Debian FreeBSD Distro? 241
antigen fiend wrote in to note that Debian Weekly News has a bit about a recent debate about a Debian port to the FreeBSD Kernel. There are several comments relating to licensing, ease of porting, and other relevant topics. Its definitely an interesting project, with a lot of political and technical snags. Any thoughts?
I just could not resist (Score:1)
I think it would a good thing. (Score:1)
There's already work being done on a Debian/GNU Hurd distribution, which in some ways is more different than FreeBSD (although it does use glibc, which helps).
Fine, but isn't GPL! (Score:1)
A few questions from the illiterate fool... (Score:1)
--
About licensing (Score:1)
Both?? (Score:2)
Debian for it's excellent desktop/workstation capabilities with their wonderful autoupdating website/dependencies thing and the total hackability of it all. The packages always work, and their dependencies aren't all funny like they often are with SuSE and RedHat,
However I use FreeBSD for my servers, yes it's not the easiest of distros and it's very hard to get anything real working (for a relative newbie like me anyway), but it works, very well in fact. It never ever crashes and it servers my websites just fine. There are very few bugs very few updates, and I can just leave it alone totally.
To combine both? Well - heaven - what else can it be. Add to it KDE 2.0 and you have the complete MS competitor for the desktop/low end server market.
Bring it on!!
Re:Fine, but isn't GPL! (Score:2)
It would give you some enemies for life though..
Re:I think it would a good thing. (Score:1)
Porting 'Ports' to Debian/Linux would make more sense...
Interesting project, but ... (Score:2)
What's the point? (Score:1)
Re:Fine, but isn't GPL! (Score:3)
The paragraph in DWN mentions the concern that it would open the doors for unscrupulous 3d parties to take their work and sell closed-source derivatives, but I don't see that as a real problem ... the FreeBSD kernel would have the BSD kernel, but much (most?) of the software that the Debian team puts in the DebianBSD distribution would still be GPL, which means FastBuck Inc. would not be able to take DebianBSD as-is and apply a closed-source license.
And now a question for the License Lawyers.
Christopher A. Bohn
just give me an up-to-date linux distro (Score:1)
Re:About licensing (Score:1)
Good idea, but glibc first would make more sense (Score:5)
We talk about freedom, yet all the major Free Software vendors are using the same kernel. If Debian pushed GNU/Linux, GNU/Hurd, and now GNU/FreeBSD, then we will eventually have the day that we can choose which kernel is the best for our purposes.
Few people use Linux for the kernel, they use it for the capabilities. I have a friend who loves his GNU/Solaris machine. Don't laugh, he administered Solaris machines for years, so his home PC has the Solaris x86 stuff on it, but all it runs is Free Software (and StarOffice).
Linux advocacy is silly. Advocating *nix, or even Free Software makes sense, but obsession with a kernel is silly. As Linus says, (I might be paraphrasing) "Linux sucks, it just sucks less than anything else out there."
The Linux kernel is fast, it is stable, but it isn't revolutionary. As I understand it, it is Microkernel-esque, although it still probably has remnants from its days as a monolithic kernel. I mean, when someone comes out with a solid exokernel, are we going to scream and yell about how Linux is still better?
Support free software, but support choice. In a free software world, we could pick our kernel without worrying about our apps breaking. This project has both technical and political merit. Although, if the BSD license allows you to do whatever you want with the code, can Debian release the GNU/FreeBSD system under the GPL? If they made the best FreeBSD distro and put it under the GPL, the license issues would be over. I'm not sure of the specifics of the BSD license, but if you can include it in proprietary licensed systems, I don't see why Debian's system couldn't be GPLed.
Debian, FreeBSD, glibc, the HURD (Score:3)
On a more general note, the more people that do porting projects like this, the more pluggable the whole system will get. Wouldn't it be great if you could select your kernel, libc, and packages independent of each other? This is what UNIX is about... choice.
It would also be great to have glibc ported to FreeBSD. I sure hope that is the route they decide to take. This would allow FreeBSDers to have better access to new applications (although I'm not sure if they care - BSD is usually about stability, not newness). Also, given the user base of Debian, this would give more competition to Linux, and give us a reason to still be competitive.
Finally, this will expose more bugs in all systems, which is always a good thing.
What's the point? (Score:1)
How much difference will it make which kernel is used? (I mean technical, not licencing, issues). How much variation is there in the performance of different kernels? Is it noticeable in daily usage?
Also, how different are the kernels that the various *BSD's use? For example, OpenBSD has a reputation for being very secure. Is that due to kernel design, or more to the way the whole distribution and its packages are configured on top of that?
Re:Fine, but isn't GPL! (Score:2)
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
3. All advertising materials mentioning features or use of this software
must display the following acknowledgement:
This product includes software developed by the University of
California, Berkeley and its contributors.
I guess they retained the original copyright, at least for the 3.3-RELEASE.
Intosi
What about their HURD project? (Score:2)
Well (Score:1)
If you think you know what the hell is really going on you're probably full of shit.
Re:What's the point? (Score:1)
--
Re:Fine, but isn't GPL! (Score:1)
Are the *BSD developers allowed to hold on the the advertising clause? The GPL certainly says that ppl are allowed to use/distribute the software according to "either version 2 of the License, (at your option) any later version."
Is their a similar clause in the BSD licence (I don't think I have ever even read the BSD licence) If there is, then the debian ppl could choose a later version than the one it was released with. ie. they could choose the one without the advertising clause.
--
MartinG.
Not Official (Score:2)
Re:Good idea, but glibc first would make more sens (Score:1)
The idea of more FreeBSD distributions (with a different userland, different package management) sounds like a horror to me. "Sorry, our product only works with Debian/FreeBSD, not FreeBSD." Shrug.
Intosi
What's the point? (Score:1)
Why on earth port Debian (an excellent system, BTW) to FreeBSD (another excellent system IMHO)? FreeBSD already has its own (outstanding) ports/package system and I totally fail to see what kind of interest is at play here.
And the absolutely worst thing is that it raises all kind of licensing questions (BSD vs GPL) and library porting (glibc vs libc) -- this thing is totally beyond me. Again: what is the interest of having Debian ported under FreeBSD?
I can understand Debian/HURD, but Debian/FreeBSD is the weirdest thing I have heard in ages. If a gentle
Ah well. As long as it's open source... =)
Re:A few questions from the illiterate fool... (Score:2)
2. Varies, but with a bit of work you can port most things
3. Nothing, other than their working methods. Debian's approach is highly parallel.
4. It would be hard. dpkg is GPL'd.
Re:Fine, but isn't GPL! (Score:1)
However, the advertising clause is incompatible with the GPL, which is why BSD-licensed software historically could not be relicensed under the GPL.
Christopher A. Bohn
Debian + FreeBSD + KDE 2.0 = _the_ desktop box (Score:3)
Debian is the best linux dist imho.
FreeBSD is the best x86 os imho.
KDE 2.0 looks _very_ promising.
And them together sounds lika a real threat to MS Win* as the desktop alternative.
The 4 s:
Speed, security, stability and style.
Its my lucky day!
Lost Carrier
Re:What's the point? (Score:2)
my feeling is that Debian/FreeBSD should base itself on glibc2.1, i.e port glibc to the FreeBSD kernel, and compile and run the whole distribution on that, *then* also include a copy of FreeBSD's libc, to be able to run regular FreeBSD programs (just like RH6 can run libc5 programs too). otherwise, Debian/FreeBSD is nothing more than FreeBSD with a bunch of GNU tools added and the installation procedure re-tooled.
OTOH, I wouldn't commit myself to a project like this for the long-term; it appeals to me more for the hack value, and for the political implications (hopefully the good ones, of showing the various factions how easily software can work together) than for actual usefulness.
Re:Sniff (Score:2)
I mean seriously - why do we need it? The FreeBSD/OpenBSD/NetBSD systems are just as painstakingly careful with their release-system as the Debian people, so it's not like we'd be getting a more mature development cycle, right?
On the other hand, there really aren't any "distros" to speak of with the BSDs. I wonder why... It would seem that a company willing to put a bit of effort into ease-of-use could become the RedHat of the BSD world. In fact, they could probably go head-to-head with RedHat in terms of OS sales. Doesn't everyone here always say "competition is good"?
-----------
"You can't shake the Devil's hand and say you're only kidding."
Re:What's the point? (Score:1)
It's the distribution as a whole that makes it secure, because everything you could want on a server has been tested.
If you add something that has not been included you remove SOME security.
But you still got the general OpenBSD design to protect you, you get mailed about changes in configfiles etc.
In the end I would say that it's the way the system has been designed that makes it so good,
and of course the auditing.
Aaaah hell yeah... (Score:1)
Naming of proposed distribution (Score:1)
I say this without malice toward anybody.
Besides the various "why"s such as "because it's there" and "to improve the portability of our software" and the like, there is a political reason to do this.
If Debian makes a BSD distribution, it would almost certainly be called "Debian GNU/BSD". This would reinforce the idea that, according to the FSF, et.al., the correct name of the Linux distributions is SoAndSo GNU/Linux. By creating a GNU/BSD, the media attention would make people pay attention to the "GNU/" prefix and would also result in comparisons to GNU/Linux.
Christopher A. Bohn
Re:A few questions from the illiterate fool... (Score:1)
2b. % make
2c. % make install
2d. debug
2e. rinse, repeat.
Christopher A. Bohn
Re:I think it would a good thing. (Score:1)
Re:Interesting project, but ... (Score:1)
Personally I can't see more than a very marginal gain. Except for perhaps hardware incompatilities or boxes that are brushing the outside of the envelope the respective kernels are much of a muchness. They work. They work as close to _all_ the time as makes no odds. It's all the other things on the box that differenciate the products. Init scripts, user space programs, system utilites. Take linux user space and put it on bsd, take the bsd use space and put it on the linux kernel if you really want - what did you gain that you didn't already have?
Debian GNU/Solaris (Score:1)
All this kernel independence surely would help a company like Sun if it wanted to base a Solaris release on the Debian system. The techie in me drools. But there would be no question of "official" support by the Debian organization, which seems to be the real sticking point with the FreeBSD effort.
Debian GNU/NT, anyone? ;-)
Hurd continues... (Score:2)
A significant merit to adding FreeBSD to the mix is that this makes Debian less and less kernel-dependent. In the long run, that makes it more and more possible for Debian to support more "UNIX variants."
Interesting, in the longer run, would be:
Thus, support for FreeBSD tomorrow may help there to be support for more unusual OS selections a couple years from now.
That seems to me to be a Good Thing.
Re:I think it would a good thing. (Score:1)
Re:About licensing (Score:1)
Re:Debian, FreeBSD, glibc, the HURD (Score:1)
UNIX is about choice... absolutely agreed. Though IMHO I'd rather people focus on GNU/Hurd instead... If I'm not mistaken, isn't the whole point of the Mach microkernel to let you run totally different interfaces on top of the same kernel? If Debian gets Hurd to work, we could even start creating a "Linux-on-Mach" interface that would allow you to run native Linux apps, a "FreeBSD-on-Mach" interface that lets you run native FreeBSD binaries, and perhaps even (shudder) a Win32 interface that lets you run native Win32 apps (akin to WINE).
Isn't this flexibility the whole reason Mach exists at all?? This would really give you freedom of choice, and perhaps even more than that, as you can simultaneously run different interfaces, all of them natively! IMNSHO this is the ultimate embodiment of the UNIX philosphy. This is why I predicted that the next big thing after Linux would be Hurd, not necessarily because of Hurd itself, but because a microkernel design gives you so much more flexibility than the traditional monolithic design. That's why I think it would be most profitable for the Debian people to work on Debian GNU/Hurd instead, if they wanted to work on anything outside Debian GNU/Linux.
DebianBSD: most likely going no where fast (Score:4)
There are several reasons I do not think DebianBSD is going to be advantageous:
1.) Most people using FreeBSD enjoy the less restrictive, more open BSD license.
2.) The FreeBSD Ports Collection is truly incredible; there little or no need or desire for a Debian package system.
3.) I personally love the fact that FreeBSD is *not* a distribution, it is a full featured, integrated Operating System. I think this allows FreeBSD to be more intelligently organized and to work more smoothly than the various Linux distributions forced to integrate an operating system beyond their control with software largely beyond their control. DebianBSD sacrifies this advantage.
4.) The licensing issues are difficult, if not impossible to resolve.
5.) Any variation on the name GNU/BSD is going to cause an uproar and permenantly doom any Debian based FreeBSD distribution.
6.) Finally, from quickly scanning the Debian mailing lists, it seems as if most of the Debian developers have no respect for FreeBSD. One called it "dying software" and others claimed it offered no advantages over Linux. While everyone is entitled to an opinion, however ill-informed and erroneous, I wonder how dedicated Debian could be to an operating system it does not like and does not respect; after all, part of the allure of working on open source software is being able to code for your own pleasure rather than someone else's.
In short, DebianBSD seems like an abortion from the start, althougth I am prepared to be surprised.
(Note, even if DebianBSD did come into existence I would never switch; Debian and the FSF irritate me with their holier-than-art-thou pronouncements on freedom. Sadly enough for DebianBSD, I have a feeling a large number of FreeBSD users will agree with me.)
Why not just try and get Potato out??? (Score:1)
The unfortunately part of this is that there are packages that have been updated in unstable, but are unusable on an unmodified slink system (slink's dhcpd-beta / unstable's dhcpd comes to mind). While not updated for security risks, it would be nice if I didn't have to update half a dozen packages and move to the 2.2.x kernel just to fix confirmed memory leak problems.
Before the 'unstable is as stable as RedHat' flames start, there have been problems with unstable (the perl problems come to mind). And for those 'well pitch in and help' flamers, I just dropped off my United Way sign-up sheet to HR and made a nice contribution to the Debian project.
Good call. (Score:2)
This is better than the opposite observation that I was going to point out, which is that there are components of Debian, such as Perl, [perl.org] Python, [python.org] and XFree86 [xfree86.org] that already use non-GPL-like licenses.
Re:Well (Score:1)
I follow the debian-bsd list, but don't do any real work. I havn't a netconnected box for this kind of work so I can't really do anything but moral support and maybe trying to answer some questions
Many has sugested OpenBSD over FreeBSD but those who did the work used FreeBSD and that how it was choosen. I don't know how much the various *BSD has in common, but jsut one BSD port will solve many political questions.
Debian isn't going to stop making Linux anymore than we are going to stop makin i386. In a perfect world (which Debian of course is :-) people shouldn't care which architecture the maintainer of a debain package is using. Autobuilders does a lot of work in Debian. And Debian is still primaly Linux users, some of us is just interested in l;eraning new worlds.
My dream is a distribution where you just do apt-get install freebsd or apt-get install linux to use that kernel. It's my personal dream and it probally not going to happend in the nex few weeks.
I hope that this project could mean better Linux support in FreeBSD (file system, emulation etc.) and also the other way around. the perfect kernel and user space toolkit doesn't exists and exchange of ideas will only be of the benefit of everybody.
then we can only hope this doesn't end in pure flaming
Re:About licensing (Score:1)
But you still using XFree86, right?
No modern distribution is anywhere near using only GPLed stuff. One big non GPLed entity is XFree86 almost used by everybody, and that doesn't make any distribution any more or any less free.
dpkg and apt will still be GPLed, so making a version of Debian based on a BSDkernel wont make Debian open for comercial takeovers any more than what Corel allready did. The key points of Debian is and will always be GPLed (jugding from the copyrightholders of dpkg and apt.)
Re:Interesting project, but ... (Score:1)
Re:Horay!! DEATH to SYSINSTALL!! (Score:1)
Procrastination, delays, tired of waiting.. (Score:1)
P.S. Don't reply about using apt-get to use the latest stuff, read: tested.
Ease of administration... (Score:2)
Interesting... I feel completely the opposite. I found the RH init scripts infinitely more logical and easier to admin than Debian's. In particular, network configuration was miles better in Red Hat with all the relevant details for each interface in a config file in /etc/sysconfig, rather than hardcoded into an rc.d script. Still, that's one of the great things about Linux (and indeed, free software in general). Everyone is free to use what they're most comfortable with.
Disclaimer: I haven't looked at Debian since 1.3, so things may have improved since then. Also, much as I like Red Hat, I freely admit it's not perfect.
Re:A few questions from the illiterate fool... (Score:1)
Sebastian
Debian GNU/NT (Score:2)
There was such a project, basically it was to be a Debian format distribution of Cygwin. It would have been very useful, given that Cygnus doesn't put in much work on making free distributions, but nothing seems to have come from it. Nobody volunteered to do actual work.
Re:Debian, FreeBSD, glibc, the HURD (Score:1)
This sounds great (Score:1)
A Debian/FreeBSD would combine the best things, IMHO.
Sebastian.
Re:Debian GNU/NT - I need this! (Score:1)
Excelent Idea! (Score:1)
Just one Word .... (Score:1)
Re:Debian GNU/NT - I need this! (Score:2)
I haven't bothered with them, as Cygwin 1.0 is close to a perfect fit for my needs. It gives me a nice Unix programming environment under NT for producing win32 executables, and since XEmacs runs natively under win32 (and is included on the CD from Cygnus), I cannot think of any X11 programs I miss.
port to cygwin ? (Score:1)
Re:Hurd-on-mach, linux-on mach, hurd-on-linux? (Score:1)
Help (Score:1)
Alas, if sysinstall touched a disk, is because you told it to. There is *no way* it could have touched it without you selecting it.
Anyway, e-mail me if you want. Or e-mail freebsd-questions@freebsd.org.
Twas a joke (Score:1)
Christopher A. Bohn
Debian: Lots of Kernels, splitted forces? (Score:2)
are not yet share holder driven (watch out for SuSE).
It is IMHO a sad fact that the Debian development
model seems slower than it would be good for Debian.
I think that Debian's receipe contains too much
democracy and not enough (benevolent) dictatorship.
Embracing additional kernels can be a nice thing
if your development model is fast and you are
among those who march in front.
Splitting up forces further when you seem to
be falling behind is IMHO the wrong thing to do.
Debian first should do some streamlining of their
development model before increasing entropy
further.
Don't get me wrong: it might be a big plus 'for
the whole thing', if different exchangable kernels
are competing against each other. But increasing
complexity might also be another nail in the
coffin if your development system is too slow to
handle it.
Re:About licensing (Score:2)
1 - Having thousands of people potentially debugging your code for you.
2 - Having complete control over the modification and distribution of your work.
I think that GPL is great for programmers, but I can't see how it can make great business sense, since it allows everyone in the world to potentially be your competitor with no cash or time outlay. Maybe Redhat will prove me wrong... Time will tell, I guess.
Re:Both?? (Score:1)
Maybe if Debian-FreeBSD can evolve in new ways that Linux has not been doing (such as new features for workstations), it would then have some success.
I am dumping my OpenLinux install for FreeBSD 3.3. I used to have 3.2, and miss it.
As a believer in the BSD licence.... (Score:1)
It *IS* market differentation for Debian, and that is what the 107+ GNU/Linux compaines are all about. So they are going to try a different thing.
1) Between GNU/Linux and *BSD, if one group comes up with a feature, or a better way of doing something, the code is either moved directly, or the idea is re-implemented under the correct license.
2) Most of the user-level code doesn't care about the kernel being BSD or GPL. So, unless it's commerical software using a group of tools, the tools on *BSD are on GNU/Linux, and GNU/Linux tools are on *BSD.
3) *BSD has the ability to run most GNU/Linux binaries. So, being *BSD lets you allready run GNU/Linux stuff.
Given these 3 things, I don't see a win for customers. (Unless, of course, the BSD kernel is more stable/better written than the GNU/Linux kernel. But such talk is considered a TROLL on
As for the argument that *bsd is dying...Sorry to say this, but there is a whole group of people who think OpenSource is doomed, that GNU/Linux is dead on the vine, etc. Ask Micro$oft. Judge Jackson thinks that OpenSource is not relavant. The people at BSD/OS mock 'the part time OpenSource' world. Such people claiming *BSD is dying are no better than the blind Amiga or Macintosh advocates. They want to push down other OSes to make thier own OS look better.
Re:Sniff -- Slower? (Score:1)
Sorry, but I am at a point where I think Linux needs more time in the oven, maybe a year or two. BSD has been around for a long time, and that is why it is faster and more stabile.
Re:d00d, he's a BSD bigot! (Score:1)
Grow up!
Re:I doubt this is for the Crabby Old Men (Score:1)
Re:Ease of administration... (Score:2)
My impression has always been that redhat's init scripts are setup the way they are because it makes it easier to write gui tools to manage them.
Slackware is at the other end of the sprectrum using the old BSD style scripts where litteraly everything is in 5 to 10 scripts. This is very hard to handle with anything but vi, on the other hand, when you do get into them with vi everything you need to know is pretty much right there. I know people who cant stand any other method.
Debian's type are I think a happy medium, pretty much everything you need to know for a particular daemon/service is in one file, with a seprate file for each daemon/service.
I also like the fact that they are the closer to the style of the commercial UNIXs I've worked with.
As you said tho, the freedom to choose the type of linux that suits you is part of what makes it so great.
ports confuse you? (Score:1)
cd
make install
installing needed packages ######
installing lynx ######
lynx www.slashdot.org
now how hard is that?
Re:Horay!! - Can't be any worse than RedHat (Score:1)
However, Caldera has been very unstable for me, and am going to remove it.
Re:DebianBSD: most likely going no where fast (Score:2)
I'm curious what the difference between a "distribution" and an "Operating System" is. I see an "Operating System" as a kernel and maybe the Debian "base" section, the minimum amount of stuff you need to get something useful running. Is the issue here that the FreeBSD people maintain both kernel code and user-space code? If there are stable interfaces between the two, then that's pretty irrelevant.
Well, what's the verdict? (Score:2)
The idea of it all is very interesting, no matter the pros and cons of it.
FreeBSD and Linux users needs to wake up and learn to respect each other. Most of us know that competition is good, and FreeBSD and Linux are in direct competition in many areas.
If the BSD community does not want GPL, I can respect that. GPL is not some Holy Grail of licensing as many of you make it out to be. I believe things can be TOO open. Why? It is because things become chaotic and fragmented, slowing progression and causing all sorts of problems. You are going to see a lot of this in the Linux community over the next couple of years. (It is going to happen because commercial distros are going to try to get ahead by breaking away from the others. The lag time for others to pick up any advantages to their ideas is also enough to cause problems.)
Besides, for the vast majority of us, the license does not matter anyways. We just install it and use it. We are not modifying the libraries, etc. etc.
Debian-FreeBSD is a wake up call to us all. It's time to start repecting our differences. We are all individuals, not members of cults.
E
Re:Sniff (Score:2)
Sure, there's a BSD distro. It's well supported, solid, comes packaged for ISPs and Servers. The fine people at BSDI sell it.
I'm beginning to think the reason there's no inexpensive, mass-market distro for *BSD is because of the license. You just can't stop the suits from closing off important pieces, and jacking up the price to what the the market will bear because they can .
Look at Cygnus. They started out as a business to offer support for Open Source "products". Now, they sell closed source products as well. Why the change in business plans when, according to Cygnus, they've always enjoyed phenomenal growth and profitability? I think it's because the suits just can't stand the thought of giving everything away.
This is what's really new about running a business distributing GPL'd code. The suits have to start rethinking where they can make the money.
I'm not sure there is a lot of money to be made, long term, in being a distro vendor. A lot of people will just buy the same thing from Cheap Bytes. Ultimately, I would think that the distro vendors would get more involved in support, custom projects and training. I think RedHat knows this and this is why the Cygnus purchase makes sense.
Re:I doubt this is for the Crabby Old Men (Score:2)
Like, um, errr, um, nope I'm stuck already.
Oh wait, I remember, *BSD has more crazy fanatics, so you can feel more at home. And if someone asks "How do I delete a file?" you can call him lame, and tell him to buy a clue. The microkernel dig above from a supposed *BSD fan is hilarious "Our kernel has less features, and is therefore inherently better".
The *BSD kernels each have a lesson or two to teach Linux, but equally there
are lessons they could learn from Linux, and from each other. It's not enough
to be worth fighting over. IMHO today's *BSD userspace is poorer than the
GNU userspace, and Linux is a better platform for bleeding edge kernel
meddling.
I respect the OpenBSD crew for having the guys to actually audit code, and the adventurers doing "First Ports" to platforms like the old 68K Macs, but the trend towards fanatical hate of other Free Software projects leaves a very unpleasant taste, BSD users are pronouncing that a GPL'd branch is "Evil", while a MS proprietary branch is "Proof that BSD is truely free".
This inconsistent value set (and I don't clame it's present inside the holy temple of the core developers) is a rotten stinking thing inside the BSD community, and makes the "Anything but MS" sentiment found in some parts of the Linux camp seem healthy by comparison. While that rotten thing remains in the community I hold no hope of GNU/FreeBSD seeing the light of day.
Nick.
Re:Good idea, but glibc first would make more sens (Score:2)
And programs under one BSD aren't promised to work on ALL BSD's..
Does it gel with BSD philosophy? (Score:3)
Where I think there might be an issue is that the BSD world does not have this idea that the operating system is the kernel, but rather the operating system is the whole distribution. A debian distribution based on the FreeBSD kernel is not the Debian distribution of FreeBSD but a completely different branch, which one might call DebianBSD.
I think there are advantages to the BSD way of looking at things: security issues around OSs are not issues of the kernel, but issues about the whole system as deployed. I would rather the BSD vision of the operating system was not buried because it is not the same as the Linux vision.
Re:Package Management. (Score:2)
Re:About licensing (Score:2)
However,
I think that GPL is great for programmers, but I can't see how it can make great business sense
The GPL, IMO, sucks for programmers, because it explicitly limits their right to steal source (note: this is not a criticism of the GPL). The GPL, believe it or not, is actually best for marketers and businesses -- marketers who want to get 'into' the free software movement (good for them!) and businesses who want to get the debugging advantages of open source but who don't want to allow their competitors to use their source to build an exclusive competitive advantage.
What, you ask, about the programmers who like GPL? The answer is simple: we're also marketers. We want to position our product not so that it serves the largest number of needs, but rather so that it serves a long-term goal, the establishment of open software.
A lot of us, BTW, are seriously decieved about the purpose of the GPL -- some even say it's more free than the other licences. It's not, it's merely more practical than the other free licenses, since the other free licenses depend on the community to keep a product free, while the GPL gives the illusory impression that even if the community disappears the software will still remain free.
Foolishness.
-Billy
Re:About licensing (Score:2)
So far as the GPL goes... I think it still is really theoretical in nature. When/If someone violates it and is found to have done so and is required to pay penalties (monetary or otherwise), then it'll actually hold more weight. Until that time, it really seems like an idealism transcribed into a pseudo-software license.
Re:Good idea, but glibc first would make more sens (Score:2)
Why don't all programs built for one Linux-based operating system work for another one? For myriads of reasons, including hardware version, kernel versions, operating system versions, adminstrative strategies, and user environment. I have numerous such examples, as do we all.
...as discussed on the FreeBSD mailin lists (Score:2)
Debian GNU/BSD or Debian GNU/FreeBSD seems to be the likley name of such a product.
There was alot of worry over FreeBSD branching, but overall there seemed to be a positive reaction.
A good number of people like the idea of the .deb package under FreeBSD. just as alot would like to see the FreeBSD ports tree in a Linux Distro
A quick search of the FreeBSD mailing lists on the topic reveals some [freebsd.org].
A few choice articles can bee seen below.
The seemng start of discusion: FreeLinux [freebsd.org]
A nice answer: RE: FreeLinux (Debian/GNU BSD) [freebsd.org]
Has anyone asked the FreeBSD community? (Score:5)
FBSD is a whole lot better than Linux mainly because the distribution is centeralized. Just like only good code makes it into the Linux kernel, only good code makes it into the fbsd distribution - all of it. In fbsd, there are no 5 different libraries for a single task, and you can be sure there is a single good library all programs use. You can be sure all programs would run on your distro, since there is only one distro. In linux, you have to install 43 and half different libs for the same tasks just to be sure everything runs on it.
There is also one single distribution for fbsd, so you can be sure things work on it in any case - you can safley download binaries, unlike linux where you have binaries for redhat6-glibc2 redhat5-glibc2 redhat6-libc5 redhat5-libc5 slackware-libc5 slackware-glibc2 suse6-libc5 suse6-glibc2 and so on, you get the point. In addition to that, sources might not be compatible between distros, since they use different libs and in differend locations. Simply put, you must tweak the program to get it working, unless an archive is available for your distro especially.
When I switched from linux to fbsd, it was like entering heaven. Things are organized, code is better. No more tweaking. If I want a program, it's simply a matter of typing:
cd
make && make install
and presto, a network connection is established, the sources are downloaded, patches are applied, the code is built, and then installed. Also, all the neccessary dependancies are also downloaded and built. Or even simpler, just mount the packages cd and do a
pkg_add
All the binaries, libs, etc. will be installed, everything will be in place and registered for later uninstall, and without an error. This is all because of a centeralized distribution.
Now, if Debian (whom I do respect for their linux distro) were to make a Debian GNU/FreeBSD distro, things wouldn't work this way. First of all, glibc would probably be ported so fbsd would start having seperate binaries for glibc and fbsd libc. And probably seperate bins for debian and the official distro. In addition to that, debian would start using it's own sources for linux programs, which means ports loses a lot of it's value. And
Now, a point that *must* be emphasize: freebsd is an *operating system*, not a kernel. Linux is a kernel, fbsd is an os. Read this sentance ten times: fbsd is an os, linux is a kernel. Making a freebsd debian distro would be like taking window's solitair and building a linux distro around it. fbsd would lose alot since people would start using it just for it's kernel, and not the beutiful system it is.
It's been mentioned that this way in an open unix system you would be able to choose your kernel, libc, etc. But this is a plain wrong view! Have you seen a new linux distro with just libc5 or glibc2 on it? NO! You would need to have ALL the kernels, and ALL the libcs this way. This would also make things worst for the linux community... if libcs become interoperable, linux binaries are opt to show up with the fbsd libc, which means 3 libc's for linux, which means even more mess.
One last point I want to make is that the support gnu, debian, et al get from linus in the linux kernel might not repeat itself in fbsd if debian won't make sure the fbsd community wants this step - and my guess is we don't.
Please try to consider all sides to this before taking such drastic steps. Thanks.
--
Oren Sarig
sarig@bezeqint.net
The More, The Merrier (Score:2)
As for the licensing, I well and truly believe that the lion's share of the daily users of these systems could not possibly care one wee whit less about the distinctions in the licences.
Finally, just some would prefer an operating system with an FSF-derived user environment but a BSD kernel, others would prefer the world the other way around. You read about what progress they're making toward this goal in active threads the BSD newsgroups today.
But please, please consider those threads read-only. Don't flame. Just read. In particular, don't have a coronary when you hear about how like Microsoft's dirty tricks some people find GNU's "embrace-and-extend" and anti-POSIX strategies.
Just let people have what they want to have. Sure, a BSD kernel and FSF non-kernel would end up making one more free Unix operating system than we had before, but likewise would a Linux kernel (I don't believe the FSF owns it yet, right?) combined with a BSD non-kernel.
So what? The Linuxes are so splintered and disorganized now that people would never even notice another one. Sure, you'd be upping the number of BSD-based operating systems by a far bigger jump than you'd be upping the number of Linux-based operating systems.
(4+1)/4 is a a bigger number than (122+1)/122--or whatever--is. Strangely, the complaints about the smaller number involved here dramatically outnumber those about the larger one. Isn't that peculiar? Try to resist.
TRY TO RESIST!
Let's just let everybody have what they want to have, ok? Unix is Unix. It's not Microsoft. Isn't that enough?
(BSD + Linux) ? BSD : Linux (Score:2)
Other people I've seen talk about this have used the term "FSF taint" instead. The kernel that Linux-based operating systems use is, for the most part, pretty well respected by other Unix programmers. The same cannot be universally said for the non-kernel clutter that various vendors slap together and call an operating system. So really, I'd try to avoid the term "Linux taint". I don't think it's really what you meant, and it won't get you in as much trouble. Try using the specific operating system, like "Redhat taint" or "Corel taint", or the specific political group, like "FSF taint", because that way you won't seem to be slamming the fine work that Linux Torvalds has done.
Re:The BSD kernel only, please! (Score:4)
Actually, there are other important things, too, but none of them are related to making the moronic masses happy. Sometimes they're about making adminstrators, programmers, engineers, and scientists happy; in other words, people who aren't afraid of thinking and who aren't afraid of learning, people who aren't looking so much for a flashy new toaster without an instruction manual or a redundant TV set to babysit the kiddies, but rather for a complete system that does precisely what it was designed to do, one solidly integrated and tested by time.
Other venues exist for the less technophilic--like hiring a secretary. :-)
Debian GNU-free BSD (Score:2)
Considering how GNU-free BSD already is (modulo in most cases the compiler, but that's largely irrelevant to the user's experience, and doesn't produce infectious output), this always sounds strangely redundantly, so I always have to rescan a few times until the words jumble back into place. :-)
Re:(BSD + Linux) ? BSD : Linux (Score:2)
Also, you might consider whether to eschew "Linux distributions" and employ instead more explicative and honest terms such as any of the following, in decreasing order by formality:
- Linux-based operating systems
- Linux-based OSes
- Linuxes
- Linuces
The lattermost suggestions is derived by playful hyperapplication of the rule seen in Latin crux / cruces, or in index / indices. So I see, so I see.Re:(BSD + Linux) ? BSD : Linux (Score:2)
Why the blazes don't these stupid TEXTAREA fields pop me into $EDITOR? The so-called editing capabilities are cretinous in the extreme, and, for some reason doubtless due to lazy fingers on my part, from time to time manage to delete the whole area and destroy significant work with no recourse to an undo. May the bowels of a thousand plague-infested camels burst in full projectile fury upon the Winix-minded instigators of this moronic and miserable mis-feature!
Re:(BSD + Linux) ? BSD : Linux (Score:2)
Re:Naming of proposed distribution (Score:2)
Given that, apparently, the "FreeBSD-based Debian" would use only the kernel and C library from FreeBSD, and would use their own utilities (and, presumably, other libraries) instead of the BSD ones, they wouldn't have to do any more work, over time, to "keep the GNU there" than to keep it in Debian GNU/Linux.
Re:Both?? (Score:2)
...and I use FreeBSD on my home desktop machine. I don't consider Debian (or most other Linux distributions) to inherently be desktop-only OSes (there might be a desktop-only distribution out there somewhere, for all I know), or consider any of the BSDs to inherently be server-only OSes....
Re:Naming of proposed distribution (Score:2)
True - but that's not a consequence of the BSD folk using fewer and fewer GNU utilities over time, if the Debian folk aren't using the BSD utilities, except maybe to the extent that ports of GNU utilities to FreeBSD become less actively maintained.
Re: Text widgets as inadequate text editors (OT) (Score:2)
Yes, the fork/exec interface isn't ideally suited to that.
Well, it could spawn it with CreateProcess() (IE 4.0, at least on Solaris, doesn't handle mail or news itself, but punts to whatever mail or news reader you tell it to - which, amusingly enough, makes it arguably more UNIXy than Netscape on UNIX, at least in that regard!), but that interface isn't, as far as I know, any more friendly towards that sort of pluggable text-editor functionality than is fork/exec. A COM/OLE editor interface might be better suited to that.
I have the impression that in some message about a mailer for GNOME that they were thinking of making a Bonobo interface for text-editor widgets, so that different widgets could be plugged in.
If they do so, I'm curious whether it'd be the first desktop environment/toolkit to do so; I have the impression something such as that could be done in Windows, or in KDE, but I don't know whether anybody's actually done so.
(The Andrew toolkit had, as its text widget, a fairly powerful editor, which had, I think, some amount of vi compatibility available atop its more EMACSish base, but if you're used to a particular editor, that still might not be all you'd want.)
Re:Has anyone asked the FreeBSD community? (Score:2)
First, you're trying to make a point sentimentally and not logically. Lets focus on the facts for a moment. There are two (that's 21.75 times less than 43 and a half) libc libraries for Linux, libc5 and libc6. All the Linux distributions are based on libc6 now, so for what it's worth, there's exactly *one* libc for Linux, and that is libc6. If you download a binary linked against libc6, and you have a recent distribution, as well as the other libraries the program is linked against, *it will run*. These are the facts.
Next. You say that there are many Linux distributions and it creates confusion. You completely omit the fact that there are three distinct distributions of 386/BSD, called NetBSD, OpenBSD, and FreeBSD. (Actually they're a little further splintered than Linux distributions since they are developed independently, on the source-code level.) You choose only to talk about FreeBSD and disregard the others, in which case you can also choose to only talk about e.g. Debian GNU/Linux and disregard the others. Now, everything you said about the FreeBSD packaging system is also applicable to Debian's apt system. Including automatic dependency resolving (be it libc versions or any other libraries and software). You're again trying to make use of sentimental arguments instead of facts. Please make a factual point.
While we're at it, FreeBSD is an operating system, Debian GNU/Linux is an operating system; the FreeBSD kernel is a kernel, Linux is a kernel. Repeat this sentence ten times. What's the point again?
Re:Has anyone asked the FreeBSD community? (Score:2)
Except if you use glibc 2 or 2.1, or libc6.. and there are times when upgrading from libc5 to 6 is a pain.. unless you like having the two versions on hand...
Next. You say that there are many Linux distributions and it creates confusion. You completely omit the fact that there are three distinct distributions of 386/BSD, called NetBSD, OpenBSD, and FreeBSD. (Actually they're a little further splintered than Linux distributions since they are developed independently, on the source-code level.) You choose only to talk about FreeBSD and disregard the others, in which case you can also choose to only talk about e.g. Debian GNU/Linux and disregard the others. Now, everything you said about the FreeBSD packaging system is also applicable to Debian's apt system. Including automatic dependency resolving (be it libc versions or any other libraries and software). You're again trying to make use of sentimental arguments instead of facts. Please make a factual point.
But you forget, they are completely differnet OS's.. with different kernels. don't forget bsd/os in your list. But because of the BSD4.4 compatability, it's easy to make a static binary, which works for those extreeme cases of need. As for apt-get, it's a bit backwards since debian doesn't maintain a central list of current packages that work together. All the packages in /usr/ports work together and stay current.
About his point about the kernel and the rest of the OS, it's all centralized. When the new kernels come out, the rest of the system gets updated too, so one thing doesn't break each other. Such as ipfw... there was a change where the kernel and the ipfw executable were updated at the same time.
---
Re:Debian + FreeBSD + KDE 2.0 = _the_ desktop box (Score:2)
---
Am i the only one... (Score:2)
It's messy to have to reboot to install tetrinet under leenux, reboot to fbsd and be able to run a good version of X with truetype support...
---
Re:Debian + FreeBSD + KDE 2.0 = _the_ desktop box (Score:2)
---
Re:Am i the only one... (Score:2)
---