Debian + FreeBSD 7.2 Kernel - Benchmarking Debian's GNU/kFreeBSD

Status
Not open for further replies.
killasmurf86 said:
Never say never

True, but apparently you don't know Debian. They have a plethora of projects, many people eager to talk all day long, but the have got a marginal output of real useful stuff. Apart from that, the Linux kernel is GPL, the FreeBSD kernel is BSDL. You cannot change the license from a GPL driver to BSDL, so they have to change the FreeBSD kernel to GPL. The latter is possible, but then it would be impossible to port it back. It's that easy ...
 
kpedersen said:
I believe Debian's GNU/kFreeBSD to be detrimental to the FreeBSD project because instead of bright people working solely on FreeBSD, their hard work is now going to be spread between the two projects.

This is only valid if you assume that developer resources (people) are leaving FreeBSD for Debian GNU/kFreeBSD. I haven't heard anyone doing that.

As for the result of the hard work done by FreeBSD and the other BSD's (and in fact, any project with a license as liberal as the BSD license) - anyone can take that work do make something else of it (and this happens all the time). Still, in several decades this hasn't killed any of those projects.

You can see this happening in linux and one of the most beneficial points of FreeBSD is everything is developed in one place (one tree).

I disagree - diversity is good for any open source project.
Your assumption seems to be that if FreeBSD was the only open source OS project, everybody would work on that project. It doesn't work that way - people's motivations are as diverse as people themselves.
 
oliverh said:
True, but apparently you don't know Debian. They have a plethora of projects, many people eager to talk all day long, but the have got a marginal output of real useful stuff. Apart from that, the Linux kernel is GPL, the FreeBSD kernel is BSDL. You cannot change the license from a GPL driver to BSDL, so they have to change the FreeBSD kernel to GPL. The latter is possible, but then it would be impossible to port it back. It's that easy ...

The license holder is the only one who can change the license; BSDL is not public domain. But that's not really a problem for device drivers unless they are a part of the kernel. There are some drivers in ports that are GPL, and the user agrees to that license by installing them.

So if Debian GPL drivers are ported to run with a FreeBSD kernel, they can still be used, just not included under a BSDL.
 
tingo said:
I disagree - diversity is good for any open source project.

I dis-disagree. "Diversity" is a flawed concept that hurts projects. It is often better to have one way of doing things that is 90% effective than to have 6 months of fighting over which of two or more methods that are 95% effective, while getting nothing accomplished in that time.

FreeBSD is a classic example. There is one implementation, everyone knows what to expect, and a great deal of success is seen from it.
 
wblock said:
The license holder is the only one who can change the license; BSDL is not public domain. But that's not really a problem for device drivers unless they are a part of the kernel. There are some drivers in ports that are GPL, and the user agrees to that license by installing them.

So if Debian GPL drivers are ported to run with a FreeBSD kernel, they can still be used, just not included under a BSDL.

Yes I know, I didn't mean "change of license" but any changes to the BSDL kernel under the GPL are lost vice versa. So how do you want to use a GPL licensed driver in FreeBSD (aka FreeBSD OS) without violating the GPL? It's lost, they can use it, they can change it. Yes they cannot change the BSDL, but they can do whatever they want with it while mentioning the source and leaving the BSDL text intact. It's a one-way root of development.
 
oliverh said:
Yes I know, I didn't mean "change of license" but any changes to the BSDL kernel under the GPL are lost vice versa. So how do you want to use a GPL licensed driver in FreeBSD (aka FreeBSD OS) without violating the GPL? It's lost, they can use it, they can change it. Yes they cannot change the BSDL, but they can do whatever they want with it while mentioning the source and leaving the BSDL text intact. It's a one-way root of development.

You can use ports to import stuff that can not be destributed with the system, kernel modules as well.
 
>You can use ports to import stuff that can not be destributed with the system, kernel modules as well.

Yes it's possible, you can even use the Linuxulator to run some Linux software, you can even shoot some bullet into your toe, almost anything is possible. But alas, we're talking of _Debian_ and a port? Somebody has to maintain such a port and that's a hard task to stay compatible to kernel changes. FreeBSD has got a plethora of construction sites, like Xorg, etc.pp. I don't want to blow your dreams, but e.g. Debian is known for lots of talk and less action in the last years. Cutting out a kernel and attaching a GNU userland to it is trivial compared to kernel hacking. In my opinion it's defective by design.
 
I think it's useful having such different versions of the userland. If there is a large advantage to the debian variant then that points to an area where FreeBSD can improve.
 
davidgurvich said:
I think it's useful having such different versions of the userland. If there is a large advantage to the debian variant then that points to an area where FreeBSD can improve.

I don't see any advantage in Debian over FreeBSD. I'm using Debian since about 10 years at work (server), I would be happy to get rid of it. Usually we're trying to get rid of the last reminders of GNU in BSD, because of the license and because of the lousy quality.

By the way, guess why Google's Android is using the majority of BSD-userland, mksh as shell or libc from OpenBSD/NetBSD? Because of quality.
 
Stupidity seems to run rampant with computer geeks. In the event you are wondering, being stupid means remaining voluntarily ignorant. The licensing used for the Android system has nothing to do with anyone's ideology; it's a business decision. Few, if any, corporations choose software and hardware on the basis of political ideology; it's a decision motivated by the profit margin. When the company begins to falter, it may return to using different licensing, software, and hardware; this is because they are there to save their own collective arses.
People create software projects according to their desires, not yours. Those which have the knowledge of such will choose the license they want. If there is a problem, let them deal with it.
Have you noticed that most developers will not come on to forums and only answer mailing list replies in a direct fashion? Maybe this is due to them knowing that someone will bring up politics.
For every operating system I use, the development and evolution interests me much more than licensing. It's great that the source code is available. When such isn't available, I'll hack it. It is my equipment for my use and I am not selling anything to anyone.

Now all of you silly little children can go back to sticking your tongues out at one another.
 
Most of you are asking the wrong question. It's not why FreeBSD needs Debian, it's other way around.
As you can see on their wiki, there a lot of sane reasons to prefer Debian/BSD over Debian/Linux, but reasons to prefer Debian/BSD over FreeBSD are purely subjective - if you like apt more than ports/packages, if you prefer GNU version of UNIX userland tools over BSD ones, if you want your system to be blob-free, etc.

I can clearly see why Debian folks want this project - FreeBSD kernel has distinct advantages over Linux and that's nicely listed, ZFS, devfs, overall lighter and cleaner kernel, no initrd bullshit, OSS, etc...my former company had a lot of solutions developed around Debian Linux, and there were some problems regarding the kernel, backward compatibility etc. I never managed to push something on FreeBSD because other guys were strict Linux guys; however they don't care for the kernel itself, they want to use familiar userland with familiar hierarchy. This project, when it gets to release phase, could replace Debian Linux installations.

So, Debian administrators that want superior storage mechanisms or any other advantage that FreeBSD provides over Linux on kernel level would simply switch to this distribution.

This project won't affect FreeBSD itself, it's not a competition to FreeBSD or a way to get FreeBSD with easy tools. It's simply an option for Debian users. On the other hand, we get more eyes and hands on FreeBSD kernel tree, and something could be backported in future.

To conclude, i don't see any reason to dislike this project, be it from angle of Debian/Linux or FreeBSD user. However, the name is quite muddy, and without investigating further, one doesn't know what he gets from the name itself. A lot of people think they will automatically get Linux drivers on FreeBSD. Biggest computer magazine in Croatia had a small article on their website about "Linux getting ZFS support". Of course, that confusion came from incompetent article writer that didn't investigate what Debian/kFreeBSD really was. Naturally i wrote some comments there, in really harsh tone, but it will steer out the confusion and misinformation about "Linux getting ZFS".
 
@sossego

>The licensing used for the Android system has nothing to do with anyone's ideology; it's a business decision. Few, if any, corporations choose software and hardware on the basis of political ideology

What the heck you're talking of? I'm not talking of ideology, I'm talking of technology. Yes I'm confident in FreeBSD and e.g. OpenBSD and guess what I even like Slackware. I don't like Debian because of technical merits - I don't hate it, if this is your opinion. Is this so hard to understand in a _FreeBSD_ forum? I don't have a problem to talk about other operating systems, but it's the wrong place to question the opinion of FreeBSD users about their operating system of choice!

>Stupidity seems to run rampant with computer geeks.

Just get lost and think about your manners.
 
>I can clearly see why Debian folks want this project - FreeBSD kernel has distinct advantages over Linux and that's nicely listed, ZFS, devfs, overall lighter and cleaner kernel,

@zare

This is surely a valid point of view, but another is just flexibility. They're also eager to get Hurd to run: http://www.debian.org/ports/hurd/ or long time ago they had some interest in NetBSD: http://www.debian.org/ports/netbsd/ and so on. They like to tinker with technology and there is nothing wrong with it. But they lost their focus on most projects during the years. That's my point of view and that's the reason why I don't see much in this project.
 
oliverh said:
What the heck you're talking of?
oliverh said:
By the way, guess why Google's Android is using the majority of BSD-userland, mksh as shell or libc from OpenBSD/NetBSD? Because of quality.

No. It is because of a business decision. It's because of the bottom line. Google is a corporation. The purpose of corporations is to make a profit.
As like other companies, having a closed product gives them an advantage in the marketplace.

oliverh said:
Just get lost and think about your manners.
Being in a FreeBSD forum does not excuse you from researching. If what Debian does or what your company uses bothers you, then change it. Complaining about it here won't change anything. All that your comment tells me is that you don't want to look at every reason for a businesses decision.


Criticism works both ways. Learn to take it.
 
@oliverh,

This is surely a valid point of view, but another is just flexibility. They're also eager to get Hurd to run: http://www.debian.org/ports/hurd/ or long time ago they had some interest in NetBSD: http://www.debian.org/ports/netbsd/ and so on. They like to tinker with technology and there is nothing wrong with it. But they lost their focus on most projects during the years. That's my point of view and that's the reason why I don't see much in this project.

Of course, Debian considers itself a complete operating system. Or operating environment, if i may. I'm aware of all their projects. Hurd version lost focus because Hurd lost focus. That kernel will never get out of alpha phase. Regarding NetBSD port, IMHO the biggest decision factor behind Debian/kFreeBSD is ZFS. devfs is an advantage too. Both of those are unavailable on NetBSD.

I can't predict what will happen to this project, so your opinion is fully valid, seeing that Debian's ports outside Debian/Linux are not doing good. However i'm leaving benefit of a doubt, mostly because of ZFS.

No. It is because of a business decision. It's because of the bottom line. Google is a corporation. The purpose of corporations is to make a profit.
As like other companies, having a closed product gives them an advantage in the marketplace.

I don't see how closing down BSD licensed stuff such as main C library and shell would make a significant impact on profit. After all, there's no proof that Google did substantial modifications to those programs.

Android is nothing out of the ordinary; more or less standard Linux system with dalvik virtual machine. There's no need for Google to close down open sourced components.
 
Zare said:
I don't see how closing down BSD licensed stuff such as main C library and shell would make a significant impact on profit. After all, there's no proof that Google did substantial modifications to those programs.

Android is nothing out of the ordinary; more or less standard Linux system with dalvik virtual machine. There's no need for Google to close down open sourced components.


Point taken.
 
Well, that was fun.

Anyway, we're not going to address any Debian GNU/kFreeBSD issues or problems on the FreeBSD Forums (see the Sticky), so we might as well stop bickering here about what it means to whom under which circumstances under which license or business model. There are certainly more interesting developments within FreeBSD itself to focus our efforts and energy on.

Closed.
 
Status
Not open for further replies.
Back
Top