Home > X11 > GNOME and KDE on Hurd GNOME and KDE on Hurd Submitted by Seo Sanghyeon 2005-06-16 X11 35 Comments Michael Banck reports about porting of GNOME and KDE to GNU Hurd: GNOME screenshot and KDE screenshot About The Author Eugenia Loli Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker. Follow me on Twitter @EugeniaLoli 35 Comments 2005-06-16 10:37 am what is great about hurd? 2005-06-16 10:39 am Not only for this when they finally have got some way to make an user interface…It looks like hurd is the way to go soon-when the linux kernel is so much used…and If it find the ways to the desktops…And the user scale will increase..and the science had said that Windows has so much problems cause it’s big userscale-if this will be in the way for linux? I don’t know..but when and if that day come – I will move to use the hurd kernel. 2005-06-16 10:44 am I appreciate that not everyone can speak english, but seriously, that is unintelligable. I wonder if thats the result of an online translator? Back on topic, hows HURDs performance these days? Anyone here have any first hand experience with running it? 2005-06-16 11:09 am I’m impressed. 2005-06-16 11:11 am “what is great about hurd?” people use to ask the same thing about linux! http://www.gnu.org/software/hurd/hurd.html http://www.nl.debian.org/ports/hurd/ Well, it is truly GNU so no arguments whether about what to call it which will certainly ease a little strain off of RMS! little info too: While many people call the GNU system GNU/Hurd this is not strictly true. The kernel is GNU Mach not the Hurd. The Hurd is a series of servers which run on top of the microkernel, GNU Mach. Both the Hurd and GNU Mach are part of the GNU project while the Linux kernel is an independent project. 2005-06-16 11:12 am No one speak english on heart, or very very few people As Anne Mc Caffrey would say, we all speak *Basic*, kind of an english made by people from many countries who have to understand each other. It reminds me some english people upseted in Cambridge, when we speaking “english” together (Italian, spanish, etc..)… So don’t be so “english” it is perfectly easy to understand. dikation is just saying that “when linux will be used by too many people, it’ll we be interesting to move to an other kernel in order to avoid mass adoption problems”. I imagine, that such problem can be virus, malware, etc… By the way, I’m interested too about hurd benchmarking. 2005-06-16 12:53 pm The HURD is a beyond Unix, microkernel OS. It’s really flexible, in that you can replace parts of it with your own, this is useful if you tune you OS for a specific purpose. The filing system translators are cool, there is one with witch you can mount FTP servers locally, you could probably write similar translators for SVN, CVS and so on. You have an expanded user system, there is an anonymous/guest user with limited privileges, and root isn’t as powerful as on Unix, you can protect your services from him. The HURD has some cool features but I guess we will only see what it’s capable to do when it’s mostly done. 2005-06-16 1:07 pm …after a longgggg time 2005-06-16 1:18 pm — “So don’t be so “english” it is perfectly easy to understand.” I don’t expect everyone to know english or even speak it well, but don’t tell a native english speaker what is “perfectly easy to understand”, because it was not. Your comment wasn’t either, though it was understandable. Im not “being english” as you put it, I don’t care wether that guy knows english or not, he can post random characters for all I care, but if a post is very difficult or impossible to understand, it just IS. You can through your political correctness around all you like, it doesn’t make the post any easier to understand. 2005-06-16 1:34 pm i don’t want to hurt any developers feelings, but just because it can run kde and gnome doesn’t mean it will be released this decade 2005-06-16 2:14 pm well, be patient… now is just a short, compared against the infinity of the future. 2005-06-16 3:22 pm Does it prove anything? 🙂 It’s just GNOME and KDE, and not Hurd at all. 2005-06-16 3:25 pm Sometimes the truth is ugly, and sometimes constructive criticism is harsh but looking at all the posts here pending review, I must say somebody went overboard. 2005-06-16 3:33 pm This proves something alright. In the past the HURD had problems with threaded apps, this was a big problem in getting larger apps to run, like for example Gnome. So having this type of stuff run on the HURD proves that threading problems and lots and lots of other problems faced by the HURD programmers in getting large apps running are now taken care of. If you can’t see that these screenshots are from the HURD look at the views on the filing system. P.S.: E17 runs on the HURD to 2005-06-16 3:54 pm HURD sounds GREAT. In practice I wonder if it will ever have even good performance. I hope it will! I really do. I love the servers –> microkernel concept a lot. @johnlein: How does E17 perform on a HURD machine? Less sluggish than GNOME/KDE? Didn’t I recently see an article on why Mac OS X Server does not compete well against Linux? It had to do with the fact that the web and database servers need to make use of kernel resources, but kernel threads are limited (instead of using server/app level threads)? Are such things being taken into account? 2005-06-16 4:31 pm >there is one with witch you can mount FTP servers locally, >you could probably write similar translators for SVN, CVS and >so on. Can do this on Linux using AVFS, at the moment it supports floppies, tar and gzip files, zip, bzip2, ar and rar files, ftp sessions, http, webdav, rsh/rcp, ssh/scp sites. The KDE and GNOME file managers do the same functionality as well. >You have an expanded user system, there is an >anonymous/guest user with limited privileges, and root isn’t > as powerful as on Unix, you can protect your services from him. You can do the same thing with SELinux, not becoming standard for many Distros. Linux will not have a traditional root account for much longer. 2005-06-16 4:34 pm I believe HURD has moved to the L4 micro-kernel as Mach was not optimal for the HURD design. http://www.gnu.org/software/hurd/hurd-l4.html#TOCwhat They do need to update their website though. 2005-06-16 5:10 pm You can do lots of things with Linux that you can do or will be able to do with the HURD, it’s just more painful. About AVFS, I can’t find much about it, there is a project on sourceforge that links to FUSE but if AVFS has become FUSE I must tell you that there is no chance of that being included in mainline. About SELinux, from what I’ve understood SELinuxes intention is to limit user from doing malicious things. The HURD tries to limit root from doing malicious things. Anyway, like I said, we will have to see what the HURD can do once it’s usable. It’s very flexible and most of what I could do probably hasn’t been dreamed up yet. 2005-06-16 5:32 pm I thought that Hurd was slowly getting into the console-only stage! But now I’m definetly impressed. Two more years and I guess we’ll have a complete Hurd-based desktop (Debian K20?) 2005-06-16 5:36 pm I don’t know. As far as I know, nobody made any benchmarks yet. Personally I would guess E17 is faster since it’s not as heavy as either Gnome or KDE. 2005-06-16 7:26 pm I’m so very glad I’m living now in these interesting times. When HURD is finally ready for consumption — by folks other than its developers — I’m guessing there’s going to be a pretty huge schism: The GNU/Linux users (who’ll still call themselves “Linux users”) and the GNU users. Will GNU be similar to use as today’s GNU/Linux? I’m guessing the commands and shell will be the same, but what will change from a user’s perspective? For example, will file permissions work the same with GNU? Will ldd/ldconfig still work the same? Will top and ps give the same sort of output they do now, or will it look more like BeOS (with all those threads and teams)? What will change from the user’s perspective? How will it be different from the application developer’s perspective? 2005-06-16 7:30 pm Wikipedia drops in with: http://en.wikipedia.org/wiki/GNU_Hurd To add my $0.02, I’m excited and glad to hear the exposure Hurd is getting on OSnews lately. It sounds like the average linux desktop user will not at first notice the impact of this new and forward-thinking OS, except that (here comes the list of qualifiers) in time, perhaps, if the Hurd catches on, the speed of OS development will be thrust into high gear due to the modular design of the Hurd. My guess is that Hurd will be easier for kernel developers to experiment with and extend. What we may end up here is with an open source QNX, similar in somes ways perhaps with BEOS, and able to run the whole docket of linux apps – and that’s just the start. Who knows where it’ll go from there. I’ll conclude with saying that we seem to on the cusp of some really cool stuff: Xen, Vanderpool/Pacifica, Quemu and others on the virtualization front, Xgl and mesa-solo on the next-generation accelerated graphics front, and Hurd, SkyOS, [insert fav OS here], and of course linux on the OS front. All in all the future looks quite interesting indeed. 2005-06-16 7:35 pm Of particular interest (to me at least), is the section on device drivers: http://www.l4ka.org/projects/virtualization/drivers.php 2005-06-16 7:43 pm > You can *through* your political correctness around all you like that’s coming from presumably ‘native english speaker’. 2005-06-16 9:30 pm Well there is a difference in mistakes in spelling or constructing completely weird sentences. (i’m still mind boggled about it’s meaning) Even yoda couldn’t top that first post. 2005-06-16 9:55 pm actually the SECOND post and yes i am sorry but it is TOTALLY beyond comprehension 2005-06-16 10:36 pm maybe you should just calm down, everybody we learn english by using english, so how do you expect us to learn if everybody beats us up after the first two sentence… ON TOPIC: Well concrats to the hurd devs. i always had the intention to get debian/hurd running as my desktop, and with KDE available one of my last excuses has gone. Seems like i need a new harddrive … ARGGHHHH 2005-06-16 11:35 pm being able to run such vaired complex systems as those within kde and gnome is a very good indication that HURD is coming along very very well. the critical pace has been acheived now – the only way is up. finally, an an alternative to the BSDs when it comes to “done by deisgn” 2005-06-17 12:43 am To add my $0.02, I’m excited and glad to hear the exposure Hurd is getting on OSnews lately. It sounds like the average linux desktop user will not at first notice the impact of this new and forward-thinking OS I’m confused, how is the HURD new or forward thinking? It’s older than Linux and there is certainly nothing new about microkernels. except that (here comes the list of qualifiers) in time, perhaps, if the Hurd catches on, the speed of OS development will be thrust into high gear due to the modular design of the Hurd. HURD has been modular for the past 15 years and you couldn’t call it’s development fast. I don’t know why it’ll suddenly speed up now. My guess is that Hurd will be easier for kernel developers to experiment with and extend. What we may end up here is with an open source QNX, similar in somes ways perhaps with BEOS, and able to run the whole docket of linux apps – and that’s just the start. Who knows where it’ll go from there. If it was easier for kernel developers to experiment with, they wouldn’t all be experimenting with Linux and the various BSDs. I’ll conclude with saying that we seem to on the cusp of some really cool stuff: Xen, Vanderpool/Pacifica, Quemu and others on the virtualization front, Xgl and mesa-solo on the next-generation accelerated graphics front, and Hurd, SkyOS, [insert fav OS here], and of course linux on the OS front. All in all the future looks quite interesting indeed. Sometimes I truly wonder about the people who post on this site (I’m not directing this specifically at you Howie S.). The Hurd is old. Older than Linux. Older than FreeBSD. Is there a particular reason that people get so excited and act like this is some new and upcoming kernel? 2005-06-17 2:12 am Congratulations! The L4 microkernel coupled with the advent dual and multiple core processors could potentially allow for a lot of flexiblity in implementing new initiatives as well as great optimization potential for specific needs: a la Core Audio or Core Video in Mac OS 10.4 This is a great foundation to build on. 2005-06-17 8:30 am Will GNU have drivers? alot of these drivers are for linux. and plug into the linux kernel? what of things like the wacom tablet driver for linux? will it have to be re-written? to be a gnu/hurd driver? 2005-06-17 9:00 am How exactly are you measuring age? The way I see it of the pair, FreeBSD, Linux and the HURD, the HURD is the youngest. FreeBSD has it’s roots in BSD (obviously) witch has it’s roots in Unix so it is definitely the oldest of them with the full age of Unix baring on it. The HURD was sitting a long time before any coding began waiting for a free version of Mach so you can’t count that time. And when you remove that time from the HURDS age Linux becomes older. Of course if you measure age in man-years that the programmers have put into the HURD with it’s handful of developers and Linux with it’s hundreds of developers then the HURD is still younger. By the way, the design idea of the HURD has changed a bit when the developers decided to port to L4, so that’s not really that old either. 2005-06-17 12:43 pm It is not *that* impressive considering this news is not about the L4 kernel but the Mach one (look at the result of the uname command on the KDE screenshot), which has existed for many years already. But it is still cool though. 2005-06-17 1:52 pm if we arent impressed by this then we shouldnt be impressed with some other kernel that has become popular lately… 2005-06-18 3:21 pm I want to thank you for responding to my post, and also for your “I’m not directing this specifically at you Howie S.” comment. It’s all to easy to offend (or feel offended) by comments someone posts online. I’m glad not to be in your direct line of fire, and respect your softer, more respectful approach. Your wrote: I’m confused, how is the HURD new or forward thinking? It’s older than Linux and there is certainly nothing new about microkernels. … HURD has been modular for the past 15 years and you couldn’t call it’s development fast. I don’t know why it’ll suddenly speed up now. … If it was easier for kernel developers to experiment with, they wouldn’t all be experimenting with Linux and the various BSDs. … The Hurd is old. Older than Linux. Older than FreeBSD. Is there a particular reason that people get so excited and act like this is some new and upcoming kernel? In regards to your comments about Hurd being an old project, I think johnlein comments well to the point about Hurd’s real age, and I will also add a few comments of my own. L4 was developed in 1996, which makes it certainly newer than Linux. Additionally, the transition from Mach to L4 was basically like rewriting Hurd from scratch, due to the vast differences of these two microkernels as well as to the significant about of resources previously handled in Mach that were needing to be reimplimented in Hurd. (By the way, this abstraction of functionality from the microkernel out to the userland Hurd servers was a key reason for the transition to L4 in the first place.) In regards to your comments about the current pace of Hurd development, the difference lately is that there are much more fundamentals in place. We now have Hurd servers running on L4ka – stuff like memory management, we also have a L4 port of (known as “L4Linux”) to play with, Qemu and Bochs also open up some doors, as well Michael Banck’s diary entries show GNOME and KDE progressing quite nicely. Things are definitely speeding up. I think we can conclude that Hurd development is producing more that we can “see”, and is doing so at a pace that has not been seen before. Finally, as regards Hurd’s modular design, wikinews writes “The greater modularity and abstraction of a microkernel approach means that the microkernel itself does not need constant modification as is seen in the Linux kernel today, since it provides only the very minimum of services, and does so very carefully.” (see: http://en.wikinews.org/wiki/GNU_Hurd_operating_system:_first_user_p…) Hurd’s design enables the much of the OS’s functionality to be outside the kernel, in userland, and does so in such a way that allows portions of this functionality to proceed independent of other aspects, minimizing the interference caused as these servers undergo internal changes. Well, I hope that helps clear some things up, and address some of your concerns. There is also a good article on GNU/Hurd in the British magazine Linux User and Developer (http://www.linuxuser.co.uk/) currently on the newsstands, if you are so interested.