 The 12-slide city of EU, it's got some Debian in it. Go back, please. We've got the Oktoberfest, of course, which we will enjoy in short. Steve, see you there. Yeah, go on. Some things which I will tell you about. First thing is how the whole project developed over time, why the decision for Linux has been made, some small facts about the budget and what the status quo is. The history of the project is mostly, we've got a very, very, very complicated infrastructure in Munich. It has really everything in it from on the client and server side, you can imagine, but mostly Windows NT4 on the clients, some Windows 2000, but that's very less. And that's the main problem we had when there are applications. We didn't know what we had to do with those applications which aren't supported by Microsoft anymore or the actual companies who developed them. We've got 14,000 workstations and 21 organizations called departments. So they got their own IT, we're the main IT provider. Yeah, and as I said, many, many different products and support concepts, no consistent thing on that part. Yeah, that's a very, very long history. I'm very often asked, yeah, when are you done with the project but actually the decision like you can read there is, yeah, has been made in 2004. It's both the beginning of 2004, 2001. We just did a study. I think it was made by AnuVention. We did conception in 2003, 2004. I joined in the beginning, January of 2005. That time I've been a developer and since 2007 our first version of the client based on DB and Sarge and Edge has been released. Yeah, some facts about the aim of the old project. I'll skip it pretty fast. More independence, yeah. What else? It's a most important argument we're after. More equality of opportunities, more competition. You can just give orders to anyone you like, anybody who has access to the source code and is able to do some programming can help us. So it's a definite advantage for us. Controlling of course, complete open source code. It's very important for us because we're maintaining some parts for ourselves, security and of course we want to avoid proprietary software to protect our personal data. Decision has been made not only for the Linux client but also for the Office Suite which we have chosen OpenOffice for. That's a short overview of the cost over the budget. That's pretty decent budget. Very important fact about the whole migration is of course the user training. We don't leave the user alone. We did a detailed requirement analysis and training concept. Everybody can have the training on the job with an e-learning system or advantage courses from the admin to the normal user. There's been some scientific evaluation by the Munich LMU and the e-learning solution. It's called the LIMUX Lernwelt e-learning world if you want so. In English is approved and awarded with the Aurelia Award in 2007. The numbers up to date are 14,000 workstations. Are using free software and some kind. We have some standards. It's Firefox of course. It's Thunderbird with iStove. We've got 8,500 employees already trained to use free software in training courses and we are e-learning. 8,000 workstations are using OpenOffice.org. ODF is going to be the standard interchange format in 2009 so we can switch off the support for, but we won't, we will have ODF definitely preferred. I don't know if when we can actually switch off the support for Microsoft documents or whatever they pull out next. We have 1,000. I think the number is actually higher because the slides are referring to the first year of the migration. Now we're already in the second year and of course at the moment we are preparing the 2.0 release for our client and yeah, it's summer break so when the IT departments come back from the summer break, we'll roll out some pretty great departments so that number will definitely increase very fast. Some short facts about the OpenOffice migration. 16,000 objects to be graded. It's macros, templates, formulas. Yeah, we have many redundancies for the same problems. It's those custom macros you just meet everywhere and of course, yeah, the problems with file exchange between Microsoft Office and OpenOffice is a very great problem. First, very, very, very great thing. I'm proud of personally is our Walmux which we developed internally. It's a letterhead system. It's a form system. It's a text module system which supports some workflows you have in, I could say governmental environment so it's not perfectly suited for the home user but definitely usable because, for example, the serial letter function is more advanced than the OpenOffice one is. The thing I'm proud of is that since I think it's two months, the Walmux is free software. We finally have found one license we can rely on. It's called the UPL, the European Union Public License. It's GPL version two compatible and that's also the license we will release parts of the source code of the modifications and patches we developed while we took Debian and fitted it for our needs. So actually this is our better test if you want so to get a license and to publish things in cooperation with Oso that's a platform, we released the Walmux. Skip that part, it's just about ODF. Skip the part two. Yeah, that's some facts about the client also. We had different file servers for user and system management. It's, yeah, it's been in parts in the LOP system. It has been Samba servers for file servers, had been NetApp filers, everything you can imagine. Different solutions for software distribution by feed or by software. There's been everything you can imagine also. We have lots of business applications. I think those 300 are not accurate. There must be many, many more. At least it feels so. And yeah, different operating support processes. That's one thing we built our actual development process after ITAL and the support process. So yeah, that's another thing that's pretty good too. That's just a short overview of how we manage our clients. On one side is the GOSA2 system and config management that the web-based management tool from Gonykos. It's a German company. We've put pretty much money and time into that project. It's already contained in the Debian repositories, but in an older version. I think if that changes, it's a very reliable and interesting solution to manage and deploy software and client machines in cooperative environments. So yeah, it's just an overview. If you want some details, just get back to me after the talk or tomorrow or it's Sunday, so we can talk about it or write an email. Yeah, another big factor, important factor is of course the usability of our client. We've put pretty much effort into it. And finally, TIFF IT certified our client for usability so that users can actually work with it and work with it better than the previous client, which was just inconsistent, different, and was very application-centric. For example, in the Linux client, you won't find open office writer in the menu. You actually find it, but the first thing you find is text writing application. So the users get focused on what they do with the application and yeah, we don't hide the application name, but that's not in the focus. Third big part of the project besides office and the client is of course those applications which are in some way application-specific developed for Microsoft Windows or Certified or whatever. So we got several migration solutions, how we deal with that. It's from web-based to Java-based solutions like the Walmart's to virtualization, VMware, Wine or Terminal Server or yeah, just let that the MPC on Windows and concentrate on much more, let's say yeah, on the 80% we want to reach and not focus on those little them solutions which will die anyways, let's say that. Lessons learned while we have been working on the project, standards are very, very important, especially open standards because you have to rely on it to develop and do your solutions in the process and in the software development process. Yeah, take input from user seriously, that's definitely a big part too and create small units like 50 PCs, we started with yeah, we'll do a couple of hundred but the problems in the beginning or the fear of problems in the beginning is simply too great, yeah, you just take one little department with like 50 uncomplicated clients and migrate that one and the people will see it's actually no problem, the people mostly like Linux client, they see no difference and how they work, they have their browser, they already know, we have Firefox and Thunderbird already deployed on Windows so there's no problem also and of course to keep communication up and involve the people, get people from the departments to work with you, meet regularly and always have one or two years open for the input. Yeah, the complexity was higher, that's no surprise I think, that's what sometimes happens in IT projects. Organizational solutions, yeah, okay, you need a second view, a revision of what is possible, we are not alone, that's one thing I really enjoyed while experiencing we're not alone, we have many other fighters beside us who are working in the open source community, you of course did a very big part of this fight for us when you developed Debian and we are very thankful for that. So if you have some questions in general, please get back directly to me, check those websites, check the developer blog, planetlimax.org, it's unofficial, it's what we do in our private time, it's mostly the developers and yeah, they will find the real news not on the official site of course. So I think that's it for the facts, how much time do we have left? Ah, okay, we're pretty good in time. So I think we can start with the both. Are you all okay with the copy session? I've prepared some ideas how we could work together in the future to get yeah, some joint effort going on. So if you want to, can you read it? Okay, okay, I just read it from up to down. Some ideas we had to get some more input from us to Debian and of course the other way around. So one thing is of course from our perspective let's become DDs or at least maintainers. That's one thing that's already in progress. I just applied two weeks ago and also Robert did. So that's one thing I think that's definitely a good idea to get some more input to Debian and back. Distribute our zero letter solution, the Walmart's with Debian, if you want so. I think we can do it. If you say, yeah, we have already so much work to do, I could maintain it, we're already working on it. So, Andres? I have a question. So you have written there, if you want so, what do you think, who is you? Yeah, you Debian. Yeah, but who is you? Well, the point is if you as a maintainer or as a maintainer in the maintainer queue think it is useful in Debian, don't ask. It is a bureaucracy, those who do something decide whether it gets done. And so you are also Debian and so you are, this is magical you, okay? Thank you, yes, absolutely, you're right. I should actually have written, if you don't want us to do so raise your hand because we're already in progress. So yeah, the next point is maintain Goza within Debian actively. The state at the moment is, I wouldn't say in an ancient version, but it's a stable version which speaks for Debian but it's not a version we would use at the city of Munich and we have put very much effort and money into the next version which is at the moment and still in better but will be released very soon. So I could do that part too. Of course with another maintainer who helps me because I have to get used to things and generally spoken more communication and interaction. An idea was probably to put our block feeds to planet Debian just to get more interaction from there. I don't know who is in charge for that. Please get in touch with me. I think it's, could be. You don't touch anybody. Yeah. Every week and every week. Yeah, okay. That's great. Good to know. So, yeah, I think I just said it. Okay, I'll write about it later after this talk. Yeah, be less derivative and more CDD like. Question is how we've been talking about it on Tuesday I think of on the drivers round table Andreas gave us some pretty good input I think and also there's been some thinking how to do it. We have some custom things we have to look about but I think it can be done and in the long term we have to do it to not become another Ubuntu or something like that. And yeah, from the technical point of view we just open our repository with the custom patches and yeah, let you decide if you want to use it or what to do it. If it could be for use for anybody out there in the world the UPL actually makes this now possible for us. We had definitely a problem before the UPL was published because we didn't know at the governmental institution how to deal with production of software because we're not allowed to do publishing or there was no experience in how governmental institutions publishing software because it's not very common. And yeah, the final idea I had was do BSPs on a regular basis. We have plenty of room in Munich so you're very welcome. And if the next release approaches I think for Lenny it's hopefully too late. Yeah, just get back to me or the email address and we'll have another BSP. So that's for my ideas if you have any other ideas just edit or could you say explain in more detail how you currently differ from Senator Debian or from your sources. So how much of your current client is customized and how much is just the standard stuff? It's a pretty small subset actually. We've got some patches which we filed bugs for also in the LibNSS libraries, some bugs which we discovered while working with our big LAP directory. So we have all users in the LAP directory. We've also all computers in the LAP directory. We have also our printers in the directory. So we've just had some issues with groups, group memberships and we have, yeah actually we have a new kernel just because we need to support new hardware in some way and at the moment we have new open office version. I think it's two for one with the Sun JDK but I hope we can give that part since Open JDK is published. So it's getting smaller and smaller and I want it to be as small as it can be. So currently your biggest changes are additions. Do you install additional software from? Addition patches. So we're just using let's say a little bit more flexibility than if you really add some software to Debian. I have some further question. I felt that several countries like I know from Brazil, from Venezuela and also Bhutan are also trying to do some government-driven efforts to build the distribution and I think most of them are deviant. Could you imagine that there is some cooperation to make things more fluently and so? Yeah, definitely. We're open for every collaboration for everybody who wants to benefit from our experiences. So we get visits regularly from all over the world. It's just amazing. You sit in the office in Munich and there's some delegation coming from Africa, from Japan, from the Netherlands, from Belgium. That's just, it's crazy but it's usual in Munich in the office where we sit. So I would prefer that really to get more people involved. Speaking of that, is there actually some collaboration between you and the people from the Extremadura region? Or... We've, as far as I know, we've met with them one time. Not me, but we did some exchange and Florian Schiesl, that's one of two project leaders. And he, yeah, it's been a year ago, I think, but he talked to them but we have no actual collaboration. I'd really appreciate it. We have some collaboration with Germany's Ministry of Foreign Affairs. So you probably were just read about it and yeah, that's the only actual collaboration we do at the moment. Another question, if it comes to Extremadura, could you imagine that Munich could sponsor some different developer meetings like Extremadura is doing? So just providing some place to sleep, something to eat and then work? Sure, absolutely. Why not? Yeah, no, actually, that's what we did the first time with the box-pushing party and that's what I would definitely do in the future. So just that the people come to one place and sponsor some pizza and yeah, do some hacking. In the back. You just mentioned very big migrations of Microsoft documents to open office. Are these tools, or which tools did you use for these conversions and are these publicly available or available in Debian? Actually, we did not use any tools to migrate those. It's mostly templates, macros and formulas but we're re-implementing them with the Walmox, which just does it pretty well. It kills about 20% of all those documents you talked about and of course we did just count those documents and look which one are just doubles or whatever else and this, yeah, it's just decreasing but we have no tool which is just taking all templates in or macros and throws them out in open office format. That's simply not working. The second question, you mentioned that you have your own version of open office relying on the Sun Java version. So I did see you have many resources, many development resources available. Why not help Debian with getting the free software stack around open office including OpenJDK better integrated into Debian? So behave much more like Ubuntu. It depends what you say when you say many development resources. We've got two office developers which are working on the Walmox and we've got two developers working on the Linux client. So we'll actually, we'll get eight new developers but we don't have them at the moment. So that's mostly the problem. We don't have the resources. Of course, if there's some strategic things we could profit off to just get it running with OpenJDK and give the part of building open office ourselves and packaging it. So we would do that, of course. Yeah, you asked about our comments and our questions and so it would be interesting. Do you have any comments? I mean, are there things that would make your life easier if they were solved differently in Debian or stuff like that? So do you have any comments you want to share with us? Yeah, sure, but it's mostly no technical issues. So I myself, I'm a technician. I can work with Debian, I can package. It's mostly no problem but the hardest thing is I think to get involved in the community, to get into the community. That's a pretty hard barrier. I think we have a cool team in Munich with some cool people. So this will be no problem, but I think other corporations or other governmental institutions will have them, I would call it mental barrier to get in touch if they have problem. They will just, yeah, okay. The technicians looks if Debian work for them and if it doesn't work, the boss says, no, we won't do it because you don't get our support and yeah, that's mostly a problem to get into Debian. So no further input at this moment. Steve. I'm aware that there's a team in Germany looking to host Debconf at some time in the future. Can we, are you allowed to be involved? Yeah, sure. Cool. Definitely. Yeah, you can all sleep at home. Where I live. I think what time is it? How important are the release dates of Debian for your client? At the moment, it's not really important because yeah, at this point in time, we're using somewhat in between Edge and Lenny. That's, yeah, it's a Lenny kernel with some patches and I think it's even the Lenny open office, I don't know, the version could match. So I think we will switch to time-based releases as Debian does or is doing. At the moment, we didn't do that. We just said, okay, what do you need for the next release? We have that much resources, it just didn't work out. So it will become more important. Regarding the business application you're using, can you just, there's a little bit more, what kind of problem you had? How many you could actually run in wine or some just, wow. Yeah, we got like those definitely unusual applications like the de-versions of archives with justice tags or yeah, every application you can imagine. Networker clients to do backups and yeah, what's those administrative application called that's not working pretty good in wine? You know it? Know what I'm talking about? Yeah, Net Support Manager, that's a pretty common application and of course those SAP Java clients which and yeah, not working that good under Linux or the web clients are much better and they're only working with the IE in of course and only certified for IE. We even did package one version of Mozilla browser 1.75 if I remember right because that was the last version of Mozilla certified for use with this web application. So it gets pretty weird at this point. Yeah, it's so much it's overwhelming. Ask me, just name it me, we will have it for sure. So any other questions? Yeah, I'm looking for the foreign, German foreign office and what I'm wondering is how we can defend against the big enemy. That's, I think this is the biggest problem we have because Microsoft just tried directly talk to our minister fortunately without success and now the minister is exchanging letters with the mayor from Munich which I'm much more happy about. And I have seen it in the office open XML debate where I have been very deeply involved. The open source and the open standards community started very late to defend the open standards and even then Microsoft left the war and they didn't defend their own open document format anymore because Microsoft just was so overwhelming. So I don't have a solution but I think that's quite the biggest problem we have when we want to migrate large parts of the world to open source. So it's not a technical problem, it's a problem of defending against a company that loses market and yeah. Yeah, I definitely have to say that most of the Linux project is a political part. So as I said, it's a small subset of changes we had to do to Debian in some way or another way to get it working for us but in general it just worked out and the great thing about the project in Munich is that I have some feeling that we can get them movement in here. So just work together with you guys and build another front, not against Microsoft but against proprietary standards and yeah, what keeps us from working efficiently. In our daily jobs. Any other questions? Steve? So what are the challenges you see coming up in the next few years? For us or for Debian in general? Well, for you first. Hello? Yeah, for you first. Keep our politicians satisfied, keep our IT satisfied and get our customers happier than they ever were. I think that's the easiest part because at the moment they just don't know. Do you make fun of me behind my back? Yeah. Okay, and how's about on the wider arena? What, did you ask me something? Yeah, yeah. And how about challenges for the wider community? I think what Mark Shuddle was said on Tuesday, one thing that encouraged me in some way was that when he said that we have to be aware of people in the outer world, outside of Debian have other objectives in using and developing Debian than the core Debian does. So not just do those nifty hacks we all love and like and where you can just spend nights on it, just to get it running somehow. The other side is some people need Debian for their daily jobs. It's not just us. It's hosts us all around the world. It's universities, whatever you want. So I think that's one important point and that's also one point I wanted to address when I said to get into Debian, to get Debian to acknowledge that there's really, yeah, I won't say a profession out there which has other needs, but raise the awareness. Cool. So how many of your team are currently actually in the new maintainer process or working towards it at the moment? It's two of them. Yeah, we actually have got four developers at the site on Tuesday. One is just, yeah, doing ticket work, prioritizing and that stuff. And Robert and me, we are currently two weeks ago, we applied and then new maintainer queue. Cool, good luck. Thank you. It's working, yeah. I told you in private, but I think we can also open that it's probably a good idea to hire Debian developer explicitly to some announcement on Debian lists and so it's probably a good idea. And perhaps also for other government-driven stuff then we get some work and you get some experts. Yeah, we'll do that, definitely, in the near future. You mostly told about, if I get it right, about the workstations, is there, I think there were a couple of things left. First, you still have a lot of applications that were not ported, that they are still running on wine or something. Yeah, or Windows PCs, whatever, yeah. Uh-huh, yeah. And the second one is, what about the server part? Do you still using NT or? Luckily, we never had any Windows server so we just could skip that part. It's okay, it's local government but it's not that pervert. We have many servers, we have AIX servers, we have BS2000 mainframes and yeah, but no Windows on the server side, luckily. Yeah, so if there are no other questions. Ah, can I give Kuna the microphone, please? I hope operating the camera doesn't exclude me from the audience. No, definitely not. I just didn't see you. No, I know, I know. No, well, you mentioned that this job is mainly a great part, a political one, so a couple of questions come to my mind on this. First, how dependent are you from the government continuing to be the same party or something like that? Second, how much can you influence? For example, well, in Europe there have been many very important developments regarding politics, the patents or the standardization of OOXML and whatnot, so how much can the experience in Linux be used as a political pressure? For the first question, not changing the party at the moment would definitely help, of course. We all know that if politicians change their minds and they do it frequently, but not in our case, luckily. So the project works pretty well. The IT works pretty well. The actual development at the moment works pretty well, so I don't see a point in changing anything. We were just approved eight more developers, so it's more or less for me a commitment to say we're doing it and we're doing it more than ever. And yeah, the question of experience, we're exchanging frequently and we'll do it further, of course. I hope and I think definitely that other communities, if it's countries, cities or whatever, can, of course, learn much from our experience. We had some pain to learn in our way. Any other questions? Guna, if not, I want to thank you all for attending. Okay, one last question. Do you think it would be possible to get the original written investment sheets and all the things that were used to convince the decision maker to actually switch to Linux to make them available online or some of the information available online so it can be used by all the... Actually the initial study and of course all of our slides are online. If you're with it, the urlmunic.com slash limux, you will find the study from Univention which led to the decision. And if you want some special information or details which you don't find online, just get in touch with us. You're very welcome. So I just say thank you very much. It's been a joy to come to Argentina after surviving the jet lag, of course. It's been a pain in the ass. And I would be glad to attend next step come with some more news and some more involvement from Munich in Debian. Thank you.