Tech Over Tea - Gnome Board President & EndlessOS CEO | Robert McQueen

Episode Date: June 7, 2023

Today we have the one and only Robert McQueen, who at one point helped to cofound Collabora, but nowadays is the CEO of EndlessOS and the GNOME Board President but he has worn other hats over his years. ==========Guest Links========== Twitter: https://twitter.com/ramcq Website: https://ramcq.net/ OpenUK: https://openuk.uk/ EndlessOS: https://www.endlessos.org/ ==========Support The Show========== ► Patreon: https://www.patreon.com/brodierobertson ► Paypal: https://www.paypal.me/BrodieRobertsonVideo ► Amazon USA: https://amzn.to/3d5gykF ► Other Methods: https://cointr.ee/brodierobertson =========Video Platforms========== 🎥 YouTube: https://www.youtube.com/channel/UCBq5p-xOla8xhnrbhu8AIAg =========Audio Release========= 🎵 RSS: https://anchor.fm/s/149fd51c/podcast/rss 🎵 Apple Podcast:https://podcasts.apple.com/us/podcast/tech-over-tea/id1501727953 🎵 Spotify: https://open.spotify.com/show/3IfFpfzlLo7OPsEnl4gbdM 🎵 Google Podcast: https://www.google.com/podcasts?feed=aHR0cHM6Ly9hbmNob3IuZm0vcy8xNDlmZDUxYy9wb2RjYXN0L3Jzcw== 🎵 Anchor: https://anchor.fm/tech-over-tea ==========Social Media========== 🎤 Discord:https://discord.gg/PkMRVn9 🐦 Twitter: https://twitter.com/TechOverTeaShow 📷 Instagram: https://www.instagram.com/techovertea/ 🌐 Mastodon:https://mastodon.social/web/accounts/1093345 ==========Credits========== 🎨 Channel Art: All my art has was created by Supercozman https://twitter.com/Supercozman https://www.instagram.com/supercozman_draws/ DISCLOSURE: Wherever possible I use referral links, which means if you click one of the links in this video or description and make a purchase we may receive a small commission or other compensation.

one good morning good day and good evening this episode has been a nightmare to start i didn't realize when um we're supposed to be doing it but we're here now robert mcqueen welcome to the show uh just give a brief introduction of yourself before something breaks again hi there great to be here thanks for having me on um. So yeah, I'm here in sunny Cambridge in the UK. I am, well, a couple of different things. You've worn very many hats, that's for sure. We have hats. Yeah, I'm the CEO of the NSOS Foundation. So actually, NSOS has been going for just over 11 years now. I joined about eight years ago. And in 2020, we switched over to being a nonprofit. So we have kind of grant-based funding. We work on digital access, digital equity. So helping people overcome barriers like device affordability

and availability of infrastructure. So connectivity to the internet. Does software work for them? Does it respect their privacy? Can they learn stuff? Does it speak their language? And trying to come up with answers to all of those so that people can kind of access this 21st century knowledge economy.

And as part of that, we do endosrs. So we can bring together a bunch of apps and content and games and things into one Linux based system, which is grown based and it's kind of batteries included. You don't need the internet, everything's built in. And so that actually took us to contribute and develop and actually do a lot of work directly in GNOME. And I'm on the board of the GNOME Foundation as the president. And actually, as part of that work, working with GNOME, we helped to start up Fl fat hub because we kind of wanted to see a thriving app ecosystem in NSOS, but also more widely in a free open source desktop in KDE,

you know, mint cinnamon, everything else, long list of things here. Yeah. I was, I was curious about how, I was curious about how that connection to Flathub came about.

Because I've known... You're one of those people like George, where I knew about your work, but I didn't know about you. I was curious how that connection to Flathub came about. But that actually makes a lot of sense. Because Flathub... When did Flathub come about? That would have been... how many years ago now?

2017. 2017. Oh, wow. It's been around for a while. Yeah. So five gone on six years. Yeah. So Endless took a really early bet on a couple of things.

Firstly, on OS3. Because the first idea behind Endless was actually a bit like the Samsung DeX or, you know, the Motorola Ultrix that you would have a mobile phone that if you plug it in, you would be able to put a keyboard and it would go into a TV. And we kind of figured out that basically the economics on that didn't work because a phone that's powerful enough to do that costs more than an inexpensive phone and an inexpensive computer.

So actually, that was the idea. And we're like, oh, yeah, smartphones will be able to do this. But actually, at the time that Endos was starting, they weren't able to do it. And it didn't add up, basically, to get a $600, $700 smartphone out to people so that they could save money on buying a PC. You're like, eh. That was 10 years ago, yes? That was going back around 10 years yeah so um so endless actually

started working on its own arm devices this is when waspy pi was just appearing so we i mean we we made our own hardware okay so this is the Endless Mini. But this is going back in history. So we had our own OS that ran on our own devices. And so we were having to put together the ARM platform support, port the desktop onto these lousy Rackdrivers from ARM, all of this stuff back in the annals of history. But we wanted a really uh reliable kind of base system so so this idea that you know android and chrome os and many systems have where you have this you know core like immutable base system you upgrade from os version 7 to os version 8 there's no packages there's no windows update there's no rebooting like it's just like it runs

and it works um so we we had os tree in kind of EndlessOS from the very start, kind of 2015, 2017 time, so 15, 16. And we've been running with that since then, kind of uninterrupted. We started off with Ubuntu, we switched to Debian at some point, so you could issue a command and switch distro. You switch from 32-bit to 64-bit, and switch distro, you switch from 32 bit to 64 bit, and just rolling forwards. So that made a lot of sense for the core OS, and it made it very reliable, and it meant we could test one thing,

we could ship it out to people, we could just run a couple of branches, one for Intel CPUs, one for the ARM devices that we had. What it didn't solve for us then was applications. So the state of the art then was your choice of PPAs, CFERs, all of this stuff. And like, yeah. So we actually had our own app bundling system,

which was a terrible thing that put zip files in slash endless. And it was like, it was bad. Let's just forget that. So we needed to replace it with something that was a bit less um uh nuts um and um we we saw the uh some of the early work on what became flatback but it was xtg app that um ah yes i thought okay this is this is cool this is a direction that makes perfect sense because we already had the immutable base system.

We were already in that place of wanting the applications to be contained. The other thing that Endless had is we had very much two teams, one that was working on the core OS itself and one that was working on our SDK and apps. So these are what we called knowledge apps, offline copies offline copies of wikipedia and health data news like you name it we had different pipelines and different ways of formatting that a lot of investment into gtk to enable different layouts and constraint-based layouts and things that later went into gtk4 um to build the kind of graphical language that we wanted to to in our sdk um but those guys were like you know modifying the widget set and the applications.

And they wanted to release every, you know, two or three months. And then the OS team were like, well, actually six months would kind of make the most sense for us because, you know, we can get the new kernel and the new kernel release and whatever. And six months was a well-established model at that point with like Ubuntu and things like that. Yeah. Yeah. So that was the cadence of the upstream that they were working with. Whereas people who were putting patches in GTK and then landing stuff on top of it and then wanting to ship it a month later.

And obviously, so the apps could use the new stuff we were doing. Just the idea of being able to separate those and say, okay, great, we have a different release cycle. This is the OS. It's going to be pretty bulletproof. We're going to release it when it's ready, keep it tested, keep up with those upstreams. And then the apps can then just pull in different versions of runtimes and SDKs and things. So the Flatback model made a lot of sense for us because it allowed us to basically give these teams their own cadence. Before, the SDK team would be trying to pull the OS team forward and the OS team would be trying to pull the SDK team backwards.

And actually just switching to Flatpak allowed us to kind of restructure the team even further. So we were able to separate the apps and the SDK team because the SDK team could just keep releasing SDKs and then the apps team could choose whichever SDK to base on. So actually adopting Flatpak changed the org chart in Endless because it allowed us to develop all of the components at their logical speed. I think of it like a Fourier transform. This is probably a really bad analogy, but they're like different frequencies. Like an OS is quite a low frequency thing. It has some high frequency overtones,

like hardware support and things you kind of need to layer in. And then, you know, developing your widget sets is like a, you know, kind of middle frequency. And then your applications are like another high frequency. So you know what? I think you're onto something. That is a really bad analogy. I'll get my coat. But yeah, so there are logical cadences for these different things. No, I get what my coat. But yeah, so there are logical cadences

for these different things. No, I get what you're saying. Using confinement and being able to put, you know, one-times in sandbox things allowed us to deliver things at their kind of natural pace. And we had a whole bunch of third-party applications. Like we had some folks from Egalia

who were like rebuilding devs into this weird endless format. Okay, well well we don't want to do that anymore actually if these are third party applications we kind of want there to just be like an app store and like you know a place we don't want to build an endless app store and try to go around getting people to submit their whatever open source pool game or whatever and put it into our store that's not value add for us what we want to do is just make this better for everyone. And it'll be there for us. It'll be there for everyone else who needs it.

And we'll get the most reach and the most contributions from having that, you know, just be its own thing. So I think we kind of just took over. That was like a GTK hack fest in London or something. And Alex Larson was there and I went and crashed it. And we ended up spending half the time talking about basically Flatpak and how we could start Flathub and kind of what would be necessary for us to have like a kind of logical destination

for people who wanted to publish Flatpaks, people who wanted to find them and download them and install them. So that's how I ended up doing work on Flathub and kind of why it makes sense for Endless to have me spend my time doing that. Do you recall how long XGF and then Flatpak have been around before Flathub became a thing? Not without Googling. Okay, fair enough.

Three years, maybe four. Sorry, I think, you know, there were a couple of attempts at kind of XGF, and I think there was something called Click before it. Yeah, yeah, I do remember this. Yeah, so I think he kind of came, worked on it and stepped away and came back to it a couple of times before it kind of got to a level of maturity.

Like, oh, hey, we have a runtime and you can actually build some Glowmaps on this. This kind of works, you know. And it kind of picked up a bit of momentum kind of when we jumped onto it as well. I think we kind of smashed up a bit of momentum kind of when we jumped onto it as well it's like i think we kind of smashed ahead and and made some stuff work and worked out some of these funny details around os tree and yeah os tree nowadays has picked up a lot of steam with things like um

you know vanilla vanilla try like vanilla oh why do i trip over that one vanilla os there we go blend os and fedora is doing their thing with um fedora 39 they're gonna have like the official images as like a part of like like an actual part of the thing they're distributing um what was the state of os tree like back when you first started getting involved with it was it it anywhere near as powerful as it is today? I think at its heart, it was like a very well-engineered thing that it wasn't too hard for Enders to pick it up and say, okay, great, we can actually take this into production. I'm pretty sure that Enders was the first into production

with an OS3-based OS. Oh, wow. The company that I worked for before Endless was Calabra. They were working on... You didn't just work for Calabra. Calabra worked for me. You were a co-founder of Calabra.

We did some consulting work for Endless before I left Calabro and joined Endless. So some of the internal R&D stuff that Calabro was doing was based around OS3, and Endless was kind of the first customer that we set up OS3 for, and it just worked super well for what Endless was doing. And I'm sure that Calabro's got some other OS3 stuff going on,

but I keep track of every last thing that they're doing. But certainly in the desktop space um endos os kind of got out with os3 first and at that point it was kind of mostly it was from gnome continuous so that's where colin walters had worked on the kind of continuous integration this build system to try and move away from jh build and have a system that would just produce binaries of like this is the latest kind of head of everything that's going on in gnome just to make um the integration and the testing um you know on this thesis that's if it's not being built then it's not being run and if it's not being run it's not being tested and if it's not being tested it's lousy basically so so

how do you have you collapsed that entire pipeline down to like you know commit and then runnable binary or even runnable desktop or even runnable os so going continuous was was working on that and the os tree was basically a side effect of just trying to build an efficient kind of layered build system with caching and binaries and you know reusing stuff so i had all these properties that made it really great for managing the bootable os um We actually never used any of the build stuff. NSRS was always built by basically just taking Debian and putting it through a woodchipper and then just getting all the binaries out. Debian's a great way for arranging everything neatly if you like describing software in text files. But there's lots of people who are great at that and they do all of that. And then you can just show up and like, okay, thanks.

Put the binaries in and carry on. I'm kind of post distro at this point. It doesn't really matter to me what the distro is. It's like, okay, there's this thing that makes binaries. They do some security. It's lovely. Thanks for the installer. Okay, great.

Well, especially if you're dealing with a lot of flat packs, like the underlying distro doesn't really matter for the most part. Yeah, it's actually really freeing. Like, you know, we're seeing things like Fedora and Silverblue that put together a GNOME desktop, and they do that using RPMs and the stuff that they're doing. The RPM OST stuff is really cool just because it kind of solves one of the problems we had at Embers.

Basically, it was a one-way transformation. So we had devs in the build system. We uh obs that can open build studio um from uh suza with various patches to make it work properly with debian and because debian's special and well it's not our game based right so this is right right right impedance mismatch um so just trying to get all that stuff to work and then dumping it out into an OS tree. There's kind of like a hidden command on NSOS that kind of puts the Debian back into the OS tree a little bit and makes it all writable. And then you can just use apps and just grab stuff. And obviously this is like breaks the seal, avoids the warranty, like abandon all you enter.

But for developments, I'm just making incremental changes and testing stuff. It's really handy. RPM OS 3 kind of like systematizes that. You're allowed to add RPMs into your base system. And then you still have something that RPM OS 3 can understand. It can upgrade. It can do the rebases for you and things. So that's super cool stuff. And I can see why it's found like a good kind of match with core os and the

cloud stuff that um conan walters you know division of 2000 engineers or whatever is doing now one of the things the uh the so nsos doesn't just have like the distro itself there's also this endless key thing what is that i saw it the page. I didn't really understand it completely. Yeah. So that comes from kind of a change in our philosophy moving from a startup, I say startup, but it ran for six or seven years, but certainly a kind of investment backed for-profit company where our idea was basically that we'd solved all the problems in NSOS. And then if you've got nsos then we would be getting all of the solutions out to all of the people who use nsos so it's kind of like a sort of winner takes all thing maximizing stakeholder value like you can get you know

your content and your apps into our os and if you're running our os you're getting everything that we've kind of thought of which which makes sense for a kind of investment strategy where you want to have a nice graph that goes up and to the right and people count money in and it's great. In terms of a nonprofit where actually we don't specifically care about people using our product, we care about the outcome of people using our product. So we care about are people accessing learning resources or playing games to teach them themselves? Are people able to use their computer offline? Are people able to trust their desktop? So these questions are kind of, they're at a meta level. It's not just build the

product, ship the product. It's more to do with kind of just untangling some of the threads of what we're doing and working out what's the logical way to get our work on the desktop to as many people as possible actually just getting people to ship and so to use nsos is not the easiest way to do that going to gnome and then spending our time and effort you know researching and helping to contribute and the stuff that george's has been doing and um you know we've had various designers working in the design team and actually kind of helping to research and improve the ground desktop for everyone. That takes longer, probably costs us a bit more than just putting some patches. I mean, it's a near-term, long-term thing.

In long-term, going upstream saves you money. But in the short term, if you want it to look just like that, it's quicker to just patch it or add an extension or something. And then this is going to zigzag between those. But this idea of taking everything that we were doing and saying, okay, great, to get this desktop out to as many people and make it as usable as possible. There's, you know, some tens of thousands of NSOS users,

but there are probably some tens of millions of GNOME users, right? So we can go into a much bigger audience if we take our work over there. And we kind of applied the same logic to all the other things we were doing as well. So, you know, we had these applications and a whole pipeline for downloading websites and putting them into applications and things. We thought, well, if we support that on different platforms and not just MSOS, then we can extend the reach of that idea that you can use storage and kind of offline content as a way to correct for Internet availability. Right. Whether that's infrastructure or cost or reliability or any combination of those.

So it's just about kind of taking the different ideas that Endless had, everything built into one box and just saying, okay, how do we logically take that idea out to more people and deliver more impact with that concept? So that's why some of the things are kind of playing outside of NSOS, and we still have NSOS that brings everything together of in in a sort of serving suggestion from us so you've been involved in endless os for how long now seven years seven years about to come up to eight yeah so i suppose i started a little bit before i

started because i was working as a consultant so right it's probably eight at this point but certainly seven as a as a member of staff so had you always been interested in doing something that has the sort of effect that endless os has was this something that you realized you wanted to do later in life or how did you sort of make your way to doing this? Have you thought about this question before? Is there a better place to start? I mean, no. I mean, if you'd asked me 15 years ago or 20 years ago or something,

I would have said that, you know, free software is going to change the world. And as long as everyone gets a written copy of the GPL along with their television, then, you know, their life will be amazing. And working in Calabria for 10 years, and we would go into these tech companies, and they would be doing the most ghastly things with their software, and their strategy would make no sense. And they would ask us to come and solve their problems for them. And we're like, well, your problem is your strategy makes no sense. Your SDK doesn't work. Your developers hate you.

Your product's not that great. You're marking it wrong. You're a little bit late. Your competitors are faster than you. And here's the project you asked us to do. And we would get paid. We would do this great work. We would put patches in GStream or Wayland or whatever.

But then eventually you kind of see that you're kind of building the, you know, another terrible analogy or whatever building the uh you know another terrible analogy whatever but you know this kind of this square is green this is working and we've delivered that and we've done a great job and we've got you know a great team working on it but everything else is just like on fire and it's never going to work so you could you could do a completely successful project deliver some really amazing technical work get paid for it you know pay the team,

keep the lights on, and still know that it was going to affect nothing, basically. It doesn't matter whether we optimize WebKit in this TV. If we do a great job of that, it's still a web browser in a TV. And effectively, no one wants a web browser in a TV, but everyone has to have one because everyone else has got one. So they could just license Opera or get brave to put some crypto mining in your TV or whatever, right? Because everyone's just going to turn that off and never connect their TV to the internet because no one actually wanted smart TV. And yet smart TV, whatever, pays this money to Calabra and do this project. I think, well, even though the source code to this TV is available, functionally, no one reflashes their TV.

Right. Or to a first approximation, like 0.0000. Yeah, there's going to be one person who's like, I've reflashed my TV. It's like, we're not talking about you. I went to the office and I asserted my TPL rights and I got the bootloader and I have now installed Manjaro on my television or whatever, right?

Yeah. You know, all power to to to that person but in the grand scheme of things I was just thinking I'd be very lucky in life I you know I'm here in Cambridge I studied in Cambridge uh you know nice house lovely family um you know there's there's many things that have gone well for me and I'm sitting on top of a lot of kind of privilege and opportunity that so many people in the world don't have and so i kind of wanted to find a way to you know not just make it about the software and the source code and things but make it about in a sense the the ideology underlying that like why do we care about people being able to get the source code or trust their computer or whatever it's to do with empowerment it's to do with accountability and it's to do with what

enders uh talks about digital agency and so this idea of are you controlling technology or is technology controlling you um and these are kind of much bigger questions at stake do people know how computers impact their lives or how decisions that affect their lives and make using computers? Do people know how they've done with AI or machine learning or whatever? And even like there's a whole thing there of people researching and using and creating stuff and actually understanding how it works is like a separate research stream. So there's like,

there's a research team at open AI who are using like an older version of gpt to explain how the newer version of gpt works this kind of stuff and like and that's choosing you know your benefits application or whether you get credit or you know how your kind of healthcare request is being processed through the system whatever so just technology is this big thing that just happens to so many people. And so this idea of empowering people with kind of accountability and understanding just became much more important to me as I realized like I'm very lucky in that I, you know, hey, technology has bought me all of these great things, but also I understand a lot about technology and how it impacts people. So this idea of just taking that out and and

kind of trying to improve the opportunities that people had and improve the kind of trust people have in psychology that was a good answer but hey if you want twitter on your fridge or web browser your tv or whatever like you know go for it i for it. I don't. I really don't. No one ever did, it turned out. I saw this post, I think it was like a month or so ago, where someone was like, they're getting ready to have a barbecue and they couldn't because the barbecue needed an update. It needed to connect to the internet and do a system update or they couldn't turn it on.

It's like, no just it just cooks food i don't need i don't need a computer in this stop no please please stop i mean yeah i have friends who who's uh they have a tap in their office that has like software and you know like they it wouldn't give them water because it needed like an over the hour update yeah like really that's that's an i love it um so one thing i was curious about um so you are the ceo of the endless os foundation what i i understand like the that being a position but i have no idea what that entails as a job what do you actually do in this position like i have no idea i don't think anyone has an idea what a ceo does they just have this like

the ceo is just this person that exists i just say you've described my Monday morning like, oh my God, what am I doing? So functionally, like day to day, I'm managing the leadership team in Endless. So we've got different departments that report to me. So we have an engineering team that does sort of technology and product and design stuff. We have a kind a program outreach team. So we have a VP of programs, and we've got people who go out and talk to partners and try and set up partnerships and deployments

where people will use the stuff that we build and we'll help them, they'll understand it, roll it out in different communities. We have a commercial team that's working on opening channels for our laptop financing program. So we did some pilots in the US and now we're kind of trying to build some new while building scale some some pilots in Guatemala and Mexico. And so

there are this very recently we have a chief of staff. But before that, the operations staff, the finance and our kind of people team reporting to me so there's managing stuff um i think there's also uh kind of making our making our strategy you know work and make sense and that's kind of ongoing conversation between um kind of our founder um matt dalio and kind of the the philanthropy team that kind of funds us and our board and our leadership team and trying to kind of synthesize these different inputs and figure out like what is a good thing for us to do in terms of our programs and our partners what to do with impact um what do we

think is cool and worth doing from a product and an engineering perspective and what do those different stakeholders think is going to work and what do they want to do so it's kind of a an ongoing conversation of kind of hearing and figuring out these different inputs and then kind of updating the plan and setting the goals for the team and then helping to lead the team to achieve those goals. And, you know, all of those sort of trade-offs of management of like stepping back and letting people solve the problems or figuring out when people need more specific support and how I can step in or how I can bring people in to make sure that the team is kind of productive. I don't know if that's a good answer, but a bit of operations, a bit of strategy, a bit of finance, a bit of management, a little bit of just leadership by saying, like, we should go over here because it seems like the best thing to do um kind of some

some combination uh of those it certainly gives me a better answer than i had before i i said i didn't know i i understand that ceo existed as a position but i never bothered to actually look into it i think most people just assume the CEO is just a figurehead that doesn't do anything so look that's that's that's as good of answer as uh I could uh I could expect yeah I mean it depends on what zoom level you say I mean there's one zoom level where like I go to a lot of meetings um I occasionally write some documents I comment on a lot of documents sometimes I open open a spreadsheet. There's a bit of slack. And then a pretty terrible email, but that happens somewhere in the day as well.

But yeah, it's all about just trying to get everyone to understand the goals, understand what we need to get to the goals, making sure they have it and making sure that we all think it's a good idea and we're all moving in the same direction. Right. Well, let's shift direction over to gnome so i presume that you had been involved in gnome for a while before you you know became the board president how long had you actually been maybe as like you know even just using gnome or doing like development work

like how long had you actually been involved in gnome in some way uh i mean certainly over 20 years yeah so i was the start of gnome basically yeah not not long after i mean i think the first version I used was 1.2. So go back a while. I remember the exciting transition to GTK2 and how that broke everything. But I originally came from the Debian community as a package maintainer. And I had a file server under the stairs, completely unnecessarily. And I was running Samba, and I had KDE on it, but I didn't have enough RAM, or something, I don't know, I can't remember. Or, oh, I had a Red Hat 5.2 CD.

So that was what got me my first copy of Linux. And then I wanted KDE, but I couldn't get it, and so I switched to Debian, and then found this KDE, and then I ended up back on GNOME, because that didn't work either something like that um but um i've been so i've been using them you know 25 or something i don't know something like that years um from very early on and i remember xamarin being really cool and exciting when it came out and like i can get like a shiny globe with extra monkeys um i'm sure there was a business plan as

well but it was all about the monkeys really um and you know all of that stuff you know i was a teenager and that was just exciting stuff happening and i could just download the cool software and have a fiddle with it but i actually started developing uh on what was then called game, now called Pigeon, into Messenger stuff. That's what kind of led me into some of the conversations that we got in contact with Nokia, and this was around DevCon for 2005, and they were doing this crazy Linux open source phone, not phone, tablet thingy.

And they had a really cool strategy, which is basically just to go find open source communities and individuals who are working on stuff that seemed useful to them. And then setting up these contracts with them to basically go out and see what they could get and bring into their platform from people who are already working in the space.

So Collabra just kind of replicated that model in the sense of, you know, connecting to all of these different open source developers and different technology spaces and setting up kind of a business model around solving the business problems using open source community and setting up this kind of bi-directional kind of contribution and being able to reuse the open source stuff in the technology space. But yeah, I was already working on GTK stuff then. And I think when we set Colabora up in

2005, that's when I kind of showed up to ended up sort of helping to maintain some Dbus bits and pieces and I think trying to convince everyone to use telepathy, everyone's favorite inter-messaging abstraction framework. I'm also kind of post messaging abstraction. I don't think you should do it. Gatewaying is terrible. Multi-protocol UIs are really hard to get right.

And probably just picking a protocol and using it is easier. So consider- I was young. Before we get into the collaborative stuff because i don't want to talk about that as well what initially um drew you to linux because surely there was a point where you weren't using linux then um going back a few years now aren't we yeah i mean i i had i had windows on on my computers before i had um linux on it and it was a combination of two things i think it was mostly star trek role-playing um but but through a couple

of different uh words yeah uh so uh irc real-time role-playing games um and uh trying to set up was it mugs or mushes or you know these kind of uh text-based multiplayer things that would run like on some bsd server and everyone would tell that in um and that just building and running those meant i spent a lot of my time like talenting or maybe ssh i don't know into these shell servers and trying to run that stuff um and also then spending a lot of time trying to customize my desktop so that it looked like a sort of like shell where you could put your own but basically describe your entire desktop with like pix maps and like hot zones and like relocate the trade components and just replace explorer.exe because that was lame um and you know windows just wasn't ready for me you know i just couldn't do those things on

on a windows system um so it's a combination of those two just got me into using kind of stuff and then i just wanted to um keep learning about servers and how they worked i i printed out a whole bunch of the Linux documentation project stuff. So I'd go on vacation with my parents and print out 100 pages of reading how to set up your Linux print server. I was very exciting as a teenager, as you can see. That's worked out for you today, hasn't it? Yeah.

My parents didn't have to worry about me staying up too late drinking it was more yeah just you know on isc staying up too late reading linux documentation yes that i i like to ask people what their story is for getting into linux because it's always a different answer but that one's extra different yeah I don't know I probably shouldn't say you can you can dig around and find some of the like websites for the way back machine the role-playing games that I was running so Collabora came about in what year was that? That was 2005.

So that's the year that I graduated from university. Oh, oh. I had a job offer from Intel Research to go and work on Zen. Okay. So the Cambridge Computer Lab had what was called Intel Upstairs, the research team who had a kind of wing of the building. called Intel Upstairs, the research team who had a kind of wing of the building. And I actually interviewed there and had a job offer there.

And then I met these Debian folks in the pub who were like, oh, yeah, we're going to a Debian conference and we're going to go meet with Nokia and talk about this VoIP stuff and want to come along. So, yeah, we designed some software on a napkin, pitched it to them. And they're like, yeah, that's know come come back and start next month like oh okay we're gonna use dbus so um you turned down a job at intel as you were just leaving your study to go and start

a thing with Nokia that you didn't know was going to become what it is today. It's just let's do this thing. Yeah. Sure. I was like, you know, I didn't have that much to lose.

If this doesn't work, I could probably figure out. I can go get a real job. Yeah, fair enough. That's been true for 20 years. work i could probably figure out i can go get a real job yeah yeah fair enough and yeah that's been true for 20 years if it doesn't work out i'll go get a real job i'm still here well nowadays a lot of people recognize the name collabora because the work that they've done alongside valve on like improving kde proving things for the Steam Deck, things like that. But what was it like in those early days? Besides this Nokia contract that appeared out of nowhere from a napkin,

what sort of work was, obviously, can't say anything that can't be said, obviously, but, like, what sort of work was being done back then? Well, in the sort of early few years, I mean, we had this really great kind of honeymoon period with Nokia and they were kind of going from their 770 internet tablet and then they were kind of gradually tacking towards producing a phone OS and the Nokia N900 being like a really proud moment for Collabro because we had stuff in the media stack, the graphics stack, the desktop compositor, the messaging and calls, like Collabro was all over that. And behind each of those things that we were doing with Nokia, we were kind of building a

team and a capability inside Collabro that we were then trying to find additional opportunities. So Nokia was kind of our, what's the kind of like flagship client that allowed us to grow the team and the capability. But we could also see that Nokia was kind of weird. They seemed to have this very internal kind of land grab combative kind of approach. They had multiple branches of Symbian OS. They had multiple different open source operating systems internally like they had at least two linuxes if not three and then like some some guy went and set up like this bsd team and there's a lot of like kind of corporate like weird like oh we're gonna merge these two divisions and then like this vp is

gonna go and take over this thing so that he can shut it down and move the resources over to his. And it was like kind of four-dimensional chess with offices and teams and operating systems. And it's just like, isn't there someone in the head office who's made a decision? Yeah, that sounds like a CEO thing. Yeah. Kind of not that great at software strategy, it turns out. But actually, it's a completely different, I mean, you see this with car manufacturers now. Like, you know, Tesla have turned the whole kind of car discussion into like, okay, we have a battery and a motor and then a computer.

And everything else is software. Yeah, Tesla's not a car company. They just happen to put software in a car. The real car companies can make... It's a computer on wheels. they just happen to put software in a car. The real car companies can make... It's a computer on wheels. Yeah. When Tesla brought out their car and made electric cars actually popular, then the real car companies can come in and actually make a car.

Tesla just gets everyone interested in it. Yeah. The same thing happens to Nokia, that basically all of their hardware, different form factors and shapes and buttons and switches and all of that stuff was just rendered completely irrelevant. Because the entire discussion was moved to this is now a slab of glass and everything is about the software. And they were just not ready for that. 2005, that's like what, two? How many years before the iPhone? Like one or three?

Yeah, it's like one or three years like two or three years yeah so we were we were part of the like rebel alliance inside nokia trying to like make some actually good software come out and be relevant before nokia just trips over their own shoelaces getting out of bed in the morning um but ultimately the shoelaces won because nokia had just tied themselves up so horrifically badly internally you know they multiple like three different versions of symbian right your series 60 series 80 series 90. another os two linux os or something else something else and you know just no coherence with making decisions about what to do and what not to do with software. And, you know, obviously we felt and the team inside Nokia, they built a really amazing team with, you know, international collaborators and contractors and open source people.

And, you know, it was a really cool place and time to be. But you could kind of see how different that was to the way that the rest of Nokia worked. So they just didn't feel like something that would last. So our priority at Calabro was trying to sort of use our skills, kind of overlapping between the open source world and the corporate world, what were they trying to do and kind of deliver the kind of weird open source thing in a way that companies could get their head around. It wasn't really a question of whether they wanted to use open source thing in a way that companies could get their head around. It wasn't really a question of whether they wanted to use open source because they got that from the economic perspective. Like someone else has written it for us and we can take it and we can save money.

It was more to do with like actually delivering successfully. So being able to kind of understand roadmaps, influence things, work upstream and kind of get the kind of mid to long-term efficiencies of making good decisions we had some very interesting um uh contracts over the time we actually had to fire one of our clients which was a uh large korean device manufacturer okay okay i think yeah make a number of devices including lots of phones yeah i think we can work this way a number of devices, including lots of phones. Yeah, I think we can work this way. They had a version

of GTK that they forked and been working on for, I don't know, three or four years with just a number of misapprehensions about how computers worked. And they came to us just like, we want to do open source.

Can you upstream this? Like, I'm going to go with a no. Could you explain why you did this and what you were trying to achieve? And then we'll take it from there. And then it was like, okay, we'll write you a report about what we think you were trying to do and then we can write you a proposal about maybe just doing that again but yeah so this the this kind of throwing engineers at it and like yeah there's there were these things like the kind of limo conference linux in mobile and there was some vp for motorola and she would fly in and she would give the exact same presentation and she would fly

out again she's like yeah we love linux we love open source we've got 2500 developers in our bangalore office woo linux like huh you'd think they'd send in a few more patches if they've got 2,000 people working on it because I didn't see any actually which is remarkable and you see how well that you know efficient and on-point software strategy worked for Motorola. Oh yeah I don't know what that company's doing so the last time I paid attention to last time time I paid attention to Nokia and Motorola, they were making... Either cheat... It was when there were some Windows phones coming out.

Motorola got acquired by Google for like 20 cents. Yeah. And Nokia got acquired by Microsoft. Yeah, they were doing the Microsoft phones for a while and then the Windows phones. And then Motorola makes mid-range Android phones that are just basically shipping. It's basically like the cheap Pixel.

That's what it is. It ships like a stock Android. Mid-range Pixels. Yeah, yeah, yeah. I don't even know if they're doing that. I don't know if they've just become... Yeah, I don't know. I've not paid attention to them for a while.

I did have a Motorola... They were like nokia they had a product roadmap that had like 6 000 different handsets and this one is for you know mid to high end mobile professionals on tuesdays and like this one is the same except it has a flippy gizmo and this one has a slightly longer antenna and like and then you have different sales teams who have different targets and different geographies and things like how does anyone like even understand this and it turns out that google acquired them like we're gonna make a phone yep like wow mind blown they're crazy about that isn't it you simplify your product stack people actually understand what's going on yeah i'm not fight with themselves

over who gets the cornflakes at breakfast but yeah it sounds like companies for you it sounds like there was a lot of like sort of not made here syndrome between the departments where like each of them wanted to be the one that did it but then someone would come in and they would be the one who wants to do it and someone else would move stuff around it just it just all tries like devour itself yeah and and it's um it's essentially well software engineering has a lot to answer for because um even just the term suggests that it's a kind of purely sort of analytical and mathematical kind of empirical process. But the reality is that your kind of social design

of how your organization is structured and how you communicate and how you set targets and measure them and things is reflected in the software that you produce. And it's probably more like coral or something, but it grows and then you shell behind. Here we go. I'm sorry, I'm terrible at metaphors. No, it's probably more like coral or something but it grows and then here we go i'm sorry i'm terrible let's see where this one goes but but the way that the organism works reflects the environment it creates and leaves behind right so so if it grows this way it grows this shape then that's the shape of the shell uh-huh and then the growing is here and

whatever so if you have like a really weird organization structure, you set goals in a really weird way, and you don't invest in certain ways, then you end up with also really weird software that reflects it. So you have, you know, this kind of fragility, or you're not ready for requirements to change. And so you big companies, which have these really complicated structures, kind of stamp those really complicated structures on the software that they make. I think it's Conway's Law. Don't quote me, but check on Wikipedia. But it's basically like the software architecture

will reflect the organizational architecture. Yes, Conway's Law. That creates it. And Nokia was like a worked example of Conway's Law. Some of my most profitable days as a consultant would be like in helsinki and they'd be like have this conversation and like you already have a team who's working on solving this problem they are three doors that way let's go and talk to them and they're like wow we have an api for single sign-on yes you yeah meet you see he's the single sign-on he can he can hook you up let's do that i'm like yeah my work here is done um and in a sense like when you and that was samsung's problem as well except like

tenfold, they would approach things with this massive hierarchy. And there would be like a guy whose job was to write the software for the button widget. And like he would maintain the button and then like, well, you can't really, if you've preselected the way that your organization architecture works and your software architecture, you can't like, like a tappable button. Like, know, that's a different person. We need to restructure the team or whatever. So this is really weird. What happens when that guy leaves? Right. No more buttons. No more. Yeah. So yeah, lots of really strange kind of organizational

dynamics. And in a sense, that was the that we we could see being these kind of outside consultants and i think that's that's something that um we were trying to do a lot is kind of have a bigger conversation with customers it wasn't just about their software and the open source and things it was more to do with um the way that they approached engineering internally and the way that they aligned that with the way that open source software and the community process and governance and things worked outside. So in a sense, more to do with kind of strategy and from the strategy, how you can then effectively set up the tools and the team to succeed and to be efficient. And, and, you know, be ready for the kind of business conditions to change as well.

But not, not all customers wanted that. They, you know, if you were talking to a software person, they wanted to buy some software or buy some services. And then you could, you know, there'd be no situations I was describing earlier. And like, this is going nowhere. I mean, we'll do the job. We'll give you the code. So obviously, like what you're doing now at EndlessOS is very different from what you're doing when you were doing that work at Collabora. But I would imagine that seeing how these organizations are structured has given you some insight on like what you don't want to do being in that like management leadership position.

Yeah, I mean, it's hard, right because you it's easy to point fun at others but sure it's also pretty hard to to um to set up the you know the the right conditions for success i mean that's kind of the art of of management is basically if you're good at it it looks like you're not doing anything right because in a a sense, you've got the team set up and you've got the team that you need. They have everything they need. They've got clear inputs and outputs and goals. And then they are empowered to go and solve

all of the things that they need to solve. And so if you get everything around the edge, all the interfaces and all of those things well defined, then the team is then able to organize themselves. And we're having a sort of an experiment at the moment all of the those things well defined then team is then able to organize themselves and we um we're having a sort of an experiment at moments in in endless because we have had this problem of um you know we had a larger team when we were endless mobile kind of for profit and we had to reduce the team size because we collect money as a non-profit and and obviously these all you know um really horrible situations but we kind of end up with like,

okay, we now have one person who really knows how this component or this software works. And now that team that used to be seven people is now four people. And actually, if you give three or four goals to that team of three or four people, then everyone is working on a separate thing. And you kind of lose some of this idea

of the team coming together and working on a separate thing. And you kind of lose some of this idea of the team coming together and working on the organization goals and sort of solving all those things. So we have an experiment now, which was actually the idea of Daniel Drake, who's our engineering VP, that we would actually put all of our projects on hold apart from one of them.

And well, to the greatest extent, as on hold as you can get, given that we've got customers that need to maintain stuff and whatever, and have the entire team in terms of development working on one product all together. So we're doing that for six months. And so we're in our first semester

working on Endless Key with the whole team. So we have kind of a platform, an OS team, and we have like an apps, more kind of UI user with the whole team. So we have kind of a platform with the OS team and we have like an apps, more kind of UI user-facing experience team. And they're both working on EndersKey, sort of the built system, the infrastructure, the content recommendations, the presentation, the UI design.

And that's an experiment we're doing at the moment to try and kind of allow people to go and understand our priorities at a higher level, but then to solve some of the micro detail themselves, because that gives them the opportunity to research and to choose how things are done. And if you do too much kind of centralized packing, then you end up kind of taking away people's opportunity to, I guess, you know, feel ownership in what we're doing. If it's just like a task comes off the queue and you do something,

then you're kind of just in a kind of conveyor belt situation. You're not able to kind of exercise some of that creativity in problem solving. So we're trying to provide more high level context and less kind of micro detail and kind of give away some of the planning and let the team sort of have more involvement in choosing their tasks and priorities

in the way that they approach things. That, sadly, that was a really good speech you went on, but I was having some connection issues during that. It's like a good portion of it just got completely cut out, which is really... If I could get Google Meet to set up working properly, I would have loved to use that. But yeah, I'll see what would have loved to use that.

But, yeah. I'll see what I can salvage from that. Hopefully that works. Do you want me to go back and answer the question again? No, no, you can hear. It was fine most of the way, but there was a couple of points in there where it got a little bit sadicky. I think it was still...

You could still follow along with it, but, yeah. Jitsi can be a bit rough sometimes. That's why I tend, it's sort of like my backup of the backups. Yeah. Whatever. Tech, as I said earlier, I hate computers. They're awful.

As I said earlier, I hate computers. They're awful. Yeah, so my wife is a human resources director, and then we have some evenings where I'm just like, computers are terrible. I hate computers. And she's like, people are terrible. I hate people.

Like, swap? Okay. Oh, no. You know what? You can keep them. I'll take the tech. You can keep the people. It turns out if you're trying to make big software you just have to do both so one thing we didn't get super into was the um

the stuff with you on the gnome board um what does the gnome board actually do um What does the Canone Board actually do? So. I'm just like, oh, no. Rumbled. No. There's some obvious legal stuff. So the directors, so every nonprofit is a corporation.

The directors of the corporation are responsible for making sure that the resources are being used appropriately we talk about fiduciary duty stuff like looking at the budget deciding how much money we've got how much money we're going to spend and i think more more specifically it's to do with the organization strategy and priorities so there's a team um we're in the process of recruiting for a new executive director. So we re-advertised the role actually last week. So if anyone knows any non-profit leading people who understand open source, please do get in touch. But there's a team of five or six staff members who work for the foundation um with you know different responsibilities and we the board should be kind of agreeing with them basically what the foundation should focus on

the best ways for us to support the community the best ways for us to get the message of um gnome and free software and the open source desktop and get that out into the world and so it's it's kind of holding holding the organization to account that we are doing good things with the money and the resources in a sense entrusted to us by the community and the donors that support the foundation um and making sure that there's a reasonable plan that translates from you know dollars in the bank accounts to stuff that we do and why we think that's the right thing to do to promote the mission and to support the project

and the organization. Yeah, okay. Makes sense. I have things on here. I thought I was going to be able to lead that into something. I didn't think this through too far. I mean, day-to to day, when I look at Germstaffer, it's not to do with recruiting ED.

I'm initially pushing my pet project, which is to stand up FatHub into a slightly separate organization. Okay. Partly because we're working on adding um donations and payments into that hub yes which changes some of the you know we need to have a stripe account we need to collect sales taxes and you know there's some difference like legal responsibilities and liabilities and there is a small very small but there is a small chance that we would sell lots and lots of proprietary software on FlatHub. And it would end up doing something that was no longer consistent with the

mission of the GNOME Foundation. And so having it in a separate legal structure, so we're thinking FlatHub NLC basically, allows us to basically protect the GNOME Foundation by moving FlatHub into a separate corporation or selling it or donating it to the Linux Foundation or whatever. It gives us options to kind of separate what FlatHub is doing from what GNOME is doing. So in a sense, GNOME is incubating FlatHub and we can provide some of our resources and help infrastructure, sysadmin and legal structure. But FlatHub has enough autonomy that we can have our own kind of governance process. So we're looking at having like a little advisory board or sort of governing board that brings

together Gnome and KDE to choose and sort of Flathub representatives so that we can sort of choose how Flathub spends its money and what we prioritise, rather than it being just a Gnome thing or just a KDE thing. But it's actually a separate organisation that has its own kind of governance and transparency. Hopefully not elections. Try and figure out something that's a little easier, but at least so that you know who's making the decisions

and you can see what they thought about and why. And then once we've got that legal structure, it gives us, you know, we can set up the agreements to upload stuff to Fat Herb and what that means in terms of collecting payments and if you need to demand a refund or whatever. Like all of those things we can set up, you know we can set up the agreements to upload stuff to that hub and what that means in terms of collecting payments and if you need to demand a refund or whatever like all of those things we can set up you know to and from this separate legal entity um so it creates a bit of kind of back office work but it sets up it protects the ground foundation from doing something that's

a bit different to what we're currently doing i mean ultimately we're doing it because the likely case and what we hope will happen is this improves the quality of software and availability of software for free open source desktop so we think it's a good thing having more apps and that's a that's not a globe thing that's a whole kind of linux community thing um which is why we're working in partnership with kd to set it up um and it's a wait now my sentence is just completely gone it's not just tech problems today it's also out of brain error yeah so being able to set it up as a partnership and being able to promote those things we think will be beneficial for you know users and developers and also just trying to open some new economic models.

So the Glenn Foundation has done a lot of work on diversity and inclusion and trying to remove barriers of entry to our community. And just this idea of being able to kind of work on free and open source software and have it kind of help to pay your bills. So I'm not saying it's the answer and it's going to work straight away, but at least being able to say we're opening the door to more people if there are ways for people to charge money or get donations or we're giving the options of economic models. Exactly which will be the winning option or what people choose or how developers will use it. winning option or what people choose or how developers will use it um we're trying to not sort of tell people exactly how they should use it we're trying to give controls out so that people could kind of run their own experiments and kind of open up that conversation with their users and

with the community and that's yeah trying to not make money a sort of bad word and you know actually say more people can do this and more people can develop software that solves their problems if we give them options to and also could have put a roof over their head or pay the board plan bill yeah monosyllation of foss is one of these things that's it's been a challenge since the existence of foss basically like there there are a couple of these companies that make it work but most of them that make it work either do paid development work for and have some like you know corporate contracts they have like they offer um support like red hat um actually those are the two major models that work when we're talking on the smaller scale

like we're talking like donations and nowadays you know you have recurring donation platforms like patreon coffee things like that uh github sponsors but actually having a model that is sustainable is it's difficult because as much as a lot of people in the the free software space like to say you know it's free as in freedom not free in beer. There is a lot of people that also do treat it as free as in beer and they don't want to give any money at all. It's, you know, this is just out there. I'm going to have it. So it's definitely a challenge that needs to be dealt with in some way.

I'm not the person to deal with, that's for sure. But... Yeah, I mean but no software is free right someone pays for it in their time and in in most places that means that you know that there's a cost to their time or there's a lost opportunity cost to their time if we're talking in sort of cold-hearted capitalist terms um but um there are lots of people who find free and open source software exactly because it's free because that's the option that gives them you know the applications they need without having to pirate something or use kind of crack copies or whatever it allows them to engage with technology in a more

meaningful way and so the the trade-off is you know there are people for who could easily pay for software and do. I don't want to make something clear. Like there's definitely people out there that don't have the resource to pay for software. And that's, I think it's great those people actually have access to software, but, you know. And that's really important to preserve.

Absolutely. Whilst over here, there are some people who have an Azente surplus. They could pay for the software and how do you kind of find a model that allows us to reallocate this surplus to the people who make the software without making the software inaccessible to the people yeah yeah and they really need that software to be uh to be free for them and so it's it's kind of balancing those two things and i think um the it's tricky right

because in a sense free software as opposed to open source is intentionally it's counterculture it's individualistic it's designed to preserve individual freedoms against the big company that will come and steal your stuff and make it proprietary and run it on a web service and steal your data and then rent it back to you and train your air models on it and all of these hypothetical things um definitely don't happen never never happen um so the i think you know one hope i have is that that the girl foundation and the fact that we are a kind of well-established um non-profit a kind of trusted community no everyone knows gnome within the freedom door space and so we say okay this is someone I can trust to take a position on this and actually talk about moving money around. But actually, we'll kind of give them some space to

figure this out and to be a trusted actor in the space, rather than a corporation trying to like, give me all the apps and give me all the money. And that could be but yeah um so just being just being a neutral actor in this space and saying well we're doing this because we think it's beneficial for these people who are excluded on financial terms and we think it's beneficial for our kind of shared values and mission to make cool software and get get it out to more people and you know do that with freedom and trust built in. Yeah. The, um... Think before you're trying to speak. The whole app store thing that Flatpak could possibly... Like, here we go.

Okay, we're going to think of this one through. I've got a lot of things going through my head right now. I have seen some people concerned over, you know, what Flatpak could possibly become if it wasn't this, like, neutral entity where, you know, it could become, like, a lockdown app store, like, you know, the Microsoft store, like, you know, app stores that are out there. I think that's definitely a concern that people

will have with something like this existing, especially as you are bringing in these monetization systems. Obviously, the goal is to be open about everything that's happening. The goal is to be this platform for free and open source software and also proprietary software to be distributed out to the users. But that idea of neutrality, I think it's very important to have as that key focus because, as I'm saying, we're on a loop now.

Yeah. So what if flat have turned evil? Yeah, there we go. Thank you. It's one in the morning right now. My brain's on hold. Well, I mean, that's part of the point of establishing it as a nonprofit. Exactly.

So you know who's behind it and in a sense why they're making the decisions they're making. I mean, there's too much about FlatHub now, which is like, Rob thinks we should go this way. Right. And that's not a sustainable model. That's a good way to start things. It's not a good way to run things. And, you know, there's too much kind of dead reckoning. And, you know, that's something we started doing

at the Linux App Summit this year as focus groups and actually having a conversation with app developers and understanding, you know, what was important to them. But ultimately, if Fl flat hub turns evil then flat hub will fail because it's a matter of maintaining the trust between the developers and the users and well the concern is that it gets it's this platform that becomes really big and then it turns evil like not not it it just becomes evil now that that's sort of where obviously it's like it's a's like... I read too many

weird takes from people on the internet. You have better things to do with your time. My entire business is looking at dumb people saying dumb things about Linux. I'm sure you've got better things to do than worry about that stuff. I don't know.

I maybe made this poor joke at some point when I was standing for board election. But if I was part of some evil Microsoft or Google conspiracy to take over the Linux desktop, you would think after 20 years, they would have called it in. Or if it is a conspiracy, it's a really poorly run one. Maybe that's what they want you to think well yeah I mean the tech didn't arrive yet

I don't know I think that's why the transparency and the trust is the important thing because I don't think our community that's based on shared values and what we're trying to achieve doesn't work if there isn't trust right um and so that's what we're trying to kind of build and communicate and the reason that i'm here talking about this is so that people understand well you know who is

rob and why is he doing these things with fat herb and should i you know care about fat herb or should i trust it or what can i rely on fat herb to do and i'm just being clear about the decisions we're making and why we're making them and in a sense if there are things we don't know or things that we might screw up then that's a shared journey we can go on together you know we we heard really clearly in the uh this app summit that there were there were a few people in the room who said i i will never require a payment to download my application um because i do not want to make it not accessible to people. And also because it's free and open source,

if I don't give people the best version from FatHub, they'll get a worse version from somewhere else. They'll get the version from the distro from 18 months ago or whatever. So I actually want people to get this version. And they were much more interested in being able to, this kind of patreon this supporter model and people you know um kind of donating not buying the software

that had already been written but kind of supporting and contributing to the next version and maybe you know helping to prioritize features or just being part of a community that kind of buys into this software and what it lets you do and so we had that kind of pretty clear feedback um so uh i don't know what we would make evil in that and if we if we jack our fees up somebody will make not fat hub right you know that we we don't have these kind of platform surfaces to kind of play people off against each other in the same way that's um uh that you know social networks and search engines and advertisers have. Corey Doctro has these kind of blog posts and things he's talking about in shitification. And this is basically like once you have all the users, then you can basically squeeze down the advertisers.

And once you have all the advertisers, then you can squeeze down the users. And basically the experience gets worse and worse and worse for everyone in the ecosystem because the switching costs are really high. And literally all the software we've produced for Flathub and all of the infrastructure, even the like OpenShift recipes, like if you have your own Kubernetes cluster, you can stand up your own Flathub by copying and pasting what we have. And then you have, you know, like, you know, not flat hub. Round hub? I don't know. So we don't have surfaces that we are, like, you know, grabbing and then playing people off against each other. Like, the preconditions for us to be evil and become evil and remain evil are, like, just not there. Right. There's no incentive for us to do

that and there's no way that we could succeed in doing that over time because people would just stop using it and yeah now one legitimate concern with monetization is okay so we're talking about like say the gimp developers they have the flat pack and they're like we want to monetize the flat pack that's a pretty simple idea like you're like, we want to monetize the Flatpak. That's a pretty simple idea. Like, you know, they should be able to monetize the Flatpak. But where it gets weird is when you have, you know, not really like a fork of a project, it's just a third-party package of Flatpak. Do you have that person able to monetize that Flatpak? Like, even though they're not the developers of the project, but they are the developers of the flat pack should they need to have some sort of relationship with the project like how how does this sort of need like how do you think this should be handled

yeah that that's that's one of the reasons why we've we developed and launched the verification features that have just come online um a couple of months ago when we kind of switched over the the internet's worst kept secret the beta beta Flathub site, now actually on flathub.org. And that verification process will be a requirement before you can access any donations or payment features. And so whoever's logged into the kind of Flathub developer console needs to then show that they are also a developer involved in this software. And that's done from a sort of computer level,

that's done with domain names and GitHub or GitLab accounts. GitLab, GitLab, GitHub and GitLab accounts. And it's not even 1 a.m. for me. It's just Tuesday. There's also layers on top of that obviously like what name the application has what icon it has and so i we are working on as a prerequisite to turning on any kind of payment or donation features we will also if you change the description or the the the icon right of the application it will go into a review queue. And at least initially, a human being will look at it and say, yes, this is still the GIMP or like, actually, no, this is not the GIMP.

And just do some eyeballing and make sure that we're making representations to the users about, yes, this is verified and this is actually who you think it is. Because, yeah, those problems are pretty hard to solve otherwise. There isn't really another kind of empirically true fact other than like someone who made this software and uploaded it to the platform. Obviously that software is going to be a patchwork quilt that could be a fork with this other one.

Minecraft launchers seem to have like a whole subsection that they kind of fork pretty much constantly my son has, you know, we had to move from PolyMC, that was like Unhealed, now we're on PrismMC yeah, that story

right? I just caught up with that but I realised I only joined my discord a while ago that was fun um he doesn't like me emails hmm yeah he definitely doesn't like me um because i there wasn't much coverage like outside of the linux space but i i was like one of the Linux space, but I was one of the channels that covered it in the FOSS context, and I went deep. I dug

up everything that was going on with this project. Yeah, man, that was a fun time. I miss that. I miss that experience. Yeah, we still get emails like, why are you shipping this? It's terrible. You should redirect people.

They're like, he controls the domain. There's not much we can do about that. So it is what it is. We can't get in play on everything unless it's a DMCA tech request or someone's copyright is being violated. And it's unfortunate because it kind of underlines the need to worry about some of these more kind of tedious details, things like fiscal sponsorship, who owns your trademarks, how

they license, those kind of things. I mean, when projects get to a certain size, you know, for a number of developers and things, you need to think a little bit about those legal structures. But there's lots of organizations to choose from. I mean, with my GNOME hat on, I've been looking at how can we sort of formalize. So we launched the GNOME Circle initiative two or three years back. And it's one of the things that I, one of the quite cool things that I think I've done on the GNOME board was kind of help to sort of create that and start it. And it's just been really cool to see.

I think it's 70, 80, like really neat applications coming in and forming this kind of, you know, wider community of developers and apps around the core desktop experience. And so we've been looking at how do we start to offer maybe things like donations, you know, handling and fiscal sponsorship stuff, basically like legal support or trademarks or those things so that we actually extend like a bigger umbrella out to those projects. But there's lots of great nonprofits in the space that will do that um as well so you can use open perspective if you just want a website and like take some donations and things but um there are there are others that will go out and enforce the gpl on your behalf like conservancy and folks

in between um i guess pi he'll sort of keep a bank account and reimburse expenses and things there's lots of different options that are available to set up the legal structure to go with your projects. And I kind of have a secret. Well, it won't be secret anymore. But my hope is that we can integrate between FlatHub and Open Collective. OK. And because one of the problems that we have with FlatHub using Stripe is there's basically a list of countries that Stripe likes. Yeah.

And if you're in one of the countries that Stripe likes, then you can set up your account and you can receive payments and yada yada. And the EU is there, UK, US, Australia, whatever. It's just not very representative of the global South. And even as a US non-profit, and that's more endless, even just wire transferring to Brazil or India or something creates creates and that's that's a lot of people right that's uh it creates really big uh barriers just for um kind of compliance and taxes and treasury controls and all of this stuff um and actually being able

to hold the money on behalf of the project and then they can reimburse expenses or buy you know equipment or reimburse airfares or whatever um that obviously opens it up and makes it a lot more flexible for how people can actually make use of these donations and payment stuff even if you just forget about like the whole worldwide thing just the different states in the u.s that all handle taxes completely differently yes yeah yeah stripe handles that very well you can tick a button and stripe you put your tax codes in and if it's business software or entertainment software if it's downloaded or if it's on a disc or whatever you tell stripe all of that there's actually a bit

in the flathead web app that basically tries to infer tax codes from desktop file descriptions like this is a game we'll give it this tax code uh is a game, we'll give it this tax code. This is productivity software, we'll give it this tax code. And then so that we can actually put the right tax codes in and then it will collect the right sales tax and then it will generate the right- And you don't get a letter from the US tax collection agency.

Yes. I mean, you normally get 50k or 100k that you can do before having to do any sales tax filings. So we can keep an eye on it and see. If it goes really popular in Nevada, we might have to file. Actually, I think Nevada has no sales tax, but anyway. Yay. Stripe does that for you. Yeah, yeah. So the open collective thing, you're going to talk about the potential partnership between that.

Yeah, so there is an API. So you can, as a, it might just be more work to do on the GitHub side, but I kind of quite like a way, if you're a developer and you can kind of sign up and say, look, I've already got this open collective, then kind of attach the app and the open collective together so the donations go into your collective.

Ah, yeah. Because then if you're in, one of the Gnome Ventures is in Paraguay and it's really hard to send money to Paraguay and he's like, I'd like to collect donations for Flatsil. This is Martin. And he could set up an open collective

and he could collect donations with VATSEAL and they would stay in the US. They'd be collected from wherever they are and they'd go to the US and then he could reimburse expenses or submit invoices or whatever best supports the project. So I think it just opens up economic opportunities for people to kind of handle their

product expenses and donations in a better way. Yeah, like, also just, it just brings everything together. Like, the problem with having, like, this is a problem that I have with managing my stuff. I have my YouTube AdSense, I have my Patreon,

I have my LiberoPay, I have this, I have that, and dealing with all of these different things is kind of a nightmare. If you can have everything going to one thing, that is so much easier to manage. Yeah. I have this idea, I mean, I thought about some Libera Pay as well. If you've got supported content in Flathub, could we can we you know if you've got supporter content in flat hub how do we kind of passport from another system to let you access the supporter content that you put in flat hub so can you like pair your flat hub account with your like libero account so that you know if someone

registers their app and we recognize that they're a contributor there and they let you access the supporter stuff but it's like yeah i mean that would be cool it's on it's on the list i wouldn't that they're a contributor there and they let you access the support staff in FlatHub. That would be cool. It's on the list. I wouldn't say it's near the top of the list, but it would be an interesting thing to look at. Or could we add that as a payment option and then we'd integrate in both directions? Because I think that's important. I think

FlatHub, knowing which applications that you use, that could be cool. You could just download and install all the FlatPaks that you want on a new system um and then knowing which ones you're a supporter of whether you support them on flat herb or you support them on a different platform i think that's kind of a nice goal to look towards it's it's a ways of off yeah we just need to to you know take the steps that we have for this year and maybe that's next year, right, Matt? First goal is getting the monetization stuff set up and ready for everyone to use.

Then you can go from there. Yeah. I think the next thing we'll do is the direct uploads because that actually, we don't have to wait for the legal setup and some of the kind of paperwork stuff and lawyers um we um that's more of a moderation problem then you got to make sure that things are being dealt with if they are you know shouldn't be on flat hub they're maybe not malware but like just like other things that you know you need to deal with yeah there's some automated checking and things that the build bot currently

does so um folks who upload directly which i think is kde the free desktop sdk um obs and firefox okay sure they basically bypass a bunch of checking that the build bot does. We're moving those checks into the repo manager so if you get a token to upload, then that kind of automated checking and potentially manual checking still happens for everything that goes in. We kind of want to set up some transparency in terms of where stuff is built. So, you know, not just having a binary off your laptop and then uploading it, but actually like, okay, can you please show us where this was built? Is it a CI flow? Is it a GitHub action? Can you please link to where this was

built so people can go and see the logs and they understand basically what's the supply chain? Obviously not from like a hundred percent guaranteed thing, but at least being able to show some transparency of where things are being uploaded from. So we don't want to lose that because at least on Flathub currently, you know that the source is in GitHub, it goes through the build part and then it goes into Flathub. So you can like find where it came from and what was built. So we kind of don't want to lose that property when we sort of open up to different build systems. Equally that property when we open up to different build systems. Equally, we have to open up to different build systems because it's pretty much impossible to make Electron or Rust or Python. You just get into a big argument between one build system and the Fatback build system. You can do it, but the number of people who can and want to do it is five.

number of people who can and want to do it is like five with the transparency of the bills doesn't that get a bit weird when we're talking about proprietary applications being packaged oh yeah all bets are off like proprietary application binary from space you either trust them or you don't i'm mostly thinking about not sort of turning the free and open source applications into that. So it's like, oh, it's just a binary from someone's laptop and apparently this is the source code to it, but I have no way of knowing. Right, right. So we would still like to lean Flathub towards things being built in public infrastructure that we can actually then see the build process that was used.

So some of that is a little bit of social pressure of like, okay, please link to the build logs and that will link to them. And there'll be a metadata field in the upstream so it can go into KD Discover or Chrome software, actually, maybe not build logs, but at least that it's recorded with each build and encourage, give people, this is the GitHub action. Copy, paste this. That's a great way to build an Electron app. And this is how you should build it and upload it.

And maybe try and upstream some of those things so that it becomes really easy to have a recommended flow of how to get, you know, a Rust app into FatHub. And if you follow these steps, people can see that you're following these steps and they can see it running on a kind of, you know, inspectable system. Right, right. Huh. So right now, what is the process of actually getting an app onto Flathub? Because I have no idea about that. So it's basically a GitHub pull request. So you can make a branch and send it into the Flathub.

There's a Flathub slash Flathub repository. And your Flatpak manifest and kind of any metadata and stuff gets reviewed by a team of basically two or three people who are absolute heroes. Thank you, Hubert. Thank you, Bart. Basically just keeping everything flow. flow um and uh just checking that the the permissions are saying the description is not terrible um the you know icons and screenshots and some of those things so some some of it is a little bit technical checking um and just following best practices and how the flatback is put together and some of it is then kind of metadata and descriptions and making sure that it will show

up properly and it will work properly in the web app and the different software centers um so after a bit of checking then basically your bots will take that branch and just make it into a new repository right and so each uh each flat pack um the ones that aren't directly uploaded has its own repository on github and then you can just see um the kind of commit history and there's a build bot that will pick up as if you each commit or pull request it it will do a build, either a production build or a scratch build, and you can test it and that will tell the build bot to publish it. Simple enough. One last thing I didn't want to get

into that we didn't touch on is the one other hat you wear the newest hat the Open UK stuff oh yeah actually I'm on sabbatical at the moment from Open UK just for the six month

period, there's a lot of family stuff going on towards the end of last year so it's just one thing too many yeah I mean family stuff going on towards the end of last year. So I was just like, this is one thing too many. Open UK was really small and not doing a lot three or four years ago. And Amanda Brock joined us as the CEO and took a real leap into it with both feet and said, OK, well, what can we do to actually bring together? There's a lot of open source developments and technology and open hardware and open data. And there's a lot of open source and developments and technology and open hardware

and open data. And there's a lot of stuff and innovation happening in the UK. But there isn't really one place that people can go to kind of come up with a strategy, come up with one voice, come up with, you know, what should we tell the government about open data? Or what should we do to try and improve the economic opportunities or stem the losses from Brexit or you know these kind of things like what does it mean to set a good digital strategy so just creating those forums to bring together some people looking at policy people looking at education like can people learn about open source in school or can they you know study open source in college

um uh what else is going on there's there so there's definitely a very active policy group there's an active education group which includes some museum stuff um and a big theme over the past year or two has been sustainability as well so kind of trying to align um your at least kind of advocate for why open source and open hardware designs and open data and energy data and things can actually help move the economy to a more sustainable place in terms of kind of climate impacts and the quality of access and those things. So we're using Open UK as a kind of sort of focal point of working groups to kind of advocate for those things.

Things like blueprints for a greener data center. You are a lot of stuff going on there. And I'm not doing that. There is a very great board and team and many, many volunteers and organizations involved in the Open UK. I was for a while the chair and I handed that over just because there were as you say slightly chimney hats right yeah yeah yeah keep keep keeping an eye on that i'm hoping to get back into it but um we'll see how the year unfolds i have four kids yeah we definitely hear them in the background uh a couple of times throughout the episode yeah they didn't come in. No, that's good.

No, I think that's really cool. So they weren't, you said they weren't really doing too much before, like a couple of years back. Yeah, it's been just a really impressive kind of reboot and just really bringing it together. There's some reports going on. Actually, if you're involved in an organization that uses open source there is a reports um questionnaire survey thing that is not a very coherent sentence um there's been there's been annual reports that open uk has been doing yes um about um state of open and uh the uk's part in the kind of open source landscape and there's huge like disproportionate percentage

like per capita or per head or whatever open source innovation and development data in the uk and just looking at you know what's the size of that you know community and that kind of corporate space and what does it mean for the economy i jinxed it ah good good i was waiting for something to happen i'm surprised it took that long I was waiting for something to happen I'm surprised it took that long me too we have our nanny here my wife works full time as well

it is legit hard to look after four kids they age between three and nine I am not surprised whatsoever I don't have kids myself but I do have some family members that have kids around that same age

and yeah it seems like they have a lot of fun dealing with that anyway so UK viewers and listeners if you use open source software in the UK, OpenUK would be very appreciative if you could have a look for our State of Open 2023 survey and input into our third report about open source, open hardware, open data in the UK.

Yes, go and do that. So with all the different hats you wear, wear all the management sort of stuff you do nowadays how how often do you manage to like do you even try to get involved like in the weeds of getting involved in projects now is it more just like you know doing your your higher level stuff like i i looked at your your github and you know i saw a bit there's a bit of there's a bit of stuff in there yeah um i mean i i definitely remember a point when i sort of i i try and still use the nooks and play around with stuff enough that i understand what's going on technically so i don't just sort of turn up and just spout complete nonsense right you don't want to be one of those managers that are like talking about like yeah um

like it looks great on powerpoint and then like you know it doesn't actually work in reality so i i try and kind of maintain enough detail and and you know that's either hands-on or you know by working on stuff with the team or talking through and you know so i i still read code i still look at what's going on i still run my own mail server you know i still use wechat i mostly got my head around docker kubernetes is still like a deep mystery um so i i try and do enough technical stuff that i know what's going on in enough detail that i can kind of sit across these different worlds of understanding what's going on sort of technically in the desktop and just in its space generally and bring that into the different kind of conversations I'm having. So if I'm looking at strategy or kind of product investment

or those things, making those decisions for Endless, then I kind of know what's likely to work or not work or what is a useful direction to consider right um and so i i sort of have that kind of cto ish input into what we're doing in endless um and for gnome obviously that the same thing is true i i don't think the the gnome foundation and its directors would could have maintained the engagement trust respect of your gnome community if we were just better off talking to some whatever. Yeah, I'm probably free to talk about stuff. And it didn't actually come down to, well, how does this affect the project and how does this help the Grom desktop and these things?

So being able to sort of stay grounded in the technology, but then actually look at how to apply that and to kind of improve the impact and the reach of what we're doing. So yeah, I try and keep it real a little bit, but I don't have a lot of spare time for that. So it's just sort of here and there from time to time. It's always weird when you see those people that are in the management positions and they're talking about something as if it's still like 10, 15 years ago

and they fully understand everything happening back then. But like the current state it's like pipe wire? What's a pipe wire? What does that mean? Weyland? What is Weyland? I've never heard of this thing Yeah Try and keep track of that

I think that's ultimately for the best because yeah as you were saying I think that's ultimately for the best because yeah as you were saying I think yeah being in that CTO-ish position I think is probably how do I say this I don't know what I'm trying to say actually

I don't know where I was going with that yeah I'm not sure what I was trying to say there um we will it's a double-edged sword because i i think if if i do too much technology strategy or whatever then i um i'm taking away the opportunity of my vp of engineering and those guys to to kind of have input into our products and technology strategy so i i may sometimes i might weigh in on something and sometimes i might kind of stay quiet and see right you know where what what ideas come out i i often try and speak last in a meeting so that i'm not sort of setting

a direction like well my boss's boss said this so i'm going to stay quiet or whatever and you know it's it's important to just to include and allow other people to to get involved with with their different approaches so obviously you like the position you're in considering how many management positions you've taken on like clearly you you have a thing for management um but do you miss those days where you were like working in the weeds get it getting there in the repos doing all the the hard work yourself um yes and no it's a little bit i i really enjoy working together with people and there is at some very very base level, actually implementing a piece of technology just by itself is or can be quite a lonely thing.

And in some ways it can also just be like, you know, person versus computer. Like you have the idea, you solve the problem, and then the rest of it is just arguing with computer to make it happen. And it's actually the solving the problem and working with others to solve bigger and um and it's actually the solving the problem and working with others to solve bigger and more complicated problems and stuff that's the bit that i enjoy the

most right and there are other people who are much better at the other parts of it the arguing with computer and actually making the software match the goal and the the ideas and so i'm i could have quite happy to live kind of vicariously and sort of joining the team, but not have to be the one that's writing the code. But I'm bringing different perspectives and different ideas. But I'm still part of that creative process and that problem solving and kind of direction setting and things. That's where, that's what I enjoy the most. I enjoy the, you know, the boss and the conferences and those things. And, you know, our kind of internal planning meetings or off sites or whatever we do at Embers.

those things and our kind of internal planning meetings or offsites or whatever we do at Embers. Those things I really enjoy because I actually like working with people to solve, figure stuff out and come up with ideas and kind of work together. So I miss it a little bit and that's why I will kind of doodle around, but equally I'm not as good at coding and operating a computer as I used to be. So I also just don't have a lot of time and patience to sit and debug some kind of gnarly stuff.

I'm just like, this doesn't work. I'm just going to have to set it aside and ask someone for help or just do something different. My son keeps on my toes because he's trying to load all these Minecraft mods and set up a server with a plugin that lets the Nintendo Switch connect to it as well. I get a little queue of sysadmin tasks coming from my nine-year-old. You have your mail server as well, so that keeps you on your toes. That works really well.

Mail Cow. One stock recomposed and a mail server pops out. It's amazing. It was basically like the mail server I was running, except completely automated. So I manually got Sogo and Postfix and Dovecart and blah, blah, blah, blah, blah. And then I looked around like I need something that just does this for me.

And Mailcow basically had the mail system I was already running, but just like in a Docker container and we haven't sorted out the web UI. Even I can get my head around that. So yeah, it was good. Mail servers are something I just... I'm good. I've got other things I'll spend my time on.

But hey, if it's that easy, I might have to check it out. Yeah, MailCow, I recommend it. I've certainly considered it but like every time i'd looked at like doing the mail service stuff like all of the stuff i could find was like you know do this and this and this and this and this and keep going going and going copy paste this for like no reason yeah to be fair it was a couple of years back so it was it probably would have been before like docker and stuff was really taking off and really becoming this like massive thing.

Yeah, so I would imagine it's definitely a lot easier, right? Definitely a lot easier now. Yeah. Yeah, there are some that sort of take a computer and then just mangle the configuration files and install everything. And I would have said it was a bit neater. So just having a mail system as a bunch of containers seems, made sense to me. It did all the spam and DKIM and SPF and whatever alphabet soup is necessary to actually receive an email or have Hotmail not spam filter you immediately.

have hotmail not you know spam filter you immediately uh immediately yeah well i'm taking up uh a lot of your time um so i think we should probably start to wrap this up it's been like almost like we've been talking for about two hours but it's been like hour 40 we're recording this geez um a little bit of editing oh i should let you get to bed as well no it's all good what I usually do with the podcast assuming there's nothing you want me to cut out I upload the entire thing as a block and then I cut out clips and upload those as a separate thing

so I'm sure I'll find some fun things that we can have as dedicated topics there yeah so some fun things that we can have as dedicated topics there. Yeah. So let people know where they can find you online, where if there's anything that you particularly want to, like, direct people towards. I know you mentioned the State of Open survey before, so that again.

Go check that out if you're in the UK. Yeah, so my blog is where I post kind of the most interesting stuff that we're doing over at FatHerb. So that's rabq.net or you can find it on PlanetGram. And nsos.org is the NSOS website. But in a sense, people watching your show are probably not necessarily the ones that need what Ennis is doing but

go and donate to your favourite open source project and say thank you to the developers, they really appreciate it Absolutely That all you wanted to mention? Yeah, I think so Okay, sweet As for me, the main channel is

Brodie Robinson, I do Linux videos there six-ish days a week. Sometimes not Linux videos. Sometimes I'm talking about Instagram doing their activity pub thing where it connects into the Fediverse and all that fancy stuff. That's cool. Yeah. If they actually do it, it's from leaked marketing slides,

so it could all just be nonsense. But I'm going to talk about it as if it's completely true because that makes it more fun. I've got the gaming channel where i am playing through yakuza zero and where what when is it right now it's not towards the end of may uh i'll be playing final fantasy 16 very soon probably in like a week from when this comes out and if you're listening to the audio version of this the video version is available on youtube at tech over t and if you're watching the video version you can find the podcast on any podcast platform there is an rss feed stick in your favorite app i like antenna pod it's pretty good and yeah um i'll give you the final word what do you want to say how do you want to end the show

uh i'd always put people in the spot. Freedom to open source software is awesome. We have the opportunity to change the world and change the way that people interact with technology. So go out, tell your friends, support the creators and authors and do what you can to be a kind and respectful advocate. We really appreciate everyone's support.

Awesome. Cool. Sweet. Thank you very much. It's been great. I'll let you get to sleep.