What I am most excited for in COSMIC is the promise of tiling in a full DE. I like the idea that you can switch back and forth.
I started trying it out a month or so ago. Still pretty incomplete. Promising though.
The fact that it may drive the Rust GUI ecosystem forward is exciting as well. I do not need to see everything re-written in Rust but it will be great if Rust is a realistic option for new app dev.
Tiling
It’s actually really good. I’ve been running the prealpha at times, and I’ve had no issues with tiling.
I’m missing 2 things from a real tiler: sloppy focus (WIP), and static workspaces.
So what sets COSMIC apart from the rest of the gang?
Wayland only, built upon only new code, pure rust as much as possible, native tiling, very fast apps
Full desktop environment with decent window tiling.
I am curious when they will release it as a full GNOME replacement, because that is a crazy task. At the current state, COSMIC is not ready at all. Even though it is already awesome.
How far is it to be daily drivable, in your opinion? Like, crazy far or just far?
I have a few machines running Pop. It’s exceptionally good IMO. It’s like an extremely refined Ubuntu. It’s one of my fav distros.
Is that your experience with the OS or cosmic?
Yes, I agree. Pop!_OS gets a lot of hate for some reason, but it’s actually a really, really good distro.
I was asking about COSMIC though, since I’m really looking forward to try it!
I have not yet done any extensive reading on what exactly distinguishes cosmic from the general Gnome DE.
It has no relation with GNOME
I run it on many of my devices, but I am absolutely waiting this one out to see just how useful it is, what’s missing, what’s not, and until it’s ready to be a daily driver. Very exciting.
Modern design they say? It still looks like 2010. They can’t even get the spacings and paddings right.
The project is motivated by “I like Rust, lets make a whole desktop in it” not by good UX.
Depends on your point of view.
Their motivation was “we have a vision for our UX and GNOME won’t let us do it — so let’s write our own.”
It was only after deciding to write their own that they decided to write it in Rust.
They like Rust, but that is not what motivated them to make COSMIC.
My view is that if the goal was to effectively make good software they wouldn’t start from scratch.
If they used wlroots the desktop would be usable today with a good feature set.
If they used Qt or GTK they would have feature rich well supported software. (GTK4 could have been an improvement for them, it’s designed around being minimal and having platform libraries implement design choices)
They didn’t take a practical approach imo. You could argue its a long term investment but because of it it’s probably years off of feature parity. The only upside today is… it’s written in Rust.
Sometimes old software just has too much legacy spaghetti written in to really build from though. Starting from scratch gives new ideas room to breathe and grow that might otherwise be impossible to implement in the previous framework—which while probably useful can also be stifling. See the reason why Wayland is being written to replace Xorg.
They dix not build the compositor from scratch, they built it on top of smithay, a library similar to wlroots but written in Rust.
I don’t know if you’ve actually tried to use GTK or QT, but it’s insanely painful. There is a reason almost all apps are written in Electron. Native GUI toolkits suck. If they had used GTK they would have still had an outdated and hard to maintain toolkit, and to deal with Gnome politics. Using GTK was actually the initial idea.
If we want Linux Desktop to succeed, at some point we have to build tools that people want to use. I’m glad they’re doing it.
You can’t announce a “drop.” It just drops.
They just did. What are you gonna do about it?
Does it fix window scaling if you have 2 monitors with different resolution settings?
More importantly to me, can blind users even install the OS.
All current mainstream distros now use Wayland, which has broken screen reading, so the OS cannot be installed.
Fuck. That is way more important.
Honestly, it’s not as important. These projects are working with very limited resources, typically dependent on free labour. Accessibility is incredibly hard to get right and half arsing it isn’t going to work. The priority should be pushing out a reliable, working prototype that people want to use. Once that’s accomplished you can refocus on expanding the features.
Demand for reliable multi monitor support is going to be far higher than screen reading capabilities.
