It's Rust based and brand new (more secure, less bagage, Wayland native, no fragile plugin system, etc). I'm really looking forward to it.
I love Gnome but I want (at least) quarter tiling, but rather have the flexible tiling KDE offers (where you drag windows into predefined areas and they snap, preferably windows always open where they are dragged once). I want speed, I want it to get out of my way. My current flow is: Open window, hit win + left arrow, open another, hit win + right arrow, have 2 windows together on ultra wide. I do the same on 2 or three other desktops. And then I start working.
Oh, something nicer than Network Manager would be nice, something simple with super simple WireGuard VPN integration, etc.
Currently I'm still very happy on (daily driver) Gnome on NixOS, but will surely check this out soon.
> I love Gnome but I want (at least) quarter tiling
This is what I've used Pop!'s Gnome extension (https://github.com/pop-os/shell) for in the past. I don't think it'll receive much love after Cosmic is released (after all, Gnome itself is being dropped) but if it still works for whatever version of Gnome you use, it may be worth checking out.
After using FancyZones on Windows, Gnome's lack of tiling management on my ultrawide has become a bit of an annoyance for me. Unfortunately, my copy of Pop Shell broke at some point.
> no fragile plugin system
While I don't like Gnome's tendency to slow down or crash, I do very much like the plugin system. Things like GSConnect and various smaller tweaks improve my Gnome experience a lot.
I hope Cosmic does expose some kind of plugin system, though hopefully one that's not as prone to lag and crashes.
The plugin system has many issue. A single failed plugin takes all of them down. And its not a stable interface, each version threatens to break many plugins.
Its absolutely annoying. You have configured something you like, next gnome version they are gone.
Cosmic plugins are their own processes (generally written in Rust) using very fast Wayland and be much more stable. In fact, a much larger amount of the functionality already lives in plugins. Its a game changer in the long term.
It's alpha. Of course not.
But the fact that is has one at all, is an improvement over GNOME where the plugin "api" is just javascript injected into the main gnome shell process.
The main reason why System76 decided to build Cosmic in the first place was to avoid having to spend an inordinate amount of time (I heard 50%) fixing their extensions any time Gnome made a new release.
The fact that they were able to release this alpha of Cosmic DE in one year just goes to show how much better a use of time it is.
That's really nice, I'm using it now and it works well.
You probably know this already, but I'm saying it anyway (hope you don't mind): The intuitive keyboard shortcuts would be win-right-arrow (then) up arrow for the top right corner tiling, like Windows does quarter tiling.
Also, Gnome (half) tiling has the windows sliding, your extension jumps them, and requires extra button (ctrl) ootb.
Very small things, and I will for sure keep the extension installed and use it, thank you!
KDE can set windows to start at a location, or remember the last position. the only downside is some applications like steam do not differenciate. Every steam window is just steamwebhelper so I cannot remember my friend list position
Gnome comes with a lot of baggage both technical and organizational, this has lead to situations where Pop_OS! wasn't able to manage/change/improve things like they think they needed to do hence why they started to create Cosmic. Just be clear this was a business decision by system76 to some degree, not just some "I don't like it so I created something even if it doesn't make sense" decision.
it looks like Gnome because it's for people which had been using Gnome so far, but it's just similar not the same and likely will only diverge more over the next many years assuming it succeeds
Isn't the business decision exactly whats being questioned? Not saying it is a bad decision, but I did have the same question. Was it really necessary to write a completely new DE just to make some tweaks easier to implement? Would love to hear more details about this.
At first glance, it seems like a truly massive effort for a marginal improvement, that could have been made with a fraction of the cost. I'm probably wrong though, hopefully.
The problem was that the Gnome crowd is going increasingly in a different direction. Mutter already doesnt support a number of otherwise universal protocol. The way theming work is compelely changed. There are lots of other reason why staying with gnome was an increasing problem for them. All those reasons are only growing bigger.
So to support what they wanted and what they promise their costumers they would have to do increasingly more and more work, in direct oppositon the the gnome community. Essentially eventially forking it. And then they have to maintain a huge legacy codebase.
And they didnt wirte a completly new one. The compositor was already a project. So was Iced. They use lots of other existing project in the Rust ecosystem.
Now they are on their own, free to build what they want, rather then endlessly pushing a rock up a mountain. They can now make more fundamental changed that would never have been possible in Gnome. They are also free to build the community they want rather then deal with the Gnome community. This also makes them a player in the Wayland community, another voice that can push protocols.
The situation between System76 and Gnome had become untenable. The modifications they wanted to make to the shell were too many and the Gnome devs don't want to support any of it.
I read somewhere that 50% of the development time on Pop_Os was spent simply un-breaking the Gnome shell extension between releases of GNOME.
Now they can put that time into creating what they actually want.
