 Thanks for having me. It's great to be one of the most exciting immigration and it might be interesting for all of you. Of course, it is one of the immigration projects that still exists and lives. And hasn't been killed for some reason or another. So I just want to give you some ideas of what was done, why it was done, the techniques behind it, and perhaps an outlook to the future of this project. So let's get started. That's about me. I'm working as freelancing software consultancy for food software. It's very easy to make a living out of that. And since I live in Cologne, sometimes I celebrate Mardi Gras, hence the nice picture. Okay. I built my presentation like a Shakespeare drama with three seats past the present and the future with the past. I just want to get you into how they, and why they decided to migrate to a living space environment. Okay. So the main trigger has been support for Windows and for this running out. And they didn't really know what to do. So they had some companies make some plan how, what could be done in the future. And like Linux, and since Susan and IBM were doing this investigation, they came up that the costs were almost equal. So the costs were not a factor. But they decided to become vendor independent. That was our main strategy goal. They didn't want to depend on one vendor telling them when to upgrade their operating system and how to do it and how much money to pay for it. So they officially decided in 2003 to make the Linux migration on the desktop. And the defined goals were to move 80% of all of the workplaces towards Linux that were roughly 12,000. And of course, I wanted to get rid of vendor documents. And they had to go to strengthen local, small, medium businesses. So to keep the money in Munich or in the area or in Germany by paying smaller companies to do the actual implementation, which was a very nice goal. Of course, they had to take care about all these custom applications and they use for texting docs and I don't know, planning on bicycle rides. So they had to work on that to make them either run in the web browser or use something like Java that is available on many platforms. So I just want to give you a small overview of the structure of the city of Munich. Munich is one of the most beautiful German cities. I think it's the third largest with 1.4 million people living in it. And the administration is rather complex. There are 13 different departments. And at the beginning of this project, they had 13 local kings, local IT kings. And they all had their own IT infrastructure. So some people were running novels, some people were running, I don't know, some Windows emulation server stuff and some people bought hardware here and some people bought hardware there. So it was rather complex. But no real Windows server, which is a good thing if you want to migrate towards the limit space this time. And they also didn't have a single high service, Windows in the print server. But something they did have and it was an X500 directory server. And that was very useful for us when we did the implementation of the client and the infrastructure to actually deploy it and roll it out. We at this point have just released the client version 4.0. It's the first client based on KDE 4.0. And I will just give you some facts about what's in it. The original client was based on Dimeon. And for some reason or another they switched it into which is not rather good because we have these more stable releases and these clients get deployed by a solution that also uses 3.0 software. All the configuration stuff gets stored into a directory server. So we are using this directory server that already exists for users and put all of our system config stuff in it. And we have a tool called GOSA that is like the front end for the directory server to configure systems and groups and printers and which users get switch shares and stuff like that and which applications get installed for which user group and so on. So it's a bit complex but rather flexible and you can do any magic you can possibly think of because you can also put scripts in it for systems or scripts for users or scripts for group. So you can just go crazy with this. And for the actual deployment we use FILE which was initially a script set for rolling out and clusters, Linux based clusters. And we brought Linux based clients with it and it's working fairly well. So as I already told you all this stuff gets stored into a directory server which is a bit unusual but it's working. So what change? And since we are using a Debian based distribution of course all installations are packet based so you can really configure which Debian package gets on which version. And if you do a new installation it will take about 50 minutes. It will roll out about two weeks of data on the client and then it's ready to use. And that is amazing if asked because a few we also have some Windows clients now it takes about two hours. And we have a nifty little script called Profile Sync so users can get their person or their person set up like wallpapers and configuration of KDE specific things or all the rest of the specific settings. They can synchronize to the actual machine they are locked in so we do provide roaming profiles and that is very nice. So it doesn't really matter where you log on you always get your own environment. Any questions so far about the technical things yet? It's okay, okay, thank you. Just to ask you how long do you take the time to implement what the 15,000 pieces in a municity to contract the IMOX as to rather implement it? That is very hard to answer. Because technically it's no problem at all but as I told you before and there were 13 different departments and they still exist and they have completely different demands. For example, the Department of Cultural Things is very easy to handle. They just use their computer for checking emails and browsing the web and writing simple documents. But other departments are very complex and they use hundreds or even thousands of special applications. And with these departments and the migration it was really painful in the took years to do it. But technically it's like a snap of a finger. Very easy. But it's hard, I can't tell you how long it took. The first client was ready in 2006 and the official end of the project was I think in the end of 2012. So it took six years to reach the project goal which was longer than expected. Does it answer your question? Six years to go. You are still maintaining it. Yeah, now the project stopped and we are now into mainline and we are maintaining and developing new things and stuff like that. But it's not a project anymore. Yeah, it's good. Thank you very much. Okay, so if you read the Linux in the press or somewhere else, it's not just a client which is very easy to implement actually but it's a whole ecosystem. The system and conflict management and deployment mechanism. So it's more than a client. It's really a whole setup to easily deploy and configure in the best desktop settings you need. We made one small picture for you so you can see how everything depends on each other. It's a bit complex but if you want, I can explain you perhaps later. Just a few things about the brand new client. I already told you in space when you went to 12.04 and there's a nice thing. They opened up people who invented the hardware enablement stack. So the back port for you, the kernel and the graphics package for all the releases. So you can support newer hardware with your older release which is very much appreciated. Of course, the users have enablement stack we use kernel 3.5.13, KDEF 4.10, Firefox extended service releases, for a standard and 31 is available, same for Thunderbird. And we have one dinosaur that is already calendar. Which is a very nice thing. It still uses motive anymore but hopefully we are switching to co-op soon because it will be some more than user-friendly environment for mail and kernel. Okay, so much for the present. It's no magic and the client itself is just plain kubuntu. And there are some special applications, obviously, the CDO for Munich needs but there is nothing special really bothered. Except a set of scripts to configure the actual shares of the printers and stuff like that. Which are written in a puppet a little bit hamster and user stuff. So this project got a lot of attention and attention in the world I guess when it started. Now it's very quiet with audit and it switched from project mode to mainline mode. For me it's hard to guess what the future will bring. I hope it will be prosperous. But I just want to figure out some options. One is obviously not a good option, Romain and Julieta. Because it's a city administration, there are many people that want something for you. So the people that actually develop this client, they get lots of jobs that are not related to the client. It's so they really work on stuff they shouldn't really do. And one thing that can happen that some key applications like for example, LibreOffice goes down because the code is very hard to maintain. It's like millions of lines of travel. And we don't know if it's useful in two years time. We hope so and we heavily contribute and patch on LibreOffice. But we'll see, it's a factor you have to consider. Another thing, every year or two some press campaign starts triggered by, I don't know, and then everybody is very much excited and lots of talk and the politicians might switch side and say, okay, we don't want this new stuff anymore. It's causing us trouble and negative press and let's go back to where we came from. Another danger, as you can see, we also deploy Windows-based clients now. We have about 2,000, I guess. And it's possible that Windows takes over the Linux resources, we will see. And one problem we have set with vendors for special applications. They don't know what to do with Linux. You know what it is sometimes. They make some applications running on a visual basic, ancient version. And now it has somehow to be transferred to Linux and sometimes it's a very small company with two or three people and two people left. So it's a problem. Of course, if you have such a huge amount of Linux-based test talks, you need a three-in-one software strategy because your whole IT must be built around these clients. And if you don't have this or if you don't communicate clearly what should be done and what should not be done, you run into trouble, that's for sure. So in the end, all that. But I will end with a positive outlook because Linux is well established. 15,000 workstations don't go away easily and the administrators like it because they can very easily configure everything they want and it's just working most of the times. And installations are quick and painless so the administrators like that. If you ever did an administrator like Active Directory and stuff, you know what I'm talking about. Most users don't really care as long as they can do their work. So they don't care whether it's GNOME or KDE or XFC or whatever, they don't care. They have to browse the web, they have to write their letters and they have to send emails and get emails. So they usually do not complain. If they complain, they complain about something not related to their desktop environment. And something I really hope for is that some bigger communities or cities or whatever will follow this example. So the city of Munich will not be alone because it's very hard to fight on your own all the time. And once some cities or communities or whatever are followed, they can have some synergies because they can make some, I don't know, frameworks, even frameworks for their needs and share the code and work on that. So I very much hope that there will be floors. And I also hope that free and open source software becomes common in the near future. So everybody is using it and everybody is already using it on their Android phones or on their Samsung TV. So it's nothing special but it's just normal IT strategy to implement and use free open source at the end of my talk. Thank you very much.