not in Debian Could you just fire some IP in Debian backtracking systems? So we let we know that It is able to patch you in Debian Just a suggestion So to restate the question to make sure I understand you're asking that when we package something new in one of our derivatives We always file an RFP in Debian for it That thing is it's able to patch tree in Debian, but still missing David and only in the derived derivations. I think it's possible to fire IFP bug right It is possible at least for Ubuntu We found that for many of the packages that were only in Ubuntu We had a very hard time finding anybody who would turn that into an IP or want to be a maintainer We have a list and we could generate the set of RFP bugs if that's desired From a Debian perspective turned out to me as that RFP bugs are in most cases not very helpful because if you don't find an ITPO then You get don't get the package in so the best idea is if you see that you need some package Just push it yourself to Debian But because finally it it's you save some work afterwards if it's inside even and I think you're mostly doing so At least for Ubuntu. We've switched so that almost all new packages go into Debian first because of that Yeah, yeah so I Yeah, there's a further question and by the mic is handed over that I will give it another answer from from another talk Just be bold if you need some package do it yourself do an ITP instead of an RFP Try it Yeah, I'm Stefan Rivera also involved in Debian and Ubuntu and I recall at UDS. We had some conversation about filing off Request for adoptions rather than RFPs for packages that are in Ubuntu that we haven't gotten to Debian yet But really should have that might be a reasonable approach for derivatives because the packaging is all they just need someone to adopt It and maintain it in Debian This is the other I think any other question you have five minutes left, so Everything is clear early things fine, and so we can proceed as before You have one close to empty to-do list and Yeah, whatever some some statements at the end or We just finish and It's related to the question you asked about NMUs Since we are built tails based on Debian stable We are more interested in where we are not more interested with as much interested in what happens after the package has been fixed in unstable for us, it's mainly that it flows to to testing properly and Someone can upload it and squeeze back ports currently so the Policy about uploading packages to back port was somehow clarified a few days ago at the back ports buff In particularly for packages that are installable from unstable or testing without rebuilding It was made clear that it was possible and welcome to upload back ports of these packages to the back ports Debian service In case there is some kind of user base who would benefit from it not just only for your or my own convenience, but if there is some user base that would benefit from it and My question is do we consider their Debian stable based derivative to be Valuable user base would benefit from having the package installable from back ports rather than Doing what we do currently Which is adding the unstable sources to our apps apps configuration Which means shipping using having a longer build time larger indexes and Have the need to maintain pain in configuration and It's really annoying for us. I would be glad to hear some kind of yes as an answer to this question I can't answer the question about back ports. I admit I am I'm should have See the the back ports both, but but I didn't and in my opinion, it is quite a good chance to to get a Less weak system than unstable with with the latest and greater software, but I'm Can't answer this question. So I'm not really competent. I'm not sure if the rolling Releases might help you in this case above rolling that interrolling We have not decided yet if we're going to use running when it eventually happens to to exist I'm not sure currently I quite like our stable based builds and building image from testing Mainly to to be ready when it's released as the as the new Debian release so that we can ship We shipped squeeze based images early in April this year for example and also to help Debian and to help the race process to to go on but Probably we using rolling as a basis for tails would be a bit too much overhead for us Do you consider using rolling for grammar? I'm not sure yet depends How the software we depend on? Can be considered as? Fresh enough and I'm looking forward to see it In reality and give it a try, but I'm not sure either Our main Well, we didn't discuss it that much yet So I'm going to try from to talk for myself my main interest in using rolling as a basis for tails would be Better new hardware support mainly for excerpt, but it was already mentioned for an invention So I think this is a last question because we have to finish somehow, but I would like to your column first Yeah, yeah Yeah, give comments Colin perfectly So on from the other perspective for given that Ubuntu normally works from unstable but what we've done in the past is When we're preparing a long-term support release. We have sometimes merged from from testing rather than from unstable and That's sort of the other perspective and that's we're still discussing whether we're going to do that in the future It's it's had its ups and its downs so We need to stop here now because the video team told us not to talk over 45 minutes Times left and I would really like to invite you to look at the next talk of the DPL Who is exactly about this topic? What are his ideas and all ideas of other deviant people how we can enhance Everything in the relations of to the derivatives. Thank you for attending. Thank you for again and organizing it