Dutch governments builds alternative for Microsoft based on NixOS

Several European governments are experimenting with NixOS-based systems as a way to reduce dependence on Microsoft and other US tech vendors, with the Netherlands’ DAWO project highlighted alongside French and German initiatives. Commenters see NixOS’s declarative, reproducible configuration as a strong fit for centrally managed public-sector desktops, but question challenges around usability, device management, and especially replacing entrenched tools like Microsoft 365 and Excel. The thread also touches on software forges (e.g., Forgejo/Codeberg) and AI-generated code policies, as well as broader concerns about digital sovereignty, regulatory leverage, and the risks of continued reliance on large US platforms.

Motivation: Digital Autonomy and “Post‑Microsoft”

  • Many see DAWO as part of a broader EU push to reduce dependence on US tech (Microsoft, Office 365, Azure), driven by fears of surveillance, lock‑in, and “enshittification” (ads, telemetry, shifting terms).
  • Dutch effort is discussed alongside German (openDesk) and French (La Suite, Securix/Bureautix) projects; some hope for EU‑wide collaboration, others expect fragmentation due to differing bureaucracies.
  • Political backdrop: strong criticism of US policy, sanctions risk, and Trump‑era unpredictability; several comments frame this as a national‑security and sovereignty issue, not just cost or ideology.

Why NixOS (vs “just Linux”, Guix, etc.)

  • NixOS praised for declarative, reproducible configuration and easy “custom distro” creation, ideal for large managed fleets and quick rebuilds after failures.
  • Compared to Guix, commenters cite Nix’s larger ecosystem, better package coverage, and smoother fit with systemd‑based software; Guix is admired conceptually but seen as less practical today.
  • Some are already running NixOS with Lix (a drop‑in Nix implementation).

Practical Challenges: Management, Office, Ecosystem

  • Manageability: Nix’s model (central configs pushed to machines, possibly immutable systems) is seen as a natural MDM/Group Policy replacement; tools like Sextant (GUI for NixOS management), osquery/Fleet, and SextantFleet are mentioned.
  • Skeptics worry about public‑sector capacity: hard to hire/retain people who can own Nix tooling versus more familiar Windows stacks.
  • Biggest pain point identified is the office suite, especially Excel‑class functionality and seamless document exchange with external partners using .docx/.xlsx. LibreOffice/Collabora and national office suites are promising but not fully drop‑in.
  • Some propose running Microsoft Office via Wine/Proton on NixOS as a transitional compromise.

NixOS Experience and Usability

  • Enthusiasts report very stable, “solid‑state” systems and say LLMs significantly lower the learning curve for writing Nix configs and packaging.
  • Pain points: unusual language, flakes and overlays complexity, packaging non‑trivial apps, and the rigidity of fully declarative systems when debugging.
  • Debate on whether NixOS is too “expert‑oriented”; counter‑argument is that ordinary staff use GNOME/KDE and web apps, while IT handles configs.

Code Hosting, AI Policy, and Federation

  • DAWO hosts code on a government Forgejo instance and mirrors “community parts” to Codeberg.
  • Large subthread on Codeberg’s policy against projects that are mostly AI‑generated:
    • Supporters see it as protecting resources and avoiding legal/security gray areas.
    • Critics find it vague, chilling for AI‑assisted projects, and potentially untenable as AI usage grows.
  • Related discussion about federated forges (Forgejo federation, ForgeFed, Radicle, Tangled), moderation politics, and whether federation actually reconciles divergent policies (e.g., on AI or content).