Repository navigation
Replies: 3 comments 1 reply
|
the default tools like file editor, file browser should be opt-in, not opt-out. With all respect to people working on omarchy, the last thing I want to do - to report file browser bugs. I use omarchy for my 9-5 work, not to tinker with new things |
|
I recognize that accommodating every user is quite difficult. It may be time to invite users to make choices, either during setup or later by opting in or out of certain experiences. |
|
Thanks for writing this up. One thing it shows is that Omarchy's defaults now serve a much wider audience than when the project started. Folio replacing Nautilus, Plain as the default editor, and the macOS-style quick launcher bar are all reasonable choices, but they move away from the original keyboard-and-terminal-first identity. As the defaults grow to fit more kinds of users, it seems fair to also let people choose which of them they want. Two suggestions that keep the defaults exactly as they are: 1. Make removing an app fully clean The usual answer to "I don't want app X" is that it's easy to uninstall, and the multi-select remove menu does help. In practice, though, removing an app often leaves things behind: keybindings that launch nothing, menu and launcher entries, web app shortcuts, and config files. If removing an app also removed everything Omarchy added for it, the "just uninstall it" answer would hold up completely. (#7801 covers part of this for keybindings.) 2. An optional "Customize" step in the installer
People who want Omarchy as a whole keep getting it with zero extra decisions; people who'd otherwise spend the first hour removing things can skip that. The installer already reads separate package lists, so this could build on the current structure rather than adding a second variant to maintain. |
Uh oh!
There was an error while loading. Please reload this page.
I've been meaning to write about this for a while. The recent decision to replace Nautilus with Folio, along with making Plain the default editor on new installations, finally gave me a reason to sit down and put my thoughts together.
Those changes aren't really the problem on their own. They're part of a larger shift I've been noticing over the last few releases, and I'm starting to wonder whether Omarchy is moving away from the things that made me want to use it in the first place.
When I first discovered Omarchy, what caught my attention wasn't just the appearance. It was the whole approach. A well-configured Arch and Hyprland setup, built around the keyboard, terminal applications, and a collection of powerful, established Linux tools. It felt like someone had taken all the tedious work out of setting up that environment without sacrificing what made it appealing.
I liked that Omarchy wasn't particularly interested in behaving like Windows or macOS. It had its own way of doing things, and you were encouraged to get comfortable with it.
Lately, I'm not so sure that's still the direction.
What happened to the original philosophy?
I went back and reread DHH's original introduction to Omarchy from June 2025.
What struck me was how much of his excitement came from the things that made Arch and Hyprland different. The modularity, the freedom to change almost anything, the keyboard-driven workflow. The idea was to make that experience accessible without spending days configuring everything yourself.
Even the official Omarchy manualstill says:
And:
That's the Omarchy I signed up for.
Compare that with the direction of Omarchy 4.5. We now have traditional floating workspaces, window title bars, a dock, and window minimization. DHH has described these features as a way to help people coming from Windows and macOS feel more comfortable.
Tiling is still the default, of course. I'm not claiming otherwise. But the change in emphasis is hard to ignore.
There's a meaningful difference between helping people learn a new way of using their computers and gradually making that experience resemble the one they're already used to.
I thought Omarchy was doing the former. Increasingly, it seems interested in the latter.
Quickshell was a bigger change than it first appeared
For me, Omarchy 4.0 was probably the biggest turning point.
I genuinely liked how the older desktop was put together. Waybar for the bar, Walker for launching applications, Mako for notifications, and separate utilities handling their own responsibilities.
Even something as ordinary as connecting to Wi-Fi or managing Bluetooth meant opening a terminal-based interface like Impala or BlueTUI. Wiremix handled audio in a similar way.
That might sound unnecessarily complicated to someone who prefers traditional desktop environments, but it was one of my favorite things about Omarchy.
These were small, focused programs. You could understand what each one did, configure it independently, or replace it if you preferred something else. The desktop felt like a collection of tools working together rather than one large application trying to handle everything.
Quickshell changed that quite substantially.
Now the bar, notifications, launcher, network controls, Bluetooth, and much of the surrounding desktop experience live inside a single integrated graphical shell.
I can see why the developers went in that direction. The interface is more consistent, and maintaining a unified system probably makes certain things easier. Quickshell itself is clearly capable software.
But I preferred the old approach.
It's not about refusing to use graphical applications. It's about what kind of environment Omarchy is encouraging people to use.
I keep seeing the argument that the new interface is still keyboard-driven, which is true. But keyboard-driven and terminal-first aren't interchangeable concepts.
A graphical panel that supports keyboard navigation doesn't give me the same experience as a dedicated terminal application.
And I don't think wanting to preserve that distinction is unreasonable, especially for a project that made such a point of embracing Linux's own way of doing things.
Why are we replacing established tools with new Omarchy applications?
This is another trend I find increasingly difficult to understand.
Omarchy has been building quite a collection of its own applications: Hype, Signed, Disktree, Monologue, Gliff, Plain, Folio, and others.
I have nothing against these projects existing. Some are genuinely useful, and I'm glad people are developing them.
What bothers me is the growing tendency to make them the preferred choice over software that's been around for years.
Take the move from Neovim to Plain.
I know Neovim isn't being removed, and existing users keep their preferences. Still, making a simple graphical editor the default for new installations seems like an odd decision for a distribution that originally put so much emphasis on terminal workflows.
Neovim has a learning curve. That's part of the deal. It's also an incredibly powerful tool once you get comfortable with it.
Omarchy used to feel like the kind of environment that encouraged you to learn those tools. Now it seems more interested in making sure you don't have to.
The Nautilus-to-Folio transition bothers me even more.
Nautilus is mature, reliable, and has an established ecosystem around it. Folio is still in its very early stages. Maybe it'll eventually become an excellent file manager, but I don't see why existing users needed to be moved over to it already.
My own update actually failed because of this change.
Folio installed successfully, but Omarchy couldn't remove Nautilus because I had ghostty-nautilus installed, which still depended on nautilus-python. I had to deal with the dependency myself before I could finish updating.
It's not a catastrophic problem, but it's a good example of why replacing established applications shouldn't be treated as a trivial change.
I didn't ask to replace my file manager. I was just updating my system.
And that brings me to the larger question: are these new applications becoming defaults because they're genuinely better at what they do, or because they're developed within the Omarchy ecosystem and fit its design language?
I suspect visual consistency and integration play a significant role, and I don't think those should outweigh maturity, functionality, and reliability.
I'd much rather see these applications offered alongside existing tools, with replacements happening once there's a compelling reason to make the switch.
I'm starting to wonder who Omarchy is being built for
None of the individual changes would bother me that much in isolation.
A graphical network panel? Fine. A dock? Sure. Floating windows? Useful sometimes. Video wallpapers, polished animations, integrated settings, custom applications? There's an audience for all of that.
But put everything together and the result feels quite different from the Omarchy I originally installed.
We went from a terminal-heavy, modular Hyprland setup to an increasingly complete desktop environment with its own shell, applications, interface conventions, and a growing list of features intended to make it more approachable.
There's also the increasing emphasis on Omarchy as an "agentic OS."
I use AI coding tools regularly, so I can appreciate the work being done there. Integrating local models and coding agents into the desktop is genuinely interesting.
Still, it's another area where Omarchy seems to be building more of its own infrastructure around workflows that previously relied on independent tools.
At what point does the desktop itself become more important than the tools it's supposed to help us use?
I don't think every Linux project needs to appeal to the widest possible audience. KDE and GNOME already do an excellent job of providing comprehensive graphical desktop environments.
What I liked about Omarchy was that it wasn't trying to compete with them on those terms.
It had a different audience and a different set of priorities.
I worry that in trying to make Omarchy appealing to more people, we're slowly losing some of that character.
Strong defaults are fine. Repeatedly changing my setup isn't.
I understand the whole omakase philosophy. DHH picks the defaults, and that's part of what makes Omarchy what it is.
I don't expect every application choice to be put to a community vote, and I certainly don't want an installer asking me fifty questions before I can use my computer.
But I think there's a line between providing an opinionated starting point and repeatedly changing the environment people have already built around it.
Choosing Folio for a fresh installation is one thing. Removing Nautilus from an existing installation, along with integrations people might be using, is another.
The same goes for keybindings, application associations, and desktop components.
After I've spent time getting my system the way I like it, I want updates to improve that system, not continually push it toward someone else's latest preferences.
I know Arch-based systems require some maintenance. That's fine. What gets frustrating is having to revisit working parts of my setup because the distribution has decided to replace them.
At some point, it feels like I'm spending more effort preserving my workflow than improving it.
That wasn't what attracted me to Omarchy.
I'm not asking for Omarchy to stop evolving
I don't want this to come across as a demand to return everything to how it was a year ago.
The early releases weren't perfect. Some of the newer features are useful, and there are good reasons to develop custom software where existing solutions fall short.
But I miss the balance the project used to have.
Back then, it felt like the terminal and the wider Linux ecosystem were the foundation, with a beautifully configured desktop built around them.
Now it increasingly feels like the integrated desktop is the foundation, and the tools are expected to fit into it.
I'd like to see the original terminal-first experience remain a serious priority. Keep the TUI workflows easy to access. Respect existing application choices during upgrades. Let new first-party applications prove themselves before replacing mature alternatives.
And please don't make people fight the distribution just to keep using the tools they already prefer.
I'm not asking for a separate edition of Omarchy or a return to the old minimal installer. There should be ways to preserve that style of working without splitting the project in two.
I'd really like to hear DHH's thoughts on this
What I'd most like to understand is how DHH sees Omarchy's identity today.
His original enthusiasm for Arch and Hyprland seemed rooted in discovering a different way of using a computer. More recently, especially with Omarchy 4.5, there's been a noticeable effort to accommodate people coming from more conventional desktop environments.
Is the terminal-heavy, modular approach still a defining part of what Omarchy is supposed to be?
Or is the longer-term ambition to build a more integrated, general-purpose desktop that appeals to a much broader audience?
I think that's worth clarifying, because the answer affects how people like me decide whether Omarchy is still the right environment for them.
I still enjoy using it. There are plenty of things it does exceptionally well, and I appreciate the work that has gone into it.
But I don't want to pretend I haven't noticed the direction it's taking.
What originally made Omarchy exciting to me was that it wasn't trying to make Linux feel like everything else. It made the things that were different about Linux feel exciting and worth learning.
I'd hate to see that gradually disappear in the pursuit of a more polished, more approachable desktop.
I'm curious whether other longtime users have had similar thoughts, or whether most people actually prefer the direction things are going.
And I'd especially appreciate hearing DHH's perspective.
References
All reactions