Slashdot Mirror
GNU Hurd 0.6 Released
jrepin writes It has been roughly a year and a half since the last release of the GNU Hurd operating system, so it may be of interest to some readers that GNU Hurd 0.6 has been released along with GNU Mach 1.5 (the microkernel that Hurd runs on) and GNU MIG 1.5 (the Mach Interface Generator, which generates code to handle remote procedure calls). New features include procfs and random translators; cleanups and stylistic fixes, some of which came from static analysis; message dispatching improvements; integer hashing performance improvements; a split of the init server into a startup server and an init program based on System V init; and more.
http://xkcd.com/1508/
It was rumored that both users could be hurd rejoicing.
HURD : a [H]urd of [U]nix [R]eplacing system[D]s
I should use this sig to advertise my book ISBN-13 : 978-1501515132.
What? No systemd ?
Recycle PCs and build a wireless community network www.hillsborough.org.nz
USB support only a few years ago? And there are people who expect Hurd to be taken seriously? Mind boggling...
These days you don't see the same hype around microkernals that you did back then. So we should probably warn the HURD team: If your boner for microkernals lasts more than 25 years, you should probably consult a physician.
I'm trying to teach myself to set people on fire with my mind... Is it hot in here?
dude, shut up
"Here Lies Philip J. Fry, named for his uncle, to carry on his spirit"
They announced work on Hurd when I was still in university. I've worked a career, ended up disabled, retired, and spent years on a pet project since then, producing 13 point releases. Over 30 years have gone by.
Yet they've still only reached release 0.6? So one decimal point release every FIVE YEARS?
Jesus.
Stick a fork in this project.
It's done -- as in dead. Pushing up daisies. Pining for the fjords. Defunct. Deceased. Non functional.
It's not even worthy of being called a pipe dream any more. Even "Duke Nukem' Forever" beat them to the punch, and everyone gave up on that project long before it was released.
I do not fail; I succeed at finding out what does not work.
Writing drivers is hard mmmkay...
Do not look into laser with remaining eye.
6 years ago I went to a speech given by Richard Stallman. The very first question someone asked was about the status of Hurd. He sighed, then proceeded to answer
They announced work on Hurd when I was still in university. I've worked a career, ended up disabled, retired, and spent years on a pet project since then, producing 13 point releases. Over 30 years have gone by.
Yet they've still only reached release 0.6? So one decimal point release every FIVE YEARS?
Jesus.
Stick a fork in this project.
It's done -- as in dead. Pushing up daisies. Pining for the fjords. Defunct. Deceased. Non functional.
It's not even worthy of being called a pipe dream any more. Even "Duke Nukem' Forever" beat them to the punch, and everyone gave up on that project long before it was released.
What do you want to name the fork? I vote for flock.
I don't think that there ever was supposed to be a Duke Nukem Forever. It was supposed to be a joke, as in, no game development would ever happen, so we would be waiting forever. Unfortunately someone else ended up with the rights to the franchise and didn't realize this.
Do not look into laser with remaining eye.
Even if it had been released as 1.0 it would still be as useless and irrelevant as it is at 0.6. Hurd users can likely be counted on a single hand at this point in time.
It's too bad Linus wasted all this time making a temporary kernel that was just going to be surpassed a mere 24+? years later by the official GNU kernel. Nothing stings more than when the code you write isn't being used.
Sure, but even toy OSes like MenuetOS have had USB support for longer than Hurd.
Well Hurd did not get the developer attention that Linux got. Obviously, this means that progress in Hurd is going to be slower than Linux.
MenuetOS has less than a handful of developers and yet has had USB support for at least 7 years now.
Writing drivers isn't hard.
But for the broader acceptance of an OS, one needs a whole shitload of them.
In the past, a computer with a half dozen devices was "packed". Today? A cheap tiny ARM SoC easily runs up to 30+ built-in devices.
All hope abandon ye who enter here.
I'll answer with two questions:
Why do you care what people do with their spare time and who the hell are you to tell people what's important to them?
But what was his answer?
From the GNU Hurd Wiki page:
It's time [to] explain the meaning of "Hurd". "Hurd" stands for "Hird of Unix-Replacing Daemons". And, then, "Hird" stands for "Hurd of Interfaces Representing Depth". We have here, to my knowledge, the first software to be named by a pair of mutually recursive acronyms.
—Thomas (then Michael) Bushnell
If it weren't for deadlines, nothing would be late.
To be fair, they know it's not going to replace the Linux kernel now, so it's merely an experimental project. Much like many of the Microsoft experimental kernels, except those eventually are put to death and the relevant guts are sucked into the next revision of Windows. (see Singularity OS--an OS that does away with hardware virtual memory and the translation lookaside buffer)
Who the hell works on the 99% of open source software that isn't popular, and why do they care? Because they do.
Oh I don't know, that was one of the best flame bait comments I've seen in a while. It was damn near poetic.
XML is a known as a key material required to create SMD: Software of Mass Destruction
I got the joke. I just didn't laugh.
If it weren't for deadlines, nothing would be late.
I think it's a good question, and they're both questions; no one suggested that they stop working on the project.
Why is it a good question? Why do we still read human interest stories? I think it's fascinating to know who is doing this work, and why.
--Jim (me)
He hasn't finished his answer yet. Check back in a few years.
Even Plan 9 was being used commercially years ago. It's disturbing to think that plan 9 is being used more than HURD.
Why don't you go help, then? Oh right, it's easier to just sit on the sideline and throw bricks at people trying to actually do something.
The reasons for disliking it vary, but there is at least one common thread among those who are upset about it: Systemd is being shoved down their throats, in that several of the most widely used, widely loved, deeply entrenched linux distributions have announced that they are adopting it. Many people who use those distributions do so for very good reasons, and since there are no equivalent alternatives, these people are being forced to either accept systemd (which will cause them unwanted trouble) or migrate to an inferior distribution (which will also cause them unwanted trouble). That kind of thing is enough to piss anyone off.
Why don't you go help, then?
Because it would be a complete waste of time. No one uses the Hurd and no one ever will.
is being shoved down their throats
Making no comment on the quality, intention or target of the parent, but regarding what I quote above: Is this not supposed to drive you to the greatest virtue of OSS?
Isn't the whole idea between (F)OSS to make it your own? In modern parlance: fork it. You don't like systemd? Fine, fork the distro and do whatever you want with it.
What I've gather from the whole systemd hatred is that a lot of people hate it, and would rather spend their time ranting about their hate than either fixing their complaints or forking a distro that meets their desires.
I'm a developer whose code sometimes runs on Linux. As long as my code runs on any distro, I have no horse in the race as far as how the system boots
Perhaps the recursion in "Hurd" is an endless one. Which would adequately explain why the development process never runs to completion. OTOH, if they had named it something that's merely vain rather than recursive like...I dunno..."Ricux", maybe it would be finished by now.
Hurd is still moving along but it is not a priority for the GNU project. The reason why it's not a prioty is because the problem of a free software (Unix-like) kernel is solved with Linux and the various BSD kernels. The FSF can direct the priority awareness to issues like free computer aided drafting, free hardware drivers for graphics chips and free BIOS.
Cool goalpost shifting. Stop trying to make excuse for the fact that the Hurd team is simply incompetent.
The Hurd team underestimated the amount of work it takes to replace the Unix kernel in the Hurd way. It turns out that it's exceedingly complex to manage timing, locking, racing and threading issues for multi-threaded multi-server systems that are written in C++; such issues are non-existant in Linux given that everything exists in the same "space".
Who the hell works on the 99% of open source software that isn't popular, and why do they care? Because they do.
Show me some proof that anyone cares enough to drive GNU/Hurd to a 1.0 release.
Richard Stallman founded the GNU project in September 1983 with an aim to create a free GNU operating system. Initially the components required for kernel and development were written: editors, shell, compiler and all the others. By 1989, GPL came into being and the only major component missing was the kernel.
In 2010, after twenty years under development, Stallman said that he was ''not very optimistic about the GNU Hurd. It makes some progress, but to be really superior it would require solving a lot of deep problems'', but added that ''finishing it is not crucial'' for the GNU system because a free kernel already existed (Linux), and completing Hurd would not address the main remaining problem for a free operating system: device support.
GNU Hurd
Try writing an opengl driver for a modern gpu..
The problem is when you fork your own distro you quickly discover that using systemd is the easiest way to maintain it. It isn't a coincidence that medium and small distros like Arch Linux picked it up in addition to the big boys. Unit-files save package maintainers boatloads of time they used to spend having to writing and maintaining initscripts a lot of which is copypasta boilerplate anyway but its usually distro specific copypasta.
This is the source of a lot of the strife in my opinion. The people who actually do work to maintain distros like systemd, the users not so much.
We'll have a real, honest-to-goodness 1995 level OS in time for 2045!
Chas - The one, the only.
THANK GOD!!!
It's just the nature of the "pie in the sky" architecture that they wanted. When they actually complete their mission, Hurd's system will be far more flexible, reusable and interconnected than any Unix system in existance. I like to think of Hurd as Plan 9++ as Plan 9 is a very flexible and straightforward "Unixy" system but Hurd takes that flexibility to the next level allowing users to do fantastic things without the need to communicate with a systems administrator to get things done. One example is that ordinary users can actually run instances of Hurd within Hurd itself allowing users to configure different Hurd instances for different purposes - an idea that's very similar to virtual machines and machine instances.
A part of me wants to write a monolithic "Unix interface server" in order to kick Hurd progression into a faster speed but there would be the irony of having a monolithic system in Hurd that was intented to give ultimate flexibility to the user.
well, then there's the oss written for some foundation or another for extracting pocket money from said foundation.... happens more than you would think.
world was created 5 seconds before this post as it is.
They announced work on Hurd when I was still in university. I've worked a career, ended up disabled, retired, and spent years on a pet project since then, producing 13 point releases. Over 30 years have gone by.
Maybe, what with all the time you have on your hands, you could have taken the time to research your facts.
There are two major distributions (Arch and Debian) that use Hurd as a kernel. Right now, I am running a GNOME desktop on Hurd.
No, it ain't Linux/BSD, but hardly a Duke Nukem scenario.
The version number is a poor argument. Enlightenment was at 0.16 when I was in college 15 years ago and 0.19 today, but progress has been made and people still use it. The problem with hurd is it's not in a realistically usable state yet.
This space intentionally left blank
I suspect that most casual users run Hurd in a VM. Really not much need for USB.
But now it has it. Rejoice!
I'm running GNOME desktop on Debian/Hurd. Works just fine.
101 posts and not a single one with technical content. Somebody should create a slashdot post generator, with modules producing output of these kinds:
- internet meme repeater ("year of Linux on the desktop", "stallman eats his own toes", "thou shalt not compare to nazi");
- xkcd repeater (its output is prefixed by the string "obligatory" and displays a strong prevalence of this one);
- project deprecator ("this software is so stupid, I could write a better one with one arm tied behind my back, except I'm too smart to actually do it");
- Google/Apple/Microsoft PR ("it's not Google who kills kittens! It's their subcontractors!");
and, last but not least,
- Slashdot deprecator ("slashdot is no longer a nice site to read these days").
Been a while since I had to do it, but I don't recall it being that hard. I did have to look at a book.
But that's not the point. If systemd was just an init system and nothing but an init system, so help it God, people could either take it or leave it. They'd have the choice, one way or the other.
The problem is that when it has dependencies on your window manager, microwave & healthcare plan you can't simply unplug it and replace it with something else.
Confucius say, "Find worm in apple - bad. Find half a worm - worse."
opengl driver
3D graphics is an outlier in the driver development.
But for a useable desktop, 2D graphics is sufficient. The 2D is commonly supported via GPU's ROM and as such implementing a 2D driver isn't hard.
Even 3D in itself isn't as hard. The problem is that games require (A) lots of edge cases optimized and (B) huge number of acceleration features implemented.
All hope abandon ye who enter here.
Ah, it's turned into a "let'smake our selves feel important by shitting on someone't hobby" day.
Yeah USB only a few years ago. Their goal is to write a microkernel OS and figure out how to make it work well for a UNIX like system with far more felxibility. The feature list and malleability of the system is impressive.
If they spent all their times on drivers and none on the base OS, they'd have yet another OS which is quite similar to all the others out there in terms of features. Their goal is not to get acceptance from random bitter blowhards like yourself on the internet.
End result: they've contributed more to the world than you ever will.
SJW n. One who posts facts.
This is where the exercise of free will kicks in. If you cannot contemplate learning something new, stick with what you have or choose a dist that chooses to do stuff the old way.
It'd be a lot easier if HURD attracted developers to work on it. The reason it is stuck in the mud is because looked up what pragmatism meant in a dictionary and decided it would be having none of that.
If the problem is solved, then move the development focus elsewhere. And I thought that FSF didn't actually like the idea of GPL2-only things?
However, the FSF development focuses are odd and abandoned.
For years a Skype-replacement was asked for. Apart from game-server-oriented once like Mumble, I can't see a viable OS alternative that works cross-platform, with video, etc. We could say that Jabber's use by Google Chat was the replacement but - again - why are we then continuing on the others?
For years, a Flash replacement was asked for. Last I checked, Gnash could do some things but nowhere near all.
Coreboot was asked for, but it's a REALLY niche project and whether that supports UEFI etc. I don't know.
Google Earth replacements, Matlab replacements, OpenDWG library (odd choice, I think), Oracle Forms replacements, all suffer the same problems.
And "automatic transcribing" is just silly and shouldn't be a priority at all.
The FSF's priorities are basically ignored and people work on what they enjoy working on. That's great and all, but HURD has been a long-time coming and can just barely run something approaching a Linux distribution. Given that it's replacing a kernel only - and userspace can be things like Debian software etc. - that seems an awful long time. Don't even get me started on filesystem or hardware support for Hurd.
Sometimes you just have to say "Oh well", and cut off the project. The people working on it would be much more useful on other projects that may well end up on machines worldwide rather than niche toy projects that have taken decades to get close to fruition. Not all of them would move, of course, but at least some of that talent could re-focus to other projects that are of greater utility.
Linus (impatient with the pace of HURD) developed a quick and dirty kernel that a Unix user land could be built on top of. He took a lot of shortcuts, he didn't think too much about portability and basically just made a beeline for the end line - to get a shell and hence other stuff running over a kernel. The kernel filled out and became portable as the project gained momentum and volunteers.
Whereas HURD got stuck up its own ass for correctness and politics. And that's even before Linux existed as a thing. It's hardly a surprise that when Linux did appear that people jumped ship.
It's true there was a debate about micro kernel designs but that alone doesn't explain HURD's failure.
Does it now support HDDs larger than 2 GB? I'm not even joking here.
Last time I heard (like 10 years ago or so) it was a theorists wet dream but basically unusable.
What's the state of things with Hurd nwo? Is it usefull already?
What are big steps Hurd still needs to take to be ready for prime time?
What are the plans? When are we there?
Please note: I have no problem replacing Unix with something better, like ome coolDMI thing where everything isn't a file but an object and the system is cleanly designed from top to bottom and back. Top notch but everything modifiable. But it has to be real-world usable and useful. Until then I'm sticking with *nix derivates such as OS X on Apple hardware or some x86 Linux like Debian or Ubuntu on ThinkPads.
Could someone give some enlightenment on this issue?
We suffer more in our imagination than in reality. - Seneca
Not every open source project has to be going somewhere. If people enjoy working on it as a hobby, good for them. I've written metric tonnes of code that has no real purpose or value beyond the enjoyment I got from creating it.
const int one = 65536; (Silvermoon, Texture.cs)
SJW, n: "Someone I don't like, and by the way I'm a fuckwit" - AC
For such a tiny developer base, writing drivers is a wasted effort if hardware has a Linux driver available under a compatible license.
What they ought focus on is a compatibility sandbox to load an arbitrary Linux kernel module and only then write *native* drivers where it makes architectural sense for performance or security reasons.
Makes Perl 6 look a bit rushed :)
No, it does not. The Window Manager have dependency on systemd, not the other way around. If the Gnome developers deciding that they need systemd, it's not systemd fault. You are free to open a bug ticket, but in the end it's the decision of the Gnome developers to use a particular technology.
http://www.mueller-public.de - My site http://www.anr-institute.com/ - Advanced Natural Research Institute
If the problem is solved, then move the development focus elsewhere. And I thought that FSF didn't actually like the idea of GPL2-only things?
You can't just force contributors out of a project. Because you know, freedom and shit
It may be irrelevant now, but it could become very relevant from one day to another if Microsoft decided to attack Linux, either directly or indirectly indirectly (e.g. by funding SCO again), in order to grab royalties and thereby delay their own inevitable demise. That's not such an unlikely scenario, and it seems good to have something else up one's sleeve...
No, but you can say "This isn't a priority any more, we don't think GNU Hurd got where it needs to be, we think your time is better spent elsewhere".
It won't be long before it's either forked (and thus no longer a burden) or abandoned.
"GNU Hurd runs on 32-bit x86 machines. A version running on 64-bit x86 (x86_64) machines is in progress." Are you kidding me, it's still just a 32 bit application? The way this is going, everything else will be 128 bit by the time they go to 64.
init? Why not systemd?
http://alternatives.rzero.com/
I fail to see how forking GNU would help.
http://alternatives.rzero.com/
Yeah, but they were the ones griping about Linux and claiming that they were going to replace Linux.
I do not fail; I succeed at finding out what does not work.
I'm sure the other Hurd user must be might impressed.
What do you mean shifting?
You first claimed all the Hurds problems were lack of developers yet I can name dozens of toy OSes with USB support that predate Hurd's support by years if not almost a decade. Then you just claimed it was all due to their design, yet MINIX3 has pretty much the same design has supported USB for most of its life.
Yeah, but they were the ones griping about Linux and claiming that they were going to replace Linux.
[citation needed]
SJW n. One who posts facts.
In chess, this is known as gambit accepted. Perhaps the joke is too subtle for your understanding.
Contribute to civilization: ari.aynrand.org/donate
No, people would move to many other choices long before HURD.
The early hype around microkernels was that they could emulate other OSes, and that's what led to IBM's attempt. It turns out what killed that and all the other attempts was primarily the fact that to emulate an OS, you essentially had to re-implement the target OS on top of the microkernel - that is, the microkernel did not end up replacing much if anything in your emulated OS. So it was just an added abstraction layer that you would need as part of the OS anyway, and you saved nothing.
That in itself wouldn't have been bad - a waste of time, but not otherwise bad. But microkernels do have inherent performance problems. Specifically any time you have tightly coupled components that have to share data, in a monolithic kernel you share the data and use locks to protect it, in a microkernel it's passed by messages. Even when optimized to just buffer copying, that can add up to a lot when the data is huge, like a video display.
The overhead normally is not really big, so in itself a microkernel can be competitive with a monolithic kernel. There are also technical advantages to compensate. I'm sure you know QNX (used in Blackberry 10, the more reliable automobile entertainment systems, etc.) is an example of that. The problem is the overhead just makes your emulated OS worse, any microkernel advantages are not part of your monolithic kernel code above it, and that makes it pointless. It's just as much work to implement, there's little to no saving, and it's measurably slower.
So in summary, the hype around the stupid idea of emulated OSes tainted the idea of microkernels, but there's nothing wrong with microkernels themselves.
No, but you can say "This isn't a priority any more, we don't think GNU Hurd got where it needs to be, we think your time is better spent elsewhere".
Sure, but that's essentially it's current state, and everybody knows that. But some people do want to experiment with a microkernel still, and so here we are.
This is where the exercise of free will kicks in. If you cannot contemplate learning something new, stick with what you have or choose a dist that chooses to do stuff the old way.
Thanks. You just provided yet another example of the same old fallacious argument. It completely ignores two important facts: The only alternative distributions left are either so anemic in their overall support that they would utterly fail as substitutes (debian's software archive is second to none) or are so different that a migration of any significant size would be unreasonably painful (let's see how long it takes you to migrate 50 machines to gentoo, let alone maintain it with 500 machines). Anyone who makes the argument you just made either has no grasp of the real-world issues, or is being disingenuous.
All it takes is the motivation, a group of likeminded individuals and the willpower to deliver a dist that does not use systemd. I expect most packages in the debian universe have no deps on systemd and therefore no work required to support those packages. So we're talking system packages, some daemons and maybe a few shims for edge cases.
As for why there are only 2 dists left not to have gone to systemd, perhaps that should serve as a clue in itself.
It's not as bad as the 4.x transition, but it's definitely true that the KDE devs have been somewhat bad at communicating that it's still at the stage where things are being reimplemented, so by no means is everything there yet. Honestly though the change from 4.x to 5.x has been far smaller than most, and apart from stuff missing because it just hasn't been gotten to yet, I can't think of much that's really regressed---at least as of the latest version of Plasma 5 that went into the Kubuntu beta last week or so, which also finally handles high-DPI without weirdness. But honestly, it's another transitionary period, if you don't want things randomly changing on you then don't follow it via a rolling release, because things WILL be broken and there WILL be wonkiness as the development goes through various systems and functionality and rewrites them for the newer frameworks.
Systemd, however, is another matter. It seems great on mobile and embedded devices, but I entirely agree with you that the layers of abstraction and automation make it really hard to figure out what the fuck is going on, especially when something is going wrong. I guess that's kindof their goal, though, at least indirectly, since the "GnomeOS" guys seem to want to emulate Apple, and that's exactly how Apple stuff is, it "just works" up until it "just fucking doesn't WTF" and it's just a black box of pain.
I remember sigs. Oh, a simpler time!
All it takes is the motivation, a group of likeminded individuals and the willpower to deliver a dist that does not use systemd. I expect most packages in the debian universe have no deps on systemd and therefore no work required to support those packages. So we're talking system packages, some daemons and maybe a few shims for edge cases.
You're implying that it would be easy. I'd like to think you're right. One group has already announced such a derivative. I'd love to see it succeed, but I'm not holding my breath. Maintaining a linux distribution as well as debian does, including timely security updates, package builds, downstream bug tracking, release management, and uniformity across so many installations as to form a vast support community, is a much bigger job than one might think. There's also the issue of various unrelated but popular packages developing dependencies on systemd, which means any such derivative distro would also be in the business of developing and maintaining forked versions of those packages; also not a trivial task. I guess we shall see.
As for why there are only 2 dists left not to have gone to systemd, perhaps that should serve as a clue in itself.
Many of them seem to be derivative distros that simply don't want to diverge from their upstream distribution's init system, so they have little choice. (Counting them as independent decision makers would be dishonest.) As for the upstream distros, I think it's more telling to note how very divided their communities were in the vote for/against systemd. A strong argument could be made that anything so integral to the core of an OS distribution should not be replaced with something so divisive to the community.
Speaking for myself, I'm a bit disappointed in the loudest factions of this disagreement. Most of what I see in these discussions is two mobs of people pushing for a decision *right now* (meaning this year, or next). The voices shouting "we need systemd!" or "we need nothing of the kind!" dominate the discussion, while a third option seems to have been forgotten: How about waiting until something can be developed that offers important core improvements over sysv init, but isn't as invasive as systemd? Most of us can obviously get by just fine with our existing init systems for a little longer; we've been doing it for years. The uproar over the topic is surely enough to motivate the development (or modification) of an init system that most of the community would find suitable. I'd love to see that happen.
In any case, I think the original question here has been answered. :)
What if they're generally the same devs? Because the core ones are all part of the same group that works at Red Hat, hence the (overblown) conspiracy theories.
I remember sigs. Oh, a simpler time!