30 Years On, HURD Lives: GNU Updates Open Source Unix Kernel
The latest version of GNU HURD is out. If you're asking, "What is GNU HURD?" you're probably in good company. But as the open source kernel that was supposed to do what Linux ended up doing—provide the core for a cross-platform, Unix-like operating system whose code would be freely shared—the HURD is important. That it is still being actively developed three decades after its launch is worth remarking.
The latest version of GNU HURD is out. If you’re asking, “What is GNU HURD?” you’re probably in good company. But as the open source kernel that was supposed to do what Linux ended up doing—provide the core for a cross-platform, Unix-like operating system whose code would be freely shared—the HURD is important. That it is still being actively developed three decades after its launch is worth remarking.
The HURD was the kernel that Richard Stallman and other developers began writing in the 1980s to serve as the core of the operating system they aimed to build called GNU. Like Linux, HURD is designed to be the core of a Unix-like platform. Unlike Linux, HURD adopts a microkernel architecture, which some developers in the 1980s deemed superior to monolithic kernels, the alternative. (Even Linus Torvalds, who created the Linux kernel, wrote in 1992 that “microkernels are nicer,” despite his having adopted a monolithic design for Linux.)
Although the GNU programmers succeeded in building most of the development tools and other basic components they needed to create an operating system, they famously failed to bring the HURD kernel anywhere close to completion within the time frame they had originally envisioned. As a result, the GNU OS remained, for a while, something like a car without an engine: All the auxiliary parts were in place, but there was no kernel to hold them together and produce a system that was actually functional.
That all changed when Torvalds and a community of programmers working over the Internet began building the Linux kernel in the early 1990s. Linux provided the central missing piece that GNU lacked, leading to the introduction of the first distributions of “GNU/Linux”—a term that, to the consternation of some Free software purists, is no longer in widespread use. Most people today simply talk about “Linux distributions,” ignoring the bit about GNU’s role in helping to build the open source ecosystem.
Despite all of the above, HURD development continued even after it became clear that the Linux kernel was, practically speaking, a more viable alternative, and the HURD lives on today. Just this month, HURD developers rolled out version 0.6 of the software, which brought updates to a few tools included in the kernel and minor changes to the architecture of various of the servers that comprise HURD’s microkernel system.
Practically speaking, HURD 0.6 does not matter much at all for the open source ecosystem. Indeed, it does not even work on 64-bit machines yet. But the fact that the HURD developers are still churning out new code attests to the enormous diversity of the open source world, where even a project that basically failed twenty years ago can remain active as long as at least a few programmers are still interested in it.
‘Most people today simply
‘Most people today simply talk about “Linux distributions,” ignoring the bit about GNU’s role in helping to build the open source ecosystem.’
Maybe so, but putting GNU first is IMHO equally wrong.
Nope
Nope
Nope
Nope
Surely ‘free software’
Surely ‘free software’ incorporates the right to freely call it whatever you want to call it? I don’t recall reading anything in the GPL that says my OS has to be called GNU-anything.
(This is a distinct matter from providing credit for the pieces you used)
Oh, and the “g” in “gnu” is silent. So there.
I believe the HURD has some
I believe the HURD has some interesting technical features. And I expect it’s a bit smaller than the Linux kernel at this point. I can imagine if I were the kind of person who wanted to take courses in operating system design, I might want to hack on it; there would be more opportunity to as it were get my hands dirty and deal with basic issues than with Linux.
A 30 Year old kernel, that is
A 30 Year old kernel, that is still not usable.
“Even Linus Torvalds, who created the Linux kernel, wrote in 1992 that “microkernels are nicer,” ”
It is only “nicer” if it works. The Linux kernel does!
“Practically speaking, HURD 0.6 does not matter much at all for the open source ecosystem.”
Agreed! Just because I call it “Linux” instead of “GNU/Linux” does NOT mean that I don’t recognize all the contributions that the Free Software Foundation has given to whatever you call it. Nor do I ignore the massive amount of software, protocols, technologies, and other contributions that all the other individuals and organizations have given to us, over all the decades of computing in the 20th and 21st centuries.
Perhaps the SystemD folks
Perhaps the SystemD folks could port their abomination to Hurd, and bring it to a complete stop rather than it’s current crawl…..
It might have failed as a
It might have failed as a widely used kernel, but for a programmer’s work it is success. This means there are more options now and depending on what you want, you can always make a choice. This also opens a door for people to join the hack to building more robust and diverse kernels. I think they have done a good work and you never know what might happen in the future. Kudos !!!
Exactly — if only more
Exactly — if only more people had your view on this matter!
As far as I am concerned, there is not such thing as a “success” from the eyes of a programmer. Success only matters from a business perspective where money is important. Without making money, the project fails. For GNU/Hurd, failure it irrelevent because it is a Free Software project, dedicated by programmers who LOVE what they do. They work on a project they enjoy (mostly) in their own time!
Looking at it from this point, GNU/Hurd is successful!
Also, as you pointed out — this provides us further with choices. Sure, The linux kernel is still going to be my main choice at this moment in time. However, the Hurd will improve as time passes. One day, one of my machines may be better suited for using Hurd. One day!
Sure I understand that not everyone will agree with me 100% here. However, for those negative a**holes on here regarding projects like Hurd can seriouslty do one!
Well done Hurd team for your hard work. Same thing applies to many other great projects even I have heard of. Oh… well, they are probably not successful. I take it back! 😉
“Practically speaking, HURD
“Practically speaking, HURD 0.6 does not matter much at all for the open source ecosystem. Indeed, it does not even work on 64-bit machines yet.”
Yes, you’re right. But, once a production-grade version of Hurd is released, getting up to the point of full-fledged GNU/Hurd based distributions of the likes of Ubuntu and Fedora shouldn’t take a lot of time.
We just need to figure out how to make a stable, production-grade release of Hurd. And the rest, are just details.