Updates to SteamVR to fix their Linux-specific bugs, like broken room view, lighthouse power management/firmware update, inconsistent performance and reprojection issues
OpenComposite to have a longer list of working games through it
More polished Wine/Proton Wayland driver
Implementation of Windows Spatial Audio in Wine
Better handling of audio sample rates/allowing adjustment of sample rate per device
Hardware video acceleration in Electron, ex. when screen-sharing on Discord
In no particular order:
PikaOS 5. I want to see this project flourish, and I think they bring some much-need UX innovations to certain GUI tools (their system update interface is the best I've seen so far). I also love that they've dumped Ubuntu in order to do the CachyOS optimization thing upon a Debian base while still keeping everything bleeding edge.
Improved default keyring services in KDE. kwallet is kinda messy, and some people have pointed out that their use of blowfish is behind current best practices. On the flipside, using PGP means entering your password twice to unlock your keyring, so the experience is just not great out of the box.
Total Linux desktop share at 3%.
More/Frequent upstream gaming improvements from the Valve x Arch joint effort.
Nvidia integration parity with AMD
Open source Nvidia driver (as long as we're wishing)
Creative software to just announce they they'll think about Linux.
No particular order to these.
Rainbow six siege support, GOG Linux program, SteamOS publicly released and 7% marketshare.
HDR support and better maintainers are the things I can remember now.
Please for all that is good in the world, let the Linux foundation invest the majority of their funds in the Linux kernel first. Then please have a marketing budget bigger than what's left.. A paltry 2% of their funding going to kernel stuff is an absolute joke.
The kernel also badly needs to be rewritten in rust. Not all at once but piecemeal. The rust tooling is lightyears ahead of whatever is being used now, getting into a rust project is also so much easier. There's another project, the name of which I can remember, doing a rewrite and building it is a simple as cargo build. Seriously, that's the level of simplicity I want, not whatever bullshit I have to do now and follow a guide, take my own notes on how to make it work on my specific distro, fail a few times, and stare at unreadable GCC errors.
Also please please please stop using a mailing list (or at least make it optional). Holy moly are you losing a bunch of next gen devs through that. Regardless of how controversial the CoC was, it finally changed the tone a little, which is nice, but mailing lists are seriously archaic. The absolute minority uses them. It's another big thing hampering kernel development (at least to me). They are a terrible experience through and through that make it difficult to follow discussions, diffs difficult to read and interact with (leave comments on lines and respond to them), and spams the inbox.
Please let kernel development move into this century.
I'd like to see herbstluftwm ported to Wayland, so I can do an apples-to-apples comparison, and be ready to switch should I ever need to. It'd also be nice if, when I did try Wayland, there wasn't something completely borked about it that causes me to switch back within a day.
While I'm wishing for impossible things, for Linus to admit he was wrong, and that there next major release will be a proper microkernel, where module crashes don't force reboots and zombie processes can be cleaned out.
Working rootkit anti-cheat, so I can dump Windows.
all 39 comments