 The future of fedora is fedora It's the title of this talk, but what does it even mean? So Fedora turns 19 this year and that's not normally a significant number But in the US it typically occurs during the first year of college And that's right about the time when you realize that college is Well, it's a lot like what you've been doing before and it's totally different You're still going to class Still doing homework Still making new friends, and you're still hanging out with your existing friends, but you're doing it all in a new context You see in the US college is usually a residential experience And so it's all totally different while still being the same And that's what this means. It means that we're gonna keep doing what we've always been doing But we're gonna do it all very differently So with that said Let me start by saying good time of day to all of you. My name is Brian Exelbeard I've been in and around the fordora project for a while You may remember me from such things as Didn't we get rid of him and get the much more awesome Marie instead or no, that's not been that's that other guy And who can forget amazing talks that I have given on topics such as what does red hat want? So today I want to walk through a couple of themes with you and My goal is to expose some possibilities and I hope to inspire you to co-create the future of the Dora To give you a few milestones I want to start by talking a little more about this title and what it means and Then I want to talk about the plane that for Dora exists in and I want to talk a bit about our bits See what I did there Next let's talk about Contribution and then I'll bring it home by talking about people. I Want to give you fair warning. I don't expect to talk for the whole hour Right after this talk We're going for a social break in the fedora museum and I'd love to use the balance of any time that I have To be a catalyst for conversation there And lastly, there are no slides in this talk So feel free to lean back and not watch the video. I will not shock you with a diagram you need to see My friend Adam Shyamalik helped me prepare for this talk and he suggests that you all stare at a lava lamp So do with that information as you wish The future of fedora is fedora So how can I be so sure about this I'm sure about this because we've done this before in 2013 we shifted from the board model to the council model The door Lennox and stop but we transformed how we got there We created space for new ideas and new outputs Critically this change Changed the way we think about strategic thinking It moved from near the bottom of the fedora boards responsibilities list to the top of the fedora council's place If you listened to Matthew Miller yesterday, you know that there's an effort going on right now Where the council is trying to figure out the next long-term vision for strategy for this project? It will be all about doing what we've always done and doing it differently and It will not succeed unless you yes, you Participate the governance changes that were made The governance changes that were made Also continued the work that had been done to empower groups in fedora to make decisions in their scope without needing additional permissions or layers of bureaucracy and With this change we essentially kept on doing what we'd always done. We kept doing it differently The second reason that I am so confident about this is that I've seen it from a lot of sides I had the privilege of being the second fedora community action and impact coordinator the FK I Worked with many of you to transform the project into where it was when Marie became FK Whether those transformations were successful or not they were all aimed at easing processes and encouraging contributions They were focused on aligning the defaults and easiest paths of our project to align with the goals of our project But not make other things impossible This can be seen in everything that we did together whether it was the move to transparent budgeting or the attempted reboot of the ambassador program So I've seen this project change and grow and now I now work directly for the Red Hat Enterprise Linux business unit I am literally on the best team at Red Hat the business and ecosystem strategy team best as an aside, I have to tell you that Gunnar Hellexen the vice president and general manager of the rail business unit was two days ago days old When he realized that our team name had been selected so that he had to constantly refer to us as his best team Now you tell me that marketing and naming don't matter Anyway as a strategist my role is focused on the community and connecting the community to the rail business Unit I'll go a bit more into depth in that but suffice it to say I am watching you And I like what I'm seeing because I've observed that the product project continues to change and still be the same and This this change this continued change, but say the same is super healthy This is the kind of transformation that keeps the project true to its goals But redefines the finish line so it can keep moving forward The strength of this transformation is actually part of the reason why I'm not giving a talk this year That's called what does Red Hat want? The other reason that I'm not giving that talk is it's not very interesting For one thing it's really hard to give a satisfying answer to that broad of a question And if you think that it's difficult for the Fedora project to come up with a three-year vision With a council driving that change you cannot imagine how hard it is for Red Hat to define goals for a community This is triply hard if you will Because we know that the best communities have the freedom to grow organically and go after their own goals Not have goals externally imposed upon them Matthew was very direct yesterday when he said Red Hat's best work happens when Red Hat doesn't do all the thinking A third reason that I am so confident that the future of Fedora is Fedora There's chat hold on Okay, so the third reason that I am so happy They're so confident that the future of Fedora is Fedora is because of the creation of Centos stream When Red Hat built that code base and put it in the Centos project What Red Hat really did was we doubled down on the Fedora model When the Centos project adopted it the way that they have it showed that even other communities understand what's going on because if you think about it if deriving a product From this process wasn't working and it wasn't adding huge value We would not have set up another version of this process again to add even more value to the end result and Other people are watching us Go look at what our friends and Susa have been doing. It looks a lot like they paid attention to what Fedora has been Successful, but let's talk about this plane that Fedora occupies And by playing here, I mean the mathematical construct. I don't mean an airplane We are all in this together, but let's not get too literal Now I don't want to talk about this from the perspective of what Red Hat wants Frankly, who cares? I mean that's Red Hat's problem It is also my day job. So if you really want to talk about it, we can find another venue to have that conversation But I would like to consider this plane and one easy way to consider it is from the perspective of an ecosystem And there's an awesome graphic that Melissa Lautenheiser a friend of mine who's a product marketing manager in Rel Made I used it in a talk that I gave last year at Nest I chose not to use slides this year. So you can't see this graphic But the graphic consists of three sets of circles that describe the way that code flows through our ecosystem Fedora Linux feeds into our stream, which feeds well And the code literally flows back and forth. There is this, you know, infinity symbol s thing throughout this graphic And I love this graphic Because it helps people see how the pieces fit together and it really brings a visual image to the idea of upstream first But I also don't like this graphic Because this graphic leads a lot out First it focuses a lot on code and code is interesting But it's ignoring people and people are way more interesting and we're going to talk about people later But the graphic also ignores the purpose or reason for each of these code trees You see Fedora isn't a place where code simply incubates before Centaur stream picks it up It's a dynamic place where code is brought together to solve specific problems This is the key to what Matthew was talking about yesterday And this is also why What does red hat want is a terrible talk Sure, some of it is my public speaking skills But a lot of it is that because what red hat really wants is for Fedora to know what it wants We know those best open source projects are sponsored and enabled They're kept independent And because red hat product ties, you know, they're kept independently as possible I should say But because red hat product ties is open source code The closer that code gets to product the less independence that may be possible And you can see that Look at how committers are selected for Fedora Linux and versus Centaur stream And before anybody complains or goes and writes a register article about me Look at the rest of the Centaur's project first and see how that's compensated for But that independence is also sometimes challenged by legal issues Candidly people don't sue Fedora. They would sue red hat These issues really don't happen. They're few and far between and so this really isn't a thing But it's something we have to remember as an edge case But in all of this red hat firmly believes that we should make sure that Fedora and frankly any project we sponsor or participate in has freedom Freedom it sounds a lot like an F. I've heard mentioned around here before And that freedom It starts when you allow the project to find its own goals and to find its own futures Goals imposed from the outside as I said, they're not going to work So if we can't talk about goals imposed from the outside, what can we talk about with those goals? Well We can talk about ideas for them And since I'm kind of here on behalf of red hat, I'll talk a little bit about red hat's opinion You see red hat does have an opinion about what Fedora should do, but it's not really a want or say We kind of have the building blocks of a vision for what could be in a universe and Fedora is part of that And and I know about this vision and these building blocks Because gunner helixon and ron pacheco told me that I had to figure this out or that that was my job Like go do this now And so this is where I am with that and your feedback is welcome and I hope it will help the project as well Um Fedora simtas and rel as I said they exist on a plane You probably thought I forgot about the plane, but I did not forget about the plane The plane has two dimensions x and y and so it can accommodate a big tent Yes, tensor three dimensional. I get it, but it has a big tent effect and everybody is on the plane The linux kernels on the plane the derivatives are on the plane seuss is on the plane even debbie and is there Everyone is there the plane is the place to be But at red hat when we look at the plane we tend to focus on rel And the antecedents of rel the things that came before rel. So these are the outputs of fedora and zindals And when we start to consider those things we have to make a choice We can think about them from our perspective as red hat or from the perspective of a third party Well, we already know what our perspective looks like. This is the code diagram that I mentioned earlier It's that direct path from upstream to product It is also something that third parties have made very clear to us is generally of mild to no interest at best So I had to really think about this from the perspective of these third parties And this has led me to create something that I am calling the value segmentation model That is a super fancy name I am super proud of it And I am told that if I can get promoted to the consulting grade of my job tier That I get to call it the vsm and everybody has to do that Unfortunately, I'm not at the consulting grade. So we're going to stick with the law this value segmentation model It describes how third parties view interaction with an operating system and how they view interaction with the producers or vendors behind those operating systems And we're going to use this model internally in red hat to Define the next generation of operating system offerings In this sentence offering is basically a fancy word for product. There's some nuance there. So it's technically an offering And because I have to write this beast I decided to solve some other problems and anybody who's ever worked with me knows that I can't just work on one problem I have to reorder the whole universe But I did limit myself here and two of the problems that I am going to solve with this model are Well, one a Matthew Miller problem See Matthew occasionally shows up in my meetings and he's like, hey um It's me Have you forgotten about me? Have you forgotten about us? You don't write you don't call And and the truth of the matter is that we haven't forgotten about fedora But when you combine the trifecta of the freedom fedora has The resourcing that fedora gets like the community funding and other things that are not subject to like a renewal process And the fact that fedora works. I mean it really just works Well, you're not on the radar And that's terrible. That's really bad. You should be on the radar. We should be focusing more on you for non-code flow reasons So I want to fix that ideally with this model The second problem that I want to tackle with this model Is that our merry band of siblings fedora centos and rel we currently go out into the world and we're out there alone most of the time We don't usually work together at all um You know, sometimes we show up at the same conferences embarrassingly when I was fk We would show up to the same conferences and not know we were showing up to the same conferences um We don't really work together outside of you know this shared code um The progress that fedora and centos are making where they're starting to share infrastructure This is really exciting to me and I love seeing that progress. I want to see even more of that progress But let's be honest shared infrastructure is basically just shared code at a different level And I think the core of this problem is that every element of our family Has to make a decision about how to solve everyone's problems We don't tell an end-to-end story to third parties I don't know if you can hear the helicopter, but it's awesome Um, we don't tell an end-to-end story to third parties that lets them make their best choices We should do that And this lack of a story It creates a situation where it can look like the projects are in competition with each other Um, or the product is in competition with one of the projects or both Um, it creates adversarial feelings and this is bad This is also very very bad So my first attempt at fixing this um is to make it part of this value segmentation model And in fact, I put it in the first segment of my value segmentation model A segment that is called wait for it. I know you like creative names The first segment It's called community Let me read you a little bit from my current working draft and keep in mind This is part of a multi page business language ease document that will become doctrine of the realm be you This is the segment. I think you might like All right, um our community projects fedora and centos must be considered as part of our overall structure Doing this ties the entire ecosystem together and avoids us having a disconnect Where the product is in competition with the projects or seems to exist on a different plane of reality However, the realm be you Rightfully views the projects through the lens of their service to product And this means that we want them to show up in adjacent use cases to product As a part of this the realm be you may encourage or support the presence of community outputs in some places And be ambivalent or even opposed to their presence in other places And I think the critical part is this one Well fedora linux and centos stream are very different They represent a user workload requirement set that's related I think these users are likely to need a desktop not a workstation key difference Have tolerance for fast-moving software that experiences continuous or frequent updates Which may or may not bring new interfaces or features or avi gods These users may want to only consume open source software at the project level Or these users may wish to engage in custom building of an operating system at a low level In an early stage And I think that talks about the space Where the fedora project and the centos project and red hat should meet to figure out the beginnings of that story But I really want to drill in on one of these points It's the third one that I had the wish to only consume open source software at the project level Now this is a fancy set of words that I wrote which basically means somebody wants to show up and say Give me the bits and have a nice day I may come back if I have questions. I know you may not have answers And that's fine. Like that is a great model for people who wish to consume open source They take what they need they use it. They hopefully show back up and give feedback ask questions and idea like contribute But how did we get to the point where I felt like this was a value segment? Well, we got here because The operating system is boring Wait, that's not right Ben cotton is screaming at his monitor right now because he and I went over this talk at one point He was like, it's not that the operating system is boring. The operating system is reliable And he's right Fedora Linux is reliable Fedora like most open source projects is building a lot of value into the bits that are shipped This value is manifest in things like choice and reliability This is the end result of where we started open source projects Are creating an amazing amount of value in the communities they build and in the bits they ship This combination the chat is distracting. I'm sorry. This combination creates the reliability of open source We need to keep doing the things we're doing to ensure that that reliability continues But we need to do things differently in order to innovate So let's dig into the value first and I promise I'll come back to the non bits bit in a few minutes As I said earlier, um open source projects are creating amazing value in their bits And the expression of this value of Fedora Linux is mostly in its bits But the bits of today are very different from the bits of yesterday In the old days before my beard was this gray or my daughter was as cute as she is Um distribution projects like Fedora focused on taking bits from the wild Trying to prove that they were halfway okay And put them into a standard size unit And in fact, this is why we have such a historical reverence for the act of packaging in the project But nowadays Bits are not being chased down in plains and valleys by people on horseback like lost cattle Instead the inputs to the Fedora project. They're generally from equally successful And critically reliable open source project Um Matthew was talking on a mailing list recently about system d and I believe ganome came up in the conversation as well And he pointed out that part of the challenge with those projects is They're providing rock solid code and rock solid setups Essentially those projects have taken over some of the work that a distribution used to have to do to make their software usable And I submit that this is a good thing because this is an example of another group doing what they've always done But doing it differently and moving their finish line forward And and this value in this bits these massive massive value in this bits is important because it's transforming all of the players on our plane Um, I like to think about it in terms of our industry And focusing specifically on the corporate side of this open source product companies are increasingly seeing their business models come under attack And that's because they've typically built products that were heavily based on the idea of delivering reliable bits And and what we've seen is that open source projects can do and do do this regularly In fact, we're starting to see projects now nibble away at some of the concepts around certification Today that's primarily because a third party vendor has stepped in and sponsored some kind of an effort But I think that's going to change over time And I recognize this is heady stuff But it continues to demonstrate Why the choices that we make in fedora about open source and how we're going to practice open source are right And why things like open core fail and why we're not worried about that in our ecosystem And this reliability of fedora linux and the others is reshaping the business landscape There have been a ton of responses to this. There's been almost whiplash level of responses to this But the one that seems to have the most staying power is called services At least services is what we're calling it today Give us some time. Um, like everything else in in this field will probably rename it to something different tomorrow But services services is a confusing word and it may be why it's time to pick a new one It's an old word with old concepts and it's now trying to do new things Kind of like how the future of fedora is fedora If you'll forgive me here, I have an analogy that I want to make it makes me super happy But it needs some setup and I hope it's worth it for you There was a car brand in the united states called oldsmobile Sadly, it was shut down in 2004 In 2004 it had been around for 107 years and at that time was the oldest remaining Surviving I should say a car brand or mark in the united states There were a couple of european brands that were older orisha and somebody else I believe Oldsmobile ran campaign in the 90s, which that's the 90s, right? So they're going to shut down soon They don't know that but clearly they're downward trending And they're trying to rebrand their cars and their campaign was this is not your father's oldsmobile Well, these are not your father's services When you think about services, you can see where they came from Our industry started with an emphasis on support as one of the key values provided by open source product companies Support is a service The other leg of the stool reliability For reliable bits. There's more there, but you can see where I'm going But as I said earlier reliable bits are now the domain of projects. They're table stakes Support value continues to be very real But there are in fact a lot of companies that now provide support for products. They didn't build What worries me a little more is they're providing support for products in which they don't even participate in the open source chain of the upstream That's concerning, but you know everybody does what they need to do The general level of knowledge about linux in the technology pool of talent is is also getting better every year Higher and higher and higher We certainly need more people. We are having hiring challenges all over the place But the pressure on open source product companies is here And so this is where this emphasis on services has come from for the last few years So I want to talk about these things to help us be grounded in the context One of the first services we saw come out was really an addressment of running the software for people Software as a service This is a pretty good idea actually in hindsight. It is blindingly obvious But uh, we have actually seen even in this short amount of time that companies are struggling here And it's because they're struggling to compete against third parties who can run the software as a service better That means that that company is running software better as a service Then the other company that actually sponsored and participated in the development of that software And you can see this if you go to your favorite cloud provider and you read their product list Look at how many of those items are open source products Projects that became products and are now being sold as a service The term of art here sass software as a service And this shouldn't surprise you if you work in the industry Um, there's a movement called dev ops and it's called dev ops for a reason because devs aren't always good at ops And ops aren't always good at devs Being devs and if that wasn't true Then one those companies would have been better at running their own services and two we wouldn't have smashed those two words together And if you're an sre be quiet, I can hear you Um, so services services as I said are for most of us not going to be this sass there there has to be more Now a lot of really smart people are working on this Um, and what we're realizing and by the way those people are like wicked smart like way smarter than them Um, but I have the privilege of sitting near them virtually um What we're realizing in this work is that the value of an open source product company needs to layer on top of the bits It has to be above it can't be expressed in the bits And so what that means is in driving this back to fedora is The companies need to be doing things that a project like fedora can't do well Things like lifecycle Like clockwork fedora releases a new release every six months fedora is reliable After 13 months we el that release because we don't have the contributor bandwidth to go further We're also not interested in going for further If you need more time, you may find that there is a product company that can sell you an extension on that code base Not for fedora, but for other things And this is the tip of an iceberg of things where companies can help out because projects aren't able to do that Fedora comes with a lot of metadata like an amazing amount of metadata But a vendor may be able to provide more specialized metadata that is harder for a community to generate or harder for a community to guarantee Opinionation can become a service Opinionation is interesting in fedora because when I look at opinionation at fedora, I see two stories Fedora Linux is has made a lot of hard choices a lot of opinionation has gone into this about what the software will do or won't do What's going to be on or off by default And we've also had to make choices that result in some things being easy to do and some things being hard to do But if you look at other areas of opinionation we're struggling If you go look at the fedora developer portal You'll find a note that if you're developing a python you may find yourself needing software from pi pi um, and it isn't necessarily going to be packaged And the developer portal goes on to help you understand how to deal with a virtual environment run pip and all that fancy stuff to get your software Nowhere does it tell you how to figure out if that software was actually already packaged in fedora And I had some direct conversations with those folks While I was fk and I remember sitting down with them over a beverage of choice and they finally just looked and they said that's Here's the deal Everybody needs to get their software from from pi pi We simply cannot package it all and keep up fast at all and a vendor May be able to step in and figure out opinionation on how that software can be consumed better from a variety of sources Because the project at least isn't there yet And I point these things out because if we go back to this reliability concept When projects started reliability wasn't a hundred percent something they could do now it is I think that over time we're going to find that Projects will continue to move their finish lines forward continue doing things differently While staying the same and these services will slowly fall down the stack And that's a good thing the pressure on companies is a good thing So I have to tell you this is my first keynote. Hopefully not my last keynote I got a couple piece of advice One of those pieces of advice was that you should be controversial So let's talk about contribution Open source has never been more prevalent or more successful That was not controversial controversial bit is this 100 percent of open source contribution is because of companies Okay I'll go with almost 100 percent because absolutes are hard. So almost 100 percent of open source contribution is because of companies Companies pay the salary of almost every contributor Lots of contributors are doing the work a company needs They're not showing up in communities to scratch their own edge At least not directly They aren't doing so indirectly Their company has an edge and they are a professional scratcher They are directed in the kinds of contributions to make Now I see that luna does not agree with me And that's cool Because to luna I say I can hear you screaming at your monitor You're saying that's easy for you to say you work for redhead. I don't work for a tech company I'm an insurance agent. No, I don't actually know if luna is an insurance agent But I hear you luna And to you I'm going to say this Not to you personally, but luna's will say to everyone who's an insurance agent or not You're right That you're not directed in what contribution effort you're making But in many cases you may be ignoring privilege You see some company is providing People with the stability and the income that allows them to make contributions in their spare time Instead of for example working a second job And if some of you are students the same thing applies Many of us in our student years will later discover that those were the richest times of our life Our needs were often low so the money didn't matter as much And our ability to do with what we do what we wanted to do was high As we get older we may earn more money But trust me it comes with three-year-olds who interrupt your talks to have their diapers changed I hope I did not just jinx this Um, but back to the controversy Almost a hundred percent of open source contributions are because of profits And this is bad this starts the exclusion of lots and lots of people So one of my challenges to fedora is that we need to figure out how we're going to change this So let's talk about jobs for a second. I love it when I hear about contributors getting jobs I also don't like it when I hear about contributors getting jobs It isn't that I don't want people to get jobs But someone getting a job is often the start of one of these cycles of doing everything the same but differently And sometimes when those folks get jobs they wind up working at say a company that works in open source And suddenly they're in a bunch of internal meetings And those meetings can be horrible for open source They can destroy the ability to participate for those who contribute in their spare time You see this occasionally when a decision is reached But there doesn't seem to have been any conversation and everyone is strangely aligned This is likely because they hashed out the decision by discussing it in an internal meeting somewhere forgetting that this was a community issue My good friends in the open source program office at red hat spend a lot of time at red hat saying Shouldn't this be in pound fedora? Shouldn't this be on discourse? Shouldn't this literally be anywhere else in meetings? Um, I don't know how many of you you saw it, but like this was very personal for me um Recently on twitter There was uh the first pr main going around I got sucked into it and I went to the little site and looked up my first pr on github Wow It is a hot mess So my first public pr was a documentation fix to the docker project, which is now called mobi The fix added a reference to the rel installation instructions It goes back and forth about some content issues and they were valid And there are some concerns raised about providing a link that goes off their documentation site But if you read the pr closely Somewhere in the middle A docker person says hey, can you get so and so to review this? And my response to this is yeah, so and so seen it. They work on my team at red hat And later in that same pr I'm asked what the status of the pr is and I reply with its on hold pending internal meetings You see this one little pr had actually managed to get big enough that docker.com And red hats direct relationship folks were actually discussing it This is bad Um, we see this all the time though. I bet if we go looking for it So this is a challenge that we have to work on and we have to think about our spare time contributors there Because they can feel like they're perpetually being blocked out of the box. They just can't seem to get ahead You know, they they get up one Saturday morning. It's a beautiful day They start working on a pr code changes documentation changes wherever get about halfway through it set it up leave themselves working notes Come back the next Saturday And during that week a conversation opened on a discussion the decision was taken Changes were implemented and that pr is now mostly invalid and worse During that week somebody who's got a full-time job and open source may have really needed that pr finished So they helpfully finished it for them Forever starting and never finishing is not a good time So we need to think about what we can do to fix this But let's move on to another topic packaging We all know why packaging is important standardized units have been proven time and time again To give greater efficiency and scale you can look at the real world and technology world example of containers to see this Packaging creates the opportunity to manage software with tools to do audits to use metadata to make decisions But my thesis is I submit to you that packaging is a wasted activity I challenge the door to figure out how to stop squandering committing Commit our effort on something that can be done by a script It's not going to be a hundred percent But I think we already have the blueprints for how to do this in the packet project We have amazing albeit complex and occasionally potentially fragile build pipelines that can feed testing pipelines That could be cloud instance pipelines. It's amazing what we can do with automation What could the fedora project accomplished if fedora linux required 90 percent less human packaging effort? I talked a little about pi pi earlier and how fedora wasn't able to package at all I know a lot of people have taken a run at this problem I've had the privilege of working with them and listening to the stories of woe and terror of trying to figure out how to solve this But I think that the world continues to evolve and do the same thing but differently and maybe there's some more ideas here We know that pi pi and other ecosystems are starting to try and handle quality control issues on their side How can fedora leverage that work? To see if we can make it easier for people to consume these ecosystems More importantly, what can fedora show those ecosystems about maintaining software quality and maintaining software? ensuring quality We have a lot of knowledge in this space like a lot a lot of knowledge and we should share it with these folks We shouldn't make them go through the things that we've already been through to learn the lessons. We already know And then there's accessibility Now accessibility is not just about ally or a11 y the the mnemonic for accessibility I'll come back to those in a bit But it's also about being where your users are when they need you Where does fedora need to be to help folks? Now we can all sit around and blame red hat for not signing legal agreements or whatever But a lot seems to get done without those Who's the next hardware vendor that's going to ship fedora by default? What upstream projects should we be going to enable to have the best possible experience testing and building fedora first? How do we get examples of solving problems the fedora way out there? Thinking about the distribution itself. Matthew touched on a lot of these topics yesterday But think about how we can transform fedora from being a solid desktop with some alternatives Some self assembly kits floating around near it Into a string of solid pre-built opinionated but modifiable outputs People want solutions not more work even when the more work comes with directions People want to swap out the parts occasionally and fedora is in a great place to enable this. It's open source Now this doesn't mean that I'm saying we should Exclusively cater to specific people's needs or specific groups needs But I do think we need to think about who our audience is It really isn't everyone when you think about it from a broad sense That value segmentation model, which I think is amazing Gives us some hints. It shows us that we can benefit by providing high enablement to certain types of groups While leaving room for everyone else who wants to come along for the ride Let's help people in specific ways while leaving others the tools they require to self solve and to join our community as they wish Innovation is is another key component of what fedora can do and and I think fedora in many ways Is the most uniquely positioned of the operating systems that I've been talking about to do it I should say fedora lex is the most uniquely positioned for fedora project in and of it And we already know this I'm sure everyone here saw and loved the piece about fedora being the new Ubuntu Um, which struck me the most about that was that fedora is now a main and I'm grateful to matthew for pointing that out Because in that headline ubuntu was the main and fedora now is the main so well done team But for me a key takeaway of this was the author saying that they felt like fedora was getting features first That sounds like a couple of apps I keep here and brought up around this joint So my question is is there more innovation waiting to be unleashed What about os tree? We're doing it. We're using it. Is it the future? I happen to personally think so and I'd love to see a smart distro prove that or disprove that And then we get to things like modularity blood packs There's also app image snap and probably for other technology someone has come up with while I was talking What is the problem here? What are we solving? Why is this so hard? It is hard But you know what fedora is the right place to solve this The project has the right kind of people the brain power the smarts the contributors The force of bringing people together and we have the right thing to test it in fedora linux to figure out that solution But let's leave this Let's talk about people Now I happen to like people. I like this topic a lot Frankly, I would not be here if you all were not also people matter And fedora is for everyone Matthew likes to remind us a lot and I think it's worth repeating over and over again that fedora Is an operating system and fedora is a community. It's the people It's people who have lived work and grown together Fedora is the mature lived example Fedora is 19 years old We should be able to drive and drink in most countries like we are the mature lived example We should buy an electric guitar But more importantly, we know mentoring We know how to make decisions We know why diversity matters and we are learning how to make it happen These are areas where fedora continues to grow prove and innovate How can we package this up For others to consume to consume in the open source way as my friend karsten wade would say How does fedora lead in this space? You remember earlier. I said that a hundred percent of open source contribution is because of companies Fedora has a lot of experience managing a company's objectives when they intersect a project's decision And frankly, we have a lot of experience with how bad this can go Like bad bad like acrimonious level bad Payback to modularity This isn't an indictment of either side and I'm not going to touch on the fact that red hat employees were on both sides of this issue Candidly, sometimes at the same time Same employee But what I will say is that fedora can teach the world how to handle this in a healthier way And trust me the world needs it The world needs it Let's look at some projects that are rising up fast out there. Let's pick on kubernetes The kubernetes project will happily give you a chart Rank order of the of the corporate affiliation of their contributors ranked by the number of people participating Independent is number four Numbers one two three and five are all major software companies that you have heard of one of them sponsors fedora So how can we share with the kubernetes project the strategies that we have learned about balancing corporate objectives with the natural goals of the community How can we teach them or share with them ways to do that without having fights? Or lots of internal meetings that leak all over their community? And let's think about what individuals Let's think about people whose needs aren't being met and how we can help them Is the entire contribution pipeline we use friendly for the blind? What about people with limited mobility? What can we do to make things more accessible? This is the a11 y accessible This isn't about just the operating system. It's literally about the tooling our community everything And it comes in a lot of flavors I've had the privilege recently to do a lot of conversations or have a lot of conversations with folks in china And a common complaint that i'm hearing there is that there is nothing in chinese To explain the governance of the projects There's a feeling from projects that if you have no english or limited english You are not welcome to participate even as a well-intentioned bug reporter or user And china is just one example of i'm sure many places where this feeling is And there's a second problem Which is the contributors who happen to be bilingual with a second language of english show up And everyone in the open source project is very welcome and excited they're here and suggest they become translators And that's not what they came here for so we shouldn't buttonhole people either But that's like that's a different talk and a different rant But we need to think about how to make our project more accessible And speaking of accessibility I want to talk a little bit about financial accessibility um Not everybody's financial wherewithal is as good as maybe we'd like it to be for them And fedora can't solve this problem with jobs But we can think about things like whether the processes we're using require high speed internet Or if the processes that we use require say a modern pc This doesn't mean we should fall back on email and irc or eliminate options to local builds Candidly those are terrible unfriendly ideas for completely unrelated reasons But it does mean we need to be thoughtful about what we require Have us an option and how we communicate Let's think a little bit about time zones Is all the decision making meetings and conversations seem to happen in just one geography or time zone It's not a good idea to talk bad about your host But um did anybody notice that nest is centered on the u.s. Eastern time zone I definitely did because the first time I was asked to give this talk was the middle of saturday And I love you all but i'm not going to miss daddy time for no good reason And I can give this talk on a friday. Let's make it work and to marie's credit we did as you can see And marie I'm going to call you out here and I know how hard this job is I have watched you do this job so much better than I ever did But I have to point out nest 2021 had the same time zone by us Now I won't belabor this point further But I do think that if we want to double our contributor base We're going to have to re-examine how we meet communicate And define thing or excuse me decide things I think we're going to need to learn a lot about asynchronous communication or in many cases learn it again This is a topic I happen to be personally very passionate about so you can you can find me later. I won't keep Going on about it. But um, I'll ask you some questions to think about for the future along the lines of What can we do to push things to be more asynchronous to let everyone participate easily? Think about whether your team really needs that weekly IRC meeting at 5 p.m. On fridays So in conclusion the future of fedora is fedora Fedora already knows how to transform We know how to keep doing what we've been doing and we know how to do it differently going forward We're going to need to keep embracing change because that's what lets us move the finish line And embracing that change means being strategic So everyone here needs to help the council to find the strategy And I suggest that that strategy should figure out what space on that plane fedora wants to occupy I suggest that fedora focuses itself not as a producer of an operating system But it is an option that third parties customers if you will can choose That doesn't mean we shouldn't focus also on growing contribution I also suggest that fedora focus on that first component of the value segment model Lean into being a project Lean into the value of bits as a solution to specific problems Lean into engagement with folks who are looking to do experiments and build custom operating systems Lean into project consumers who will later become project contributors And enable more usage of fedora throughout the upstream Go where your users need you to be This space is actually huge All of this is a fancy way of saying double down on the reliability of bits Double down on choice Embrace new ways of producing software Package up your learnings Help new projects share a lot Help new projects and share a lot Matthew Miller pointed out yesterday that multiple communities have been so grateful for the fedora code of conduct as a starting point for their own conversation Let's make more documents like that. Let's all contribute to the open source way and make it a better document for guidance Candidly, let's figure out how the next modularity doesn't become acrimonious As coke used to say let's teach the world to sing in perfect harmony Think about people How can we make the project more accessible accessible in a physical sense and accessible in a participation sense Accessibility is something we can do Asynchronous is something we can do Cross-language patterns are something we can figure out Thank you for having me It's been amazing to get to think about these topics Even if I had to procrastinate a bit to get them all out of my head And I hope I can come back next year and every year as a speaker and a participant The future of fedora is fedora So let's go talk in the museum There's a social hour program going on right after this and I'd like to talk to you about this and see where we can take it So please go to the museum. Thank you