Title: How to be a good downstream Abstract: Taking a released tarball and packaging it for a distribution is a standard and well understood process for about every distribution package. But there's more to 'packaging than just tinkering with the source: Having good connections to 'upstream' makes life easier for both sides in the long run. This talk will present some of the non-technical aspects of the 'upstream-downstream' relationship and what a 'downstream' should and should not do. Speaker: Thomas Weber (Debian) Bio: Thomas Weber is a Debian developer and the current leader of the Debian Octave Group. Working with Linux since 2000, he mostly deals with packaging GNU Octave related packages for Debian. Time: 30 Note: not after 16:00 sunday Title: Swimming Upstream: How Distributions Help Open Source Communities Abstract: This talk will discuss the ways that distributions (such as Fedora) work collaboratively with upstream communities, and the value that distributions bring. It will also explore ways that distributions can improve their communication and collaboration with upstream projects. Speaker: Jared K. Smith (Fedora) Bio: (see below) Time: 60 Title: Ada in Debian and other distributions Abstract: This talk will: * briefly describe Ludovic's work maintaining the Ada part of GCC as part of the larger Debian-GCC team; * describe the Debian Policy for Ada packages; why one is needed and why it is written the way it is; * compare (in a constructive way, obviously) Debian with other distributions that also have good support for Ada, in particular Gentoo and FreeBSD, and with complementary third-party efforts such as the GNU Ada project on SourceForge[2] and MacAda[3] "GNAT for Macintosh"; * propose coordination between distributions on how to package Ada software. Followed by a live demonstration Speakers: Ludovic Brenta (Debian), Miguel Telleria de Esteban Bio: Miguel Telleria de Esteban - Universidad de Cantabria, Spain - telleriam@unican.es Miguel Telleria de Esteban is a Free Software engineer, Ph.D student and computer science researcher from the Cantabria region in the north of Spain. He started using Debian GNU/Linux in 2002 and keeps collaborating ever since with Linux User Groups BxLUG (Brussels) and Linuca (Cantabria region, Spain). He discovered Ada in 1998 through the lectures of Prof Michael Gonzalez Harbour in Cantabria and pursuited it a year later with the Software Engineering course of Prof. Alfred Strohmeier's lab at the EPFL. After a 5 year period of IT consulting work in Brussels (where he discovered Free Software), he returned to the University of Cantabria to start a research career on Real-Time systems in the same lab where he was taught Ada for the first time. Ludovic Brenta - Debian, Belgium - ludovic@ludovic-brenta.org Ludovic Brenta has been programming since 1989 and using GNU/Linux since 1994. He graduated from INSA Lyon in industrial engineering in 1996 and has been a software engineer ever since. In 2002, dissatisfied with the languages he used, he started looking for safer alternatives and discovered Ada, which he taught himself with help from the Free Software community. He started giving back in 2003 when he adopted most of the Ada packages in Debian and has been an official Debian Developer since 2006. Time: 60 - 90 Title: Zentyal Abstract: Zentyal (formerly eBox Platform) is a Linux small business server, that can act as Gateway, Infrastructure Manager, Unified Threat Manager, Office Server, Unified Communication Server or a combination of them. These functionalities are tightly integrated, automating most tasks, avoiding mistakes and saving time for system administrators. Zentyal is based on Ubuntu server, being able to manage services in a computer network like advanced routing, firewall and traffic shaping, DHCP, DNS, RADIUS, VPN, proxy, IDS, mail, file and printer sharing, VoIP, IM and more. The main challenge of the project has been going further than simple configuration providing integration between the managed services. Modules like firewall, DNS, CA, webserver or mail offer their services to other modules which can be configured to work together. It offers a layer of abstraction and management, with concepts like services, network objects, events or master-slave LDAP architecture between others. To archive these goals a specific service management and integration MVC framework was developed and improved during the 6 years project life. This framework has proved how easy it's to configure a new service or improve an existing module, experience we shared with the Perl community on last year YAPC::EU. This talk will present the project, its framework, how it integrates with Ubuntu and the plans to upload to Debian too. Also the difficulties that exist dealing with packaging policies and security profiles and how the project attempts to resolve them: configuration file updates: fstab, nsswitch, pam, etc. Speaker: Jorge Salamero Sanz (Debian) Bio: Jorge Salamero is a 25 years old, spanish software developer and system administrator. He studied Computer Science at the University of Zaragoza where he was involved in the Free Software University group. After his studies he worked in the Gluz2 (University of Zaragoza Linux Distribution) development, as system administrator in the GFN (Numerical Fluid Dynamics reseach group) and in the University IT Department. Since 2008 he is a Debian Developer. Currently he is working on Zentyal as systems analyst but also involved in development, training and community tasks. Time: 60 Title: Gentoo Q&A session Abstract: Do you have anything on your mind related to Gentoo? It can be technical, organizational etc. I will do my best to answer any question the time allows for and there should also be other Gentoo developers around if I am not able to. If you email me topics in advance (to betelgeuse@gentoo.org) I will spend time to research the topic before FOSDEM. As a backup plan in case the community is silent I will present the latest EAPI 4 and possible future technical directions for ebuilds. If you are interested in joining a Gentoo group dinner the group will come together at the end of the session. Please mail jmbsvicetto@gentoo.org well in advance if you are interested in attending. Speaker: Petteri Räty (Gentoo) Bio: Petteri has been a Gentoo developer since 2005. He has been part of the Gentoo council since 2007 and is currently the lead of the Developer Relations project. He is nearing the end of his studies at Aalto University School of Science and is involved in a couple software startups. Time: 60 - 30 Constraint: preferably last slot on Saturday. Title: Debian GNU/kFreeBSD - apt-get install freebsd? Speaker: Axel Beckert (Debian) Link: http://noone.org/talks/kfreebsd/ Abstract: This talk will show you what Debian GNU/kFreeBSD exactly is, what the differences between Debian GNU/kFreeBSD and the well known Debian GNU/Linux are, what difficulties the developers had, especially with regards to packaging and infrastructure, and what design decisions where made to get so far that Debian GNU/kFreeBSD finally has been accepted as Debian release architecture and what you can do with it but not with Debian GNU/Linux or with plain FreeBSD. It's also gives a little overview over similar projects. (There are more than you may expect. :-) Bio: Axel Beckert (Master of Computer Science, born 1975) studied Computer Science with minor Biology at the University of Saarland. Despite a master thesis in Artifical Intelligence, he always had a faible for networks, especially the world wide web, and Unix. Since 1995 he worked with different Unix derivatives and defenestrated his last Windows just a few years later. From 2002 to 2006 he worked as software developer with focus on web applications and accessibility at the Embperl forge ECOS. Since 2006 he works as Senior System Administrator of Linux and FreeBSD servers at the Dept. of Physics at ETH Zurich. He's a member of the Debian and FreeWRT distributions as well as with some other Open Source projects like Blosxom or Conkeror. Besides that he's board member of the Linux User Group Switzerland and co-moderator of the Swiss radio show and podcast Hackerfunk. Time: 60 - 30 Note: has lightning and mozilla talk, too. Title: Model Checking the Linux Kernel ? Speaker: Alexandre LISSY (Mandriva) Abstract: Linux Kernel evolves rapidly and, as every software, is subject to bugs. Some people are trying to find defects in the code by searching for known patterns and known misuse, documented either as part of the kernel API usage or after bugs being known. At Mandriva, and as foundation of my PhD subject, we want to explore another way which is well-known but complex: Model-Checking. Main goal of this talk will be to introduce the subject and present the state of the art around kernel code checking. Bio: Engaged in free software since a while with (French)Mozilla mainly around localisation (ChatZilla and now Addons.Mozilla.Org), Participation in various other free projects such as Kerrighed (RBT scheduler), and now PhD student between Mandriva and Tours' University Computer Science laboratory. Time: 30 Note: has lightning talk, too Title: Mancoosi tools for the analysis and quality assurance of FOSS distributions Speaker: Ralf Treinen (Debian) Abstract: we will present a set of tools for the analysis and quality assurance of FOSS distributions. These tools analyse meta-data of packages (like dependencies, build-dependencies, conflicts, etc), and answer questions like: - which packages are installable and which are not? - for which packages is it possible to install a build environment? - which packages are absolutely mandatory for the installation of a package (which might have dependencies with alternatives, or on virtual packages)? - which packages will certainly become uninstallable when I upgrade a certain package? These tools have been developed by the Mancoosi project, continuing the work started during its predecessor project EDOS. This will also include a demo of the new "Weather Service" for distributions. Time: 60 Title: Distribution collaboration manifesto Speakers: Stefano Zacchiroli (Debian) & Jared Smith (Fedora) Type: Panel discussion (with intro) Abstract: The FOSS distribution communities work toward very similar goals: each distribution aims at providing a coherent set of (Free) softwares to their users. A good deal of work (and patches!) can be shared across distributions and in part already is. The more we collaborate, among us and with our (common!) upstreams, the more we can improve Free Software as a whole. Sometimes, however, social obstacles get in the way of collaboration: volunteers don't like each other or each other distro; they might also easily get "religious" about the everlasting question "which is the best distro out there?". Understanding *why* we should collaborate across distros is the first needed step in improving *how* we collaborate. - Would it be useful to establish a "distribution collaboration manifesto" on how we should collaborate across distros and with our upstream? - On which shared values should such a manifesto be based? At least the following seem to be reasonable goals: (1) avoiding re-inventing the very same wheels in different distros; (2) increase productive communications on topics of shared interest. - ("Respect" has been proposed, but shouldn't we rather speak of "software freedom", "civility", and "collaboration"?) - *If* the goal of the manifesto is worthwhile, how can we practically work towards it? We would like to discuss these topics with all distribution which will be attending the distribution mini-conference! Bio: Stefano Zacchiroli: Stefano Zacchiroli has been a Debian Developer since 2001, working in areas of the project as varied as the maintenance of editors, programming languages compilers and libraries, math and XML software, as well as distribution-wide Quality Assurance. Stefano is currently serving as Debian Project Leader. Jared Smith: Jared Smith has been involved in various free and open-source software communities for over a decade. His interests include VoIce over IP, relational databases, programming, DocBook and other XML publishing tools, technical training, and systems administration. Jared currently serves as the Fedora Project Leader, and resides in Virginia, USA with his wife and two children. Time: 30 Title: Configuration data upgrade during package upgrade Abstract: For a casual user, editing configuration files in /etc in an intimidating task. Manually managing configuration upgrade is even worse as it requires good knowledge to merge current configuration data with new data coming from new packages. Config::Model was designed to make user's life easier by providing a configuration GUI and handle configuration upgrades transparently. This presentation will explain how to provide these capabilities and will cover: * the main notions or Config::Model (config tree and model) * how to create a model * how package and configuration upgrade is performed * how to specify upgrade feature in the configuration model * explain how Config::Model can be used from other languages through the fuse interface. * Show a demo. Important: This presentation does *not* require Perl knowledge Speaker: Dominique Dumont (Debian) Bio: Dominique is based in Grenoble, France. He's an active Perl module author since 1996, mainly on Tk modules and now on his Config::Model pet project. Dominique has also been using Debian since 2000. He's contributed occasionally to Alsa and DVB projects to debug his home cinema PC. He's now member of Debian-Perl team and takes care of packaging all Config::Model modules and others Perl modules. Time: 60 Constraint: no later than 15:30 sunday, perhaps other talk too. Title: Deploying sssd Bio: Based in Zurich, Marcus is a member of the CentOS project where he contributes in terms of artwork, wiki content and gardening and as an ambassador. He is also active in the Fedora Project in the roles of ambassador and package maintainer. He works as a linux systems engineer at the ETH Zurich where the biggest part of his job is to provide, maintain and support the central linux systems lifecycle management solutions for Fedora, CentOS and Red Hat Enterprise Linux. To be able to fully accomplish his job he also need to perform intense testing/QA and come up with lots of new ideas to improve those products. Speaker: Marcus Moeller (CentOS) Abstract: The complexity of configuring client authentication with Kerberos or LDAP is well known. With sssd the whole process has been simplified. We are going to take a look at the framework and supported identity and authentication provides. Some examples will show the capability of sssd. Note: through Ralph Angenendt Time: 30 Speaker: Dieter Plaetinck (Arch) Title: Can we build a simple, cross-distribution installation framework? Abstract: First, this talk will entail the philosophy behind the shellscript-based Arch Linux installation framework (AIF), as well as some implementation information and lessons learned. Second, I want to initiate a discussion on how multiple distributions could work together maintaining an installation framework following a "keep it simple" philosophy, and how feasible it would be to do so. In my experience, quite some necessities, issues and features are distribution-independent. Specific distributions could be supported with plugins or feature branches. Bio: In daily life, Dieter works on a high-performance search engine project at the university of Ghent, in his spare time he prefers building simple tools to make his life easier, usually related to his distro of choice, Arch Linux. He maintains the installation framework and builds the Arch release media. Until recently, he led the Uzbl browser project. His preferred languages are bash and python. He also plays drums. Time: 30 - 60 Note: preferably between 10 and 12 Title: piuparts.debian.org: dealing with several k of policy violations Speaker: Holger Levsen (Debian) Abstract: piuparts.debian.org tests all binary packages in Debian whether they are good citizens or not. This has produced roughly 90k logfiles, showing roughly 10k policy violations, though less than 500-1000 are realistically fixable (and a real problem). This talk will feature a short intro into piuparts(.debian.org), followed by a presentation of the challenges faced and shall end with a discussion about the strategies and goals for making wheezy a even better release. Bio: (see FOSDEM 2009) Time: 60 - 30 Title: Gentoo's RelEng and catalyst use Speaker: Jorge Manuel B. S. Vicetto Abstract: A short introduction to Gentoo's Release Engineering[1] team and its current model to produce media to install Gentoo. An introduction to the the catalyst[2][3] tool and a description of the current configuration to produce our official stages and isos, specifically the weekly autobuilds. A short demonstration of how to configure and use catalyst to create stages and isos on one's hardware. [1] - http://www.gentoo.org/proj/en/releng/ [2] - http://www.gentoo.org/proj/en/releng/catalyst/index.xml [3] - http://git.overlays.gentoo.org/gitweb/?p=proj/catalyst.git;a=summary Bio: Jorge Manuel B. S. Vicetto is a Net/Sys admin for the Angra do Heroísmo Hospital at Terceira Island, Azores. He used Linux for the first time in 1995, switched to Gentoo on 2004 and became a forums moderator on 2006. After getting his Gentoo Developer cloak, he has worked for a few teams including User Relations, KDE, Desktop-Effects, SPARC, Developer Relations, Elections and Release Engineering. He's currently serving in the Gentoo Council, focusing on the weekly autobuilds for amd64/x86, helping out with mysql and doing some occasional KDE work. Time: 60 - 30 Note: Preferably not too early Title: Gentoo's Reform and Future (debate) Abstract: The purpose of this talk is to present the view of a few Gentoo Developers that are part of the Council, Foundation and Developer Relations about the future of Gentoo and its reform. We'd like to start a debate involving other Gentoo Developers and interested attendees about this. Speakers: Jorge Manuel B. S. Vicetto, Petteri Räty, Roy Bamford, (possibly other speakers) Note: preferably right next to the Q&A one Time: 60 Title: Distributed Compilation of RPMs Speakers: Brian Schueler and Geerd-Dietger Hoffmann Abstract: Software packages are steadily increasing in size (lines of code) and many tools do not take advantage of the multi-core paradigm and still only compile on one core. This has the effect that a huge increase in compilation time is taking place. Some packages depend on other packages which have to be built beforehand. The easy and time/resource consuming method is to build all packages in a continuous loop and skip failures till everything is finished or do it by hand. The better method is to build the packages in the correct order from the start in only one pass. This implies some analysis beforehand to construct this order. In the talk a tool will be described that does exactly this for rpm packages. After the build order is defined the actual build can take place. On the Operating System level this can include hundreds of rpm and take a very long time. (Days is not unusual) To shorten this, the build process can be separated on numerous machines. Especially compilation jobs can be distributed over a cluster. The icecream distributed compiler is introduced in this talk to enable this. It takes care of the right gcc compiler for the right CPU architectures being used on the nodes, which can be of different Linux distributions (i.e. Debian, SuSE, or something else). Some complexity is added as the main compilation has to be done in a chroot and this needs to be synchronised. It is also possible to do this in the Cloud by using specific Virtual Machines which will also be described in the talk. The idea is that in the end we can take a repository of source rpms, process them on a cluster in a highly distributed manor and end up with binary packages Bios: Brian: Engineer employed at the Beuth University of Applied Sciences inBerlin. Laboratory for process and systems technology - main subjects:networks, autonomous mobile robots, computer vision, process technology, field bus systems Homepage: http://public.beuth-hochschule.de/~bschueler Geerd-Dietger Hoffmann: has just started his PhD at the Technical University of Berlin. Switching between industry, including CERN as well as IBM, and completing his BSc in Bournemouth and MSc at University College London with 1st class honours he has engaged in a wide range of topics. He is the creator of the award winning Objic programming language and involved in various Open Source projects. In recent presentations and papers he is showing a growing interest in the future of Cloud Computing. Homepage: http://www.ribalba.de Time: 30 Title: Towards the Comparative Analysis of Evolving Libre Software Distributions' Environments* Speaker: Leandro Doctors Bio: Leandro Doctors has been involved in the Libre Software Community of San Juan (Argentina) since 2002. In that year, he co-founded (and chaired for three years) the first Libre Software Users' Group of this Argentine Province. In early 2008, he graduated as Licenciado (the equivalent of an European Masters) in Information Systems at the National University of San Juan. As soon as he could, he (looked for and) took the opportunity to travel. After a beautiful one-year-long internship in Lisbon, Portugal, Leandro finally realized he didn't enjoy what most people would call a “normal” job (at least, not for the moment). Finally, in late 2009, Leandro received an offer to “escape” the “industry” while combining work and pleasure. Since then, he is a full time PhD student at the University of Mons (Belgium). His research topic involves the comparative analysis of the evolution of libre software distributions' environments. Altough Leandro is currently out of shape, he loves tango dancing, language learning and travelling. He also likes trekking (andinismo) and swimming. Abstract: Most empirical studies on the evolution and quality of libre software focus on *individual* software projects. Instead, we aim to empirically study, analyse and compare software *distributions*. We are interested in understanding how the environment that surrounds the considered systems affects their evolving quality. Therefore, we take into account different points of view from different stakeholders. Our goals are: *Help developers to asses, control and improve the quality of the project they work on (in this case, software distributions) *Advise users on quality criterion when choosing a software distribution *Gain a better understanding of evolving software environments* As a pilot case, we use and analyse package-related metrics for Debian and Ubuntu using data extracted from the Ultimate Debian Database (UDD). We observed an increasing amount of sharing between both projects' communities since the establishment of the Ubuntu project, in 2004. Even if the standard UDD has been our first choice as both data source and analysis tool, it has not been as effective as we first thought. We are now looking for other data sources and (reusing existing and/or building new) analysis tools that allow us to extend this study to other distributions. *We define a project's environment as the set of all possible artifacts, entities (communities, projects), processes (development and business models) and other aspects that might influence the system under study. Time: 60 Speaker: Dominique Dumont Bio: (see above) Title: Sharing package description (and translations) between distros. Abstract: This session proposes to share concrete ideas, gather requirements and wishes (and hopefully volunteers) towards an implementation to help collaboration between distros teams. (The ones that don't talk much together otherwise). The basic idea revolves around creating a central DB that would collect package description. (ok, that's not new). To use this data, a model (à la Config::Model) will be proposed that: - use the common descriptions through a DB backend - represents the package description within the distro's packaging system (Such model is already available in Debian's libconfig-model-perl package. A model for RedHat's spec file would be required to implement a proof of concept) - the common data would be used as default value for the distro's package description with Config::Model's "compute" mechanism. This system could also be used for data that are common between distros like license, summary, upstream home page. One thing is still to be defined: how to update the "common" descriptions and translation. This could be done through Config::Model or by another means (wiki, trusted team ...). Discussing this point at FOSDEM could be useful. Unfortunately, the author probably will not have time to implement a prototype until FOSDEM, but the idea can be discussed to see if other distro could be interested. To start the discussion, a short presentation of Config::Model capabilities will be done (mostly around the compute feature to get data from the common pool to distro teams). Important: - This presentation does *not* require Perl knowledge Time: 60 (or 30 if right after the first) Constraint: no later than 15:30 sunday, perhaps other talk too. Bernhard Wiedemann (bernhardout@lsmod.de) Bio: Bernhard M. Wiedemann is a long-time free software enthusiast and software developer. During the last decade he contributed to coreboot, qemu, VirtualBox, openSUSE and other projects. Having joined the openSUSE Testing Core Team in mid-2009, he started writing OS-autoinst (an automated testsuite for whole operating-systems) in April 2010 that has since helped to discover dozens of bugs. He is now employed by SUSE LINUX Products GmbH. Title: How to make QA-engineers start drooling - or -, How automated testing can help distributions. Abstract: Testing is an important task. But testing daily development-snapshots of a Linux-distribution would mean testing the same things every day, which would be pretty tiresome to testers. Yet, no one likes to break his system with untested code. This presentation will be about my solution to this problem, using automated testing with KVM for openQA.openSUSE.org to guarantee some minimum quality and about how the framework can be adapted to test other distributions like Fedora-Rawhide. It will also have installation videos which are cool, fun and useful. Time: 30 - 60 Speaker: Dominik Heidler / Duncan Mac-Vicar P. Bio: Dominik is an open source software developer working for SuSE Linux since 2008. He is a member of the YaST-Team and develops ZYpp, the package management system of openSUSE. Title: ZYpp your distro Abstract: ZYpp is an opensource project originated in the openSUSE distribution which groups the most advanced and fast dependency solver (satsolver), a package management library and a command line package management tool (zypper). Over the years ZYpp has matured to become one of openSUSE most appreciated features. This talk will show ZYpp's main features and work being done to have it running on other rpm-based distributions. Time: not specified, 30 Speaker: Adrian Schroeter Bio: Title: Once source to rule all binaries Abstract: After it is possible for users of software to easily escape the famous dependency hell with smart and user-friendly package managers this talk will show you how free and open source software developers can escape the next circle - The Free Software Distribution Hell. In today's Linux market you have dozens of distributions and the same amount of ways to package software. Users don't want to bother with source tarballs anymore and expect a ready to install binary from you. Standardisation is still in its early stages and only possible for a very limited subset of what distributions offer because after all they strive for differentiators to be ahead of the competition. If you want to provide your software to your users today you have to have it for Fedora, Ubuntu, Debian, Mandriva, openSUSE and possibly more. Of course there are 3 different versions for each distribution out there building for at least 2 different architectures. Just to ship your software, you need to create all these packages - Welcome to the Free Software Distribution Hell. But don't worry, we will show you how to escape it with the openSUSE Buildservice, an open and complete packaging platform. Time: not specified, 30 Speaker: Jos Poortvliet Bio: Jos Poortvliet is since 6 months openSUSE community manager for SUSE Linux. Before that he was active as volunteer in the international KDE community as team lead for the marketing team. He coordinated the communication around releases, the creation and distribution of marketing materials, visiting and organizing conferences and staying in contact with the press. In his 'working life' he has had jobs at a variety of companies like the dutch telco KPN, Royal Bank of Schotland and the dutch Department of Eductation. Title: Amazing openSUSE Abstract: openSUSE has developed both great infrastructure technology and a powerful enduser product; where is it going now? Over 5 years ago Novell decided to turn their new asset, SuSE Linux into a community distribution. It was a long and laborious process but since almost 2 years now the development processes have been opened up to the community. We now speak of openSUSE, a community which is now searching for it's direction. The project has state-of-the-art infrastructure at its disposal and develops a modern, stable and powerful linux distribution. Meanwhile, communication and marketing are it's weak points - innovative technologies like the openSUSE Build Service and SUSE Studio do not get the exposure they deserve. In this talk an overview is given of openSUSE's history, the developments in the community and the latest openSUSE technology is presented. Time: 60 - 30 Speaker: Hans de Goede and Michal Hrušecký Bio: Hans de Goede: Hans has been a Linux developer since 1996, working on a wide variety of projects, including maintaining 200 packages in Fedora, various hwmon kernel drivers, rewriting and merging many out of tree webcam drivers into the mainline kernel, libv4l a userspace library to transparently handle the decompression of many proprietary webcam video formats in userspace, anaconda the Fedora / Red Hat installer, parted (the partition tool) and lately Spice and usb-redirection under qemu. He has also been a long-time Fedora contributor and maintainer of many packages in Fedora. As such he regularly works together with people from other distributions and often see problems with projects where upstream is dead in some way. Michal Hrušecký: Michal started contributing seriously back to the open source in 2007 by participating in Hacking & Development (project to run Linux on Palm PDA - some kernel patching and some OpenEmbedded packaging and bugfixing). He's also administrating school computer laboratory since 2007 (maintaining a little Gentoo overlay for that purpose). Joined Novell/SUSE Linux in 2008 and started there as a packager maintaining mainly MySQL, Palm and sblim related packages. Since 2009 he's part of openSUSE Booster team, so in addition to the packaging he's giving talks, doing workshops and working on projects like openSUSE Connect, openSUSE Paste or the wiki rewrite. Title: Downstream packaging collaboration, patch sharing or start a new upstream? Abstract: Most distributions contain packages where upstream is completely dead or very close to it, yet there still is a significant user group for the packages in question, be it directly or through other packages depending on them. But even when upstream is alive, packages often needs some additional patching. All distributions seem to end up carrying some custom distro grown patchset, with some patches being shared and other patches being unique per distro. Clearly not an ideal situation and one which most distributions try to avoid by feeding all necessary changes upstream and aiming for having packages which use the upstream provided code as is. Clearly the usual model of sharing code fixes / integration work done at the distribution level between distributions by pushing it upstream does not work for projects where there is no upstream. And even with upstream around, distributions sometimes needs to fix bugs in versions that upstream no longer care about or fix bugs that are not important/considered feature upstream and it may take a long time to get them fixed there. In the end we can't avoid all patching. But we can try to avoid duplicating the work by better sharing the results. The plan for this sessions is to give a short presentation sketching the problem, show two possible approaches to help solving this problem together with few generally helpful ideas. After presenting these hopefully somewhat structured discussion with the audience about this topic is planned. Constraint: Hans does another talk 14:00 - 15:30 Sunday. Time: 60 Title: Pardus and Pisi: Distribution from Scratch Abstract : Pardus is an open source, freely distributable, GPL licensed user-friendly GNU/Linux distribution with many usability, hardware compatibility, stability and security enhancements. The project is sponsored by the Turkish National Research Institute of Electronics and Cryptology (UEKAE). PiSi (Packages Installed Successfully, as Intended) is the package management system of Pardus. It is the tool that installs, upgrades and removes the software packages successfully, as intended. PiSi stores and handles the dependencies for the other packages, libraries and COMAR tasks. You may install and use software without being aware of the underlying technical details, not even the name of the application. Tasks such as opening a new file type or serving web pages are just a click away with PiSi. Some distinctive features of Pisi package management system are the following: * Implemented in Python * Efficient and small * Package sources are written in XML and Python * Integrates low-level and high-level package operations (dependency resolution) * Framework approach to build applications and tools upon * Comprehensive CLI and a user-friendly Qt GUI (distributed separately) * Extremely simple and fast package construction Speaker: Gökçen Eraslan Bio: Gökçen Eraslan is open source software developer working for Pardus since 2006. He is package maintainer and one of the 2 release managers of Pardus 2011 which is the next stable release of Pardus, also contributes to KDE, Firefox and LibreOffice projects. He is still working as full-time developer for Pardus. Time: not specified, 30 Title: COMAR: A new approach to system management Abstract: COMAR is COnfiguration MAnageR of Pardus Linux, which is a DBus service that helps the installed software operate flawlessly. COMAR knows the tasks that can be provided by each application, together with the functionality they depend on and other information. Different applications may adapt themselves according to the presence and capability of their peers. It provides to achieve privileged operations for a normal user by asking users privileges to PolicyKit about the operation. These privileges defined by the program itself and its possible to define atomic privileges for each method. It makes system management very flexible and extendable. For end-users Pardus provides a series of graphical management tools which are written in Python using PyQt/PyKDE and uses COMAR calls for the operations at all. Speaker: Gökmen Göksel (Pardus) Bio: Gökmen is open source software developer working for Pardus since 2006. He developed the Pardus' earlier versions installer called YALI and also lots of graphical user interfaces for management family of Pardus. He also developed some of KDE based technologies for Pardus. He is still working as full-time developer for Pardus. Time: not specified, 30 Title: Example of cross-distro collaboration: Application Installer Abstract: Nearly all distributions are slowly moving towards making it easier to deal with applications, and get information about those. However, many of us are working in our own corners: unfortunately, this is often what happens with many topics that would move faster if they were looked from a cross-distro perspective. A small team of people from various distributions will meet in January to discuss what can be done in a cross-distro way. This talk will present the conclusions from this meeting, and will come back on the experience of organizing work in a cross-distro way. Bio: When he’s not procrastinating or eating ice cream, Vincent Untz shows his real face: he is an active Free Software enthusiast, GNOME lover, and advocate, and openSUSE Booster. When describing his contributions, Vincent finds it easier to declare he is a “touche-à-tout”, working on various (some say random) areas of the desktop: code, release management, promotion, bug triage, specifications… Vincent is working at Novell. Speaker: Vincent Untz Time: 30 Title: Using NixOS for declarative deployment and testing Presenter: Sander van der Burg Abstract: Linux distributions are often used in a distributed setting, such as web application environments containing web servers, databases, proxy servers which are all connected to each other. Manually installing and configuring networks of machines is often a complex and time-consuming process. Upgrading is even more difficult, as this may render (parts) of a system inaccessible. Furthermore, it is also difficult to test a distributed environment, since it is usually too expensive to set up a test environment faithfully reproducing a production environment. NixOS is a Linux distribution built on top of the Nix package manager, a purely functional package manager. In NixOS, a system configuration is built from a declarative specification capturing the complete system configuration. Moreover, because of the Nix heritage, a user can safely and efficiently install multiple configurations next to each other without interference. This allows users to perform atomic upgrades and rollbacks of a complete Linux distribution. In this talk we explain how we can use NixOS to manage a network of machines by using a declarative specification of a network of NixOS configurations. We can use this specification to: (1) Automatically, reliably and efficiently install or upgrade a network of physical machines (2) Deploy virtual machines in a cloud infrastructure (3) Generate networks of virtual machines to automatically test a distributed environment with complex environmental dependencies. These virtual machine instances use components stored on the host machine, so that the same components can be shared across multiple machines and no disk images are required, making virtual machine instances relatively cheap. Bio: Sander van der Burg is a PhD student at the Software Engineering Research Group, Delft University of Technology, The Netherlands, working on the Pull Deployment of Services project. He is an active contributor of the NixOS project, a GNU/Linux distribution built around the Nix package manager since 2007. As part of his PhD research he has developed Disnix, a distributed deployment extension built on top of the Nix package manager, to deploy service-oriented systems in networks consisting of machines running various operating systems and architectures. Time: 60 - 30 -------- Talks past this point were submitted after the deadline had passed, or were submitted with a bit of an optionality to it. They're listed in order of submission, and will only be accepted on the schedule if, after all of the above talks have been scheduled, there is still time left to have them. Speaker: Bdale Garbee Abstract: HP and Community Linux distributions Note: optional talk (filler if the schedule doesn't get full otherwise) Speaker: Anne Nicolas Abstract: Mageia project Note: past deadline, first on waiting list. Speaker: Stefano Zacchiroli Title: Who the bloody hell cares about Debian? Note: past deadline, second on waiting list. Speaker: Paulo Trezentos Abstract: apt-pbo Note: past deadline, forwarded from lightning talks with some others. Time: 30 Speaker: John Thomson Title: Improving Rollback in Linux via DSL approach & distributing meta-data Note: past deadline, forwarded from lightning talks with some others. Time: 30 Speaker: André Guerreiro Title: Gumby: a package dependency Browser/Viewer for Linux Note: past deadline, forwarded from lightning talks with some others. Time: 30 Speaker: Sami Anttila Title: Introducing the Flexible Worker Pool Note: past deadline, forwarded from lightning talks with three others. Time: 30