Episode 10 - Talking Open Source with Paul O'Shannessy
Joel chats with Meta Open Source Program Engineering Manager Paul O'Shannessy, about all things open source, his journey at Meta and through open source, and a couple of other fun topics.
Summary​
Welcome to the 10th episode of The Diff audio podcast, and the first episode of The Diff video series. In this episode, recorded Thursday October 28, 2021, join your host Joel Marcey as he chats with Meta Open Source Program Engineering Manager Paul O'Shannessy about all the innovation he's played a part in during his decade in open source. Stay tuned for the rapid-fire question section where the two cover such controversial topics as pineapple on pizza and Star Trek vs. Star Wars.
Video Clips​
The Diff: Talking Open Source with Paul O'Shannessy​
The Diff: Supporting Open Source at Scale​
Social Accounts​
Episode Transcript​
Talking Open Source with Paul O'Shannessy
[00:00:00] Joel Marcey: we're back. I'm Joel Marcy, after a bit of hiatus, the diff is back with a new series called behind the scenes where we will explore the stories and journeys of members of the open source community discussing topics that may not necessarily be in the public. Not only will we have the normal podcast audio.
[00:00:24] We are also now rolling out video for this particular series. Yes. You will be able to actually watch us discuss cool things. Oh yeah. And my company has a new name. It's called meta. Our first guest in this pilot episode is Paul O'Shannessy engineering manager for meta open sources, tooling team. In a nutshell, his team develops the software that allows me to actually do open source in a productive, efficient, and secure manner.
[00:00:48] Let's talk to Paul about how he goes about supporting thousands of engineers who do open source at the company.
[00:00:57] Hey, Paul, thank you so much for agreeing to be the [00:01:00] first guest on our new, behind the scenes series for the deaf. And this is our pilot episode. So you're our first ever guest. Uh, I'm super excited to have you kick this off for us. Uh, could you please quickly introduce yourself?
[00:01:11] Paul O'Shannessy: Yeah, I'm really excited to be here, Joel.
[00:01:13] My name is Paul O'Shannessy. I'm an engineering manager here at Facebook supporting our open source tools. I've been doing this team for about five years. And before that doing other open source work here at Facebook. So, um, I've really been in this open source space here for a long time.
[00:01:30] Joel Marcey: Awesome. Yeah, thanks for the background.
[00:01:32] But before we get into the swing of things on all things open source, um, there's some breaking news as we record this on Thursday, October 28th, 2021 or 2021. Um, I don't know if people have noticed on the whiteboard, um, Facebook has changed its name and we are now called. So there's that, that happened on the day we record this initial pilot podcast.
[00:01:56] Uh, any, any immediate reaction to the name, change, Paul?
[00:01:59] Paul O'Shannessy: I [00:02:00] think that was a fun surprise today. Uh, the name really fits with the direction that the company has been going and the direction we see ourselves going in the next, I don't know, 10, 20, 30 years. So I think the name actually makes a lot of sense.
[00:02:15] Um, I think it took a lot of us by surprise, even though. There was a leak last week. Um, maybe people had already guessed this thing.
[00:02:26] Joel Marcey: Yeah. Um, I always wondered how big of a secret it actually was, but, um, now it's official, uh, uh, I feel like personally the biggest issue for me is remembering. That I actually work for a company called Metta and not Facebook.
[00:02:41] And I kind of likened it to how, you know, in January of a new year, you're writing a year on a check and you write the previous year. And so that's going to be me. Someone's going to ask me where I work. And for some time I'm going to say Facebook. Oh, I mean, Metta. So it's going to take a pit to break that cycle.
[00:02:59] Um, [00:03:00] And I kind of realized, I just talked about writing checks, which is kind of old school, but, uh, those still exist. Not everything's done electronically still, but, uh, yeah,
[00:03:09] Paul O'Shannessy: I write a few checks a year and honestly I write data notes. So until honestly, just a couple of weeks ago, I was writing 20, 20 still, uh, many of those notes.
[00:03:20] So I think this, this whole transition to manager. Take me
[00:03:24] Joel Marcey: awhile as well. Yeah. Yeah. We're 10 years. I mean we're 10 months in to 2021. Paul 2020 is a little
[00:03:31] Paul O'Shannessy: embarrassing when I made that realization. It's
[00:03:33] Joel Marcey: okay. You know, it's been kind of a weird couple of years, so it's still March
[00:03:38] Paul O'Shannessy: 20th.
[00:03:39] Joel Marcey: Exactly. Um, okay. So on that breaking news, now we can get to the actual meat of the, of why we're here.
[00:03:46] Um, can you talk a little bit, Paul, about your current role as it pertains to open source? I know you're, you've mentioned you're an engineering manager on the team, but can you kind of dive into that a little bit?
[00:03:57] Paul O'Shannessy: Yeah. So on my current team, we [00:04:00] built a lot of the tools. Teams like PI torch, react native, and many other teams here at Facebook used to do open source well, or better than they would have otherwise.
[00:04:10] So some of that involves code sinking, taking commits from our internal repositories and sinking those out to get hub, enabling internal developers, to retain workflows that they're really familiar with. We have many people who don't even know that they're contributing code to source because we just automate a lot of that behind the scenes.
[00:04:30] So this enables us to keep communities up to date as, as we make changes, as well as kind of improve the, the velocity that folks can move. And so we, we do a lot with GitHub. We do a lot with other tools out that exists out in the community. And so we really, really try to make it as easy as possible for folks at Facebook to do open source, but also make it as easy as possible for folks outside of Facebook to contribute to Facebook open source.[00:05:00]
[00:05:01] I just said Facebook, like 10 times there. I think I'm going to have to say Metta.
[00:05:06] Joel Marcey: I was just about to bring that up. I'm like, how many times are we going to say Facebook before we say Mehta? And we might go through this whole thing saying Facebook. Um, cause it's literally the same day. That's right. Um, I mean, it seems like a good, I mean, it seems like a good goal.
[00:05:21] Um, you know, to both enable. Folks at the company to do open-source as well, as well as, you know, let people outside the company do open source as well. It's a very good mission for your team. I know you had mentioned that in your intro that you've been at phase Metta was Facebook now, Metta. Um, you've been at the company for over 10 years, I guess.
[00:05:42] Um, and you've been, but you've been an engineering manager, you know, for the last four to five years. Can you. Summarize your journey at the company that led you to this team, um, as, as it relates to open-source. So you must have done open source before joining the team that probably gave you a little bit of a headstart [00:06:00] into understanding what the needs and requirements would be.
[00:06:03] Um, can you kind of talk about that a little bit?
[00:06:05] Paul O'Shannessy: Yeah. First off, only been here nine and a half years. Not quite 10. I'm looking forward to double digits next year. Sorry. Yes. And, uh, I've actually been a manager for a little over three years now. Uh, and so got to go through that whole experience of being an individual contributor and making that, that role switch over to engineering manager in an entirely different job.
[00:06:28] Um, but my, my route into open source here at Facebook, um, when I first started, I was on a product team, kind of had a lot of fun, doing very typical products. But my passion has always been around frameworks and developer tools. And so I heard about this thing called react. Uh, the team I was on had started trying to use it.
[00:06:48] It didn't quite fit with our use case at the time, but I found it really exciting. And so I jumped over, had the opportunity to work with Jordan Walker. Photometry Gino Lucchino he's going to kill me. [00:07:00] Um, Uh, I'm not
[00:07:02] Joel Marcey: mad around anymore. She's done
[00:07:03] Paul O'Shannessy: it for three years.
[00:07:06] Joel Marcey: Yeah. He's he's not the company. So he won't, he won't kill you immediately when he listens to this,
[00:07:12] Paul O'Shannessy: he might.
[00:07:13] Um, so I got the opportunity to work on react and I joined that team just before we open source. And actually one of the first things that Tom asked me, I was like, Hey, do you think we could get react ready to open source in the next two months? Uh, leading up to JS comp. So I kind of switched teams in March of 2013 and may end of may.
[00:07:41] 2013 is when we open source reacting. So we had this big sprint to go open source, had a huge amount of fun. Um, you know, it wasn't as huge as success when we first opened sources that has become. Um, so I got the opportunity to stay on that team for a number of years, worked on community-related things, quantum [00:08:00] tooling, you know, every issue inside and out for better or worse, and kind of got to have a lot of fun doing this.
[00:08:08] Um, I learned a huge amount about building communities, building projects, and how they interact with those communities and building things in ways that people outside of the core team can really build on and use those. And from there.
[00:08:26] Joel Marcey: Oh, no continue. I was going to ask you a question about react, but please continue.
[00:08:30] Okay.
[00:08:30] Paul O'Shannessy: Yes. Save all questions for the end, please. Uh, so, uh, from there it actually expanded, uh, into the open source tooling team. This is a fairly young team at the time we had, um, two other folks on the team had been kind of a, a bootstrapped effort to start building more tools for more teams doing open source.
[00:08:52] So one of the first projects around that was actually. Um, which we've open-sourced, uh, it's the main tool we still use that helps [00:09:00] sync our commits from internal repositories out to get hub and anywhere else we might want to sync them. And we've since then built up a whole bunch of other tools to help maintainers.
[00:09:11] Now keep an eye on stack overflow, or respond to vulnerabilities as they're reported by GitHub. Just do a lot of that integration type work and. As I mentioned after a couple years on the team, I actually made the move over to engineering manager. So got to build a lot of the stuff and now get to tell us the folks that I support that that problem you're seeing is actually a direct result of code that I wrote.
[00:09:36] And
[00:09:38] Joel Marcey: it's for better or worse for better or worse. Right. Right. Um, I was thinking when you were mentioning react, how amazing it must feel to like, you know, you say. Tried to open sources. They wanted you to open source react code in two months, and you did that, but just how that project has blown up over the last 78 years to becoming not just like the [00:10:00] most utilized and popular project that Metta as ever, um, uh, produced.
[00:10:06] But. One of the top in the, in the world, um, from a technology perspective. Absolutely. And imagine that you're pretty proud of
[00:10:14] Paul O'Shannessy: that. I am, uh, it's been really incredible to, to see, and, you know, I've crossed after last year, sometime year before, more time on the open source team than on the react team. So I've gotten actually a lot more distance from it than I had in the past.
[00:10:31] Uh, and in, uh, in my current role, I don't get the opportunity to write as much code as I should. So I actually get to talk to a lot of folks who know way more about react than I do now. And that's awesome. I love seeing how others have taken on the taken up that, that effort as well as just, you know, people are smarter than me.
[00:10:52] That's
[00:10:52] Joel Marcey: great. And that pie of people who know react is now. Humongous. [00:11:00] I mean on job descriptions everywhere, if you know, you know, requirement react, you know, it's a thing on job description. You gotta know it. Um, so Metta. I'm doing it. I'm going to do it. Netta. Metta Metta is a big company that takes open source, like really seriously.
[00:11:18] And you kind of discussed a little bit that, you know, previously here, um, we have hundreds of projects, thousands of engineers all contributing to, to the program. I imagine there's some challenges. Um, can you describe the challenges in supporting and open source program as large as Meadows?
[00:11:36] Paul O'Shannessy: Honestly scale is the hardest part.
[00:11:39] So some of those are technical challenges. How do we scale our tools to support a growing number of projects and a growing number of people contributing to those projects? And I mentioned some of the code thinking pieces. We'd do a lot with the GitHub API and using GitHub, web Webhooks and responding to all of these events that are happening in our projects.
[00:11:59][00:12:00] So we have a bunch of technical challenges around keeping up with all of that activity. Um, that's one of the big ones. I think one of the other ones is around of developer behavior. Um, you know, we have thousands of engineers here at Facebook. There are many thousands of people contributing to our projects.
[00:12:18] So as we try to kind of introduce new tools or new changes to workflows, kind of bringing people along as, as we make those shifts can sometimes be a challenge. Looking looking a little bit broader and there are a lot of challenges with projects themselves and kind of figuring out how my team and others can support those projects in, in good ways.
[00:12:43] So, um, security is a really big one. I mentioned get hub vulnerabilities reported through GitHub. Um, security is a super, super vital part of the open source ecosystem. And so how do we. Deploy security fixes. How do we ensure that we're [00:13:00] responding to all of those reports in healthy and fast ways? And so there's a lot of ways where we're also influencing developer behavior in that way.
[00:13:13] Joel Marcey: Is it fair to say? So you, you know, you mentioned a few of these challenges. Is it fair to say that the way to resolve these challenges is throwing more people and resources at these problems? Or is it coming up with creative ways and quote unquote algorithms to try to solve these challenges? Or is it both like more people will help that?
[00:13:34] Or is it basically coming up with new ways to solve problems?
[00:13:38] Paul O'Shannessy: I think it's both. Over my time here, I've seen most of our engineering teams are really like as lean as they can be. So, you know, when I was on the reacting, we were a handful of people and there were millions of people outside the company using react.
[00:13:54] And so like the scale or the ratio of like engineers that [00:14:00] any internal engineer or person working on, one of our projects supports is aspirin. Yeah. So sometimes adding more people can help. Um, other times it is kind of coming up with creative ways or, or really making it easy. I think, you know, we've talked a lot in the engineering world around like the pit of success.
[00:14:22] How do you bring people when we talk about API APIs, how do you bring people to that point where it's really easy to do the right thing? Um, I like to think about that in terms of tools as well. How do we. Really make it easy to go fix that security vulnerability or, you know, go respond to that pull request and do a code review.
[00:14:44] So I do think it is a combination of both.
[00:14:48] Joel Marcey: So I was going to ask you this question, but I think you might've gave us the hint of the answer in a previous response. Um, you mentioned ship it, I think a couple of questions ago, [00:15:00] I was gonna ask you what the first tool you developed for Metta open was, is it, was it ship it, um, or was it something else?
[00:15:06] If it was shipping, maybe you can talk about another tool that's maybe being used today, um, that you, that you developed.
[00:15:14] Paul O'Shannessy: Yeah, so ship. It is probably the first one I worked on on this team. And so kind of something that really supports the entirety of the open source program. Um,
[00:15:24] Joel Marcey: and again, ship it sinks code from internal, your internal Kobe's out to, out to get up.
[00:15:30] Right. So it could be your internal Kobe's could be material and there's conversion mechanisms to take that code over and bring it over to get. Exactly.
[00:15:38] Paul O'Shannessy: Uh, so it's, it's a piece of software that takes commits from one repository, interprets them, kind of strips them back and rewrites them out to a different format or the same format.
[00:15:49] Did it fun things with retaining author information? Um, this is also how we take pull requests from community members, bring them into our internal repositories [00:16:00] and then sync them back out, maintaining that author information. It's really important. Not only for blame purposes, but really to give everybody the credit that they are deserved.
[00:16:12] Um, so should, it's definitely something that's been core to our team. Everybody on the team has contributed to, I spent a bunch of time when I first joined the team, working on that.
[00:16:21] Joel Marcey: Um, yeah, you want to look to it? I even helped
[00:16:24] Paul O'Shannessy: Joel, you know, a thing or two for those who don't know, Joel used to be on the HHVM.
[00:16:31] And worked on some of the early versions of ship it when it was just for that single team, which actually brings, that's why my current team makes this, we really work to take some of these tools that individual teams have built for themselves, um, and take them and broaden them out so that they can support hundreds of open source projects.
[00:16:53] Give them the support that.
[00:16:55] Joel Marcey: And that's like a big thing, right? Because even five years ago, or [00:17:00] six years ago, when the first version of ship it was, um, being developed for HHVM, you could have, you could have siloed tools for individual projects because the number of projects that we were dealing with back then, it was not nothing compared to what we're dealing with today.
[00:17:16] And, you know, your team is to help make that scale. You know, that's really good.
[00:17:22] Paul O'Shannessy: Um, going back to your question though, about what was the first thing I built. Yeah. You know, I mentioned react earlier. I think I'm going to lean in on that one. I think when I joined the react team and we did that sprint to go work on open sourcing it, um, I got to help build a lot of that initial tooling, our build process.
[00:17:43] Almost non-existent internally. We were very integrated into Facebook, internal tooling and bill set ups. And so one of the first things I got to do was like, how do we build this with grunt? We used grunt at the time. Cause it was still cool in 2013. Uh, it's still a great [00:18:00] tool. I think it's been a long time since I've looked.
[00:18:02] Um, and so that was, that's probably one of the first things I actually built for our open source program.
[00:18:10] Joel Marcey: And that tool that you build back then isn't being used now? No,
[00:18:15] Paul O'Shannessy: I think even I had replaced it when I was still on the team. Uh, and it has since evolved as others have, have been working on it.
[00:18:25] Joel Marcey: Got it.
[00:18:25] Um, I kind of want to tell a quick story before I get into my next question and not do a tooling necessarily, but it's something to do with what you did that really impressed me at one point it's kind of a behind the scenes thing. Um, we were, uh, Basically revamping the meta open source website. And, uh, I remember, I think it was four or five years ago at an F eight that, uh, you decided that you didn't like the current.
[00:18:56] Revamp was going and basically at [00:19:00] our booth at F eight, you rewrote our entire website. Uh, I think in with Jekyll at the time. And you did that through one day or night, uh, at eight. And, um, that ended up actually ended up being a springboard into how our website became what it is today through. Um, you know, we have this tool called DocuSign us now, which was based upon, you know, kind of.
[00:19:23] Brought about because of all of our websites, they were in Jekyll, but it was, it was amazing to watch you do that. Uh, uh, what was it, 2017 or 2016 when you did that? It was pretty. And I just want to make sure you get a little bit of credit for that. Cause that was one of the most amazing feats of engineering.
[00:19:37] And I think I see,
[00:19:38] Paul O'Shannessy: so thank you Paul. Yeah. Uh, it was a static cation they'll page and make changes. I had to make it easy. Uh, had to have a, have a bill.
[00:19:49] Joel Marcey: Yeah, I wish we had the code for that still. We might somewhere, but, um, it was, it was amazing. We have it somewhere on my computer. Yeah, you can, you can, you can bring that up in any, uh, [00:20:00] any interview or podcast you, uh, ever go on that and it goes pretty cool.
[00:20:03] So, um, anyway, keep going. Uh, so. This might be a little more of a, uh, kind of a philosophical question, but, um, is there any hard lesson that you've learned over your time in open source? So like for example, is there a decision or effort that you thought at the time was like, yeah, this has to be the way to go, but in hindsight you would do it differently.
[00:20:24] You would do something else.
[00:20:29] Paul O'Shannessy: I, I don't know that I have any singular decisions. No we're particularly difficult at the time. I think I know I've made a number of mistakes and I think, you know, in open source at Metta, we have kind of changed how we've approached some things in open source. Um, so I have personal mistakes. We'll ignore those for the moment.
[00:20:54] But I think when I, one of the things that comes to mind is actually around open-source licensing. Now, one of the [00:21:00] decisions, um, the company had made years ago was like, how. Do opensource and balance, you know, patents and things like that. And so, you know, we had this whole, um, BSD plus patents license situation, um, which was mostly fine for a while.
[00:21:19] And I think as the community grew and as the kind of influence of open source around the world continue to grow, there are a lot more people involved in that. A lot of changes around how people were thinking about open source. So a couple of years ago, we actually changed our licensing on almost all of our projects away from this, this BSD plus patents non-standard licensing to.
[00:21:48] A really permissive MIT license. We removed that whole patents clause and, you know, BSD and MIT are fairly similar. So it was not a dramatic change in terms of, of the actual [00:22:00] licensing. Um, but it gave a lot of people, the confidence to trust our open source and not, you know, use it, not, uh, stop using it out of fear of legal ramifications.
[00:22:13] So I think that's one where we've learned a really good lesson. Um, that, you know, we have to find the right balance of what's right. For, you know, defending and kind of protecting the company versus what's right. To just do for open source.
[00:22:28] Joel Marcey: I think, um, on that note, we think we have a blog post somewhere.
[00:22:32] If I remember correctly, um, about that whole change, um,
[00:22:36] Paul O'Shannessy: my. So like, if you go and look at the react history, I'm pretty sure I'm the one who made the change to introduce the patents clause. Uh, so I guess this one both falls under personal land company lessons to learn.
[00:22:51] Joel Marcey: Yeah. But I think, um, in the end, you know, with it, like with any lesson, um, you know, always good intention to start and, you know, circumstances on the [00:23:00] ground that you might not have considered, um, forced you to rethink something.
[00:23:04] You take that lesson and learn from it and do something even better, which I think in the case you were talking about with the BSD plus patents to MIT, um, and we do all their licenses to, you know, Apache now, everybody. Um, I think that was, you know, that's a, you know, a good lesson that, you know, ended up with.
[00:23:20] Positive result in the end. So yeah,
[00:23:24] Paul O'Shannessy: I think there's actually I'm if you don't mind, I think there's another really important lesson that I've, I've learned over my entire career in opensource. Um, and that's really around community. I think community is a really important factor that often gets, um, discounted.
[00:23:39] When you talk about open source projects and, you know, for react, for example, we really look at how. Grew in the community and really some critical people in the community who gave it some additional exposure or investment that we were ourselves were unable to make. And so this isn't a mistake. It's not a hard [00:24:00] lesson, but I think it's a really important one for people to think about when they are open sourcing their own projects, really think about how you want to grow that community, what you want it to be and what tools you can use along the way to kind of get there.
[00:24:14] You know, codes of conduct are really important. How do you respond to issues and pull requests and like what tone do you take? Um, I think it's really important to kind of help people, you know, make the project that you want it to be and make the community what you want it to be.
[00:24:30] Joel Marcey: That's an excellent point, Paul.
[00:24:32] Yeah. Thanks for, thanks for bringing that up. Um, I mean, community, uh, in the end is pretty much all the reason why we do open source to begin with and, um, you know, any, anything that you can do to. Uh, enable the community in a, you know, in a impactful way or a way that allows them to feel included in the work that we do.
[00:24:53] I mean, it's always going to be a win, I think for us. So, yeah, it's a really good point. Um, I was thinking [00:25:00] about something, uh, You know, given, um, the name change, um, you know, now Metta, um, you know, one of the things that, you know, going back to challenges, one of the things that's going to be, uh, a challenge, um, maybe short-term challenges that we're going to, um, have to figure out how the name change affects, you know, some of the tooling and, uh, you know, uh, social media, um, that we have going on.
[00:25:22] That's another thing we can put on our plate. That's a challenge. If you change your name of your company, there's going to be some ramifications on the backend of how things might work. So indeed,
[00:25:31] Paul O'Shannessy: uh, we can't just do a string replace everywhere for no change Facebook to Metta.
[00:25:37] Joel Marcey: Yeah. So this is like real world problem solving, um, and real world behind the scenes of how open source programs actually work.
[00:25:43] Can we actually do stuff in the back office because.
[00:25:47] Paul O'Shannessy: Very fun. Scrambling message thread going on this morning. It's probably still going on right now. Uh, around all the things we need to chase down.
[00:25:56] Joel Marcey: Yes, yes. And we will be doing that for a couple of days, I think, [00:26:00] at least. So, um, uh, I want to ask you, um, so open source.
[00:26:07] I mean, it's been around a while. Um, you know, Before I started in my career and stuff, but I think open source has really taken off as a, just like, as a global thing, you know, like many companies want to do it. There's a ton of projects. Get hub obviously has allowed for the expansion of open source and, you know, all sorts of these, um, platforms and tools.
[00:26:28] Um, so companies or entities, or even like, uh, you know, people, you know, they might want. Manage some sort of open-source program. Um, what advice now that you, not that you are an engineering manager on the meadow consorts team, um, what advice would you give to those that are looking to start or manage their own open source programs?
[00:26:50] Paul O'Shannessy: It's a really good question. Um, I think every organization is going to be a little bit different. Um, there's a lot of trade-offs to think about. Do you have kind of security trade-offs, you know, risks of [00:27:00] putting. Source code that may never have been visible before onto public servers or public services.
[00:27:07] And so there's a lot from the security side to think about. And so that's where my head has been a lot recently, so I it's top of mind for me. So I do think that's a really important one. I think the other one that you're balancing that with is like ease of use. How easy is it to do open source? Uh, so one of the tools that my team supports is this new project where.
[00:27:29] So in order to create a new project, we have a form internally. Talk about the name of the project that you want to use. Um, why you want to open source it, who inside the company is going to be supporting it, a number of other questions. And that kind of runs through a little bit of the approval process.
[00:27:46] And then you've got a repository. So this replaced a kind of long. Process from before. Whereas a lot of individuals like handing tasks off to the next person. So automating a lot of that in order to make it [00:28:00] really easy to do. And it has resulted in people being able to do that really easily. Um, and my team to go build additional tools.
[00:28:10] So, you know, there's always tooling to build, but I think it's a, it's a balance that you need to strike. Give people the ability to try to start new projects, balanced that with legal security, privacy risks, and kind of come out to a place where, um, you just find what works for your Oregon.
[00:28:31] Joel Marcey: So some people might think, wait, you mean, you just don't create a new project on kid hub and, you know, go, go to go to town on development.
[00:28:39] There's actually processes involved. Um, you know, I imagine for a one person shop or a couple person company that might be the way you could do things, but I imagine as you get to bigger and bigger companies, you know, there's always processes that you have to go through in order to do something. You're saying, you know, think about that, you know, [00:29:00] depending on the size of your company, how, how much process you want to put on the actual people, open-sourcing the project and how, you know, how much easier you can make it to do something.
[00:29:10] Is that kind of what you're talking about?
[00:29:12] Paul O'Shannessy: Absolutely. I think if we were a much smaller company and honestly, at prior points in our company's history, you know, we've leaned much closer to the, like, you can just go and get up and press a button and have a repository. And as we've scaled and we'd need to make other considerations, we've had to kind of shift how we do that.
[00:29:31] Um, but yes, if I were a much smaller company today, I definitely recommend lean as much as possible on what tools exist out there today, either via GitHub or, you know, there are projects coming out of like Linux foundation for CLA tracking, if you use CLS at all. So there are a number of options out there to really jumpstart your own open source project program on.
[00:29:55] Joel Marcey: What would you recommend folks look at, um, initiatives like [00:30:00] to do? Um, you know, um, basically, you know, the, to do group is this group, a consortium of people from different companies and, and, and, um, places that get together to talk about how. How to run an open source program. And I think they even have like a framework to help you do that.
[00:30:15] Do you, do you recommend looking at those types of things or do you recommend sitting, you know, kind of understanding your specific needs and kind of developing something tailored to that versus trying to, you know, shoehorn your own program into some vanilla framework or something like that?
[00:30:32] Paul O'Shannessy: I think the resources that to do group has, are really excellent.
[00:30:36] They may not be tailored specifically to your organization. So I think make those your own, I think there's a number of resources there around, like, how would you convince your company to start an open source program and how might you structure it? I think there are every organization is going to be different where that program reports in the organization, is it direct to the CTO?
[00:30:58] To somewhere else. [00:31:00] Um, those are questions you have to answer for your organization. Cases you'll have to make for yourself, but I do think kind of learning what other companies are doing, learning what other resources may exist is a really good option. And then kind of tailoring it to, to your own specific.
[00:31:19] Joel Marcey: Excellent. Yep. A hundred percent agree. Okay, Paul, um, enough of the serious stuff we're gonna, I think, um, think we're going to have a little bit of fun to close out our discussion here, so let's do it. This is the pilot pilot episode, and I'm going to initially call this the rapid fire response portion.
[00:31:38] Okay. And, um, so basically I'm going to ask a quick question or have some sort of comparison and you need to give your answer or preference. Okay. All right. Are you ready for this? Let's do it. Okay. Metta or Facebook? Metta. Good answer. BIM [00:32:00] or Emacs.
[00:32:02] Paul O'Shannessy: Although to be honest, mostly BS code these days.
[00:32:06] Joel Marcey: Yeah, same here, but I, I, you know, IDE to text-based, uh, you know, terminal base editor kind of different.
[00:32:13] Paul O'Shannessy: So I still open Mack them all the time. My, the text manipulation is still ingrained in my fingers.
[00:32:21] Joel Marcey: Yeah, exactly. Yep. Um, Chrome or Firefox.
[00:32:27] Paul O'Shannessy: I didn't say it today, but I used to work in Zillow. He used to work on Firefox. I was
[00:32:30] Joel Marcey: just gonna say, is that a bias from the previous time at Mozilla is a good browser.
[00:32:36] Yes, it is a very good browser. Um, tabs or spaces spaces two or four,
[00:32:44] Paul O'Shannessy: two. It used to be a four,
[00:32:47] Joel Marcey: but yeah, I mean some languages. There their style guidelines demand for, so, you know
[00:32:53] Paul O'Shannessy: what that's, that's true. I will use whatever style is actually used in the file I'm editing. So I like to respect, you [00:33:00] know, you started an open source project that uses four spaces.
[00:33:03] All right. Uses tabs. I'll use tabs. Uh, there's a lot of great tools. Editor config is one and other, I think a lot of the modern editors actually just will inspect the existing style and try to follow it.
[00:33:16] Joel Marcey: Yeah. There's like settings to be as code. You can do that. Automatically, you know, make your code match to what previous code had or that sort of stuff.
[00:33:24] So, um, GIF or Jeff, it's a hard
[00:33:30] Paul O'Shannessy: GGL. Okay. Your name is not in goal, is it?
[00:33:35] Joel Marcey: It is not goal. I've been called worse, but yeah, it's not. Alright. Alright. Alright. Um, pineapple on pizza.
[00:33:46] Paul O'Shannessy: When appropriate. I do enjoy a ham and pineapple pizza. Am I going to throw ham and mushroom on the same pizza? Probably not. Ooh, sorry.
[00:33:56] Pineapple and mushroom. Hammond. Mushroom. Yeah, that goes together. [00:34:00] But pirate machine mushroom. I don't think so. I
[00:34:02] Joel Marcey: don't think I've ever had it. I would try it, but yeah, it's not, it wouldn't be my immediate choice, but Hammond mushroom for sure. Ham and pineapple, for sure. Yeah.
[00:34:10] Paul O'Shannessy: Yeah. I like sweet, acidic.
[00:34:13] Works. It
[00:34:14] Joel Marcey: works. It works really well actually. Um, I don't know if they have it in Italy. Like if you go to Italy and say, you want a pint, a Hawaiian pizza or pineapple and a pizza, they might get a little offended. I'm not sure.
[00:34:25] Paul O'Shannessy: I mean, I'm sure you can find one. Yeah.
[00:34:29] Joel Marcey: All right. Beat your mountains.
[00:34:35] Paul O'Shannessy: Both. Both. If I have to choose one.
[00:34:43] I like to vacation at the beach. I mean, I live in Seattle. So as beaches are rare to begin with a warm beaches for a few weeks in the summer. Uh, so I will enjoy the beach mostly out of the rarity. We have a lot of [00:35:00] mountains here. I see them every day. Uh, I enjoy visiting and staying there, but I think if I had to choose one today,
[00:35:08] Joel Marcey: Okay.
[00:35:09] Um, I think, um, maybe mountains maybe, but yes, it probably depends on your mood at the time mood, obviously it's mountains, you know,
[00:35:20] Paul O'Shannessy: um, lay down in the sun. Probably not the mountain that
[00:35:24] Joel Marcey: and the mountains. Not the mountains. Not, yeah. Especially if the mountain has some trees, hard to lay in the sun, the trees, but okay.
[00:35:31] This might be the most controversial one. Okay. All star wars or star Trek.
[00:35:42] Paul O'Shannessy: These, this is a hard Munjal
[00:35:45] Joel Marcey: member. Could be millions of people listening to this answer.
[00:35:50] Paul O'Shannessy: Yes. Um, well, if there are millions of people listening, then I will accept any hate I get because that's more people knowing I [00:36:00] exist. Um, so let's see. I grew up watching star Trek the next generation with my dad. So I'm going to go with star Trek.
[00:36:13] I really enjoy star wars.
[00:36:17] Joel Marcey: Um, but if I had to pick
[00:36:18] Paul O'Shannessy: one
[00:36:20] Joel Marcey: star Trek, I'm the opposite of you, but I'm also the same rationale. Like I would choose if, if one had to go, uh, it had, you know, a two star Trek to go. Yeah, because I love star wars, but I don't want star Trek to go, yes, let's have both, you know what, we're going to end it and say that both can still exist and they can coexist, um, piece of reject
[00:36:44] Paul O'Shannessy: the premise of that question.
[00:36:46] Joel Marcey: I might ask it again to other folks and see what kind of responses to get, because there could be someone who has a real powerful opinion for one or the other. And that would be an interesting conversation this year about that. So next
[00:36:58] Paul O'Shannessy: time you have Cammie on, I'm pretty sure. [00:37:00] Her answer is keep star Trek mostly because I don't think she's seen any star wars.
[00:37:05] Joel Marcey: Yeah. Yeah. Uh, for those who don't know, Cammy is one of, uh, the developer advocates on our, on the meta open source team. And if she does come on the show, I will ask her about that and ask her why she hadn't seen any of the star wars. If that's indeed the case, then we can have a real fun discussion. Well, that's it, Paul, I, thanks.
[00:37:25] Thanks so much for coming on. And, uh, this was really a fun discussion. Um, enlightening to me, actually, I actually learned a few tidbits and we know each other really well. I hope, um, the audience, um, kind of understands what goes, goes on a bit and the, you know, behind the scenes of managing an open source program.
[00:37:41] And, um, I'm glad you do what you do. You make things a lot better for, for the program at Metta. And, um, yeah, I really appreciate it. And I really appreciate the time that you took to discuss this with me. Thanks.
[00:37:52] Paul O'Shannessy: Yeah, thank you for having me, Joel. And I look forward to the next season of the diff slash behind the scenes.
[00:37:58] Yeah, it'll be fun. [00:38:00]
[00:38:00] Joel Marcey: Yeah. This is the pilot episode. Hopefully there's a ton more. Um, thanks again. Talk to you soon, Paul Nigel. Bye. Hi everyone. This is Joel Marcy creator and host of the Def. If you like what you heard today on our new, behind the scenes series, tell your friends like it, share it, review it, learn more about our [email protected].
[00:38:20] And if the content you heard today, or if any of our podcasts peak your interest. Checkout Facebook, careers.com to learn more about the challenges we're solving and how we are running an open source program at scale .