Yserver: A modern X11 server written in Rust

Original Article ↗ Hacker News Discussion ↗

Multiple Screens vs. Virtual Screen Model

  • Several comments clarify that “multiple screens” here means old-style X11 screens (:0.0, :0.1), effectively separate framebuffers with shared cursor/VRAM.
  • Most modern setups use one large virtual screen via Xinerama/RandR, with each monitor mapped into it; windows can move freely without apps knowing about multiple monitors.
  • Legacy multi-screen is considered obscure and awkward: apps must manage separate windows per screen, so almost none support it.
  • Some niche use cases still rely on it (flight simulators, psychophysics experiments, “zaphod” setups), leading to criticism of calling it “legacy” and dropping it outright.
  • Confusion arises between “screens”, “monitors”, and “virtual displays”; several comments clarify terminology.

Mixed DPI and Refresh Rates

  • One side claims X’s unified virtual screen model forces a common refresh rate and makes mixed DPI hard, and touts Wayland as superior here.
  • Others counter with examples showing Xorg can handle per-monitor DPI and refresh rates via xrandr, at least on some distros and hardware.
  • Mixed DPI on X often relies on monitor DPI attributes plus toolkit support (Qt with specific env vars), or manual scaling tricks using xrandr, with quality/performance tradeoffs.
  • Debate continues over whether Wayland truly “handles” mixed DPI better or mainly hides complexity via compositor-side scaling.

Networked GUI / Remote Access

  • Multiple commenters miss X11’s built-in “GUI over the network” and view Wayland’s lack of transparency as a regression, leaving VNC/RDP-style solutions.
  • Others point to Wayland tools like waypipe, RDP integrations in major desktops, and various remote protocols, though headless/remote-login support is still described as immature or fragile.
  • Some argue GUI-over-network is now niche compared to screen sharing, explaining its low priority.

AI-Generated Code and Project Value

  • Strong disagreement over AI usage: some label the project “slop” and assume little human review; others push back, arguing AI-assisted coding can raise quality, reduce drudgery, and enable exploratory projects like this.
  • One perspective: even if the end result is not production-grade, the learning journey (e.g., exploring X11 internals with AI help) has value.
  • Calls appear for clearer disclosure of AI involvement to set expectations.

Other Points

  • A question about NVIDIA support is raised but not answered in detail.
  • Some mention confusion with an old “Y Window System” project; others see the naming overlap as irrelevant given that project’s age.
  • A few users report successfully compiling and running Yserver with common desktop environments, though with quirks (e.g., compositor issues, manual TTY start).
  • The broader X11 vs Wayland debate surfaces repeatedly, with strong views on both sides about maturity, regressions, and priorities.