Since we entered a new year, we also entered a new quarter, and that means a new quarterly report from the Hurd, the project that aims to, to this day, developer a kernel for the GNU operating system. Over the course of the fourth quarter of 2025, an important undertaking has been to port dhcpcd to Hurd, which will ultimately bring IPv6 support to Hurd. For now, the port only supports IPv4, only works on Ethernet, and is still generally quite limited when it comes to its functionality. It’s a great start, though, and an amazing effort.
Furthermore, Q4 2025 also saw improvements in symmetric multiprocessing support on x86, not exactly a small feat. There’s a ton of work left to be done, but progress is being made and that’s important considering today’s processor landscape. There’s also the usual load of fixes, smaller improvements, and changes all over the operating system, and the report makes it clear that Debian’s recent announcement that APT will start requiring Rust is not a major issue for Hurd, as it already has a Rust port.
I started using Linux in 1993. And since then i’ve heard about Hurd “being worked on” or “making progress”. Sometimes even “big progress”.
I’m still not impressed.
Hurd. 36 years in the making and very little to show for it. At this point in time it is clear this microkernel is someone’s hobby plaything. At this pace, maybe it will be running on a 2025 machine a year before the heat death of the universe. GNU got lucky with Linux. I wish I could still get excited about Hurd news, but at this point in time Hurd is not an also ran but a no go.
Contrast this with Redox. Announced shy of 10 years ago and I can see me trying this out before 2030. If the FSF is serious about having their own kernel, they better start a new project from scratch and go for speed and usefulness. Hurd won’t deliver.
Your comment (and orzel’s as well) look like GNU Hurd still being actively developed is something that hurts whole OSS movement.
No, it doesn’t hurt (hurd?) anyone, but it is also a futile endeavor for anyone but the hobbyists wasting their time on this. Good for them, but I don’t need to know how they spend their time. After 36 years it’s clear that Hurd won’t deliver a useable kernel ever. Even ReactOS has more activity and achieves more than Hurd, while being a very slow project.
I guess what I am trying to say is this; report on Hurd when they reach version 1.1. Anything before is dead kernel walking.
And you’re requesting limiting such reports from from the site that has a name “osnews”.
I understand you might not be interested in some news that appearing here, but information about small and hobby OSes are published on Osnews since the beginning to this day on the regular basis. And for some reason, HURD frequently receives such rants, as it should close its development, or hide somewhere underground.
Even if in the beginning FSF had intended HURD to be a ‘proper’ replacement for UNIX kernel, and still endorses the project, it has s no longer the same goal, and any progress here comes not from the FSF attempts to ‘oust’ Linux, but from the fact that there are volunteers who are interested in doing this. It’s legitimate reason for them to work on it, and for Thom to inform us about Hurd.