That would be completely wrong. See the Fedora Miracle Spin, for example. The project (miracle-wm
) is still a work in progress, and yet Fedora already officially offers a Spin for it. What you're describing would only be true if Fedora was switching to COSMIC as the official desktop for Fedora.
That just depends on your use case. I'd imagine you'd only get 5-15% performance difference. Choice of programming language and the quality of the code makes a bigger impact on performance. Case in point, Rust's png library is 1.8x faster than the C libpng system library. Of course, only Rust applications benefit from that, unless the C libpng maintainers decide to adopt Rust's png library as their implementation.
There would be no reason to delay anything regardless of it being beta or not. Fedora is always doing rolling release updates of software that lends well to that paradigm, and COSMIC is one of those. It's not like GNOME or KDE where you have to carefully and meticulously manage 100 system libraries used by the whole ecosystem. System dependencies in the COSMIC ecosystem are virtually non-existent. If they really wanted to, they could just wait a few weeks to release the spin. But I don't think there's any reason to.
Not ready to release yet
Wayland compositors use IPC over a UNIX socket to communicate with Wayland clients. To increase security and enable sandboxed applet support, COSMIC applets use the security-context protocol for their IPC connection to the compositor. To be an applet, COSMIC applications use the layer-shell protocol to behave as an applet. Neither of which were made for COSMIC. Some other Wayland compositors support these protocols. You can see which compositors support the protocols at the bottom of the wayland.app protocol pages.
In practice, because Rust libraries are always statically-linked, the MPL-2.0 is equivalent to the LGPL in spirit. Meanwhile, because of the static linking restrictions in the LGPL, the LGPL is effectively no different from the GPL. Hence, you're going to find a lot of open source copyleft projects from the Rust ecosystem preferring either GPL or MPL-2.0, where MPL-2.0 is used in libraries where LGPL would have used previously in C projects. Dynamic linking is essentially going the way of the Dodo.
The Linux kernel already allows proprietary modules via DKMS, and a handful of vendors have been using this for decades, so this is no different. Case in point: NVIDIA driver, and Android vendor drivers.
All source code in Rust is statically-linked when compiled, which thereby renders the LGPL no different from the GPL in practice. For Rust, the MPL-2.0 is a better license because it does not have the linking restriction.
Niri is also based on the smithay library we use for COSMIC, so there's some collaborative work between COSMIC and Niri on Smithay.
applets live in their own process and communicate via Wayland protocols (behind a COSMIC API)
Even better. A COSMIC API was not necessary since Wayland protocols already exist for this (layer-shell and security-context).
Alpha 5 is actually releasing next week. I'd also disagree with "very alpha" as most of the beta milestone is finished now. Regardless, I don't understand where the disconnect is. A spin can release at any time. Doesn't matter when COSMIC Epoch 1 releases.