Home > Debian > Debian GNU/FreeBSD Progress Debian GNU/FreeBSD Progress Eugenia Loli 2003-06-10 Debian 89 Comments The Debian GNU/FreeBSD port has recently made significant progress using the GNU C library as a base instead of FreeBSD’s libc. This project tries to port the Debian system, which is based on GNU, on top of the FreeBSD OS. About The Author Eugenia Loli Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker. Follow me on Twitter @EugeniaLoli 89 Comments 2003-06-10 5:40 pm Anonymous This is great news, I’d love to use Debian distribution based on Darwin with my iBook! By now I use it with the current Woody Debian GNU/Linux, but I’d like to see if the Darwin kernel can squease more power out of my (now kind of) old G3 processor! Debian sure is the best distro ever, but the PPC ports need to some work to be up with the x86 ports. And, please, sorry my poor english. I’m from Brazil, and so not a native english speaker (or writter for that matter 2003-06-10 5:41 pm Anonymous Stop tainting *BSD with GNU shi.. err crap. 🙂 2003-06-10 5:46 pm Anonymous I’m not one to berate other open source projects… But why in the h-e-double-hockey-sticks would someone want to use this? Is this just a “proof of concept” sorta thing or are there actual reasons for doing this? One of FreeBSD’s greatest strenghts IMHO is it’s userland, why would i want to replace that with the GNU userland? Kernel-wise, FreeBSD’s kernel is quite nice and more advanced than Linux in certain areas, but I wouldn’t say remarkably so. 2003-06-10 5:55 pm Anonymous Z-e-a-l-o-t-s GaNew! GaNew! Everything must be GaNew! 2003-06-10 6:00 pm Anonymous this sounds cool in theory especially for ppc. i’d love to see what this can do after a bit more development. this could be a good combo of stable debian and simple freebsd… 2003-06-10 6:04 pm Anonymous “One of FreeBSD’s greatest strenghts IMHO is it’s userland, why would i want to replace that with the GNU userland?” First, not everyone agrees with you. It’s a personal preference kind of thing. In the same way, some people might prefer FreeBSD’s IP stack to Linux’s, I don’t know. On a more practical level, people who know Debian but want to use FreeBSD instead of Linux may now do so without having to re-learn all the commands and different methodologies. The FreeBSD port will “feel” like a Debian system in every respect, with only the kernel being different. Consistency is a clear benefit here. As well as this one, there are ports in progress to bring Debian to (at least) the Hurd and NetBSD. The experience gained by this port will make porting to other kernels easier, especially NetBSD. Each port increases the level of choice for the Debian user base, and since each new kernel will have its specific strengths and characteristics, this is all to the good. I suspect there’s also a “because we can” element involved as well. I applaud their efforts, and look forward to trying out this port in the near future. 2003-06-10 6:07 pm Anonymous Why on earth would you want to run the GNU userland that comes with Linux on top of your FreeBSD kernel? You can still run all the GNU applications you like under FreeBSD. FreeBSD is not just a kernel, its strength lies in it being a complete, mature Operating System. 2003-06-10 6:09 pm Anonymous Debiand apt-get system is no way as slick or powerful as FreeBSD’s packages and ports system. This is a really really silly idea. If you want FreeBSD, get FreeBSD. if you want Debian, get Debian. 2003-06-10 6:17 pm Anonymous Debian packages undergo much more strenuous testing than the ports tree, any day. See the Unstable/Testing/Stable transition process, for one thing, the Bug Tracking System (http://bugs.debian.org) and the entire triple-tiered maintenance apparatus. The Debian packages are also far more integrated with each other and the distribution as a whole thanks to the strict adherence to Policy (http://www.debian.org/doc/debian-policy) for which FreeBSD has no equivalent. Whether or not you prefer the FreeBSD way of doing things is irrelevant; the point is that each is clearly different and each has its strengths and weaknesses. Now you can pick and choose from each one, whilst being able to run the FreeBSD kernel in either case. Sounds good to me. 2003-06-10 6:20 pm Anonymous Which is reason #1 why I use FreeBSD over Linux… I absolutely cannot stand the GNU userspace. By this we can also guess that they are replacing the heap implementation from FreeBSD’s C library with the glibc one. This implementation integrates tightly with FreeBSD’s VMM to provide the best mix of performance and security. The FreeBSD heap implementation does a great deal of consistency checking which the GNU one does not, which has kept FreeBSD from being vulnerable to a number of double free buffer overflow exploits and so forth which Linux systems have been vulnerable to, such as the double free vulnerability in zlib. 2003-06-10 6:21 pm Anonymous hey, sup with debian/netbsd that started sometime ago. It seems like it would be more useful..but really similar to the debian/freebsd. More useful because it could run on hardware that nobody has anymore 2003-06-10 6:24 pm Anonymous “Debiand apt-get system is no way as slick or powerful as FreeBSD’s packages and ports system.” Well, thank you for that unbiased, objective appraisal of the respective strengths and weaknesses of each system with regards to each other. 😛 “This is a really really silly idea.” Wow, yet another healthy dollop of measured objectivity! “If you want FreeBSD, get FreeBSD. if you want Debian, get Debian.” Or alternatively (as this porting effort allows you to do) use Debian, but with the FreeBSD kernel. Then you can mix-and-match the advantages of the FreeBSD kernel with the advantages of the Debian userspace and package system. Ain’t choice great? 2003-06-10 6:28 pm Anonymous I find it curious that high-and-mighty BSD zealots claim that BSD code is ‘freer’ than (and thus superior to) the GPL by not imposing redistribution clauses — then turn around and bitch when someone does whatever they want with some BSDed code. I won’t quote anyone specific, because I can’t be sure that those who are bitching about this are also BSD code maintainers. But the whole point of the BSD licence is to attach as few strings to the code as possible, yes? If you don’t like the BSD licence, stop using it (either as a coder or as a user/admin) and stop bitching (as a whiner). And I got news for ya: there’s already plenty of GPL code that has already “infected” your precious FreeBSD, by freebsd.org’s own hands. So long as my great aunt is still able to reliably access her e-mail, she doesn’t care. And neither do I. Regards, GG 2003-06-10 6:29 pm Anonymous http://www.debian.org/ports/netbsd/ They’re rather remiss about updating the web page, but development continues, with occasional status updates to the debian-bsd mailing list (http://lists.debian.org/debian-bsd). I agree that NetBSD would be equally cool, but in any case, the various *BSD ports should be able to share code and ideas freely with each other. They’re all sharing the same mailing list, after all 😀 2003-06-10 6:34 pm Anonymous GNU tainted or BSD? 2003-06-10 6:44 pm Anonymous Why is it that Debian needs the FreeBSD kernel? I am confused, if the GPLed OSes were so great, then why even involve yourself with FreeBSD. And if FreeBSD is so great, then why in the hell would anyone using it want a GNU userspace? It totally makes no sense. The only thing it shows is how shitty linux is, and how shitty these people want the BSDs to be. I’ll take my BSD with its regular userland. 2003-06-10 6:49 pm Anonymous The only new component in this port is the kernel. The userland will stay as whatever license it already is (it’s not all “GNU-tainted” [great, another BSD zealot], it can be any license that meets the Debian Free Software Guidelines (http://www.debian.org/social_contract.html). They’re not going to relicense the FreeBSD kernel – it’s already DFSG-Free, and what would be gained by it? The *changes* that Debian incorporates into the FreeBSD kernel once its in the Debian package system are another matter, e.g. bugfixes are another matter. The subject of what license they should be under was covered on the debian-legal list long ago, see http://lists.debian.org/debian-devel/1999/debian-devel-199902/msg01…. Effectively, it’s entirely up to whoever writes those changes to release them under whatever license he chooses, but, and I quote from the above debian-legal message, “the courteous thing to do is to honor the upstream author’s choice of license.” So I’d imagine they’d stick with the BSD license for the changes, because a) that way they’d be contributing back to the FreeBSD community and they don’t want to fork for political reasons and b) code management is so much easier without merging different licenses with different requirements into the mix. 2003-06-10 7:12 pm Anonymous When do i get to have debian on top of the NT kernel? I’d really like to use a userspace that’s constantly out of date with the ability to have MS security holes. Can get that too? 2003-06-10 7:17 pm Anonymous LinBSD was started to port the NetBSD userland to the Linux kernel 2003-06-10 7:21 pm Anonymous And “The most prodigiously wasted open source effort- Award” goes to Debian! for their work on porting GNU userland on FreeBSD. ewk. BSD userland is one of the most pleasurable thing about any of the BSDs… 2003-06-10 7:24 pm Anonymous No, because the NT kernel is not DFSG-free (http://www.debian.org/social_contract.html)… 😉 2003-06-10 7:27 pm Anonymous I would definitely waste my time porting my OS userland to a superior OS userland. No offense, but aren’t there better things the Debian project could do with their manpower? Like possibly make Debian a competitive distro again? No offense but I cringe at the thought of running something called “unstable” and the Debian-based OSes don’t count b/c they are NOT pure Debian itself. And the installer!!! *SHAKES* please, instead of galevanting about with useless porting endeavours such as GNU/Hurd, NetBSD, and now FreeBSD, why don’t you go do something useful like make a decent friggin installer for Debian 2003-06-10 7:33 pm Anonymous When I first found out about this I realy wanted to go the other way with it… that is a freebsd userland runing on a linux kernal… is there any one else out there that feels this would be a much better mix than what debian is doing with freebsd/netbsd now? 2003-06-10 7:35 pm Anonymous “And the installer!!! *SHAKES* please, instead of galevanting about with useless porting endeavours such as GNU/Hurd, NetBSD, and now FreeBSD, why don’t you go do something useful like make a decent friggin installer for Debian” Because if you made things better and/or easier, then you would be competing with mainstream distributions, more people would start using your distribution and then you wouldn’t be near as “cool” anymore. 2003-06-10 7:39 pm Anonymous understaffed??? How bout you read up on that before making such a claim. Dont like the word “unstable”? Then run Debian Sarge, i dont see the word “unstable”, and its still “pure debian”, but with more recent packages. Want even more recent, run Debian Sid. Debian has the nicest installer i’ve ever used. If you dont like it, dont use it. But it’s by FAR the most powerful i’ve ever used. I’m managed to install Debian on boxes that would refuse to install any other linux distro. The only part that i find leaves to be desired is having to pick your modules, but i generally dont care, just insert the module for my network card, and end up building myself the latest kernel from kernel.org within the hour. As for Debian-based OSes not counting cause they aren’t “pure”… Thats what i thought until i started playing around with Libranet on a friend’s box. It IS debian, but with a bit of sugarcoating. 2003-06-10 7:40 pm Anonymous I’d have to agree with ya…porting the BSD userland to Linux would be kinda nice for the hardware that FreeBSD doesn’t support yet, however I’d say it would be a far more difficult road. The GNU userland was designed to be somewhat portable and is not nearly as kernel-tied as the BSD userlands are. However I’m sure such a port COULD be done if there was reasonable demand for it. I wonder how many API calls would have to be implemented in the Linux kernel for such a project and how many lines of code would have to be changed in the BSD userland. Interesting to say the least. 2003-06-10 7:44 pm Anonymous ROFL! This comes up every single time… Debian is a volunteer distribution. Broadly speaking (and this is of course an oversimplification, because Debian is such a complex entity) people work on whatever they feel like working on, and for many different reasons. You cannot tell them what to do; or rather, you can rage impotently and make an arse of yourself as you are now, but you’re just pissing into the wind. And that’s the way it should be – who are you to decide what any volunteer should do with his time? “why don’t you go do something useful like make a decent friggin installer for Debian” Since Debian is made up of volunteers, ordinary people just like you (well no, that’s not fair on them, but you get my point) I could equally point the finger at you and say “Why don’t YOU do it?” There are people working on it, and like the rest of Debian, it’ll be released when it’s ready. Of course, if you help them rather than attack them it’ll be done that much sooner. It’s an entirely open process – just start coding and contributing. “aren’t there better things the Debian project could do with their manpower” LMAO! Just who, exactly, is it that you think is in a position to co-ordinate the expenditure of manpower? 2003-06-10 7:44 pm Anonymous I just see the point for anything other than gee whiz factor. Then again gee whiz factor goes a long way in the tech world. 2003-06-10 7:47 pm Anonymous Because if you made things better and/or easier, then you would be competing with mainstream distributions, more people would start using your distribution and then you wouldn’t be near as “cool” anymore. Just continue to run your distro nearly untested packages, where the packages are not made to integrate properly into the system. But hey, you’re running newer software, and thats ALL that matters. I don’t know anyone who runs debian to be cool, it just WORKS better than most distros. I dont care for GNOME2, nor for KDE3, they look purdy, but they’re not for me. So i’ll stick with my “outdated” software that i can trust. 2003-06-10 7:50 pm Anonymous Funny…the BSDs seem to do a damn good job coordinating man-power for being a large complex volunteer driven organization. I don’t see them wasting their efforts on useless endeavours, like porting their userland to a different kernel, instead, if they see something they like from a different kernel, they add it to their own friggin kernel. And I’m all of a sudden ignorant for suggesting that what this team of people is doing is a waste of time? Guess what, that is my OPINION, and I am entitled to it. This is what I would like to call constructive criticism. Thank you, please accept my comments for what they are, my opinions which you may or may not take with a grain of salt. 2003-06-10 7:52 pm Anonymous “This is what I would like to call constructive criticism.” Eh? You think that you ranting “This is utterly pointless. What bollocks. What a total waste of time!” is constructive? What does it accomplish other than venting your spleen? Constructive criticism is when you identify a problem and help solve it, or offer a helpful suggestion. Utter condemnation does not qualify. 2003-06-10 7:59 pm Anonymous Just to respond to the anonymous troll: “Funny…the BSDs seem to do a damn good job coordinating man-power for being a large complex volunteer driven organization.” And yet there are areas where they lag appreciably behind, e.g. the hardware support is not nearly as good as Linux. Let me mirror you by saying: I THINK THEY SHOULD HAVE WORKED ON THIS INSTEAD OF [INSERT FEATURE HERE] THAT I THINK IS CRAP. WAAAAH! WAAAAH! WAAAAH! See, I can be “constructive” too! 😛 Now, back to a mature discussion: No volunteer project can exercise a great amount of control over its developers. Individual BSD developers, as with Debian developers, identify problems that they would personally like to solve for whatever reason and go to it. There is no “central command”. No-one *orders* them or gets to set their priorities for them, and if they did, they would tell them to f*** off. And rightly so. 2003-06-10 8:01 pm Anonymous So because software is newer it’s automatically less stable? ROFL…let’s all go back to DOS 1.0 because it never crashes…LOL I understand the need for testing and integration, but FreeBSD does a more than adequate job of this. Their base install, even on their 5.x branch is the most stable base I have ever seen. And they still manage to include up-to-date components. A good deal of testing IMHO should fall on the package maintainers instead and leave the distro’s to test their base installs instead. Also, if a package is found to be broken in the ports tree, it is immediately marked as such and you have to manually enable it before you can build it. And as for integration, the BSDs are much more integrated than any Linux distro could ever dream of being, because the kernel and userland are maintained by a single project. Just because Debian doesn’t test a package, does not leave that package “untested” -bytes256 2003-06-10 8:09 pm Anonymous “if a package is found to be broken in the ports tree, it is immediately marked as such” Rubbish! 😀 That depends entirely on the popularity/importance of the port in question. “So because software is newer it’s automatically less stable?” Yes, in the sense that no package in Stable is allowed to have a bug that’s severe enough to be considered “Release Critical” (the highest severity). This takes a good deal of testing to achieve, but the downside is a slow release cycle. I personally think it’s worth it. Those who agree are the ones that use Debian – those who don’t, don’t. It’s not “automatically” less stable if it’s newer, but Debian’s view is that it must be subjected to intensive testing to ensure that this is the case. 2003-06-10 8:15 pm Anonymous even though im not involved in coding for any of the projects… i personally love the bsd userland… and i love the bsd licence… that said… one of the reasons i do love the bsd license is that except for taking credit for code which isnt yours you can more or less do what you want with it… whether you agree with debian’s decisions or not here… the way they are handling it is well within their rights… there are places to defend the bsd way… but complaining here isnt really helping anything.. just making us look as nutty as some of the more rabid gpl-ites 2003-06-10 8:21 pm Anonymous All I can visualize in my head is Beastie giving it to the penguin from behind…. It would seem to me that once this was done, it would no longer be linux at all considering linux is just the kernel whereas *BSD is the whole OS (kernel+userland)…. 2003-06-10 8:25 pm Anonymous So because software is newer it’s automatically less stable? ROFL…let’s all go back to DOS 1.0 because it never crashes…LOL Heh, way to try to undermind my point. Seriously though… what software should i trust more? The one that has been proven to work time and time again, or the one that just came out? Why should i be an un-willing betatester? I use my computer to get work done, applications crashing isnt something i tolerate. I’m willing to wait for software to get into Debian Stable, and I think it’s worth it. There are SOME apps that i grab from Sarge, but very few. I have no huge problem with the BSDs, I dont bash them, and i have quite a bit of respect for them. When i compare debian to ‘distros’, i mean linux distributions. 2003-06-10 8:29 pm Anonymous Lol, that’s a hilarious image. 😀 “It would seem to me that once this was done, it would no longer be linux at all considering” You’re spot-on correct. It’ll still be *Debian*, but no longer Linux – Debian GNU/FreeBSD instead of Debian GNU/Linux. Debian hopes to become kernel-agnostic whilst still retaining consistency within itself – thus the use of the Debian userland throughout. 2003-06-10 8:33 pm Anonymous It seems that some people here are under the impression that the Debian GNU/*BSD projects are major undertakings initiated by the Debian project and are taking away precious resources from more pressing work. They’re not. All projects (counting the discontinued Debian GNU/OpenBSD effort) were each initiated by a single individual under the Debian umbrella and the number of significant contributors can be counted on five fingers, tops. Just take a look at the mailinglist archive for the BSD porting effort: http://lists.debian.org/debian-bsd/ Not a whole lot going on. And yeah, I think they’re dumb projects, but as neither has any impact on Debian GNU/Linux, I really couldn’t care less. People are free to work on whatever they want to. 2003-06-10 8:35 pm Anonymous So because software is newer it’s automatically less stable? Of course not. The distinction is tested known-to-work software versus untested ought-to-work software. 2003-06-10 8:45 pm Anonymous I don’t see how this can be a bad thing? Debian is simply trying to step away from dependance on the Linux kernel. They will still depend on GNU, but thats all. Any kernel that complies with the DFSG will likely at least be attempted, and as such they are currently working on NetBSD, FreeBSD, and GNU HURD as replacements for Linux. The plan as far as I can tell, it to make the kernel as insignificant as any other single package. They are supposed to be able to be drop-in replacements at some point down the road. I can’t imagine anything better! I like FreeBSD, I like NetBSD, I like Linux, and I like GNU HURD in theory. One day, through Debian, I will be able to mix and match as I please. After all, whats truely the difference between the 4 operating systems? The userland, and my favorate userland is Debian. I can think of infinite reasons why porting Debian to different kernels is a good idea, I can’t think of any reason why its a bad idea. I say, thankyou so very much to each of the groups developing these ports, I can’t wait to reak the benifits, keep up the great work 2003-06-10 8:59 pm Anonymous I agree with you about the manpower effort, but I have to question: “but as neither has any impact on Debian GNU/Linux” Of course it does. 🙂 Quite apart from anything else, once the first non-Linux port is in a reasonably-completed state, it won’t *be* Debian GNU/Linux any more. The home page et al will be amended to refer to just “Debian” instead. The implications of the FreeBSD port *itself* may not be that significant (though it’s really to early to pass judgement) but Debian’s becoming kernel-agnostic is surely a pretty important moment in its evolution. It’ll be interesting to see what comes of it. 2003-06-10 9:11 pm Anonymous the structure of osnews comments lends itself to people not reading other’s comments. 2003-06-10 9:23 pm Anonymous Yes. 🙁 A threading option would make it easy to see which comments are in response to what, for one thing. It’s easy to miss important messages when you’re reading quickly. The lack of a preview or edit facility is also a problem, because one has to manually proof the message giving someone else more time to post something similar to what you’re writing in the meantime. Lol, and of course there’s always the problem that you walk away from your PC for a couple of minutes to take a leak or something, finish your half-completed message, post it, to discover to your horror that someone has posted something very similar in the interim. 2003-06-10 9:35 pm Anonymous Debian has many problems even with the development of its main branches; like the awfully slow release cycle. Shouldn’t the project try to concentrate the development efforts on relevant things more instead of dispersing the focus all the more to dozens of subprojects àla Debian/this&that? Well, the Debian/FreeBSD may actually be one of the more reasonable subprojects, but what’s the point of having an official list like Debian/Win32 (lists.debian.org/debian-win32/)? Obviously odd ideas like Debian/Win32 just bring enough joy to the geeks themselves so that they are even officially embraced by the project… If yet another subproject of Debian is really needed, why not have something more useful like Debian/i686 instead? And when someone makes perfectly reasonable and very well-argumented suggestions to essentially improve Debian in its main functionality as a server OS, the suggestions may just be neglegted by the project elite: Adamantix/Trusted Debian (http://www.trusteddebian.org/). So, Debian/Win32 is important, but Trusted Debian aka Adamantix is not…?? 2003-06-10 10:36 pm Anonymous First off, I don’t know why the Adamantix guy is pissed that he had to remove “Debian” from the name of his project. Debian’s worried that that would set a precedent so you’d have some commercial vendor selling something called “Debian Improved” or whatever, which they couldn’t defend against because they could be shown to not have defended their trademark. Second, he claims that he’s been ignored. That’s rich, since I’ve spoken to the guy and he admits he never even broached the subject on -devel, just emailed some random people in the project who certainly aren’t in charge. I spoke to him about it on irc, and he said he’d think about mentioning it on -devel. Nothing yet. To be honest (though the Adamantix devs haven’t been smart enough to work this out, apparently) each and every component of Adamantix has been proposed for inclusion in Debian individually in the past. Search the archives at http://lists.debian.org/google.html for yourself. RSBAC support is possible to add, certainly, and anyway, SELinux is being integrated already. But as for the kernel patches, they’d involve recompilation of the entire archive, as well as porting to the other architectures – Adamantix is x86-only. That’s an absolutely staggering effort. Again, note how Adamantix only supports a tiny fraction of Debian’s packages, and only on one architecture. Even if it were to be taken on board, would not be undertaken until after the next release is out. Second, Debian is a general-purpose OS, and the consensus view is currently that those who aren’t in high-security environments would not be willing to take the performance hit entailed as well as the other problems (e.g. trouble with XFree86 is mentioned in the Adamantix FAQ – I’m sure the Desktop sub-project would be really chuffed at *that*). Conclusion: Integration is just not practical at the current time. If you disagree, please raise it on debian-devel so that they can convince you, as the Adamantix devs can’t be bothered to, instead preferring to snipe at Debian on their homepage (re: the name change) and in their FAQ without even coming to -devel to discuss the issue or (apparently) to search the archives for past related discussions. 2003-06-10 10:47 pm Anonymous “what’s the point of having an official list like Debian/Win32 (lists.debian.org/debian-win32/)? Obviously odd ideas like Debian/Win32 just bring enough joy to the geeks themselves so that they are even officially embraced by the project…” I personally think that having the Debian package management system and the Debian package archive running without a hitch in Cygwin would be fantastic. It’d make running Free Software on Windows so much easier, thus promoting the cause of Free Software. Sounds pretty simple to me. If you think they should spend their time elsewhere, go raise it on said mailing list and I’m pretty sure they’ll convince you of the merits of the project. “why not have something more useful like Debian/i686 instead?” This has been brought up on the lists countless times. Please check http://lists.debian.org/google.html for yourself. Essentially, they feel that creating a whole new port when the i386 port already runs just fine would be archive bloat. Especially since it’s only worthwhile recompiling major packages like X, Glibc, KDE, major apps like that. And with the apt-build tool, it’s now easy to do just that – targetted recompilation of whatever you select with whatever optimizations you desire. A reasonable compromise which allows you to gain i686 performance on your system without having to maintain a whole new port for i686. And i586. And… You get the picture, I’m sure. 2003-06-10 11:29 pm Anonymous Good points, thanks for correcting me on those details. Adamantix and 1686 port were just examples though. But bringing even more complicated issues to Debian would really be a bad idea, true. My main point was that, IMHO the problem of Debian seems to be the lack of focus, what so often troubles the open source/free software world. Maybe Debian may already have too many subprojects, maybe even too many supported architectures, so that the project, for example, could even keep up with the release cycle of its competitors (= MOST other GNU/Linux distributions). A new Debian release tends to be old already when born (and testing/unstable are NOT officially supported nor recommened to be used). That is one big reason why Debian is really more suitable to be used as a server OS than as a desktop OS. GNU/win32: “It’d make running Free Software on Windows so much easier, thus promoting the cause of Free Software.” Why not first concentrate on solving the big problems of the main product (Debian??) first, like the awfully long release cycle of Debian – before trying to conquer the whole world outside? As to Adamantix formerly Trusted Debian: they claim that the name change “was demanded by the Debian project, who claim Debian is a trademark”. 2003-06-11 12:30 am Anonymous Your points are likewise good. The long release cycle is undeniably one of the negative aspects of Debian’s focus on a tried-and-true, well-tested, well-integrated distro. That said, I don’t think it’s possible to bring the release cycle in line with other distributions. It ought to be possible to keep Unstable reasonably up-to-date with current software (as it is now), and to make a new Stable release every one to one-and-a-half years. Whilst I agree that this is a potential problem, the release cycle is one of the areas of Debian that undergoes regular overhaul; Woody was the first release made using the new Testing distribution, for example. Check out the Release Manager’s “Woody retrospective and Sarge introspective” which he made shortly after Woody’s release: http://lists.debian.org/debian-project/2002/debian-project-200207/m…. “Why not first concentrate on solving the big problems of the main product (Debian??) first, like the awfully long release cycle of Debian – before trying to conquer the whole world outside?” Personally, I agree with you in that I do think the release process is more important, though I still think Debian Cygwin would be wonderful. That said, the developers working on -win32 obviously don’t, and their opinion is the only one that matters since they’re volunteering to do the work. It’s also important to remember that quite a few developers don’t think about the project globally, or even act in unison – they might use Debian because its resources and infrastructure make the project or task they have in mind so much easier. That’s the reason lots of maintainers maintain their respective packages, for example. Anyway, each one will have a different motivation for his contribution, and each one a different rationale. This is just the way it is. I don’t think this is a problem that can be solved in any way. Besides, were anyone to show that -win32 were actively *slowing* the release process, it would probably be dropped. No one has, most likely because it’s not. “(and testing/unstable are NOT officially supported nor recommened to be used).” I don’t know what you mean by “not officially supported” – they’re part of Debian, they’re just constantly in flux. Unstable is where the development takes place, and Testing is Unstable apps ~10 days after they remain with no Release-Critical bugs. They each have their respective problems – Testing has security bugs (it’s at least ten days behind unstable, remember, and this *includes* security patches) and Unstable, whilst it doesn’t break too often, requires expertise to fix when it does. Neither Testing nor Unstable are supported by the Security Team, but the developers usually fix their own respective packages in Unstable pretty promptly when problems show up. But there’s hope on the horizon of getting Testing usable as a day-to-day distribution. Several people expressed an interest on -devel about maintaining security updates for Testing. This could be one solution to the problem of the slow release cycle – use Testing in the interim if you think that Stable is too old and the new functionality is more important than the stability. “Debian is really more suitable to be used as a server OS than as a desktop OS.” Perhaps, though the Desktop subproject is making headway here. I confess that I run Unstable on my own desktop, which probably says a lot. But for the *corporate* desktop where stability and ease of support reign supreme, and the admins maintain the systems, I’d say Debian Stable is the ideal choice. I personally think the best bet is to get security updates going for Testing, so people can use that with confidence on their machines. Aside from the security aspect, it’s certainly stable enough for desktop use, and it’s at least more up-to-date than Stable. If anyone wants to help with that, now would be the time to volunteer; as I say, it only recently was mentioned on -devel. “As to Adamantix formerly Trusted Debian: they claim that the name change “was demanded by the Debian project, who claim Debian is a trademark”.” I did a bit of digging in the archives, and Debian *is* a trademark. Check here for the initial announcement from way back in ’98: http://www.debian.org/News/1998/19980306a. I suspect (hope!) that Debian would have phrased it as a request rather than a demand. The FAQ shows the Adamantix devs’ venom towards Debian for “ignoring them” (again, they didn’t even bother posting to -devel, so this argument is hogwash), so I don’t trust them to be objective. If not, it’s rather an extreme reaction, but Debian are within their rights to do it. If one were to enquire on -legal, I’m sure a fuller explanation would be forthcoming – I’m only guessing here. 2003-06-11 12:46 am Anonymous Lol… Jesus, that was a long last post. OSNews: please implement a preview function. Writing that was torture, though maybe I should’ve written it in an editor and pasted it into the window afterwards. @Metic: Sorry, I forgot to address your point about whether there are too many architectures. Check out this thread: http://lists.debian.org/debian-user/2002/debian-user-200206/msg0030… which asks that very question, as well as whether time would not be better spent elsewhere, on the mainstream architectures. As one might suspect about such a loaded question, it spawned a tremendous discussion. In an *extremely* oversimplified nutshell: the consensus view was “No” in response to both questions, on the grounds that Debian should not leave the other archictures (together with their existing userbase) out in the cold (indeed, Debian is for a fair few architectures the only main-stream distribution that supports them) and that in any case it has yet to be proven that the release process would be much faster even if it were only x86 supported. Despite having the largest userbase, the x86 port has just as many problems/hiccups as the other architectures. If anyone’s interested, go read the thread. It’s very interesting. 🙂 2003-06-11 2:10 am Anonymous freebsd is great . imho freebsd performs excellent for servers but linux is where the action is . more developers are working on linux . that’s why i like it, specifically gentoo. and btw fyi gentoo’s portage is a hybrid of bsd ports and debian packaging system 2003-06-11 3:14 am Anonymous I use Debian/alpha on a daily basis as my main box is a DEC Alpha box, and i really dont think i’d be able to run that box if Debian dropped support for Alphas. I’ve tried the other free OSes for Alphas out there, like Gentoo and the BSDs, and I found that Debian ran best. The 2nd closest was FreeBSD, which i think would have ran alright, if only my ISP didnt bastardize ftp access. Supporting more architectures doesnt slow down anything. It’s different people who support them. Maybe those people wouldnt be on the Debian team if they couldn’t do that. I was never on any debian mailing list until i got my alpha, when i joined the debian-alpha ml. There’s a good number of very qualified people that work on the alpha port. I love knowing that Debian will run on pretty much any computer that i’d want to buy, even an ipaq. 2003-06-11 4:00 am Anonymous Look, this GNU/FreeBSD product does 2 concrete REAL things: 1) It gives the GNU/Linux croud somewhere to flee to away from SCO. Want to take the wind right out of SCO’s sails? Stop using the Linux kernel. If there is no one using it, and no money in it, SCO will go away even faster. Rather than whine, pack up and move. On your way out, be sure to salt the earth so SCO will find the soil unable to grow the money tree. 2) Once and for all, the debates about which kernel is faster and more stable can now be answered :=) To the FreeBSD is better cabal….don’t scream it too loud or the GNU/Linux rabble will all want some 😉 The journey to FreeBSD takes steps, and if going from GNU/Linux -> FreeBSD takes GNU/Linux -> GNU/FreeBSD -> FreeBSD, so mote it be. 2003-06-11 4:21 am Anonymous Debian has trademarked “Debian”. There for only projects officially endorsed should hold that name. Debian is based on the beleifs of the DFSG. The only reason for the “GNU” part in there name is that they were once the official GNU operating system (and thus the HURD port is perhaps the most advanced version of the HURD). Debian doesn’t want to depend on Linux. Debian is not Linux, its a system for managing packages. Linux is a package, FreeBSD is too, and so is NetBSD, and the HURD. DFSG (Debian Free Software Guidelines, also the basis for the simular article relating to Free Software Foundation) outlines the packages that can be accepted as part of the “main” debian system. They also permit certain other packages available in “contrib” and “non-free”. However the system will never depend on any such package etc. As for all those saying “Debian doesn’t support testing and sid”, well it doesn’t really support woody either, not in the way many are suggesting. Woody gets security patches, but sid and testing do to, they just aren’t advertised. See the stable guidelines dictate that no new software be added to the stable branch once it is declared stable. It is the reason for backports, and its probably part of the reason woody is the first to have 3 branches (potato only had stable and unstable). Debian Sid is far less unstable then Red Hat, SuSE, or any other distro I have tried however. Yes, Sid is basically alpha, and sarge is basically beta. This doesn’t make Debian outdated though, it simply means you KNOW you can depend on stable, and you know sarge will usually be fairly stable also. You also know if you don’t pay attention to the newsgroups, sid will likely break once a week. My Sid box has been up for 9 months now however, and is still performing well. It has X 4.3, every package is as new as any other distro. I fail to see how having the newest software could possibly mean “outdated”. In comparison, Red Hat, SuSE and many others rush there distros out, and FreeBSD didn’t release a thing for over 2 years. I think many people need to take a look at the Debian list of tools. After all, thats all Debian is, its a collection of tools that handle fetching, installing, and configuring packages. Its a distro, not an OS. The Linux people do a fine job at writting kernels, as do many other groups, Debian just allows you to use the best tools available for using them! And hell, if you like a tool, and its not available at an apt source, package it and upload it… thats all the other 900 maintainers did. (of course its a little harder to do this for other branches though 2003-06-11 7:13 am Anonymous Are they really *that* different? Linux has at least 10x the manpower of the BSD’s combined, so why not port the GNU userland to the BSD world? Does FreeBSD or NetBSD really have enough Developers to port the BSD userlands to Linux? Ok, I thought not. I will say this, which I was taught a mere 3 years ago, the BSD’s attract more mature programmers, while Linux attracts more programmers. Many BSD’s spew the lines that their code is more mature or that it is more robust, how is this possible without the programmers being better? The core of the matter is that they are both *OPEN* Unix Derivatives/Clones. *PERIOD*. They both try to accomplish the same things. Linux has more programmers, any BSD advocate that disagrees is foolish. Any BSD’er that asks how or why (they end up being the same question don’t they?) is being negligent. So the question to me at least is why not? Why limit a system to one kernel, stagnation == death. 2003-06-11 7:20 am Anonymous A NetBSD port would be great, as so far the NetBSD kernel autodetected every piece of hardware correctly, unlike FreeBSD or Linux. The problem with NetBSD is that there are not so many packages available, and that their versions are quite old. KDE 3.0.x is not that up-to-date. If with Debian more packages come available, that would only be great. 2003-06-11 7:32 am Anonymous Want debian with the power of FreeBSD? Get GENTOO :-)) http://www.gentoo.org 2003-06-11 7:46 am Anonymous Want debian with the power of FreeBSD? Get GENTOO :-)) Gentoo users has got to be the most clueless people around. 2003-06-11 8:28 am Anonymous This is one of those projects that will lead to nothing because it simply adds nothing. Why would I want to swap my stable and efficient FreeBSD, something that took years of fine-tuning and polishing, for this untested and complicated half breed ? I would never even consider it. And neither would any serious power user (like yahoo.com ) . If your really would like to try something outrageous, go port Debian userland on a PDP-11 running Multics. 2003-06-11 9:33 am Anonymous Umm, Debian is being ported to FreeBSD (hence Debian GNU/FreeBSD)… Debian is available for FreeBSD (see linux_base-debian). I fail to see how this is a “verses” discussion? 2003-06-11 10:31 am Anonymous beeing a newb, and with my english not beeing all that good at all, i was wondering what do you guys mean by “userland.” can someone explain this to me? 2003-06-11 10:53 am Anonymous DEFINITION of ‘userland’ From Jargon File (4.3.3, 20 Sep 2002) userland n. Anywhere outside the kernel. “That code belongs in userland.” This term has been in common use among Unix kernel hackers since at least 1985, and may have have originated in that community.The earliest sighting was reported from the usenet group net.unix-wizards. 2003-06-11 11:12 am Anonymous (this is to *zealots*, not all BSD users. Unfortunately, the majority of posts by opponents of this seem to be of the typical “Screw the GPL and Linux” troll variety.) ATTENTION! This is not geared towards you. The Linux community has known for a long time already that the mention of any project seeking to combine the two systems will be met with rabid howling on your part. We know you think yourselves ultra 31337. Don’t worry, this isn’t being done to take away your sense of self-satisfaction. This is a project for people with open minds. There are a number of comparisons which show that BSD is better in some areas than Linux (native CD-burning, for instance). Many people, believe it or not, consider the APT system to be superior to ports and pkgs. They also do not want the overhead of learning a whole new UNIX, but do want the added stability. This is why this project is created. Free Software/Open Source is all about choice. Don’t tell these developers to fuck off because you don’t like GNU. They like GNU *and* BSD, and being open-minded, do not see the problem with an effort such as this. You are stuck in a narrow corridor of thinking, like a small child always needs one thing to the utter exclusion of the other. Respect people’s choices. When was the last time you went to a Windows forum and heard, “You’re using Win2k? That means you are completely inferior to me and my XP. Don’t bother having children, you pathetic slug.” ? Syntaxis, thank you. 2003-06-11 1:29 pm Anonymous Porting Debian userland to run on the FreeBSD kernel. At first, I could not imagine why one would want to do this or having done it who would use the resulting system. After a few minutes of reading comments here at osnews.com, it seems that the unstated reason for the port is that there is a population of Debian users/developers that like Debian and are dissatisified with Linux (referring to the kernel) to the point that they will spend their time and effort porting the Debian/GNU userland off of Linux (referring to the kernel) and onto the FreeBSD kernel. Wouldnt the time and effort be better spent making Linux (referring to the kernel) good enough that people would not need to port Debian/GNU to run on the FreeBSD kernel? 2003-06-11 1:42 pm Anonymous Free Software/Open Source is all about choice. Don’t tell these developers to fuck off because you don’t like GNU. And telling them to fuck off is free speech my dear friend. 2003-06-11 1:45 pm Anonymous I would say, most of these posts are not so much “against” Debian, it’s just somewhat out of shock that someone would want to create/use this project. Not to speak for everyone in this thread, but I think most of us “31337 BSDers” (btw i’d say from my experience that Debian people tend to have this mentality more than BSD people do, but i’m not gonna go there) are just wondering why? What is the motivation? There is the obvious geek mantra “because we can” and the “you ask me why? i ask you why not?” Well, those are obvious motivations for any project, the question for us is, why is this deemed interesting/necessary? I would say many of these posts are not out-right attacks, just interrogations, and they continue because we haven’t heard any new responses to these basic obvious questions. -bytes256 2003-06-11 2:04 pm Anonymous “they continue because we haven’t heard any new responses to these basic obvious questions.” How about addressing all the responses that have been made already, rather than just asking the question again. “Why” and derivatives have been asked many times in this thread; people, including myself, have answered, and their responses have been largely ignored. “it’s just somewhat out of shock that someone would want to create/use this project.” Ummmmm… okay. I think that says it all, really. If this “shocks” you, you really need to get a life; I think that’s something of an over-reaction, to say the least. What’s shocking about it? The fact that someone could think that the Debian userland is better? “I would say many of these posts are not out-right attacks, just interrogations” Most of them aren’t even posing questions, just trolls posting variants of “This is a complete and utter waste of time.” Moreover, not one of them backed up their statement beyond something like “because the BSD userland is so much more cool.” That’s not debate by any stretch of the imagination, that’s sheer zealotry. “I think most of us “31337 BSDers” (btw i’d say from my experience that Debian people tend to have this mentality more than BSD people do” Who cares? Either way, I hope we can agree that no matter what “side” they think they’re on, such rabid, close-minded zealots contribute nothing useful to a discussion. 2003-06-11 2:05 pm Anonymous Then telling YOU to fuck off for doing so is free speech by the same argument, and is thus perfectly acceptable. 🙂 Sorry, you can’t have it both ways! 2003-06-11 2:24 pm Anonymous “it’s just somewhat out of shock that someone would want to create/use this project.” Ummmmm… okay. I think that says it all, really. If this “shocks” you, you really need to get a life; I think that’s something of an over-reaction, to say the least. What’s shocking about it? The fact that someone could think that the Debian userland is better? Umm no offense, well actually yes offense…who’s the one that needs to get a life, what percentage of these comments on this thread are by you? Guess what, we all don’t have an undying love for Debian, I have many strikes against it personally, which is my right, and I don’t ask everyone to love FreeBSD in an undying way. And btw, the questions continue to be asked because the questions continue to be poorly or completely un-answered. I still have yet to understand what this truly achieves, all of the GNU userland will run with few if any modifications under FreeBSD as is, in fact to compile much of anything out of the ports tree, you have to have a considerable chunk of it on your system. For example, I know of bash, gm4, gtar, gmake, and gcc just to name a few. You can always build the important parts of the GNU toolchain with a “g” prefix and use aliases in your .bashrc if you want a GNU userland under FreeBSD, and what does apt-get bring to FreeBSD? pkg_add -r isn’t just as easy to use or how about the beautiful ports system? I just don’t understand what Debian has to offer FreeBSD that isn’t already available. Like a previous poster said, I see some things that the Debian project could gain from the FreeBSD userland. If you can answer me that I will shutup and acknowledge that this is a valuable project. However, aside from the initial Gee-Whiz factor, I see this as being no more useful than being able to run NetBSD on a Sega DreamCast. -bytes256 2003-06-11 2:54 pm Anonymous Debian’s APT has a cool GUI front-end called Synaptic. I’d like to see something similar in FreeBSD. 2003-06-11 2:57 pm Anonymous Hey…that is a pretty good idea 😀 I’m surprised nobody’s made one yet…hmmmm. Are you thinking for both packages and ports? Like integration with portupgrade and the like? 2003-06-11 2:58 pm Anonymous “Umm no offense, well actually yes offense…who’s the one that needs to get a life, what percentage of these comments on this thread are by you?” Ooh, touche! 😀 Lol – in all seriousness, I retract that part of my statement, that was unfair. But, to elaborate on my stance a bit more: I just don’t find it useful when people make statements like “I find it shocking that one could want to do this.” To my mind (and yes, I know this is not a perfect analogy) it’s rather like someone saying that they find it shocking that someone likes strawberry ice-cream rather than vanilla, which is their *own* favourite. There’s no law against it, but it would certainly be a strange thing to say. Rather a pointless contribution, as well – it’s voicing an opinion, but not actually adding anything useful – it leads to no substantive further discussion, it’s a dead-end, and it *cannot* be rebutted because it consists wholly of opinion. Besides, I don’t love Debian unconditionally. If you’d bothered to actually read the comments I wrote, you’d see that I myself elaborated on a few problems that have to be faced, e.g. the *slow* release cycle. I do, however, love the *consistency* of the distribution. “what does apt-get bring to FreeBSD?” and “I just don’t understand what Debian has to offer FreeBSD” You’re looking at it from the point of view of FreeBSD. Why? This is a *Debian* project, not a FreeBSD project. FreeBSD isn’t even participating. Surely it’s obvious, therefore, that the focus of it is going to be Debian? Debian’s doing this primarily because they think that the FreeBSD kernel will add something to the distribution. Undeniably, it will. Debian will become kernel-agnostic, a pretty big milestone as I said in a previous comment, and in addition it will gain a damn fine kernel to offer alongside Linux. As a side benefit, the FreeBSD kernel will probably gain some more exposure in the server space. Undeniably it will be installed on more systems than it would have otherwise, and this is a gain to FreeBSD: additional testing of one of its core components. Also, it’s exceedingly valuable if you already know how to administrate a Debian Linux machine; when this port is complete, you’ll automatically know how to administrate a Debian FreeBSD machine. Wahey, 2 OSes for the price of one and all that… 😉 This is the benefit of userland consistency throughout the distribution, irrespective of what kernel resides at the core. “I see some things that the Debian project could gain from the FreeBSD userland.” I do, too. Undeniably, there are some things that Debian could gain from the FreeBSD userland. And vice-versa, too. As I’ve said before in this thread, both the FreeBSD and Debian userlands have their own respective strengths and weaknesses. That still doesn’t diminish the fact that the FreeBSD *kernel* also has something to offer to Debian, and that’s why they’re doing this port. Your argument that bringing the kernel into the distribution is worthless because bringing the userland in would be even better (highly debatable, especially from the point of view of existing Debian-using admins, again read my point about consistency) is logically unsound. 2003-06-11 3:13 pm Anonymous Sorry I was being so thick-headed earlier. I now understand the purpose of this exercise. I retract any statements I may have made that were personal attacks on you or Debian, or this particular sub-project of Debian. You’re right, I was looking at this from a FreeBSD perspective which as you said is unfair. From a FreeBSD perspective this makes little sense…however from a Debian perspective it makes A LOT of sense. And “shocked” isn’t really the right word for me to have used…surprised would have been a better choice. You’ve convinced me Syntaxis, 180-degree-turn-around for me. The more users of the FreeBSD kernel the better, and perhaps they will discover some bugs and limitations, and make some important contributions for the betterment of both communities. Afterall, that is the point of the BSD license isn’t it? Use the software however you see fit, so long as you give credit where credit is due 😀 Besides, the Debian project appears to be a bunch of good citizens overall from what I’ve seen when it comes to these issues, as long as they don’t fork the FreeBSD kernel and GPL this fork I have no issues with it. Anyway, thank you for revealing the error of my ways. Perhaps I’ll give this thing a try when it’s all finished, just for the heckuvit. -bytes256 2003-06-11 3:13 pm Anonymous Why not Gentoo? Cause Debian supports 11 architectures, and Gentoo only really supports 2? Dont say “but they support alpha too…”, cause in reality they dont. There’s a reason it only took 2 days to make that “port”, it’s not done. When Debian claims to support an architecture, they SUPPORT it, they produce packages for it, and treat it like all the other ports. There weren’t any binary packages to download for Gentoo (when i tried the alpha port), and the sources wouldnt build properly, then there’s the “all the packages are masked” problem…gee..fun. I’ve had a very sour experience with Gentoo and i dont look at that project the same way anymore. 2003-06-11 4:58 pm Anonymous It is fine in itself that there are all these Debian subprojects like Debian GNU/FreeBSD and Debian GNU/Win32, and that Debian supports so many architectures. But it may mean that the project as a whole may be lacking clear goals and focus, which shows, for example, as slow development of desktop features when compared to many competitors. And of course the need to equally support all those architectures slows down the release cycle to some extent (if the versions for archs are not separated more). That’s why there are many commercial Debian-based distros which do what the Debian project itself seems to be unable to accomplish. Debian is sure a fine example of the “bazaar” development model, but for so big a project a bit more “cathedral” approach, i.e. better focus and concentration of efforts, might not hurt, if some real breakthroughs are really wanted. I don’t know if that is really possible for a project like Debian though, and if the developers would accept it? Debian is a very good server OS, but as a desktop OS it has lots of problems. Just read the reviews like http://www.debianplanet.org/node.php?id=831 & http://distrowatch.serve-you.net/dwres.php?resource=review-debian to find out. If Debian wants to be a good desktop OS too, it just needs to answer the demands of common desktop users (not just Debian devs/hobbyists). These features come to my mind: faster release cycle, GUI configuration tools àla YAST of SuSE, better multimedia support, main concentration on the architectures that most desktop users use (i386-i686 & Macs), and maybe also the separation of other architectures more from the main development branch if it is needed. That’s what commecial desktop GNU/Linux developers like SuSE and Redhat do because they have to really listen to their users. Their developers can’t just develop software to their own needs and likings. And, personally I wouldn’t mind having some top class security features even for my desktop too, àla Adamantix, even though it would cost a little in speed (the problems maybe related to the Adamantix project itself are another story). 2003-06-11 5:02 pm Anonymous “Debian’s APT has a cool GUI front-end called Synaptic. I’d like to see something similar in FreeBSD.” For a GUI front-end there’s portsman under sysutils in FreeBSD. I’m not entirely sure this is the same thing as Synaptic. I freely admit that I’ve yet to keep Debian running for more than a week. So I really can’t say how the two compare. Somehow I always kill it…. 2003-06-11 5:40 pm Anonymous “For a GUI front-end there’s portsman under sysutils in FreeBSD.” Thanx. I’ll try it. 2003-06-11 6:11 pm Anonymous “slow development of desktop features when compared to many competitors” This is a fair criticism, though rock-solid stability is in my eyes (yes, I know I’m not a typical desktop user) a good trade-off. Since Stable doesn’t change until the next release (i.e. every 12-18 months) this is inevitable. As I said earlier, if Testing gets security updates, desktop users will have the choice of running that. An alternative possibility is the Desktop subproject which is focusing specifically on bringing Debian to the masses. “And of course the need to equally support all those architectures slows down the release cycle to some extent (if the versions for archs are not separated more).” Actually, if you think about it, supporting them *unequally* would be an administrative nightmare. If the ports were allowed to declare new versions at their own discretion, there’d be no cohesion whatsoever, and the Security Team would be completely swamped, having to support so many different versions of each package. It’d also completely wreck the consistency of Debian (whatever port you’re running, it’ll “feel” the same as any of the others). “That’s why there are many commercial Debian-based distros which do what the Debian project itself seems to be unable to accomplish.” And quite a few of them have vanished into the ether (Corel, Storm Linux, Progeny, that’s three off the top of my head) while Debian has endured. Also, none of the commercial variants have anything remotely approaching the Debian userbase. Plus, they serve as encouragement for Debian to incorporate said value-added features (or Free equivalents) themselves. It’s all good. 🙂 “for so big a project a bit more “cathedral” approach, i.e. better focus and concentration of efforts, might not hurt, if some real breakthroughs are really wanted. I don’t know if that is really possible for a project like Debian though, and if the developers would accept it?” No on both counts. 😀 Besides, it’d destroy Debian. As you rightly say, Debian embodies the Bazaar development model. With each release it’s predicted that there won’t be another because Debian will fall in the meantime, drowned in its own bureaucracy. Whilst it’s not an excuse for complacency, Debian’s still here. Issues raised *do* tend to get solved. “If Debian wants to be a good desktop OS too, it just needs to answer the demands of common desktop users (not just Debian devs/hobbyists)” If they make “demands”, they certainly won’t get anywhere. Requests are another matter; developers are generally pretty responsive to those. Anyway, desktop support is coming, albeit slowly. Apart from anything else, I suspect quite a few devs maintain corporate desktops, so they’ll have an incentive to work on desktop features alright. By the way, most of the developers are professional admins, so they’re not really “hobbyists”. For many, maintenance of their package(s) is directly linked to their professional careers as well as their own usage. “faster release cycle” They’re working on it. I think a yearly release is about the best it can possibly ever get, though, and that’s being optimistic. A year and a half is more feasible. “GUI configuration tools àla YAST of SuSE” Hardware detection and a more user-friendly installation interface will come when the new installer is completed (work progresses apace). As far as GUI configuration tools within the distribution are concerned, I’d personally rather leave these to KDE and GNOME et al rather than make them distro-specific. “better multimedia support” This all looks fine to me in Testing and Unstable. Again, Stable won’t be updated until the next release. There’s always unofficial repositories at http://www.apt-get.org, though I admit that’s not an ideal solution. “main concentration on the architectures that most desktop users use (i386-i686 & Macs)” I posted a thread from -devel about this in one of my other comments. Arguments against it include: for some architectures, Debian is the only distro that supports them. Also, cutting out the other architectures won’t get their developers working on mac or x86 instead – it’ll just get them leaving Debian altogether. Nor will those two architectures necessarily release faster as a result – they still have their share of bugs and hold-ups despite having the largest userbase, and the biggest delays in Unstable, e.g. transitions from one major library version to another, usually apply to all architectures and are fixed at the same time on all of them. “the separation of other architectures more from the main development branch if it is needed.” For the reasons above, I don’t think it’s needed, and if we want to maintain consistency and security updates throughout, the architectures have to be kept in sync. “That’s what commecial desktop GNU/Linux developers like SuSE and Redhat do because they have to really listen to their users. Their developers can’t just develop software to their own needs and likings.” Lol – this is obviously complete nonsense, or else Debian wouldn’t have any following at all. It’s just words. There is no collusion to prevent these features being added to Debian, it’s just that no-one has volunteered to actually do the work. As I said before, anyone can become a Debian Developer, and you can contribute code even without being one. “personally I wouldn’t mind having some top class security features even for my desktop too, àla Adamantix” Then use Adamantix. 🙂 I covered this in an earlier comment, btw. It’s just not practical to integrate, and Adamantix themselves don’t even have XFree86 working properly yet. 2003-06-11 7:56 pm Anonymous Would people bitch and moan about somebody working on something you can use for free. I have no real interest in Debian on top of FreeBSD, but I am not going to complain about it. It is something I have the option of trying if I like (although I’m sure I won’t). I can understand someone saying they don’t plan on using something, but the reactino to certain projects on this messageboard is ridiculous. I usually don’t agree with making people register to post. But I think something has to be done about the “Anonymous” posters here. They ruin a good discussion by flooding it with garbage. 2003-06-11 8:06 pm Anonymous No no no… You are talking about a different kind of stability, it has nothing to do with this. It has nothing to do with wether it crashes or not. If it is constantly changing, it is “unstable”, as in, moving. Basically it’s “design” is not finalized, and might change considerably at any moment, it’s a moving target, but still stable enough (in your terms) to be usable. “Stable” on the other hand, is finalized, the only changes it will undergo will be in the form of minor bugfixes or security updates, it’s not a moving target, it is still, it is stable :p 2003-06-11 10:36 pm Anonymous Good points again…;) But still, the release cycle & especially the desktop development of Debian just seem slow when compared to most other Linux distros. Why? I only tried to find reasons for that, and I still think that the too loose focus of the Debian project is really the main reason (and maybe not much can be do with it as Syntaxis comments seem to suggest too?). About Debian-based alternatives: Howcome is it that Knoppix has got a good automatic hardware support in a very little time, and with a very small developer community while Debian proper still hasn’t got anything even close? Howcome Libranet has (besides of good hardware auto-detection) developed a good central admin tool Xadminmenu though it has only a couple of developers? And howcome it needs Xandros to show us what a really easy to use Debian-based desktop can be? Those people and companies focus their efforts towards those kinds of goals while Debian does not. Of course all those projects mentioned above may thank Debian for much. Debian is a great project and I wouldn’t be writing so much about it here if I didn’t care for it. Just saying: “don’t worry, these things are being developed”, may not get people happy if the development seems awfully slow. But sure, lets hope that the Debian desktop project, and all the other dozens of Debian sub-projects, will someday solve the problems of Debian. 2003-06-11 11:49 pm Anonymous What you don’t understand is that numerous forks and -based distros are in the very concept of Debian. Essentially, the APT system is under a BSD license–you can use it and do whatever you want with it, but Debian itself will always remain free. The main branch is not desktop oriented. Knoppix is great. If you want all of Debian + a great desktop, use Knoppix. How is this in any way a problem? (also, please remember that Adminmenu is proprietary.) 2003-06-12 12:17 am Anonymous “Howcome is it that Knoppix has got a good automatic hardware support in a very little time, and with a very small developer community while Debian proper still hasn’t got anything even close?” Actually, it has. Check out PGI (http://hackers.progeny.com/pgi/) for an example. It’s even in the Debian package tree, as well. People should be able to create PGI install images for i386 with no problem. There’s even a sample ISO and a great tutorial on PGI’s site. Of course, there is no way that this thing will suffice for Debian’s official installer as-is. It doesn’t support all the various different install methods that are required, and it needs to be ported to *all* of the 11 architectures that Debian supports. “Howcome Libranet has (besides of good hardware auto-detection) developed a good central admin tool Xadminmenu though it has only a couple of developers?” I think this is becoming obsolete as the Gnome and KDE control panels and apt frontends like Synaptic mature. “And howcome it needs Xandros to show us what a really easy to use Debian-based desktop can be?” I agree that their file manager above all is really slick. That said, it uses KDE 2.2.2, so there’s already a trade-off between the functionality they’ve added to the file manager etc and the functionality that KDE have since added to the entire DE in the 3.x branch. Once again, nothing’s perfect, and there’s always a trade-off of some kind. 😀 “Those people and companies focus their efforts towards those kinds of goals while Debian does not.” Sure. When you pay a developer, you get to tell him what to do. Of course, the trade-off is that the fruits of their labours are closed-source and comparatively expensive… “Just saying: “don’t worry, these things are being developed”, may not get people happy if the development seems awfully slow.” That’s an extremely good point. 😀 You shouldn’t just take my word for it. Take advantage of Debian’s philosophy of open-ness and go and check it out for yourself. All the code is freely available and so is access to the mailing lists and bug tracking system. The only realistic way of speeding up the process is to contribute to it in some form. Coding would be the most obvious example, but testers are always needed, as are people to write documentation. And again, since the process is open and *everyone*, official Debian Developer or otherwise, has the ability to contribute and yet the work is still not completed… who is it that should shoulder the blame? I’m not really sure that there’s an answer to this question. “But sure, lets hope that the Debian desktop project, and all the other dozens of Debian sub-projects, will someday solve the problems of Debian.” I remain similarly optimistic. 🙂 2003-06-12 12:40 am Anonymous “Essentially, the APT system is under a BSD license” Just a minor correction: Apt is actually licensed under the GNU GPL, with an exception to allow linking to QT (this was added to legalise what Corel were already doing). But yeah, it’s pretty free. 🙂 “If you want all of Debian + a great desktop, use Knoppix. How is this in any way a problem?” I agree with this. It’s a good solution in that it will work just fine for most people’s requirements, and it’s simple to install and set up. It’s what I recommend to all my newbie friends. 🙂 Unfortunately, it’s not the ideal solution in a couple of ways. Having Debian-based projects outside of the Debian umbrella is sub-optimal because it leads to wasteful duplication of effort. To counter this, some developers are working on a way for Debian users (and Knoppix) to automatically generate their own LiveCD, with a choice of whatever packages they desire, straight from the Debian package mirrors (and a choice of from Stable, Testing and Unstable as well). This would minimise the duplication of effort of Knoppix and all its various knock-offs having to maintain their own packages, sources and infrastructure; they have to do this, because currently, some pretty gross hacks are required to get a Debian system to even work in a LiveCD environment. 2003-06-12 8:05 am Anonymous Is it some sort of hack of jigdo? That would be a good idea. 2003-06-12 10:28 am Anonymous “The main branch is not desktop oriented.” Just what I’ve been trying to say here. Debian is a good server OS, and a nice example of loose voluntary collaboration, but there may be and are better desktop oriented alternatives, and it probably remains like that. A big ship is always slower to turn than a smaller one. It is very difficult to make a supposedly universal all around OS, not to mention so loose voluntary project like Debian, to be as slick and usable in a certain field (like desktop) than what a dedicated OS (like Xandros, Knoppix or Libranet) can do in its own special field. I suppose Knoppix could be called an up-to-date desktop Debian, although it is really a Live CD which shows as many problems when you use it. I remember to have met several dependancy issues when I tried to install some common new versions of desktop software to my Knoppix on a harddisk. Knoppix mixes software from various sources which hardly makes updating it an ideal solution especially for a relatively new user. Also to have a good and up-to-date hardware auto-detection in Debian (perhaps with plug-in support) that supports all the architectures that Debian must support is a really huge and difficult task to accomplish. Let me again say that I think Debian is a great project in itself. But it is not the best choice for all the possible purposes and will not be. Debian is, however, advertised as a universal OS for all purposes, even more than what MS does. Somehow I feel that Debian tries to be everything, everywhere to everyone – which may lead to a bit too loose focus, don’t you think? 2003-06-12 4:02 pm Anonymous “there may be and are better desktop oriented alternatives, and it probably remains like that.” Certainly. 🙂 Nothing’s ever perfect, and it’s all subjective, anyway. “Better” will mean something different to different people. If you want a *stable* desktop (e.g. the corporate desktop, or someone who really does use their PC as just a tool) Debian is absolutely perfect for the job. On the other hand, if you want glitz and glamour, others are further ahead. You’d have to decide which is more important. No-one offers both. “Knoppix mixes software from various sources which hardly makes updating it an ideal solution especially for a relatively new user.” True. 🙂 Integrating LiveCD support into Debian proper would be a great step towards solving this problem. @Greg and all others interested in this: check the various relevant threads at http://mailman.linuxtag.org/pipermail/debian-knoppix/2003-March/thr… which discuss how this integration might be achieved. A nice summary of the proposed roadmap can be found at http://mailman.linuxtag.org/pipermail/debian-knoppix/2003-March/002…. “Somehow I feel that Debian tries to be everything, everywhere to everyone – which may lead to a bit too loose focus, don’t you think?” This is a valid point of view. And of course, the term “Universal Operating System” is rather a tired cliche – you’re right, nothing can be all things to all people. But as it stands, Debian can do (or can be made to do – see Knoppix as an example) an awful lot of different things pretty darn well. Until viable alternatives for each and every one of those things come into being which at the very least share Debian’s dedication to a) integration (again, see Policy at http://www.debian.org/doc/debian-policy) and b) Freedom as enshrined in the Constitution (http://www.debian.org/devel/constitution), Social Contract and Debian Free Software Guidelines (http://www.debian.org/social_contract.html) I think Debian will be around to stay. 🙂 Those two unique virtues (there are more, but now’s not the time to list them all) can make up for a lot of shortcomings by themselves, in my mind. And of course (to take the other end of the scale), if you decide to use a different tailored distribution for every single different purpose you want to use a PC for, you’ll have to learn loads of different ways of doing things which is highly inefficient. Much simpler to just learn Debian which can do them all pretty well, with a consistent userland and feel throughout. You’re right. 🙂 Debian’s not for everyone, and never can be. But I think the way they’re going about it means that they’ll get as close to a “Universal Operating System” as its possible to be, whilst still retaining their focus on software freedom, stability and integration. The focus on these things is the lynchpin that holds the whole massive Project together. Of course, the only way to know for sure how it’s going to turn out is to wait and see. Or better yet, get involved, contribute and try to play a part in actively shaping its future. I certainly can’t predict the future, but as I say, I personally remain pretty optimistic that the size of the Project won’t cause it to collapse in on itself just yet. 🙂 2003-06-12 4:03 pm Anonymous Debian developers have their own interests, and what they are good at. Debian was one of the first distro’s of Linux, and they set forth a very precise picture of what a distro should be. They beleived a distro should make it easier then it was at the time to put together a system comprising of Linux with GNU tools. They do commit patches etc to packages in Sid, but for the most part, its about just that, packaging a usable system. Those of you that state perhaps it would be better for them to contribute to other upstream projects, thats not what they are interested in. The first port away from Linux was started in 99, so this isn’t a new thing. Bruce Perens was probably responsible for that one, he decided that Debian should be ported to the HURD, people went along with it because he was the project lead. Some wanted it to be reality, many didn’t like the idea, and those people decided to devote themselves to making sure HURD and Linux weren’t the only Linux kernels available. The project as a whole went along with this too. Now today, the FreeBSD port works mostly, the HURD project is only a little bit behind, and the NetBSD project is moving along thanks to both of the other ports. Those who say “Why”, and “who would use it”, the obvious answeres are “cuz they feel like it”, and “the people working on it”. I for one will be testing all ports if and when I am able to apt-get the different kernels, for no other reason then to test the performance. I can’t wait for the day when this is possible. I can’t stand FreeBSD’s userland, but I want to use FreeBSD, I want to see what its got to offer. I also wish to try to HURD port, I want to see how it performs as a microkernel compared with monolithic kernels mostly. Above all though, I want to be able to use the best tool for the job, without having a steep learning curve. AFAIK, there is also an OpenBSD kernel port in the works also, so in the future it will be possible to have: Linux – Desktop FreeBSD – Server OpenBSD – Proxy, Firewall NetBSD – Any obscure archs lying around And they will all look and feel the same to the user, I won’t have to learn 4 different userlands, I will be using the same basic userland, and filesystem layouts on them all. Debian consists of only a group of userland tools, and a group of 900+ maintainers of native packages. Only packages used by Debians users are available for Debian, which I think is a great thing, you only get what others find usefull. Its also very easy to install and completly remove packages with the tools available, and most are packaged in such a way that they will work right away, only needed constomization, not full configuration. These are the reasons for the port, the Debian community is responsible for everything they wish to include as part of Debian. Every package is maintained by someone who actually uses that package. And above all else, most find it fun to search for bugs in their packages, and try to fix them. Debian also is a democracy… don’t like the direction of the project in general? Every 2 years there are elections for a new project lead. There are also people in charge of most every part of Debian. Its very organized. Above all though, its an operating system by geeks for geeks, not a corporation making an OS for the tech-illiterate. They add features they want implemented. If they feel something is missing, they attempt to implement it, hence sub-projects like debian-desktop – for ease of use, debian-installer – for an easier installer, debian-freebsd – for those that see merits in the FreeBSD kernel, debian-hurd – for GNU zealots mostly, debian-netbsd – for those that like the idea of portability, debian-openbsd – for those beleive in building an OS around security, and of course debian-w32 – for those that wanted (its no longer being worked on) to run Debian on Windows. The people responsible for each sub-project have their reasons, but above all, like most everything to do with Debian in general, its about the people who use Debian working to implement something they beleive will improve Debian. Nothing has been worked on for any other reason. If you ask me, noone here has any right to doubt them, they are doing something they want to do, simple as that. Its definatly *NOT* about “who will use this”, they will, and thats all they care about! 2003-06-12 4:11 pm Anonymous Umm, I made a few spelling and grammatical errors, and I meant “Woody” not “Sid” that gets patches etc. Sorry about that, I didn’t notice the errors when I proof read the post first time. Anyways, my points are valid, so please don’t overlook the points due to the small errors, thankyou in advance.