Slashdot Mirror
GNU Hurd 0.5, GNU Mach 1.4, GNU MIG 1.4 Released
jrepin writes "Which day could be better suited for publishing a set of Hurd package releases than the GNU project's 30th birthday? These new releases bundle bug fixes and enhancements done since the last releases more than a decade ago; really too many (both years and improvements) to list them individually, The GNU Hurd is the GNU project's replacement for the Unix kernel. It is a collection of servers that run on the Mach microkernel to implement file systems, network protocols, file access control, and other features that are implemented by the Unix kernel or similar kernels (such as Linux)."
30 years for Hurd 0.5, so 1.0 will be available in 2043?
I deny that I have not avoided attaining the opposite of that which I do not want.
"Development of the Hurd has proceeded slowly." (http://en.wikipedia.org/wiki/GNU_Hurd)
As per http://www.gnu.org/software/hurd/hurd/status.html: " It may not be ready for production use, as there are still some bugs and missing features".
Exactly how long has it been like this ? I tracked this project for about a decade until I concluded it would never be ready for production - over a decade ago.
Oh, I don't know maybe some day in the early 90's. Back when it would have been useful to me. /kidding only a little.
Well.. maybe. Or Maybe not. But Definitely not sort of.
I've heard it said that time moves on and that maybe even Linux won't last forever. Wether or not that is true I Believe *nix in general will be around for a long long time yet. Fast forward a decade or two - despite Hurd and it's sllloowww.... development being a bit of a joke, wouldn't it be something if we are all actually using Hurd in the future? Stranger things have happened.
Brought to you by Carl's Junior.
Fork a BSD variant, license it under the GPL, package it with GNU stuff, call it Hurd 1.0.
What exactly is relevant about Hurd now? The OS landscape has changed and people have moved on. This is really a non-story, aside from the humor value.
...on my PowerPC 620...
Lawrence Person (lawrencepersonh@gmailh.com (remove all "h"s to mail)
http://www.lawrenceperson.com/
Comment removed based on user account deletion
...of a kernel that doesn't actually work. Except on Stallman's PC.
Stallman does not possess such devices he runs and developed emacs on a unix VM inside his brain! After realizing that all unix passwords and attempts to hide source code in a binary were useless. The concept of a conceptual computer without passwords and accessed only by obscure command macros written in C exploded from his mind and POOF we had emacs. This was then enhanced by interpreting the commands in binary form but it only worked for those who spoke with a lisp. Then all this went out the Windows when a stiff DOSe of source code was obscured by means of non standard compilers and suddenly word and data processing binaries could easily be obfuscated by hiding the source.
Others tried to change this situation by judiciously applying rubber to source code and the resulting LateXT could be stretched into a usable FLEXable shape, at least until a Bison shat on the source. Stallman HURD about this change in how binaries were now being used and created and GNU for certain that he would have to come out of his brain and actually become the Kernel in charge of parsing things at the source. Because he still insists upon compiling source only in his brain before creating binaries the resulting OS kernel has been extremely slow to take shape because debugging it has given him nightmares whenever he actually sleeps in fact the that the sleep command causes instant dreams that bring him back to the Bill Gates rants he witnessed at computer club meetings in the 1970's.
This message was not sent from an iPhone because Peter Sellers really was a deviated prevert without a dime for the call
IPv6 support in pfinet, based on Linux 2.2.14.
30 years to get to v0.5? Time to adopt Scrum.
You have to admit with the recent news of Cyanogenmod selling out, there is an opening in the phone market for a truly free OS not controlled by some mega-corp.
I love the idea of GNU Hurd but it's just not progressing like it should.
Does anyone know why this project is stalling so much?
I'm looking forward to the Alpha port, though I'm also hearing good things about Itanium.
It was my understanding that everyone had HURD...
What does GNU Hurd have that Minix 3 does not? They are both microkernels except Minix 3 looks more mature.
GNU Hurd uses a sane license. Minix 3 uses a BSD license, which is unfree.
That's hilarious. (I don't mean just the bit where, like so many FSF fanboys, but not RMS, you can't grasp the difference between "Free" and "Copyleft"; I mean the argument as a whole.)
So HURD's only benefit (that you can think of) is that if some evil company wants to take a microkernel-based UNIX-like, make their changes, and distribute the result without source... they'll be forced to go with Minix 3 (the one that GP says "looks more mature", which you don't seem to dispute) instead of HURD (which you can't or won't explain any technical benefit of)? Yeah, I think they'll just go with Minix 3, same as they would no matter how you licensed HURD.
things tend to go slow. Real slow. If you want things now, now, now, pay the man/men. It is free, as in someone-else-will-do-it, so you get what you, that's right, didn't pay for.
Fortunately, eventually people found this hobby project worth paying for, although I think it proved its worth before the big money started pouring in.
There are, of course, some other hobby projects that also manage to support a little more hardware than the Hurd does without huge amounts of money poured into them.
You have to include the original copyright notice and the terms that pertain to that code. But you can certainly distribute BSD code within/alongside your own code under your own license. That's kinda the point of the BSD license.
Just sprinkle few patches throughout the codebase that are difficult to strip out, and your new project becomes de facto GPL.
It's the other way around.
Minix3 is a Pure Microkernel Multiserver system, using a custom microkernel. The kernel is small and everything else runs contained in userspace processes.
HURD is a hybrid system, which means drivers run in kernelspace, negating most of the benefits of using a microkernel in the first place. This is a consequence of using Mach, which has severe performance issues which pretty much exclude its use on a pure microkernel architecture system.
Other Mach-based systems such as Darwin (used on OSX and IOS) are also hybrids, for the same reasons.
But Mach is pretty much an ancient, archaic microkernel. L4 showed the world that the whole microkernel thing can be done in a reasonably efficient manner.
Gnu Hurd or Minix 3
And never has. Apple got a Mach *based* kernel working by having complete hardware control and not having to develop drivers for alternative hardware. The Linux, kernel, and BSD kernels that are atually in use on broad sets of kernels, accepted the risk of having the drivers inside the kernel so that they could get them to actually *work*.
As a 'product' it may be a dismal failure, but the work getting to it has clearly been not and we every day enjoy the 'collateral successes'.
For all his faults, RMS did help the 'movement' in incalculable ways.
---- Booth was a patriot ----
That's how we roll.
Says who? Cyanogenmod taking money from the public is fantastic and means more time working on the project we love.
- Michael T. Babcock (Yes, I blog)
Bullshit. It's as free as it gets. It's the definition of free. You can do anything with BSD material, including proprietize derived works. That conflicts with some people's concept of subsuming derivative work and making it as open as the original free bits, but that doesn't allow redefining "free" as "free, and forcing yours to be free, too".
And if Mach's architecture entailed a high overhead on a 16 MHz single core 16020, it would be utterly negligible on today's 3 GHz four core CPU.
And why, pray tell, is developing reliable drivers for a microkernel more difficult than for a monolithic kernel? Logic would dictate entirely the reverse.
and don't forget about the PPC port on that site, with posts all the way into 2003, that project is on fire by HURD standards. We'll finally be able to run HURD on our Macs........*!* oh wait.......
Why should he be worried since he has the source code? Another bombshell for you guys. His systems did not even have passwords until an untrustworthy worker started fucking around.
Only the State obtains its revenue by coercion. - Murray Rothbard
Comment removed based on user account deletion
I once noted that the best CPU for pure FSF/GNU stuff would be a VLIW CPU. Since all the dynamic analysis is done by the compiler, every time a CPU undergoes changes, backwards compatibility should be broken. Bigger register set? Recompile. More branch units? Recompile. More ALUs? Recompile. With all that recompiling that is needed, the FSF goals of software being completely liberated in terms of the source code being always available would be more successfully enforced.
Not that that would apply to the Itanium, since that CPU has become more RISC in Itanium 3 than it initially was.
His PC doesn't run HURD - it runs gNewSense, a distro of what he calls 'Libre-Linux'
Actually, the perfect pure GNU system - HURD on an Itanium (or a real VLIW CPU), w/ Emacs on top of it as well as GNOME 3. Only impurity here - X11, which isn't GPL3. Maybe they can try and do a GPL variation of either X11 or Wayland or Mir, whichever they prefer.
I'm curious. How much time and money would using the goo.im updater save for work on other things?
Yes. It would be the perfect cpu for a gentoo style OS. But from what I heard, Itanium failed to live up to its promise because the compilers, particularly gcc. weren't that good at optimization.
not sure why this got modded down. Has anyone ever tried to setting up three monitors? I recently converted my workstation from Windows which worked quite well for gaming and multi-monitor support. On-board nvidia gpu had one monitor and the amd/ati pci card connected the other two. Can't do that under X. So now I'll have to buy another PCI card. But wait, commercial drivers are buggy and Xscreensaver is a mess, I can't use text virtual consoles, but I use open source drivers which fixes that, steam wont run. Running Ubuntu? Having the fear of god every time a system update or reboot is not ideal, but about half the time something bad will happen. BSD? nothing is written for it. I've been running X in some form on some machine since the early 90s on a 386. For a while, the unix desktop was the way to go, but for years Windows or OS X is vastly superior in a desktop environment - you know, actual drivers exist. Thank god for virtualbox at least.
Fuck Ajit Pai
And never has. Apple got a Mach *based* kernel working by having complete hardware control and not having to develop drivers for alternative hardware.
...and by running drivers for network and storage devices, and the network protocol and file system code that uses them, in kernel mode.
Actually, one good way of slowing down GNOME would be to ask that team to port it directly to HURD, w/o going through X or Wayland or anything else. In fact, build a display server directly into either HURD or GNOME. Preferably GNOME, since that would be quicker.