Hurd Progresses - Debian GNU/Hurd by end of 2012? | Hacker News Reader

Hurd Progresses - Debian GNU/Hurd by end of 2012?

h-online.com

31 pointsb2spirit23 comments

Serious question: what's this for? It doesn't seem to be an interesting research project, and shows no sign of being useful in a pragmatic sense.

Surely a clean-sheet design started _now_ would be more interesting and perhaps even useful than a 21-year-old microkernel. Think about all the changes in architecture, networks, storage, etc.

We're in a world with consumer processors with 4-8 cores and hardware multithreading, GPUs that rival or exceed (in certain domains) the processing grunt of the CPUs, network cards that can talk to L2 cache, SSDs, etc.

Surely enough has changed that a clean-sheet design could be a lot more exciting - and even more practical - than trying to finish a design that hasn't succeeded for 2 decades. A clean-sheet design would at least be interesting - we already know you can build a workable system on top of a microkernel.

This isn't a claim to know what a clean sheet design would look like; I haven't really looked seriously at OS research in 15 years. I just strongly suspect you might do radically different, interesting designs in 2011 vs 1990.

As far as I can tell it's more of a political statement than a goal to somehow deliver a better OS. RS has never been all that happy with Linux being the poster child for free software, especially since the ideological goals of Linux are not the same for the FSF. Witness the fact that Linux is GPL2 not GPL3 and isn't going to GPL3.

At the end of the day it provides an OS to the FSF which they can control, and which isn't Linux. Even if it was in all ways equal to Linux in performance, and features (which seems doubtful) it's unlikely that it would deploy to any great extent.

Just like ReactOS is pretty much "pointless" (given that Windows costs basically nothing), Hurd is practically pointless because it competes with Linux purely on ideological grounds, and is arguably an inferior product. It simply doesn't have enough differentiating it at this point for anyone to really care.

On the other hand, maybe they'll finish it before I die, and maybe it'll contain some ground-breaking improvement, and we'll all be eating humble pie. Anything is possible. In the meantime if they want to carry on working on it, then good for them. Isn't that what open-source is all about?

>Surely a clean-sheet design started _now_ would be more interesting and perhaps even useful than a 21-year-old microkernel.

Just curious, are there any such OS in development?

http://www.minix3.org/

Sape Mullender and Noah Evans from Bell Labs announced last year that they are working on a system called Osprey. There's video of the talk floating around somewhere.

I've occasionally thought it would be an interesting project to try to organize some kind of "Let's finish Hurd for them" meet up, maybe in conjunction with some major conference that a lot of Linux kernel hackers attend.

Let's call it the Gathering of the New Herd!

I was thinking "Hacking Hurd for GNUmanity" as a play on "Habitat for Humanity".

That would be nice, but honestly I doubt it since they have been trying to release it for longer than I have been alive.

Anyway the Linux kernel is good enough, so the effort seems wasted.

Rivalry is always good, it's essential for evolution.

Let's see, we have FreeBSD, NetBSD, OpenBSD, and Solaris. Two of them have even been integrated in current GNU systems (Debian/FreeBSD and Nexenta). How does yet another kernel help?

Hurd isn't stuck in the seventies and aims to do more than re-implement a 40-year-old design.

The irony of this remark is that while it is literally correct, the truth of the matter is that Hurd is stuck in the early 90s and aims to re-implement a 20-year-old design.

... while wrapping it to make it look like an early 1970s design.

Or we all can just stick with Windows XP. Wouldn't it be great?

I played seriously with Hurd in the early 2000s, and remember really loving many of its ideas. The user space file systems (which Fuse does for other OSes) were quite cool. I can't imagine that much of the good stuff Hurd promised hasn't already been adopted by others.

It would be nice if everyone take some time off linux to help with Hurd. Participation and write 1 line of code can go a long way in helping them along :). I really want to see Hurd completed.

It might be nice, but from a purely selfish point of view I'd rather people (not just coders) spent time improving Linux (which is already in widespread use) than tinkering with Hurd.

How well does Hurd compare to Linux?

Not well. It has only a better design (in theory!) and is GNU (so politically speaking is more Free Software than Linux...).

Is Hurd licensed under GPLv3? There is a reason why Linux isn't.

I think Debian already works with HURD using Mach 4 as long as the core libraries include extra header information(1). I believe this process is called "porting"(2) in Debian lingo but involves the whole distribution, i.e. just core libraries but driver frameworks as well. When I think of porting, I think of rewriting in another language. This is a patching job to me, massive as it may be, but whatever.

My information only comes from my research made the better part of a decade ago(3) when they were seriously looking to port the microkernel to L4 from Mach 4(4), but sadly it appears that work was abandoned around 2006 [Dunno, why? It might be started again]. I suspect the low level switch had a lot to do with interest in a better kernel in OS X than what Darwin(5), XNU(6) on the Mac, could provide(7). In the Panther, pre-Leopard days there was a lot of interest in this(8) kind of project: large-scale microkernels(9).

I don't know if this effort will include the port from L4(10) with the Mach abstraction API to "save" the prior work in getting Mach to work with standard Debian without the extra header requirements.

Anyway, I'm going to get some coffee and lunch. If anyone wants to fund an idea like this, I would like to any of the number of amazing things that they are doing with non-tricyclic antidepressants these days. Just look at the good it did for Tony Soprano.

(1) http://www.debian.org/ports/hurd/hurd-install

(2) http://www.gnu.org/software/hurd/hurd/porting/guidelines.htm...

(3) http://www.shakthimaan.com/installs/debian-t41.html [Not me]

(4) http://www.gnu.org/software/hurd/microkernel/mach.html

(5) http://darwinbuild.macosforge.org/

(6) http://developer.apple.com/library/mac/#documentation/Darwin...

(7) I don't have any insider information. I'm just a developer.

(8) Mac OSX Panther for Unix Geeks, Brian Jepson, ISBN-13: 978-0596006075, Chapter 7

(9) http://www.ertos.nicta.com.au/software/darbat/

(10) http://kilobug.free.fr/hurd/pres-en/abstract/html/abstract.h...

Reply on news.ycombinator.com