 Good morning. We apologize for the delay, but we've been having a bit of a sound issue So I shall just hand over to Mark shuttle worth for his Ubuntu Q&A buff Okay, so this buff really exists because there wasn't enough time for all the questions people wanted to ask after the sort of Ubuntu annual report back so Do you if anyone has questions then we should climb straight into it We had a nice like over dinner discussion last night and there were quite a few people there So we had some good brainstorming on things that we can do and we can touch on that if if people ask questions related to that But maybe we should just climb straight in with with questions Okay, so who's got the first question? Oh, that was an easy buff. Thank you very much. We'll be back to the more readers Don Armstrong is not here But he wanted to ask question at the end he had his he had his hands up and he came up afterwards and he asked about the governance of Ubuntu and the separation between Ubuntu and canonical and So the governance structure is that the the group which is really responsible for Ubuntu is called the community council At the moment, there's one non canonical person on there. That's Benjamin Mako Hill Who many of you will know and he used to work for canonical then he went back to study at MIT But he remains on the community. Yeah Where did you find it? Where did you find it? Success Okay, so the the community council is The the kind of overall body and then there's a technical board which also has a non canonical person Matthew Garrett on it I'm perfectly happy with both of those groups ultimately ending up with them with a minority of canonical employed people Obviously that's gonna take time because it's you know people to get on to that We're not just gonna put anybody on to those groups, right? Those are very very serious groups. They they they make a lot of decisions around Ubuntu So to be on the technical board, they'd have to be you know, super good really Really demonstrated an exceptional knowledge of the whole distro how it all works together and so on be very committed to the project And if they weren't gonna be in a canonical employee They would they would then have to not want to work for canonical because if anyone was that good and that committed We would want them to work for us But it may be I expect that in time what will happen is that other companies that are building stuff around Ubuntu We'll be happy to fund people to work on Ubuntu and and in that case we may well have You know people who or students or people who you know just don't want to be employed Could easily be on the technical board or the community council Any any other questions? Yeah Just a minor one. Ubuntu is developing an additional graphic and installer. Will it be Will it be merged with DI? No, so this DI has a graphical mode now a GTK. I think it's GTK based mode Which Colin Watson did a fair amount of work on Joey will probably confirm that and Then there's a new installer which is which is totally different. It doesn't share any code whatsoever It's written in python. It has a different sort of style approach. It's much more limited effectively DI is very flexible malleable. You can do all sorts of stuff with it I saw full hands was doing stuff with zen and di so di is a much more flexible kind of installer the The graphical installer in dapper is much more limited. It's basically just do a desktop And so it's optimized just for that simple sort of use case So it's much more restricted effectively in that way It's free software. I think it's uploaded in debian as well. Any other questions? Oh, I feel like I'm let off the hook very easily Again, I want this taller for ubuntu ubuntu When will it be ready for a complex text layout languages for us india Cambodia Okay, I didn't know that it wasn't It should be Yeah, I think it's just it's just using gtk and pango So so as long as you've got all the right fund support, you should be you should be okay. It shouldn't be a problem Anybody else? Oh, that's totally easy by elege. You want to start early? Okay, yeah Yeah, um, you talked about, uh, ubuntu looking at a different Uh package Dependency resolution aside from apt. I don't remember the name smart. How does that affect? The kind of the collaboration between ubuntu and debian Shouldn't affect it at all because smart doesn't doesn't require any changes in the package formats And there are already people using smart with With debian. So I think smart is uploaded to debian and there are already people using it with with debian It's just like that middle layer of trying to figure out when you've got this whole set of packages installed And this whole set of packages in the in the archives with new versions and so on Figuring out what the what the best way to sort of the optimal way to get from here to there is And uh, so one of the things that we like about smart is that it will do things like consider Potentially every possible path of installs and removes and then installs again So it can say okay to get from here to there First i'm going to install this other thing that allows me to remove this thing Then i'm going to install that over there. So it just has um smarter Algorithms to do that. It's not perfect. I've seen it blow up doing doing transitions We tried it from like hori to breezy and stuff like that and it wasn't able to handle some of the transitions So that's what we would consider edgy Type type technology, right? It's something that we can do and if it works really well and so on Then we can help build the case for that to be part of the next debian stable release That's kind of some the sort of thing that I think Ubuntu does very well Is find find something new and interesting bring it in show that it works And then it's very easy to to get approval for adoption In in debian of new technology and I think we've already proven That that that process has worked quite well Christian Yeah You know probably what I will be talking about Mostly q&a processes about localization stuff How do you think we can handle this in the future between ubuntu and debian mostly And how do you currently handle q&a stuff in rosetta? Okay, so rosetta is this web-based translation infrastructure that we use in ubuntu And it's really designed to optimize the process of translating and passing translations upstream So we've really optimized to get the translations ultimately to upstream because for 99 percent of things I believe upstream is the right place for translations to land if a translation lands upstream Javier will have an opinion about this if it lands upstream it goes to red hat it goes to debian There's no questions about which distros it ends up in and so on there are some things of course where upstream is debian like di Right and there that depends on personal relationships really to to make sure that that that flow happens in terms of quality assurance Rosetta basically allows you to set up a team of people who can edit a particular profile through the web and When we set that up the intent was obviously that they would keep tight control over who was in those teams And I noticed the other day that one or two of the languages have more than a hundred people in the teams Which means that they've had fantastic Results if you look at how the bar chart is moving But not so fantastic results in terms of the consistency and quality of those translations So this geordi is working with those translation teams to say hold you should have a process Where people first, you know first they they demonstrate their ability to actually to do this well And then they become part of the team So we don't have a very Rosetta doesn't have a very efficient review and and approval kind of workflow process It's either you can edit it or you can sort of make suggestions But they don't become official until until someone who's in the team edits it And so it depends the quality depends a lot on how Well controlled that group has been basically Yeah, I had another question How do you intend in the future to Interface with what we will probably soon develop in Debian for infrastructure for internationalization Which will probably not be based on Rosetta for very well known reasons Uh, how do you plan to integrate with stuff like World Forge stuff just like Javier Sola over there talked about so they're really they're really specs Which we plan all of our development using specs and and there are really specs for in the Rosetta roadmap for integration with word forge Which is like a free software glossary management system So the idea is that Javier's team as word for which develops They will set up a glossary for different languages And then we can link from Rosetta straight into that glossary If word forge develops like an xml rpc api Then we can start to sort of say okay Looking at this translation over here. We've noticed that in this language words In this translation appear in the glossary, which means they have some sort of special Interpretation or you should be aware of certain additional information We can then display that directly in in Rosetta Rosetta itself currently has an as a web based interface for uploading profiles and retrieving profiles So there's nothing locked into Rosetta and we can also do an xml rpc interface to allow people programmatically to say Are there any new translations and if so, send me the profile kind of thing Ubuntu does separate between here Does separate between main universe and multiverse? Can you elaborate who can upload to multiverse? Who does the new processing there and why do you distribute un-distributable packages? For example adobe reader For example adobe reader acrobat reader acrobat reader No, it's not in debian Okay, i'm not sure the specific licensing issues there, but so the difference between Main and universe multiverse is whether or not we will do we will guarantee that we provide security updates The difference between universe and multiverse is basically Free and non-free And there is an additional there is an additional repository in dapper which is non-free stuff Which has actually been certified it's the stuff like ibm db2 and things like that vmware and so on Which is which are sort of certified where the companies actually provide support for it So that you get you get the assurance of sort of security updates and so on but by default It's not visible after an open to install you have to go in and say okay. I want to see this stuff It's like adding an extra app sources line Um, sorry you didn't answer my question. Can you um Say who can upload there which people and who does the new processing there? Yes, so if you look if you look in Launchpad.net slash distros slash ubuntu you will see a list of those components in the left hand side of the page And so for main you'll see a team ubuntu core dev for universe You'll see ubuntu dev and then if you click on those you'll see a list of all the people who who have those upload privileges basically so Um universe and multiverse. I think is the same group Which is which is every all the developers in ubuntu and then main and restricted would be would be only the core developers And all core developers are also part of the motu because of the way launchpad handles cascading teams So you can easily get a full list of all the people who have uploaded New processing is currently Okay, so there's no special difference between new processing for that one It's not exposed at the moment It's a bit of a hack behind the scenes like a command line tool as to to the new processing But I think that that's moving to a web-based interface and there'll be a there'll be a Again in that same page which describes ubuntu in launchpad. There'll be like an archive administrators person And you'll be able to see which will be a team and you'll be able to see exactly who that is So you can figure out you know always who to talk to and ubuntu is pretty good about exposing the queue So you can see everything that's been uploaded and where it is in the queue processing and so on And and and are there efforts to Remove the stuff which is not uh legal to distribute at the moment this So we have no intent to distribute anything which is illegal to distribute We may have we have redistribution agreements for a variety of things, right? So we don't put them in the distro per se. We don't stick them on the cd. It's not part of the default install There's no non-free application Which is part of the default install But there can be stuff in in commercial or multi versus or on where we where we have Redistribution agreements and things like that. So you could take that up. James troop is by and large Most persistent license expert, but you can take that up with me or with the community council If there's a specific example, uh, I hardly no I doubt hardly that heart that uh, you have one for adobe Because i'm parking adobe since nearly a year to get such a redistribution license And they only allow it to distribute on CDs, but not through any kind of network And uh, it's fact that you have it on your multiverse servers repository came Thanks for raising it Go ahead. Okay. In the world. There are various projects from government or from private industry to put Linux in the desktop or for the government Do you know in which countries they are Choosing ubuntu and if you know why they chose Um, I think ubuntu is quite popular in the countries that that are familiar with adobe in space and that are more that are more advanced effectively in terms of Of their understanding and and policies around open source and free software So I know in china for example that ubuntu has been distributed in a bunch of universities officially by the government I know that in um, indonesia. There's some consideration of of Of building a derivative based on ubuntu in most of these cases They're not actually going to ship ubuntu as it is It's just that ubuntu is a convenient starting point for a local derivative of one of one form or another Which i'm perfectly happy with right? I don't mind if they call it Something completely different change the brand change the name. I see that as being part of what ubuntu is there to to be for Spain there are a couple of the provinces and regions and regions in spain that have used ubuntu as the as the template for for their own Um, official derivatives in brazil. There's work on a multimedia version of ubuntu for use in Telecenters for music and and and sound and art and stuff like that Um, so there's all projects in various stages of production So I just quickly want to say the reason acrobat isn't currently non-free Is because adobe took so endlessly long to release a new reason At new release that there were so many security bugs patting up that we removed it But um now nobody really gives a shit So it just hasn't been uploaded to non-free again if you want to take it over and upload it to non-free. I think it can be uploaded Okay, but still if there is a potential issue there that's been raised we'll look into it right right? I did have another question. Um I've heard that there are some plans and maybe you could elaborate a little bit on that on extending launch pad to have an email interface It does already have an email interface for for example the bug tracker So you can you can create a bug by filing an email by by sending an email you can close bugs You can reassign bugs do all of that sort of stuff Um, we also have email interfaces for things like our spec tracker So if you subscribe to a spec then you can get notices of changes in its status And you can even subscribe launch pad to the wiki where the spec is and then it'll bounce wiki change notifications Out to all the subscribers of the spec. So that's kind of primitive. It's not it's not comprehensive but over time it'll it'll improve and if you have suggestions and ideas then We would definitely Want to integrate those? Oh, so we we're asking about the unauthenticated email What we what we sort of discussed last night? Okay But there was an interesting there was an interesting comment in our discussion last night where Say you want to say say you want to subscribe to changes to a particular package or bugs from a particular package in Ubuntu that does mean that you need to go and register at launch pad and So for some people that's going to be an issue because launch pad is proprietary That's like not using google or not using amazon or something like that right but fair enough So we can look and see if there is in a way a special way for dds to be able to say okay I'm not I don't have a launch pad account In other words, I I don't have a password I can't log in and update my details and so on but you can use my email address in this way Right and so you can you can if you would send me something you can send it to me But I don't want to have an account which is You know and fairly What the french would call a nuanced position But I think we could quite happily figure something out like that if that would Help people the other thing that we sort of figure we can do now is we can probably automatically send a bunch of change notifications To the debian package tracking system and if you subscribe to your debian package with the derivatives tag Then you should be able to get that stuff automatically without Subscribing to launch pad Could you please share some On ship it program Sure. Um, I think the final tell you was about Okay, the question was for statistics on the ship it program, which is where we make CDs available around the world Some some really interesting statistics. We should more CDs to iceland than to japan Uh, we should more six CDs to cunia than to japan as far as as far as I recall So some very interesting country country derivatives or variations About 15 of those CDs went to the states And so the states is actually relatively low in terms of if you think of normal technology adoption, right? So it's a very it's a very broad based kind of program I think about 200 countries Ship CDs to about 200 countries Including more CDs to pit cairn island than there are supposedly people on pit cairn island I'm not sure exactly what that means In total I think about six million breezy CDs in total, which is a lot Um Obviously there's a big win for us in dapper because we cut the whole thing down to one CD There was a big part of the drive to not have two CDs and install and a live CD. So that's why we had to do this kind of simplified installer Um, we still have the di base installer for server editions, but I don't think we will ship server edition CDs Yet another question about localization, of course Do you have currently contacts or plans for localization? Especially even in of the installer because this is the beginning of the for how you install a boon towards dbn actually for african languages And also for africans for africans Um at the moment it's all kind of community driven. We've done one One language one african language where I funded it personally basically because we wanted to figure out From scratch. What does it cost to do a desktop? Pretty well, it turns out to be about 200 to 300 thousand dollars from scratch To hire professional translators to do the whole thing. So at least we have a number now There are 500 Um african languages, so i'm not going to fund Many more on that basis Whereas etter is weak is because it's web based And because the pages are quite heavy effectively. You've got a lot of information coming and going It's not good for languages in countries where bandwidth is poor which is true across most of africa So in africa, we have to figure out a different strategy where what we will probably do is get people to work locally Right using either something like poodle or or just using kbabel or G translator or one of the local desktop editing Translation files and then upload just in one shot to risetta. So risetta there just becomes a convenient way of Keeping track of all of the different things. So if you have 10 different sprints happening covering different files Then they can still pull up one page that says how they're doing overall Yeah, I think that in the case especially for africa You may have a different case than we have for other communities and the concept of Community driven translation may not work very well because we all know that in these places Most of the people interested in technology are actually not interested in translation in their language They perfectly comfortable with english with french and most of the time So I think it has to be a different way mostly probably in involving more Institutional organization more top-down more driven on the basis of this being the state saying this is a good thing So that's one of the things that we do with canonical is we lobby very hard whenever I meet a government official It's one of the things that I say they absolutely must do right because First it's a unique thing that they can get from free software that they can't get from proprietary software Second it completely changes their ability to get technology into communities and third if they if they precede it effectively if they if they Underwrite a lot of the translation then in theory They are they are they have an incentive to distribute that work get lots of users and then something like rosetta is very good for Keeping translations up to date because once you've got 95 percent people like to sort of go It's like a wiki right go and fix that last five percent, but getting from zero to 95 percent doesn't happen Magically on its own you know how much translation work there is to do right You can go to a single page on which will bring up all the translations in dapper And we've got maybe 12 languages that are very well translated And and maybe another 20 languages that are you know There's good translations for the important stuff and then after that nothing and there are 350 languages with more than a million speakers So if we want free software Widely used in the native language Then there's just a huge amount of work to be done so Yeah, you just mentioned that you only use di for the server installed anymore But as far as I know you use di for the live cd on breezy does that mean there is a new technology For dapper to boot or you still you're absolutely right basically Yeah, that the that the core underlying live cd technology is still di Okay, but the but the it's not Okay, so apparently there isn't and maths amendment would be the right guy to talk to do an answer Just address of what it would be but if it's do you have a specific question about how the new live cd works or What I don't know I'll look up the documentation. I just wanted to know what's in it like Were you familiar with the breezy live cd setup? Yeah, I know it was di Based and then right and it was based on the casper Package and it still is now. It's just the casper packages change So if you look in the same place you can see How the moon works is based on the net ramifest tools and union fs and squash fs Okay, thanks here, okay I didn't understand this very well for the live cd installation Do you have to translate the dev installer? No, for example kubuntu Yes, this is sort of a separate issue In addition to the live cd technology itself, which used to use di modules for bootstrapping itself and no longer does The actual installer called ubiquity, which is used on the live cd Uses debbie and it uses di modules for certain parts of the installation the partitioner and so forth So it's reusing those strings. So if you translate di then we use those translations in ubiquity as well But on top of that you still have to translate Specific parts of the ubuntu or kubuntu installer Correct. There are a few strings which are unique to the graphical interface. Yeah But they're actually merged into one set of templates. So if you translate, you know, the installer of stuff in ubuntu, it's all together Good I learn something new every day Yes Regarding the the companies that have signed up or are in the website For providing support for ubuntu the marketplace. Yeah Have you gotten any feedback or comments or any Cases that they have implemented or you know any That you could share Not that I'm aware of that that the number of companies there is growing astronomically that we don't have a very cohesive program effectively for working For for for building a stronger interaction We have a we have a slightly different partnership and alliance program where we have more formal relationships with people But that marketplace program is really almost like a wiki people just come in and say, please would you put our company over there? This is our contact details and this is a short description of what we do Yeah Do you plan or do you have ever considered to So do you plan Do you plan to add something in the code of conduct about giving back Stuff to debian in terms of patches or in terms of bts or everything else So we that's a that's a really good point. I think we do have stuff there which says, you know Plain nice with the rest of the free software community. And if we don't I'd be absolutely happy to put that in We do modify the code of conduct every now and then when we do it and then Second to the code of conduct We also have a bunch of kind of guides which say what does the code of conduct mean in irc? For example, what does it mean on mailing lists what it mean in the wiki and in forums? And we could definitely have an you know, what does the code of conduct mean for developers? Which would certainly include Notifying people when you're doing work making sure that you publish the work in the right sort of you know Release early and often make sure that you work well with other people's bug tracking systems Whether that's debian whether that's that's upstream whether that's other distros And I I think we could certainly draft stuff like that up. That's again stuff to raise with the community council mark I read about the d package version 2 a paper Can you say something about the time frame when it will be seen or something between because it's mostly Ubuntu developed as far as I know Yeah, I mean there are a lot of folks who work at canonical who are interested and involved in that stack And there's no particular drive from canonical's perspective to do d package version 2 There is some interesting big Challenges that are out there that we like to see done but and obviously we have the resources to do them But they're not necessarily strategic priorities for us For example multi-arch and there are a couple of proposals floating around in debian for multi-arch If you're familiar with tolef and scott's proposal, that's that's the thing that we're most kind of comfortable with So obviously we hope that if work starts on d package version 2 it follows that line There are there are other areas where I think we could all benefit from good work Right. There are a whole bunch of different source package formats out there and The existence of those varieties partly is a good thing because they each exist to serve a particular sort of model or need Or or set of requirements or way of doing things Sometimes that's actually been a source of conflict between ubuntu and debian and other distros or dist or derivatives For example, I've often seen a case where Either an ubuntu guy has repackaged something that's already in debian because he wants to use a different packaging system Or vice versa We've had packages that existed in ubuntu and a debian guy said i'm going to create a from scratch package Because I don't like that packaging system We have a guideline for developers not to change the packaging format Unless there's an absolutely compelling reason like for example the xfce guys want to build a lot of the a lot of the GNOME applications without GNOME support So they use gtk, but they leave out the GNOME interaction stuff And in in many cases they just can't do it using the existing packaging formats That hasn't happened for dappa because we were in sort of quite tight release process when we got there They will try and consult with the debian guys But if they absolutely can't get any traction there, then they will go to a different packaging format where they can build from a Single source package the GNOME versions and the xfce versions So what would I like to see one of the things I'd like to see is a lot of thinking and work done on the source packaging side of things So that we can bring in the best of cdbs and dpatch and all of the different proposals and systems We can pen and so on that have that have been put in place And and and and kind of get people using a common set of free software tools to manage those Where one of the key features we get is is better interaction between distros right So for example, there's been some complaints about that when we automatically publish our changes versus debian that you get a monolithic diff Well, you only get a monolithic diff if the debian package Am I right here As I understand that you only get a monolithic diff if the debian package itself uses the old monolithic diff format Right. So in other words, if we can't easily pull out changes automatically, then what you get back is our change is not pulled Out automatically that's not to be mean. It's just because automatically we can't easily do anything Whereas if the the debian package is using split out diffs, then it's easy for us to say Oh, here are our extra changes and it's diff one of two diff three to four And so of course, then that's much easier for the debian maintainer to take So in the past that wasn't so important But I think in future if we look towards a world where there's lots of derivatives and we want to share Right, then having source package formats which make that easy and having General sort of agreed practices that make that easy is going to help all of us So that's that's where I'll have an interest because I want to reduce the cost of Shifting shifting stuff around right from debian into a buntu and from a buntu into debian equally So there is no real time frame Because you know Unless unless we decide to invest in it, which we haven't yet decided to do Then it's a scratch your own each kind of thing for tolef and scott and other people who want to collaborate with them on d-package too Bdell I think this will work So you were just mentioning, you know, if we end up in a world where there are lots and lots of derivatives that And it's kind of interesting because one of the things that for a long time seem to be really interesting to me about debian space is that People figured out how to do lots of flavors as I like to call them and other people have other names for them Where one of the defining characteristics is all of that work did one way or another end up within sort of the Defining boundaries of what we call debian And one of the defining characteristics of what I call the flavor was that it was sort of a proper subset of the whole in some sense It seems to me that things really are a little different when we start talking about A world in which a lot of stuff is happening beyond that boundary and we have, you know Layers of derivatives and it seems to me that some things do get more challenging and it almost It almost seems to me to put more pressure on us to figure out how to be successful about You know making sure that we don't end up with too many different places where little bits of goodness in the form of Patches for this and that end up Sort of stashed away. I wonder if you have any thoughts Since I know you have very strong feelings and have talked in keynotes and other things Technologies like distributed revision control and so forth If you have any thoughts about sort of where we are where Some of those things ought to be directed so that we don't end up In sort of, you know, the the the chaos and pathos of way too many little derivatives that are each caching too much Goodness and and lost work effectively. It's yeah, it's enormously frustrating to me when I see how how because it's hard to move Work upstream or because it's hard to move work to Debian or to Ubuntu or, you know Too red hat or wherever that a lot of work. It's either duplicated or just falls on the ground, right? These guys go charging ahead. They do a little derivative that's optimized for a particular thing They do interesting stuff and then they get bored and they move on and that work rots bit rots And that's a loss to all of us Right So I do think tools and formats and protocols and best practices are are the answer now The example I would give is the linux kernel, right the linux kernel has linus's tree Where he takes care of they agree to take care of a certain set of use cases So a certain set of architectures are in tree like zen one day will be in tree But then there's a ton of other places where linux work happens And so people maintain their own trees This used to be an absolute nightmare because there would there was no infrastructure to to move stuff around And git whatever its other flaws may be git has actually made that process a whole lot more cohesive because you can now at least Cohesively through a single interface if you know where work is being done you can go and say all right How are they different so it allows people to pursue their specialist interests But at the same time allows other people to go in and do an audit and say oh, that's interesting That's interesting. That's interesting. Pull it in beat out So you're absolutely right on the other hand I may actually have found the patch And it looks like it might be something from like last year that hasn't been pushed upstream yet So this is one of those places where good tools aren't sufficient and it comes back to this conversation I commented on at the start of my you know governance thing the other day that I had with With cliff from the linspire crowd about the fact that if we want to make things work In this kind of a context we have to figure out how to be more active active about it You know both people who've done interesting stuff pushing it Right, so But basically both sides need to agree to be willing to do a certain amount of work And then what needs to happen is agreements on how we make that work as easy as and lightweight as possible. So First everyone needs to agree that this is actually a good thing and second then what what what the responsibilities are on on each side So I think yeah, you're absolutely right. It's not just tools. It's also social Contracts and protocols and agreements and covenants I've just found you at the wound to have a marketing team in large part. So what do Will be the role of marketing teams and distributions So that that team is self-forming right that that team popped up on its own and they they do their own thing they do stuff like Announce in different countries if there's a new team that's getting together and doing work there or either on translation or on a derivative So loco teams things like that they they write up guides to you know, what's new in the distro when we do a new release It's not hugely active at this stage, but I'm very I'm very supportive of it Right, I think it's a very it's a very healthy part of the whole ecosystem and if guys want to do that Then I I think that's awesome. They should have the space to go run and do that um recently Well fairly recently we voted to um take certain gfdl documents out of main Uh, I'm just curious if a bunch who was going to follow with that decision if I'm not sure exactly How your your freeness rules relate? I know they're based on devians free software guidelines, but I don't know if it's as strict or Of view on the gfdl is we consider it free Um, but we would look on a case-by-case basis at the use of at the use of um, what are they called invariance? Right, and we've just not come across it and a case where we weren't comfortable with the way that was being used So so that's been that's been our approach. We're comfortable with the gfdl We don't have a major problem with it. So the debian's decision Is a very interesting example of how I think debian can it potentially marginalizes itself, right? I would like to say you know upstream is king if upstream ships a tarball We should express what we do as changes against that tarball Changes which other people can take or not take and I really do think that if debian wants to be successful that it should be very careful Always to preserve the ability for other people to still work with debian But be somewhat selective in what in in how they choose to do that in the same way that a debian derivative can create its own Main which includes whatever crack it wants right and people do right scola and others include You know java or whatever it is in pre-licensed change days Because that's what they feel they need to do for their set of users Now in terms of the collaboration side of things right if debian has changed the original tata gz And they different to upstream tarballs Then that becomes extremely different difficult for other distros to work with effectively So we haven't really bothered to put back Stuff that debian has taken out in dfs g free cases accepting cases where we thought it was pretty essential I think there are there might be some cases where we've added it back as a separate package or we've done something different So again, I would just say without passing judgment on what's happened in the past that one of one thing that it would be good for debian to bear in mind As it goes forward is that Is that if debian wants to be this universal environment, which is what I love about debian right lots of architectures You know lots of packages Harnessing parallelism harnessing all of those things together with debian legal is very clear-cut analysis of you know What's in what's out and so on but still leaving space for non-free right? I mean debian voted to keep non-free right? So if you're willing to do that then perhaps it would be possible also to express some of these other Policy ideas in ways which don't exclude or make life Excessively difficult for other derivatives of debian which take a different policy view So it's asking quite a difficult thing It's asking you to say here's our set of policies right here's how we implement them But we respect when other people Want to build off our work with a different set of policies The reason I would say that that's useful for debian to think about is because people are going to do it anyway Right and more importantly if you if you want to be a good partner for collaboration Understand that difference that that importance of saying here's work. I make it available under a free software license Here's what I express about my policies, but I respect your right and willingness to do things slightly differently I want to just close with with with an observation I've been really thinking hard about this social process and coming up with good examples and bad examples up in the free software community And I think the guy who in my mind has been had the most success in terms of getting his ideas and his thoughts Widely used has been Linus Torvalds, right if you look at back to 1991 or 1992 There were a number of free software kernels out there and we should ask why did linux become So enormously popular and successful and I think it's because linus considers it's a success When somebody takes the linux kernel and does something different with it Even if he wouldn't merge it into his main line, even if he thinks it's total utter crack It has been shown in days gone by that sometimes things that linus said were total utter crack two years later Got merged into the main line by linus right that technical ideas for the softical ideas and and so on should be allowed to flourish And that you know ultimately it boils down to this I think linus counts it a win when someone does something Out there with linux when they put it on a watch or a pda or a mainframe or something like that And I think ultimately we all when we do free software people are going to take it and use it in other ways And we each have a personal decision to make as to whether that's a win or a loss with debian There's no institutional view right it's a personal decision that every developer takes And I would say it's actually we all end up a lot happier if your general view is when somebody takes my stuff and does Interesting things with it does different things with it. It's not a loss to me It's part of what free software is all about and in fact if we if we all take that view Then debian is going to be successful in the same way that linux has been successful So thanks very much for the qz and a is and i'm still available. Oh and the other one thing that came out last night is that You guys should absolutely know that you can raise stuff about ubuntu either formally through the Through the various structures like the technical board if it's a technical issue of community council If it's a social issue with the code of conduct or something like that or directly to matt to myself If you see a bad interaction between a debian developer and ubuntu developer Let us know because if it's something that we can work in our community We can always do that if you see a bad interaction coming from a debian side help that guy to do better and uh and That's all folks. Thank you