 So, hello everyone. If I haven't met you, my name is Sophia Vargas. I'm a program manager in Google's open source programs office where I focus on research and analytical tasks in support of our open source programs, projects, and communities. I've also been working with the chaos community since 2020. And I'm here today to talk about building toward open source sustainability. Now, to be honest, the reason why I pitched this to this group is, right when we were planning ChaosCon, I got invited to participate in the open source sustainability forum, which was a meetup of researchers and industry professionals, foundations, project leaders to talk about what areas we needed to better understand to research open source and sustainability of open source practices and ecosystems. So I thought, well, wouldn't it be great to bring some of the ideas and topics that came up in that discussion to this group, knowing that we wanted to run this session as a discussion? So the second part of that then with this particular presentation was to share some of those concepts in hopes that it can encourage more discussion afterward. So there might not be anything in here that's super groundbreaking, but the idea is to set the stage that when we move into the discussion phase, we can have a productive start, so we don't have to kind of repeat the things that we already know, or maybe you want to challenge what I said. So it gives you more ideas of what to talk about. So I mostly built this as a way to preface the discussion that follows. So what a better way to start then with the definition? Thinking about what is sustainable and sustainability mean? Yes, I do work at Google, but I did start at Google just Googling things, trying to understand what do the internet think about sustainable and sustainability. I liked this from the Oxford Language Dictionary ability to be maintained, and then a couple of other definitions that focused more on sort of the climate themes around sustainability, maintain or support a process continuously over time, and without preventing the depletion of natural or physical resources. And yes, this has sort of a bit of a climate connotation, but I think this also still applies well to open source in the sense that if we think about sustaining open source or open source projects, we are thinking about continuously supporting or maintaining a thing over time. And ideally, by in a way that's not depleting our natural resources, which is us, the community, the people that maintain these things. So I actually thought this fit well, even though it wasn't written for open source. When we brought the same question to this open source sustainability forum, one of the main themes that came through was permanence. But I will say that was something that was contested in the room, where folks didn't like the word permanent because it implied unmoving or unchanging when in most technologies, if you don't change, that's how you become irrelevant. And so maybe the better concept was resilience or adaptability. How do you evolve to continue to be relevant? And so you are continuing to exist, which implies an element of permanence, but doing so requires this ability to adapt and to change and to evolve. The second major theme was understanding, when we talk about sustainability, who is sustainable for whom? And trying to understand the specific personas, I liked that Georg had a whole stakeholder slide because I wish I had seen that before. I oversimplified it into two buckets and yes, there are more kinds of stakeholders, but we have to start the conversation somewhere. So I like to narrow it down on the two most prevalent stakeholders that we think about around open source, those that contribute and create and those that use either personally or on behalf of their company or within their company's infrastructure. So maybe from the perspective of those that create, we are thinking about is this workload manageable? Can we continue to manage, to maintain? What will happen if I step away? Is there someone else, another resource that I can depend on to keep the work going? Maybe from the perspective of the user, they are thinking about, well, how long is this going to be supported? Or how long is this going to be a viable tool that I can use? And maybe supported to them means if there is a security incident, someone patches it. And if they don't, maybe it's time to move away from that thing because it's no longer supported or viable. The second part of this question is not just for whom, but for how long? How are people thinking about scales? Are we thinking about something being available indefinitely or maybe users are thinking about it in years for their company or maybe contributors are coming in and out of projects in months to a year? So if we start thinking about this sort of idea of scale and just a spoiler, I'm going to bring this down to measurement at the very end. And so we need to understand scale and time when we start to apply measurement. I also, if you've pretended any of my talks in the past, like to draw on images from my own home and my own cats. And so for a perspective of how long things might last, does anyone have a guess at how long Moby let that tail hang out of the box before he attacked it? Just shout out anything. Oh, very close. I want to say it was under 20 seconds. So I'm going to call that a yes. Yes. It was good. So the next thing, the next set of people I want to think about is more on the contributors. And so understanding who they are and what motivates them and why they're here as a resource. Why are they coming to your project? This is where I was going to say you Georg and a number of other researchers pursued a project. I think this was a reiteration of a project looking at the motivation of why people are working on open source. What brings them to open source and why do they stick around and does that motivation shift over time? And so I was thinking about this a lot when I was thinking about contributors as the natural resource. Why are they here in the first place? And this research suggests that they're here to improve their skills. They're here to learn. And they're here to have fun because they feel intellectually stimulated by participating in these spaces. There are clearly a number of other motivations, but those are the sort of the top things that stood out to me by this ranking. And the other part of it that's not visible on this slide is that I ran my own study last year with a completely different set of question options. And they agreed with these findings. The top things that we found were that people were there to learn. And the top reasons that they stuck around was because they were having fun participating in these spaces. So when we think about contributors as a natural resource, part of the thing that I start thinking about is why are we here and can we continue to maintain that environment that encourages people to be here? So the flip side of that is why people leave. So here I'm actually showing a slide from the study that I ran on behalf of Google. Note that the sample was comprised of open source developers, contributors, maintainers and students that self-identified in those buckets. And when we looked at reasons why people stopped contributing, lack of time was the top, not that they weren't having fun. They just didn't have enough time to contribute in the way that they wanted. But when we isolated maintainers as part of this population, we saw that they were a little bit more concerned with some of the administrative or operational components, say licensing or legal problems or conflict with contributors. And so other sort of administrative or operational things came up in terms of why they might have stepped away. We also asked about reasons why people didn't contribute to open source and lack of time also came to the forward of that of that ranking. I also want to have a quick note to other members of the KS community who pursued a research project a few years ago, looking at the community characteristics of why people join or leave communities. So this was a focus on just those community characteristics, but I thought this was a nice compliment to the work that we did. I put the GitHub repo link in here, but if there's a better one, someone on the Slack channel might be able to put it in there, just a little suggestion. Thank you to those folks that worked on that research. So why are we talking about this now? I liked, again, Gayarx call out around how pervasive open source is. You had a different stat, and I love seeing different stats that all agree to the same thing, that 70 to 100 percent of code bases contain open source software. Another flip to look at that is 90 percent or more of companies use open source in their software development life cycles. I've seen this quote from multiple different studies. I've also collected this directly in my own research. So I think I'm pretty comfortable with this number, but I actually think it's higher than that because this is usually talking about those that are actively using open source versus those that are passively using it as part of tooling or infrastructure that they purchase from a third party. But the part that really is getting to me is that I feel like maybe this is me being newish to open source, but it's clearly spent around for a long time. But the popularity of it just seems it just here everywhere in mainstream technology environments where organizations now look at open source as a viable solution. And maybe they did before but more are doing that now. So we already have a saturated way to track this how do we know if it's continuing to grow beyond that. And the part that concerns me is how do we know if the level of contribution is scaling at the level of usage. And yes, open source was became popular because of the scalability of being able to download software that didn't add additional strain to the maintainers. But if all the new users start submitting bugs and issues into the repository, then suddenly maintenance isn't scaling as well as usage. So I was trying to measure how many open source developers, those that use open source in the development practices also contribute to open source. And I haven't really been able to find a number. So I've been running a number of studies where if it's open source users, I'll ask how many also contribute back. And I'm seeing about 40% or less is kind of the number that I'm seeing. I'm going to keep asking this because I'm curious to know, is this something that scaling, is it growing, is it changing? I tried to estimate this based on other numbers that, as you can see, don't really make a lot of sense together. Estimates of 20 million global software developers or 90 million developers on GitHub, I guess I have to imagine that their definitions aren't the same or maybe multiple people have multiple accounts on GitHub. I know I do personally. But I know I'm just kind of getting the feeling that I'm not sure if the usage to contribution levels are balanced and is that going to change over time. So one of the last things I want to bring up in terms of sort of, sorry, I just want to do a time check on myself. One of the things that I wanted to think about is what are the tools available to us today to sort of think about how we proactively encourage sustainable practices in our own projects and communities. And one of the main things we have is a governance model. How do we define how rules and decisions are made? And I just listed two examples. There are probably many more where we know actively that a governance model can promote or encourage sustainable practices, say by encouraging behavior that everyone is okay with, that encourages people to stick around because we've reduced or removed toxicity in our communities. But we've also seen cases where governance models can not scale poorly with the community and create structure that's frustrating or complex or you have to go through six people to get an approval. And that could discourage contribution because of the experience of it. So as sort of a side question, notice a lot of these questions are leading up to the questions at the end. I'm curious how well governance models are designed to handle this change, the changes required to be able to adapt and to be able to be sustained. How well does your governance model handle growth, change or disruption as a way to sort of test whether or not your governance model is promoting sustainable practices or not. The last thing I wanted to bring up, I think I'm moving quickly but I think we also started early and that's good because I have a lot of questions, is that this is another personal story. I was at KubeCon three weeks ago and someone asked me what is an example of a successful open source project or what does success mean in open source. And I wasn't really sure how to answer that. So I started thinking, okay well, no one would argue that the Linux project isn't successful and so what are the characteristics of it that make it successful? Is it because it's been around for 30 years? Is it because it still has an active group of participants that are supporting the Linux kernel or is it the incredible amount of usage and popularity that's gained in the user community? Probably some element of all three. So there is a sense that sustainability in that regard is a marker that the Linux project succeeded but also kind of like Garek's quote again, thinking about how many projects don't survive, many projects fail and some create these incredible movements or gravity in the community. So I'm not actually really sure if sustainability is a marker of success because potentially it threw out an idea that created something better. And so as we think about what sustainability means to us, there was a lot of debate about this in the room at the forum again where there was some argument that maybe we should be talking about health not sustainability to understand how to improve practices in front of us immediately versus thinking about sustainability or how to make sure that this thing remains viable, remain supported and last over time. The last wrench I want to throw in is that I've been kind of talking about sustainability from the lens of projects. And I think the more that I thought about it, I wasn't really sure that was the right level to think about sustainability and open source. Maybe we should be thinking about the entire ecosystem. I think thinking again about the Linux foundation and all of the hundreds of projects that sit below it and knowing that while we like this community are separate pockets and have our own models, our own communities within it, there's a lot of overlap that happens between people, foundations, companies, funding, infrastructure that is shared across a whole number of projects. And so maybe when we think about something like sustainability we should really be thinking about the ecosystem and not any individual project. So I have a lot of discussion questions prepared and ideally if I said something that you really didn't agree with that would be a great thing to bring up in this group. But I have sort of three themes and sort of the thought that I had for how we would break up is maybe groups of three to five clusters where you're going to choose your own adventure. And the adventure is starting with sort of fundamental questions that were raised by this. Maybe starting understanding what sustainability means to you. What does it mean versus health? Maybe is this something we should care about or not is sort of what I'm adding to it now. What level should we be thinking about this in terms of project or ecosystem? And what role should the governance process have in terms of supporting or encouraging sustainable practices? So that would be that your group chooses one of these that you're more interested in and pursues that in conversation. The second theme is more specific tactics. Now there could be others that I don't have listed here. But two things that I've been thinking about a lot in my own world are how can money solve this problem or make it worse? We have a lot of change in the technology industry that's going to impact funding models across the board. So if we think well if I had a million dollars today where would I put it to what effect? Who would I give it to? How would I distribute it to the people that need it the most? Is it at the foundation level, the person level, the project level? Who maintains that? Who manages that? Who distributes that? And also thinking about the impact of providing that funding, maybe it's only coming from one company and that company might not exist anymore. Have we just created a dependence on that one company in this funding model? So kind of just thinking more about how money could solve the problem. This is a very big question again. So it could take up all your time but maybe not. The second theme was around how project leaders can signal that they're at risk. In the sense that we in the chaos community risk working group think a lot about areas to measure that would spark concern. Maybe if we remove security from the table maybe we're looking at waning participation or maintainers aren't being backfilled or so I guess that's more of a like a population thing but essentially how do we tell if maintainers are being overworked and they're not able to step away or many of these other indicators but essentially if they're at risk how do they communicate that to the people that need to hear it. We've seen a lot of posts on GitHub. I've seen blog posts but is this an effective way to reach users that might be downloading your software from a third-party package manager and not have any experience with your repository directly. So is there a way that we can reach the users and advertise how best to support the project and what they need. The final question that I would love everyone to get to no matter what questions you discuss in the first part is thinking about how to measure it. We are the metrics group after all I would love to think about if this is something that we want to measure given all the dynamics we discussed and whatever your group chooses to focus on what what will we pick to measure. What can we measure what can we measure what would you like to be able to measure and that's sort of what I would love to come into the summaries at the end of the day if we can remember what we talked about. Gary you're making a face but I think we're ready. Yes okay this is an experiment so I hope this works well. You ready. Okay stream stream paused. The Slack channel is welcome to participate on Slack and what I'm going to do is in this ends in two thirty. Oh wow but I can also end it early because what I'll do is in ten minutes I'll flip the next slide so that you have the prompt in case you want to go there. That's a long time. I don't know if we can talk that much. So I guess like pockets near you I'm going to stop talking because this is when we all start talking. All right it thank you so much it's awesome to see so much conversation happening here in this room that's what we wanted to see so thank you for participating this is fantastic. Now we have a little break I need to check my schedule when are we coming back I should have done that before coming we have like 20 minute break.