Bonjour, Le 20/03/2019 à 09:49, JP Guillonneau a écrit : > suggestions et préférences Merci beaucoup Jean-Paul, j'ai tout intégré. Amicalement, Thomas
#use wml::debian::template title="Platform for Martin Michlmayr" BARETITLE="true" NOHEADER="true" #include "$(ENGLISHDIR)/vote/style.inc" <h2>Introduction</h2> <p> I will start with a confession. For a few years now, I've been thinking that it's time to retire from the Debian project. This is partly because my interests have changed over the years, but partly because... Debian just doesn't seem all that exciting anymore. The only reason I haven't actually resigned is because Debian and the Debian community plays such an important role in my life and I don't want to let go. </p> <p> I believe my mixed feelings about Debian are not that unusual. We've seen a number of prominent contributors leave the project in recent times out of frustration or sheer boredom and I think there are many who stick around but who have essentially checked out mentally. </p> <p> Seeing that there was no candidate for the Debian Project Leader (DPL) position, which I see as a symptom of grave underlying problems, got me thinking. I — like everyone who chooses to contribute to Debian — have two options: I can resign, run away in frustration, or I can help to discuss, confront, tackle and solve these problems. </p> <h2>Debian today</h2> <p> Debian appears to be in a very peculiar situation right now. About 10-15 years ago, Debian faced an existential crisis. A number of factors contributed to Debian becoming less relevant. This includes the introduction of Ubuntu (which offered a more polished desktop at that time and long term support for servers), the move towards Apple's macOS (among the general public and even among open source developers, especially in the web space), and a general shift towards mobile devices. </p> <p> Today, things are vastly different however. Debian is more popular than ever on servers. There is a renewed understanding of the importance of Debian and a new wave of adoption in the cloud and container space where a stable operating system is key. Debian is everywhere. It runs large parts of our society's IT infrastructure and also constitutes the building blocks for many other solutions. Users and corporations appreciate not just the quality of the Debian system, but also the maturity of our community and the open nature of our development process which is not dominated by a single vendor. </p> <p> So, despite this success and importance, why does it sometimes feel that Debian has become somewhat irrelevant? Why is there no buzz around Debian that reflects Debian's success in so many areas? </p> <h2>Problems and solutions</h2> <p> I recently read Michael Stapelberg's post on <a href = "https://michael.stapelberg.ch/posts/2019-03-10-debian-winding-down/";>winding down his Debian involvement</a> in which he describes how stagnant Debian's infrastructure and processes have become. His well-reasoned post resonates with me, although I believe the problems are even bigger than that. </p> <p> The open source world has fundamentally changed in the last 5-10 years in many ways. Yet, if you look at Debian, we mostly operate the same way we did 20 years ago. Debian used to be a pioneer, a true leader. Package managers, automatic upgrades, and packages builds on 10+ architectures — they were all novel, true innovations at the time. The only significant innovation I can think of that came out of Debian in recent years are reproducible builds. Reproducible builds solve an important problem and the idea has spread beyond Debian to the whole FOSS world. </p> <p> I hate when large companies talk about being "nimble" or similar business buzz words. But looking at Debian, I finally understand what they mean — the project has evolved in a way that makes change difficult. We have failed to adopt to the new environment we find ourselves in and we're struggling to keep up with an ever-faster changing world. </p> <h3>Sustainability</h3> <p> When I got involved in free software in the 1990s, it was out of technical curiosity and a passion for software freedom. It was a hobby and I never imagined making a career out of it. But then more and more companies discovered the technical superiority of software developed with collaborative FOSS models coupled with the practical freedoms offered by FOSS and the industry took off. Many of us are nowadays fortunate to contribute to FOSS not just in our spare time but as part of our jobs. </p> <p> Today, the majority of successful FOSS projects rely on (and thrive from!) paid contributors. This is not a failure, but a reflection of the success of FOSS. Unfortunately, compared to other successful projects, the proportion of contributors who are paid is much smaller in Debian and I see that as a failure. </p> <p> Over the years, I have seen many valuable Debian contributors leave after they graduate from university and get a job (often working on FOSS, just not Debian; or working <em>with</em> Debian, but not <em>on</em> Debian). I have seen many dedicated contributors who essentially have two jobs — a paid job during the day, and an unpaid gig at Debian working long evenings and weekends. I have met many contributors who make huge sacrifices in order to contribute to Debian. Unfortunately, I have seen a lot of burnout, too (and experienced it myself)! </p> <p> If you want to work on Debian as a hobby, great! But if you want to make Debian your career, you should be able to do so. While there are some paid opportunities around Debian, I believe they are currently too scarce and that we can take a number of actions so more contributors can make Debian their careers. </p> <p> First, we can help more companies get involved in Debian. This does not mean that we will sacrifice our neutrality or focus on quality. It means that we'll bring more contributors on board who will work towards improving Debian for everyone. We already have a number of companies (Arm, credativ, Google, just to name a few) who make strategic investments in Debian and pay engineers to work on Debian. </p> <p> We can encourage more companies to get involved. We can make it easier for them to contribute, we can show companies that rely on Debian the importance of actively getting engaged (and help them develop the business case for it), and we can provide more incentives, such as acknowledging their contributions better. </p> <p> Second, grants are becoming a popular way to fund R&D projects and FOSS development. The Reproducible Builds project is funded through grants from the Linux Foundation and other organisations. I doubt they would have achieved as much as they did in relatively little time without this funding. We have a lot of smart people in Debian with good ideas — the DPL can identify grants and help them apply for funding. </p> <p> Third, we need to work with the wider FOSS community to find more sustainable ways to fund FOSS development. The OpenSSL debacle has shown that society depends on FOSS without making the necessary investments. Many independent FOSS developers struggle to make a living from their work. Platforms like Patreon exist but few can live from this since there's not enough awareness that people need to invest in and sponsor FOSS. </p> <p> I'm sure there are other ways. Overall, I believe it will become increasingly difficult for significant projects with unpaid contributors to compete with efforts where everyone is paid full-time and I think that partly explains why Debian has struggled to keep up. More fundamentally, we also have to ask ourselves what kind of project we want to be. The world around us has changed and it's time to acknowledge that. </p> <p> The Long Term Support effort organised by Raphaël Hertzog of Freexian has been spectacularly successfully. Reproducible Builds has attracted significant funding to improve an important aspect of Debian. Is it time to bring these efforts in and do them as Debian? Or should we encourage more such efforts around Debian and give them a helping hand? Is it time to use our funds for certain things? </p> <p> These are big questions we'll have to answer as a project eventually. However, we can start by inviting more companies to participate in our community and by encouraging those who depend on Debian to contribute. </p> <h3>Leadership and culture</h3> <p> I see a lack of leadership in Debian. I don't only mean DPL-style leadership, but lack of leadership in general. The beauty of FOSS is that anyone can make a contribution. You don't need to get permission to be a leader. If you have a great idea, just do it! </p> <p> Unfortunately, while we have many talented people in Debian, I think we've reached a point where people are afraid to make or propose changes, especially big, far-reaching changes. They are afraid of the resulting flame-war or other fallout (or simply tired of the cat herding involved in getting the change adopted everywhere). </p> <p> Why is it that our community makes people so afraid to speak? Why have we stalled? It seems to me that Debian has developed a number of toxic anti-patterns over the years that we have to move away from. </p> <p> What I'm going to say now will be very controversial. Debian prides itself on being the universal operating system. This is a great goal, on many levels. However, it's also important to acknowledge that you cannot do everything — that doing everything slows things down. We can't always wait for everyone, otherwise we'd never get things done. Not everything is equally as important. Sometimes it's important to set priorities and to tell people "no". Again, we're afraid to do that. The universal operating system, and the underlying culture, while laudable in practice, has become toxic in practice on some levels. </p> <p> Same with technical excellence, another laudable goal, that becomes toxic when taken to the extreme. We talk and talk... and talk. We should remember that not every battle is worth fighting for and that not every argument needs to be discussed till the bitter end. It's fine to agree to disagree sometimes. We have to remind ourselves to pause, reflect and ask: is it really worth it? Do I really have to send this email? What about the social cost? Let's stop draining energy from our community! We're not a debate club. We're here to solve problems in a practical manner and to ship solutions to users. </p> <p> In addition to more leaders, we need more cheerleaders. We have to thank people. We have to appreciate contributions. We have to create a buzz around Debian. We should celebrate, and build upon, our success. (And yes, I appreciate the irony of saying this in a platform which essentially consists of a critique of Debian. However, as DPL and as a community member in general, I aim to say "thank you" more often and to encourage other people to do the same.) </p> <h3>Influence</h3> <p> Debian plays a very special and important role in the FOSS ecosystem. We are respected and our contributions are appreciated. Debian contributors tend to be leaders in the FOSS space. We pride ourselves not only on packaging software from upstream but on maintaining good relationships. This often results in us getting involved upstream and taking on leadership roles there. You can also look at current and past board members of the Open Source Initiative (OSI) and again you'll see many Debian people. </p> <p> While Debian people play important roles everywhere, they often don't represent the Debian project. We need to learn to develop and speak as a single voice. Overall, I believe we, as a project, need to be more vocal and take a more active role in influencing the FOSS ecosystem. Debian has an incredible reputation but we don't use our clout for important change. </p> <p> While we're fairly good at influencing technical aspects, we have little sway in the business world of FOSS. And let's be honest — open source has become a big business. We're missing out of important opportunities because this is a game we don't play well. We need to take a seat at the table (and, yes, that's often an actual table or a conference call, and not email). </p> <h2>Background</h2> <p><img src="tbm.jpg" alt="Photo of Martin Michlmayr" align="right"></p> <p> Who am I and how do I see the role of the DPL? I have been a Debian Developer since 2000 and I worked on a number of areas over the years. I worked as an Application Manager, helping new contributors join the project, and also managed the New Member Front Desk. I contributed to various Quality Assurance efforts, filed bug reports and introduced the Missing in Action (MIA) process. I ported Debian to various ARM and MIPS devices and wrote documentation for it. I served as Debian Project Leader from 2003-2005. Most of my contributions to Debian recently have been indirectly through Software in the Public Interest, one of Debian's trusted organisations. </p> <p> I worked for Hewlett Packard for nine years, facilitating and leading various internal and external FOSS activities. I served on the board of the Open Source Initiative for six years and on the board of Software in the Public Interest for close to five years. I currently serve on the board of Software Freedom Conservancy, the home of the Debian Copyright Aggregation Project. Over the years I've become an open source accounting geek and I contribute to ledger, beancount, and ledger2beancount. </p> <p> There are various opinions on what kind of time commitment being DPL takes, ranging from proposals to a "leaderless" Debian to the DPL being a full-time job. Based on my experience as DPL in the past, I'm firmly in the latter camp. When I was DPL in the past, I was a student and had (what looking back seems like) unlimited time. I'm now a FOSS consultant and every hour I'd spend as DPL takes away from that work. I am, however, willing to devote significant time to Debian. (I might also set up a virtual tin cup.) </p> <p> The role of DPL is very unique. In Debian you don't typically get "promoted" in the traditional sense. You start doing the work and eventually people recognise you for it. It's all based on reputation and ideally we want a DPL who has already earned the reputation for being a leader. I gave a talk about FOSS culture at an event some time ago and I emphasised the importance of building up a good reputation. To illustrate the point, I said there's this one particular person in Debian and when this person sends an email, people will listen because they know that the email will contain genuine wisdom written in a calm tone. Despite Debian having over a thousand <em>incredible</em> contributors, this description was enough for some Debian Developers in the audience to nod their heads with a shared understanding of whom I was talking about (without ever mentioning the person's name). This is the kind of reputation you want to build up! (And there are many other contributors I could describe for their very unique ways of contributing to Debian!) </p> <p> Unfortunately, I haven't been active in Debian lately and so I don't have the reputation that I would hope to see from a DPL. While many old-timers are familiar with my work, I'm a blank canvas for many new contributors. I am aware of this, and if elected, I will work hard to earn their trust. </p> <h2>The DPL</h2> <p> The role of DPL is often described as having an internal and an external function and both are absolutely vital. I believe you can also define three roles of the DPL: the DPL as administrator, facilitator, and leader. </p> <p> While I enjoy getting things done, I also find it immensely rewarding to help other people get work done. In this way, I see the DPL as a facilitator. I just attended FOSSASIA where someone compared managing communities to conducting an orchestra. When you look at orchestra conductors, you may wonder what they are doing apart from waving their hands around. You don't see all the coordination and practice that was necessary for a flawless performance. DPLs have in the past been criticised for not being active in part because you don't see a lot of the day to day work. You only see if something doesn't get done or goes wrong! </p> <p> Facilitation includes a lot of coordination. It involves helping people do their work effectively, removing problems, connecting people, using delegation and following up. It involves listening to people. It also involves asking people to do certain tasks. You'd be surprised how effective it can be to ask someone to help out. If you don't ask, it doesn't get done. Facilitation doesn't have to be <em>reactive</em> as it's often been. I'd like to be pro-active in terms of resolving problems and identifying areas that need to be tackled. </p> <p> The external leadership role is also important. There are discussions about replacing the DPL with a board or committee, and while it's important to have these conversations and to evolve our governance, I think we should not discount various human factors. Companies, the press and others often want to talk to the leader of the project. They don't want to talk to a mailing list, or a committee and sometimes not even to a delegate. As DPL, I intend to work with delegates on forming more connections and partnerships; on building bridges with our allies. I'd also like to start an honest conversation about the state of Debian. How can we improve our culture? What are the things holding us back? </p> <p> Finally, it's important to take a step back and look at the big picture. This platform is highly critical of Debian in many ways. I should say that I'm a critical person by nature and I tend to emphasise the negative side. Fortunately, I've been able to use that personality flaw in positive ways, such as through my QA work where I identify bugs and that feedback is used to improve the software. But it's important to highlight the good aspects. Debian is so unique, so wonderful and so special in many ways. If I didn't believe in Debian, I wouldn't run in this DPL election. I acted as DPL before and know how difficult being the DPL can be sometimes. Yes, I see severe problems in Debian, but I firmly believe that together we can solve them! </p>
Attachment:
signature.asc
Description: OpenPGP digital signature