Watch Out Linux, GNU Hurd Coming 463
sfcrazy writes "Debian now has concrete plans to bring GNU Hurd to the larger community. GNU Hurd is expected to be released with the release of Debian 7.0 Wheezy towards the end of 2012 or beginning of 2013. Debian maintainer Samuel Thibault has already produced a Debian GNU/Hurd CD Set with a graphic installer which is available to download."
A random observation (Score:5, Interesting)
One of the very few people to put me on her Slashdot enemies list did so because I made a derogatory statement about the length of the HURD development process. In, as I recall, the year 2000 or 2001. It was a running joke at least five years before that.
Way to be timely and relevant, GNU.
Re:Interesting (Score:5, Interesting)
Nope. All I saw was an organization that maintains a distribution that is the basis for one of the fastest growing Linux distributions offer a distribution based on a kernel made by an organization that maintains one of the most used software compiler suites and userland tools. Please tell me where I should be looking. :P
Re:This can't be!! (Score:5, Interesting)
I first waited for this, back before FTP distribution was possibe. Stallman sent GNUsletters on xerox paper. With stamps.
GNU emacs was distributed in source. On QIC-02 tape.
And CMU Mach was to be the centrepiece of a system with the few GNU utilities.
Was that 1988? I think so.
Re:Interesting, but.. (Score:5, Interesting)
It has always been interesting for me to reflect on the HURD. The issue of mind share was crucial when comparing Linux and the HURD. Back in 1991 (or early 1992, maybe... can't remember that far back) I tried to contribute to the HURD. I had done some work on Mach in an OS course at school and was interested in playing more with it. But since I was pretty much a new grad and didn't have a proven track record working on OS kernels, the HURD team told me to take a hike. Well, they were polite about it, but were clear that they didn't want help from a nobody.
Linux was completely different. Linus may have blasted your code, but he accepted any and all help. This created mind share. For those who weren't around at the time, the whole idea of accepting work from any random joe off the street was a relatively new concept. The "Cathedral and the Bazaar" hadn't been coined by ESR yet and the normal way to do things in the free software world was to have one or two uber programmers hacking away, never seeing the light of day. Now everybody realises that a key indicator of success on a free software project is having an open and unobstructed development process.
The HURD has some good ideas, but their initial attitude killed them. Even though the team is very different now (from what I've hear, anyway -- lost interest in it more than a decade ago), there is really no chance of making a comeback, I think. Enticing eyeballs away from other projects will just be too difficult. Linus' biggest contribution to free software was *not* the Linux kernel, IMHO, but rather the development process that Linux used. He showed everyone how it should be done.
That'd still be an improvement. (Score:5, Interesting)
If that can actually be done safely and efficiently, it also means a non-free driver can't crash the kernel or fuck up other drivers. I would guess there are security implications as well.
Right now, a bug in the nVidia kernel driver on Linux could compromise the security of the entire machine, or crash the entire OS, or flip some bit in some other unconnected kernel system (or userland process), and it's hard enough to debug these things when you do have source code. So wanting an untainted kernel makes a lot of sense.
Re:A random observation (Score:4, Interesting)
Not really. Apple doesn't use Mach as a microkernel, they use it as a hardware abstraction layer. A few things in OS X use Mach ports, because they do have some advantages (e.g. being able to easily tell which process sent the message, in an unspoofable way - implementing Keychain without that is really hard). Everything else is done via the BSD subsystem. When you make a system call on a traditional Mach system, you send a Mach message. When you make a system call on XNU, you just issue a syscall / sysenter instruction and jump straight to the BSD system call handler.
Comparing HURD, which is a multi-server microkernel, to XNU, which is a monolithic kernel implemented as a single-server Mach kernel, is meaningless.