Council Manifesto of Sebastian Pipping (sping) v1.0.3, 2010-06-28 01:53 UTC+2 First, thanks for your interest! Please read on. In this document ================ - Gentoo in 2011: A vision - Building sites - Active Council - More direct democracy - New conflict resolution team (reforming DevRel) - Ownership and territories - The two sides of infra - The Git migration - Opening up documentation - Gentoo PR: Database and frequent news - Internal news (i.e. increasing transparency) - (Paid external) website re-design - Introducing templates (e.g. for bug replies) - Tone and attitude in Gentoo - Other (minor) things - Who is sping? Gentoo in 2011: A vision ======================== Gentoo's entry on Distrowatch recognizes us as an extremely flexibile distribution with a community of competent and friendly people around it. We are said to have resolved our struggles with a council practicing real leadership in 2010. We have moved to Git for the package tree and split it up to a satellite architecture thanks to proper multi-repo support from Portage 2.3. People are proposing improvements to each other, collaboration is pure fun in Gentoo. Due to that new contributors and core developers with a diversity of backgrounds join every few days. Our website reflects the replacement of our previous primary color purple to red, is shiny and well-organized leading potential contributors right into Gentoo as well as answering questions and problems of those seeking help. Our quizzes are reported to teach a lot and encourage to learn more without frightening off. People in Gentoo are experts on Bash and XSLT as we have the best tutorials including Gentoo specific sections around. Daniel Robbins joined back as a developer and quit Funtoo as he recognizes that Gentoo now is actually more fun than Funtoo, again. People fix and bump packages and documentation freely and resolve breakages in a cooperative way. "Sorry", "thanks", "explain" and "awesome" have become the most frequent words on IRC and the major mailing lists. Active Council ============== The council is a group of people elected by all developers to work on important issues. There seems to be the understanding that the council is meant to be reacting only, not pushing progress by itself. Depending on whether that's really set in stone somewhere or just a popular misconcepting we need to take one of these actions: A) Redefine the council to enable it to be active or B) Create another group working actively with adequate power or C) Just start being active as we could be Either way we will end up with a group of people that is elected by all developers working actively on the future of Gentoo. I do not believe in single leaders, but I do believe in need for joint leadership. I'm all for an active council. More direct democracy ===================== Certain questions have larger groups of both supporters and opponents in Gentoo: - Should Python 3.x be marked stable? - Should developer X be banned? - Should sexist jokes be allowed on non-public medium Y? - ... I would like to see a mean allowing to force a vote from the project as a whole (i.e. all developers) on a certain question. New conflict resolution team ============================ I have voiced my concerns about the status quo of the developer relations team recently. To my understanding the conflict resolution part of DevRel is currently not working effectively. Quoting my earlier mail I propose the following changes to DevRel: - Clearly split DevRel into groups for recruiting and conflict resolution with distinct aliases. - Let Gentoo developers vote on who's in the conflict resolution team just like we do with the council. - Disallow membership in both the conflict resolution group and the council at the same time (as the council is where issues with DevRel are taken to). With this new team in place they need to start acting on the base of the Code of Conduct, expressing warnings for instance. Demanding proper tone in Gentoo is everyone's job though, not just that of the conflict resolution team. Ownership and territories ========================= Especially with manpower of clearly below 300 active developers I doubt we should stick to a default policy of don't-touch-unless-yours-already. Alternative approaches like that rule's inversion with the addition of exceptions need to be discussed. Allowing masked non-maintainer bumps could be worth trying, too. The current rules are demotivating and block our progress on getting bugs fixed. The two sides of infra ====================== As rbu was thinking aloud to me once the infrastructure team is performing two very different kinds of tasks at the moment: - Maintenance (of existing services) - Development (of new services) While the maintenance of services (like keeping them available) is something that you need a "real admin" for, development of new services is something where advice of an admin may be helpful but is not a strong requirement. Developing new services, webtools, websites for Gentoo is something that should be possible for people who are not part of the infra team. Controlling about 50 machines centrally does make sense in various ways but does not blend well with these two kinds of jobs and to these two groups of people. While I don't need root privileges everywhere, I do need access to more places: for instance as a member of the overlays team I want to have permissions to fix the overlays machine no matter what the problem is. I propose decoupling of permissions of essential services from less essential services. I hope we can work towards this in the near future. Status quo is blocking developer potential currently. The Git migration ================= The Git migration is a good example of perfection gone too far in Gentoo. There's a good chance we were on Git with the main tree already if we would save things like thin Manifests and signed commits for the next round and switch to a Git converted tree without that without hurting future improvals. Again, credit to the original idea goes to rbu. I hope to see Robin integrate me with the conversion process. As I said to RadioTux: If this isn't done by the end of the year I don't know why. Opening up documentation ======================== People need direct commit access to documentation, independently of arrival of a Wiki. The two main reasons are: - Making patches and opening bugs on documentation is too high of an entry barrier. For instance making patches for types is something many people even start doing. - The doc team's resources are very low already We have to come up with a process that works reasonably well with the docs team. This callenge has to be addressed. Gentoo PR: Database and frequent news ===================================== People contact Gentoo PR with concerns like a list of interview questions on Gentoo. Among these are questions like "Is there a particular industry that is using your Linux distribution more than others?" Such questions are much easier to answer with a database holding - Companies using Gentoo with details - Detailed answers on repeated questions and misconceptions (sort of FAQ) Promoting and explaining Gentoo with that at hand would be much easier. Besides that database, PR is one of the points of Gentoo with a huge impact on the size of our userbase. We need to turn Gentoo PR from a delayed reply channel back into a machine producing articles and frequent news. The internet is waiting for us to come back. We need to be "louder" (i.e. more visible) to the internet. To make frequent news possible we need to come up with a way to collaborate on that front. Maybe moving that into a wiki could help. Also if we had frequent internal news keeping fellow developers up to date about progress external news could largely copy from that. Internal news (i.e. increasing transparency) ============================================ When I spoke to idl0r recently, I learned that the infra team is working on quite a few exciting things that I never read about anywhere. Keeping fellow developers up to date about your work is something I would like to see more of us do. If someone is asking me "what are people working on in Gentoo these days" I want to be able to have a good answer, at least after reading my e-mails or looking at dedicated website. Such change would have positive effect on external news and the ability to do better PR for all of us. Yes, people are interested in what you do. (Paid external) website re-design ================================= Many parts of the website of Gentoo are less than sufficient, especially the entry page. We can learn a lot from Fedora on that front: - Get Gentoo - Install Gentoo - Join Gentoo - .. The previous re-design that I saw did not go far enough with this. Some liked it, some didn't. I propose we start looking for professional designers, co-op with them on standards and XSLT so the result integrates well. The Gentoo Foundation should compensate such designer from their pool of finance. To my understanding we should be able to pay this task. The website is our face. I believe our face is worth paying a talented designer once. Introducing templates (e.g. for bug replies) ============================================ Templates work well when you don't notice there's a template. A template is a piece of text with well explained details that perfectly fits the current situation and saves re-typing the same thing over and over again with less quality. I believe templates could save us typing and sounding annoyed in a few places while transporting knowledge we'd be too lazy to type for at the same time. The overlay team already makes slight use of templates. I would like to see this applied in more places, helping both users and developers. Tone and attitude in Gentoo =========================== The feedback we received for the Gentoo booth on the Chemnitzer LinuxTag 2010 included a comment that the booth people were "competent and friendly". I want Gentoo to be a distribution with a community to be recognized as just that: competent and friendly. Being friendly can be hard work some times. I welcome everyone to join me with working for better tone and co-operation in Gentoo. Tone in Gentoo is a matter of everyone. Most of us need to learn more on that front, me too. Other (minor) things ==================== Other things that I think we could use in Gentoo (though I don't promise to work on these myself anytime soon): - An Installation checker A script possibly shipped with stage3 tarballs that checks for certain installation errors like unset root passwords or DHCP mode without a DHCP client installed. - A compile flags checker A script proposing and checking compiler flags based on sane recommendations and the given hardware and compiler. - A proper Python-based replacement for autounmask A script easing up unmasking of given packages and their dependencies for those using Gentoo stable. - An overlay version differ A service continuously syncing all registered overlays scanning for versions of packages newer than those on main Gentoo (i.e. detecting outstanding bumps) as well as versions older than those in other overlays (i.e. detecting cruft) Who is sping? ============= My name is Sebastian Pipping, I'm close to 28 years old. I'm a student of computer science at Technical University of Berlin close to finishing. rbu who studied with me before he finished had been using Gentoo long before me. Spending time with him and being dissatisfied with Ubuntu and Debian I came to a point where I felt waiting for a compiler was no longer reason enough to keep me away. So I finally gave it a try and fell in love with Gentoo instantly. The flyer of that time and the Gentoo handbook take a lot of that. Before I became a developer I participated in Summer of Code for Gentoo extending Smolt - a joint effort in collecting information on hardware setups by Fedora and SUSE - to Gentoo's interest in software setups. I joined in presenting Gentoo on fairs, turned SVGs into real Gentoo T-Shirts in co-operation with rbu of which we have sold more than 100 pieces already. That's 100 people promoting Gentoo through their clothes alone. As rbu is busy with his job these days I have been organizing the booth at LinuxTag this year. When Gunnar Wrobel left Gentoo Layman was at version 1.2.3. I took over Layman releasing 1.2.4 and 7 releases following up to the current 1.3.3. It went stable on x86 and amd64 a few days ago. rbu and I redesigned the XML format holding information of overlays allowing things like specifying different sources of Git repository. I joined the lead of Förderverein Gentoo e.V. when it was at risk of falling apart at the end of 2009. Besides Gentoo I'm a big fan of Git and Python but I start missing C++ these days, too. Teaching C to others is something I also enjoy doing - understanding pointers is key to that. Free Software is very important to me. I respect that choice is put even above that in Gentoo. Your vote for me is a vote for a more friendly, more healthy and more active Gentoo Project. Thank you! Sebastian Pipping