So it's election time again... I'm from Regensburg, Germany and by profession an experimental physicist leading a university research group. Outside physics and Gentoo I'm interested in a wide range of topics all the way from art and cultural history to scifi and cocktail mixing. (Doesn't necessarily mean that I have time for all that though.) I'm a Gentoo developer since 2010; I decided at some point that if I'm tinkering with the computer in my free time anyway I might as well do it productively. Today I mostly work on keeping toolchain stuff like glibc and binutils going, and solving odd problems here and there. Council business and occasionally Comrel eat up my free time. At some point I got interested in Gentoo history, and started assembling the council decisions index [1] and updating the Gentoo ecosystem poster [2]. Gentoo runs on our university PCs and is controlling our measurements. So, I see a solid stable set and clean upgrade paths as very important. Conversely, ~arch gives us the unique opportunity to provide our users with the bleeding edge of code, and that's something we should use, value and advertise too! Gentoo is about providing choice, not locking anyone into a single solution - so a wide range of projects is great - but on the other hand projects should ideally be structured in a way that they don't block each others' progress. There's one thing that Gentoo is really bad at, and that is making strategic decisions and actually implementing them. We discuss controversial topics on the mailing lists, but activity there is not necessarily a good indicator of how opinions are actually distributed across the developer community. If you ask 10 persons, you'll get 11 different opinions on how things should be done, and at least 4 people with completely different opinions will be fully convinced that they speak for the majority! As consequence, what Gentoo needs is clear and undisputed areas of responsibility, and *one* democratically elected steering body, the Gentoo Council, that listens to the developer community, is familiar with daily affairs, and makes decisions based on that. Electing people means you trust them to represent your interests for a year; an election is the most reliable way that we have so far to find out what the developer community actually wants. From this background I stand for a proactive council that takes responsibility for all aspects of Gentoo. * I support tamiko & mgorny's proposal of a "General Resolution" (GR) to override Council decisions, or initiate a vote of no confidence against the Council. Passing a GR should be hard, but not impossible. If a GR passes, then the Council members know that they have seriously misjudged the overall opinion of the developer community. However, if somebody tries to get a GR to pass and it doesnt work out, that's maybe also a good hint for him/her to stop claiming noisily "I alone speak for the majority". * About Council and Trustees. Ulrich has very clearly spelled out in his mail [3] how the separation of responsibilities in Gentoo was designed. (Some of) the trustees are not happy with that and want to expand their responsibilities unilaterally. Since any infighting hurts Gentoo as a whole, as a consequence I strongly prefer that *one* elected steering body becomes responsible for all aspects of Gentoo. Given below arguments, this steering body would be the Gentoo Council, whose role is constituted by GLEP 39 and which is elected by the Gentoo developers. * About the Gentoo Foundation. What we have definitely learned from the history of the Gentoo Foundation is that Gentoo developers are not good in running a business. Even if we have active trustees and officers now, there is no guarantee that this will be the case in the future, or that they will see their main focus in fulfilling their main duties as described on the foundation's own web page [4]. Which means that I fully support the idea to eventually migrate Gentoo's business side to an appropriate umbrella company such as SPI. A full transition can only be done after the finances are cleaned up, but we can start in the meantime by having further financial service providers for the Gentoo developer community, additional to the Gentoo Foundation. * About lack of manpower. Well, first of all, has anyone already noticed that over the last weeks we've been able to sign up new developers at record rate?! We have to be doing something right, though I'm not really sure what. What in my opinion most kept people from joining is the horrible public image of a few trolls dominating the public fora and of developer infighting [5]. The first has luckily stopped at the moment, the second we should still work on. What's keeping projects from functioning is that we seem to be excelling in one-man-projects (wiki, tinderbox, ...) but that there is not much cooperation between developers in many cases. So, fostering and enabling cooperation is very important! * We need to build on the strengths of Gentoo and make it cool again. Infinite adaptibility, combination of cutting edge testing and solid stable, wide arch support, ideal for software developers. This means generic public relations work (see e.g. FOSDEM), but also initiatives of developers to "do something cool", supporting and publicising that. * We need to at least try to go with the times. Yes I see the ideological and practical disadvantages of Github, but if we want to attract a large base of contributors, we need to seriously think about having modern, comparable ways to contribute! And not get stuck in eternal yesterday... just because things were done successfully in one way when you joined Gentoo 10 years ago, that doesnt mean they have to be done that way for all eternity; this attitude does not help! [*] * About the gentoo-devel mailing list. I supported switching to whitelist mode, since the activity of a handful of trolls was making Gentoo list traffic unbearable. This was a stopgap measure, since noone was presenting a *feasible* better idea. Also, it turned out a bit silly since we restricted the one list where we never had problems. Nevertheless, since then things have been quiet also elsewhere. I'm convinced that doing something to try to solve a problem is better than doing nothing and just hoping that the problem goes away. That means also sometimes trying out non-optimal solutions. When these don't work out, or when a potentially better solution becomes available, things can always be changed again, see below. * I've been working over the last months to (re-)found the Gentoo Proctors, a project dedicated to the Code of Conduct on Gentoo's public communication channels. This was an initiative of ComRel in cooperation with the Council, essentially splitting off some responsibilities from ComRel that never really should have been ComRel's responsibility from the start. [#] We're finally at the point where this can start, so expect an announcement very soon. Depending on the Proctors this may make the whitelisting mode of gentoo-devel quickly obsolete. That's it for now. Cheers! [1] https://dev.gentoo.org/~dilfridge/decisions.pdf [2] https://github.com/gentoo/gentoo-ecosystem [3] https://archives.gentoo.org/gentoo-project/message/ d88b40707e39f1989e80ee91ca8dde80 [4] https://www.gentoo.org/inside-gentoo/foundation/ [5] https://it.slashdot.org/comments.pl?sid=12290940&cid=56862526 [*] Some weeks ago, when I was talking to someone who does a lot of good work for Gentoo and whose opinion I value, I was horrified when he mentioned that "there are three ways to do multilib in Gentoo". * There is the way via pseudo-useflags, which is what nearly the entire Gentoo repository is using. That makes 1. * There is the historical code in toolchain; while it would be nice to translate that into a more modern variant, it is very hard to do (see also e.g. crossdev). Rather keep things working than break them unnecessarily, but this is a situation unique to toolchain. That makes 1.5. * There is multilib-portage. Sorry, but this train has not just left the station, it crossed the entire Rocky Mountains in the meantime. It may still be a private fun project, but on the Gentoo level it is dead, dead, dead. Insisting that it is still important just keeps us from doing more useful stuff. And this is kinda symptomatic; because each old, outdated solution is still someone's pet project, we tend to never finish any [#] Happy to elaborate on that a bit more, but this doesn't really belong in here. Please wait for the Proctors announcement mail.