153: DHH – Omarchy and Designing Your Own OS on Arch Linux

**Adam Wathan** (0:01)
Okay, David Heinemeyer-Hansen. So right now, we're recording a video podcast, and you are on Linux. This is something the internet would have you believe is not necessarily possible.

**David Heinemeyer Hansson** (0:14)
It almost wasn't possible. That's the funny thing, right? We just tried to get it going with Riverside, and there was definitely some snafu. But then we jumped into Squatcast, and your ear pods doesn't work. Computers, man.

**Adam Wathan** (0:28)
Nothing works. I don't know.

**David Heinemeyer Hansson** (0:31)
Or everything works. Everything is just an interesting mystery you got to find out. And I think that's certainly an attitude you should adopt if you're going on a Linux adventure. I think you should have that attitude for any kind of computer because they all break it sometimes. But it is probably fair to say that Linux has more of those adventures, at least in the startup phase, than you're probably used to if you have something that's been vetted by QA department of 1,500 people who bang on everything. It's a little more community focused when you do it with Linux, just as it is with all kinds of other open source. But I think what's funny about Linux is, a lot of folks have some apprehension about that, while at the same time entrusting their entire business, their net worth, their access to the Internet on Linux, Linux, Linux, Linux, on open source, open source, open source, because 95 percent of all servers in the world run Linux. Virtually every system is built with open source software, same kind of model. It is interesting that the last frontier seems to be the desktop, that that's where people go like, that's too scary to try something else than what I've been running the last 20 years.

**Adam Wathan** (1:45)
I feel like historically for me, at least, I mean, there is that element of like terror, of not feeling like you know how to make everything work. And I think we can get into it with Arch Linux especially. That sounds especially terrifying. But I think historically a lot of it has just come down to like, oh, the tools that I need to use, like equivalents don't necessarily exist on Linux, you know? And I don't really feel that's as true today. And I personally, I kind of feel that's especially because of like the increasing capabilities of the web, you know, like you can use Figma in Linux, whereas prior to that, there was probably not any great professional UI design tool you could use on Linux, you know?

**David Heinemeyer Hansson** (2:29)
It's completely changed the game. And I'd say that change of the game is a huge part of why I'm so fascinated and interested and enthusiastic about Linux. Because if you rolled the clock back 20 years, like the list of little Mac apps I used to use, like I was, I don't know, very long. The list of native Mac apps I used just before switching to Linux was yee-bick. And all of those had great alternatives on Linux. So except one, Adobe Lightroom. I still really like for my photography. And I thought like, oh, I need a Mac to run that. Well, it turns out Adobe runs great on Windows, and you can dual boot Windows on any PC. So that's just what I'm doing. I need Lightroom about-

**Adam Wathan** (3:14)
It probably runs better on Windows. I wouldn't be surprised.

**David Heinemeyer Hansson** (3:17)
It probably kind of does. I think Adobe and Apple, not always the best.

**Adam Wathan** (3:22)
There's a lot of Windows first software that I found definitely works better on Windows. Like if you use like real Excel on Mac OS, it really sucks and feels sluggish compared to running it on Windows.

**David Heinemeyer Hansson** (3:33)
It totally does. So I think if you're going to have like a backup OS for that 1.2 app that you positively need and you can't get anywhere else, it takes about 30 seconds to reboot into Windows if you have two NVMe drives in your machine. So that's a hassle. I wouldn't want to do it for an app I needed five times a day, but for something you need it once in a blue moon, like I do with Lightroom, it really isn't a big deal. But I think the largest story, as you say, it is the web. I mean, Figma is a great example. Our designers at 37signals, they mostly develop straight in HTML, straight in CSS. They don't use a lot of pre-production tools, but occasionally they do. And if they're working on graphical design, if they're working on logo stuff, it's all Figma. And Figma is just a web app. I mean, isn't that kind of incredible that that's how far we've gotten? And not only is Figma a web app, but like Teams, Microsoft Teams. I don't know a lot of people in the US who use Teams, but in Europe, it's absolutely dominant, and it just works as a web app. So does Google Meet, so does Zoom. Zoom also has a native app, but virtually everything today is on the web, which I of course find fantastic because the web is what I love. That to me is the greatest application platform. I don't actually want a ton of native stuff on my desktop. I want just a handful of tools. I want a great terminal. I want some keyboard interaction with how I set up my machine and all that stuff. But other than that, I kind of want a lot of web apps. So Omarchy, for me, the aesthetic that really hit where I went like, I actually can't use Ubuntu anymore, was that a tiling window manager allows you to reduce any window to like one, I think we were on a double border, border highlight around something, and then you just see the web surface. There is zero Chrome. There's zero browser navigation. There's zero any of that. The dash dash app equals URL, that Chrome and Brave and the other Chromium-based browsers have, is amazing in its aesthetic purity when you paired with a tiling window manager like Hyprland. And then you get something where Linux, I think, for a long time, kind of, at least my perception of Linux for a long time, I'd look at someone running Ubuntu or KDE, and I'd always go like, it's not quite right. Like the proportions aren't quite right. The font rendering always looked a little off. The different GTK toolkit stuff looked a little off. And I just went, I don't love that. And you look at what's possible with the modern setups, with stuff like Hyprland, with stuff like running Chromium, with dash dash app and all this other stuff. And you just go like, hey, this doesn't look just on par to my aesthetics, to my eyes, it looked way better, way better. And someday I have a computer that is just not only different, which I actually think is a property that's valuable in and of itself. I've been running the Mac for 20 years. Do you know what? I'm a little fucking tired of the Jonathan I aesthetic. I mean, it's a great aesthetic. It's a wonderful aesthetic. But it's like you're wearing that same plate shirt for like 20 years. You're like, maybe you should try to mix it up, wear something else just for a little bit. The novelty of it, it's funny because in technology, not without reason, novelty is often poo-pooed upon. Like, well, if it's just different, it's not better. You're doing it wrong. Jonathan Ive has a great quote in that. We're not just trying to be different, we're trying to be better. But I also think there is some value in different. There's some value in the excitement. Remember when, was it iOS 7 came out, when Jonathan Ive just redesigned everything, we were done with skeuomorphic design. And first everyone was like, oh my God, this is crazy. But also, this is kind of cool. This is new. And then they tried just recently to do it again with the glass thing, and it didn't quite have the same impact. Let's just put it like that.