Algora | Algora TV

Co-founder of Tauri, Daniel Thompson-Yvetot | Open source interviews Ep37

Algora

Watching now

00:00 Hey everyone, welcome to the Open Source Founder Podcast.

00:15 Joining us today is Daniel, co-creator of Tori and founder of Crab Nebula.

00:19 For those of you that don't know, Tori helps you build cross-platform desktop applications

00:24 and mobile applications.

00:27 Recently, a few days ago, you released 2.0, Tori 2.0 Beta, and you will give us a demo, an exclusive one.

00:34 So thank you very much.

00:35 We're super excited.

00:36 And after the demo, we can discuss also about Tori and Crab Nebula.

00:40 Thanks for joining, Daniel.

00:42 It's an absolute pleasure to be here.

00:44 Should I just start the screen share?

00:46 Yes, sir.

00:47 So what we see here on the bottom left is a classical desktop app in Tauri's dev mode,

00:55 which is basically kind of a preview of what you're going to have when you finally bundle the app.

01:01 when you finally bundle the app.

01:01 It's really useful when you're developing to get instant feedback.

01:04 It's really useful when you're developing to get instant feedback. And the beta has now stabilized Tauri's 2.0 branch

01:10 to the point where it's generally useful.

01:13 And one of the things you can do with Tauri that you couldn't do before

01:16 is mobile apps.

01:18 So on the right-hand side of the screen, you'll see an emulator.

01:23 This is an Android emulator, but iOS emulator is also available.

01:27 It's just too much going on to do all of them at the same time.

01:30 But this is kind of the neat thing.

01:33 If you're familiar with web dev,

01:36 you've got something called hot module reloading,

01:40 which means as you make a change in the developer environment here at the top, I'm using VS Code.

01:48 As you make a change there and save the file, so welcome to today's demo, and I save it.

01:58 What we get is an instant update of both the desktop app's preview version and the mobile app's preview version.

02:05 And when I first saw this, I was pretty impressed.

02:11 You know, okay, sure, you have an app that has Rust in the back end,

02:15 so I can send a message to that app on the desktop version.

02:20 So I'll just say my name is Daniel.

02:22 That's the anglicized version of my name. And I'll click greet.

02:26 And this message was sent to the Rust core behind all the Tauri apps.

02:32 And then replied back across our message boundary with the IPC protocol.

02:40 And if you notice, that didn't change the state of the mobile app.

02:45 And that's quite compelling for me because I can go to the mobile app and say, I'll just give myself my Germanized name, and I'll greet.

03:00 And the state is respectively the same.

03:08 I mean, it's respective to each instance that's running, right? We've got two different instances.

03:10 And, again, I can change the UI code with, let's just say, you know what, I'm going to get rid of the logos just because, you know, everybody knows what Tauri is.

03:26 And if you don't know what solid is, you've been living under a rock.

03:29 And I'm going to save that again.

03:31 And boop, they're gone.

03:33 And everybody is probably saying, Daniel, we know this stuff.

03:39 But what you might not recognize on a wide format screen is something that the mobile user is

03:48 going to see. So if we go to our button that says greet, and we just change it to something like

03:55 stay hello and save that, suddenly my layout changes in mobile. It doesn't really change in desktop, right? And what I find so interesting about this new opportunity for multi-platform, multi-device design is as a developer, you can work with the wireframes of your designer and all that stuff.

04:22 work with the wireframes of your designer and all that stuff.

04:25 But once you get into things, you're like, well, shoot,

04:29 we can't say say hello because it just looks like junk on mobile.

04:31 Why do you have those double? Come on. You know,

04:34 like I can't make it smaller because no one's going to be able to read it.

04:35 Right.

04:39 And so it's a great way to get really quick feedback in the development process.

04:41 And I would just like to say thank you to the,

04:44 the demo gods for not crashing my system.

04:49 So I'm going to just stop there and say thank you.

04:54 And, you know, back when we were proposing the idea to the working group,

05:02 the Tauri working group of introducing mobile, there

05:07 were some voices that said, it's a bad idea.

05:09 Don't do it.

05:10 If you're out there listening to this, you know who you are, and I still love you.

05:15 But what we discovered along the way, part of the joy of writing anything, whether it's

05:22 a novel or a code, is the discoveries you make.

05:27 And one of the discoveries that we made early is that the API design that we created works great on desktop, just like that button works great on desktop.

05:37 Say hello.

05:38 On mobile, it's different.

05:41 And the API needed to be rebuilt from the ground up. We wouldn't have known. It would have solidified. And I think that the revisions that were made to the way people interact with the core of Tauri is, for me, the most exciting part of it because it's like we're still involved in creating something valuable and we're learning

06:06 through the process right and i think that's something that that continuously encourages me

06:13 to stay involved in the project and brings people in i think that everybody who comes into the

06:20 working group uh brings in their own ideas.

06:29 And, you know, the doers end up deciding, which I think is great.

06:33 That's the right approach.

06:36 And I mean, this, what I just saw, this is a superpower and very unique.

06:41 I don't know anything else that I can do like this exact same thing and ship so fast.

06:45 So now with the beta, like a lot of things are in place.

06:46 People can just take a stab at it.

06:50 And there's a review for the stability of the release,

06:51 but it's good to go, right?

06:59 For us, beta, well, alpha, beta, and RC have specific meanings, right?

07:02 The alpha phase was quite long.

07:10 We had premises that we had to prove or disprove and ideas that needed implementation.

07:17 There's a number of security features that, I mean, you know, security is at our core.

07:23 It's one of our, you know, philosophical ground stones, you know, it's a really important thing.

07:23 And anything that touches security has to be well thought and implemented.

07:27 And alpha is that time for innovation. And, you know,

07:32 sometimes you just throw your idea out and start with something else,

07:34 you know, it just takes time.

07:36 And beta is the point,

07:39 the beta that we marked last weekend for everybody else,

07:43 that's the beginning of February, was, okay, we're happy with this design.

07:51 It works for us.

07:53 Now we have external auditors.

07:55 The company known as Radically Open Security is auditing the Tauri 2.0.

07:59 They audited the Tauri 1.0.

08:01 They've been doing it for a while.

08:03 Tauri 1.0. They've been doing it for a while. Tauri is mega complex. There's a ton of

08:06 libraries involved and they've been working on the low-level libraries while we finished the

08:11 high-level Tauri universe. And now they're able to investigate what we built for Tauri itself.

08:20 And by the way, to date, they've found no major vulnerabilities in any of the code that we've written.

08:25 It's not because we're friends.

08:28 It's because we had a great foundation to build on and an amazing group of people that have been involved.

08:36 And there was a comment today on Twitter, great news that this has been released, but it's been a disappointing dev experience as the docs have not been updated

08:45 and there's very little information in the blog

08:46 about what upgrades are required

08:48 to make the switch from alpha to beta.

08:49 Any thoughts?

08:50 And I guess the comment that I have is,

08:55 primarily, writing documentation

08:57 on top of a moving target like Tauri's alpha

08:59 is a fool's errand.

09:02 We've built tooling to help you migrate.

09:05 Tell remigrate is a command you can run

09:06 to migrate from 1.0 to 2.0.

09:09 Of course, there's rough edges

09:10 because everybody writes code differently,

09:12 but it will get the configuration right.

09:14 And now is the time for us

09:18 to really focus on the documentation.

09:21 And once the documentation is complete

09:23 and the auditors have come back and

09:25 we fixed all of the things that they found, small or big, then we mark it as a release candidate.

09:31 A release candidate means feature complete, fit for purpose. Perhaps there's something that

09:39 in all of these wild edge cases was missed. And it's a time for the community to say, yeah,

09:42 in all of these wild edge cases was missed.

09:45 And it's a time for the community to say, yeah, all right,

09:49 this RC is something that we can get behind, but it's still not stable.

09:52 And stable comes a short time after RC.

09:55 So, yeah, that's where we're at right now.

09:57 We're happy with the design.

10:01 And, of course, there's still things to iron out on the way to a stable release.

10:04 Thank you for sharing this timeline.

10:06 Well, this approach makes perfect sense.

10:10 And, you know, writing and updating the docs will be a community activity as well.

10:17 So now taking, you know, a step back, we would love to hear about your personal background all the way to co-creating Tori, if you'd like to.

10:20 If you'd like to.

10:22 Wow.

10:30 I mean, I moved to Germany in the year 2000 to study fine arts. I'd already received a bachelor of fine arts, but at the university, my work moved into media.

10:36 And my thesis project was a complicated object.

10:46 It was a real-time analog holographic projection system

10:53 with homemade video projectors.

10:55 This was back in 2004.

10:58 And the holograph was important, or hologram,

11:01 or the idea of one,

11:03 because I was presenting the notion that a digital identity

11:08 is a part of a human identity and that that digital identity was qualified enough to

11:15 defend the artist behind the artwork as it were. Needless to say, after that project, I sort of received the recognition in the community as I was active in as, you know, like that uncle who can bridge these genres of media, technology, culture, community.

11:40 And that's pretty much what I've been doing my entire career.

11:55 A lot of the time I spent in volunteerism, it's hard to know how many hours. I think it's something like about 20,000 hours of volunteer time in my lifetime and a lot of that has been in collaborative culture. And I think that

12:07 writing code is collaborative culture. I think that open source is a license,

12:14 but the community is a group of people who have aligned goals, more or less, and work together.

12:23 For me, it's the community that's inciting.

12:27 We call it like open source software.

12:32 I think of it more in the context of a digital commons that a community works on together.

12:35 And the licensing, that's all stuff that you have to do.

12:41 And community is something that you choose to do.

12:44 you have to do, and community is something that you choose to do.

12:52 And I think that that desire to be part of a community is at the root of why we do software the way we do.

13:01 If I didn't have friends in this towery community i couldn't have built it like

13:11 if it was just lucas sitting in a cafe in brazil alone he couldn't have done it like actually

13:21 without all the people who've contributed to, we couldn't have done it.

13:25 The whole notion of what a web view is was built by geniuses, giants, like HTML.

13:34 I mean, you can argue that JavaScript was a mistake, but it's still what we have.

13:39 And I feel like the conversation that we're having in these communities is a progression

13:46 built on what came before you know like um a car was built on the notion of a cart and uh

13:58 an airplane was was no maybe the airplane comes from us. I don't know.

14:10 But the community is where the strength and how this project has manifested.

14:13 That's where it all comes from.

14:14 But that's also where a lot of the challenges come from

14:16 in terms of the collective decision-making,

14:19 the governance, the sustainability.

14:21 And so how have you approached all this so far?

14:24 And what can we learn from it?

14:28 You know, one of the reasons why I think we sort of exploded in popularity had to do with

14:37 the point in time where we were doing this. Back in 2019, there was amazingly crazy divisiveness in the JavaScript community.

14:50 Like you were either on the React team, the Vue team, or the Angular team, or maybe one

14:57 of the other ones that were out there back then.

15:20 And what I feel we've been able to accomplish in the Talry community is bringing those or bridging that divide, bringing people together that wouldn't classically have anything to talk about, except at a W3C meeting about SVG standards or something.

15:25 Like, yeah, everybody has these same problems of how do I use the API?

15:26 Or I want to use this configuration. And we don't care where you're from.

15:28 Like, we don't care which stack you're using.

15:31 We want to support everybody.

15:33 And inside of the organization, we've attempted to do something similar.

15:38 From the beginning, we've worked hard to stay accountable and transparent and accessible.

15:48 And, you know, the way that we've made or worked toward our accessibility

15:55 is through our working group.

15:58 So way back when it was just a couple of us,

16:01 if it's three people in a video call,

16:05 what do you need a working group for?

16:08 It's like, good idea, bad idea.

16:10 Yeah, I'm going to do this.

16:11 You're going to do that.

16:12 All right, cool.

16:12 Let's go.

16:14 But as more and more people get involved

16:16 and the project becomes more and more complicated

16:18 and there's special activities involved,

16:22 you start to notice that people group together in little huddles.

16:27 And as more and more people joined, what do you do with a core team of 10 people, 11 people,

16:35 12 people? It stops making sense, right? So we developed an open and working group model.

16:42 Basically, the way it works is you show up somewhere in the community,

16:47 whether that's GitHub or Twitter or Discord, probably Discord,

16:51 and you start contributing to the community's knowledge base

16:57 or fixing a bug or filing a feature request, making the PR,

17:01 getting yourself involved in the community.

17:03 filing a feature request, making the PR, getting yourself involved in the community.

17:11 And generally speaking, people get recognized for that by the existing team and then recruited for the working group. It's like, hey, do you want to get closer to the team,

17:17 meet the team and work with the team? And if people accept and the team doesn't have any

17:23 vetoes, then they're brought into the working group and they get access to the private Discord channels for the team and they get access to the GitHub repos and they can branch away and make PRs and file issues, respond to other people's issues. And all of that's possible because of a principle that I like to call interrelatedness.

17:54 And it's an old technique from grassroots organization where you're faced with a dilemma.

18:02 You work with volunteers and you don't have any guarantee that they're going to stick around.

18:09 They take on a task, but you don of social pressure is exerted from within the individual

18:30 to unblock the others, right?

18:33 And it's a really simple technique.

18:35 It's called a pull request review.

18:38 Like, I can't merge.

18:42 I could abuse my authority and merge,

18:51 but I can't maintain my standing in the community and emerge my own pull requests.

18:54 And neither can anyone else, not even Lucas.

18:57 We have to have,

18:58 we have this hard requirement of getting code review.

19:02 And what that means is you're putting your team

19:06 in the position of looking at what you've done.

19:10 And that interrelatedness as a core principle

19:14 is what I believe to be the secret key

19:19 to success for open source.

19:21 As soon as you get into the benevolent dictator

19:24 for life model, it's over.

19:27 At some point, either the dictator recognizes that they're being a dictator

19:31 and decides to bow into the peer pressure of letting other people decide,

19:37 or they don't.

19:39 And when they don't, that's when projects just start to disappear

19:43 and lose relevance.

19:44 When they don't, that's when projects just start to disappear and lose relevance.

19:47 Well, it's super insightful.

19:55 And that interdependence is what also creates a lot of accountability for people and comes with autonomy as well. So that's a, you know, I want to say unique, powerful and successful model for you guys, for Tori.

20:04 and successful model for you guys, for Tori.

20:10 And shifting gears, it looked like you saw that there was some additional space to support the ongoing development of Tori, put resources behind it

20:15 and engineering bandwidth, as well as help organizations who are using Tori

20:20 to be successful through consulting services.

20:24 And so Crab nebula was created

20:27 tell us about this story so far i mean i have to backtrack a little bit because crab nebula didn't

20:33 just arise it wasn't just like one day we're like we're gonna monetize let's let's get rich um

20:40 actually it was the contrary one day we got contacted by a venture capitalist and we were like, we're not prepared for that.

20:50 As an open source group of friends, what happens if someone goes rogue and sells the idea?

20:57 How can we prevent that from happening?

20:59 How can we prevent financial bus factor?

21:02 financial bus factor.

21:09 So what we did is we became a program,

21:12 the TOWERY program within the Commons Conservancy is the official name of the project.

21:13 And it is a statute bound organization

21:19 within a Dutch nonprofit that owns or has,

21:24 holds the moral stewardship over the project.

21:27 There's an elected board of directors, of which there's seven,

21:32 and the working group is the entity entitled to qualify or disqualify people

21:42 from participating as representatives of them in the board.

21:47 And it brought another layer of organization governance to the project

21:57 that took a while to get right. I think at the beginning, we maybe tried to do too much organizational stuff with it.

22:06 But ultimately, the board is really there to fight for the trademark and to decide on how to spend the donations and to agree to partnerships and that kind of thing for which you need a legal entity or legal person.

22:27 And decentralization. Yeah, absolutely. thing for which you need a legal entity or legal person and the centralization activities yeah absolutely and and just to quickly note here to people that this may sound like confusing or arcane

22:32 or why are all these things needed like the scale of tori it's the it's the third largest

22:36 rust project on github right like it's what 70 75 000 stars right now so that these go hand-in-hand, I feel

22:46 like the scale of the community, the usage

22:48 with the requirement to have this kind of

22:49 group behind it, or that's how you approach it,

22:52 right? Sure, but I

22:54 mean, I think the flip side is

22:56 let's just pretend that

22:57 someone, myself or

22:59 Lucas, decided

23:01 as the co-creators of

23:03 Tauri to sell Tauri for 50 million to some rich company, exit with the FU money, and then that company could use its lawyers to change the licenses and the community that was grown to build this project would be cut out.

23:27 And we didn't want that.

23:28 And that was the guiding light for why we went community first.

23:35 Nevertheless, as time moves forward, you know, we've been doing this for almost five years

23:41 now.

23:42 As time moves forward, people burn out.

23:46 And people have burned out. And at a certain point, we were facing not exactly exodus,

23:52 but potentially losing critical contributors to Tauri.

24:00 And, you know, we thought about it.

24:12 And, you know, we thought about it, spent about, you know, a couple months thinking about it, talking to people.

24:25 And we ultimately created a company that served the purpose of fulfilling the needs of the community that the community was not in a position to do.

24:26 So I'll give you an example. A community based of volunteers cannot accept a contract for work based on leveraging the

24:35 goodwill of volunteers.

24:38 It might not be illegal, but it's immoral.

24:41 And you have no guarantees that the other party is going to do it.

24:45 So first of all, like, we got questions all the time.

24:49 Can you help us?

24:50 Can you do this for us?

24:51 And what ended up happening in a number of cases is people from the working group would take on a one-off contract,

24:59 which is okay, but it's definitely not sustainable,

25:01 especially because the work that needs to be done isn't on helping someone implement Tauri.

25:08 Sure, there's value in that.

25:10 But it's more about the one million lines of code in core. on every platform and building a middleware that is unified across them,

25:30 more or less, and solves the edge cases of which the Linux ecosystem has shattered into.

25:32 Like, that's the important stuff.

25:36 But there's other issues.

25:38 Like, we love GitHub.

25:41 GitHub is a for-profit company, and it provides services for open source projects,

25:46 which is great. But if you're running a proprietary project and you want to use GitHub releases,

25:53 now do you have to make your proprietary project public if you want to use GitHub releases?

25:58 How do you get metrics about those releases? How do you know that it's actually available? Right. And, you know, to that

26:08 end, we've spent a lot of time conceiving of and building a cloud for distribution of these

26:15 artifacts, not replicating GitHub. We're not doing CI. We expect you to do that yourself. But we now as Crab Nebula provide a service that you can use to distribute your

26:31 Linux, Windows, Mac binaries to everybody. Tauri is not stable 2.0 yet. So we're not

26:37 servicing mobile yet. That's going to come, of course. And, you know, we've recognized that there's also not enough time to

26:50 do R&D. Like, if you imagine the caliber of engineering that this group of people have in

27:01 them, and then you hit them Monday morning with 35 new issues over the weekend,

27:07 and you go to the Discord, and there's 120 problems that people are having,

27:11 and you're constantly playing catch-up, these thoroughbreds, they don't have time to spend a

27:20 couple of weeks on thinking about something else. And when you have the stability of a job

27:27 that says, yeah, you can work on open source, you can have new ideas. If you want to take a

27:32 couple of weeks, that's great. We want you to do that. I think there's one other compelling story

27:38 though, similar to this say hello button from the beginning demo,

27:51 when we started the company, we went to a lot of trade fairs in Europe,

27:56 in the U.S., and at one of them, I remember this so clearly.

28:00 In the afternoon after I'd talked at Rust Nation in London,

28:07 someone rushed the table, and he's like, guys, you got to help me.

28:14 I'm building a Dioxus app and I don't know how to run the Tauri updater with it because it's too tightly coupled to Tauri core.

28:17 And it took us a while to figure out how we wanted to make things better, but it also raised a really, really, really important lesson, I think, for me, for the rest of the team, and for open source in general.

28:36 And that is similar to the journey that an author has when writing a book.

28:41 You discover things along the way that you never anticipated.

28:44 author has when writing a book, you discover things along the way that you never anticipated.

28:51 We did not expect to solve a major ecosystem. And by ecosystem, I mean, everybody who's building apps. We did not expect to solve that problem with Tower. We were just trying to, you know,

28:57 make a more efficient harness for a desktop app. As it turns out, we've developed a service that would grow into a project we call

29:10 Packager at Crab Nebula that allows you to, at the moment, bundle up an Electron app and make

29:21 updates for it. You can do the same thing for Dioxus,

29:26 which, by the way, leverages some of the same ecosystem

29:29 that Tauri is built on.

29:31 Slint works.

29:33 And, you know, as more and more people visit the project,

29:37 we will discover new applications for this.

29:42 And I think the big lesson there is to

29:48 stop thinking so narrowly about you solving your own problems and try to find ways to solve them

30:00 for the greater ecosystem. And I'm not trying to do the Obama meme of giving myself a medal right now.

30:05 I think that it's really critical

30:09 to jump over your own shadow, you know,

30:11 and to say, hey, it's okay

30:16 that you're building an Electron app.

30:17 We don't hate you.

30:18 In fact, we care for you so much

30:20 that we're going to give you a system

30:21 that you can use right now with NPM

30:23 that's going to make your life easier and better.

30:28 Does Cargo Packager slot perfectly into your CI that you're going to use

30:32 to build apps and distribute them with Crab Nebula Cloud?

30:35 Oh, yeah, it's a side effect because, you know, we care about you.

30:37 We want you to make sure that you're getting the best out of your time.

30:41 And is it open source?

30:43 Absolutely.

30:43 MIT Apache 2.

30:48 You can take it, make it better.

30:57 Hopefully you contribute it back. But yeah, I think it's the right thing to do at the right time.

31:09 And I'm glad we did it. And where could this be, you know, if you let yourself potentially naively think about where this could be in some years,

31:11 along with everything else happening in the ecosystem or the other options developers have to bundle, design,

31:14 to distribute, to update their apps cross-platform on all devices and systems.

31:19 How do you picture this?

31:22 Well, I think a lot of things.

31:27 I can imagine a lot of ways that things could turn out.

31:33 I mean, don't quote me on this,

31:38 but when you start building your own web view,

31:42 you know, we're collaborating with people who are working on Servo.

31:45 And we did an experiment where we built a Servo-based web view for Rai.

31:53 I mean, at the point where, as a group, we're starting to take ownership within the larger Rust ecosystem for the user interface,

32:05 for the display, for the keyboard, for the storage.

32:10 Would you consider that at some point an operating system?

32:14 I might, but more importantly, two things.

32:20 When we launched Howery 1.0 stable,

32:23 it was the same week that Internet Explorer 11 and Adam Schell were deprecated, or their deprecation was announced.

32:32 You don't just deprecated as a project.

32:51 And are we going to dive headfirst into Apple Vision?

32:58 Thought about it.

33:00 Why not?

33:01 I mean, at the end of the day, it's just another screen with just another haptic interface. Is there something stopping us from doing it? Maybe the expense of the hardware, but at certain scales, it doesn't matter anymore. Are we going to start jumping into building Xbox games? Well, technically, Xbox does provide a WebView2 interface.

33:30 R&D at Crab Nebula? Yes. But I think there's a more important question. And that has to do with the Rust language itself. And it's something that took me a long time to learn. What a lot of

33:35 beginners in Rust struggle with is the compiler. The compiler is the thing that takes your Rust

33:40 code and turns it into something useful. And when you build broken code, well, it breaks down

33:49 and it tells you kind of what happened. And you don't understand what error codes are until you

33:53 figure out there's multiple different ways to represent strings and integers and all of that

33:59 stuff. But what TALRI does out of the gate, when you have all of the system dependencies installed, is it gives you something that works.

34:12 And by works, I mean it bundles.

34:18 You can package it up.

34:20 You can run it on your computer.

34:23 package it up. You can run it on your computer.

34:31 What that implies for the development process is understanding what done means.

34:39 Having a notion of being done, I mean, I remember the projects that I would work on based in the Node.js ecosystem and never felt like I would get done.

34:45 Like, I might catch most of the errors,

34:47 but then you have to install some service like Sentry

34:50 to make sure that you didn't miss some.

34:52 And that's why you use try-catch.

34:54 And then, like, and finally, and you never know what's going on.

34:58 You just kind of have to assume.

35:00 And I think the mental model of being done transferred over

35:04 to the idea of a framework like Tellery, well, what we've tried to do with the 2.0 is reintroduce a base set of modularity such that as new use cases, platforms, systems arise, that we can create adaptations that still work in that baseline system.

35:27 So in some contexts, when 2.0 stable lands, and don't hate on me, working group, for saying this,

35:35 from some perspectives, we're done. From other perspectives, we're absolutely not. We know that

35:42 the moment 2.0.0 lands, there's going to be a 2.0.1 because CI is hard and

35:47 Tauri is complex.

35:50 But 3.0, there are rumors of a 3.0 being thought about.

35:57 But a 4.0?

35:59 I don't know.

36:00 And what that suggests to me is that like some kinds of software, you can actually approach being done.

36:09 And what being done means is that you as an engineer, as a developer, have a chance to think about what's next.

36:15 You get to innovate again instead of dealing with issues that are things like, I filed this issue three months ago and no one responded

36:26 and I'm really angry about it, right?

36:28 Instead of like dealing with that kind of maintenance nightmare,

36:32 you get to innovate again.

36:33 And that's what I'm wishing for the working group.

36:35 And I'm going to work hard to get us to a point where we can really be done

36:43 and start all over again.

36:46 This is very compelling.

36:48 And so I'm wishing you the best along this way.

36:52 We'll be following all the progress.

36:54 While you were speaking about sort of like how things could form in the future,

36:59 what popped into my mind was something you mentioned some months ago

37:02 when also Crab Nebula was announced to the world, about a browser.

37:08 Now, adding this actual layer of complexity, what could we say here?

37:16 There's some calculations that would suggest that the creation of the Chrome browser

37:20 cost a billion US dollars.

37:22 of the Chrome browser cost a billion US dollars.

37:28 If you look at the complexity of things like the web transition API

37:33 or the simplicity of the box model

37:38 inside of floating CSS

37:43 or complying with modern ECMA script.

37:49 It's a huge task, not to mention everything surrounding accessibility

37:54 and the requirements of various nations and super nations like Europe

38:00 for protecting human identity.

38:05 Do I think that there's space for something like that?

38:09 Yeah, absolutely.

38:12 Is it something we're going to solve with 50 volunteers and five full-time

38:18 engineers?

38:19 No, no.

38:22 Are there efforts underway?

38:26 I've heard of a couple, you know,

38:28 and there's a lot of projects who are trying to reinvestigate what the web is

38:34 all about.

38:34 And I'm pretty sure that some friends are going to be involved in that

38:41 conversation.

38:42 And I'm not trying to say no, and I'm not trying to say yes.

38:47 I think that, you know, as with any massive project

38:52 with enormous human involvement,

38:57 it's important to take baby steps. I think growing a community to the scope that Tauri has after five years

39:11 took five years.

39:14 You can't open up a company tomorrow and have 5,000 employees.

39:21 It's just not possible.

39:23 Even if you have 20 lawyers and 15 HR reps,

39:26 it takes years to grow a team,

39:31 whether open source or corporate.

39:35 It has to be humanly possible to do anything.

39:38 And I think that Tauri is a step in that right direction.

39:41 I think Servo is a step in that right direction.

39:43 And I think that stepping together through toward a future where we get to

39:49 decide what the internet looks like is an exciting challenge.

39:53 Are we there yet?

39:54 No.

39:55 Can we get there?

39:57 I think it's something that humanity has proven over and over again.

40:01 You can get anywhere you imagine.

40:04 And open source is, you know,

40:07 a tool and a community of choice

40:09 to help you get there.

40:12 You know, what's the saying that sort of

40:14 you want to go fast, go alone,

40:15 want to go far, you know, go with others.

40:18 And so my question is,

40:21 with open sources in the spotlight,

40:24 so it's commercial open source,

40:26 to people who are new to software, to startups, to open source,

40:31 and for those who are considering starting a company,

40:34 starting their own projects,

40:35 what could we tell them in terms of these new realities

40:38 and maybe any piece of advice

40:40 or something to keep in mind going forward?

40:45 Get your licenses straight from the beginning.

40:48 Don't rug pull your licenses.

40:50 Don't like create a new license unless you work very closely with,

40:54 no, never mind.

40:55 It's like creating your own cartography.

40:57 You don't do it.

40:59 Like get it right from the beginning, draw a line in the sand,

41:03 get inspired about what these different open source licenses mean,

41:07 and stick to it.

41:08 Like, no one is forcing anybody who uses Tauri to use an open source license.

41:13 I know a lot of people, friends of mine,

41:15 who are building proprietary software on top of Tauri,

41:18 and that's totally legitimate.

41:20 But take a stance.

41:23 Take a stance to it. And I think there's a risk in leveraging an open source community that companies can fall into really quickly when they think they own the community.

41:41 when they think they own the community.

41:47 And I guess the other piece of advice is protect your community,

41:48 care for your community. Because, like, if you're building an open-source software company,

41:54 the community is your biggest resource,

41:58 and throwing them over will lead to, you know,

42:01 a HashiCorp exodus where all of your competitors are like

42:05 gleefully, you know, rubbing their hands.

42:08 We're just going to make the same thing except cut you out of it, right?

42:13 It's like the IBMification of Red Hat where suddenly, oh, yeah, we're just going to make

42:22 it impossible for you to build and other people are like

42:25 no open source

42:28 and specifically the

42:30 communities behind the digital commons

42:32 will hold you accountable

42:33 and it's

42:36 like it's really important to get that

42:38 messaging right and to care deeply

42:40 deeply about them

42:42 because if you don't it'll show

42:43 and I think the open source communities are important.

42:50 And it's why I do what I do.

42:54 Very well said, very well said.

42:56 And to people who are getting started,

42:57 who want to learn more about,

43:00 who want to practice, first of all, software,

43:02 who want to make a contribution,

43:04 who want to get involved. I think, software, who want to make a contribution, who want to get involved?

43:05 I think there's some basics you have to understand

43:09 about modern software development.

43:11 You have to understand what Git is, unfortunately.

43:15 Git, fortunately, unfortunately, it is the system

43:19 that pretty much everybody uses on the planet for code collaboration.

43:26 And if you're getting into open source, it is always all about collaboration. So that's number one. Number two, check your tone

43:33 at the door. Like if you're rude to other people in the community and nobody knows you, it's a

43:38 great way to get hammer banned or reported to GitHub or whatever. Like, just be a kind person.

43:45 And think of the other person that you're talking to,

43:47 the other side of the issue or the pull request, the chat message.

43:52 You know, they're people too, and they care deeply about what they're doing.

43:57 And a great place to get started, it's repeated a lot.

44:01 I think a great place to get started is in the documentation.

44:05 Read the documentation.

44:07 Because if you're going to be using the software,

44:11 you should read the docs.

44:12 Because it's too complicated to just figure out.

44:14 And sometimes the best documentation is the code.

44:18 So maybe my secret message is read the documentation

44:21 because good code is self-documenting

44:23 but should have some documentation in it. And look for mistakes. Look for things you don't understand because then

44:29 you're practicing your skill of analysis. And that skill that you develop through the time

44:39 of reading the code and understanding it will at some point surface a mistake, whether that's a spelling mistake or somebody copy pasted the wrong thing. Then learn how to clone the repository.

44:52 Well, you probably have to fork it first and then clone the fork. And then like,

44:56 you see what I'm saying? Git is complicated. You know, you fork it, then you clone it,

45:00 then you make your changes, you commit with a proper message, and then you push that commit to a fork, a branch on your fork, and then you pull request that.

45:10 There's a lot of detail here in this complicated system. Once you've done it a couple times,

45:17 it becomes second nature, and there's great tooling for it. And that's why a low friction way for beginners, absolute beginners to do it is in

45:26 documentation. And that's like, I always feel like heart warmed when someone finds a mistake

45:33 in stuff that we have live on GitHub or on the website. It's like, thank you. I appreciate it.

45:40 And, you know, as, as time goes on, um, find things that you think could be done better or even

45:47 more important bugs stuff that is broken and create an issue and teach people how to replicate

45:54 it and maybe even figure out how to fix it and then create a pull request that closes the issue

46:01 that you just created and that's the collaboration that open source software teams are looking for.

46:07 And another entry point is Discord.

46:13 I think like almost all open source projects have Discords.

46:16 So like go to the Discord, ask questions, be wrong.

46:20 Have people tell you how to be right

46:23 and participate in that community that way

46:26 if you stick around for months the organizers of that server are going to recognize it

46:31 and you know become friends join the community participate and you know i think everything else

46:41 sort of falls from there absolutely and so get those table stakes in the beginning

46:46 and then expect that you will get feedback,

46:50 you will get interaction, you will get guidance,

46:53 you will get mentorship,

46:54 like you will work with a team that builds complex software

46:56 and puts it in production.

46:58 And where else can someone get that,

47:00 especially when they're starting up in the industry?

47:05 That's absolutely the industry. That's

47:05 absolutely the case.

47:08 What do you usually tell

47:09 founders and commercial

47:11 open source founders if you get the

47:14 chance to advise people?

47:16 Because you've been around for a little while, now you're going through

47:18 this motion yourself, just trying

47:20 to extract some

47:21 more advice here.

47:25 What do you mistake. Design a company.

47:28 I think

47:29 that there's a rush to

47:32 get to product market fit

47:33 and a rush

47:36 to create a product.

47:39 And

47:40 if you missed

47:41 creating a company, which means

47:44 you know,

47:45 your taxes, it means getting a financial audit.

47:50 It means having honorable contracts with your employees.

47:55 It's like the secret sauce of startup life is to be a company.

48:02 And if you're so focused on just creating a product,

48:07 you're going to run out of runway

48:09 and you won't have built a company.

48:11 And then you're going to get trapped

48:14 in the expensive capital market game

48:16 of raising another round.

48:18 And I think the other advice that I commonly give

48:22 is figure out how much you think you need to raise for three years and double it.

48:33 And that's how much you should be raising.

48:39 Human time is expensive, and you want to keep the best people around you that you can.

48:48 And running a shoddy business is a great way to lose people.

48:53 I mean, what else?

48:56 To quickly add here, the part of creating a company is also figuring out the business activities.

49:01 I mean, building products is one of them, but there's also distributing the product

49:05 and then having an outbound motion

49:07 or catering to the inbound motion

49:09 and having customer success.

49:11 And there's all these peripheral activities

49:13 that you need to have a plan for from early on, right?

49:17 And how all of these ultimately drive revenue.

49:21 Yeah.

49:22 Also, even if you're a CEO or a CTO, Yeah. your daily job, being in touch with the team, understanding how the team dynamics work,

49:45 understanding what your code does gives you confidence when you're talking about it.

49:51 And really don't underestimate the amount of time that goes into human resources.

49:58 Like HR, don't wait.

50:01 Don't wait on bringing in someone to work with you on HR. If you have a number of founders, make sure that one of them feels compelled to care for the team.

50:25 I mean, through the community, like you very oftentimes find people that then join full time.

50:29 And so just adding that note there that should make you like. It could.

50:32 But I think one of the risks of only recruiting from open source is that people who work in open source do it for a variety of reasons.

50:46 And there is a spectrum of social ability that people are on, on one side, the other, in the middle.

50:59 And just because you're good at open source doesn't mean you're good at a startup.

51:04 Startups change priorities quickly.

51:07 They sometimes pivot.

51:10 And if you're not able to pivot along with that because you have a very clear

51:14 thought in your mind of what things should be like, maybe it isn't for you.

51:20 And I can speak with confidence that not everyone who works in open

51:25 source is going to

51:28 be finding

51:29 joy in

51:31 a corporate environment,

51:33 even if it's an open source software company.

51:38 Very interesting. Do you have

51:40 any spicy takes these days

51:41 to share with people?

51:44 Spicy takes? I think PHP is okay. any spice it takes these days to share with people and things. Spice it takes.

51:48 I think PHP is okay.

51:50 I think Python 2 is not okay.

51:53 If you're still running Python 2 and you know you are,

51:58 my favorite language is Rust.

52:06 And I can't think of anything else.

52:10 Last remarks, perhaps, or if you would like to maybe share your screen again

52:15 and show the website, the repository, give a call to action to people.

52:21 Sure.

52:21 Let me find the Tauri

52:25 app, Tauri website.

52:28 All right.

52:30 73.9

52:32 thousand stars. Screen, here

52:34 we go.

52:35 Yeah, so, I mean, this is

52:38 the Tauri repo

52:40 and

52:41 I mean,

52:45 one of the, one of the docs that I love to share is this one, the,

52:50 the architecture MD and what it,

52:53 what it does and why you should have it is that it explains from a high level

52:57 how things are glued together. And I don't know, I'm,

53:02 I'm particularly particularly proud of this because it's leading by example, right?

53:10 Like, it's one thing to introduce developers to Rust through Tauri.

53:15 It's another thing to encourage young engineers in open source to do things in healthy ways.

53:26 I mean, one of my favorite files in the repo,

53:29 I think, is probably going to be the audit workflow right there.

53:37 So we run a crone job to check to see if there are rust

53:43 or NPM vulnerabilities that we weren't aware of.

53:49 I like it.

53:51 It makes me feel good.

53:51 It makes me feel like we've got a continuous something.

53:57 If you did want to contribute and you made a change that's about to get merged,

54:02 what you might notice is that we've got this weird thing called

54:05 dot changes. This is written by a Tauri core member and now member of the Tauri board, Jacob.

54:12 He wrote a system called, we've got a pull request here from four days ago,

54:20 and we've got a little Tauri patch bug fixed regression on the JavaScript.

54:26 And it

54:27 looks simple. It's just like, okay,

54:29 you make a pull request and you change one

54:31 thing. But there's so many implications

54:33 here, right? You're responsible for documenting

54:36 your change. We

54:37 know which pull request

54:40 it was. We have the commit hash.

54:42 We know who did it. That's all

54:44 stuff you can get out of blame.

54:45 But when you go to the

54:48 Tauri

54:49 project's website, so you

54:51 go to Tauri app, and

54:54 whoa,

54:56 that's weird. All right.

54:57 You go here, and

54:59 you look

55:02 for the

55:04 releases. you look for the releases,

55:08 what you get is, you know, features, enhancement,

55:16 bug fixes, and changes, right?

55:18 So out of this practice of documenting what you do,

55:21 we get, like, you can go back in time to,

55:27 well, you know, as we were building the system out back in 0.6,

55:32 which was, I think, 2020, right before COVID hit.

55:36 And, you know, we've got a bunch of different applications,

55:40 like a 404, for example.

55:43 Or here, RIE, breaking change, update raw window handle,

55:48 create to 0.6, drive in, RIE patch.

55:54 Okay.

55:56 Nice.

55:59 Like, what this teaches us about software engineering

56:02 is to document your changes, to explain it to other people, and to make it available to the broader community.

56:08 That's what transparency is about.

56:10 That's what accessibility is about.

56:11 And that's what cooperation is about.

56:14 Thank you for sharing these best practices.

56:17 This was very well on point. people, maybe foreclosers who have not yet tried Tauri,

56:26 let's give them a little

56:28 nudge and

56:29 a promise of what they will get.

56:32 You might have already

56:34 tried Tauri and you didn't know it.

56:35 You might have already tried Tauri and you didn't know it.

56:38 I think there's a lot of people building apps

56:40 out there right now.

56:42 The hurdles of

56:43 Rust is complicated, we help you with. That's what

56:47 the community is there to do. And really, I mean, if you're interested in building on mobile

56:55 and you complain about the complexity of iOS and Android, then you haven't been building on mobile

57:01 for the past decades. It's not easy. There's a lot of stuff and we try to make it as pain-free as possible and i think you should try it out i think you should

57:12 um file issues uh close issues actually close other people's issues

57:20 perfect perfect awesome daniel thank you so much for making the time it was a pleasure to